RepoMirrors
/
mpv
mirror of https://github.com/mpv-player/mpv
1
0
Fork 0
Code Issues Releases Wiki Activity
mpv/mpvcore/m_option.c

2414 lines
66 KiB
C
Raw Normal View History

Add license header to all top-level files missing them. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@30471 b3059339-0415-0410-9bf9-f77b7e298cf2
2010-01-30 23:24:23 +00:00
/*
* This file is part of MPlayer.
*
* MPlayer is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* MPlayer is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with MPlayer; if not, write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
Doxygen attack! git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@18259 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-04-24 19:20:04 +00:00
/// \file
/// \ingroup Options
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
#include "config.h"
#include <stdlib.h>
#include <string.h>
#include <math.h>
#include <stdio.h>
#include <stdarg.h>
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
#include <limits.h>
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
#include <inttypes.h>
FreeBSD fix git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8263 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-24 01:03:16 +00:00
#include <unistd.h>
Remove internal network support This commit removes the "old" networking code in favor of libavformat's code. The code was still used for mp_http, udp, ftp, cddb. http has been mapped to libavformat's http support since approximately 6 months ago. udp and ftp have support in ffmpeg (though ftp was added only last month). cddb support is removed with this commit - it's probably not important and rarely used if at all, so we don't care about it.
2013-07-07 17:40:14 +00:00
#include <ctype.h>
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
#include <assert.h>
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
#include <libavutil/common.h>
#include <libavutil/avstring.h>
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
#include "talloc.h"
core: move contents to mpvcore (2/2) Followup commit. Fixes all the files references.
2013-08-06 20:41:30 +00:00
#include "mpvcore/mp_common.h"
#include "mpvcore/m_option.h"
#include "mpvcore/m_config.h"
#include "mpvcore/mp_msg.h"
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: commandline: print error type for parse failures If parsing an option fails, print a string corresponding to the parse function return value (M_OPT_MISSING_PARAM etc). The primary motivation is that the parsing code already outputs messages explaining most problems, but does not itself print anything in the missing parameter case. Before double-dash --options such errors were rare (or rather they resulted in the next commandline argument being silently misinterpreted as an argument to the previous option instead); but now an argument like "--ss" should give a better indication about the problem than just "Error parsing option".
2011-07-29 04:52:29 +00:00
char *m_option_strerror(int code)
{
switch (code) {
case M_OPT_UNKNOWN:
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
return mp_gtext("option not found");
options: commandline: print error type for parse failures If parsing an option fails, print a string corresponding to the parse function return value (M_OPT_MISSING_PARAM etc). The primary motivation is that the parsing code already outputs messages explaining most problems, but does not itself print anything in the missing parameter case. Before double-dash --options such errors were rare (or rather they resulted in the next commandline argument being silently misinterpreted as an argument to the previous option instead); but now an argument like "--ss" should give a better indication about the problem than just "Error parsing option".
2011-07-29 04:52:29 +00:00
case M_OPT_MISSING_PARAM:
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
return mp_gtext("option requires parameter");
options: commandline: print error type for parse failures If parsing an option fails, print a string corresponding to the parse function return value (M_OPT_MISSING_PARAM etc). The primary motivation is that the parsing code already outputs messages explaining most problems, but does not itself print anything in the missing parameter case. Before double-dash --options such errors were rare (or rather they resulted in the next commandline argument being silently misinterpreted as an argument to the previous option instead); but now an argument like "--ss" should give a better indication about the problem than just "Error parsing option".
2011-07-29 04:52:29 +00:00
case M_OPT_INVALID:
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
return mp_gtext("option parameter could not be parsed");
options: commandline: print error type for parse failures If parsing an option fails, print a string corresponding to the parse function return value (M_OPT_MISSING_PARAM etc). The primary motivation is that the parsing code already outputs messages explaining most problems, but does not itself print anything in the missing parameter case. Before double-dash --options such errors were rare (or rather they resulted in the next commandline argument being silently misinterpreted as an argument to the previous option instead); but now an argument like "--ss" should give a better indication about the problem than just "Error parsing option".
2011-07-29 04:52:29 +00:00
case M_OPT_OUT_OF_RANGE:
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
return mp_gtext("parameter is outside values allowed for option");
case M_OPT_DISALLOW_PARAM:
return mp_gtext("option doesn't take a parameter");
options: commandline: print error type for parse failures If parsing an option fails, print a string corresponding to the parse function return value (M_OPT_MISSING_PARAM etc). The primary motivation is that the parsing code already outputs messages explaining most problems, but does not itself print anything in the missing parameter case. Before double-dash --options such errors were rare (or rather they resulted in the next commandline argument being silently misinterpreted as an argument to the previous option instead); but now an argument like "--ss" should give a better indication about the problem than just "Error parsing option".
2011-07-29 04:52:29 +00:00
case M_OPT_PARSER_ERR:
default:
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
return mp_gtext("parser error");
options: commandline: print error type for parse failures If parsing an option fails, print a string corresponding to the parse function return value (M_OPT_MISSING_PARAM etc). The primary motivation is that the parsing code already outputs messages explaining most problems, but does not itself print anything in the missing parameter case. Before double-dash --options such errors were rare (or rather they resulted in the next commandline argument being silently misinterpreted as an argument to the previous option instead); but now an argument like "--ss" should give a better indication about the problem than just "Error parsing option".
2011-07-29 04:52:29 +00:00
}
}
m_option: add function to check whether parameters are required To avoid that it will be duplicated with m_option.c and m_config.c.
2013-04-21 01:18:32 +00:00
int m_option_required_params(const m_option_t *opt)
{
if (((opt->flags & M_OPT_OPTIONAL_PARAM) ||
(opt->type->flags & M_OPT_TYPE_OPTIONAL_PARAM)))
return 0;
return 1;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static const struct m_option *m_option_list_findb(const struct m_option *list,
struct bstr name)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
for (int i = 0; list[i].name; i++) {
bstr: rename bstr() function to bstr0(), and typedef bstr to struct bstr Replace all uses of bstr() with bstr0(). Also remove the ridiculous C++ workaround.
2012-07-28 21:47:42 +00:00
struct bstr lname = bstr0(list[i].name);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if ((list[i].type->flags & M_OPT_TYPE_ALLOW_WILDCARD)
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
&& bstr_endswith0(lname, "*")) {
lname.len--;
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (bstrcmp(bstr_splice(name, 0, lname.len), lname) == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return &list[i];
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
} else if (bstrcmp(lname, name) == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return &list[i];
}
return NULL;
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
const m_option_t *m_option_list_find(const m_option_t *list, const char *name)
{
bstr: rename bstr() function to bstr0(), and typedef bstr to struct bstr Replace all uses of bstr() with bstr0(). Also remove the ridiculous C++ workaround.
2012-07-28 21:47:42 +00:00
return m_option_list_findb(list, bstr0(name));
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
}
spelling and grammar fixes galore git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@10642 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-08-17 13:51:04 +00:00
// Default function that just does a memcpy
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
static void copy_opt(const m_option_t *opt, void *dst, const void *src)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
if (dst && src)
memcpy(dst, src, opt->type->size);
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
// Flag
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
#define VAL(x) (*(int *)(x))
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
static int clamp_flag(const m_option_t *opt, void *val)
{
if (VAL(val) == opt->min || VAL(val) == opt->max)
return 0;
VAL(val) = opt->min;
return M_OPT_OUT_OF_RANGE;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_flag(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
if (param.len) {
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (!bstrcmp0(param, "yes")) {
options: accept "yes" and "no" only for flags This removes the alternative values like "off", "0", "false" etc., and also the non-English versions of these. This is done for general consistency. It's better to have a single way of doing things when multiple ways don't add singificant value. Also update some choices for consistency.
2012-09-18 13:50:24 +00:00
if (dst)
VAL(dst) = opt->max;
return 1;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (!bstrcmp0(param, "no")) {
options: accept "yes" and "no" only for flags This removes the alternative values like "off", "0", "false" etc., and also the non-English versions of these. This is done for general consistency. It's better to have a single way of doing things when multiple ways don't add singificant value. Also update some choices for consistency.
2012-09-18 13:50:24 +00:00
if (dst)
VAL(dst) = opt->min;
return 1;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
}
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Invalid parameter for %.*s flag: %.*s\n",
BSTR_P(name), BSTR_P(param));
return M_OPT_INVALID;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
} else {
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (dst)
VAL(dst) = opt->max;
return 0;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static char *print_flag(const m_option_t *opt, const void *val)
{
if (VAL(val) == opt->min)
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
return talloc_strdup(NULL, "no");
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
else
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
return talloc_strdup(NULL, "yes");
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
static void add_flag(const m_option_t *opt, void *val, double add, bool wrap)
{
if (fabs(add) < 0.5)
return;
bool state = VAL(val) != opt->min;
state = wrap ? !state : add > 0;
VAL(val) = state ? opt->max : opt->min;
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_flag = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
// need yes or no in config files
.name = "Flag",
.size = sizeof(int),
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
.flags = M_OPT_TYPE_OPTIONAL_PARAM,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.parse = parse_flag,
.print = print_flag,
.copy = copy_opt,
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
.add = add_flag,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_flag,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
options: change handling of "no-" options Normally, all flag options can be negated by prepending a "no-", for example "--no-opt" becomes "--opt=no". Some flag options can't actually be negated, so add a CONF_TYPE_STORE option type to disallow the "no-" fallback. Do the same for choice options. Remove the explicit "no-" prefixed options, add "no" as choice. Move the handling of automatic "no-" options from parser-mpcmd.c to m_config.c, and use it in m_config_set_option/m_config_parse_option. This makes these options available in the config file. It also simplifies sub-option parsing, because it doesn't need to handle "no-" anymore.
2012-09-20 01:32:01 +00:00
// Single-value, write-only flag
static int parse_store(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
{
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (param.len == 0 || bstrcmp0(param, "yes") == 0) {
options: change handling of "no-" options Normally, all flag options can be negated by prepending a "no-", for example "--no-opt" becomes "--opt=no". Some flag options can't actually be negated, so add a CONF_TYPE_STORE option type to disallow the "no-" fallback. Do the same for choice options. Remove the explicit "no-" prefixed options, add "no" as choice. Move the handling of automatic "no-" options from parser-mpcmd.c to m_config.c, and use it in m_config_set_option/m_config_parse_option. This makes these options available in the config file. It also simplifies sub-option parsing, because it doesn't need to handle "no-" anymore.
2012-09-20 01:32:01 +00:00
if (dst)
VAL(dst) = opt->max;
return 0;
} else {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Invalid parameter for %.*s flag: %.*s\n",
BSTR_P(name), BSTR_P(param));
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
return M_OPT_DISALLOW_PARAM;
options: change handling of "no-" options Normally, all flag options can be negated by prepending a "no-", for example "--no-opt" becomes "--opt=no". Some flag options can't actually be negated, so add a CONF_TYPE_STORE option type to disallow the "no-" fallback. Do the same for choice options. Remove the explicit "no-" prefixed options, add "no" as choice. Move the handling of automatic "no-" options from parser-mpcmd.c to m_config.c, and use it in m_config_set_option/m_config_parse_option. This makes these options available in the config file. It also simplifies sub-option parsing, because it doesn't need to handle "no-" anymore.
2012-09-20 01:32:01 +00:00
}
}
const m_option_type_t m_option_type_store = {
// can only be activated
.name = "Flag",
.size = sizeof(int),
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
.flags = M_OPT_TYPE_OPTIONAL_PARAM,
options: change handling of "no-" options Normally, all flag options can be negated by prepending a "no-", for example "--no-opt" becomes "--opt=no". Some flag options can't actually be negated, so add a CONF_TYPE_STORE option type to disallow the "no-" fallback. Do the same for choice options. Remove the explicit "no-" prefixed options, add "no" as choice. Move the handling of automatic "no-" options from parser-mpcmd.c to m_config.c, and use it in m_config_set_option/m_config_parse_option. This makes these options available in the config file. It also simplifies sub-option parsing, because it doesn't need to handle "no-" anymore.
2012-09-20 01:32:01 +00:00
.parse = parse_store,
m_option: add missing copy callback for some option types One reaosn for this is that this broke proper handling of .defval, since m_option_copy did nothing for these option types.
2013-08-17 17:52:33 +00:00
.copy = copy_opt,
options: change handling of "no-" options Normally, all flag options can be negated by prepending a "no-", for example "--no-opt" becomes "--opt=no". Some flag options can't actually be negated, so add a CONF_TYPE_STORE option type to disallow the "no-" fallback. Do the same for choice options. Remove the explicit "no-" prefixed options, add "no" as choice. Move the handling of automatic "no-" options from parser-mpcmd.c to m_config.c, and use it in m_config_set_option/m_config_parse_option. This makes these options available in the config file. It also simplifies sub-option parsing, because it doesn't need to handle "no-" anymore.
2012-09-20 01:32:01 +00:00
};
m_options: more typesafety Change the option definition macros so that they cause compiler warnings if the type of the referenced option struct member doesn't match the type implied by the macro. The compiler warning printed isn't very telling, but it's better than silently invoking undefined behavior by violating the C strict aliasing rules. Also fix some minor cases that violate the type rules. For the option "no-aspect" we have to add a new option type to handle it properly. Some option types are hard to check, so we don't in these cases.
2013-03-01 10:21:13 +00:00
// Same for float types
#undef VAL
#define VAL(x) (*(float *)(x))
static int parse_store_float(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
{
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (param.len == 0 || bstrcmp0(param, "yes") == 0) {
m_options: more typesafety Change the option definition macros so that they cause compiler warnings if the type of the referenced option struct member doesn't match the type implied by the macro. The compiler warning printed isn't very telling, but it's better than silently invoking undefined behavior by violating the C strict aliasing rules. Also fix some minor cases that violate the type rules. For the option "no-aspect" we have to add a new option type to handle it properly. Some option types are hard to check, so we don't in these cases.
2013-03-01 10:21:13 +00:00
if (dst)
VAL(dst) = opt->max;
return 0;
} else {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Invalid parameter for %.*s flag: %.*s\n",
BSTR_P(name), BSTR_P(param));
return M_OPT_DISALLOW_PARAM;
}
}
const m_option_type_t m_option_type_float_store = {
// can only be activated
.name = "Flag",
.size = sizeof(float),
.flags = M_OPT_TYPE_OPTIONAL_PARAM,
.parse = parse_store_float,
m_option: add missing copy callback for some option types One reaosn for this is that this broke proper handling of .defval, since m_option_copy did nothing for these option types.
2013-08-17 17:52:33 +00:00
.copy = copy_opt,
m_options: more typesafety Change the option definition macros so that they cause compiler warnings if the type of the referenced option struct member doesn't match the type implied by the macro. The compiler warning printed isn't very telling, but it's better than silently invoking undefined behavior by violating the C strict aliasing rules. Also fix some minor cases that violate the type rules. For the option "no-aspect" we have to add a new option type to handle it properly. Some option types are hard to check, so we don't in these cases.
