[PEAK] access violation with py24/msvc compiler

[alexander smishlajev]

Phillip J. Eby wrote, at 30.01.2005 20:58:
> FYI, this issue is now moot; I have removed all uses of kjbuckets in 
> PEAK, and have changed setup.py to only build kjbuckets if it is 
> explicitly requested to do so (as a convenience for people with other 
> code that depends on it).

thank you.  access violation has gone.

there is new test failure, but perhaps you are already aware of it since 
it appears both in msvc and mingw32 builds:

======================================================================
FAIL: Doctest: Graph.txt
----------------------------------------------------------------------
Traceback (most recent call last):
   File "E:\Python24\Lib\site-packages\peak\util\doctest.py", line 2163, 
in runTest
     raise self.failureException(self.format_failure(new.getvalue()))
AssertionError: Failed doctest test for Graph.txt
   File "E:\python24\lib\site-packages\peak\util\Graph.txt", line 0

----------------------------------------------------------------------
File "E:\python24\lib\site-packages\peak\util\Graph.txt", line 23, in 
Graph.txt
Failed example:
     Graph([(1,2,3)])
Expected:
     Traceback (most recent call last):
     ...
     ValueError: unpack tuple of wrong size
Got:
     Traceback (most recent call last):
       File "E:\Python24\Lib\site-packages\peak\util\doctest.py", line 
1254, in __run
         compileflags, 1) in test.globs
       File "<doctest Graph.txt[4]>", line 1, in ?
         Graph([(1,2,3)])
       File "E:\Python24\Lib\site-packages\peak\util\Graph.py", line 10, 
in __init__
         for k,v in iterable:
     ValueError: too many values to unpack


----------------------------------------------------------------------
Ran 819 tests in 5.610s

FAILED (failures=1)

best wishes,
alex.