This affects property format strings like they are used in the
"show_text" input command, for --playing-msg, and other places.
To quote the documentation comment on m_properties_expand_string():
${NAME} is expanded to the value of property NAME.
If NAME starts with '=', use the raw value of the property.
${NAME:STR} expands to the property, or STR if the property is not
available.
${?NAME:STR} expands to STR if the property is available.
${!NAME:STR} expands to STR if the property is not available.
STR is recursively expanded using the same rules.
"$$" can be used to escape "$", and "$}" to escape "}".
"$>" disables parsing of "$" for the rest of the string.
Most importantly, "?(property:str)" becomes "${?property:str}".
Make the simple fallback case easier, e.g. "${property:fallback}"
instead of "${property}?(!property:fallback)".
Add the ability to escape the format meta characters. "$" is used for
escaping, because escaping with "\" is taken by the commands parser in
the layer below. "$>" can be used to disable interpretation of format
strings (of course escapes by the commands parser can't be canceled).
By default, properties which are unavailable or don't exist are turned
into a string signaling the status (e.g. "(unavailable)"), instead of
an empty string. If an empty string is desired, this has to be done
explicitly: "${property:}" (the fallback part is an empty string). Raw
properties still return an empty string on error.
m_properties_expand_string() now returns a talloc'ed pointer, instead of
a malloc'ed one.
This macro is supposed to make sure an array has enough memory allocated
for a given number of elements. Unfortunately, a condition was inverted
(I... what???). It only allocated if there was still enough memory left,
or only one additional element had to be allocated.
Since this macro was only used by MP_TARRAY_APPEND, this case was never
actually triggered. It's still utter non-sense.
If a command is not found, warn about it at loading time (just like
other command parsing errors are printed at loading time).
Add an explicit "ignore" command. input.conf instructs users to use this
command to cancel out existing mapping. This clashed with the
warning added in this commit. Make "ignore" a real command and remove
the specialcasing for it from get_cmd_from_keys(). Now "ignore" is
ignored because it's not handled in command.c.
Previously, both the command parser and property expansion
(m_properties_expand_string) handled escapes with '\'. Move all escape
handling into the command parser, and remove it from the property code.
This removes the need to escape strings twice for commands that use
property expansion.
The command parser is practically rewritten: it uses m_option for the
actual parsing, and reduces hackish C-string handling.
This is Better (tm).
The only actual change is that with M_PROPERTY_SET_STRING, the option
parser will use the property name, instead whatever was set in the
name field of the option returned by M_PROPERTY_GET_TYPE. In most cases,
these should be the same, though.
When input.conf is loaded, verify each command and print a warning if
it's invalid or uses legacy commands. This is done for both the user's
and the embedded config files.
The diff is a bit noisy, because mp_input_parse_cmd() is changed to take
a bstr as argument instead of a char*.
Now it depends on the command whether a property wraps around, or stops
at min/max valid property value.
For practically all properties, it's quite unambiguous what the "switch"
command should have done, and there's technically no need to replace it
with these new commands. More over, most properties that cycle are
boolean anyway. But it seems more orthogonal to make the difference
explicit, rather than hardcoding it. Having different commands also
makes it more explicit to the user what these commands do, both just due
to the naming, and what wrapping policy is used. The code is simpler
too.
Move the code for "switch_ratio" to the M_PROPERTY_SET case of the
"aspect" property. The rules are exactly the same, e.g. setting a ratio
smaller than 0.1 sets the pixel aspect ratio to 1:1. For now, we define
that writing "0" sets the PAR to 1:1, and disallow -1 (possibly reserve
it to reset to default aspect ratio).
Instead, communicate the new value range with M_PROPERTY_GET_TYPE. Add
M_PROPERTY_GET_WRAP to allow properties to enable cycling instead of
stopping at min/max.
This should be done by mp_property_generic_option() only. Also reindent
levels_property_helper() to make it a little bit more readable.
Remove the m_option_get_ptr() function, which doesn't really make sense
anymore.
The Doxygen-style documentation comments were nothing but bloat.
Also move mp_property_do() and mp_property_print() to command.h, where
they should belong, and fix their argument types. m_property.c/h is
supposed to be generic, while command.h provides declarations specific
to the mplayer core.
These should never be overridden by property implementations anyway,
because it would likely result in inconsistencies. The right way to do
this is adding a new m_option type.
Also some cosmetics.
Make more properties use the property-to-option bridge to reduce code
size and to enforce consistency. Some options are renamed to the same
as the properties (the property names are better in all cases).
Do some other minor cleanups. One bigger issue was memory management of
strings: M_PROPERTY_TO_STRING assumed the strings were statically
allocated, and no dynamic allocations could be returned. Fix this in
case the need for such properties arises in the future. Get rid of
m_property_string_ro(), because it's not always clear that the "action"
parameter is M_PROPERTY_SET and the string argument will be used.
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).
The m_option_value union is supposed to contain a field for each
possible option type (as long as it actually stores data). This helps
avoiding silly temporary memory alocations. Using a pointer to an union
and to a field of the union interchangeably should be allowed by
standard C.
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.
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.)
Most property implementations checked whether the "arg" parameter was
NULL. This is entirely pointless, because NULL is actually never pased.
It was inconsistently done, too. Remove the checks.
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.
This was the option parser for the off_t C type. These days, off_t is
always int64_t, so replace all its uses by int64_t and CONF_TYPE_INT64.
