Re: [cdt-dev] Committing through Gerrit

[Jeff Johnston <jjohnstn@xxxxxxxxxx>]

In Linux Tools, we use hudson/gerrit just to verify that the build works 
via tycho.  We don't run the tests.  If the build fails, the change is 
blocked.  Each change kicks off a parallel build.  Running the full test 
suite would be quite a burden with many changes posted.

The tests get run as part of the nightly build process and it is then 
possible to evaluate which ones are real or are problems that are being 
worked on.

A real reviewer can also run tests and reject if there is regression.

-- Jeff J.

On 09/17/2012 05:00 PM, Doug Schaefer wrote:
> I dunno. Do our tests pass consistently enough. I'd hate to see false
> positives rejecting changes when it's something else.
>
> On 12-09-17 4:25 PM, "Anna Dushistova" <anna.dushistova@xxxxxxxxx> wrote:
>
> > Why is it not possible for CDT to value the hudson/gerrit plugin?
> >
> > Anna.
> >
> > On Mon, Sep 17, 2012 at 11:21 PM, Doug Schaefer <dschaefer@xxxxxxx> wrote:
> > > I always download the change and do a quick test locally before
> > > committing.
> > > In a ideal world we would have hudson hooked up so we know the change
> > > passes
> > > tests automatically. But we're a ways away from that.
> > >
> > > From: Sergey Prigogin <eclipse.sprigogin@xxxxxxxxx>
> > > Reply-To: "CDT General developers list." <cdt-dev@xxxxxxxxxxx>
> > > Date: Monday, 17 September, 2012 3:03 PM
> > > To: "CDT General developers list." <cdt-dev@xxxxxxxxxxx>
> > >
> > > Subject: Re: [cdt-dev] Committing through Gerrit
> > >
> > > I find committing through Gerrit risky since CDT Gerrit is not hooked up
> > > into Hudson for running tests.
> > >
> > > -sergey
> > >
> > > On Mon, Sep 17, 2012 at 12:00 PM, Doug Schaefer <dschaefer@xxxxxxx>
> > > wrote:
> > > > That's a good idea. BTW, I thought that's what the difference between
> > > > +1
> > > > and +2 was. +1 means you're OK with the change but you need to +2 to
> > > > actually commit it, no?
> > > >
> > > > D
> > > >
> > > > On 12-09-17 2:56 PM, "Marc Khouzam" <marc.khouzam@xxxxxxxxxxxx> wrote:
> > > >
> > > > > Hi committers,
> > > > >
> > > > > as our use of Gerrit increases I thought I'd mention a practice we've
> > > > > been doing in Debug.
> > > > > When a committer reviews code from another committer, we let the
> > > > original
> > > > > committer do the
> > > > > actual commit.
> > > > >
> > > > > Therefore, in Gerrit, the reviewing committer would use the "Publish
> > > > > Comments" button
> > > > > and would leave the original committer to go to Gerrit and press
> > > > "Publish
> > > > > and Submit".
> > > > > That way, committers get to commit their own changes.
> > > > >
> > > > > I'm not sure if this practice is done throughout the CDT or not but I
> > > > > thought I would
> > > > > suggest it as a good practice if others agree.
> > > > >
> > > > > Thanks
> > > > >
> > > > > Marc
> > > > >
> > > > >
> > > > >
> > > > > _______________________________________________
> > > > > cdt-dev mailing list
> > > > > cdt-dev@xxxxxxxxxxx
> > > > > https://dev.eclipse.org/mailman/listinfo/cdt-dev
> > > >
> > > > _______________________________________________
> > > > cdt-dev mailing list
> > > > cdt-dev@xxxxxxxxxxx
> > > > https://dev.eclipse.org/mailman/listinfo/cdt-dev
> > >
> > >
> > >
> > > _______________________________________________
> > > cdt-dev mailing list
> > > cdt-dev@xxxxxxxxxxx
> > > https://dev.eclipse.org/mailman/listinfo/cdt-dev
> > _______________________________________________
> > cdt-dev mailing list
> > cdt-dev@xxxxxxxxxxx
> > https://dev.eclipse.org/mailman/listinfo/cdt-dev
>
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/cdt-dev