commit mail subject format

Aggelos Economopoulos aoiko at cc.ece.ntua.gr
Wed Dec 3 05:25:49 PST 2008


Simon 'corecode' Schubert wrote:
> Aggelos Economopoulos wrote:
>> Simon 'corecode' Schubert wrote:
>>> For anybody confused about the mail subject, let me explain the thing:
>>>
>>> DragonFly-2.1.1.5.g73473 master test/test README
>>> ^^^^^^^^^ ^^^^^ ^ ^^^^^^ ^^^^^^ ^^^^^^^^^ ^^^^^^
>>> |           |   |  |     ^branch  ^directory  ^file in directory
>>> |           |   |  abbreviated commit id
>>> that's us   |   topological distance to the release, here 5 commits
>>>           last "release"
>>>
>>> The string "2.1.1.5.g73473" is actually the output of git-describe and
>>> probably will be used in more places in the future (kernel version, ISO
>>> build rev, etc)
>>
>> Why do we need all this stuff in the message Subject? The "DragonFly-"
>> part is essentially wasted space (gee, like I need to be reminded which
>> folder I'm reading), space that could be used for something useful.
>> Also, if I want any of the rest of the info, all I need is the commit id
>> and I can use git to get it. Maaaybe having the commit id appended (so
>> it won't intrude too much) to the Subject would be useful so you can
>> speedily search for any discussions relevant to a commit, but that's
>> also debatable.
>>
>> Why can't we use the commit summary or just the branch, directory and
>> files changed? The whole idea behind the Subject: line is that you can
>> tell at a glance if you're interested in reading the mail body and I
>> don't see how this format accomplishes it... All the extra info could go
>> in the message body if people still want it to be instantly available.
>> Personally, I only care about the commitid and diffstat, but that's
>> just me.
>>
>> Another idea would be to stuff the extra info in the headers so that
>> each subscriber can select which fields to view, but I'd prefer a sane
>> format that more or less works for everybody.
>>
>> Opinions?
> 
> Exactly my thoughts.  Matt insisted on the DragonFly and the commit
> id-at-the-beginning for search/subject alignment reasons.

Search is much faster in git and also more accurate, since the repo also
contains all the CVS history. Again, the Subject: line is intended to
save people's time, it is not for arbitrary metadata.

Perhaps there is some other reason? Matt?

Aggelos






More information about the Commits mailing list