|
From: Eyal S. <eya...@gm...> - 2021-03-02 00:43:43
|
Hi all! I have been investigating https://bugs.kde.org/show_bug.cgi?id=432801 and it looks like a false-positive due to a clang optimization. Perhaps it could be fixed if there were a memcheck "expensive" version of greater-than and less-than. I don't want to dig into it if someone already determined that this is a bad idea! Has anyone done work on this before? Also: If I did the work, would a patch be accepted? Thanks, Eyal |
|
From: Eyal S. <eya...@gm...> - 2021-03-02 18:49:11
|
Hi again! I have modified the patch and added a regression test, too. You can run the test both with and without the changes to memcheck and see how memcheck's results are improved. https://github.com/eyal0/valgrind/commit/7886d69c1ce503cb6dec05b239343743f0645d2a.patch Thanks! Eyal On Mon, Mar 1, 2021 at 5:43 PM Eyal Soha <eya...@gm...> wrote: > Hi all! > > I have been investigating https://bugs.kde.org/show_bug.cgi?id=432801 and > it looks like a false-positive due to a clang optimization. Perhaps it > could be fixed if there were a memcheck "expensive" version of greater-than > and less-than. > > I don't want to dig into it if someone already determined that this is a > bad idea! Has anyone done work on this before? > > Also: If I did the work, would a patch be accepted? > > Thanks, > > Eyal > |