Git migration

[George Goldberg]

On 11 October 2010 16:30, George Kiagiadakis
<kiagiadakis.george at gmail.com> wrote:
> On Mon, Oct 11, 2010 at 4:29 PM, David Edmundson
> <david at davidedmundson.co.uk> wrote:
>> Questions:
>>
>> In Olli's talk he said they push proposed branches to their own personal
>> directory on the collabora server. This was then reviewed.
>>
>> Given we are using reviewboard, what is the process? My understanding was that
>> reviewboard showed just a single diff, which means we won't get the breakdown
>> of individual commits which is one of the main advantages of using git.
>
> I was also wondering about that, but I think that reviewboard may
> actually handle git commits if you use the post-review script [1].
> Somebody needs to try it to find out :)

This may be a problem. In upstream telepathy we just use the gitweb
view of the branch for review and send comments by irc/email/bug
report. So basically, do whatever makes sense. I don't think we need a
strict policy.

>
>> Who needs to review a patch before it can be merged? Just any one person?
>> If hypothetically I wrote a patch for libktelepathy, how do I know who wants
>> to review it before it can be merged. Say DrDanz gave it a review++, I can
>> imagine George would still want to look at it before it's merged. His silence
>> might either mean he's busy, or agrees with DrDanz... or is mid-review saying
>> my patch is rubbish. I wouldn't know the state of it.
>
> I think that just one other person is fine, as long as that one other
> person (or the author of the patch) is one of the main contributors to
> this repository (i.e. quite familiar with the code). So, in the case
> of kcall for example, it would be acceptable if you review my patches
> or I review your patches, but not if you review domme's patches for
> example, since you both are not in the main contributors. Just my
> opinion here, of course...

Yup. Use your judgement. A person who is very familiar with or wrote
the affected code is the ideal reviewer. So in the case of the call
UI, I wouldn't merge anything without gkiagia's approval, but for the
kcm, perhaps you, or me, or boiko would be suitable reviewers since we
have all worked on it to some extent (although of course, I wouldn't
be a suitable review for code I myself wrote etc). Some bits of code
are less obvious - in that case better to ask more people than less
for an opinion (and some of them might say "if $foo says it's alright,
that's good enough for me").

--
George

>
> [1]. http://www.reviewboard.org/docs/manual/dev/users/tools/post-review/
> _______________________________________________
> KDE-Telepathy mailing list
> KDE-Telepathy at kde.org
> https://mail.kde.org/mailman/listinfo/kde-telepathy
>