Code review required for commits - formal Team vote.
ira at samba.org
Wed Oct 17 06:53:19 MDT 2012
On Wed, Oct 17, 2012 at 8:43 AM, simo <idra at samba.org> wrote:
> On Wed, 2012-10-17 at 05:18 -0700, Jeremy Allison wrote:
>> On Wed, Oct 17, 2012 at 07:17:48AM -0400, Ira Cooper wrote:
>> > On Wed, Oct 17, 2012 at 3:07 AM, Jeremy Allison <jra at samba.org> wrote:
>> > > On Wed, Oct 17, 2012 at 01:00:58AM +0200, Michael Adam wrote:
>> > >> - Do we need a tracking tool for reviews so patches don't get lost?
>> > >> (this is roughly a.k.a how do I get a reviewer for my patch?)
>> > >
>> > > Right now I'm using my inbox. What would you suggest as
>> > > a different tool ? At work we use gerrit, but Simo has
>> > > some valid (Java cough, cough :-) reasons to avoid that :-).
>> > Why is Java a valid reason?
>> I'll let Simo answer if he wants, but basically he objects
>> to them re-writing a git server rather than re-using the
>> existing one. Java is just the punchline to the joke :-).
> I mostly object in gerrit taking over the master tree, I'd be fine using
> it if it were a side tool that could push *and* pull from master if/when
> we push patches there.
> Also as much as I like neat UIs I strongly prefer a tool that let's you
> do everything from the command line. Autobuild does that, but afaik
> gerrit requires WebUI interaction, to review/push patches.
> I really do not like that, I want the option of doing reviews offline
> and commit stuff in my 'reviewd' branch to be pushed once I get online.
It has a cli.
I believe you can also just pull anything up for review... So you
should be able to avoid the web ui.
I personally believe the requirements we have can be met by some
python scripts that just keep track of what e-mail threads are active
on SN, and possibly output a webpage :)
Note: I don't care if it is Gerrit, or just a set of python scripts on
sn that say "this e-mail topic is active... this one is being reviewed
by x, this one is closed."
Nobody says this has to be complex. It as complex as we want to make
it, I explicitly stuck to giving requirements for a reason. How to
implement those requirements in a non-odious way is another topic. :)
More information about the samba-technical