tech-pkg archive

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

Re: pkgsrc Commit Message Policy



On 09/18, Thomas Klausner wrote:
> On Tue, Sep 12, 2017 at 03:40:00PM +0200, Thomas Klausner wrote:
> > For several years, there have been mirrors of pkgsrc in fossil, git,
> > and hg. Standard practise when using these tools is to make the first
> > line of a commit message function as a summary that can be read
> > without the rest, such as is commonly done with "git log --oneline".
> > Note that having a new commit message policy does not imply that a
> > decision to move to a DVCS has been made, but simply that we are now
> > in a world where DVCS usage is pervasive, even for repositories that
> > are authoritatively in CVS.
> > 
> > For this reason, pkgsrc-pmc has decided on the following guidelines
> > for pkgsrc commit messages:
> > 
> > Start the commit message with a line that explains the big picture in
> > 50 characters or less. When a commit is for one package, include the
> > name of the package. For updates, include the version to which it is
> > updated.
> 
> After listening to feedback, the limit has been increased to 65:
> 
> Start the commit message with a line that explains the big picture in
> 65 characters or less. When a commit is for one package, include the
> name of the package. For updates, include the version to which it is
> updated.

Hi, Thomas!

I'm not against a limit of 65 characters, but I am curious, would you
mind sharing the rationale behind it?  For example, maybe the initial
idea of 50 characters came from Tim Pope's "A Note About Git Commit
Messages" [1] blog post?  The rationale there seems to be that the
commit summary is used in various parts of Git output, and so a short
length for the summary makes that output look nicer by avoiding wrapping
or truncating.  With the pkgsrc-pmc guideline to include the name of the
package in the commit message if the commit is for a single package,
maybe a concern was that 50 characters could be challenging when the
package name eats up part of that right off the bat?  Again, just
curious.

Thanks!

Lewis

[1] http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html


Home | Main Index | Thread Index | Old Index