GitLab

have found.

at free box content to see if it could be moov box, I hope there won't be
any misdetection. (Even when there is a spec, nobody follows it :((

* ALL: changed the prototype of input_AddES() to include enough information so we can build an "video-es", "audio-es" and "spu-es" object variable. These variables can be used by the interfaces to navigate between the elementary streams.
* modules/gui/wxwindows/menus.cpp: use the "foo-es" object variables.

big audio chunk. (Avoid 100000 samples at once :(

 * mp4: don't return a failed value when recognize a ref file, but use
a dummy Demux function that return EOF.

we should perhaps change the behavour of input_NewPacket).

* modules/demux/mp4/mp4.c: fixed an off by one error in array access.

the audio part of mov file. I'm really interested by any regression you
could find, please _test_ all .mp4/.mov file you have.

* ALL: i18n updates and fixes.

 * mp4 : removed a bad initialisation. Erased all ending spaces.
 * mpeg/audio : send i_rate to decoder (not yet usefull but ...)

informations to send.

(multiple moov headers is strictly forbidden but I 've seen such file)

decoder_fifo_t (p_waveformatex and p_bitmapinfoheader). It's
a replacement of p_demux_data as this variable is used by
some demuxers in another way. This way, I will be able to add
preliminary support of mpeg4 in mpeg2TS... (I haven't tested ogm
demuxer so if somebody wants ...)

are passed (more logical).

  * ALL: a bunch of additional warning fixes.

test if mp4,asf,avi,wav demuxers and araw,ffmpeg,faad decoders work under
big endian architectures ? (else could you give vlc log )

that don't work with it (I don't see why it doesn't work :(
 * mp4: fix a bug in timestamp calculation. Replace s/u* by s/uint*_t
 * avi : remove an useless debug message.
 * src/misc/objects.c : fix stream output object allocation.

        - fix a bug while playing last frames (in last chunk).
        - add seek support, not complete but usable.

    name but with different directory names.
  * ./Makefile: plugins get installed in different directory names.
  * ALL: additional -ansi -pedantic fixes.

    demux plugins (ps, ts, but also dvd*).
  * ./src/input/mpeg_system.c: removed MPEG-related stuff from the vlc core.
  * ./src/misc/modules.c: max recursion level is now 5.
  * ./modules/access/dvdplay/es.c: fixed a buffer overflow.

    waveout, directx, qnx, beos, win32, macosx, and the AltiVec modules.
  * ALL: removed mention of AC3 in favour of A52.
  * ./configure.in, ./Makefile*: modules can now be built deeper than 1
    directory. As a consequence, the build is even slower (but I'm fixing
    this) and make clean doesn't work anymore.

    about 2000 lines of code.

    I could not test everything on every architecture, so please report
    any module breakage (especially on BeOS, OSX and QNX).

but sometimes it's in fact MPEG-4, thx titer ).

 * mp4 : add capability to read mov with mp3 track. But it seems that
mad doesn't like too much fragmented data packets, and thus breaks the
audio.( since with '--codec dummy' I'm able to read resulting audio file
without any problem :)

 * ffmpeg and mp4: some clean up and change the way ffmpeg is
initialised.

  * include/input_ext-intf.h: killed p_es->b_audio, since i_cat already fills
    its purpose.
  * include/input_ext-intf.h: killed p_es->i_type in favour of p_es->i_fourcc.
  * include/video.h: killed p_pic->p_plane->b_margin; now i_visible_pitch is
    always valid.

  * ./src/libvlc.h: replaced --mpeg-adec and --ac3-adec with --codec; see
    vlc --help for more information.

  * include/video.h: removed fourcc values which should _not_ be global
    (thanks fenrir for convincing me).
  * ALL: video output plugins use the VLC_FOURCC macro to build their fourccs.
  * ./plugins/beos/InterfaceWindow.cpp: removed the last remaining call to
    input_ChangeES.

   Things I'm not sure about and need comments:
    - I used "mpga" and "mpgv" for MPEG (1+2) audio and video fourcc values.
    - I also introduced "spu ", "lpcm" and "ac3 ".
    - I removed for instance "MP4_GetCodec" because it wasn't really
      used except for the description string. I hope this change did not
      break anything really important.
    - fenrir, I must have broken a lot of things you wrote. Sorry in advance.

Thanks to jlj to show me this and to tell me what to check to solve it.

right).
 * cinepak : take care of grayscale option.

it ( add by hand to Makefile ... and link it with zlib -lz ) as I will
probably break it too many times. It work for some mp4(but without
audio, need aac decoder) and mov with sorenson v1(ffmpeg :).