On 26 Apr 2016, at 18:37, Nick Boldt wrote:
b) means "a PR change in Base could be used to rebuild Server,
to
verify no tests now fail"
c) means "simply aggregate the PR change in Base + all the latest
Server, Openshift, etc. bits from master" to be able to push that PR
change into a DP installer CI build
So they solve two different needs and will catch different problems.
But I'm not sure how to report back the status of any
downstream-PR-caused builds back to the original PR. Also not sure how
to contain situations where there's a pair (or trio?) of tied PRs
,
eg., an API change in Base (PR#1) and the related change in Server
(PR#2) & Openshift (PR#3). Either that's 3 builds, or it's 6 builds,
or 8 builds ... ? (Been too long since my last combinatorics course.)
so I don't follow what your problem is here since *one* PR would be the
trigger for all this so thats the one that would get the 'status'
reported - but I think we are jumping way ahead of what needs done
first, so lets start with one PR doing a build and report if that fails
the build for that repo.
/max
On Tue, Apr 26, 2016 at 11:50 AM, Max Rydahl Andersen
<manderse(a)redhat.com> wrote:
> a first.
>
> b is a separate thing and much more complex IMO
>
> c how is that different from b ?
>
> in other words, start with a :)
>
> /max
>> Hi
>>
>> I'd love to get a build triggered when I create a PR in github. That
>> would allow me to force awareness of breaking tests with a change.
>> When talking to Nick he was pointing out that there are actually 3
>> items
>> that he's after currently, where my suggestion is the 1st of these
>> aims.
>>
>> a) does the PR break my tests / want a green light on the PR page
>> b) does the PR break downstream (eg., API change in Base breaks
>> Server
>> or UI change in Server breaks Openshift)
>> c) does the PR need to be aggregated into JBT or JBDS so we can see
>> it
>> in the larger context
>>
>> a) is what I'd love to get and is good enough for me now.
>> b) is imho a nice-to-have but wasnt that much of a problem in
>> OpenShift
>> lately.
>> c) I guess is the larger picture where we want to achieve continous
>> delivery.
>>
>> Nick wants to know about the need of other component leads. So
>> please
>> speak up. a), b), c), all of them, etc?
>>
>> Cheers
>> Andre
>> _______________________________________________
>> jbosstools-dev mailing list
>> jbosstools-dev(a)lists.jboss.org
>>
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
>
>
> /max
>
http://about.me/maxandersen
> _______________________________________________
> jbosstools-dev mailing list
> jbosstools-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
--
Nick Boldt :: JBoss by Red Hat
Productization Lead :: JBoss Tools & Dev Studio
http://nick.divbyzero.com
/max
http://about.me/maxandersen