[gentoo-amd64] FVWM taskbar segfault on ~amd64

[gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

I have fvwm running on three computers, two ~x86 and one ~amd64, a
dual opteron system.  All three have the same config file.

When fvm starts on the dual opteron system, the taskbar does not show
up, and this message appears in the system log:

    fvwm2[17794]: segfault at 7fffa5878818 rip 4299a1 rsp 7fffa47c8ae0 error 4

Fvwm otherwise works fine.

Anyone else run fvwm on their ~amd64 systems?

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[Jesús Guerrero]

On Wed, 20 Feb 2008 21:24:35 -0800
felix@crowfix.com wrote:

> I have fvwm running on three computers, two ~x86 and one ~amd64, a
> dual opteron system.  All three have the same config file.
> 
> When fvm starts on the dual opteron system, the taskbar does not show
> up, and this message appears in the system log:
> 
>     fvwm2[17794]: segfault at 7fffa5878818 rip 4299a1 rsp 7fffa47c8ae0 error 4
> 
> Fvwm otherwise works fine.
> 
> Anyone else run fvwm on their ~amd64 systems?

There's a big chance that you are going to have problems
with amd64 if you use anything below 2.5.16, but, besides that, you
should really be using 2.5.24, which is the most stable version.

What version of fvwm are you using?

-- 
Jesús Guerrero <i92guboj@terra.es>
--
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Thu, Feb 21, 2008 at 06:41:22AM +0100, Jes?s Guerrero wrote:

> There's a big chance that you are going to have problems
> with amd64 if you use anything below 2.5.16, but, besides that, you
> should really be using 2.5.24, which is the most stable version.
> 
> What version of fvwm are you using?

2.5.21, the latest ebuild.  I know it's behind the fvwm site, but
other than install outside of portage, there's not much to do about
that.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[Jesús Guerrero]

On Thu, 21 Feb 2008 08:26:19 -0800
felix@crowfix.com wrote:

> On Thu, Feb 21, 2008 at 06:41:22AM +0100, Jes?s Guerrero wrote:
> 
> > There's a big chance that you are going to have problems
> > with amd64 if you use anything below 2.5.16, but, besides that, you
> > should really be using 2.5.24, which is the most stable version.
> > 
> > What version of fvwm are you using?
> 
> 2.5.21, the latest ebuild.  I know it's behind the fvwm site, but
> other than install outside of portage, there's not much to do about
> that.
> 

Well, I made an ebuild for .23 and .24 long ago, though no one seems to care
about fvwm users for quite a long time :P

http://bugs.gentoo.org/show_bug.cgi?id=191638

You can also use the live-ebuild in the devnull overlay (use layman for that).
Both ebuilds are mine and have been around for a long time. In fvwm, usually, 
the latest version (even cvs) is usually more stable than any previous one.

Fancy stuff is rarely committed, so, almost all commits are bugfixes,
and that 's why it can only get more and more stable. If you have some problem
with any of these ebuilds, just ask. Though they have been tested for a long
time and should work painlessly.

Regards.
-- 
Jesús Guerrero <i92guboj@terra.es>
--
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Thu, Feb 21, 2008 at 07:44:49PM +0100, Jes?s Guerrero wrote:

> Well, I made an ebuild for .23 and .24 long ago, though no one seems to care
> about fvwm users for quite a long time :P
> 
> http://bugs.gentoo.org/show_bug.cgi?id=191638
> 
> You can also use the live-ebuild in the devnull overlay (use layman for that).
> Both ebuilds are mine and have been around for a long time. In fvwm, usually, 
> the latest version (even cvs) is usually more stable than any previous one.
> 
> Fancy stuff is rarely committed, so, almost all commits are bugfixes,
> and that 's why it can only get more and more stable. If you have some problem
> with any of these ebuilds, just ask. Though they have been tested for a long
> time and should work painlessly.

I've not yet used overlays or layman; can I beg for a quick summary
for dummies? :-) Unfortunately, I am away from that machine for the
next week or so, and won't be able to report success or not until I
get access to it again.

Back when I ran slackware, there were some packages which I always
updated immediately because they were always good code: fvwm and
postgresql were among them.  I know postgresql lags some, which is
annoying, but I had no idea fvwm was so out of date.  Should I add a
requtes to that bug for moving it into the main ebuilds, or is there a
better way of requesting that?

