Hi,
I am using perl embedded on C + wxWidgets with a threading mechnism.
If a thread is never join, I can see clearly the amount of memory growing after
each thread run.
However, if it is joined, the amount of memory used gets freed.
I had checked that on Windows 2000.
Kind regards,
Jean-Phil
Chris Wagner wrote:
>> I'm not sure. Ending thread on Windows deallocates memory as it said in
> MSDN, but I'm not exactly sure how Perl handles
>> all this stuff.
> So that could be Windows specific? The application I made was on Solaris.
Yes, it could. From the other hand, once I had a scrip
On Tue, 28 Jul 2009, Chris Wagner wrote:
> At 06:48 PM 7/28/2009 +0300, Serguei Trouchelle wrote:
> > I'm not sure. Ending thread on Windows deallocates memory as it said
> > in MSDN, but I'm not exactly sure how Perl handles all this stuff.
>
> So that could be Windows specific? The application I
At 06:48 PM 7/28/2009 +0300, Serguei Trouchelle wrote:
>I'm not sure. Ending thread on Windows deallocates memory as it said in
MSDN, but I'm not exactly sure how Perl handles
>all this stuff.
So that could be Windows specific? The application I made was on Solaris.
I actually thought it was a m
Chris Wagner wrote:
>> Or, as your question partially suggests, use threads: ending a thread will
> release the memory back to OS.
>
> Really?
Yes, here's an example (takes about 200M of memory and releases it):
_
#!/usr/bin/perl -w
use strict;
use warnings;
use threads;
$| = 1;
su
At 01:31 AM 7/25/2009 +0300, Serguei Trouchelle wrote:
>Or, as your question partially suggests, use threads: ending a thread will
release the memory back to OS.
Really? Is that documented anywhere? Knowing that could've saved me a lot
of trouble on a massively threaded long running application
Hi all,
I'm using ActiveState Perl 5.10. Is there any way to increase pipe's
buffer size? By running the included sample script, I found that (at
least at Windows Server 2003) it is limited to 512 bytes. I am aware
that Win32::Pipe has a method which allows to alter the buffer size, but
it use
Al Mollenkopf wrote:
> working on a server/client that encrypts payloads (in chunks) for
> transmission, trying to profile payloads prior to sending but puzzled by no
> performance hit by adding compression.
> Question: Anyone have any insight why there virtually no loss of processing
> perfor