[FFmpeg-devel] Who is against GIT now? (Was: [RFC] libswscale into the FFmpeg SVN repo)

[Michael Niedermayer]

On Sun, Apr 05, 2009 at 07:08:53PM +0100, Robert Swain wrote:
> On 5/4/09 18:12, Luca Barbato wrote:
>> Diego Biurrun wrote:
>>> On Sun, Apr 05, 2009 at 06:56:37PM +0200, Luca Barbato wrote:
>>>> Diego Biurrun wrote:
>>>>> On Sun, Apr 05, 2009 at 05:28:40PM +0200, Luca Barbato wrote:
>>>>>> Michael Niedermayer wrote:
>>>>>>> and i belive there is no consensus on a switch to git ...
>>>>>> I just wonder who is against git now...
>>>>> I strongly dislike that metadata cannot be versioned or changed, for
>>>>> example commit messages.
>>>>>
>>>>> Given the often low-quality commit messages we suffer from around here
>>>>> and the importance they have when debugging or understanding the code
>>>>> I think there has to be a way to update and fix them.
>>>>>
>>>>> And I am most emphatically *not* talking about spelling errors in
>>>>> commit
>>>>> messages. Repeat after me, this is *not* about pedantic nitpicking[1].
>>>> git rebase -i on topic branches does what you'd ever dream about ^^;
>>>
>>> You misunderstand: It needs to be done in the repository, not in
>>> somebody's local tree.
>>
>> If instead of having the single tree you have more remote branches you
>> can do and publish more commits and hopefully fix those issue before
>> landing them in the master tree (if you like this workflow).
>
> It could also be noted that if we ask for git format-patch generated 
> patches, they contain the commit message. If we mandated this form of patch 
> rather than `git diff > somefile.patch` then the commit message could also 
> be reviewed.

Iam definitly in favor of commit messages in patches, independant of git,
also if its not in the patch checklist it should be added.


> It wouldn't stop subpar commit messages from core 
> developers/maintainers however but it would help in a lot of cases.

now here we have the problem because i think most of the bad commit messages
are from core developers and not applied patches.
but i didnt do a survey its just my gut feeling

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Avoid a single point of failure, be that a person or equipment.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20090405/8af8fd5f/attachment.pgp>