1
0
mirror of https://github.com/mpv-player/mpv synced 2024-12-25 00:02:13 +00:00
Commit Graph

34240 Commits

Author SHA1 Message Date
Stefano Pigozzi
6c1e21e223 stream_ftp: fix compilation with libav
Fixup commit for a04b35013a. That commit was
cherry-picked from mplayer(1) and thus assumes ffmpeg.
2012-11-22 00:05:53 +01:00
al
a04b35013a stream_ftp: support longer filenames
This change was split into 8 patches. Squash them together, as they
affect stream_ftp.c only.

stream ftp: readline: Fix off-by-one error

Even if max bytes are available read at most max - 1 bytes.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35427 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	stream/stream_ftp.c

stream ftp: readline: Always initialize output parameter buf

Only exception if passed parameter max is less than or equal
to zero. That cannot happen with the current code.

Additionally change readresp function to always copy the first
response line if the parameter rsp is non-NULL. This fixes some
error reporting that used uninitialized stack arrays.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35428 b3059339-0415-0410-9bf9-f77b7e298cf2

stream ftp: readline: Always try to read complete lines

If there is not enough space in the provided line buffer just
skip the remaining bytes until reaching EOL.

Usually we are only interested in the first 5 characters and
for everything else the (on-stack) response buffer should still
be big enough.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35429 b3059339-0415-0410-9bf9-f77b7e298cf2

stream ftp: Revise file descriptor usage

* Set unbound descriptor variables to -1
* Always test >= 0 to see if a variable refers to a valid descriptor

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35430 b3059339-0415-0410-9bf9-f77b7e298cf2

stream ftp: Only send QUIT command if connected

Do not attempt to send commands without control connections.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35431 b3059339-0415-0410-9bf9-f77b7e298cf2

stream ftp: Create buffers before opening control connection

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35432 b3059339-0415-0410-9bf9-f77b7e298cf2

stream ftp: Allocate command buffer on-heap

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35433 b3059339-0415-0410-9bf9-f77b7e298cf2

stream ftp: Increase command buffer size

Allow for more longish file names (be it because of length or more
lengthy characters).

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35434 b3059339-0415-0410-9bf9-f77b7e298cf2
2012-11-21 20:03:58 +01:00
cigaes
73618d11d4 subreader: do not skip the first char of ASS fields.
Without this change, fields that can be sometimes empty and
sometimes not, such as the Effect field, are counted in an
inconsistent way. Since the number of fields is used to find
where the text starts, it leads to internal field arriving
on the video.

Bug reported anonymously on the users mailing list.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35426 b3059339-0415-0410-9bf9-f77b7e298cf2
2012-11-21 20:00:52 +01:00
wm4
10a1ae7b72 screenshot: add subs even with vf_screenshot
Until now, screenshots with the video filter didn't add subs (unclear
whether that was an oversight or feature). Fix this and make behavior
when taking screenshots with vf_screenshot more consistent with VO
screenshots.

The change in vf_screenshot is needed, because add_subs() checks this
flag to decide whether it's allowed to mutate the image.

This commit has another user visible side effect. When taking a
screenshot each frame (using the "each-frame" mode of the screenshot
command), a normal screenshot command will stop the each-frame mode.
2012-11-21 19:58:51 +01:00
wm4
4b91861b67 video: fix bogus uses of mp_image.w/width
mp_image has this confusing distinction between the w/h and width/height
fields. w/h are the actual width and height, while width/height have a
very special meaning inside the video filter code: it's the actually
allocated width, which is also used for stride padding.

Screenshot related code abused the w/h fields to store the aspect
corrected size. Some code confused the role of w/h and width/height.

Fix these issues. For aspect corrected size, display_w/h are used, while
width/height should never be used outside vf.c internals and related
code.

This also fixes an actual bug when taking screenshots of anamorphic
video with vf_screenshot, as well as using vo_image with such videos.
2012-11-21 19:58:51 +01:00
wm4
ed01dfeff3 vo_x11: don't require framestepping to update OSD
The OSD couldn't be updated at all without frame stepping. This made the
VO unusable (In fact, vo_x11 should never be used, but it's provided as
fail-safe fallback). Implement VOCTRL_REDRAW_FRAME to deal with this.

