Exporting build results even if the build fails

[Christopher Baines]

On 29/11/15 15:47, Mattia Rizzolo wrote:
>> I tried to put some context in the commit messages, but as a more
>> concrete example, I have a Jenkins instance that runs jobs which involve
>> running pbuilder. py.test, which runs as part of the package build
>> process outputs a file (containing jUnit XML), this file is listed as a
>> additional build result, such that it gets copied to the Jenkins working
>> directory, and then read by Jenkins for display in the web interface.
>> The above patch means that the test results will be displayed when the
>> tests fail (and the package build fails), as well as when the tests all
>> pass (and the package build succeeds).
> 
> ok, I'm open to this, even if I'm a bit worried about the change of
> behaviour (I'm a fan of adding options, but maybe it's overkill in this
> case?)
> 
> Can you please open a bug, post the patches and we'll discuss it over
> there?
> In a bug report is easier for me to track requests, etc....

Done. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=806650

> btw,
> +    if [ -d "${BUILDRESULT}" ]; then
> +        export_additional_buildresults
> +    else
> +        log.e "BUILDRESULT=[$BUILDRESULT] is not a directory."
> +    fi
> 
> I don't think this is the check you wanted to perform:
> [ -d $BUILDRESULT ] is done earlier on the same script (lines 82:92) and
> the thing already fails if that directory doesn't exists before starting
> the build.

Ok, I just copied the check as I moved the code.

Looks like the check can be removed from both places at the bottom of
that file.