gnustep-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: GNUstep apps segfault on AMD64 (solved!)


From: Tim Schmielau
Subject: Re: GNUstep apps segfault on AMD64 (solved!)
Date: Sat, 28 Jun 2008 19:37:55 +0100

On 22 Jun 2008, at 20:03, Richard Frith-Macdonald wrote:
On 20 Jun 2008, at 23:50, Gregory John Casamento wrote:

I used SuSE for a while and had a similar issue. I thought that, by now, SuSE would have resolved it's 64 bit issues, but I see that's not the case. *sighs*

GNUstep works fine on Debian when compiled 64-bit, so this is definitely something that is distro-specific. I'll see what I can do to test this and determine what can be done.

On my Debian 64bit intel ffcall works fine but ffi does not.
I'm not sure that libffi works correctly on 64bit intel for anyone.


Thanks a lot for today's svn commit 26723! It's log message got my attention, and indeed a 64 bit installation from SVN using libffi solved my problem. Gorm as well as a program written by myself now run on both machines without problems.

On 26 Jun 2008, at 16:32, Manuel Guesdon wrote:
On Thu, 19 Jun 2008 11:47:02 +0100 "Schmielau, Tim" <address@hidden > wrote:

| I have two quad core 64 bit Xeon machines with SLES 10 sharing the same home directory where a fresh GNUstep installation (startup 0.20.0), gcc-4.2.4 and ffcall libraries reside. On one of the machines, GNUstep runs fine, on the other one every GNUstep application I've tried so far segfaults.

I've got a problem with nx flagged Xeons.
See http://article.gmane.org/gmane.comp.lib.gnustep.devel/6334
Problem was occuring when configuring ffcal under a normal (not root) user.

That sounds very similar to my problem. I had indeed seen your posting when searching the web for solutions, but not fully understood it. I have no root access on those machines, thus indeed had to configure ffcal as normal user. Still I got the same config.h settings as you described.

For comparison, I tried a home dir installation on one of the other machines that had always worked, using no root access at all (of course having removed the original /usr/GNUstep dir). This one also worked.

Anyways, as the problem is solved now, I'm not going to investigate further. Thanks to anybody who contributed, and especially to Richard Frith-Macdonald for the fix in svn!

Tim





reply via email to

[Prev in Thread] Current Thread [Next in Thread]