Confusion with Blame Upstream Committers

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Confusion with Blame Upstream Committers

edriscoll
Hey all -

I'm new to Hudson, but loving it.  I've a job that does an svn update, builds, and runs unit tests (triggered by watching for svn changes).  It also tars up the build and stores it as an artifact.  If this passes, it kicks off a downstream build that does more intensive testing using that build (and does the right fingerprinting to establish the dependency).  This downstream job is setup with the blame upstream committers plugin.  But I'm seeing an odd thing - often when this downstream job fails it sends email to the wrong person, specifically it sends email not to the original build that triggered it, but to the one previous.  For example, say I have the following jobs:

BuildTrunk 100
BuildTrunk 101

Each will have their own uniquely fingerprinted artifact.  Assume a downstream job of TestTrunk, BuildTrunk 101 job triggered TestTrunk 53, which uses BuildTrunk 101's artifact.  If TestTrunk fails, it will end up sending email to the committer of BuildTrunk 100.  Now one thing that is suspicious, is that in the TestTrunk 53 job result details, there is this line:

Changes in dependency
BuildTrunk #100 →  #101 (detail)

Which makes me think that somehow Hudson thinks there is some sort of dependency between BuildTrunk 100 and 101, in which case, maybe it makes sense that the committer for job 100 is getting blamed for the TestTrunk 53 job failure.  So, I'm at a lost.  I've confirmed that TestTrunk 53 is fingerprinting the artifact from BuilTrunk 101 (I've compared the fingerprints).  Any ideas why the committer for the previous BuildTrunk job is getting the email?

Thanks,
evan
Reply | Threaded
Open this post in threaded view
|

Re: Confusion with Blame Upstream Committers

Christoph Kutzinski
Can you create a JIRA issue for this and assign it to me (kutzi), so I
won't forget?
I think, I've fixed a very similar issue in the Jabber/IRC plugin some
time ago.
I'll see if I contact the blame plugin maintainer or fix it myself.


Christoph

Am 20.01.2011 01:34, schrieb edriscoll:

>
> Hey all -
>
> I'm new to Hudson, but loving it.  I've a job that does an svn update,
> builds, and runs unit tests (triggered by watching for svn changes).  It
> also tars up the build and stores it as an artifact.  If this passes, it
> kicks off a downstream build that does more intensive testing using that
> build (and does the right fingerprinting to establish the dependency).  This
> downstream job is setup with the blame upstream committers plugin.  But I'm
> seeing an odd thing - often when this downstream job fails it sends email to
> the wrong person, specifically it sends email not to the original build that
> triggered it, but to the one previous.  For example, say I have the
> following jobs:
>
> BuildTrunk 100
> BuildTrunk 101
>
> Each will have their own uniquely fingerprinted artifact.  Assume a
> downstream job of TestTrunk, BuildTrunk 101 job triggered TestTrunk 53,
> which uses BuildTrunk 101's artifact.  If TestTrunk fails, it will end up
> sending email to the committer of BuildTrunk 100.  Now one thing that is
> suspicious, is that in the TestTrunk 53 job result details, there is this
> line:
>
> Changes in dependency
> BuildTrunk #100 →  #101 (detail)
>
> Which makes me think that somehow Hudson thinks there is some sort of
> dependency between BuildTrunk 100 and 101, in which case, maybe it makes
> sense that the committer for job 100 is getting blamed for the TestTrunk 53
> job failure.  So, I'm at a lost.  I've confirmed that TestTrunk 53 is
> fingerprinting the artifact from BuilTrunk 101 (I've compared the
> fingerprints).  Any ideas why the committer for the previous BuildTrunk job
> is getting the email?
>
> Thanks,
> evan

Reply | Threaded
Open this post in threaded view
|

Re: Confusion with Blame Upstream Committers

edriscoll
I've logged this bug:
http://issues.hudson-ci.org/browse/HUDSON-8567

Note, I couldn't assign it to you as I don't have permission to do that (I think).