On Tue, Oct 05, 2004 at 12:42:10AM -0000, John Gardiner Myers wrote:
> The following test program crashes on a multiprocessor box.
...
> Perl #31838 has had remarkably little activity for a reproducable
> crasher with a proposed fix. Could someone with more knowledge of Perl
> internals review and comment on my proposed fix?
Sorry - your followup with the patch doesn't seem to have made it to the
list (or to my inbox, anyway).
I havn't looked closely at this yet, but surely nulling PL_curpm will
cause $1 etc to get lost across win32 forks, ie
/(.)/;
if (fork() == 0) {
# $1 now undef
}
Dave.
--
In my day, we used to edit the inodes by hand. With magnets.
