Eric Blake <ebl...@redhat.com> writes: > On 2/8/19 1:30 AM, Markus Armbruster wrote: >> Short story: please add >> >> [am] >> messageid = true >> >> to your .gitconfig. >> >> Long story. git-am can add a Message-id: tag. Looks like this: >> > >> >> Signed-off-by: Thomas Huth <th...@redhat.com> >> Reviewed-by: Daniel P. Berrangé <berra...@redhat.com> >> Reviewed-by: Philippe Mathieu-Daudé <phi...@redhat.com> >> Tested-by: Philippe Mathieu-Daudé <phi...@redhat.com> >> Acked-by: Alex Bennée <alex.ben...@linaro.org> >> ---> Message-id: 1549268743-18502-1-git-send-email-th...@redhat.com >> Signed-off-by: Peter Maydell <peter.mayd...@linaro.org> >> >> The Message-id identifies the patch e-mail. It makes finding the review >> thread easier and more reliable. It's also a valid key on Patchew[*]. > > I find the tag valuable enough in later git searches that I don't mind > feeding my own patches back through the mailing list to add it (patchew > helps with that, of course). But for it to become mandatory, we'd need > to enhance scripts/checkpatch.pl to enforce it.
I'm afraid checkpatch is the wrong place. When you submit v1 patches for review, there is no Message-id. Even for respins, we don't want one. It should be added when patches get applied for real, so the commit carries exactly one Message-id, and it refers back to the final version on the list. Two ideas: * Have Patchew flag pull requests lacking Message-id * Admittedly vague: some kind of git pre merge hook magic to make git-merge flag missing Message-id