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[Jesús Guerrero]

On Thu, 21 Feb 2008 11:04:21 -0800
felix@crowfix.com wrote:

> On Thu, Feb 21, 2008 at 07:44:49PM +0100, Jes?s Guerrero wrote:
> 
> > Well, I made an ebuild for .23 and .24 long ago, though no one seems to care
> > about fvwm users for quite a long time :P
> > 
> > http://bugs.gentoo.org/show_bug.cgi?id=191638
> > 
> > You can also use the live-ebuild in the devnull overlay (use layman for that).
> > Both ebuilds are mine and have been around for a long time. In fvwm, usually, 
> > the latest version (even cvs) is usually more stable than any previous one.
> > 
> > Fancy stuff is rarely committed, so, almost all commits are bugfixes,
> > and that 's why it can only get more and more stable. If you have some problem
> > with any of these ebuilds, just ask. Though they have been tested for a long
> > time and should work painlessly.
> 
> I've not yet used overlays or layman; can I beg for a quick summary
> for dummies? :-) Unfortunately, I am away from that machine for the
> next week or so, and won't be able to report success or not until I
> get access to it again.
> 
The quick picture is "emerge layman". Then you just need to edit your
make.conf and add this at the end of the file. Be sure you put it at
the end so it doesn't get overwritten by accident.

======cut here=========
source /usr/portage/local/layman/make.conf
=====/cut here=========

Then update the list of overlays with layman -f. You can now list
the overlays with layman -L, the installed ones with -l, and you
can add a new one with layman -a <overlay>. For example, to get
the devnull overlay, you would do "layman -a devnull".

From now on, you just need to run "emerge -pv fvwm" and you should
see that fvwm is going to be updated to 9999, which means that is
a live ebuild (based on cvs, svn, git or whatever).

Make sure you update your layman list regularly with -f, and that you
sync the installed overlays with "layman -S". To delete overlays that
you will no longer use, you can use -d.

> Back when I ran slackware, there were some packages which I always
> updated immediately because they were always good code: fvwm and
> postgresql were among them.  I know postgresql lags some, which is
> annoying, but I had no idea fvwm was so out of date.  Should I add a
> requtes to that bug for moving it into the main ebuilds, or is there a
> better way of requesting that?

You can add another petition at the end of that thread like many
people already did if you want... I don't know if it is worth it, though. :(

-- 
Jesús Guerrero <i92guboj@terra.es>
--
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[Sebastian Redl]

Jesús Guerrero wrote:
> Then update the list of overlays with layman -f. You can now list
> the overlays with layman -L, the installed ones with -l, and you
> can add a new one with layman -a <overlay>. For example, to get
> the devnull overlay, you would do "layman -a devnull".
>   
layman -f is deprecated. Just do a layman -L and it will update
automatically.

Note, however, that it only lists overlays that you have the necessary
tools for, i.e. an overlay managed in a Subversion repository is only
shown if you have Subversion installed, and so on. The most popular
management tools for overlays are Subversion and GIT, so be sure to
merge them, too. There's a few more: rsync is already installed, of
course, then there's Mercurial, Bazaar, and I think the Haskell overlay
is even done in Darcs.

Sebastian Redl
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Thu, Feb 21, 2008 at 08:42:42PM +0100, Jes?s Guerrero wrote:
> From now on, you just need to run "emerge -pv fvwm" and you should
> see that fvwm is going to be updated to 9999, which means that is
> a live ebuild (based on cvs, svn, git or whatever).

I see it sets up more than just fvwm, such as mplayer.  Is there some
way to tell emerge to ignore the devnull overlay except for fvwm?  I
know I can specify the exact version as in <=mplayer-1.0_rc2_p25993,
but I'd have to do that for every package which has an overlay in
devnull.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Thu, Feb 21, 2008 at 08:42:42PM +0100, Jes?s Guerrero wrote:
> Then update the list of overlays with layman -f. You can now list
> the overlays with layman -L, the installed ones with -l, and you
> can add a new one with layman -a <overlay>. For example, to get
> the devnull overlay, you would do "layman -a devnull".

Oh boy, looks like some interesting overlays there.  I can see only
too easily how I could have fun and get into trouble exploring them.
You will be the ruin of me! :-)

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[Jesús Guerrero]

On Thu, 21 Feb 2008 13:54:14 -0800
felix@crowfix.com wrote:

> On Thu, Feb 21, 2008 at 08:42:42PM +0100, Jes?s Guerrero wrote:
> > Then update the list of overlays with layman -f. You can now list
> > the overlays with layman -L, the installed ones with -l, and you
> > can add a new one with layman -a <overlay>. For example, to get
> > the devnull overlay, you would do "layman -a devnull".
> 
> Oh boy, looks like some interesting overlays there.  I can see only
> too easily how I could have fun and get into trouble exploring them.
> You will be the ruin of me! :-)

Indeed, it's dangerous if you don't know what you are doing.

