On Saturday, April 10, 2021 3:19:08 AM CEST Paul Eggert wrote: > On 4/9/21 12:50 AM, Kamil Dudka wrote: > > The warning[-Wstringop-overflow=] reports you refer to come from GCC > > actually. > Weird. The code being warned about has nothing to do with strings, and > the only string operation in the entire module is strlen and it is > compiled only on macOS.
I can see some regressions about stringop-overflow false positives in upstream bugzilla. They could be related to these reports. > And I don't get a warning from GCC or from Clang > when I run it myself on the "unfixed" version. This happens with a pre-release of gcc-11. > I suspect Coverity is using GCC (or Clang?) in some odd way, so for now > let's keep "Coverity" in the ChangeLog entry. Definitely not. I am able to reproduce it with gcc-11.0.0-0.20.fc34 or gcc-11.0.1-0.4.fc34 on a Fedora 34 VM where none of the other tools is installed. Kamil
