www-builds mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gav..." <ga...@16degrees.com.au>
Subject RE: Builds that have been failing for a while
Date Fri, 24 Sep 2010 09:44:53 GMT


> -----Original Message-----
> From: Jukka Zitting [mailto:jukka.zitting@gmail.com]
> Sent: Friday, 24 September 2010 7:25 PM
> To: builds@apache.org
> Subject: Re: Builds that have been failing for a while
> 
> Hi,
> 
> On Fri, Sep 24, 2010 at 11:05 AM, Niklas Gustavsson
> <niklas@protocol7.com> wrote:
> > These three builds are set to be checking for updates on a periodic
> > basis (polling the SCM every hour) and when upstream dependencies are
> > built.
> 
> That shouldn't be too much of a burden, or is it? It doesn't tie up
> executors like some of the other failing builds.
> 
> I'm all for disabling builds that continuously keep failing, but in
> these cases only the last build has failed, and I totally expect the
> builds to go blue again as soon as someone gets around to touching the
> codebases.
> 
> Instead of the time limit, would it make more sense to only disable
> those jobs where >n of the last builds have failed?

Depends on the trigger frequency, last n builds could be used up in one day
by some
projects and take months to reach for others.

I would suggest either a combination of both methods - perhaps time of 30
days .and.
the last 5 builds failed, or something like that?

this is a new thing that needs doing, we can't have everyone replying saying
'oh yeah
please don't disable my build due to blah ...' . Lets find a sensible
setting and
stick to it. The aim is to get people to fix their builds or they will be
disabled
until they are fixed, simple.

Gav...

> 
> BR,
> 
> Jukka Zitting



Mime
View raw message