2013-03-01 10:21:13 +00:00
};
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
// Integer
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
#undef VAL
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
static int clamp_longlong(const m_option_t *opt, void *val)
{
long long v = *(long long *)val;
int r = 0;
if ((opt->flags & M_OPT_MAX) && (v > opt->max)) {
v = opt->max;
r = M_OPT_OUT_OF_RANGE;
}
if ((opt->flags & M_OPT_MIN) && (v < opt->min)) {
v = opt->min;
r = M_OPT_OUT_OF_RANGE;
}
*(long long *)val = v;
return r;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_longlong(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
struct bstr rest;
long long tmp_int = bstrtoll(param, &rest, 10);
if (rest.len)
tmp_int = bstrtoll(param, &rest, 0);
if (rest.len) {
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"The %.*s option must be an integer: %.*s\n",
BSTR_P(name), BSTR_P(param));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_INVALID;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if ((opt->flags & M_OPT_MIN) && (tmp_int < opt->min)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"The %.*s option must be >= %d: %.*s\n",
BSTR_P(name), (int) opt->min, BSTR_P(param));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_OUT_OF_RANGE;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if ((opt->flags & M_OPT_MAX) && (tmp_int > opt->max)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"The %.*s option must be <= %d: %.*s\n",
BSTR_P(name), (int) opt->max, BSTR_P(param));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_OUT_OF_RANGE;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (dst)
*(long long *)dst = tmp_int;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 1;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
static int clamp_int(const m_option_t *opt, void *val)
{
long long tmp = *(int *)val;
int r = clamp_longlong(opt, &tmp);
*(int *)val = tmp;
return r;
}
static int clamp_int64(const m_option_t *opt, void *val)
{
long long tmp = *(int64_t *)val;
int r = clamp_longlong(opt, &tmp);
*(int64_t *)val = tmp;
return r;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_int(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
{
long long tmp;
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
int r = parse_longlong(opt, name, param, &tmp);
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (r >= 0 && dst)
*(int *)dst = tmp;
return r;
}
static int parse_int64(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
{
long long tmp;
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
int r = parse_longlong(opt, name, param, &tmp);
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (r >= 0 && dst)
*(int64_t *)dst = tmp;
return r;
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static char *print_int(const m_option_t *opt, const void *val)
{
if (opt->type->size == sizeof(int64_t))
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
return talloc_asprintf(NULL, "%"PRId64, *(const int64_t *)val);
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
return talloc_asprintf(NULL, "%d", *(const int *)val);
}
static void add_int64(const m_option_t *opt, void *val, double add, bool wrap)
{
int64_t v = *(int64_t *)val;
v = v + add;
bool is64 = opt->type->size == sizeof(int64_t);
int64_t nmin = is64 ? INT64_MIN : INT_MIN;
int64_t nmax = is64 ? INT64_MAX : INT_MAX;
int64_t min = (opt->flags & M_OPT_MIN) ? opt->min : nmin;
int64_t max = (opt->flags & M_OPT_MAX) ? opt->max : nmax;
if (v < min)
v = wrap ? max : min;
if (v > max)
v = wrap ? min : max;
*(int64_t *)val = v;
}
static void add_int(const m_option_t *opt, void *val, double add, bool wrap)
{
int64_t tmp = *(int *)val;
add_int64(opt, &tmp, add, wrap);
*(int *)val = tmp;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_int = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "Integer",
.size = sizeof(int),
.parse = parse_int,
.print = print_int,
.copy = copy_opt,
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
.add = add_int,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_int,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
Add a 64 bit integer type to the suboption parser. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@28794 b3059339-0415-0410-9bf9-f77b7e298cf2
2009-03-02 11:17:50 +00:00
const m_option_type_t m_option_type_int64 = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "Integer64",
.size = sizeof(int64_t),
.parse = parse_int64,
.print = print_int,
.copy = copy_opt,
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
.add = add_int64,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_int64,
Add a 64 bit integer type to the suboption parser. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@28794 b3059339-0415-0410-9bf9-f77b7e298cf2
2009-03-02 11:17:50 +00:00
};
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_intpair(const struct m_option *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
return M_OPT_MISSING_PARAM;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
struct bstr s = param;
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
int end = -1;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
int start = bstrtoll(s, &s, 10);
if (s.len == param.len)
goto bad;
if (s.len > 0) {
if (!bstr_startswith0(s, "-"))
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
goto bad;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
s = bstr_cut(s, 1);
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (s.len > 0)
end = bstrtoll(s, &s, 10);
if (s.len > 0)
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
goto bad;
if (dst) {
int *p = dst;
p[0] = start;
p[1] = end;
}
return 1;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
bad:
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Invalid integer range "
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"specification for option %.*s: %.*s\n",
BSTR_P(name), BSTR_P(param));
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
return M_OPT_INVALID;
}
const struct m_option_type m_option_type_intpair = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "Int[-Int]",
.size = sizeof(int[2]),
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
.parse = parse_intpair,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.copy = copy_opt,
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
};
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
static int clamp_choice(const m_option_t *opt, void *val)
{
int v = *(int *)val;
if ((opt->flags & M_OPT_MIN) && (opt->flags & M_OPT_MAX)) {
if (v >= opt->min && v <= opt->max)
return 0;
}
;
for (struct m_opt_choice_alternatives *alt = opt->priv; alt->name; alt++) {
if (alt->value == v)
return 0;
}
return M_OPT_INVALID;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_choice(const struct m_option *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
{
options: allow choice options without parameter If an m_option_type_choice option is declared with M_OPT_IMPLICIT_DEFAULT in its flags, it doesn't require a parameter. For example, if --opt is such an option, it can be invoked as "--opt=val", "-opt", or "--opt". The last two will set the option to the first choice the option declares. Note that "-opt val" (using the old option syntax) is not allowed in this case, as it would be ambiguous. Normal option parsing should be unaffected.
2012-01-06 21:26:21 +00:00
struct m_opt_choice_alternatives *alt = opt->priv;
options: remove M_OPT_IMPLICIT_DEFAULT This was to make an option without value use the option's default value (e.g. --term-osd is the same as --term-osd=auto). Make it simpler by handling this case as an empty choice. The flag was probably needed when option handling still did ambiguous argument parsing.
2012-09-03 20:44:44 +00:00
for ( ; alt->name; alt++)
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (!bstrcmp0(param, alt->name))
options: remove M_OPT_IMPLICIT_DEFAULT This was to make an option without value use the option's default value (e.g. --term-osd is the same as --term-osd=auto). Make it simpler by handling this case as an empty choice. The flag was probably needed when option handling still did ambiguous argument parsing.
2012-09-03 20:44:44 +00:00
break;
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
if (!alt->name) {
options: remove M_OPT_IMPLICIT_DEFAULT This was to make an option without value use the option's default value (e.g. --term-osd is the same as --term-osd=auto). Make it simpler by handling this case as an empty choice. The flag was probably needed when option handling still did ambiguous argument parsing.
2012-09-03 20:44:44 +00:00
if (param.len == 0)
return M_OPT_MISSING_PARAM;
options: change --loop option, and extend choice option type The --loop option takes slightly different parameters now. --loop=0 used to mean looping forever. Now it means looping is disabled (this is more logical: 2 means playing 2 more times, 1 means playing 1 more time, and 0 should mean playing not again). Now --loop=inf must be used to enable looping forever. Extend choice types to allow an optional range of integers as values. If CONF_RANGE is added to the flags of a m_option_type_choice option, m_option.min/max specify a range of allowed integer values. This can be used to remove "special" values from make integer range options. These special values are unintuitive, and sometimes expose mplayer internals to the user. The (internal) choice values can be freely mixed with the specified integer value range. If there are overlaps, the choice values are preferred for conversion to/from strings. Also make sure the extension to choice options works with properties. Add the ability to step choice properties downwards, instead of just upwards.
2012-09-05 00:17:13 +00:00
if ((opt->flags & M_OPT_MIN) && (opt->flags & M_OPT_MAX)) {
long long val;
if (parse_longlong(opt, name, param, &val) == 1) {
if (dst)
*(int *)dst = val;
return 1;
}
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Invalid value for option %.*s: %.*s\n",
BSTR_P(name), BSTR_P(param));
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Valid values are:");
for (alt = opt->priv; alt->name; alt++)
mp_msg(MSGT_CFGPARSER, MSGL_ERR, " %s", alt->name);
m_option: print valid integer range for choices on parsing errors Only some choices have an additional integer range. For those which do, printing the choices only would be confusing. E.g. --cursor-autohide accepts the choices "always", "no", or an integer value. The help text printed on option parse errors should print the accepted integer range additional to "always" and "no".
2012-11-19 00:19:12 +00:00
if ((opt->flags & M_OPT_MIN) && (opt->flags & M_OPT_MAX))
mp_msg(MSGT_CFGPARSER, MSGL_ERR, " %g-%g", opt->min, opt->max);
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "\n");
return M_OPT_INVALID;
}
if (dst)
*(int *)dst = alt->value;
options: remove M_OPT_IMPLICIT_DEFAULT This was to make an option without value use the option's default value (e.g. --term-osd is the same as --term-osd=auto). Make it simpler by handling this case as an empty choice. The flag was probably needed when option handling still did ambiguous argument parsing.
2012-09-03 20:44:44 +00:00
return 1;
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static char *print_choice(const m_option_t *opt, const void *val)
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
{
int v = *(int *)val;
struct m_opt_choice_alternatives *alt;
for (alt = opt->priv; alt->name; alt++)
if (alt->value == v)
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
return talloc_strdup(NULL, alt->name);
options: change --loop option, and extend choice option type The --loop option takes slightly different parameters now. --loop=0 used to mean looping forever. Now it means looping is disabled (this is more logical: 2 means playing 2 more times, 1 means playing 1 more time, and 0 should mean playing not again). Now --loop=inf must be used to enable looping forever. Extend choice types to allow an optional range of integers as values. If CONF_RANGE is added to the flags of a m_option_type_choice option, m_option.min/max specify a range of allowed integer values. This can be used to remove "special" values from make integer range options. These special values are unintuitive, and sometimes expose mplayer internals to the user. The (internal) choice values can be freely mixed with the specified integer value range. If there are overlaps, the choice values are preferred for conversion to/from strings. Also make sure the extension to choice options works with properties. Add the ability to step choice properties downwards, instead of just upwards.
2012-09-05 00:17:13 +00:00
if ((opt->flags & M_OPT_MIN) && (opt->flags & M_OPT_MAX)) {
if (v >= opt->min && v <= opt->max)
return talloc_asprintf(NULL, "%d", v);
}
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
abort();
}
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
static void choice_get_min_max(const struct m_option *opt, int *min, int *max)
{
assert(opt->type == &m_option_type_choice);
*min = INT_MAX;
*max = INT_MIN;
for (struct m_opt_choice_alternatives *alt = opt->priv; alt->name; alt++) {
*min = FFMIN(*min, alt->value);
*max = FFMAX(*max, alt->value);
}
if ((opt->flags & M_OPT_MIN) && (opt->flags & M_OPT_MAX)) {
*min = FFMIN(*min, opt->min);
*max = FFMAX(*max, opt->max);
}
}
static void check_choice(int dir, int val, bool *found, int *best, int choice)
{
if ((dir == -1 && (!(*found) || choice > (*best)) && choice < val) ||
(dir == +1 && (!(*found) || choice < (*best)) && choice > val))
{
*found = true;
*best = choice;
}
}
static void add_choice(const m_option_t *opt, void *val, double add, bool wrap)
{
assert(opt->type == &m_option_type_choice);
int dir = add > 0 ? +1 : -1;
bool found = false;
int ival = *(int *)val;
int best = 0; // init. value unused
if (fabs(add) < 0.5)
return;
if ((opt->flags & M_OPT_MIN) && (opt->flags & M_OPT_MAX)) {
int newval = ival + add;
if (ival >= opt->min && ival <= opt->max &&
newval >= opt->min && newval <= opt->max)
{
found = true;
best = newval;
} else {
check_choice(dir, ival, &found, &best, opt->min);
check_choice(dir, ival, &found, &best, opt->max);
}
}
for (struct m_opt_choice_alternatives *alt = opt->priv; alt->name; alt++)
check_choice(dir, ival, &found, &best, alt->value);
if (!found) {
int min, max;
choice_get_min_max(opt, &min, &max);
best = (dir == -1) ^ wrap ? min : max;
}
*(int *)val = best;
}
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
const struct m_option_type m_option_type_choice = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "String", // same as arbitrary strings in option list for now
.size = sizeof(int),
options: add "choice" option type, use for -pts-association-mode Add a "choice" option type. Options of this type take a string as input and set an int option variable to the value corresponding to the string. The string->int mapping is option-specific and is given in the option definition. Strings not found in the mapping are rejected as invalid option values. Change the option -pts-association-mode to use this new option type and accept values "auto, decoder, sort" instead of "0, 1, 2". The change in accepted values shouldn't cause problems as this option is not appropriate to use in normal user config files.
2010-12-17 23:00:28 +00:00
.parse = parse_choice,
.print = print_choice,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.copy = copy_opt,
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
.add = add_choice,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_choice,
options: move -chapter values to option struct -chapter can optionally take a range with a start and an end. Add a new option type which supports such values and use that instead of a custom per-option function. This commit also fixes a build configuration bug: before the availability of the -chapter option depended on DVD functionality being enabled in the binary, even though the option works with other sources too.
2010-04-24 17:09:31 +00:00
};
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
// Float
#undef VAL
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
#define VAL(x) (*(double *)(x))
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
static int clamp_double(const m_option_t *opt, void *val)
{
double v = VAL(val);
int r = 0;
if ((opt->flags & M_OPT_MAX) && (v > opt->max)) {
v = opt->max;
r = M_OPT_OUT_OF_RANGE;
}
if ((opt->flags & M_OPT_MIN) && (v < opt->min)) {
v = opt->min;
r = M_OPT_OUT_OF_RANGE;
}
options: reject nan, inf, 0:0, etc. for float options Same for time values.
2013-02-24 20:05:59 +00:00
if (!isfinite(v)) {
v = opt->min;
r = M_OPT_OUT_OF_RANGE;
}
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
VAL(val) = v;
return r;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_double(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
struct bstr rest;
double tmp_float = bstrtod(param, &rest);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
options: remove parsing of "," as decimal point Apparently the intention was parsing numbers reliably in presence of non-C locales. mpv is always in C locale, and not being in C locale would probably break even more things, so remove this code.
2013-02-24 20:00:17 +00:00
if (bstr_eatstart0(&rest, ":") || bstr_eatstart0(&rest, "/"))
tmp_float /= bstrtod(rest, &rest);
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (rest.len) {
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"The %.*s option must be a floating point number or a "
"ratio (numerator[:/]denominator): %.*s\n",
BSTR_P(name), BSTR_P(param));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_INVALID;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (opt->flags & M_OPT_MIN)
if (tmp_float < opt->min) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"The %.*s option must be >= %f: %.*s\n",
BSTR_P(name), opt->min, BSTR_P(param));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_OUT_OF_RANGE;
}
if (opt->flags & M_OPT_MAX)
if (tmp_float > opt->max) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"The %.*s option must be <= %f: %.*s\n",
BSTR_P(name), opt->max, BSTR_P(param));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_OUT_OF_RANGE;
}
options: reject nan, inf, 0:0, etc. for float options Same for time values.
2013-02-24 20:05:59 +00:00
if (!isfinite(tmp_float)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"The %.*s option must be a finite number: %.*s\n",
BSTR_P(name), BSTR_P(param));
return M_OPT_OUT_OF_RANGE;
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (dst)
VAL(dst) = tmp_float;
return 1;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static char *print_double(const m_option_t *opt, const void *val)
{
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
return talloc_asprintf(NULL, "%f", VAL(val));
m_option_type_double git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11948 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-02-10 14:32:06 +00:00
}
m_option: pretty-print floats with 3 pre-decimal digits instead of 2 There were complaints that ${fps} was printed as e.g. "23.98" instead of "23.976". Since there's no way to format floats exactly _and_ in a user- friendly way, just change the default precision for printing floats.
