Re: re bugs

[Andrew Bennetts <andrew-debpython@puzzling.org>]

On Sun, Nov 10, 2002 at 12:55:36PM +0200, Kai Hendry wrote:
> test.py is:
> import re
> p = re.compile('[a-z]+')
> print p
> 
> The first error message is from python2.1, python2.2 and last
> python2.3 from SID.
> 
> I think this is a serious packaging error, as it works on [GCC 2.96
> 20000731 (Red Hat Linux 7.2 2.96-109)] on linux2

It works for me, but see my comments below...

> Regards,
> -Kai Hendry
> 
> 
> hendry@bilbo:/tmp$ vim test.py

I'm not sure how this command is related to the following tracebacks.

> Traceback (most recent call last):
>   File "test.py", line 1, in ?
>     import re
> ImportError: Bad magic number in re.pyc
> 
> shell returned 1

Is there a file called "re.pyc" in the same directory as test.py?  If so,
was it possibly generated by a newer Python than the one that produced this
traceback?

> Hit ENTER or type command to continue
> Traceback (most recent call last):
>   File "test.py", line 1, in ?
>     import re
>   File "/usr/lib/python2.2/re.py", line 3, in ?
>     # If your regexps don't work well under 2.0b1, you can switch
> AttributeError: 'module' object has no attribute 'compile'
> 
> shell returned 1

That traceback is screwy.  The line that's reporting an error doesn't have
any code on it.  I suspect that is also caused by your mysterious re.pyc
with the bad magic number.

-Andrew.