[PEAK] peak test errors

Phillip J. Eby pje at telecommunity.com
Tue Nov 30 17:25:35 EST 2004


At 05:06 PM 11/30/04 -0500, Phillip J. Eby wrote:
>At 03:00 PM 11/30/04 -0500, R. David Murray wrote:
>>Sorry for the long post, but this seems to be the easiest way to answer
>>your question.  Note that I deleted my peak directory in site-packages
>>and redid the install before running the tests, just to make sure.
>>
>>Maybe it's a python2.4 issue?  I'm running a3 at the moment.
>
>The KeyError:313 looks like there must be a new syntax production for 
>tuples in the 2.4 grammar.  I'll take a closer look as soon as I can.

It was generator expressions; even though we're not parsing generator 
expressions, the construct for tuples passes through a state where it 
*could* be looking at a generator expression.  I've got a temporary 
should-fix patch in place now.


>>======================================================================
>>ERROR: testContextAttrs (peak.web.tests.test_environ.TestTraversals)
>>----------------------------------------------------------------------
>>Traceback (most recent call last):
>>   File 
>> "/usr/local/lib/python2.4/site-packages/peak/web/tests/test_environ.py", 
>> line 142, in testContextAttrs
>>     self.assertRaises(TypeError, lambda: self.getPath('/').traversedURL)
>>...
>>AttributeError: 'NoneType' object has no attribute 'startswith'
>
>Hrm.  Looks like there's been a change to posixpath.join; I'll have to see 
>what's up with that.

The change is just that it now raises AttributeError instead of TypeError; 
I've changed the test to allow either.

Anyway, all the problems you reported should now be fixed in CVS, but I 
haven't actually tested them against Python 2.4, so please let me know if 
anything is still broken.




More information about the PEAK mailing list