[wildfly-dev] Timeout errors in JCA bits in testsuite

[Darran Lofthouse]

Every single PR build we execute risks interference.

Without running every PR again after every merge every single PR is a 
race condition hoping to avoid conflicting merges.

On 29/01/16 08:02, Carlo de Wolf wrote:
> I'm not saying, do not use snapshots. I'm saying do not use SNAPSHOT. :-)
>
> Clearly the path is a shared one, so the build must use timed snapshots
> or risk interference.
>
> Carlo
>
> On 01/28/2016 12:54 PM, Tomaž Cerar wrote:
>>
>> On Thu, Jan 28, 2016 at 9:58 AM, Carlo de Wolf <cdewolf at redhat.com
>> <mailto:cdewolf at redhat.com>> wrote:
>>
>>     There should be no SNAPSHOT dependencies in any CI build. While it
>>     does build core beforehand, it might get overwritten by another CI
>>     run.
>>     //[org.wildfly.core:wildfly-core-parent] /[INFO] Installing
>>     /opt/buildAgent/work/db872761b443af46/core/pom.xml to
>>     /store/repository/org/wildfly/core/wildfly-core-parent/2.0.8.Final-SNAPSHOT/wildfly-core-parent-2.0.8.Final-SNAPSHOT.pom
>>     /
>>     Would it be possible to use timed snapshots instead?
>>
>>
>>
>> As Stuart said, whole point of this builds is to have snapshot builds.
>> Same build agent first builds core (clean install) and than builds full
>> with -Dversion.wildfly.core=<version of wildfly core previousily built>
>>
>> snapshot used here are only build agent local, not publish on some
>> outside nexus repository.
>>
>> --
>> tomaz
>>
>
>
>
> _______________________________________________
> wildfly-dev mailing list
> wildfly-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/wildfly-dev
>