[Zope-dev] zope.testing 3.8 fails in mysterious ways

[Christian Theune]

On 08/01/2009 01:35 AM, Godefroid Chapelle wrote:
> Sidnei da Silva wrote:
>> On Thu, Jul 30, 2009 at 9:33 PM, Martin Aspeli<optilude+lists at gmail.com>  wrote:
>>> Unfortunately, I've got other packages that depend on a newer
>>> zope.testing (specifically, collective.testcaselayer). But I thought
>>> zope.testing aimed to be able to run any "valid" tests, so it sounds
>>> like a bug in zope.testing regardless, at least since every other test
>>> I've run in the same instance work fine.
>>
>> The traceback you pasted shows an UnboundLocalError. What about
>> looking at the source and figuring out why that local variable is not
>> defined?
>>
>> -- Sidnei
>
> I bumped into the same bug while working on the same package...
>
> I allowed myself to add the missing assignment.
> However, I'd like someone to check over my shoulder.
> http://mail.zope.org/pipermail/checkins/2009-July/036586.html
>
> Further, this bug is triggered by another one happening when running
> tests with buildout in plone.z3cform current trunk (r102411).
>
> The test runner is trying to spawn a subprocess with the --resume-layer
> argument.  However, the --resume-layer argument is not accepted by the
> subprocess.
>
> This produces an error in a format not foreseen by the error parser.
> (which then triggered the UnboundLocalError).
>
> It would be nice if someone with a better understanding of the
> testrunner could take a look and understand why the --resume-layer stuff
> happens. Christian ?

Can you make a bug out of that? I won't be able to look at it for a 
while but I'd also not have it forgotten.

Christian

-- 
Christian Theune · ct at gocept.com
gocept gmbh & co. kg · forsterstraße 29 · 06112 halle (saale) · germany
http://gocept.com · tel +49 345 1229889 0 · fax +49 345 1229889 1
Zope and Plone consulting and development