I like the [MINOR] prefix because it makes it easy to identify simple commits (via grep or ctrl+f), the same way [CALCITE-1234] makes it easy to find commits related to [CALCITE-1234]. I also like that it maintains the "[...]" styling at the beginning of the commit message.
Neither of these reasons is strong enough for me to say I oppose, just some minor (heh) counter-arguments. -Tanner On Tue, Jan 2, 2024 at 1:05 PM Julian Hyde <jh...@apache.org> wrote: > Ralph Waldo Emerson once wrote: “A foolish consistency is the > hobgoblin of little minds, adored by little statesmen and philosophers > and divines." > > That said, people tend to bring conventions from other projects to > Calcite, and we end up with chaos. By which I mean, lots of > self-expression, but no standards, and therefore commit messages that > have lower information content, and more work for the release manager > coercing them into a consistent change log. > > In Calcite we have not used '[MINOR]' as a prefix to minor commits. If > it is minor, it doesn't need a jira case, and doesn't need a prefix. > But a few commits with [MINOR] crept in, starting about a year ago. > Once or twice, I asked people to remove them, but the PRs had already > been merged. > > Any objections if I add a lint rule to fail the build if the commit > message contains [MINOR]? > > While I'm there, any other standards we should enforce? > > Julian >
