On Thu, 31 Jan 2019 at 13:30, Iain Sandoe <idsan...@googlemail.com> wrote: > > > > On 29 Jan 2019, at 10:22, Christophe Lyon <christophe.l...@linaro.org> > > wrote: > > > > On Tue, 29 Jan 2019 at 11:12, Iain Sandoe <idsan...@googlemail.com> wrote: > >> > <SNIP> > > >>> Did it work for you? > >> > >> So far it’s looking nice… (anything that makes the dg output more stable > >> is Good in my book) > >> > >> … I am not sure (not checked thoroughly) but I think I saw one case in > >> libgomp testing where it didn’t pick up the timeout, will poke some more > >> at that if i see it again. > > > > If you still have the logs, you can run the script manually. > > So it seems that the only library testsuite that uses dg-extract-results is > libstdc++v3. This explains why the script wasn’t applied to libgomp. It > DTRT if applied manually. Good news. I didn't imagine there were differences in the way results are extracted between our various tools/libs.
> I guess an edit of the various library testsuite Makefile(s) to generate the > original .sum file as a temp followed by dg-extract-results to the final file > would work. Will try it when I have a few moments (unless you get to it > first ;) ) Please do, I don't plan to check libgomp. Thanks Christophe > > Iain >
