Hi, On Fri, Oct 28, 2011 at 8:19 AM, Janne Grunau <[email protected]> wrote: > On Fri, Oct 28, 2011 at 08:11:53AM -0700, Ronald S. Bultje wrote: >> On Fri, Oct 28, 2011 at 7:18 AM, Kostya Shishkov >> <[email protected]> wrote: >> > On Fri, Oct 28, 2011 at 07:13:33AM -0700, Ronald S. Bultje wrote: >> >> Hi, >> >> >> >> This fixes false positives of has_codec_delay_been_guessed() for >> >> streams where not every input picture generates an output picture, >> >> such as interlaced H264. >> >> >> >> Fixes 3 samples from the conformance suite without requiring >> >> -std_compliance strict (which I'd like to kill): >> >> MAIN/CVFI2_Sony_H.jsv >> >> MAIN/CVNLFI2_Sony_H.jsv >> >> FREXT/HVLCFI0_Sony_B/HVLCFI0_Sony_B.264 >> > >> > I'd say that AVStream looks like a public structure, so adding something >> > to it >> > requires a bump and mentioning in documentation. But the approach seems >> > sane. >> >> Moved to private data. >> >> Ronald > >> From 405de967266162b01bf21c35d398fbe801ce30dd Mon Sep 17 00:00:00 2001 >> From: Ronald S. Bultje <[email protected]> >> Date: Thu, 27 Oct 2011 20:37:34 -0700 >> Subject: [PATCH 1/2] lavf: use number of output pictures for delay checks. >> >> This fixes false positives of has_codec_delay_been_guessed() for >> streams where not every input picture generates an output picture, >> such as interlaced H264. >> --- >> libavformat/avformat.h | 5 +++++ >> libavformat/utils.c | 4 +++- >> 2 files changed, 8 insertions(+), 1 deletions(-) >> >> diff --git a/libavformat/avformat.h b/libavformat/avformat.h >> index 1b67ee6..ccc6c14 100644 >> --- a/libavformat/avformat.h >> +++ b/libavformat/avformat.h >> @@ -630,6 +630,10 @@ typedef struct AVStream { >> int codec_info_nb_frames; >> >> /** >> + * Number of frames that have been returned during try_decode_frame() >> + */ >> + >> + /** > > move or remove > >> * Stream informations used internally by av_find_stream_info() >> */ >> #define MAX_STD_TIMEBASES (60*12+5) >> @@ -639,6 +643,7 @@ typedef struct AVStream { >> int duration_count; >> double duration_error[MAX_STD_TIMEBASES]; >> int64_t codec_info_duration; >> + int nb_decoded_frames; > > I still don't see why it is useful to have it avstream if it's only used > from try_decode_frame
Don't want each function to have a million arguments. Admittedly the whole info struct could become a function argument, but it was decided long ago that that was not to be, so I'm just following that decision. Ronald _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
