RE: HANDLE leak

[Steve Croall <SCroall@staffware.com>]

Thanks.  Are there any disadvantages to running the threads detached?  Is
there any reason why this is not the default behaviour?  I wouldn't be
surprised if that was just the way it was :-)
 
Steve.  
 

-----Original Message-----
From: Geoff Stevens [ mailto:geoff.stevens@macro4.com ]
Sent: 12 October 1999 17:04
To: 'pthreads-win32@sourceware.cygnus.com'
Subject: FW: HANDLE leak



Steve, 

I think you should either create your threads detached (set
PTHREAD_CREATE_DETACHED in the attributes) or do a pthread_join for each of
them in main().

At the moment the threads you create are waiting for a join after
pthread_exit, and their Win32 threads (and thus their handles) cannot be
released.

I'm using detached threads with the same levels of the library and NT and
see no increase in handle usage after all my threads have terminated.

Best, 
Geoff Stevens   

-----Original Message----- 
From:   Steve Croall [SMTP:SCroall@staffware.com] 
Sent:   12 October 1999 16:24 
To:     geoff.stevens@macro4.com 
Subject:        HANDLE leak 

The following small program shows the HANDLE count in "Performance Monitor",

for the executable when running, increasing at an alarming rate.  Is this 
normal or am I not doing something to clear-up?  I am using the 1999-09-17 
snap-shot on Windows NT 4, service pack 4. 

#include <stdio.h> 
#include <stdlib.h> 
#include <conio.h> 

#include "pthread.h" 

void *thrd_func (void *i) 
{ 
        Sleep (500); 

        pthread_exit ((void *) 1); 
} 

int main (int argc, char **argv) 
{ 
        int             i = 0; 
        pthread_t       t_id; 

        while (!_kbhit ()) 
        { 
                Sleep (1000); 
                pthread_create (&t_id, NULL, (void *) thrd_func, NULL); 
        } 

        Sleep (1000); 
} 


Regards, 

        Steve Croall ( SCroall@staffware.com ) 


Chicago law prohibits eating in a place that is on fire.