On Wed, Aug 3, 2016 at 3:31 PM, Martin Liška <mli...@suse.cz> wrote: > Hello. > > As I've going through all PRs related to gcov-profile, I've noticed this PR. > Current implementation of cycle detection in gcov is very poor, leading to > extreme run time > for cases like mentioned in the PR (which does not contain a cycle). Thank to > Joshua, I've > grabbed his patch and removed the scaffolding (classes: Arc, Block, ...) he > did. After doing that > the patch is quite subtle and fast (of course). > > The patch survives gcov.exp regression tests and I also verified than *.gcov > is identical before > and after the patch for Inkscape: > > $ find . -name '*.gcov' | wc -l > 10752 > > I'm also thinking about adding [1] to test-suite, however it would require > implementation 'timeout' > argument in gcov.exp. Does it worth doing?
We usually add such tests without any "timeout" and expect we'll notice if compile-time goes through the roof for them. Richard. > Ready to install? > Thanks, > Martin > > [1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67992#c5
