On Tue, Apr 14, 2015 at 3:43 AM Diego Biurrun <di...@biurrun.de> wrote:
> On Fri, Apr 10, 2015 at 10:36:03AM -0700, Timothy Gu wrote: > > FATE currently discards this information anyway, so why waste the > > disk space? > > --- a/tests/fate-run.sh > > +++ b/tests/fate-run.sh > > @@ -211,7 +211,13 @@ else > > > > -echo "${test}:${sig:-$err}:$($base64 <$cmpfile):$($base64 <$errfile)" > >$repfile > > +if [ $err -eq 0 ]; then > > + unset cmpo erro > > +else > > + cmpo="$($base64 <$cmpfile)" > > + erro="$($base64 <$errfile)" > > +fi > > +echo "${test}:${sig:-$err}:$cmpo:$erro" >$repfile > > This changes the format of the FATE report files. No. Consider three tests, the first one succeeds without any output, the second one outputs some debug information, and the third fails. Before: test1:0:: test2:0:adsfasdfdsafdsafdsFADFadsf==: test3:-2:asfasfwer34fsdadf:adsfrqwfdsaadsf After: test1:0:: test2:0:: test3:-2:asfasfwer34fsdadf:adsfrqwfdsaadsf To be honest, I don't think it's worth the trouble. > On FFmpeg where this patch is already deployed: 248K 20150410161959 248K 20150410194845 56K 20150410231712 56K 20150411024535 204K 20150410161959/report.xz 204K 20150410194845/report.xz 12K 20150410231712/report.xz 12K 20150411024535/report.xz You can do the math, but I think it is definitely worth it. Timothy _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel