I'm not sure if I understand your use case. What is wrong about the
current implementation? Currently a build remains unstable until all new
warnings are fixed (see
https://issues.jenkins-ci.org/browse/JENKINS-5147). This is the same
behavior as for unit tests. Here also a build remains unstable until all
tests are running again.
Ulli
On 04/02/2012 11:08 AM, Johannes Wienke wrote:
> Hi,
>
> sorry for pushing this, but has anyone got an idea how to solve this?
>
> Kind regards,
> Johannes
>
> Am 15.02.2012 10:34 schrieb Johannes Wienke:
>> Hey everyone,
>>
>> we have a job which performs some custom coding style checking on some
>> files. As the result format my custom checker produces checkstyle XML
>> output.
>>
>> What we would like to have is that committers are notified by mail each
>> time they introduce a new style warning.
>>
>> I could achieve something like that by using the advanced settings of
>> the checkstyle plugin for jenkins and setting the unstable threshold for
>> new warnings to 0 ("Status thresholds (New warnings)"). This now marks
>> the build as unstable each time I commit a new warning.
>>
>> However, what confuses me is that the build remains unstable until this
>> warning is actually fixed. I would have though that the build gets only
>> unstable once and if I rebuild the project I am back to stable. I do not
>> want to use the delta computation as it might prevent mails for new
>> warnings from being sent.
>>
>> Can anyone explain this behavior to me?
>>
>> As I might produce some output with my custom checker, which could not
>> be fully compatible with the expectations of the checkstyle pluign, it
>> would also be good to know which rules are applied by the plugin to
>> decide if a warning is new or not.
>>
>> Thanks for any help.
>> Johannes
>>
>