2013-03-25 23:18:16 +00:00
static char *print_double_f3(const m_option_t *opt, const void *val)
commands: generally handle property formatting with m_option Use the m_option code by default to format property values, instead of having separate code in m_property. To facilitate that, add a pretty_print callback to option types. These format values in a more human readable and user friendly way, as opposed to the print callback, which produces parseable values. This also changes the strings used with flags. Instead of "enabled" and "disabled", flags are formatted as "yes" and "no". (We could use the pretty_print callback to deal with this, but we don't for consistency.)
2012-09-18 13:31:46 +00:00
{
m_option: pretty-print floats with 3 pre-decimal digits instead of 2 There were complaints that ${fps} was printed as e.g. "23.98" instead of "23.976". Since there's no way to format floats exactly _and_ in a user- friendly way, just change the default precision for printing floats.
2013-03-25 23:18:16 +00:00
return talloc_asprintf(NULL, "%.3f", VAL(val));
commands: generally handle property formatting with m_option Use the m_option code by default to format property values, instead of having separate code in m_property. To facilitate that, add a pretty_print callback to option types. These format values in a more human readable and user friendly way, as opposed to the print callback, which produces parseable values. This also changes the strings used with flags. Instead of "enabled" and "disabled", flags are formatted as "yes" and "no". (We could use the pretty_print callback to deal with this, but we don't for consistency.)
2012-09-18 13:31:46 +00:00
}
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
static void add_double(const m_option_t *opt, void *val, double add, bool wrap)
{
double v = VAL(val);
v = v + add;
double min = (opt->flags & M_OPT_MIN) ? opt->min : -INFINITY;
double max = (opt->flags & M_OPT_MAX) ? opt->max : +INFINITY;
if (v < min)
v = wrap ? max : min;
if (v > max)
v = wrap ? min : max;
VAL(val) = v;
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_double = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
// double precision float or ratio (numerator[:/]denominator)
.name = "Double",
.size = sizeof(double),
.parse = parse_double,
.print = print_double,
m_option: pretty-print floats with 3 pre-decimal digits instead of 2 There were complaints that ${fps} was printed as e.g. "23.98" instead of "23.976". Since there's no way to format floats exactly _and_ in a user- friendly way, just change the default precision for printing floats.
2013-03-25 23:18:16 +00:00
.pretty_print = print_double_f3,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.copy = copy_opt,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_double,
m_option: make "add speed 0.1" command work Was broken since the speed property was switched from float to double.
2013-08-19 20:30:44 +00:00
.add = add_double,
m_option_type_double git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11948 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-02-10 14:32:06 +00:00
};
#undef VAL
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
#define VAL(x) (*(float *)(x))
m_option_type_double git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11948 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-02-10 14:32:06 +00:00
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
static int clamp_float(const m_option_t *opt, void *val)
{
double tmp = VAL(val);
int r = clamp_double(opt, &tmp);
VAL(val) = tmp;
return r;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_float(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
m_option_type_double git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11948 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-02-10 14:32:06 +00:00
double tmp;
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
int r = parse_double(opt, name, param, &tmp);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (r == 1 && dst)
VAL(dst) = tmp;
100l (cvs commit + tired ...) git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11951 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-02-10 22:58:30 +00:00
return r;
m_option_type_double git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11948 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-02-10 14:32:06 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static char *print_float(const m_option_t *opt, const void *val)
{
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
return talloc_asprintf(NULL, "%f", VAL(val));
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
m_option: pretty-print floats with 3 pre-decimal digits instead of 2 There were complaints that ${fps} was printed as e.g. "23.98" instead of "23.976". Since there's no way to format floats exactly _and_ in a user- friendly way, just change the default precision for printing floats.
2013-03-25 23:18:16 +00:00
static char *print_float_f3(const m_option_t *opt, const void *val)
commands: generally handle property formatting with m_option Use the m_option code by default to format property values, instead of having separate code in m_property. To facilitate that, add a pretty_print callback to option types. These format values in a more human readable and user friendly way, as opposed to the print callback, which produces parseable values. This also changes the strings used with flags. Instead of "enabled" and "disabled", flags are formatted as "yes" and "no". (We could use the pretty_print callback to deal with this, but we don't for consistency.)
2012-09-18 13:31:46 +00:00
{
m_option: pretty-print floats with 3 pre-decimal digits instead of 2 There were complaints that ${fps} was printed as e.g. "23.98" instead of "23.976". Since there's no way to format floats exactly _and_ in a user- friendly way, just change the default precision for printing floats.
2013-03-25 23:18:16 +00:00
return talloc_asprintf(NULL, "%.3f", VAL(val));
commands: generally handle property formatting with m_option Use the m_option code by default to format property values, instead of having separate code in m_property. To facilitate that, add a pretty_print callback to option types. These format values in a more human readable and user friendly way, as opposed to the print callback, which produces parseable values. This also changes the strings used with flags. Instead of "enabled" and "disabled", flags are formatted as "yes" and "no". (We could use the pretty_print callback to deal with this, but we don't for consistency.)
2012-09-18 13:31:46 +00:00
}
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
static void add_float(const m_option_t *opt, void *val, double add, bool wrap)
{
double tmp = VAL(val);
add_double(opt, &tmp, add, wrap);
VAL(val) = tmp;
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_float = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
// floating point number or ratio (numerator[:/]denominator)
.name = "Float",
.size = sizeof(float),
.parse = parse_float,
.print = print_float,
m_option: pretty-print floats with 3 pre-decimal digits instead of 2 There were complaints that ${fps} was printed as e.g. "23.98" instead of "23.976". Since there's no way to format floats exactly _and_ in a user- friendly way, just change the default precision for printing floats.
2013-03-25 23:18:16 +00:00
.pretty_print = print_float_f3,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.copy = copy_opt,
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
.add = add_float,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_float,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
///////////// String
#undef VAL
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
#define VAL(x) (*(char **)(x))
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: parse C-style escapes for some options Being able to insert newline characters ("\n") is useful for --osd-status-msg, and possibly also for anything that prints to the terminal. Espcially --term-osd-esc looks relatively useless without being able to specify escapes. Maybe parsing escapes should happen during command line / config parsing instead (for all options).
2013-02-16 22:24:10 +00:00
static char *unescape_string(void *talloc_ctx, bstr str)
{
char *res = talloc_strdup(talloc_ctx, "");
while (str.len) {
bstr rest;
bool esc = bstr_split_tok(str, "\\", &str, &rest);
res = talloc_strndup_append_buffer(res, str.start, str.len);
if (esc) {
if (!mp_parse_escape(&rest, &res)) {
talloc_free(res);
return NULL;
}
}
str = rest;
}
return res;
}
static char *escape_string(char *str0)
{
char *res = talloc_strdup(NULL, "");
bstr str = bstr0(str0);
while (str.len) {
bstr rest;
bool esc = bstr_split_tok(str, "\\", &str, &rest);
res = talloc_strndup_append_buffer(res, str.start, str.len);
if (esc)
res = talloc_strdup_append_buffer(res, "\\\\");
str = rest;
}
return res;
}
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
static int clamp_str(const m_option_t *opt, void *val)
{
char *v = VAL(val);
int len = v ? strlen(v) : 0;
if ((opt->flags & M_OPT_MIN) && (len < opt->min))
return M_OPT_OUT_OF_RANGE;
if ((opt->flags & M_OPT_MAX) && (len > opt->max))
return M_OPT_OUT_OF_RANGE;
return 0;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_str(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: parse C-style escapes for some options Being able to insert newline characters ("\n") is useful for --osd-status-msg, and possibly also for anything that prints to the terminal. Espcially --term-osd-esc looks relatively useless without being able to specify escapes. Maybe parsing escapes should happen during command line / config parsing instead (for all options).
2013-02-16 22:24:10 +00:00
int r = 1;
void *tmp = talloc_new(NULL);
if (param.start == NULL) {
r = M_OPT_MISSING_PARAM;
goto exit;
}
m_option: fix -vo opengl lscale validation OPT_STRING_VALIDATE actually did nothing. This made -vo opengl crash or misbehave when passing an invalid value for the lscale, cscale or 3dlut- size (the only users for this option type). The code added with this commit was either blatantly forgotten with the commit introducing this option type, or somehow lost.
2013-05-31 19:22:06 +00:00
m_opt_string_validate_fn validate = opt->priv;
if (validate) {
r = validate(opt, name, param);
if (r < 0)
goto exit;
}
options: parse C-style escapes for some options Being able to insert newline characters ("\n") is useful for --osd-status-msg, and possibly also for anything that prints to the terminal. Espcially --term-osd-esc looks relatively useless without being able to specify escapes. Maybe parsing escapes should happen during command line / config parsing instead (for all options).
2013-02-16 22:24:10 +00:00
if (opt->flags & M_OPT_PARSE_ESCAPES) {
char *res = unescape_string(tmp, param);
if (!res) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Parameter has broken escapes: %.*s\n", BSTR_P(param));
r = M_OPT_INVALID;
goto exit;
}
param = bstr0(res);
}
options: fix specifying string options without parameter Specifying a string option with no parameter, as in "--dumpfile" with no '=', erroneously set the corresponding variable to NULL. Fix this to give an error about missing parameter instead. Suboption parsing explicitly treated empty option values as if the option had been specified with no value (no '='). Thus it was not possible to specify empty strings as values. I think this behavior was originally added only because of other limitations in the old implementation. Remove it, so that suboptions now behave the same as top-level ones in this regard. Document the NULL-distinguishing property of bstrdup0() that the code depends on, and also make bstrdup() behave consistently.
2012-06-29 02:14:26 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if ((opt->flags & M_OPT_MIN) && (param.len < opt->min)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Parameter must be >= %d chars: %.*s\n",
(int) opt->min, BSTR_P(param));
options: parse C-style escapes for some options Being able to insert newline characters ("\n") is useful for --osd-status-msg, and possibly also for anything that prints to the terminal. Espcially --term-osd-esc looks relatively useless without being able to specify escapes. Maybe parsing escapes should happen during command line / config parsing instead (for all options).
2013-02-16 22:24:10 +00:00
r = M_OPT_OUT_OF_RANGE;
goto exit;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if ((opt->flags & M_OPT_MAX) && (param.len > opt->max)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Parameter must be <= %d chars: %.*s\n",
(int) opt->max, BSTR_P(param));
options: parse C-style escapes for some options Being able to insert newline characters ("\n") is useful for --osd-status-msg, and possibly also for anything that prints to the terminal. Espcially --term-osd-esc looks relatively useless without being able to specify escapes. Maybe parsing escapes should happen during command line / config parsing instead (for all options).
2013-02-16 22:24:10 +00:00
r = M_OPT_OUT_OF_RANGE;
goto exit;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (dst) {
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(VAL(dst));
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
VAL(dst) = bstrdup0(NULL, param);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: parse C-style escapes for some options Being able to insert newline characters ("\n") is useful for --osd-status-msg, and possibly also for anything that prints to the terminal. Espcially --term-osd-esc looks relatively useless without being able to specify escapes. Maybe parsing escapes should happen during command line / config parsing instead (for all options).
2013-02-16 22:24:10 +00:00
exit:
talloc_free(tmp);
return r;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static char *print_str(const m_option_t *opt, const void *val)
{
options: parse C-style escapes for some options Being able to insert newline characters ("\n") is useful for --osd-status-msg, and possibly also for anything that prints to the terminal. Espcially --term-osd-esc looks relatively useless without being able to specify escapes. Maybe parsing escapes should happen during command line / config parsing instead (for all options).
2013-02-16 22:24:10 +00:00
bool need_escape = opt->flags & M_OPT_PARSE_ESCAPES;
char *s = val ? VAL(val) : NULL;
return s ? (need_escape ? escape_string(s) : talloc_strdup(NULL, s)) : NULL;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
static void copy_str(const m_option_t *opt, void *dst, const void *src)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
if (dst && src) {
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(VAL(dst));
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
VAL(dst) = talloc_strdup(NULL, VAL(src));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static void free_str(void *src)
{
if (src && VAL(src)) {
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(VAL(src));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
VAL(src) = NULL;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_string = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "String",
.size = sizeof(char *),
.flags = M_OPT_TYPE_DYNAMIC,
.parse = parse_str,
.print = print_str,
.copy = copy_str,
.free = free_str,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_str,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
//////////// String list
#undef VAL
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
#define VAL(x) (*(char ***)(x))
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
#define OP_NONE 0
#define OP_ADD 1
#define OP_PRE 2
#define OP_DEL 3
#define OP_CLR 4
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
#define OP_TOGGLE 5
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static void free_str_list(void *dst)
{
char **d;
int i;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (!dst || !VAL(dst))
return;
d = VAL(dst);
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
for (i = 0; d[i] != NULL; i++)
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(d[i]);
talloc_free(d);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
VAL(dst) = NULL;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
static int str_list_add(char **add, int n, void *dst, int pre)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
if (!dst)
return M_OPT_PARSER_ERR;
m_option: minor simplification Relict of mplayer's C89 compatibility?
2012-10-31 23:50:08 +00:00
char **lst = VAL(dst);
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
m_option: minor simplification Relict of mplayer's C89 compatibility?
2012-10-31 23:50:08 +00:00
int ln;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
for (ln = 0; lst && lst[ln]; ln++)
/**/;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
lst = talloc_realloc(NULL, lst, char *, n + ln + 1);
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (pre) {
memmove(&lst[n], lst, ln * sizeof(char *));
memcpy(lst, add, n * sizeof(char *));
} else
memcpy(&lst[ln], add, n * sizeof(char *));
// (re-)add NULL-termination
lst[ln + n] = NULL;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(add);
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
VAL(dst) = lst;
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 1;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
static int str_list_del(char **del, int n, void *dst)
{
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
char **lst, *ep;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
int i, ln, s;
long idx;
if (!dst)
return M_OPT_PARSER_ERR;
lst = VAL(dst);
for (ln = 0; lst && lst[ln]; ln++)
/**/;
s = ln;
for (i = 0; del[i] != NULL; i++) {
idx = strtol(del[i], &ep, 0);
if (*ep) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Invalid index: %s\n", del[i]);
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(del[i]);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
continue;
}
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(del[i]);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (idx < 0 || idx >= ln) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Index %ld is out of range.\n", idx);
continue;
} else if (!lst[idx])
continue;
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(lst[idx]);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
lst[idx] = NULL;
s--;
}
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(del);
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (s == 0) {
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(lst);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
VAL(dst) = NULL;
return 1;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
// Don't bother shrinking the list allocation
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
for (i = 0, n = 0; i < ln; i++) {
if (!lst[i])
continue;
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
lst[n] = lst[i];
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
n++;
}
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
lst[s] = NULL;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 1;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static struct bstr get_nextsep(struct bstr *ptr, char sep, bool modify)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
struct bstr str = *ptr;
struct bstr orig = str;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
for (;;) {
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
int idx = bstrchr(str, sep);
if (idx > 0 && str.start[idx - 1] == '\\') {
if (modify) {
memmove(str.start + idx - 1, str.start + idx, str.len - idx);
str.len--;
str = bstr_cut(str, idx);
} else
str = bstr_cut(str, idx + 1);
} else {
str = bstr_cut(str, idx < 0 ? str.len : idx);
str_list separator backslash-escaping git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22151 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-05 18:03:12 +00:00
break;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
}
str_list separator backslash-escaping git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22151 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-05 18:03:12 +00:00
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
*ptr = str;
return bstr_splice(orig, 0, str.start - orig.start);
str_list separator backslash-escaping git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22151 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-05 18:03:12 +00:00
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_str_list(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
char **res;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
int op = OP_NONE;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
int len = strlen(opt->name);
if (opt->name[len - 1] == '*' && (name.len > len - 1)) {
struct bstr suffix = bstr_cut(name, len - 1);
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (bstrcmp0(suffix, "-add") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_ADD;
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
else if (bstrcmp0(suffix, "-pre") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_PRE;
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
else if (bstrcmp0(suffix, "-del") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_DEL;
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
else if (bstrcmp0(suffix, "-clr") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_CLR;
else
return M_OPT_UNKNOWN;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
// Clear the list ??
if (op == OP_CLR) {
if (dst)
free_str_list(dst);
return 0;
}
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
// All other ops need a param
m_option: allow setting empty lists There's no reason why this should be forbidden.