Most overlays are safe, but there are several that could be a big pain if
misused.

About the fvwm thing... well, the easiest thing I can think of would be
this: don't use layman. Sync the overlay into your home dir, for example:

cd ~
mkdir svn
cd svn
# for the next command you need subversion, emerge it if
# you haven't done it yet
svn http://svn.engelkotzen.net/public/devnull

Now you have the overlay in your home, but it is not too useful there. So,
you need to make your personal overlay, edit /etc/make.conf, and add this
BEFORE the other line I gave you above:

PORTDIR_OVERLAY="${PORTDIR}/local"

Now, this:

cd /usr/portage
mkdir local
cd local
mkdir x11-wm
cd x11-wm
ln -s /home/i92guboj/svn/devnull/x11-wm/fvwm .

Now, you have your own personal overlay, with only one package, but that
package is into your home dir. The contents of your /usr/portage/local/x11-wm/fvwm
overlay will be a link to the real contents in your home.

What is the advantage of this? Well, portage now can see your new fvwm package,
and you can update the devnull overlay with a simple command, but the rest of
package into devnull will not interfere with your system.

To update the overlay, you do

cd ~/svn/devnull
svn update

And you will have the latest stuff in that overlay (but only the fvwm stuff, or
whatever you link into your local overlay will be used.

You can as well just create the /usr/portage/local/x11-wm/ dir, set up the make.conf
as I told you above, and then cd into /usr/portage/local/x11-wm/ and then 
do "svn co http://svn.engelkotzen.net/public/devnull/x11-wm/fvwm" to update
only that little bit of the tree.

This second way can be more convenient because you save some bandwidth, since
you only download the stuff that you want.

As you see, you have many alternatives. The problem with either of this,
is that you are the responsible to update the ebuild when necesary via svn.

If you have questions, just ask.
-- 
Jesús Guerrero <i92guboj@terra.es>

--
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Thu, Feb 21, 2008 at 11:40:29PM +0100, Jes?s Guerrero wrote:

> As you see, you have many alternatives. The problem with either of this,
> is that you are the responsible to update the ebuild when necesary via svn.
> 
> If you have questions, just ask.

I just may do that :-)  I will see how the new fvwm works probably
next Wednesday, and go from there.

Sure appreciate all the hand holding.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Thu, Feb 21, 2008 at 03:00:50PM -0800, felix@crowfix.com wrote:
> On Thu, Feb 21, 2008 at 11:40:29PM +0100, Jes?s Guerrero wrote:
> 
> > As you see, you have many alternatives. The problem with either of this,
> > is that you are the responsible to update the ebuild when necesary via svn.
> > 
> > If you have questions, just ask.
> 
> I just may do that :-)  I will see how the new fvwm works probably
> next Wednesday, and go from there.

Well, no luck.  taskbar still segfaults:

[  165.614093] FvwmTaskBar[14701]: segfault at fffffff7 rip 2b7c4a6c52b5 rsp 7fff637048c0 error 4

[ebuild   R   ] x11-wm/fvwm-9999  USE="bidi doc gtk imlib nls perl png readline session stroke svg tk truetype xinerama xpm -debug -extras -rplay -shape -translucency" 0 kB [1]

I noticed that the fvwm overlay version is not very useful.  Mplayer,
for instance, which is also part of the devnull overlay, has a useful
version:

[ebuild   R   ] media-video/mplayer-2008.0103

fvwm --help is only slightly more useful:

fvwm 2.5.25 (from cvs) compiled on Feb 21 2008 at 20:06:03
with support for: ReadLine, Stroke, XPM, PNG, SVG, XShm, SM, Bidi text, Xinerama, XRender, XCursor, XFT, NLS

How does it track the version internally well enough to know when
updates are available?

And back to fvwm, should I report this to the gentoo maintainer (who
seems to be seriously out of date) or to the fvwm project -- does
gentoo add its own bells and whistles which would make a gentoo bug
report less useful?

