[Python-Dev] valgrind and python?

[Marcelo Matus]

hmmm... I tried what you said:

- I used the last cvs 2.2 version (cvs upd -r release22-maint)

- I never load my module

- I used gcc 3.1.1

- I even passed the option --workaround-gcc296-bugs=yes


but still I'v got the same kind of error report.....
maybe the compiler is too new, but I can go back to
gcc 2.96 (because of the c++ support).

So, in the meantime, I will just ignore the reports......


Thanks

Marcelo.



Neal Norwitz wrote:

>Marcelo Matus wrote:
>  
>
>>Did somebody already test python 2.2.1 using valgrind 1.0.1?
>>    
>>
>
>I have used valgrind on python, but mostly the latest CVS version,
>not 2.2.1.  valgrind and python were built with gcc 2.96 (redhat 7.2).
>I also use purify from time to time when it works (pretty rare).
>
>  
>
>>because testing one of my own modules, I get the following
>>recurrent error report whenever I import something:
>>
>>==19951== Conditional jump or move depends on uninitialised value(s)
>>==19951==    at 0x8094B85: find_module (in
>>/home/mmatus/oss2/gcc3/bin/python)
>>==19951==    by 0x8095DE2: import_submodule (Python/import.c:1887)
>>==19951==    by 0x80959B8: load_next (Python/import.c:1752)
>>==19951==    by 0x8097608: import_module_ex (Python/import.c:1603)
>>
>>and  I don't know if I can ignore it or if this is a real python error.
>>    
>>
>
>Unlikely a python problem.  I just tried valgrind with the current
>CVS version of 2.2.1+ (what will eventually become 2.2.2).
>There were no problems reported doing import math.
>
>You can try several things to fix the warning:
> * run without your module
> * use a different compiler (version)
> * run the CVS version: cvs upd -r release22-maint
> * pass --workaround-gcc296-bugs=yes option to valgrind
>
>Neal
>  
>