2013-06-07 15:05:07 +00:00
if (param.len == 0 && op != OP_NONE)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
// custom type for "profile" calls this but uses ->priv for something else
char separator = opt->type == &m_option_type_string_list && opt->priv ?
*(char *)opt->priv : OPTION_LIST_SEPARATOR;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
int n = 0;
struct bstr str = param;
while (str.len) {
get_nextsep(&str, separator, 0);
str = bstr_cut(str, 1);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
n++;
}
m_option: allow setting empty lists There's no reason why this should be forbidden.
2013-06-07 15:05:07 +00:00
if (n == 0 && op != OP_NONE)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_INVALID;
if (((opt->flags & M_OPT_MIN) && (n < opt->min)) ||
((opt->flags & M_OPT_MAX) && (n > opt->max)))
return M_OPT_OUT_OF_RANGE;
if (!dst)
return 1;
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
res = talloc_array(NULL, char *, n + 2);
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
str = bstrdup(NULL, param);
char *ptr = str.start;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
n = 0;
options: fix failure to parse trailing ',' in string list A trailing separator in string list options was ignored after recent commit e873d703e9 ("options: change option parsing to use bstr"), which broke uses such as "-vo vdpau,". Fix.
2011-07-31 17:25:22 +00:00
while (1) {
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
struct bstr el = get_nextsep(&str, separator, 1);
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
res[n] = bstrdup0(NULL, el);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
n++;
options: fix failure to parse trailing ',' in string list A trailing separator in string list options was ignored after recent commit e873d703e9 ("options: change option parsing to use bstr"), which broke uses such as "-vo vdpau,". Fix.
2011-07-31 17:25:22 +00:00
if (!str.len)
break;
str = bstr_cut(str, 1);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
res[n] = NULL;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
talloc_free(ptr);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
switch (op) {
case OP_ADD:
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
return str_list_add(res, n, dst, 0);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
case OP_PRE:
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
return str_list_add(res, n, dst, 1);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
case OP_DEL:
return str_list_del(res, n, dst);
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (VAL(dst))
free_str_list(dst);
VAL(dst) = res;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
m_option: allow setting empty lists There's no reason why this should be forbidden.
2013-06-07 15:05:07 +00:00
if (!res[0])
free_str_list(dst);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 1;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
static void copy_str_list(const m_option_t *opt, void *dst, const void *src)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
int n;
char **d, **s;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (!(dst && src))
return;
s = VAL(src);
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (VAL(dst))
free_str_list(dst);
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (!s) {
VAL(dst) = NULL;
return;
}
Fix the Gui with NEW_CONFIG git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8169 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 12:39:05 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
for (n = 0; s[n] != NULL; n++)
/* NOTHING */;
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
d = talloc_array(NULL, char *, n + 1);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
for (; n >= 0; n--)
options: revert passing around talloc contexts This reverts commit 48f0692ab9 "options: make option struct the talloc parent of options". This made things actually more complicated. It introduced a new parameter to the option parse and copy functions, which was used inconsistently. Some code passed a parent, some not. Morever, you have to call m_option_free() anyway, because not all options actually respect the talloc parent. There is also the question whether passing NULL as parent is supposed to work, or if you still have to implement m_config_free(). On the other hand, this simplifies nothing. I assume the intention was being able to free all option values with a single talloc_free() call, but the same goal can be reached by simply freeing the m_config struct. (The m_config talloc destructor will free each option values.) Get rid of the talloc parent context parameter. This essentially reverts commit 48f0692ab9 ("options: make option struct the talloc parent of options"). In video_out.c, make the VO priv struct the talloc parent for the m_config object, so that destroying the VO will free the options. The ability to free the m_config struct and all its managed options was introduced in commit 89a17bcda6c.
2012-08-04 09:14:38 +00:00
d[n] = talloc_strdup(NULL, s[n]);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
VAL(dst) = d;
}
static char *print_str_list(const m_option_t *opt, const void *src)
{
char **lst = NULL;
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
char *ret = NULL;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (!(src && VAL(src)))
return NULL;
lst = VAL(src);
commands: change property mechanism to use talloc strings
2011-07-03 17:04:21 +00:00
for (int i = 0; lst[i]; i++) {
if (ret)
ret = talloc_strdup_append_buffer(ret, ",");
ret = talloc_strdup_append_buffer(ret, lst[i]);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
return ret;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_string_list = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
/* A list of strings separated by ','.
* Option with a name ending in '*' permits using the following suffixes:
* -add: Add the given parameters at the end of the list.
* -pre: Add the given parameters at the beginning of the list.
* -del: Remove the entry at the given indices.
* -clr: Clear the list.
* e.g: -vf-add flip,mirror -vf-del 2,5
*/
.name = "String list",
.size = sizeof(char **),
.flags = M_OPT_TYPE_DYNAMIC | M_OPT_TYPE_ALLOW_WILDCARD,
.parse = parse_str_list,
.print = print_str_list,
.copy = copy_str_list,
.free = free_str_list,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
/////////////////// Print
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_print(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
if (opt->type == CONF_TYPE_PRINT) {
mp_msg(MSGT_CFGPARSER, MSGL_INFO, "%s", mp_gtext(opt->p));
} else {
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
char *name0 = bstrdup0(NULL, name);
char *param0 = bstrdup0(NULL, param);
int r = ((m_opt_func_full_t) opt->p)(opt, name0, param0);
talloc_free(name0);
talloc_free(param0);
return r;
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (opt->priv == NULL)
return M_OPT_EXIT;
return 0;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_print = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "Print",
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
.flags = M_OPT_TYPE_OPTIONAL_PARAM,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.parse = parse_print,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
const m_option_type_t m_option_type_print_func_param = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "Print",
.flags = M_OPT_TYPE_ALLOW_WILDCARD,
.parse = parse_print,
New option type to print help text with a function. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@17471 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-01-24 11:14:13 +00:00
};
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
const m_option_type_t m_option_type_print_func = {
.name = "Print",
options: unify single dash and double dash options There were two option syntax variations: "old": -opt value "new": --opt=value "-opt=value" was invalid, and "--opt value" meant "--opt=" followed by a separate option "value" (i.e. interpreted as filename). There isn't really any reason to do this. The "old" syntax used to be ambiguous (you had to call the option parser to know whether the following argument is an option value or a new option), but that has been removed. Further, using "=" in the option string is always unambiguous. Since the distinction between the two option variants is confusing, just remove the difference and allow "--opt value" and "-opt=value". To make this easier, do some other cleanups as well (e.g. avoid having to do a manual lookup of the option just to check for M_OPT_PRE_PARSE, which somehow ended up with finally getting rid of the m_config.mode member). Error reporting is still a mess, and we opt for reporting too many rather than too few errors to the user. There shouldn't be many user-visible changes. The --framedrop and --term-osd options now always require parameters. The --mute option is intentionally made ambiguous: it works like a flag option, but a value can be passed to it explicitly ("--mute=auto"). If the interpretation of the option is ambiguous (like "--mute auto"), the second string is interpreted as separate option or filename. (Normal flag options are actually ambiguous in this way too.)
2012-09-21 07:22:25 +00:00
.flags = M_OPT_TYPE_ALLOW_WILDCARD | M_OPT_TYPE_OPTIONAL_PARAM,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
.parse = parse_print,
};
pphelp git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8737 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-01-03 12:40:10 +00:00
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
/////////////////////// Subconfig
#undef VAL
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
#define VAL(x) (*(char ***)(x))
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
// Read s sub-option name, or a positional sub-opt value.
// Return 0 on succes, M_OPT_ error code otherwise.
// optname is for error reporting.
static int read_subparam(bstr optname, bstr *str, bstr *out_subparam)
{
bstr p = *str;
bstr subparam = {0};
if (bstr_eatstart0(&p, "\"")) {
int optlen = bstrcspn(p, "\"");
subparam = bstr_splice(p, 0, optlen);
p = bstr_cut(p, optlen);
if (!bstr_startswith0(p, "\"")) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Terminating '\"' missing for '%.*s'\n",
BSTR_P(optname));
return M_OPT_INVALID;
}
p = bstr_cut(p, 1);
options: allow using [ ] for quoting in sub-options This is an attempt to make quoting of sub-option values less awkward, even if it works only with some shells. This is needed mainly for vf_lavfi. Also update the vf_lavfi manpage section.
2013-04-26 14:52:54 +00:00
} else if (bstr_eatstart0(&p, "[")) {
if (!bstr_split_tok(p, "]", &subparam, &p)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Terminating ']' missing for '%.*s'\n",
BSTR_P(optname));
return M_OPT_INVALID;
}
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
} else if (bstr_eatstart0(&p, "%")) {
int optlen = bstrtoll(p, &p, 0);
if (!bstr_startswith0(p, "%") || (optlen > p.len - 1)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Invalid length %d for '%.*s'\n",
optlen, BSTR_P(optname));
return M_OPT_INVALID;
}
subparam = bstr_splice(p, 1, optlen + 1);
p = bstr_cut(p, optlen + 1);
} else {
// Skip until the next character that could possibly be a meta
// character in option parsing.
options: allow using [ ] for quoting in sub-options This is an attempt to make quoting of sub-option values less awkward, even if it works only with some shells. This is needed mainly for vf_lavfi. Also update the vf_lavfi manpage section.
2013-04-26 14:52:54 +00:00
int optlen = bstrcspn(p, ":=,\\%\"'[]");
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
subparam = bstr_splice(p, 0, optlen);
p = bstr_cut(p, optlen);
}
*str = p;
*out_subparam = subparam;
return 0;
}
m_option: split out sub-config parsing
2013-04-20 22:45:22 +00:00
// Return 0 on success, otherwise error code
// On success, set *out_name and *out_val, and advance *str
// out_val.start is NULL if there was no parameter.
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
// optname is for error reporting.
static int split_subconf(bstr optname, bstr *str, bstr *out_name, bstr *out_val)
m_option: split out sub-config parsing
2013-04-20 22:45:22 +00:00
{
bstr p = *str;
bstr subparam = {0};
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
bstr subopt;
int r = read_subparam(optname, &p, &subopt);
if (r < 0)
return r;
m_option: split out sub-config parsing
2013-04-20 22:45:22 +00:00
if (bstr_eatstart0(&p, "=")) {
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
r = read_subparam(subopt, &p, &subparam);
if (r < 0)
return r;
m_option: split out sub-config parsing
2013-04-20 22:45:22 +00:00
}
*str = p;
*out_name = subopt;
*out_val = subparam;
return 0;
}
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_subconf(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: support parsing values into substructs Add an alternate mode for option parser objects (struct m_config) which is not inherently tied to any particular instance of an option value struct. Instead, this type or parsers can be used to initialize defaults in or parse values into a struct given as a parameter. They do not have the save slot functionality used for main player configuration. The new functionality will be used to replace the separate subopt_helper.c parsing code that is currently used to parse per-object suboptions in VOs etc. Previously, option default values were handled by initializing them in external code before creating a parser. This initialization was done with constants even for dynamically-allocated types like strings. Because trying to free a pointer to a constant would cause a crash when trying to replace the default with another value, parser initialization code then replaced all the original defaults with dynamically-allocated copies. This replace-with-copy behavior is no longer supported for new-style options; instead the option definition itself may contain a default value (new OPTDEF macros), and the new function m_config_initialize() is used to set all options to their default values. Convert the existing initialized dynamically allocated options in main config (the string options --dumpfile, --term-osd-esc, --input=conf) to use this. Other non-dynamic ones could be later converted to use this style of initialization too. There's currently no public call to free all dynamically allocated options in a given option struct because I intend to use talloc functionality for that (make them children of the struct and free with it).
2012-05-17 00:31:11 +00:00
int nr = 0;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
char **lst = NULL;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
struct bstr p = param;
while (p.len) {
m_option: split out sub-config parsing
2013-04-20 22:45:22 +00:00
bstr subopt, subparam;
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
int r = split_subconf(name, &p, &subopt, &subparam);
m_option: split out sub-config parsing
2013-04-20 22:45:22 +00:00
if (r < 0)
return r;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (bstr_startswith0(p, ":"))
p = bstr_cut(p, 1);
else if (p.len > 0) {
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
"Incorrect termination for '%.*s'\n", BSTR_P(subopt));
Add support for suboption escaping via both "" and %n%str syntax git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@16742 b3059339-0415-0410-9bf9-f77b7e298cf2
2005-10-12 16:35:37 +00:00
return M_OPT_INVALID;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (dst) {
lst = talloc_realloc(NULL, lst, char *, 2 * (nr + 2));
m_option: split out sub-config parsing
2013-04-20 22:45:22 +00:00
lst[2 * nr] = bstrto0(lst, subopt);
lst[2 * nr + 1] = bstrto0(lst, subparam);
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
memset(&lst[2 * (nr + 1)], 0, 2 * sizeof(char *));
nr++;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
}
if (dst)
VAL(dst) = lst;
return 1;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_subconfig = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
// The syntax is -option opt1=foo:flag:opt2=blah
.name = "Subconfig",
.flags = M_OPT_TYPE_HAS_CHILD,
.parse = parse_subconf,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
m_config: support auto-allocated sub-structs Given your option struct has a field that is a pointer to another struct, this commit allows you to declare options that write into that other struct. The code in m_config will dereference the pointer field on its own if such an option is accessed. If the field is NULL on initialization of the containing m_config, the struct is automatically allocated. OPT_SUBSTRUCT() can be used to declare such a field. struct m_sub_options is used to describe the pointed-to struct, and includes size and defaults if the struct has to be allocated by m_config.
2012-08-06 15:45:17 +00:00
const m_option_type_t m_option_type_subconfig_struct = {
.name = "Subconfig",
.flags = M_OPT_TYPE_HAS_CHILD | M_OPT_TYPE_USE_SUBSTRUCT,
.parse = parse_subconf,
};
m_option: add color option type This accepts HTML-style hex colors in the form #RRGGBB. It's also possible to provide an alpha component with #AARRGGBB. Each 2-digit group is a hex number, which gives the color value from 0-255 (e.g. There is existing code in subassconvert.c, which parses HTML-style color values in SRT subs. This is not used: it's probably better if option parsing is completely separate from code specific to certain subtitle formats, even if a little code is duplicated.
