The guidelines written by Chris Beams are designed to fit well within the
gui for github. He specifically says try for 50 but then try to keep 69
chars as a hard-limit because the github gui will truncate at 69 chars and
append ellipse "..." at the end. This mostly affects non-apache committers
creating pull requests.

I favor loose adherence to these guidelines:

1st line should be limited to 50-69 chars if possible. Subsequent lines
should be wrapped at 72 chars.

-Kirk


On Wed, Aug 17, 2016 at 4:47 PM, Dan Smith <dsm...@pivotal.io> wrote:

> Yeah, 50 chars seems a bit short. I think I've been aiming for 72
> personally.
>
> -Dan
>
> On Wed, Aug 17, 2016 at 4:37 PM, Kirk Lund <kl...@pivotal.io> wrote:
>
> > Some examples from feature/GEODE-1781-02 which are my latest unmerged
> > commits...
> >
> > 1) 1st line is 62 chars. To shorten to 50: "GEODE-1799: fix javadocs on
> > CacheServerStats" or "GEODE-1799: change javadocs from Bridge to Cache"
> >
> > commit 0a07db189c8b928976ed6554499f15b6a64e1633
> > Author: Kirk Lund <kl...@apache.org>
> > Date:   Wed Aug 17 16:27:52 2016 -0700
> >
> >     GEODE-1799: change javadocs from Bridge Server to Cache Server
> >
> > 2) 1st line is 80 chars. To shorten to 50: "GEODE-1781: exclude class
> from
> > AnaylzeSerializableJUnitTest"
> >
> > commit 55c5df5e4cd6228be1617fb1e92d8d955a703b08
> > Author: Kirk Lund <kl...@apache.org>
> > Date:   Tue Aug 16 09:25:33 2016 -0700
> >
> >     GEODE-1781: exclude LinuxProcFsStatistics$CPU from
> > AnaylzeSerializableJUnitTest
> >
> > 3) 1st line is 59 chars. To shorten to 50: "GEODE-1782: fix stat resource
> > equals"
> > (the later lines meet our guidelines)
> >
> > commit 7dc1ce68a483f993adeb613893073d8a7c88a9b7
> > Author: Kirk Lund <kl...@apache.org>
> > Date:   Mon Aug 15 18:35:45 2016 -0700
> >
> >     GEODE-1782: include start timestamp in stat resource equals
> >
> >     * stat resources with different time stamps should not be equal
> >
> >     * StatArchiveWithConsecutiveResourceInstGenerator generates gfs with
> >     multiple stat resources of same name but different times
> >
> >     * StatArchiveWithConsecutiveResourceInstIntegrationTest confirms
> >     existence of bug GEODE-1782: StatArchiveReader ignores later stats
> >     resource with same name as closed stats resource
> >
> >     * ResourceInstTest verifies the underlying issue and fix in
> >     StatArchiveReader.ResourceInst.equals and the fix
> >
> > 4) I got this one down to 48 chars!
> > (the later lines meet our guidelines)
> >
> > commit 115070123ec15638ca1189a7349938c35e0d51e0
> > Author: Kirk Lund <klund@Kirks-MacBook-Pro.local>
> > Date:   Mon Aug 15 18:26:16 2016 -0700
> >
> >     GEODE-1781: refactor internal statistics classes
> >
> >     * move internal statistics classes into
> com.gemstone.gemfire.internal.
> >     statistics
> >
> >     * move statistics tests into com.gemstone.gemfire.internal.
> statistics
> >
> >     * modify tests to include integration and distributed in names
> >
> >     * modify tests to use TemporaryFolder and TestName rules
> >
> > On Wed, Aug 17, 2016 at 4:22 PM, Kenneth Howe <kh...@pivotal.io> wrote:
> >
> > > Agree with Kirk, 50 chars is really short by the time you use up the
> > first
> > > 12 characters for the Jira tag. If we’re going to have a guideline, I’d
> > > rather be longer - somewhat arbitrarily I’d probably make it 20-30
> chars
> > > more. It’s been a long time since text listings were intended to fit
> on a
> > > 80x24 dumb terminal, so I don’t see a need to restrict the commit
> message
> > > headers so severely.
> > >
> > > I do use the —online option embedded in a local alias I use to look at
> a
> > > history list of my local repo.
> > >
> > > Ken
> > >
