Thank you for discussing about this. Michael Niedermayer wrote in <20141114203251.GX32286@nb4> >> + ctx->has_palette = 0; >> + ret = avio_open2(&ifo, p, AVIO_FLAG_READ, NULL, NULL); > >libavcodec does not depend on libavformat, thus libavcodec cannot >use avio_* directly. I should have realized/mentioned this when >suggesting avio
O.K. I understood using avio-* stuffs breaks independency of libavcodec. Thanks for pointing it out. Nicolas George wrote in <20141114224120.ga3...@phare.normalesup.org> >Le quartidi 24 brumaire, an CCXXIII, Michael Niedermayer a écrit : >> libavformat OTOH could read the IFO file with avio_open2/avio_read >> into extradata or some newly added ifo specific side data >> and dvdsubdec.c could then parse the palette out of that. > >That would work. It can be even reasonably simple: dvdsubdec already accepts >extradata with the Vobsub IDX syntax (the same way it is stored as codec >private data in Matroska). Unless I am mistaken, it would suffice to use a >BPrint buffer to write "palette: XXXXXX,XXXXXX..." and use >avpriv_bprint_to_extradata() afterwards. > >OTOH, if Shin-ichi is not motivated to change the patch once again, I >believe the version with fopen() is acceptable if marked experimental / >temporary in the docs. I hope I could do this in more proper way. But I'm quite new to inside ffmepg and my skill is immature. In addition, I have not enough spare time for this currently. I would like to learn more inside ffmpeg and look into how to use extradata method when I have time. Anyway, I revised the last patch using fopen() again and also marked it experimental in `decoders.texi'. Please review it and merge it if acceptable. I really appreceatate many help from you all. Thanks! -- TOYAMA Shin-ichi mailto:sh...@wmail.plala.or.jp
0001-New-option-for-obtaining-global-palette-from-.IFO-fi.patch
Description: Binary data
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
