01-07-11 18.20, skrev Doug MacEachern på [EMAIL PROTECTED] följande:
the following program does not leak with 5.6.1, with bleedperl leaks,
example:
% ./clone_leak -MCGI -e0
Scalars leaked: -4
Scalars leaked: 4
PID VSZ COMMAND
3673 4428 clone_leak
I have seen this this problem on and
Ok, I am so totally lost with regards to weakrefs.
It seems like a weakref realy also is a weakref to itself because the
referants backref magic array is a list of the weakrefs, but they are not
refcounted.
However this breaks badly with regard to duplicating the tree because the
refcount of
This patch changes shared to unique after Dougs suggestion. (Sorry I had lost that
email and couldn't remeber if it was my
imaginiation or real).
Is this documented anywhere? On another note, how the heck are we refcounting these?
It doesn't seem to be atomic refcounts.
Artur
diff -ur
01-06-23 15.54, skrev Jarkko Hietaniemi på [EMAIL PROTECTED] följande:
Is the patch to pp_regcomp from Richard Soderberg realy applied?
In my local copy it is applied and the 634 test passes on threaded win32 and
unix!
This one?
Index: perl/pp_ctl.c
--- perl/pp_ctl.c.~1~ Sat Jun
01-06-10 09.55, skrev Kirill E. Shpitsa på [EMAIL PROTECTED]
följande:
Hi!
Consider the following snippet running on Win32 platform:
use Win32::Event;
$event=Win32::Event-new(1,1,'');
# waitpid((fork() or exit),0);
print waiting...;
$event-wait() or die $^E;
print done\n;
not
Hi,
I have understood that the current ithread implmentation is unsafe because
of the regex engine. The only notes I could find on this is ActiveStates
release notes on the pseudo forking.
My very limited understand of the issue is that the regex engine changes the
op tree for certain cases
01-05-04 18.31, skrev Dan Sugalski på [EMAIL PROTECTED] följande:
At 06:27 PM 5/4/2001 +0200, Artur Bergman wrote:
ITHREADS has nothing to do with mutexes. It's there to emulate fork on
Win32. That it can be used for other things is a reasonably nice side
effect, but as Alan has pointed