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[Jesús Guerrero]

On Thu, 28 Feb 2008 00:33:11 -0800
felix@crowfix.com wrote:

> On Thu, Feb 21, 2008 at 03:00:50PM -0800, felix@crowfix.com wrote:
> > On Thu, Feb 21, 2008 at 11:40:29PM +0100, Jes?s Guerrero wrote:
> > 
> > > As you see, you have many alternatives. The problem with either of this,
> > > is that you are the responsible to update the ebuild when necesary via svn.
> > > 
> > > If you have questions, just ask.
> > 
> > I just may do that :-)  I will see how the new fvwm works probably
> > next Wednesday, and go from there.
> 
> Well, no luck.  taskbar still segfaults:
> 
> [  165.614093] FvwmTaskBar[14701]: segfault at fffffff7 rip 2b7c4a6c52b5 rsp 7fff637048c0 error 4
> 
> [ebuild   R   ] x11-wm/fvwm-9999  USE="bidi doc gtk imlib nls perl png readline session stroke svg tk truetype xinerama xpm -debug -extras -rplay -shape -translucency" 0 kB [1]
> 
Then, it might be an fvwm bug. Report it at fvwm-workers@fvwm.org, which is
the dev mailing list for fvwm, so it can be tracked and fixed.

> I noticed that the fvwm overlay version is not very useful.  Mplayer,
> for instance, which is also part of the devnull overlay, has a useful
> version:

Mplayer has a version number because it's a snapshot with a given date. The fvwm is
a live ebuild, which means that each time you emerge it, it download the latest code
from cvs. So, if you emerge it again the next week, it will automatically update your
fvwm if something has changed in cvs.

Cvs/svn/git and similar ebuilds are always numbered as 9999. That's why.


> 
> [ebuild   R   ] media-video/mplayer-2008.0103
> 
> fvwm --help is only slightly more useful:
> 
> fvwm 2.5.25 (from cvs) compiled on Feb 21 2008 at 20:06:03
> with support for: ReadLine, Stroke, XPM, PNG, SVG, XShm, SM, Bidi text, Xinerama, XRender, XCursor, XFT, NLS
> 
> How does it track the version internally well enough to know when
> updates are available?

It doesn't. You need to keep trak of the live ebuilds manually. There's an unofficial
script called live-ebuilds-update or something like that. You can search the gentoo forums.
I saw it around there but can't give you a link right now.

> And back to fvwm, should I report this to the gentoo maintainer (who
> seems to be seriously out of date) or to the fvwm project -- does
> gentoo add its own bells and whistles which would make a gentoo bug
> report less useful?

Gentoo doesn't add too much stuff. Just a couple of patches. But it would  help if you can
reproduce your problem with a vanilla build of fvwm. Without patches. So we can confirm that
it's an fvwm problem.

With my fvwm cvs ebuild you can do this, to get a vanilla build:

USE="-extras -translucency" emerge =fvwm-9999

-- 
Jesús Guerrero <i92guboj@terra.es>
--
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Thu, Feb 28, 2008 at 07:57:42PM +0100, Jes?s Guerrero wrote:

> Gentoo doesn't add too much stuff. Just a couple of patches. But it
> would help if you can reproduce your problem with a vanilla build of
> fvwm. Without patches. So we can confirm that it's an fvwm problem.

> With my fvwm cvs ebuild you can do this, to get a vanilla build:
> 
> USE="-extras -translucency" emerge =fvwm-9999

OK, it's compiling right now.  All I have for an error is the segfault
line in syslog.  Should I run fvwm with any particular args to get a
core dump or anything useful?  I am also willing to try special
builds, but you'd have to hold my hand for whatever gentoo tricks that
would require.  Or running it under gdb, I could probably do that, but
I am not sure how one would go about that since it is a window manager.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[Jesús Guerrero]

On Thu, 28 Feb 2008 22:48:29 -0800
felix@crowfix.com wrote:

> On Thu, Feb 28, 2008 at 07:57:42PM +0100, Jes?s Guerrero wrote:
> 
> > Gentoo doesn't add too much stuff. Just a couple of patches. But it
> > would help if you can reproduce your problem with a vanilla build of
> > fvwm. Without patches. So we can confirm that it's an fvwm problem.
> 
> > With my fvwm cvs ebuild you can do this, to get a vanilla build:
> > 
> > USE="-extras -translucency" emerge =fvwm-9999

First, lets see if the new vanilla build gives you the same problem. In
that case you can compile with USE="debug" to get a debug build of fvwm.

