Rockbox mail archiveSubject: Re: jdgordon: r28078 - trunk/apps/radio
Re: jdgordon: r28078 - trunk/apps/radio
From: Paul Louden <paulthenerd_at_gmail.com>
Date: Wed, 15 Sep 2010 01:46:33 -0500
On 9/15/2010 1:38 AM, Jonathan Gordon wrote:
> Its alot easier when someone has 37 commits to find a bad one....
> http://www.ohloh.net/p/rockbox/commits/22027901 "Fix a bug concerning
> difficulty, pointed out by DerPapst in IRC.". No context at all, not
> even the name of the plugin.
Let's see. The message explains what the commit did, and considering
only one file is touched, the context seems pretty obvious. Care to
explain how "bugger" provides any context?
You're still basically trying to avoid the point - commit messages
should explain what happened. You haven't justified why they don't need
to yet. Are you going to continue avoiding answering that question?
There's no difference in the commit log between a trivial commit and a
non-trivial one unless the commit message makes it clear it's trivial -
you do this by writing a clear commit message that says "add missing
semicolon" or similar. If someone has to click on the file to know why
the commit happened, the log message didn't do its job.
Instead of playing "I've committed more than you, so I'll bring up how
few commits you have, and try to single one out" (by the way Karl, if
you're still reading, I expect you to jump in here and say JdGordon
should've singled me out in private too, unless you're planning on
following the hypocritical path since at this point we've got someone
*explicitly* singling someone out, rather than just the implication you
jumped on before) why don't you just discuss the *point* rather than
playing games around it?
The question is simply "what is the purpose of commit messages?" Why not
answer what *you* think they are for JdGordon?
Received on 2010-09-15