> > > > On Aug 17, 2016, at 3:45 PM, Kevin Duling <kdul...@pivotal.io>
> wrote:
> > > >
> > > > The format is very similar to the one most other git shops I've
> worked
> > in
> > > > before use.  I don't believe we ever had formal length limits.
> > > Typically,
> > > > it was:
> > > >
> > > > <JIRAPROJECT>-####: <Jira Ticket Summary>
> > > >>
> > > > blank line
> > > >
> > > > <brief description of fix, usually matching what was placed in the
> > > ticket>
> > > >
> > > >
> > > > The Atlassian plugin for IDEA automates a lot of this.  There are
> > limits
> > > on
> > > > the length of a jira ticket summary, but I'm not sure what that is.
> I
> > > ran
> > > > in to it when I did my round of CI.
> > > >
> > > > I don't see a reason to change anything except maybe stress that he
> > > lengths
> > > > are a guideline, not a hard & fast rule.  If more room is needed to
> > write
> > > > good information, it shouldn't be truncated as it's not unknown to
> move
> > > > away from a given ticket system.
> > > >
> > > > On Wed, Aug 17, 2016 at 3:38 PM, Kirk Lund <kl...@pivotal.io> wrote:
> > > >
> > > >> 50 chars including "GEODE-nnnn: " is awfully short.
> > > >> http://chris.beams.io/posts/git-commit/ does say that's just a
> > general
> > > >> rule
> > > >> of thumb and not a hard limit. The author's reasoning seems to be
> > > >> specifically for using "git log --oneline" -- does anyone use that
> > > option
> > > >> with git log? I don't.
> > > >>
> > > >> I guess another option is to not have to have a guideline if we
> don't
> > > want
> > > >> one... our current git log messages are reasonable and make sense.
> > > >>
> > > >> -Kirk
> > > >>
> > > >>
> > > >> On Wed, Aug 17, 2016 at 3:21 PM, Kirk Lund <kl...@pivotal.io>
> wrote:
> > > >>
> > > >>> Here's the git commit message guidelines we discussed and voted on
> > last
> > > >>> year. I just checked and my own git commit message line lengths
> have
> > > >> grown
> > > >>> beyond what we decided to use. Most other are also not following
> this
> > > >>> guideline.
> > > >>>
> > > >>> Here's the list of folks who voted last year along with their vote:
> > > >>>
> > > >>> Anthony Baker +1
> > > >>> Vincent Ford +1
> > > >>> William Markito +1
> > > >>> arghya sadhu +1
> > > >>>
> > > >>> Do we want to reaffirm this guideline or should it change?
> > > >>>
> > > >>> -Kirk
> > > >>>
> > > >>> ---------- Forwarded message ----------
> > > >>> From: Kirk Lund <kl...@pivotal.io>
> > > >>> Date: Wed, Aug 5, 2015 at 3:18 PM
> > > >>> Subject: git commit messages
> > > >>> To: dev@geode.incubator.apache.org
> > > >>>
> > > >>>
> > > >>> Several of us were discussing http://chris.beams.io/posts/
> > git-commit/
> > > --
> > > >>> there are a couple other really good articles about git commit
> > messages
> > > >> and
> > > >>> below is the message style I've been trying to follow.
> > > >>>
> > > >>> http://chris.beams.io/posts/git-commit/
> > > >>> http://www.laurencegellert.com/2013/07/how-to-write-a-
> proper-commit-
> > > >>> message/
> > > >>> http://tbaggery.com/2008/04/19/a-note-about-git-commit-
> messages.html
> > > >>>
> > > >>> GEODE-nn: Begin capitalized and 50 chars or less
> > > >>>
> > > >>> More detailed explanation with linefeeds to wrap at 72 characters
> > after
> > > >>> a blank line following the summary.
> > > >>>
> > > >>> Further paragraphs come after blank lines.
> > > >>>
> > > >>> - Bullet points are okay, too
> > > >>>
> > > >>> - Typically a hyphen or asterisk is used for the bullet, followed
> by
> > a
> > > >>>  single space, with blank lines in between, but conventions vary
> here
> > > >>>
> > > >>> - Use a hanging indent
> > > >>>
> > > >>>
> > > >>>
> > > >>>
> > > >>
> > >
> > >
> >
>

Reply via email to