Unfortunately, this exposes issues with draw_bmp's subtitle rendering:
areas that are not covered by subtitles at all are changed as well. The
"backup" mechanism provided by draw_bmp doesn't copy these areas, which
turns up as quite visible artifacts. (These are included when taking
screenshots too.)
2012-11-21 19:58:16 +01:00
wm4
4757f9d0ce vo_x11: add screenshot support
Although vo_x11 shouldn't be used, it's a good thing to have screenshot
support in absolutely all interactive VOs. (Except vo_caca, but that is
literally a joke.)
2012-11-21 19:56:59 +01:00
wm4
ea4332daf4 vo_xv: don't require frame stepping to remove OSD or subs
In order to improve performance, vo_xv didn't create a backup of the
video frame before drawing OSD and subtitles during normal playback. It
required the frontend to do frame stepping if it wanted to redraw the
OSD, but no backup of the video frame was available. (Consider the
following use case: enable the OSD permanently with --osd-level=3, then
pause during playback and do something that shows an OSD message. The
player will advance the video by one frame at the time the new OSD
message is first drawn.)

This also meant that taking a screenshot during playback with vo_xv
would include OSD and subtitles in the resulting image.

Fix this by always creating a backup before drawing OSD or subtitles.
In order to avoid having to create a full copy of the whole image frame,
introduce a complex scheme that tries to backup only the changed
regions.

It's unclear whether the additional complexity in draw_bmp.c for
backing up only the changed areas of the frame is worth it. Possibly
a simpler implementation would suffice, such as tracking only Y ranges
of changed image data, or even just copying the full frame.

vo_xv's get_screenshot() now always creates a copy in order not to
modify the currently displayed frame.
2012-11-21 19:56:59 +01:00
Stefano Pigozzi
264b5124e7 cfg-config: avoid warning with -Wincompatible-pointer-types
This was throwing off a warning with clang. Add a cast to (void *) like many
other options do.
2012-11-21 14:10:33 +01:00
Stefano Pigozzi
9125ba0a4f cocoa_common: create window in the correct screen 2012-11-21 13:42:55 +01:00
wm4
670a4a85a9 gl_common: improve Mesa compatibility for GL3 context creation
Don't request CONTEXT_FORWARD_COMPATIBLE when creating a GL3 context on
X11. This improves compatibility with some Mesa drivers.

Setting this bit was originally intended to make the code compatible
to newer (future, possibly not yet existing) OpenGL implementations.
These implementations wouldn't have to support deprecated legacy GL
features, which we know that we don't use, and could return a context
with a higher OpenGL version than requested.

In practice, this didn't work out. CONTEXT_FORWARD_COMPATIBLE requires
the driver to explicitly disable deprecated functionality. If it doesn't
do that, it has to reject context creation with that flag set. And it
turns out there are Mesa drivers which suffer from this caveat. See [1].

This probably also means that a GL application with a fixed set of
required GL features (such as a subset of GL 3.2 core, like vo_opengl.c)
may need to probe several GL versions if drivers really start dropping
legacy GL.

On Windows, CONTEXT_FORWARD_COMPATIBLE is still set. It is not known
if this is ideal, but fortunately there is no Mesa on Windows.

CONTEXT_CORE_PROFILE is still always set. For requested GL versions
lower than 3.2, this flag is ignored. Should we requires higher versions
later, we want it to be set, so don't remove it.

[1] https://bugs.freedesktop.org/show_bug.cgi?id=57241
2012-11-20 18:00:16 +01:00
wm4
ef75d0ff26 core: minor cleanups
Enable printf format warnings for set_osd_[t]msg.
Remove the pointless assertion in mplayer.c (the assertion proved that
the following NULL check is probably pointless, but leave that check
anyway for robustness - it's not really clear whether it's needed).
2012-11-20 18:00:16 +01:00
wm4
2cdbaaf31c osd: fix OSD status symbol display in some cases
The playback status symbol in the OSD status display on video (such as
displayed when seeking or with the show_progress input command)
sometimes kept displaying the last seek, without resetting the symbol.
(For example: disable the OSD, seek, enable the OSD, run show_progress;
but also other cases.)

