On 24 Jul, Pete French punched keys in this particular order: > Its not bogus - the trouble is that you cant tell at the time malloc returns > whether the pointer will be useable or not. You only find that out when > you try and use it, and whether theres any space or not depends oon what > else may have munched up (or released) memory between you making the call > to malloc() and actually writing to the location returned.
This looks similar to the problem with mktemp(). Maybe it is possible to solve this in a similar way? For example, by allocating memory and filling it in the same call? That probably would mean that all the software should switch to a new way of allocating memory, but it's a start... -- Alban Hertroys http://solfertje.student.utwente.nl - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - This person has performed an illegal operation and will be shot down. _______________________________________________ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "[EMAIL PROTECTED]"