Fix the --sstep option. It used CONF_TYPE_INT with an off_t variable,
which will result in invalid memory accesses. Make it use type double
instead, which seems to make more sense for this option.
Rename both the option and property to "osd-level", which fits a bit
better with the general naming scheme. Make it a choice instead of an
integer range. I failed to come up with good names for the various
levels, so leave them as-is.
Remove the useless property handler for the "loop" property too.
Replace --hardframedrop with --framedrop=hard. Rename the framedrop
property from "framedropping" to "framedrop" for the sake of making
command line options have the same name as their corresponding
property. Change the property to accept choice values instead of
numeric values.
Remove unused/forgotten auto_quality variable.
osd_show_[property_]text => show_text
osd_show_progression => show_progress
show_text, osd_show_property_text and osd_show_text both map to the
code for the previous osd_show_property_text. The only special thing
about osd_show_text is that you don't need to escape "$". Also,
unfortunately osd_show_property_text requires escaping things twice,
one time for the command parser, and the other time for the property
formatting code, while osd_show_text needed only one level of escaping.
Use "-" instead of "_" in property names. The intent is that property
names and options names should be the same (if they refer to the same
thing), and options use "-" as word separator.
Rename some other properties too, e.g. "switch_audio" -> "audio".
Add a way to translate the old property names to the new ones, similar
to the input command legacy bridge.
Update input.conf. Use the new property names, and don't use legacy
commands.
Most input commands had their own policy whether to display an OSD
message for user feedback or not. Some commands had two variants, one
that showed an OSD message and one that didn't (e.g. step_property_osd
and step_property).
Change it such that all commands show a message on the OSD. Add a
"no-osd" modifier that disables OSD for that command. Rename the
"step_property" and "step_property_osd" command to "switch", and rename
"set_property" and "set_property_osd" to "set".
Note that commands which haven't used OSD before still don't use OSD.
That will possibly be fixed later. (E.g. "screenshot" could display an
OSD message instead of just printing a message on the terminal.)
The chapter and edition properties still produce OSD messages even with
"no-osd", because they don't map so well to the property_osd_display[]
mechanism.
Make the definition more compact and less confusing by omitting rarely
used fields. Avoid having to initialize osd_id with -1 in every entry
by making 0 an unused OSD ID instead.
There are many input commands which are redundant to properties. They
were parsed like normal commands, but set_property_command() in
command.c handled them automatically using the property mechanism. This
still required having the command specifications around, and the code in
command.c was quite messy.
Replace this with a text based replacement mechanism. Some corner cases
are not handled: commands of form "seek_chapter 3 1" are supposed to set
the "chapter" property to 3. This use is probably rare, and doesn't show
up in the default input.conf.
The reason compatibility is kept is because breaking input.conf is quite
annoying, so a minimal effort is made to avoid this. Currently we print
an annoying warning every time a legacy command is used, though.
Also add a compatibility entry for "pt_step", which was removed some
time ago. Variations in whitespace are not handled, but it's good enough
to deal with old input.conf entries.
These have been replaced by properties. Also remove some other slave-
mode specific get commands that can be replaced by property uses.
The get_metadata() function didn't actually contain anything useful,
and just replicated code from other parts of mplayer.
ao_dsound.c depended on the same configure check as vo_directx.c, which
was removed in commit 0e2c48a3ce. This accidentally disabled
inclusion of ao_dsound.
Fix it by adding a new check. Also, move it below ao_portaudio on the
auto-select list, as ao_dsound is considered deprecated.
Unrelated to that, move ao_lavc below ao_null to prevent it from being
auto-selected.
The change detection signalled a full re-upload on initialization, but
no texture reallocation. Binding the uninitialized buffer caused an
OpenGL error.
Fixes error messages on start with "opengl-hq".
GL_RGB16 doesn't seem to work universally (e.g. Intel). Use GL_RGB by
default, and use GL_RGB16 for "opengl-hq" only.
This may require users of Intel GPUs to manually experiment with the
fbo-format suboption when using "opengl-hq", as GL_RGB16 doesn't seem to
work there in some cases (black screen).
It's not really known whether PBO use causes problems of any kind (most
likely not). They should slightly increase performance. Use them by
default with "opengl-hq".
Even though PBOs don't have anything to do with rendering quality,
"opengl-hq" provides a test bed for features that should be enabled by
default, but aren't out of fear for regressions.
Change the default settings for vo_opengl to highest performance and
compatibility, but lowest quality. Use bilinear as default scaler.
Add "opengl-hq" as alias for high quality settings. This alias uses
exactly the same settings as vo_opengl did before this commit.
Normally, we don't want to call glGetError() all the time, because this
supposedly causes slowdowns. (I could not measure any on Linux with
nVidia binary drivers; maybe it's due to the fact that we have only a
few, expensive calls per frame.)
However, having to ask users to add the "debug" suboption when trying to
diagnose problems is very annoying. Since most errors happen during
initialization only, enabling it for the first 5 frames only is an
interesting compromise.
This renames vo_gl3 to vo_opengl, and makes it the default. The old
vo_gl is still available under "opengl-old".
We keep "gl3" as alias to "opengl" for short-term compatibility.
For OSX/Cocoa, the autoprobe order changes (prefer the "opengl" over
"opengl-old").
Remove "gl_nosw". This was a compatibility alias for "opengl-old", and
there's no point in keeping it.
Now both GL3 (for vo_gl3) and legacy context context creation (for
vo_gl) use the same code to create the X window. Only actual GL context
creation is different.