2012-11-17 18:58:57 +00:00
static int parse_color(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
{
if (param.len == 0)
return M_OPT_MISSING_PARAM;
bstr val = param;
struct m_color color = {0};
if (bstr_eatstart0(&val, "#")) {
// #[AA]RRGGBB
if (val.len != 6 && val.len != 8)
goto error;
bool has_alpha = val.len == 8;
uint32_t c = bstrtoll(val, &val, 16);
if (val.len)
goto error;
color = (struct m_color) {
(c >> 16) & 0xFF,
(c >> 8) & 0xFF,
c & 0xFF,
has_alpha ? (c >> 24) & 0xFF : 0xFF,
};
} else {
goto error;
}
if (dst)
*((struct m_color *)dst) = color;
return 1;
error:
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Option %.*s: invalid color: '%.*s'\n",
BSTR_P(name), BSTR_P(param));
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Valid colors must be in the form #RRRGGBB or #AARRGGBB (in hex)\n");
return M_OPT_INVALID;
}
const m_option_type_t m_option_type_color = {
.name = "Color",
.size = sizeof(struct m_color),
.parse = parse_color,
m_option: add missing copy callback for some option types One reaosn for this is that this broke proper handling of .defval, since m_option_copy did nothing for these option types.
2013-08-17 17:52:33 +00:00
.copy = copy_opt,
m_option: add color option type This accepts HTML-style hex colors in the form #RRGGBB. It's also possible to provide an alpha component with #AARRGGBB. Each 2-digit group is a hex number, which gives the color value from 0-255 (e.g. There is existing code in subassconvert.c, which parses HTML-style color values in SRT subs. This is not used: it's probably better if option parsing is completely separate from code specific to certain subtitle formats, even if a little code is duplicated.
2012-11-17 18:58:57 +00:00
};
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
// Parse a >=0 number starting at s. Set s to the string following the number.
// If the number ends with '%', eat that and set *out_per to true, but only
// if the number is between 0-100; if not, don't eat anything, even the number.
static bool eat_num_per(bstr *s, int *out_num, bool *out_per)
{
bstr rest;
long long v = bstrtoll(*s, &rest, 10);
if (s->len == rest.len || v < INT_MIN || v > INT_MAX)
return false;
*out_num = v;
*out_per = false;
*s = rest;
if (bstr_eatstart0(&rest, "%") && v >= 0 && v <= 100) {
*out_per = true;
*s = rest;
}
return true;
}
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
static bool parse_geometry_str(struct m_geometry *gm, bstr s)
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
{
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
*gm = (struct m_geometry) { .x = INT_MIN, .y = INT_MIN };
if (s.len == 0)
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
return true;
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
// Approximate grammar:
// [W[xH]][{+-}X{+-}Y] | [X:Y]
// (meaning: [optional] {one character of} one|alternative)
// Every number can be followed by '%'
int num;
bool per;
#define READ_NUM(F, F_PER) do { \
if (!eat_num_per(&s, &num, &per)) \
goto error; \
gm->F = num; \
gm->F_PER = per; \
} while(0)
#define READ_SIGN(F) do { \
if (bstr_eatstart0(&s, "+")) { \
gm->F = false; \
} else if (bstr_eatstart0(&s, "-")) {\
gm->F = true; \
} else goto error; \
} while(0)
if (bstrchr(s, ':') < 0) {
gm->wh_valid = true;
if (!bstr_startswith0(s, "+") && !bstr_startswith0(s, "-")) {
READ_NUM(w, w_per);
if (bstr_eatstart0(&s, "x"))
READ_NUM(h, h_per);
}
if (s.len > 0) {
gm->xy_valid = true;
READ_SIGN(x_sign);
READ_NUM(x, x_per);
READ_SIGN(y_sign);
READ_NUM(y, y_per);
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
}
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
} else {
gm->xy_valid = true;
READ_NUM(x, x_per);
if (!bstr_eatstart0(&s, ":"))
goto error;
READ_NUM(y, y_per);
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
}
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
return s.len == 0;
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
error:
return false;
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
}
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
#undef READ_NUM
#undef READ_SIGN
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
// xpos,ypos: position of the left upper corner
// widw,widh: width and height of the window
// scrw,scrh: width and height of the current screen
// The input parameters should be set to a centered window (default fallbacks).
void m_geometry_apply(int *xpos, int *ypos, int *widw, int *widh,
int scrw, int scrh, struct m_geometry *gm)
{
if (gm->wh_valid) {
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
int prew = *widw, preh = *widh;
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
if (gm->w > 0)
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
*widw = gm->w_per ? scrw * (gm->w / 100.0) : gm->w;
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
if (gm->h > 0)
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
*widh = gm->h_per ? scrh * (gm->h / 100.0) : gm->h;
// keep aspect if the other value is not set
double asp = (double)prew / preh;
if (gm->w > 0 && !(gm->h > 0)) {
*widh = *widw / asp;
} else if (!(gm->w > 0) && gm->h > 0) {
*widw = *widh * asp;
}
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
}
if (gm->xy_valid) {
if (gm->x != INT_MIN) {
*xpos = gm->x;
if (gm->x_per)
*xpos = (scrw - *widw) * (*xpos / 100.0);
if (gm->x_sign)
*xpos = scrw - *widw - *xpos;
}
if (gm->y != INT_MIN) {
*ypos = gm->y;
if (gm->y_per)
*ypos = (scrh - *widh) * (*ypos / 100.0);
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
if (gm->y_sign)
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
*ypos = scrh - *widh - *ypos;
}
}
}
static int parse_geometry(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
{
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
struct m_geometry gm;
if (!parse_geometry_str(&gm, param))
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
goto error;
if (dst)
*((struct m_geometry *)dst) = gm;
return 1;
error:
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Option %.*s: invalid geometry: '%.*s'\n",
BSTR_P(name), BSTR_P(param));
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: allow using % for width and height in --geometry Now all numbers in the --geometry specification can take percentages. Rewrite the parsing of --geometry, because adjusting the sscanf() mess would require adding all the combinations of using and not using %. As a side effect, using % and pixel values can be freely mixed. Keep the aspect if only one of width or height is set. This is more useful in general. Note: there is one semantic change: --geometry=num used to mean setting the window X position, but now it means setting the window width. Apparently this was a mplayer-specific feature (not part of standard X geometry specifications), and it doesn't look like an overly useful feature, so we are fine with breaking it. In general, the new parsing should still adhere to standard X geometry specification (as used by XParseGeometry()).
2013-01-23 09:56:27 +00:00
"Valid format: [W[%%][xH[%%]]][{+-}X[%%]{+-}Y[%%]] | [X[%%]:Y[%%]]\n");
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
return M_OPT_INVALID;
}
const m_option_type_t m_option_type_geometry = {
.name = "Window geometry",
.size = sizeof(struct m_geometry),
.parse = parse_geometry,
m_option: add missing copy callback for some option types One reaosn for this is that this broke proper handling of .defval, since m_option_copy did nothing for these option types.
2013-08-17 17:52:33 +00:00
.copy = copy_opt,
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
};
video: add --autofit and --autofit-larger options --autofit=WxH sets the window size to a maximum width and/or height, without changing the window's aspect ratio. --autofit-larger=WxH does the same, but only if the video size is actually larger than the window size that would result when using the --autofit=WxH option with the same arguments.
2013-01-23 09:56:36 +00:00
static int parse_size_box(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
{
struct m_geometry gm;
if (!parse_geometry_str(&gm, param))
goto error;
if (gm.xy_valid)
goto error;
if (dst)
*((struct m_geometry *)dst) = gm;
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
video: add --autofit and --autofit-larger options --autofit=WxH sets the window size to a maximum width and/or height, without changing the window's aspect ratio. --autofit-larger=WxH does the same, but only if the video size is actually larger than the window size that would result when using the --autofit=WxH option with the same arguments.
2013-01-23 09:56:36 +00:00
return 1;
error:
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Option %.*s: invalid size: '%.*s'\n",
BSTR_P(name), BSTR_P(param));
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Valid format: W[%%][xH[%%]] or empty string\n");
return M_OPT_INVALID;
}
const m_option_type_t m_option_type_size_box = {
.name = "Window size",
.size = sizeof(struct m_geometry),
.parse = parse_size_box,
m_option: add missing copy callback for some option types One reaosn for this is that this broke proper handling of .defval, since m_option_copy did nothing for these option types.
2013-08-17 17:52:33 +00:00
.copy = copy_opt,
video: add --autofit and --autofit-larger options --autofit=WxH sets the window size to a maximum width and/or height, without changing the window's aspect ratio. --autofit-larger=WxH does the same, but only if the video size is actually larger than the window size that would result when using the --autofit=WxH option with the same arguments.
2013-01-23 09:56:36 +00:00
};
options: move -geometry parsing to m_option.c This also means the option is verified on program start, not when the VO is created. The actual code becomes a bit more complex, because the screen width/height is not available at program start. The actual parsing code is still the same, with its unusual sscanf() usage.
2013-01-23 09:56:11 +00:00
Rename directories, move files (step 2 of 2) Finish renaming directories and moving files. Adjust all include statements to make the previous commit compile. The two commits are separate, because git is bad at tracking renames and content changes at the same time. Also take this as an opportunity to remove the separation between "common" and "mplayer" sources in the Makefile. ("common" used to be shared between mplayer and mencoder.)
2012-11-09 00:06:43 +00:00
#include "video/img_format.h"
Improve the imgfmt option type git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9601 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 20:49:39 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_imgfmt(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (!bstrcmp0(param, "help")) {
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_INFO, "Available formats:");
options, codecs.conf, img_format: unify imgfmt name handling Remove the duplication of image format name lists from codec-cfg.c and img_format.c. Remove the list of "long" image format names from img_format.c. One user visible change is that now mplayer won't print "long" format names anymore: e.g. instead of "Planar 420P 10-bit little-endian", the name "420p10le" is used. This is consistent with the names used by the option parser, and also less noisy. Partially based on mplayer2 commit f98e47574de15, with some differences.
2012-08-28 21:58:48 +00:00
for (int i = 0; mp_imgfmt_list[i].name; i++)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_INFO, " %s", mp_imgfmt_list[i].name);
mp_msg(MSGT_CFGPARSER, MSGL_INFO, "\n");
return M_OPT_EXIT - 1;
}
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
options, codecs.conf, img_format: unify imgfmt name handling Remove the duplication of image format name lists from codec-cfg.c and img_format.c. Remove the list of "long" image format names from img_format.c. One user visible change is that now mplayer won't print "long" format names anymore: e.g. instead of "Planar 420P 10-bit little-endian", the name "420p10le" is used. This is consistent with the names used by the option parser, and also less noisy. Partially based on mplayer2 commit f98e47574de15, with some differences.
2012-08-28 21:58:48 +00:00
unsigned int fmt = mp_imgfmt_from_name(param, false);
if (!fmt) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Option %.*s: unknown format name: '%.*s'\n",
BSTR_P(name), BSTR_P(param));
return M_OPT_INVALID;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
if (dst)
*((uint32_t *)dst) = fmt;
return 1;
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_imgfmt = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
// Please report any missing colorspaces
.name = "Image format",
.size = sizeof(uint32_t),
.parse = parse_imgfmt,
.copy = copy_opt,
New config system + cleanup of header inter dependency git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8165 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-11-12 01:56:42 +00:00
};
video: decouple internal pixel formats from FourCCs mplayer's video chain traditionally used FourCCs for pixel formats. For example, it used IMGFMT_YV12 for 4:2:0 YUV, which was defined to the string 'YV12' interpreted as unsigned int. Additionally, it used to encode information into the numeric values of some formats. The RGB formats had their bit depth and endian encoded into the least significant byte. Extended planar formats (420P10 etc.) had chroma shift, endian, and component bit depth encoded. (This has been removed in recent commits.) Replace the FourCC mess with a simple enum. Remove all the redundant formats like YV12/I420/IYUV. Replace some image format names by something more intuitive, most importantly IMGFMT_YV12 -> IMGFMT_420P. Add img_fourcc.h, which contains the old IDs for code that actually uses FourCCs. Change the way demuxers, that output raw video, identify the video format: they set either MP_FOURCC_RAWVIDEO or MP_FOURCC_IMGFMT to request the rawvideo decoder, and sh_video->imgfmt specifies the pixel format. Like the previous hack, this is supposed to avoid the need for a complete codecs.cfg entry per format, or other lookup tables. (Note that the RGB raw video FourCCs mostly rely on ffmpeg's mappings for NUT raw video, but this is still considered better than adding a raw video decoder - even if trivial, it would be full of annoying lookup tables.) The TV code has not been tested. Some corrective changes regarding endian and other image format flags creep in.
2012-12-23 19:03:30 +00:00
static int parse_fourcc(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
{
if (param.len == 0)
return M_OPT_MISSING_PARAM;
unsigned int value;
if (param.len == 4) {
uint8_t *s = param.start;
value = s[0] | (s[1] << 8) | (s[2] << 16) | (s[3] << 24);
} else {
bstr rest;
value = bstrtoll(param, &rest, 16);
if (rest.len != 0) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Option %.*s: invalid FourCC: '%.*s'\n",
BSTR_P(name), BSTR_P(param));
return M_OPT_INVALID;
}
}
if (dst)
*((unsigned int *)dst) = value;
return 1;
}
const m_option_type_t m_option_type_fourcc = {
.name = "FourCC",
.size = sizeof(unsigned int),
.parse = parse_fourcc,
.copy = copy_opt,
};
Rename directories, move files (step 2 of 2) Finish renaming directories and moving files. Adjust all include statements to make the previous commit compile. The two commits are separate, because git is bad at tracking renames and content changes at the same time. Also take this as an opportunity to remove the separation between "common" and "mplayer" sources in the Makefile. ("common" used to be shared between mplayer and mencoder.)
2012-11-09 00:06:43 +00:00
#include "audio/format.h"
CONF_TYPE_AFMT similar to CONF_TYPE_IMGFMT git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14247 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-12-27 17:32:54 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_afmt(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (!bstrcmp0(param, "help")) {
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_INFO, "Available formats:");
options, libaf: unify audio format name handling Remove the duplicated list of audio format names from m_option.c. Remove "long" audio format names, and let af_fmt2str() print the usual short format names. The long names were overly verbose, and were actually rarely user visible. The only difference between af_fmt2str() and af_fmt2str_short() is that the former prints unknown format values as hexadecimal string as fallback.
2012-08-28 22:34:21 +00:00
for (int i = 0; af_fmtstr_table[i].name; i++)
mp_msg(MSGT_CFGPARSER, MSGL_INFO, " %s", af_fmtstr_table[i].name);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_INFO, "\n");
return M_OPT_EXIT - 1;
}
options, libaf: unify audio format name handling Remove the duplicated list of audio format names from m_option.c. Remove "long" audio format names, and let af_fmt2str() print the usual short format names. The long names were overly verbose, and were actually rarely user visible. The only difference between af_fmt2str() and af_fmt2str_short() is that the former prints unknown format values as hexadecimal string as fallback.
2012-08-28 22:34:21 +00:00
int fmt = af_str2fmt_short(param);
audio: make internal audio format 0 an invalid format Having to use -1 for that is generally quite annoying. Audio formats are created from bitmasks, and it can't be excluded that 0 is not a valid format. Fix this by adjusting AF_FORMAT_I so that it is never 0. Along with AF_FORMAT_F and the special formats, all valid formats are covered and guaranteed to be non-0. It's possible that this commit will cause some regressions, as the check for invalid audio formats changes a bit.
