[PATCH] vd_ffmpeg: crash on changing colorspace and extradata handling
Hi, attached are to patches to vd_ffmpeg. The first one avoids a crash when the colorspace changes during decoding (e.g. mf://*.jpg when the jpgs use different chroma subsampling, like 00.jpg and 04.jpg in incoming). This is done by properly reinitializing the vo in this case, so that the get_buffer function returns large enough buffers. The second patch passes extradata on as default, instead of having this large list of all codecs that need it. This one may be more reason for discussion (and testing) I guess. Greetings, Reimar Döffinger
Hi, On Thu, Mar 23, 2006 at 11:40:12AM +0100, Reimar D?ffinger wrote:
The second patch passes extradata on as default, instead of having this large list of all codecs that need it. This one may be more reason for discussion (and testing) I guess.
I forgot to mention, maybe a switch()-case on sh->format might be better than the proposed if-then-else construct (though it would be a bit bigger diff I guess). As said, comments welcome. Greetings, Reimar Döffinger
Hi, On Thu, Mar 23, 2006 at 11:40:12AM +0100, Reimar D?ffinger wrote:
attached are to patches to vd_ffmpeg. The first one avoids a crash when the colorspace changes during decoding (e.g. mf://*.jpg when the jpgs use different chroma subsampling, like 00.jpg and 04.jpg in incoming). This is done by properly reinitializing the vo in this case, so that the get_buffer function returns large enough buffers. The second patch passes extradata on as default, instead of having this large list of all codecs that need it. This one may be more reason for discussion (and testing) I guess.
Any comments? Otherwise I intend to apply the first patch tomorrow. Greetings, Reimar Döffinger
Hi, On Thu, Mar 23, 2006 at 11:40:12AM +0100, Reimar D?ffinger wrote:
attached are to patches to vd_ffmpeg. The first one avoids a crash when the colorspace changes during decoding (e.g. mf://*.jpg when the jpgs
Applied.
The second patch passes extradata on as default, instead of having this large list of all codecs that need it. This one may be more reason for discussion (and testing) I guess.
Atteched is a version that I like better. Please comment, because otherwise I'll apply in a few days (because I will need something like it to support my NuppelVideo decoder in ffmpeg *g*) Greetings, Reimar Döffinger
Hi, On Mon, Mar 27, 2006 at 11:46:48PM +0200, Reimar D?ffinger wrote:
Atteched is a version that I like better. Please comment, because otherwise I'll apply in a few days (because I will need something like it to support my NuppelVideo decoder in ffmpeg *g*)
Applied. Greetings, Reimar Döffinger
participants (1)
-
Reimar Döffinger