The main reason for that was the code clearing the OSD bar is also
responsible for clearing the osd_function (which stores the playback
symbol). If no OSD bar was set, the osd_function was never reset.

Fix by always setting the timer for clearing the OSD bar and the
osd_function whenever the osd_function is set. Clearing the OSD bar
when it wasn't set is OK. If the OSD bar is set some time after
osd_function is set, the timer is overwritten - that's a good thing,
as it makes both disappear from the screen at exactly the same time.

Always reset osd_function to 0 and determine the playback status
explicitly from mpctx->paused when displaying the status on screen.
2012-11-20 18:00:16 +01:00
wm4
0e63702ef8 command: use yes/no instead of enabled/disabled for consistency
The idea is that the OSD uses the same names as the options, if
possible.
2012-11-20 18:00:16 +01:00
wm4
62e78fab79 configure: make --enable-debug default
This may result in larger binaries by default, and should be harmless
otherwise. Users are advised to use "make install-strip" if they want
binaries without debug symbols.
2012-11-20 18:00:16 +01:00
wm4
fe97ab6059 configure: add --disable-optimization, change --enable-debug semantics
--disable-optimization removes -O2 from CFLAGS.

Now --enable-debug only adds -g to CFLAGS, and doesn't disable
optimization anymore.

As an obscure feature, --enable-optimization=<n> adds -O<n> to CFLAGS.

Also remove stray $def_debug from configure.
2012-11-20 18:00:16 +01:00
wm4
77a77d8869 mplayer: disable auto-loading of external codecs.conf
Do not load codecs.conf files located in $PREFIX/etc/mpv/ or ~/.mpv/.
There really is no use for this, other than possibly breaking things.
It's still possible to use --codecs-file explicitly to load an external
config file, and this option can be used in ~/.mpv/config.

While we're at it, remove the global codecs_file variable, and another
unused variable.
2012-11-20 18:00:15 +01:00
wm4
3d41fb8e48 manpage: minor fixes
The typo in options.rst was introduced in a54088.

--hardframedrop mentioned in mpv.rst has been merged with --framedrop.
2012-11-20 18:00:15 +01:00
wm4
967e1e8290 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".
2012-11-20 18:00:15 +01:00
wm4
80270218cb osd: make the OSD and sub font more customizable
Make more aspects of the OSD font customizable. This also affects the
font used for unstyled subtitles (such as SRT), or when using the
--no-ass option. This adds back some customizability that was lost with
commit 74e7a1 (osd: use libass for OSD rendering).

Removed options:
--ass-border-color
--ass-color
--font
--subfont
--subfont-text-scale

Added options:
--osd-color
--osd-border
--osd-back-color
--osd-shadow-color
--osd-font
--osd-font-size
--osd-border-size
--osd-margin-x
--osd-margin-y
--osd-shadow-offset
--osd-spacing
--sub-scale

The font size is now specified in pixels as it would be rendered on a
window with a height of 720 pixels. OSD and subtitles are always scaled
with the window height, so specifying or expecting an absolute font
size doesn't make sense.

Such scaled pixel units are used to specify font border etc. as well.

(Note: the font size is directly passed to libass. How the fonts are
actually rasterized is outside of our control, but in theory ASS font
sizes map to "script" pixels and then are scaled to screen size.)

The default settings should be about the same, with slight difference
due to rounding to the new scales.

The OSD and subtitle fonts are not separately configurable. It has
limited use and would double the number of newly added options, which
would be more confusing than helpful. It could be easily added later,
should the need arise.

Other small details that change:
- ASS_Style.Encoding is not set to -1 for subs anymore
  (assuming subs use VSFilter direction in -no-ass mode too)
- use a different WrapStyle for OSD
- ASS forced styles are not applied to OSD
2012-11-20 18:00:15 +01:00
wm4
d23b620a89 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.
2012-11-20 18:00:15 +01:00
wm4
dcd2435e79 vf_softpulldown: fix possible regression
The lines added with this commit were accidentally removed in commit
a0d759, which was a pure cleanup commit merged from mplayer-svn. The
difference between mplayer-svn and the mplayer2 base is that this
filter has been made PTS aware in mplayer2.

