GitLab

- This is needed for proper seek/pause/fast forward/...
- The decoder/packetizer do not need to scale packet length anymore as the
decoder thread do it.

It will avoid useless picture copy.

This is the issue exploited by the anime groups.
We now skip anything in { }
Note that ass/ssa cannot have a valid string with text { text } text.

I removed:
 i_x/i_y: redondant with region i_x/i_y. Now a subpicture always represent the
 whole video.
 i_flags: unused anywhere in the core.

 I may have broken a few video filters (osdmenu mainly). I will fix it if you
report the problem and the way to reproduce it.

This is most likely exploitable with malicious sub files.

There are real-life cases. There may also be differences in punctuations
and character sets. Besides the VLC core will anyway apply gettext to
these strings, so they are translatable in any case...

Note that _protocol abbreviations_ and friends are not localisable!

Fix potential segfault (check malloc return value).

This improves readability a lot and prevent us from reinventing the wheel.

Reported by Drew Yao

    use size_t instead of int to represent sizes
    removes unused parameters from:
        block_BytestreamInit()
        mpeg4_audio/LOASSyncInfo()
        cinepak/cinepak_Getv1()
        speex/SendPacket()
        subsass/ParseColor()
        ffmpeg/postprocess/InitPostproc() & PostprocPict()
        dvbsub/YuvaYuvp() & encode_pixel_line_{2,4,8}bp()
        faad/DoReordering()

vlc_block_helper.h:
    use size_t, removes unused parameter from block_BytestreamInit()

struct filter_t:
    removes unused parameters from pf_picture_link() & pf_picture_unlink()
    
cmml: fix a segfault (p_item isn't an input item, despite the name) present for some months, proving that nobody did use that code ?

Missing some cases that I could not test.
Also ffmpeg/chroma.c is locked, so I can but leave it broken.
Fix is for the remaining modules is obvious and the issue is easy to detect.