2013-08-25 16:23:40 +00:00
if (!fmt) {
options, libaf: unify audio format name handling Remove the duplicated list of audio format names from m_option.c. Remove "long" audio format names, and let af_fmt2str() print the usual short format names. The long names were overly verbose, and were actually rarely user visible. The only difference between af_fmt2str() and af_fmt2str_short() is that the former prints unknown format values as hexadecimal string as fallback.
2012-08-28 22:34:21 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Option %.*s: unknown format name: '%.*s'\n",
BSTR_P(name), BSTR_P(param));
return M_OPT_INVALID;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
CONF_TYPE_AFMT similar to CONF_TYPE_IMGFMT git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14247 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-12-27 17:32:54 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (dst)
*((uint32_t *)dst) = fmt;
CONF_TYPE_AFMT similar to CONF_TYPE_IMGFMT git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14247 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-12-27 17:32:54 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 1;
CONF_TYPE_AFMT similar to CONF_TYPE_IMGFMT git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14247 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-12-27 17:32:54 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_afmt = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
// Please report any missing formats
.name = "Audio format",
.size = sizeof(uint32_t),
.parse = parse_afmt,
.copy = copy_opt,
CONF_TYPE_AFMT similar to CONF_TYPE_IMGFMT git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14247 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-12-27 17:32:54 +00:00
};
audio: add channel map API Unused, will be used in the following commits. Let chmap.h define the number of maximum channels, because that is most convenient.
2013-04-05 13:25:09 +00:00
#include "audio/chmap.h"
static int parse_chmap(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
{
// min>0: at least min channels, min=0: empty ok, min=-1: invalid ok
int min_ch = (opt->flags & M_OPT_MIN) ? opt->min : 1;
if (bstr_equals0(param, "help")) {
mp_chmap_print_help(MSGT_CFGPARSER, MSGL_INFO);
return M_OPT_EXIT - 1;
}
if (param.len == 0 && min_ch >= 1)
return M_OPT_MISSING_PARAM;
struct mp_chmap res = {0};
if (!mp_chmap_from_str(&res, param)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Error parsing channel layout: %.*s\n", BSTR_P(param));
return M_OPT_INVALID;
}
if ((min_ch > 0 && !mp_chmap_is_valid(&res)) ||
(min_ch >= 0 && mp_chmap_is_empty(&res)))
{
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Invalid channel layout: %.*s\n", BSTR_P(param));
return M_OPT_INVALID;
}
m_option: fix segfault in parse_chmap
2013-05-14 12:45:06 +00:00
if (dst)
*(struct mp_chmap *)dst = res;
audio: add channel map API Unused, will be used in the following commits. Let chmap.h define the number of maximum channels, because that is most convenient.
2013-04-05 13:25:09 +00:00
return 1;
}
const m_option_type_t m_option_type_chmap = {
.name = "Audio channels or channel map",
.size = sizeof(struct mp_chmap *),
.parse = parse_chmap,
.copy = copy_opt,
};
CONF_TYPE_AFMT similar to CONF_TYPE_IMGFMT git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@14247 b3059339-0415-0410-9bf9-f77b7e298cf2
2004-12-27 17:32:54 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_timestring(struct bstr str, double *time, char endchar)
Move parsing of the -ss option to the option code. Also fixes a memory leak of the parameter string. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22330 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-24 07:11:57 +00:00
{
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
int a, b, len;
double d;
*time = 0; /* ensure initialization for error cases */
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (bstr_sscanf(str, "%d:%d:%lf%n", &a, &b, &d, &len) >= 3)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
*time = 3600 * a + 60 * b + d;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
else if (bstr_sscanf(str, "%d:%lf%n", &a, &d, &len) >= 2)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
*time = 60 * a + d;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
else if (bstr_sscanf(str, "%lf%n", &d, &len) >= 1)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
*time = d;
else
return 0; /* unsupported time format */
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (len < str.len && str.start[len] != endchar)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 0; /* invalid extra characters at the end */
options: reject nan, inf, 0:0, etc. for float options Same for time values.
2013-02-24 20:05:59 +00:00
if (!isfinite(*time))
return 0;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return len;
Move parsing of the -ss option to the option code. Also fixes a memory leak of the parameter string. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22330 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-24 07:11:57 +00:00
}
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
Move parsing of the -ss option to the option code. Also fixes a memory leak of the parameter string. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22330 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-24 07:11:57 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_time(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
Move parsing of the -ss option to the option code. Also fixes a memory leak of the parameter string. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22330 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-24 07:11:57 +00:00
{
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
double time;
remove C99-ism (mixed declaration and code), 10l to Uoti Urpala for breaking GCC-2.95 support yet again even though any post-2.95 GCC version print a clear warning about it Patch by Carl Eugen Hoyos % cehoyos A ag P or P at % git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22374 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-28 09:18:42 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (!parse_timestring(param, &time, 0)) {
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Option %.*s: invalid time: '%.*s'\n",
BSTR_P(name), BSTR_P(param));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_INVALID;
}
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (dst)
*(double *)dst = time;
return 1;
Move parsing of the -ss option to the option code. Also fixes a memory leak of the parameter string. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22330 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-24 07:11:57 +00:00
}
commands: generally handle property formatting with m_option Use the m_option code by default to format property values, instead of having separate code in m_property. To facilitate that, add a pretty_print callback to option types. These format values in a more human readable and user friendly way, as opposed to the print callback, which produces parseable values. This also changes the strings used with flags. Instead of "enabled" and "disabled", flags are formatted as "yes" and "no". (We could use the pretty_print callback to deal with this, but we don't for consistency.)
2012-09-18 13:31:46 +00:00
static char *pretty_print_time(const m_option_t *opt, const void *val)
{
return mp_format_time(*(double *)val, false);
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_time = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "Time",
.size = sizeof(double),
.parse = parse_time,
.print = print_double,
commands: generally handle property formatting with m_option Use the m_option code by default to format property values, instead of having separate code in m_property. To facilitate that, add a pretty_print callback to option types. These format values in a more human readable and user friendly way, as opposed to the print callback, which produces parseable values. This also changes the strings used with flags. Instead of "enabled" and "disabled", flags are formatted as "yes" and "no". (We could use the pretty_print callback to deal with this, but we don't for consistency.)
2012-09-18 13:31:46 +00:00
.pretty_print = pretty_print_time,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.copy = copy_opt,
commands: handle property stepping in a generic way Instead of forcing each property implementation implement its own logic for M_PROPERTY_STEP_UP/M_PROPERTY_STEP_DOWN, handle it in the generic property code. Rename the M_PROPERTY_STEP_UP command to M_PROPERTY_SWITCH (the other property command, M_PROPERTY_STEP_DOWN, isn't needed anymore: stepping downwards is done by passing a negative argument). Always use double as argument type; it makes the code easier, and covers all property types. Move the code which does the actual type-specific value stepping to m_option.c (the idea is that m_option handles types). Some properties still have custom handlers implemented with M_PROPERTY_SWITCH. They can't be mapped to the generic mechanism, because their value range is dynamic or entirely unknown. For some properties, the default step stride is changed to 1. This is no issue, because the default bindings in input.conf all use an explicit stride in the affected cases.
2012-09-18 12:00:08 +00:00
.add = add_double,
commands: handle property clamping in m_option Instead of clamping property values to the valid range in each property implementation, handle it in the property layer. The functionality to handle clamping for each type is in m_option.c. It's not really clear whether this is really needed. Normally, the raw values for M_PROPERTY_SET come only from m_option_type.parse (setting properties as string) or from m_option_parse.add (using the "switch" input command). However, since this was already done before, and since we _really_ want to be sure only to write valid values, add this code anyway. The newly added warnings/error messages should never actually be printed during normal operation and are for debugging (if they happen, we definitely want to see them).
2012-09-18 15:05:11 +00:00
.clamp = clamp_double,
Move parsing of the -ss option to the option code. Also fixes a memory leak of the parameter string. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@22330 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-02-24 07:11:57 +00:00
};
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
// Relative time
after a long time, finally i could add -endpos option to mplayer executable. as oded told me on 1006 02 24, i applied it, blame me if some problem occurs, i hope not, since i tried it for a while.... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@19979 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-09-25 16:47:56 +00:00
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
static int parse_rel_time(const m_option_t *opt, struct bstr name,
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
struct m_rel_time t = {0};
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (param.len == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_MISSING_PARAM;
after a long time, finally i could add -endpos option to mplayer executable. as oded told me on 1006 02 24, i applied it, blame me if some problem occurs, i hope not, since i tried it for a while.... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@19979 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-09-25 16:47:56 +00:00
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
// Percent pos
m_option: fix braindead --start, --length, --end option parsing bugs The option type m_option_type_rel_time was completely broken. It interpreted everything starting with a number as percent position. This is because sscanf() semantics are idiotic (trailing string doesn't need to be matched), and due to my own idiocy this was overlooked when testing. Fix by considering sscanf() evil and not using it. (bstr_sscanf() is a straight wrapper around sscanf()). Even if the percent code was fixed, there was another bug: it always interpreted times as negative (starting from end for --start). Fix the basic logic.
2012-11-17 23:19:43 +00:00
if (bstr_endswith0(param, "%")) {
double percent = bstrtod(bstr_splice(param, 0, -1), &param);
if (param.len == 0 && percent >= 0 && percent <= 100) {
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
t.type = REL_TIME_PERCENT;
t.pos = percent;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
goto out;
}
after a long time, finally i could add -endpos option to mplayer executable. as oded told me on 1006 02 24, i applied it, blame me if some problem occurs, i hope not, since i tried it for a while.... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@19979 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-09-25 16:47:56 +00:00
}
options: support chapters for --start and --end The --start and --end switch now accept a chapter number. The chapter number is prefixed with '#', e.g. "--start=#2" jumps to chapter 2. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise. (This was supposed to be added in 51503a, but apparently the fixup commit adding it was lost in a rebase. This might also be the reason for the mess-up fixed in 394285.)
2012-11-18 17:02:14 +00:00
// Chapter pos
if (bstr_startswith0(param, "#")) {
int chapter = bstrtoll(bstr_cut(param, 1), &param, 10);
if (param.len == 0 && chapter >= 1) {
t.type = REL_TIME_CHAPTER;
t.pos = chapter - 1;
goto out;
}
}
m_option: fix braindead --start, --length, --end option parsing bugs The option type m_option_type_rel_time was completely broken. It interpreted everything starting with a number as percent position. This is because sscanf() semantics are idiotic (trailing string doesn't need to be matched), and due to my own idiocy this was overlooked when testing. Fix by considering sscanf() evil and not using it. (bstr_sscanf() is a straight wrapper around sscanf()). Even if the percent code was fixed, there was another bug: it always interpreted times as negative (starting from end for --start). Fix the basic logic.
2012-11-17 23:19:43 +00:00
bool sign = bstr_eatstart0(&param, "-");
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
double time;
if (parse_timestring(param, &time, 0)) {
t.type = sign ? REL_TIME_NEGATIVE : REL_TIME_ABSOLUTE;
t.pos = time;
goto out;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
options: support chapters for --start and --end The --start and --end switch now accept a chapter number. The chapter number is prefixed with '#', e.g. "--start=#2" jumps to chapter 2. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise. (This was supposed to be added in 51503a, but apparently the fixup commit adding it was lost in a rebase. This might also be the reason for the mess-up fixed in 394285.)
2012-11-18 17:02:14 +00:00
"Option %.*s: invalid time or position: '%.*s'\n",
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
BSTR_P(name), BSTR_P(param));
return M_OPT_INVALID;
Fix broken parse_time_size, it would cause MPlayer to parse its parameter twice, e.g. "mplayer file -endpos 01" would try to play the file "01", too git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@20009 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-09-29 23:23:41 +00:00
out:
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (dst)
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
*(struct m_rel_time *)dst = t;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 1;
after a long time, finally i could add -endpos option to mplayer executable. as oded told me on 1006 02 24, i applied it, blame me if some problem occurs, i hope not, since i tried it for a while.... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@19979 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-09-25 16:47:56 +00:00
}
options: rename -ss and -endpos, allow relative times Rename the -ss option to -start, and -endpos to -length. Add a -end option. The -end option always specifies an absolute end time, as opposed to -endpos/-length. All these options (--start, --end, --length) now accept relative times. Percent positions (e.g. "--start=30%") are interpreted as fractions of the file duration. Negative times (e.g. "--start=-1:00) are interpreted relative to the end of the file. Chapters (e.g. "--start=#3") yield the chapter's time position. The chapter support might be able to replace --chapter completely, but for now I am not sure how well this works out with e.g. DVDs and BDs, and a separate --chapter option is useful interface-wise.
2012-11-15 17:49:17 +00:00
const m_option_type_t m_option_type_rel_time = {
.name = "Relative time or percent position",
.size = sizeof(struct m_rel_time),
.parse = parse_rel_time,
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.copy = copy_opt,
after a long time, finally i could add -endpos option to mplayer executable. as oded told me on 1006 02 24, i applied it, blame me if some problem occurs, i hope not, since i tried it for a while.... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@19979 b3059339-0415-0410-9bf9-f77b7e298cf2
2006-09-25 16:47:56 +00:00
};
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
spelling and grammar fixes galore git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@10642 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-08-17 13:51:04 +00:00
//// Objects (i.e. filters, etc) settings
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
#undef VAL
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
#define VAL(x) (*(m_obj_settings_t **)(x))
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
bool m_obj_list_find(struct m_obj_desc *dst, const struct m_obj_list *l,
bstr name)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
for (int i = 0; ; i++) {
if (!l->get_desc(dst, i))
break;
if (bstr_equals0(name, dst->name))
return true;
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
}
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
if (l->aliases) {
for (int i = 0; l->aliases[i][0]; i++) {
const char *aname = l->aliases[i][0];
const char *alias = l->aliases[i][1];
const char *opts = l->aliases[i][2];
if (bstr_equals0(name, aname) &&
m_obj_list_find(dst, l, bstr0(alias)))
{
if (opts) {
dst->init_options = opts;
} else {
// Assume it's deprecated in this case.
// Also, it's used by the VO code only, so whatever.
mp_msg(MSGT_CFGPARSER, MSGL_WARN,
"VO driver '%s' has been replaced with '%s'!\n",
aname, alias);
}
return true;
}
}
}
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
return false;
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
}
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
static void obj_setting_free(m_obj_settings_t *item)
{
talloc_free(item->name);
talloc_free(item->label);
free_str_list(&(item->attribs));
}
// If at least one item has a label, compare labels only - otherwise ignore them.
static bool obj_setting_equals(m_obj_settings_t *a, m_obj_settings_t *b)
{
bstr la = bstr0(a->label), lb = bstr0(b->label);
if (la.len || lb.len)
return bstr_equals(la, lb);
if (strcmp(a->name, b->name) != 0)
return false;
int a_attr_count = 0;
while (a->attribs && a->attribs[a_attr_count])
a_attr_count++;
int b_attr_count = 0;
while (b->attribs && b->attribs[b_attr_count])
b_attr_count++;
if (a_attr_count != b_attr_count)
return false;
for (int n = 0; n < a_attr_count; n++) {
if (strcmp(a->attribs[n], b->attribs[n]) != 0)
return false;
}
return true;
}
static int obj_settings_list_num_items(m_obj_settings_t *obj_list)
{
int num = 0;
while (obj_list && obj_list[num].name)
num++;
return num;
}
static void obj_settings_list_del_at(m_obj_settings_t **p_obj_list, int idx)
{
m_obj_settings_t *obj_list = *p_obj_list;
int num = obj_settings_list_num_items(obj_list);
assert(idx >= 0 && idx < num);
obj_setting_free(&obj_list[idx]);
// Note: the NULL-terminating element is moved down as part of this
memmove(&obj_list[idx], &obj_list[idx + 1],
sizeof(m_obj_settings_t) * (num - idx));
*p_obj_list = talloc_realloc(NULL, obj_list, struct m_obj_settings, num);
}
// Insert such that *p_obj_list[idx] is set to item.
