../../third_party/libaom/source/libaom/av1/common/arm/compound_convolve_neon_dotprod.c:40:19: error: always_inline function 'vdotq_s32' requires target feature 'dotprod', but would be inlined into function 'convolve4_4_2d_h' that is compiled without support for 'dotprod' ../../third_party/libaom/source/libaom/av1/common/arm/compound_convolve_neon_dotprod.c:67:12: error: always_inline function 'vdotq_s32' requires target feature 'dotprod', but would be inlined into function 'convolve8_8_2d_h' that is compiled without support for 'dotprod' ...
Created attachment 885878 [details] Build log
Created attachment 885879 [details] pqv log
Created attachment 885880 [details] Info log
Created attachment 885881 [details] Environment
(mentioned on irc, but just to have it on the bug too) This is likely another runtime-detection feature usage bit, and to build it it explicitly pass -march=armv8.2-a+dotprod (which is fine in this context, even if your cpu does not support this) -- but then the user's -march=armv8-a+crc is passed *after* and overrides that resulting in a build failure. In qtwebengine:6 been using a patch (chromium/01-flags.patch) to invert this to allow chromium to override flags (except the ones we don't want like -O<level>), including e.g. -mno-avx followed by -mavx which removes the need to filter these. As a workaround for now, you could simply abstain from passing -march on this package, I assume it will build fine save for being a bit less optimized.
I am busy emerging chromium with the work around you suggested. It may take a few days to complete before I can report back.
Created attachment 886404 [details] New Build Log I tried to build it without the -march defined. It went further but failed wit a new error. I have include the new build log
The latest build of chromium emerged successfully
