[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Problem with libart backend
From: |
Charles Philip Chan |
Subject: |
Re: Problem with libart backend |
Date: |
Sat, 26 Oct 2002 12:31:51 -0400 |
Circa Sat. Oct. 26, 2002 at 16:31:04 +0200, a lone cry was heard from
Alexander Malmberg <alexander@malmberg.org> in the wasteland called the
Internet:
> Sounds like the shared memory code is failing without noticing it.
> Which version of GNUstep is this? What system? There was a similar
> problem on some systems if a segment was marked for destruction too
> soon, but it was fixed 2002-08-23.
I am on a x86 Linux system with kernel 2.4.19. The version of GNUstep is
cvs dated today.
> Some things worth trying are (in back/Source/x11/XWindowBuffer.m):
> * line 174, change 0700 to 0777 (a potential security problem, though)
This worked.
> * line 226, remove the shmctl() call (will leak shared segments;
> remove them with ipcs or equivalent)
This doesn't work.
> As a temporary workaround, you can comment out the if
> on line 147 so the non-shm code always gets used instead.
Didn't try that since I would rather use shm on a local machine.
Charles
--
"Whip me. Beat me. Make me maintain AIX."
(By Stephan Zielinski)
- Problem with libart backend, Charles Philip Chan, 2002/10/25
- Re: Problem with libart backend, Jeff Teunissen, 2002/10/25
- Re: Problem with libart backend, Charles Philip Chan, 2002/10/26
- Re: Problem with libart backend, Alexander Malmberg, 2002/10/26
- Re: Problem with libart backend,
Charles Philip Chan <=
- Re: Problem with libart backend, Alexander Malmberg, 2002/10/26
- Re: Problem with libart backend, Charles Philip Chan, 2002/10/26
- Re: Problem with libart backend, Alexander Malmberg, 2002/10/27
- Re: Problem with libart backend, Charles Philip Chan, 2002/10/27
gnustep-resources (Was: Re: Problem with libart backend), Stefan Urbanek, 2002/10/26