> Anyway, what do you guys think? Is it good? Are there too many > false-positives?
> Should the suppressions be moved back to `AM_CPPCHECKFLAG`s not to alter the > code, although it makes it harder to maintain? I'm fine with having the suppressions inline. I don't remember if it was me starting to put them into `AM_CPPCHECKFLAG` or suggested by someone. In any way, usually in Python code, I put them inline. > * Use a version that doesn't require the exhaustive checking (@eht16 why do > we use latest version in the first place? I see > [1562ca0](https://github.com/geany/geany-plugins/commit/1562ca07500505ab396fe99ce19cb211d1c2c8fa) > but no rationale… and if it's that the version in Ubuntu 20.04 is too old > for something, maybe using 22.04 would be enough?) In the CI we used the cppcheck version available in the Ubuntu release the runner is using. This version tends to be a bit outdated. The nightly builds for Debian Sid use rather new versions (by the design of Sid) and those caused various new findings or other problems which caused the nightly builds to fail. This probably bothers only me but it does. And so I try to fix or workaround the issues or find a hack to make cppcheck happy. This is how #1197, #1310 and maybe more cppcheck related changes emerged. In general I think for static code checkers it is fine to use recent versions and not rely on versions provided by the distribution to benefit from improved and new checks more quickly than the CI runner images are updated. > * we could avoid running cppcheck in the distcheck phase, I'm not sure it has > any value… does it? Good idea, I don't think it gives any value on distcheck. > Or it got removed in the version of #1197 that was eventually merged. Not > sure but I think when @eht16 upped the version of cppcheck in CI the > `--library=gtk` got built in or something, so it wasn't included in what got > merged. Might mean it won't work for older versions of `cppcheck` though. I don't mind at all if we use `--library=gtk` or not. AFAIR this was one way to make recent cppcheck versions happy. If there is a better alternative, this is fine. So, after all, whatever makes the CI, the builds on Debian Sid and @b4n happy, is fine for me :D. -- Reply to this email directly or view it on GitHub: https://github.com/geany/geany-plugins/pull/1346#issuecomment-2088790813 You are receiving this because you are subscribed to this thread. Message ID: <geany/geany-plugins/pull/1346/c2088790...@github.com>
