On 11/27/2019 5:19 PM, Michael Niedermayer wrote: > On Wed, Nov 27, 2019 at 01:17:34PM -0300, James Almer wrote: >> On 11/27/2019 1:09 PM, Michael Niedermayer wrote: >>> On Wed, Nov 27, 2019 at 11:41:09AM -0300, James Almer wrote: >>>> On 9/3/2019 10:49 AM, James Almer wrote: >>>>> On 8/30/2019 1:39 PM, James Almer wrote: >>>>>> Signed-off-by: James Almer <[email protected]> >>>>>> --- >>> >>>>>> Untested. >>> >>> i can test if it applies >>> >>> >>>>>> I'm also not sure how to add the FFMPEG_BSF define to tools/Makefile, >>>>>> and have it coexist with FFMPEG_DECODER. Assuming it's needed. >>> >>> maybe something like >>> tools/target_bsf_%_fuzzer.o: tools/target_dec_fuzzer.c >>> $(COMPILE_C) -DFFMPEG_BSF=$* >>> >>> but maybe i misunderdtand ... >> >> But then what's the decoder used for it? Where is FFMPEG_CODEC defined >> in the absence of FFMPEG_DECODER? > > do we need a decoder ? > the code could take the codec_id from the bsf codec_ids
Not using a decoder would require writing a different tool just for bsfs, because the way it's implemented in this patch, it processes a packet with a bsf before passing it to a decoder. Also, some bsfs support more than one codec, so we still need to pass two arguments: a bfs, and a codec. We can't just pick the first codec_id reported by the bsf. So i again wonder how exactly does the glue work here, and how can we pass two values to any given fuzzer tool in tools/Makefile. _______________________________________________ ffmpeg-devel mailing list [email protected] https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email [email protected] with subject "unsubscribe".
