[PEAK] Current cvs peak errors

Phillip J. Eby pje at telecommunity.com
Wed Dec 1 11:09:56 EST 2004

At 11:25 AM 12/1/04 +0100, Alain Poirier wrote:

>Here I've got these 2 errors (python 2.3.4, fresh peak install from CVS
>today) :

Thanks for the report.

>ERROR: checkSchemaLookup (peak.config.tests.ZConfigTests)
>ValueError: unknown url
>type: /usr/lib/python2.3/site-packages/ZConfig/doc/schema.dtd

This one appears to be that XML problem I was talking about, that appears 
to be broken in Python 2.3 and fixed in Python 2.4 (since the 2.4 users 
haven't reported it, and I don't get it in 2.2).  I believe it's due to a 
Python bug regarding system ID vs public ID in XML processing.  I should 
probably track it down further and put in some version checking code to try 
to work around it, although I don't know if it's actually possible without 
patching ZConfig.

The error is relatively harmless, however, as it only breaks XML files with 
DTD (DOCTYPE) declarations, which I only have in there so that jEdit will 
be able to do syntax highlighting, tag completion, and DTD validation when 
I'm editing the XML.  That's why I haven't bothered pursuing it before.

>FAIL: testHeaderFormats (wsgiref.tests.test_handlers.HandlerTests)
>Traceback (most recent call last):
>   File "/usr/lib/python2.3/site-packages/wsgiref/tests/test_handlers.py",
>line 203, in testHeaderFormats
>     (stdpat%(version,sw), h.stdout.getvalue())
>   File "/usr/lib/python2.3/unittest.py", line 278, in failUnless
>     if not expr: raise self.failureException, msg
>AssertionError: ('HTTP/1.0 200 OK\\r\\nDate: \\w{3} \\w{3} \\d{2} \\d\\d:
>\\d\\d:\\d\\d \\d{4}\\r\\nServer: FooBar/1.0\r\nContent-Length:
>0\\r\\n\\r\\n', 'HTTP/1.0 200 OK\r\nDate: Wed Dec  1 10:12:34
>2004\r\nServer: FooBar/1.0\r\nContent-Length: 0\r\n\r\n')

It looks like the date formatting in 2.3/2.4 is also different from 2.2; 
this is looking for "Dec 01" rather than "Dec  1".  I'll modify the regular 
expression accordingly.  Luckily, this is just a too-brittle test, rather 
than an actual problem with the code.

More information about the PEAK mailing list