Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT sources DOT redhat DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin AT sources DOT redhat DOT com>
List-Help: <mailto:cygwin-help AT sources DOT redhat DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner AT sources DOT redhat DOT com
Delivered-To: mailing list cygwin AT sources DOT redhat DOT com
From: "Gerrit P. Haase" <gp AT familiehaase DOT de>
Organization: Esse keine toten Tiere
To: cygwin AT cygwin DOT com
Date: Tue, 9 Oct 2001 23:43:15 +0200
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: Perl 5.7.2
Reply-to: cygwin AT cygwin DOT com
Message-ID: <3BC38B93.11178.2FA66410@localhost>
In-reply-to: <3BC30401.2D09A965@rowman.com>
X-mailer: Pegasus Mail for Win32 (v3.12cDE)
X-Hops: 1
X-Sender: 320081107336-0001 AT t-dialin DOT net

John Peacock schrieb am 2001-10-09, 10:04:

>The problem I am seeing under Cygwin is an aberration, both for Perl 
>and Cygwin.  I had clean debug builds as late as last month.  It only 
>has problems under Perl-Built-For-Debugging (and three minor test 
>failures I have not gotten around to digging into).  Almost all tests 
>succeed (even with PBFD); the cores only happen when exiting the 
>program (during cleanup apparently).

Can you figure out when the stackdumps occur the first time?

I found this in the changes of July:
____________________________________________________________________________
[ 11364] By: jhi                                   on 2001/07/13  21:21:23
        Log: Move both usemymalloc and d_bincompat5005 to the Platfom
             section from the Compiler section.
     Branch: perl
	   ! myconfig.SH
____________________________________________________________________________
[ 11188] By: jhi                                   on 2001/07/07  15:12:36
        Log: Subject: Re: bleedperl perl_alloc..free leakage
             From: Radu Greab <radu AT netsoft DOT ro>
             Date: Fri, 6 Jul 2001 19:48:54 +0300
             Message-ID: <15173 DOT 60406 DOT 308246 DOT 456184 AT ix DOT netsoft DOT ro>
             
             Leak-be-gone.
     Branch: perl
	   ! op.c
____________________________________________________________________________
[ 11152] By: gsar                                  on 2001/07/05  00:52:57
        Log: fix the binary compatibility issue when building with/without
             usemymalloc by exporting Perl_malloc() et al as simple wrappers
             around the system functions (this allows most extensions built
             using one mode to coexist with perls built in the other mode)
             
             XXX the Perl_mfree() wrapper might need to do return(free()) on
             platforms where Free_t isn't "void"
     Branch: perl
	   ! embed.h embed.pl makedef.pl proto.h util.c
____________________________________________________________________________
[ 11151] By: gsar                                  on 2001/07/05  00:42:49
        Log: perl built with USE_ITHREADS can deadlock during fork() or backticks
             since it doesn't ensure threads other than the one calling fork()
             aren't holding any locks; the fix is to use pthread_atfork() to
             hold global locks
             
             building perl with -Dusemymalloc exacerbates the problem since
             Perl_malloc() holds a mutex, and perl's exec() calls New()
             
             XXX the code in win32thread.h may be needed on platforms that have
             no pthread_atfork()
     Branch: perl
	   ! perl.c thread.h win32/win32thread.h

But they seem to be very 'old'.

Gerrit


-- 
=^..^=

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/