Also remove the redundant initialization of vf->priv->state.
2012-11-20 18:00:15 +01:00
wm4
2a353381f3 core: fix crash when video filter returns inf as PTS
When a video filter returned inf as PTS, the player crashed. One
reason for this was that decode_audio() was called with a negative
minlen parameter, which at some point caused it to call a memory
allocation function with a ridiculous value, triggering an out of
memory code path in talloc.c. (talloc.c has been modified to abort()
on out of memory situations.)

Fix this by sanity checking minlen in decode_audio(). (The check
against outbuf->len always succeeded, because it's an unsigned
comparison.)

Make an existing sanity check in mplayer.c more robust: check for NaN
too, which happens if the video PTS is inf.

This happened with "-vf pullup,softpulldown" (but is not triggered when
the following commit is applied).
2012-11-20 18:00:15 +01:00
wm4
6f6dfc5163 mplayer: fix potential issue when ao_play() fails
ao_play() can fail; in that case a negative error code is returned.
This error code is returned by write_to_ao() in turn. The function
fill_audio_out_buffers(), which calls write_to_ao(), doesn't check for
any error codes, and will likely trigger the assertion following the
function call. Change write_to_ao() to return 0 on failure to hopefully
prevent crashes when AOs fail.
2012-11-20 18:00:15 +01:00
wm4
9085b85729 stream: fix dvd:// + cache crashing
The language string was dynamically allocated, which completely fails
if the cache is forked (which it usually is). Change it back to a fixed
length string, like the original code had it.
2012-11-20 18:00:15 +01:00
wm4
efaa73cc73 stream, demux_lavf: minor cleanup for stream size code 2012-11-20 18:00:15 +01:00
wm4
ddffcce678 stream, demux: replace off_t with int64_t
On reasonable systems, these types were the same anyway. Even on
unreasonable systems (seriously, which?), this may reduce potential
breakage.
2012-11-20 18:00:15 +01:00
reimar
51dac4e070 stream: change STREAM_CTRL_GET_SIZE argument type to uint64_t
Update endpos each time libavformat asks for it.

Fixes playback of still downloading files to not stop before we
really reached the end.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35107 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	libmpdemux/demux_lavf.c

Change STREAM_CTRL_GET_SIZE argument type from off_t to
uint64_t.
Also fix the incorrect type of the uint64_res variable.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35360 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	libmpdemux/demux_lavf.c
	libmpdemux/muxer_lavf.c

Note: also merges the "forgotten" cache support from r35107.
2012-11-20 18:00:14 +01:00
upsuper
7759e565b7 asf_streaming: remove broken MMSU support code
Comment out unused code

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35359 b3059339-0415-0410-9bf9-f77b7e298cf2

Removing mmsu-related code

MMSU was never supported by MPlayer, and the protocol has been
deprecated. Since the code is not in use at all, Removing it
should not break anything.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35420 b3059339-0415-0410-9bf9-f77b7e298cf2
2012-11-20 18:00:14 +01:00
reimar
3f85094d4e Fix potential bugs and issues, general cleanups
Most of these are reimar fixing issues found by Coverity static
analyzer, and possibly some more cleanup commits independent from
this.

Since these commits are rather noisy, squash them all together.

Try to make code a bit clearer.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35294 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	audio/out/ao_alsa.c

Check the correct variable for NULL.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35323 b3059339-0415-0410-9bf9-f77b7e298cf2

Remove pointless unreachable code (the loop condition already checks
the 0xff case).

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35325 b3059339-0415-0410-9bf9-f77b7e298cf2

Fix typo that might have caused reading beyond the string end.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35326 b3059339-0415-0410-9bf9-f77b7e298cf2

Do not needlessly use "long" types.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35331 b3059339-0415-0410-9bf9-f77b7e298cf2

Use AV_RB32 to avoid sign extension issues and validate offset before using it.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35332 b3059339-0415-0410-9bf9-f77b7e298cf2

Remove nonsense casts.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35343 b3059339-0415-0410-9bf9-f77b7e298cf2

Fix crash in case sh_audio allocation failed.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35348 b3059339-0415-0410-9bf9-f77b7e298cf2

Fix potential NULL dereference.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35351 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	libmpcodecs/ad_ffmpeg.c

Note: Slightly modified.