// If idx < 0, set idx = count + idx + 1 (i.e. -1 inserts it as last element).
// Memory referenced by *item is not copied.
static void obj_settings_list_insert_at(m_obj_settings_t **p_obj_list, int idx,
m_obj_settings_t *item)
{
int num = obj_settings_list_num_items(*p_obj_list);
if (idx < 0)
idx = num + idx + 1;
assert(idx >= 0 && idx <= num);
*p_obj_list = talloc_realloc(NULL, *p_obj_list, struct m_obj_settings,
num + 2);
memmove(*p_obj_list + idx + 1, *p_obj_list + idx,
(num - idx) * sizeof(m_obj_settings_t));
(*p_obj_list)[idx] = *item;
(*p_obj_list)[num + 1] = (m_obj_settings_t){0};
}
static int obj_settings_list_find_by_label(m_obj_settings_t *obj_list,
bstr label)
{
for (int n = 0; obj_list && obj_list[n].name; n++) {
if (label.len && bstr_equals0(label, obj_list[n].label))
return n;
}
return -1;
}
static int obj_settings_list_find_by_label0(m_obj_settings_t *obj_list,
const char *label)
{
return obj_settings_list_find_by_label(obj_list, bstr0(label));
}
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
static int obj_settings_find_by_content(m_obj_settings_t *obj_list,
m_obj_settings_t *item)
{
for (int n = 0; obj_list && obj_list[n].name; n++) {
if (obj_setting_equals(&obj_list[n], item))
return n;
}
return -1;
}
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
static void free_obj_settings_list(void *dst)
{
int n;
m_obj_settings_t *d;
if (!dst || !VAL(dst))
return;
d = VAL(dst);
for (n = 0; d[n].name; n++)
obj_setting_free(&d[n]);
talloc_free(d);
VAL(dst) = NULL;
}
static void copy_obj_settings_list(const m_option_t *opt, void *dst,
const void *src)
{
m_obj_settings_t *d, *s;
int n;
if (!(dst && src))
return;
s = VAL(src);
if (VAL(dst))
free_obj_settings_list(dst);
if (!s)
return;
for (n = 0; s[n].name; n++)
/* NOP */;
d = talloc_array(NULL, struct m_obj_settings, n + 1);
for (n = 0; s[n].name; n++) {
d[n].name = talloc_strdup(NULL, s[n].name);
d[n].label = talloc_strdup(NULL, s[n].label);
d[n].attribs = NULL;
copy_str_list(NULL, &(d[n].attribs), &(s[n].attribs));
}
d[n].name = NULL;
d[n].label = NULL;
d[n].attribs = NULL;
VAL(dst) = d;
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
// Consider -vf a=b=c:d=e. This verifies "b"="c" and "d"="e" and that the
// option names/values are correct. Try to determine whether an option
// without '=' sets a flag, or whether it's a positional argument.
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
static int get_obj_param(bstr opt_name, bstr obj_name, struct m_config *config,
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
bstr name, bstr val, int flags, int *nold,
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
bstr *out_name, bstr *out_val)
{
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
int r;
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
if (!config)
return 0; // skip
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
// va.start != NULL => of the form name=val (not positional)
// If it's just "name", and the associated option exists and is a flag,
// don't accept it as positional argument.
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
if (val.start || m_config_option_requires_param(config, name) == 0) {
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
r = m_config_set_option_ext(config, name, val, flags);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (r < 0) {
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
if (r == M_OPT_UNKNOWN) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Option %.*s: %.*s doesn't have a %.*s parameter.\n",
BSTR_P(opt_name), BSTR_P(obj_name), BSTR_P(name));
return M_OPT_UNKNOWN;
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (r > M_OPT_EXIT)
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Option %.*s: "
"Error while parsing %.*s parameter %.*s (%.*s)\n",
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
BSTR_P(opt_name), BSTR_P(obj_name), BSTR_P(name),
BSTR_P(val));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return r;
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
*out_name = name;
*out_val = val;
return 1;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
} else {
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
val = name;
// positional fields
if (val.len == 0) { // Empty field, count it and go on
(*nold)++;
m_option: fix positional sub-option skipping Empty sub-option parameters mean the sub-option should be skipped, e.g. -vf gradfun=:10 sets the second option (by position) to 10. This was broken in commit 04f1e2d.
2013-04-23 12:21:08 +00:00
return 0;
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
}
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
const char *opt = m_config_get_positional_option(config, *nold);
if (!opt) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Option %.*s: %.*s has only %d "
"params, so you can't give more than %d unnamed params.\n",
BSTR_P(opt_name), BSTR_P(obj_name), *nold, *nold);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_OUT_OF_RANGE;
}
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
r = m_config_set_option_ext(config, bstr0(opt), val, flags);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (r < 0) {
if (r > M_OPT_EXIT)
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Option %.*s: "
"Error while parsing %.*s parameter %s (%.*s)\n",
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
BSTR_P(opt_name), BSTR_P(obj_name), opt, BSTR_P(val));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return r;
}
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
*out_name = bstr0(opt);
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
*out_val = val;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
(*nold)++;
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
return 1;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
// Consider -vf a=b:c:d. This parses "b:c:d" into name/value pairs, stored as
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
// linear array in *_ret. In particular, config contains what options a the
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
// object takes, and verifies the option values as well.
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
// If config is NULL, all parameters are accepted without checking.
// _ret set to NULL can be used for checking-only.
// flags can contain any M_SETOPT_* flag.
int m_obj_parse_sub_config(struct bstr opt_name, struct bstr name,
struct bstr *pstr, struct m_config *config,
int flags, char ***ret)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
int nold = 0;
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
char **args = NULL;
int num_args = 0;
int r = 1;
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
if (ret) {
args = *ret;
while (args && args[num_args])
num_args++;
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
while (pstr->len > 0) {
bstr fname, fval;
m_option: allow quoted positional parameters for -vf This allows things like: '--vf=lavfi="gradfun=20:30"' Adjust the documentation for vf_lavfi to make the example less verbose. As an unrelated change, add a general description to vf_lavfi.
2013-04-23 12:10:43 +00:00
r = split_subconf(opt_name, pstr, &fname, &fval);
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
if (r < 0)
goto exit;
if (bstr_equals0(fname, "help"))
goto print_help;
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
r = get_obj_param(opt_name, name, config, fname, fval, flags, &nold,
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
&fname, &fval);
if (r < 0)
goto exit;
m_option: fix positional sub-option skipping Empty sub-option parameters mean the sub-option should be skipped, e.g. -vf gradfun=:10 sets the second option (by position) to 10. This was broken in commit 04f1e2d.
2013-04-23 12:21:08 +00:00
if (r > 0 && ret) {
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
MP_TARRAY_APPEND(NULL, args, num_args, bstrto0(NULL, fname));
MP_TARRAY_APPEND(NULL, args, num_args, bstrto0(NULL, fval));
m_option: fix memory leak on error path
2013-07-21 18:26:56 +00:00
MP_TARRAY_APPEND(NULL, args, num_args, NULL);
MP_TARRAY_APPEND(NULL, args, num_args, NULL);
num_args -= 2;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
if (!bstr_eatstart0(pstr, ":"))
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
break;
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
if (ret) {
if (num_args > 0) {
*ret = args;
args = NULL;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
} else {
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
*ret = NULL;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
}
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
goto exit;
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
print_help: ;
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
if (config) {
m_config_print_option_list(config);
} else {
mp_msg(MSGT_CFGPARSER, MSGL_WARN, "Option %.*s doesn't exist.\n",
BSTR_P(opt_name));
}
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
r = M_OPT_EXIT - 1;
exit:
m_option: fix memory leak on error path
2013-07-21 18:26:56 +00:00
free_str_list(&args);
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
return r;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
// Characters which may appear in a filter name
#define NAMECH "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_-"
// Parse one item, e.g. -vf a=b:c:d,e=f:g => parse a=b:c:d into "a" and "b:c:d"
static int parse_obj_settings(struct bstr opt, struct bstr *pstr,
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
const struct m_obj_list *list,
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
m_obj_settings_t **_ret)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
int r;
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
char **plist = NULL;
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
struct m_obj_desc desc;
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
bstr label = {0};
if (bstr_eatstart0(pstr, "@")) {
if (!bstr_split_tok(*pstr, ":", &label, pstr)) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Option %.*s: ':' expected after label.\n", BSTR_P(opt));
return M_OPT_INVALID;
}
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
bool has_param = false;
int idx = bstrspn(*pstr, NAMECH);
bstr str = bstr_splice(*pstr, 0, idx);
*pstr = bstr_cut(*pstr, idx);
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
// video filters use "=", VOs use ":"
if (bstr_eatstart0(pstr, "=") || bstr_eatstart0(pstr, ":"))
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
has_param = true;
Remove trailing whitespace from most files
2009-07-06 23:26:13 +00:00
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
bool legacy = false;
m_option: fix skipping of missing -vo/-ao entries The -vo/-ao options support skipping of unknown entries for the sake of allowing using the same config on possibly very different systems, which have different sets of VO/AOs available. Unfortunately skipping didn't quite work, possibly a rebase error when this was originally committed. Fix it.
2013-07-31 18:44:56 +00:00
bool skip = false;
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
if (m_obj_list_find(&desc, list, str)) {
legacy = !desc.priv_size && list->legacy_hacks;
} else {
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
if (!list->allow_unknown_entries) {
mp_msg(MSGT_CFGPARSER, MSGL_ERR, "Option %.*s: %.*s doesn't exist.\n",
BSTR_P(opt), BSTR_P(str));
return M_OPT_INVALID;
}
desc = (struct m_obj_desc){0};
m_option: fix skipping of missing -vo/-ao entries The -vo/-ao options support skipping of unknown entries for the sake of allowing using the same config on possibly very different systems, which have different sets of VO/AOs available. Unfortunately skipping didn't quite work, possibly a rebase error when this was originally committed. Fix it.
2013-07-31 18:44:56 +00:00
skip = true;
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
if (has_param) {
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
if (legacy) {
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
// Should perhaps be parsed as escape-able string. But this is a
// compatibility path, so it's not worth the trouble.
int next = bstrcspn(*pstr, ",");
bstr param = bstr_splice(*pstr, 0, next);
*pstr = bstr_cut(*pstr, next);
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (!bstrcmp0(param, "help")) {
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_WARN,
m_config: improve option list output a bit This is printed with --list-options or e.g. --vf=lavfi=help. Note that in theory, the options should be able to print their own help, and we shouldn't special case certain types (like m_option_type_choice in the commit). But that is too hairy for now, so we don't do it.
2013-07-21 18:16:46 +00:00
"Option %.*s: %.*s has no option description.\n",
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
BSTR_P(opt), BSTR_P(str));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_EXIT - 1;
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
if (_ret) {
plist = talloc_zero_array(NULL, char *, 4);
plist[0] = talloc_strdup(NULL, "_oldargs_");
plist[1] = bstrto0(NULL, param);
}
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
} else {
options: handle presets directly in m_config This means that "mpv -vo opengl-hq:help" will actually show the correct defaults, instead those of plain vo_opengl.
2013-07-31 18:45:06 +00:00
struct m_config *config = NULL;
if (!skip)
config = m_config_from_obj_desc(NULL, &desc);
r = m_obj_parse_sub_config(opt, str, pstr, config,
M_SETOPT_CHECK_ONLY,
_ret ? &plist : NULL);
talloc_free(config);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (r < 0)
return r;
}
}
if (!_ret)
return 1;
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
m_obj_settings_t item = {
.name = bstrto0(NULL, str),
.label = bstrdup0(NULL, label),
.attribs = plist,
};
obj_settings_list_insert_at(_ret, -1, &item);
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
return 1;
}
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
// Parse a single entry for -vf-del (return 0 if not applicable)
// mark_del is bounded by the number of items in dst
static int parse_obj_settings_del(struct bstr opt_name, struct bstr *param,
void *dst, bool *mark_del)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
bstr s = *param;
if (bstr_eatstart0(&s, "@")) {
// '@name:' -> parse as normal filter entry
// '@name,' or '@name<end>' -> parse here
int idx = bstrspn(s, NAMECH);
bstr label = bstr_splice(s, 0, idx);
s = bstr_cut(s, idx);
if (bstr_startswith0(s, ":"))
return 0;
if (dst) {
int label_index = obj_settings_list_find_by_label(VAL(dst), label);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
if (label_index >= 0) {
mark_del[label_index] = true;
} else {
mp_msg(MSGT_CFGPARSER, MSGL_WARN,
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
"Option %.*s: item label @%.*s not found.\n",
BSTR_P(opt_name), BSTR_P(label));
m_option: allow using -vf-del with a name Until now, -vf-del required a list of indexes. This was a bit inconvenient, so add support for using filter names too. Also simplify the code a bit, doing the change would have been too painful otherwise.
2013-05-18 07:27:16 +00:00
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
*param = s;
return 1;
}
m_option: allow using -vf-del with a name Until now, -vf-del required a list of indexes. This was a bit inconvenient, so add support for using filter names too. Also simplify the code a bit, doing the change would have been too painful otherwise.
2013-05-18 07:27:16 +00:00
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
bstr rest;
long long id = bstrtoll(s, &rest, 0);
if (rest.len == s.len)
return 0;
m_option: allow using -vf-del with a name Until now, -vf-del required a list of indexes. This was a bit inconvenient, so add support for using filter names too. Also simplify the code a bit, doing the change would have been too painful otherwise.
2013-05-18 07:27:16 +00:00
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
if (dst) {
int num = obj_settings_list_num_items(VAL(dst));
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (id < 0)
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
id = num + id;
Add suffix support to the object list type. So now -vf-clr destroy the list -vf-add filer1=blah,filter append these 2 filters, -vf-pre ... will 'prepend' them. Finnaly -vf-del 2,3,-1 will delete the filters at at given indexs It start from 0, negative number start from the end of the list (so -1 is the last one). git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9914 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-04-12 13:45:43 +00:00
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
if (id >= 0 && id < num) {
mark_del[id] = true;
} else {
m_option: allow using -vf-del with a name Until now, -vf-del required a list of indexes. This was a bit inconvenient, so add support for using filter names too. Also simplify the code a bit, doing the change would have been too painful otherwise.
2013-05-18 07:27:16 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_WARN,
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
"Option %.*s: Index %lld is out of range.\n",
BSTR_P(opt_name), id);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
Add suffix support to the object list type. So now -vf-clr destroy the list -vf-add filer1=blah,filter append these 2 filters, -vf-pre ... will 'prepend' them. Finnaly -vf-del 2,3,-1 will delete the filters at at given indexs It start from 0, negative number start from the end of the list (so -1 is the last one). git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9914 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-04-12 13:45:43 +00:00
}
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
*param = rest;
m_option: fix -vf-del in profiles We don't bother with option verification (as it happens in profiles), because it's barely possible.
