| Message ID | Zy+KxjmZtc1KZfY5@tucnak |
|---|---|
| State | New |
| Headers | show |
| Series | [committed] contrib: Add 2 further ignored commits | expand |
On Nov 9, 2024, Jakub Jelinek <jakub@redhat.com> wrote: > r15-4998 and r15-5004 had wrong commit message, add those to > ignored commits. Ugh, sorry and thanks. Was that .c vs .cc only, or was there anything else? I'm surprised the commit-time checker didn't catch them. It used to, and that was very helpful to avoid typos in filenames.
On Sun, Nov 10, 2024 at 01:30:06PM -0300, Alexandre Oliva wrote: > On Nov 9, 2024, Jakub Jelinek <jakub@redhat.com> wrote: > > > r15-4998 and r15-5004 had wrong commit message, add those to > > ignored commits. > > Ugh, sorry and thanks. > Was that .c vs .cc only, or was there anything else? I think so. > I'm surprised the commit-time checker didn't catch them. I'm surprised too, but don't want to try to push further broken commits just to double check that. ;) > It used to, and that was very helpful to avoid typos in filenames. Yes. And I think it usually still does, I had one commit rejected because of such a reason recently. Jakub
On Sun, 10 Nov 2024, Jakub Jelinek wrote: > On Sun, Nov 10, 2024 at 01:30:06PM -0300, Alexandre Oliva wrote: > > On Nov 9, 2024, Jakub Jelinek <jakub@redhat.com> wrote: > > > > > r15-4998 and r15-5004 had wrong commit message, add those to > > > ignored commits. > > > > Ugh, sorry and thanks. > > Was that .c vs .cc only, or was there anything else? > > I think so. > > > I'm surprised the commit-time checker didn't catch them. > > I'm surprised too, but don't want to try to push further broken commits just > to double check that. ;) I think this is the known loophole where branch-specific checks only get applied to commits that are new to the repository, not to commits that were previously pushed to a branch with laxer checks. This has been fixed - see https://github.com/AdaCore/git-hooks/issues/20 - but we need to update the copy of the hooks used by GCC (there are multiple copies on sourceware, I think), which means checking for any incompatible changes since the version currently in use that might require updates to the hook configuration. (And then we'd need to use the new facilities introduced by the fix for that issue to ensure the branch-specific checks are applied to all commits added to relevant branches, not just ones new to the repository; it's not automatic even with updated hooks.)
On Nov 10, 2024, Jakub Jelinek <jakub@redhat.com> wrote: > On Sun, Nov 10, 2024 at 01:30:06PM -0300, Alexandre Oliva wrote: >> I'm surprised the commit-time checker didn't catch them. > I'm surprised too, but don't want to try to push further broken commits just > to double check that. ;) You can leave that testing to me ;-) As in, I'm sure I'll eventually try to push commits with CL typos again, and then we'll know ;-) >> It used to, and that was very helpful to avoid typos in filenames. > Yes. > And I think it usually still does, I had one commit rejected because of such > a reason recently. Hmm... I have a theory. Since it'd been a while since I'd refreshed the tree as I posted the patchset, I refreshed it, pushed to my testing branch, and ran a test build before pushing the same commits onto the trunk. Could it be that the checker didn't test the first push because it was for an open branch, and then didn't test the second push because the commits were not new?
diff --git a/contrib/gcc-changelog/git_update_version.py b/contrib/gcc-changelog/git_update_version.py index ec06fc965f8..005dcca769a 100755 --- a/contrib/gcc-changelog/git_update_version.py +++ b/contrib/gcc-changelog/git_update_version.py @@ -42,7 +42,9 @@ ignored_commits = { '040e5b0edbca861196d9e2ea2af5e805769c8d5d', '8057f9aa1f7e70490064de796d7a8d42d446caf8', '109f1b28fc94c93096506e3df0c25e331cef19d0', - '39f81924d88e3cc197fc3df74204c9b5e01e12f7'} + '39f81924d88e3cc197fc3df74204c9b5e01e12f7', + '8e6a25b01becf449d54154b7e83de5f4955cba37', + '13cf22eb557eb5e3d796822247d8d4957bdb25da'} FORMAT = '%(asctime)s:%(levelname)s:%(name)s:%(message)s' logging.basicConfig(level=logging.INFO, format=FORMAT,