[Python-Dev] Re: python vs. valgrind (original) (raw)
Tim Peters tim.peters at gmail.com
Mon Aug 30 19:45:33 CEST 2004
- Previous message: [Python-Dev] Re: python vs. valgrind
- Next message: [Python-Dev] Volunteer for next bug day?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
[Fredrik Lundh]
(but I must be tired; can anyone explain how "making the function last" prevents the code from looking at uninitialized data?)
It doesn't. obmalloc can read uninitialized memory, and by design (see REAMDE.valgrind). Putting the function last was, as Neal said in his comment there. to prevent gcc from inlining it. That in turn is required so that the Valgrind suppression file can name the function.
- Previous message: [Python-Dev] Re: python vs. valgrind
- Next message: [Python-Dev] Volunteer for next bug day?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]