[vtk-developers] Continuous Build Defect Process

David Cole dlrdave at aol.com
Mon Feb 18 10:04:27 EST 2013


You could consider gerrit as a sort of next... even though it's not a single named/shared branch, it's a similar thing conceptually. Stuff doesn't get merged to master till it passes the 'next' gauntlet (gerrit code review and CDash at home dashboards for VTK). Passing that means convincing peers, and making dashboards green.
 

 

 

-----Original Message-----
From: Pat Marion <pat.marion at kitware.com>
To: David Cole <dlrdave at aol.com>
Cc: bill.lorensen <bill.lorensen at gmail.com>; vtk-developers <vtk-developers at vtk.org>
Sent: Mon, Feb 18, 2013 2:52 am
Subject: Re: [vtk-developers] Continuous Build Defect Process


To clarify, I meant a next/master branch model *in addition* to code review, not as replacement.

Pat


On Mon, Feb 18, 2013 at 2:02 PM, David Cole <dlrdave at aol.com> wrote:

Developer diligence is always required regardless of your workflow or development model.

I would tend to lean more toward Bill's statement than Pat's.

With the CDash at home actually green (very soon, very soon) it should be an automatic reason to reject a topic if it introduces warnings, errors or test failures on the CDash at home dashboards.

 
Even after merging to master, dashboard observation is required to verify that it didn't break anything on an as-yet-untested combination of platforms and options.


D


 

 

-----Original Message-----
From: Bill Lorensen <bill.lorensen at gmail.com>
To: Pat Marion <pat.marion at kitware.com>
Cc: VTK Developers <vtk-developers at vtk.org>
Sent: Sun, Feb 17, 2013 7:12 pm
Subject: Re: [vtk-developers] Continuous Build Defect Process


I do not see how next/master is better than a good gerrit review.

On Sunday, February 17, 2013, Pat Marion  wrote:

This is great, Bill.  Thanks to everyone who helped out!

The statement on the wiki "Once reduced to 0, developer diligence is needed to keep the defects to 0. The burden is on the Gerrit reviewers"  I agree with the first statement, but not the second.  I think that the best way to keep the number at 0 is to have a next & master branch model.

Pat


On Sat, Feb 16, 2013 at 6:15 AM, Bill Lorensen <bill.lorensen at gmail.com> wrote:

Folks,

Some of you may be interested in the process we used to reduce the
Continuous build defects to 0:
http://itk.org/Wiki/VTK/SoftwareQuality/ContinuousBuildTestFailures

Thanks to all who contributed gerrit topics.

Bill
_______________________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtk-developers






-- 
Unpaid intern in BillsBasement at noware dot com


 
_______________________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtk-developers


 




 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20130218/35ac03e1/attachment.html>


More information about the vtk-developers mailing list