Quoting Derek Buitenhuis (2021-08-09 15:43:02) > On 8/9/2021 11:29 AM, Anton Khirnov wrote: > > diff --git a/libavfilter/Makefile b/libavfilter/Makefile > > index b0348ccfa3..27dd0c4b47 100644 > > --- a/libavfilter/Makefile > > +++ b/libavfilter/Makefile > > @@ -559,7 +559,8 @@ SKIPHEADERS-$(CONFIG_VULKAN) += vulkan.h > > > > OBJS-$(CONFIG_LIBGLSLANG) += glslang.o > > > > -TOOLS = graph2dot > > +TOOLS = graph2dot \ > > + spherical_compare > > Is there a reason it needs a new tool rather than ffmpeg.c?
I do not believe every single testing-only feature needs to be stuffed into ffmpeg.c. It's big and complicated enough already, while tests should ideally be simple and test just the thing they are supposed to test. > > > +frame 0 > > +lavfi.ssim360.Y=0.97 > > +lavfi.ssim360.U=1.00 > > +lavfi.ssim360.V=1.00 > > +lavfi.ssim360.All=1.00 > > +lavfi.ssim360.dB=25.19 > > Is it wise to do a non-fuzzy compare of floats? One could hope that truncating to two decimal places might not break, but I suppose it's still possible. Guess I could change the tool to parse the strings and compare them properly. -- Anton Khirnov _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
