Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[] Fwd: Re: I'm wondering if it's time to look at CDO releng stuff


Moving the discussion to the mdt-ocl dev list.

-------- Original Message --------
Subject: Re: I'm wondering if it's time to look at CDO releng stuff
Date: Fri, 15 Feb 2013 14:22:41 +0000
From: Ed Willink <ed@xxxxxxxxxxxxx>
To: Adolfo Sanchez-Barbudo Herrera <asbh500@xxxxxxxxxx>


Please use the mdt-ocl-dev list for all normal project discussion.

Our discussions should be open for all to view and particpate in.



On 15/02/2013 13:45, Adolfo Sanchez-Barbudo Herrera wrote:
We all should build against milestones repos of projects on which we
depend, and the result should be contributed to the aggregator. I dont
know what other do, but we do it correctly:

1. Wait until projects we depend build their S-build, whose result
will surely be in their milestones repository and it will be
contributed to the aggregator.
2. Build our S-buiild using their milestones repository.

Your proposal sounds like let's use some Integration repos instead.
For every project on which we depend ? for some of them (for instance
+2 projects) ? No thanks, we are doing it in the right way which is
usually  practical since at +1 we only have to wait for EMF and UML
and at +3 every project on which we depend must be done at +2.

This doesn't prevent the potential problem in which an S-build of any
project on which we depend, introduce a last minute breaking change,
then the aggregator breaks, then we need to build a new S-build, etc

If we all do 1. and 2. well this will never happens although the
potential breakage won't avoid stressing you in the last minute anyway.

Moreover, building at +2 undoubtedly reduces releng work, and
introducing that potential last minute breakage is not usual and the
consequences apart of the stress for breaking the aggregator is simply
having to produce MXa, Mxb, etc. So if you find reasonable building at
+2, let's communicate that to PMC and obtain their feedback.

My preliminary analysis of the CDO releng adoption is done at

Adolfo Sanchez-Barbudo Herrera
Engineering Doctorate Student
University Of York
Follow me @adolfosbh

On 15/02/2013 10:55, Ed Willink wrote:
Hi Adolfo

    +2 only seems like a really good idea.

The only inconvenience is that there should be a lot of +2 projects
"waiting" for us and in spite of being at +2 we would still require to
wait for other projects e.g. Xtex
It's not absolutely necessary to wait. Projects should provide an
I-build of all API changes in advance, so a build against a recent
I-build should be the same as a pedantic S-build.



No virus found in this message.
Checked by AVG -
Version: 2013.0.2899 / Virus Database: 2639/6103 - Release Date: 02/14/13

Back to the top