You cannot force developers to call their java version the exact same as
everyone else... there is always some reason why people name them
differently... so you will always need a mapping layer... a smart mapping
layer is easy to config.


On 10 March 2014 21:47, Les Mikesell <[email protected]> wrote:

> On Mon, Mar 10, 2014 at 4:26 PM, Stephen Connolly
> <[email protected]> wrote:
> >
> >> Nothing at all wrong with the idea.  But I can take any mapping of one
> >> thing to another and mistype it or misunderstand it - or have the
> >> computer put stuff in the wrong character set so what I see isn't what
> >> jenkins will try to map.  What I want is a way to have the mapping
> >> done and check the results (by executing them to the extent possible)
> >> before committing it.  And I do call mapping one thing to another
> >> magic if there isn't a way to trace your way through it.
> >
> >
> > How can you verify that the environment you typed in to your build job of
> > "java-1.7" and "maven-3.2.1" is correctly mapped to the tool installer
> names
> > you have typed into the CI system of "java-1.7" and "maven-3.2.1"
>
> That's equally awkward.  I do that by reading the console log of the
> failed jobs.  But I'd rather not be doing even more of that.
>
> > Where would it make sense - except on the CI server - to validate those
> > labels?
>
> Basically I want things to run on a developer's desktop the same as
> they do in jenkins.  So the jenkins environment settings and tool
> locations are in fact only supposed to make up for any differences
> between a local developer's layout and the node where jenkins
> dispatches the job.  They shouldn't be some new thing that doesn't
> really exist.
>
> > The fancier mapping is config in Jenkins to allow Jenkins to understand
> that
> > e.g. "jdk-1.7" means "java-1.7"... in other words this is about telling
> > jenkins what the labels in CI mean... that is config that only makes
> sense
> > to the CI server.
>
> Yes, but we need a way to cut through all the unnecessary abstraction
> layers back to what a developer actually does locally.   And I'm sure
> a programmer's answer to that would be that we need another layer of
> abstraction...
>
> --
>   Les Mikesell
>      [email protected]
>
> --
> You received this message because you are subscribed to the Google Groups
> "Jenkins Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to