Fix malloc failure check to check the correct variable.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35353 b3059339-0415-0410-9bf9-f77b7e298cf2

Avoid code duplication and pointless casts.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35363 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	stream/tv.c

Error out if an invalid channel list name was specified
instead of continuing and reading outside array bounds
all over the place.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35364 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	stream/tv.c

Make array "static const".

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35365 b3059339-0415-0410-9bf9-f77b7e298cf2

Properly free resources even when encountering many
parse errors.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35367 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	parser-cfg.c

Avoid leaks in error handling.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35380 b3059339-0415-0410-9bf9-f77b7e298cf2

Do not do sign comparisons on "char" type which can be both signed or unsigned.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35381 b3059339-0415-0410-9bf9-f77b7e298cf2

Free cookies file data after parsing it.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35382 b3059339-0415-0410-9bf9-f77b7e298cf2

http_set_field only makes a copy of the string, so we still need to
free it.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35383 b3059339-0415-0410-9bf9-f77b7e298cf2

check4proxies does not modify input URL, so mark it const.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35390 b3059339-0415-0410-9bf9-f77b7e298cf2

Remove proxy "support" from stream_rtp and stream_upd, trying
to use a http proxy for UDP connections makes no sense.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35394 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	stream/stream_rtp.c
	stream/stream_udp.c

Add url_new_with_proxy function to reduce code duplication and memleaks.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35395 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	stream/pnm.c
	stream/stream_live555.c
	stream/stream_nemesi.c
	stream/stream_rtsp.c

Fix off-by-one errors in file descriptor validity checks.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35402 b3059339-0415-0410-9bf9-f77b7e298cf2

Remove pointless cast.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35403 b3059339-0415-0410-9bf9-f77b7e298cf2

Abort when opening the file failed instead of calling
"write" with an invalid descriptor.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35404 b3059339-0415-0410-9bf9-f77b7e298cf2

Remove pointless local variable.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@35411 b3059339-0415-0410-9bf9-f77b7e298cf2

Conflicts:
	stream/http.c
2012-11-20 18:00:14 +01:00
wm4
f5e2ee5138 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.)
2012-11-20 18:00:11 +01:00
wm4
39428525c3 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.
2012-11-18 00:25:25 +01:00
wm4
af8ded53db sd_ass: fix ASS subtitles coming from demux_lavf
libavformat demuxes ASS subtitles as complete ASS event, rather than
Matroska-mangled events without time codes.
2012-11-16 22:54:22 +01:00
wm4
4e8e7d91dd command: export A/V sync difference as "avsync" property
This is the same as on the status line after the "A-V: ".
2012-11-16 21:21:16 +01:00
wm4
7a1396b6ca demux_mf: allow displaying single image files, various cleanups
Enable autoprobing for demux_mf, so that image files can be directly
displayed with e.g. "mpv file.jpg --pause". (The --pause switch is
needed to prevent the window from closing immediately.)

Since demux_mf doesn't have any real file format probing and goes by
file extension only, move the demuxer down the demuxer list to ensure
it's checked last. (ffmpeg's demux_mf equivalent, "image2", probes by
file extensions too, and there doesn't seem to be anything that can
probe typical image file formats from binary data.)

Remove the --mf "w" and "h" suboptions. Don't pass the width/height to
the video stream header. Both of these are useless, because the decoder
reads the real image size at a later point from the file headers.
Remove setting the BITMAPINFOHEADER as well, as vd_lavc doesn't need
this.

Enable --correct-pts by default. This fixes displaying a single image
with vo_vdpau (as mentioned by uau).

Keep around a pointer to the sh_video stream header instead of
accessing demuxer->video->sh_video. Fixes a crash when deselecting the
video track.

Note that the format probing is incorrect when opening images from HTTP
locations. File extensions don't have to match the actual file format.
A correct implementation would require to check the MIME type, or to
probe the binary data correctly.
2012-11-16 21:21:16 +01:00
wm4
f7163c8065 subtitles: improve support for libavformat demuxed subtitles
Make demux_lavf not error out if no video or audio track is present.
This allows opening subtitle files with the demuxer.