2013-05-20 22:04:46 +00:00
return 1;
Add suffix support to the object list type. So now -vf-clr destroy the list -vf-add filer1=blah,filter append these 2 filters, -vf-pre ... will 'prepend' them. Finnaly -vf-del 2,3,-1 will delete the filters at at given indexs It start from 0, negative number start from the end of the list (so -1 is the last one). git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9914 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-04-12 13:45:43 +00:00
}
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
static int parse_obj_settings_list(const m_option_t *opt, struct bstr name,
options: get rid of ambiguous option parsing Options parsing used to be ambiguous, as in the splitting into option and values pairs was ambiguous. Example: -option -something It wasn't clear whether -option actually takes an argument or not. The string "-something" could either be a separate option, or an argument to "-option". The code had to call the option specific parser function to resolve this. This made everything complicated and didn't even have a real use. There was only one case where this was actually used: string lists (m_option_type_string_list) and options based on it. That is because this option type actually turns a single option into a proxy for several real arguments, e.g. "vf*" can handle "-vf-add" and "-vf-clr". Options suffixed with "-clr" are the only options of this group which take no arguments. This is ambiguous only with the "old syntax" (as shown above). The "new" option syntax always puts option name and value into same argument. (E.g. "--option=--something" or "--option" "--something".) Simplify the code by making it statically known whether an option takes a parameter or not with the flag M_OPT_TYPE_OLD_SYNTAX_NO_PARAM. If it's set, the option parser assumes the option takes no argument. The only real ambiguity left, string list options that end on "-clr", are special cased in the parser. Remove some duplication of the logic in the command line parser by moving all argument splitting logic into split_opt(). (It's arguable whether that can be considered code duplication, but now the code is a bit simpler anyway. This might be subjective.) Remove the "ambiguous" parameter from all option parsing related code. Make m_config unaware of the pre-parsing concept. Make most CONF_NOCFG options also CONF_GLOBAL (except those explicitly usable as per-file options.)
2012-08-05 21:34:28 +00:00
struct bstr param, void *dst)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
{
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
int len = strlen(opt->name);
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
m_obj_settings_t *res = NULL;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
int op = OP_NONE;
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
bool *mark_del = NULL;
int num_items = obj_settings_list_num_items(dst ? VAL(dst) : 0);
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
struct m_obj_list *ol = opt->priv;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
assert(opt->priv);
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (opt->name[len - 1] == '*' && (name.len > len - 1)) {
struct bstr suffix = bstr_cut(name, len - 1);
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
if (bstrcmp0(suffix, "-add") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_ADD;
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
else if (bstrcmp0(suffix, "-set") == 0)
op = OP_NONE;
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
else if (bstrcmp0(suffix, "-pre") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_PRE;
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
else if (bstrcmp0(suffix, "-del") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_DEL;
options: use case-sensitive comparsion for options This is better for consistency, and also allows using -V as alias for --version.
2013-05-15 13:28:29 +00:00
else if (bstrcmp0(suffix, "-clr") == 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
op = OP_CLR;
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
else if (bstrcmp0(suffix, "-toggle") == 0)
op = OP_TOGGLE;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
else {
char prefix[len];
strncpy(prefix, opt->name, len - 1);
prefix[len - 1] = '\0';
options: commandline: don't print video filter list twice Video filter help output was printed twice since it used a MSGT type that was not suppressed during preparsing. Fix. This removes a "ID_VIDEO_FILTERS" line that was printed --identify, but I'm not aware of anything using that and other things like audio filters have no such ID_ output either. Also change some printf() calls to mp_msg() in code printing filter-specific help.
2011-07-31 17:40:05 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_ERR,
"Option %.*s: unknown postfix %.*s\n"
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
"Supported postfixes are:\n"
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
" %s-set\n"
" Overwrite the old list with the given list\n\n"
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
" %s-add\n"
" Append the given list to the current list\n\n"
" %s-pre\n"
" Prepend the given list to the current list\n\n"
" %s-del x,y,...\n"
" Remove the given elements. Take the list element index (starting from 0).\n"
m_option: allow using -vf-del with a name Until now, -vf-del required a list of indexes. This was a bit inconvenient, so add support for using filter names too. Also simplify the code a bit, doing the change would have been too painful otherwise.
2013-05-18 07:27:16 +00:00
" Negative index can be used (i.e. -1 is the last element).\n"
" Filter names work as well.\n\n"
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
" %s-clr\n"
" Clear the current list.\n",
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
BSTR_P(name), BSTR_P(suffix), prefix, prefix, prefix, prefix, prefix);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_UNKNOWN;
}
}
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
options: change option parsing to use bstr Using bstr allows simpler parsing code, especially because it avoids the need to modify or copy strings just to terminate extracted substrings.
2011-07-28 08:07:47 +00:00
if (!bstrcmp0(param, "help")) {
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_INFO, "Available %s:\n", ol->description);
for (int n = 0; ; n++) {
struct m_obj_desc desc;
if (!ol->get_desc(&desc, n))
break;
options: hide encoding AO/VO in help output These can't be used manually. Encoding is enabled with -o instead, and the encoding AO/VO is selected using internal mechanisms.
2013-07-21 20:03:53 +00:00
if (!desc.hidden) {
mp_msg(MSGT_CFGPARSER, MSGL_INFO, " %-15s: %s\n",
desc.name, desc.description);
}
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
}
options: commandline: don't print video filter list twice Video filter help output was printed twice since it used a MSGT type that was not suppressed during preparsing. Fix. This removes a "ID_VIDEO_FILTERS" line that was printed --identify, but I'm not aware of anything using that and other things like audio filters have no such ID_ output either. Also change some printf() calls to mp_msg() in code printing filter-specific help.
2011-07-31 17:40:05 +00:00
mp_msg(MSGT_CFGPARSER, MSGL_INFO, "\n");
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return M_OPT_EXIT - 1;
}
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
if (op == OP_CLR) {
if (dst)
free_obj_settings_list(dst);
return 0;
} else if (op == OP_DEL) {
mark_del = talloc_zero_array(NULL, bool, num_items + 1);
}
if (op != OP_NONE && param.len == 0)
return M_OPT_MISSING_PARAM;
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
while (param.len > 0) {
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
int r = 0;
if (op == OP_DEL)
r = parse_obj_settings_del(name, &param, dst, mark_del);
if (r == 0) {
options: use m_config for options instead of m_struct For some reason, both m_config and m_struct are somewhat similar, except that m_config is much more powerful. m_config is used for VOs and some other things, so to unify them. We plan to kick out m_struct and use m_config for everything. (Unfortunately, m_config is also a bit more bloated, so this commit isn't all that great, but it will allow to reduce the option parser mess somewhat.) This commit also switches all video filters to use the option macros. One reason is that m_struct and m_config, even though they both use m_option, store the offsets of the option fields differently (sigh...), meaning the options defined for either are incompatible. It's easier to switch everything in one go. This commit will allow using the -vf option parser for other things, like VOs and AOs.
2013-07-21 17:33:08 +00:00
r = parse_obj_settings(name, &param, ol, dst ? &res : NULL);
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
}
m_option: redo code for parsing -vf to accept quotes Parsing sub-configs (like --rawvideo=subopts or the suboptions for --vo=opengl:subopts) was completely different from the -vf parsing code for a variety of reasons. This change at least makes -vf use the same splitter code as sub-config options. The main improvement is that -vf now accepts quotes, so you can write things like: -vf 'lavfi=graph="gradfun=10:20"' (The '' quotes are for shell escaping.) This is a rather big and intrusive change. Trying some -vf lines from etc/encoding-example-profiles.conf seems to confirm it still works. This also attempts to unify one subtle difference in handling of positional arguments. One consequence is that a minor detail changes. Sub-configs don't know positional arguments, and something like "-- opt=sub1=val1:sub2" means that sub2 has to be a flag option. In -vf parsing, sub2 would be a positional option value. To remove this conflict and to facilitate actual unification of the parsers in the future, the sub2 will be considered a flag option if and only if such a flag option exists. Otherwise, it's considered a value for a positional option. E.g. if there's a filter "foo" with a string option "sopt" and a flag option "fopt", the behavior of the following changes: -vf foo=fopt Before this commit, this would set "sopt=fopt" in the filter. Now, it enables the fopt flag, and the sopt option remains unset. This is not an actual problem to my knowledge.
2013-04-21 01:33:13 +00:00
if (r < 0)
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return r;
options: add some features to video filter parser This is in preparation of making VOs and AOs use the parser which originally was for video filters only. The --vo and --ao options have several very annoying features, which are added here: - They can skip unknown video outputs (might be useful if a config file is supposed to work on several systems, where not all VOs/AOs are available everywhere) - The trailing "," in "-vo a,b," was significant, and meant that if "a" and "b" don't work, try the normal autoprobe order as fallback - There were deprecated VO names (like "gl3" and "gl"), which have to be handled with the option parser - Separating VO/VF names and options is different ("-vf foo=opts" vs. "-vo foo:opts") - vo_opengl.c provides opengl-hq as opengl + preset options
2013-07-21 19:17:25 +00:00
if (param.len > 0) {
const char sep[2] = {OPTION_LIST_SEPARATOR, 0};
if (!bstr_eatstart0(&param, sep))
return M_OPT_INVALID;
if (param.len == 0) {
if (!ol->allow_trailer)
return M_OPT_INVALID;
if (dst) {
m_obj_settings_t item = {
.name = talloc_strdup(NULL, ""),
};
obj_settings_list_insert_at(&res, -1, &item);
}
}
}
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
m_option: allow -vf "" With the current semantics, there's no reason to disallow this. (Although in my opinion, -vf should rather map to -vf-add than -vf-set, however that is an independent issue from this change.)
2013-05-18 15:44:21 +00:00
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
if (dst) {
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
m_obj_settings_t *list = VAL(dst);
if (op == OP_PRE) {
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
int prepend_counter = 0;
for (int n = 0; res && res[n].name; n++) {
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
int label = obj_settings_list_find_by_label0(list, res[n].label);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
if (label < 0) {
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
obj_settings_list_insert_at(&list, prepend_counter, &res[n]);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
prepend_counter++;
} else {
// Prefer replacement semantics, instead of actually
// prepending.
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
obj_setting_free(&list[label]);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
list[label] = res[n];
}
}
talloc_free(res);
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
} else if (op == OP_ADD) {
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
for (int n = 0; res && res[n].name; n++) {
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
int label = obj_settings_list_find_by_label0(list, res[n].label);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
if (label < 0) {
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
obj_settings_list_insert_at(&list, -1, &res[n]);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
} else {
// Prefer replacement semantics, instead of actually
// appending.
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
obj_setting_free(&list[label]);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
list[label] = res[n];
}
}
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
talloc_free(res);
} else if (op == OP_TOGGLE) {
for (int n = 0; res && res[n].name; n++) {
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
int found = obj_settings_find_by_content(list, &res[n]);
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
if (found < 0) {
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
obj_settings_list_insert_at(&list, -1, &res[n]);
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
} else {
m_option: move code around Move the helper functions for parsing -vf, and rename some.
2013-05-22 21:25:59 +00:00
obj_settings_list_del_at(&list, found);
obj_setting_free(&res[n]);
m_option: add -vf-toggle Works like -vf-add, except if a filter already exists and has the same parameters, it's removed instead of added. Not really useful on the command line itself, but will make sense for runtime filter changing in the following commit.
2013-05-18 09:31:03 +00:00
}
}
options: allocate dynamic options with talloc Allocate data structures for dynamic option values with talloc. Hopefully there's no code left that tries to modify those dynamic option values outside the option parser and relies on them being malloc-allocated. Currently talloc functionality isn't used much and the allocations are not hierarchical, so the main practical use for now is just to allow very easy checking for memory leaks.
2011-07-05 22:24:24 +00:00
talloc_free(res);
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
} else if (op == OP_DEL) {
for (int n = num_items - 1; n >= 0; n--) {
if (mark_del[n])
obj_settings_list_del_at(&list, n);
}
for (int n = 0; res && res[n].name; n++) {
int found = obj_settings_find_by_content(list, &res[n]);
if (found < 0) {
mp_msg(MSGT_CFGPARSER, MSGL_WARN,
m_option.c: fix typo in message
2013-08-02 15:06:49 +00:00
"Option %.*s: Item not found\n", BSTR_P(name));
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
} else {
obj_settings_list_del_at(&list, found);
}
}
free_obj_settings_list(&res);
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
} else {
assert(op == OP_NONE);
free_obj_settings_list(&list);
list = res;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
m_option, vf: add label support Can be used to refer to filters by name. Intended to be used when the filter chain is changed at runtime. A label can be assigned to a filter by prefixing it with '@name:', where 'name' is an user-chosen identifier. For example, a filter added with '-vf-add @label1:gradfun=123' can be removed with '-vf-del @label1'. If a filter with an already existing label is added, the existing filter is replaced with the new filter (this happens for both -vf-add and -vf-pre). If a filter is replaced, the new filter takes the position of the old filter, instead of being appended/prepended to the filter chain as usual. For -vf-toggle, labels are compared if at least one of the filters has a label; otherwise they are compared by filter name and arguments (like before). This means two filters are never considered equal if one has a label and the other one does not.
2013-05-22 21:19:57 +00:00
VAL(dst) = list;
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
}
m_option: allow removing filter entries by content with -vf-del If a complete filter description is passed to -vf-del, search for an existing filter with the same label or the same name/arguments, and delete it. The rules for filter entry equality are the same as with the -vf-toggle option. E.g. -vf-add gradfun=123:gradfun=456 -vf-del gradfun=456 does what you would expect.
2013-05-22 21:30:53 +00:00
talloc_free(mark_del);
cosmetics: reformat demux_lavf.c, m_option.[ch] and m_property.c
2011-07-03 12:42:04 +00:00
return 1;
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
}
correct const usage in the option handling code so that tables can be declared const and moved from .data to .text/rodata sections. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@24994 b3059339-0415-0410-9bf9-f77b7e298cf2
2007-11-09 06:50:53 +00:00
const m_option_type_t m_option_type_obj_settings_list = {
options: simplify option parsing/setting machinery Each option type had three separate operations to copy option values between memory locations: copy between general memory locations ("copy"), copy from general memory to active configuration of the program ("set"), and in the other direction ("save"). No normal option depends on this distinction any more. Change everything to define and use a single "copy" operation only. Change the special options "include" and "profile", which depended on hacky option types, to be special-cased directly in option parsing instead. Remove the now unused option types m_option_type_func and m_option_type_func_param.
2012-05-07 23:10:39 +00:00
.name = "Object settings list",
.size = sizeof(m_obj_settings_t *),
.flags = M_OPT_TYPE_DYNAMIC | M_OPT_TYPE_ALLOW_WILDCARD,
.parse = parse_obj_settings_list,
.copy = copy_obj_settings_list,
.free = free_obj_settings_list,
Add the new -vf option wich is the same as vop in reverse order. Syntax is we decided, so you can give the nomes or not with both vop and vf. vf take precedence over vop. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9594 b3059339-0415-0410-9bf9-f77b7e298cf2
2003-03-15 18:01:02 +00:00
};