git.net

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: calcite-avatica git commit: [CALCITE-2412] Add appveyor.yml to have tests on Windows against jdk1.8, jdk9, jdk10 Add Appveyor badge Add -DskipDockerCheck because of CALCITE-2385 and to make it sync with travis.yml


I don’t think it’s possible to automate. Crafting good messages is an art, not a science. Which is not to say that developers can’t get good at it, with a little practice.

I’m sorry if you feel intimidated by my critiques. The fact that people can criticize commits after the event is a good thing - it allows us to move faster, because PRs do not have to be made 100% perfect before they are submitted. I try to make my criticism as gently as possible, and bear in mind that everyone has this project’s best interests in mind.

Regarding https://git-wip-us.apache.org/repos/asf?p=calcite.git;a=commit;h=53e15af6c5e8e782b2edcd7f5bf4f5f32225d110. I wish that you had committed that change, and I was a bit surprised that you, as a committer, had not already committed it. It would have saved me 2 hours over the weekend reviewing and revising it. I included both contributors’ names in the message because I had squashed together two commits from different authors. I don’t know whether I broke guidelines, but I was acting in good faith.

Julian


> On Nov 19, 2018, at 10:55 AM, Vladimir Sitnikov <sitnikov.vladimir@xxxxxxxxx> wrote:
> 
>> Sorry to harp on to everyone about commit messages. But a clear code
> history is essential for future contributors, and clear release notes are
> essential for our users.
> 
> Sorry for highjacking the thread, however it would be really great if the
> verification could be automated.
> The fear of "Julian claiming the commit message not being good enough" is
> very real for me, and it is part of the reason I asked Michael to commit
> "[CALCITE-2266] Implement SQL:2016 JSON functions" (and he was kind enough
> :) ! )
> 
> PS. I'm not sure if
> https://git-wip-us.apache.org/repos/asf?p=calcite.git;a=commit;h=53e15af6c5e8e782b2edcd7f5bf4f5f32225d110
> is
> in line with "standards".
> Should committers (e.g. "Vladimir Sitnikov") be placed to the commit
> headline (first line)?
> I thought we don't place committer names to the headline, however the Art
> of Calcite Commit Message is close to black magic to me.
> 
> Vladimir