Improve the test whether subtitles read from demuxers must do explicit
packet reads. (I'm not sure whether always doing these reads could have
bad effects, such as reading too many audio and video packets at once,
so be conservative.)
2012-11-16 21:21:16 +01:00
wm4
589bda26ce mplayer: refactor: move computation of a/v sync
The computation for the A/V sync value was inside print_status(). Move
it into its own function; this makes things simpler and gets rid of some
minor dead code.
2012-11-16 21:21:16 +01:00
wm4
f1175cd905 core: add --keep-open, which doesn't close the file on EOF
The --keep-open option causes mpv not to close the current file.
Instead, it will pause, and allow the user to seek around. When
seeking beyond the end of the file, mpv does a precise seek back to
the previous last known position that produced video output.

In some corner cases, mpv might not be able to produce video output at
all, despite having created a VO. (Possibly when only 1 frame could be
decoded, but the video filter chain queues frames. Then a VO would be
created, without sending an actual video frame to the VO.) In these
cases, the VO window will not redraw, not even OSD.

Based on a patch by coax [1].

[1] http://devel.mplayer2.org/ticket/210#comment:4
2012-11-16 21:21:15 +01:00
wm4
b7052b431c command: add sub_reload and sub_remove commands
sub_remove remove an external subtitle track, for whatever this may be
needed.

sub_reload removes and re-adds an external subtitle track.

Also rename sub_load to sub_add, because that seems to be more in line
with sub_remove.
2012-11-16 21:21:15 +01:00
wm4
dd7dc2ee3d subreader: replace sub_free() by talloc destructor
Makes it less annoying to free the sub_data.
2012-11-16 21:21:15 +01:00
wm4
25a098fe78 options: add --mute for setting initial audio mute status
Similar to --volume. Takes this as opportunity to move the variable
corresponding to --volume into MPOpts.
2012-11-16 21:21:15 +01:00
wm4
1197e13c2f options: rename -sub-fuzziness to -autosub-match, change option values
"--autosub-match" is close to "--autosub", and reflects what this
option does slightly better. Replace the magic number option values
with choices:

--sub-fuzziness=0  becomes --autosub-match=exact
--sub-fuzziness=1  becomes --autosub-match=fuzzy
--sub-fuzziness=2  becomes --autosub-match=all
2012-11-16 21:21:15 +01:00
wm4
2b851c6ed3 options: rename --cursor-autohide, replace magic number values
--cursor-autohide-delay=-2   becomes --cursor-autohide=always
--cursor-autohide-delay=-1   becomes --cursor-autohide=no
--cursor-autohide-delay=123  becomes --cursor-autohide=123
2012-11-16 21:21:15 +01:00
wm4
269f7df18b options: --field-dominance: replace magic number values with choices
--field-dominance=-1  becomes --field-dominance=auto
--field-dominance=0   becomes --field-dominance=top
--field-dominance=1   becomes --field-dominance=bottom
2012-11-16 21:21:15 +01:00
wm4
51503a0577 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.
2012-11-16 21:21:15 +01:00
wm4
b4b86e9286 cookies: don't read cookie files from ancient browsers
Remove the code that attempted to read cookie files from well-known
browser locations. This code was written for ancient browsers, and only
knew about Mozilla and Netscape. While it's possible that these browsers
are still alive and still use the same config locations and cookie file
formats, the only Mozilla-based browser that still matters is Firefox.
Firefox uses a sqlite database for cookies, located in a slightly
different config path.

Just remove this code.
2012-11-16 21:21:15 +01:00
wm4
db45c8771a options: respect --no-msgcolor during early program start
Using --no-msgcolor, error messages that happened before "really"
parsing the command line were still printed in color. Add the
CONF_PRE_PARSE flag to make this option take effect as early as
possible.
2012-11-16 21:21:15 +01:00
wm4
2628ff6224 options: remove --display
Was used to set the X11 display. XDisplayName(NULL) does the same,
using the DISPLAY environment variable instead.
2012-11-16 21:21:15 +01:00
wm4
a540881571 options: rename --xineramascreen to --screen, remove magic values
--xineramascreen=-2 becomes --screen=all
--xineramascreen=-1 becomes --screen=current
2012-11-16 21:21:15 +01:00
wm4
dc0e610677 options: remove --osdlevel in favor of --osd-level
--osdlevel has been documented as renamed/removed. Not actually
removing it was probably an oversight.
2012-11-16 21:21:15 +01:00