geode-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kirk Lund <>
Subject Re: [discuss] Should we evaluate at commit messages as part of PR review?
Date Fri, 14 Sep 2018 22:05:51 GMT
*The commit message should follow imperative style.* The wiki page seems to
be missing this even though we agreed to it several times on the dev-list
over the last 3 years. I'll add this to the wiki page.

You can then say *"If I apply this commit, then it will..."* for any of the
git commits. For an explanation, see *


*Use this:*

*GEODE-xxxx: Fix failing CompositePropertySourceTest*

Instead of:
GEODE-xxxx: Fixing failing CompositePropertySourceTest
GEODE-xxxx: Fixed failing CompositePropertySourceTest

And *definitely* instead of:
GEODE-xxxx: failing CompositePropertySourceTest
GEODE-xxxx: polishing stuff
GEODE-xxxx: CompositePropertySourceTest is failing intermittently

Just to be clear, these last 5 are examples of how you should NOT word the
commit message.

On Fri, Sep 14, 2018 at 12:11 PM, Alexander Murmann <>

> I do find it very helpful to have the ticket number at the beginning of the
> title. It makes it really easy to scan the output of `git log --oneline` or
> GitX to see what tickets happened recently or since a certain tag.
> On Fri, Sep 14, 2018 at 11:55 AM, Bradford Boyle <>
> wrote:
> > How would people feel about removing the requirement to include the
> > "GEODE-XXXX: " prefix in the summary line? That accounts for about 25% of
> > the 52 character limit. We could move it to the first non-summary line of
> > the commit message.
> >
> > --Bradford
> >

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message