Front page | perl.perl5.porters |
Postings from August 2011
perlbug@plan9.de wrote: >If stack changes are indeed a problem, then the problem cannot be in >encoding.xs, but is likely in perlio.c, which calls load_module without >PUSHSTACK/POPSTACK, or in some other layer. There are many places that are liable to hold lurking stack-reallocation bugs. Stacks don't move very often, so a tacit assumption that the stack won't move will rarely bite. We could do with a debugging option that forcibly moves the stack at every opportunity (every time EXTEND is called). That would go nicely with valgrind. -zeframThread Previous