classification
Title: pprint.pformat requires __repr__
Type: Stage:
Components: None Versions:
process
Status: closed Resolution: fixed
Dependencies: Superseder:
Assigned To: Nosy List: cmoad, dkuhlman, doerwalter (3)
Priority: normal Keywords

Created on 2004-11-12 21:36 by dkuhlman, last changed 2005-08-10 07:07 by doerwalter.

Files
File name Uploaded Description Edit Remove
screen3.txt dkuhlman, 2004-11-12 21:49
Messages (6)
msg23117 - (view) Author: Dave Kuhlman (dkuhlman) Date: 2004-11-12 21:36 
Not having a __repr__ attribute seem to cause problems for:

    pprint.pformat

in Python 2.4.  This problem goes away when I use a Zope instance for 
Python 2.3.

The error message is generated when I try to access a private
workspace in the CPS interface.

I inserted a bit of debugging code, and I've attached a file that
contains the debugging print-out of the data that caused the error
and also the traceback.

I'm using:

    CPS-3.2.1
    Zope-2.7.3-0
    Python 2.4b2 built from source

on Linux.

I've attached a file containing a bit of debugging information and also 
the traceback.

Dave
msg23118 - (view) Author: Walter Dörwald (doerwalter) Date: 2004-11-15 13:53 
Logged In: YES 
user_id=89016

This should be fixed in pprint.py 1.31 in current CVS. Thanks 
for the report.
msg23119 - (view) Author: Charles Moad (cmoad) Date: 2005-08-08 19:08 
Logged In: YES 
user_id=705459

This problem affects matplotlib still.  Here is a short example:

import pprint
from matplotlib.transforms import unit_bbox
b = unit_bbox()
pprint.pformat(b)
msg23120 - (view) Author: Walter Dörwald (doerwalter) Date: 2005-08-09 18:30 
Logged In: YES 
user_id=89016

Which version of Python are you using? Can you post a stack
trace?
msg23121 - (view) Author: Charles Moad (cmoad) Date: 2005-08-09 18:34 
Logged In: YES 
user_id=705459

This is python 2.4.1 in gentoo linux, compiled from source.
 Other python 2.4 linux reported the same affect.  I don't
know how to get a stack trace when python seg faults.

We fixed the issue in matplolib's cvs by adding __repr__
support to all the wrapped cxx objects.
msg23122 - (view) Author: Walter Dörwald (doerwalter) Date: 2005-08-10 07:07 
Logged In: YES 
user_id=89016

I didn't realize that Python segfaults. I don't think your
problem is related to this bug.

To get a stacktrace for the segfault you should compile a
debug version of Python (via configure --with-pydebug &&
make), then start gdb with it (gdb ./python) then start the
program (run), with gives you a Python prompt, where you can
import your failing code. After the segfault you should be
back in the debugger, where you can display a stacktrace via
the bt command.
History
Date User Action Args
2004-11-12 21:36:38dkuhlmancreate