Re: Valgrind and Python

on 11/10/2007 08:38 AM Esa A E Peuha wrote :
Running Python 2.5.1 under Valgrind is interesting; just starting it and
then pressing Ctrl-D produces this:

==27082== ERROR SUMMARY: 713 errors from 56 contexts (suppressed: 10 from 1)
==27082== malloc/free: in use at exit: 1,243,153 bytes in 508 blocks.
==27082== malloc/free: 3,002 allocs, 2,494 frees, 2,748,487 bytes allocated.
==27082== For counts of detected errors, rerun with: -v
==27082== searching for pointers to 508 not-freed blocks.
==27082== checked 1,399,984 bytes.
==27082== ==27082== LEAK SUMMARY:
==27082== definitely lost: 0 bytes in 0 blocks.
==27082== possibly lost: 17,072 bytes in 58 blocks.
==27082== still reachable: 1,226,081 bytes in 450 blocks.
==27082== suppressed: 0 bytes in 0 blocks.
==27082== Reachable blocks (those to which a pointer was found) are not shown.
==27082== To see them, rerun with: --show-reachable=yes

A lot of those 713 errors occur in the various deallocation functions.


really wired!!! i tried valgrind --leak-check=yes python on Python 2.4.3

it showed me even 839 errors,

==4375== ERROR SUMMARY: 839 errors from 108 contexts (suppressed: 38 from 1)
==4375== malloc/free: in use at exit: 710,604 bytes in 242 blocks.
==4375== malloc/free: 1,971 allocs, 1,729 frees, 1,541,112 bytes allocated.
==4375== For counts of detected errors, rerun with: -v
==4375== searching for pointers to 242 not-freed blocks.
==4375== checked 626,920 bytes.


I didn't get any error when using valgrind with perl or php!!!!

does that mean that python's interpretor is a memory beast?

.

Relevant Pages

  • Re: can someone explain the candela a little?
    ... It's just that different measurements give incompatible ... including those that the CODATA value was taken from. ... Esa Peuha ...
    (sci.physics.research)
  • Re: what makes it true?
    ... Esa A E Peuha writes: ... > The completeness theorem "if a statement is true in every model of PA, ... Prev by Date: ...
    (sci.math)
  • Re: sin x / x tends to 1...
    ... Esa A E Peuha writes: ... Oops. ... Esa Peuha ... Prev by Date: ...
    (sci.math)
  • Valgrind and Python
    ... Running Python 2.5.1 under Valgrind is interesting; ... then pressing Ctrl-D produces this: ... A lot of those 713 errors occur in the various deallocation functions. ...
    (comp.lang.python)
  • Re: what makes it true?
    ... Esa A E Peuha writes: ... It is not a consequence of the completeness theorem that PA proves ... Prev by Date: ...
    (sci.math)
Loading