Hello Michael, I have replaced 'ready' flag name to 'engine_ready' and added a corresponded comment. Regarding ff_qsv_prepare() looks like it is need not at all, I have modified ff_qsv_decode_init() instead. So patch become even simple, thank.
Please review. NOTE The ff_qsv_decode_init() should be another subject of refactoring, it should be split at least to 3 parts: -qsv/mfx session creation (common with qsvenc.c so this part should be moved to qsv.c or qsv_internal.c) -header detection (because it calls several times in common case and here can be a necessity to re-detect sequence header on a fly) -encoder creation. But I plan to take it some later, currently I have got couple critical bug fixes in my queue of patches. Thursday, July 23, 2015, 4:18:13 AM, you wrote: >> + int ready; MN> I think this should be documented, its not immedeatly clear from MN> just the name "ready" what exactly it refers to MN> maybe a more verbose variable name would make it also more immedeatly MN> obvious what it means MN> maybe something like MN> "int prepared" and ff_qsv_prepare() but maybe that is not the MN> best suggestion i dont know MN> also ff_qsv_prepare() should be documented what it exactly does/means MN> [...] -- Best regards, Ivan mailto:ivan.us...@nablet.com
0001-libavcodec-qsvdec_h264.c-refactoring-functional-of-q.patch
Description: Binary data
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