> 
> OK, it's compiling right now.  All I have for an error is the segfault
> line in syslog.  Should I run fvwm with any particular args to get a
> core dump or anything useful?  I am also willing to try special
> builds, but you'd have to hold my hand for whatever gentoo tricks that
> would require.  Or running it under gdb, I could probably do that, but
> I am not sure how one would go about that since it is a window manager.

Look here, particularly sections df1 and df3.
http://www.starshine.org/xteddy/thomas/fvwm/fvwmchanfaq.html#df1

If we can reproduce the same error with the vanilla build, then the
backtrace would belong to the fvwm-workers@fvwm.org mailing list, where
the fvwm can give you a better advice and the bug would probably be hunted
and tracked much more faster than I could do.

-- 
Jesús Guerrero <i92guboj@terra.es>
--
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] FVWM taskbar segfault on ~amd64

[felix]

On Fri, Feb 29, 2008 at 03:55:41PM +0100, Jes?s Guerrero wrote:
> First, lets see if the new vanilla build gives you the same problem. In
> that case you can compile with USE="debug" to get a debug build of fvwm.

> Look here, particularly sections df1 and df3.
> http://www.starshine.org/xteddy/thomas/fvwm/fvwmchanfaq.html#df1
> 
> If we can reproduce the same error with the vanilla build, then the
> backtrace would belong to the fvwm-workers@fvwm.org mailing list, where
> the fvwm can give you a better advice and the bug would probably be hunted
> and tracked much more faster than I could do.

Yes, TaskBar still segfaults.

Now I feel like some dinosaurish fool here.  I cannot get a core
dump.  ulimit -c 90000 or -c 1 or -c unlimited tells me

-bash: ulimit: core file size: cannot modify limit: Operation not permitted

Setting ulimit in /etc/profile doesn't give me a core dump, in $HOME
or /tmp or anywhere that I can find.  Bash's help ulimit says nothing
about this.  I can't find anything in /etc/sysctl.conf which disables
core files.  Changing /etc/security/limits.conf didn't help.

I have written threaded apps, database apps, mail handlers, all sorts
of stuff, and it's been so long since I used a core dump that I feel
positively ancient.  Good grief.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list

[gentoo-amd64] Re: FVWM taskbar segfault on ~amd64

[Duncan]

felix@crowfix.com posted 20080229170308.GA6121@crowfix.com, excerpted
below, on  Fri, 29 Feb 2008 09:03:08 -0800:

> Now I feel like some dinosaurish fool here.  I cannot get a core dump. 
> ulimit -c 90000 or -c 1 or -c unlimited tells me
> 
> -bash: ulimit: core file size: cannot modify limit: Operation not
> permitted
> 
> Setting ulimit in /etc/profile doesn't give me a core dump, in $HOME or
> /tmp or anywhere that I can find.  Bash's help ulimit says nothing about
> this.  I can't find anything in /etc/sysctl.conf which disables core
> files.  Changing /etc/security/limits.conf didn't help.

You probably checked it but if you mentioned it I missed it.  Does your
kernel include corefile functionality?  I have that option turned
entirely off, here.  If likewise there, that might explain not being
able to turn it on from user mode.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-amd64@lists.gentoo.org mailing list

Re: [gentoo-amd64] Re: FVWM taskbar segfault on ~amd64

[felix]

On Sat, Mar 01, 2008 at 01:55:28AM +0000, Duncan wrote:
> felix@crowfix.com posted 20080229170308.GA6121@crowfix.com, excerpted
> below, on  Fri, 29 Feb 2008 09:03:08 -0800:
> 
> > Now I feel like some dinosaurish fool here.  I cannot get a core dump. 
> > ulimit -c 90000 or -c 1 or -c unlimited tells me
> > 
> > -bash: ulimit: core file size: cannot modify limit: Operation not
> > permitted
> > 
> > Setting ulimit in /etc/profile doesn't give me a core dump, in $HOME or
> > /tmp or anywhere that I can find.  Bash's help ulimit says nothing about
> > this.  I can't find anything in /etc/sysctl.conf which disables core
> > files.  Changing /etc/security/limits.conf didn't help.
> 
> You probably checked it but if you mentioned it I missed it.  Does your
> kernel include corefile functionality?  I have that option turned
> entirely off, here.  If likewise there, that might explain not being
> able to turn it on from user mode.

I hadn't remembered there was such a kernel option, but I just checked
and CONFIG_ELF_CORE is set.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
-- 
gentoo-amd64@lists.gentoo.org mailing list