This is going to sound strange, no doubt, but I recently upgraded my
laptop (Dell Latitude D600) to Fedora Core 2 (Linux). Mostly it's been
great, with a few things that didn't work under Fedora 1 now working.
The funny thing is, if I start SLIME from within Emacs, using SBCL as my
Lisp, it crashes the laptop so badly that I have no option but to hit the
power switch. I guess something in the chain of SLIME, SBCL, Emacs,
XOrg, is tickling a bug in the kernel. I'd love to provide a good bug
report to someone who can fix it, but so far starting SLIME is the only
way I've found to trigger the bug. Starting SBCL from a terminal doesn't
trigger it, and neither does just starting Lisp as an inferior process
from Emacs.
Any ideas?
Cheers,
Bill.
--
Dr. William Bland.
It would not be too unfair to any language to refer to Java as a
stripped down Lisp or Smalltalk with a C syntax. (Ken Anderson).
<····@abstractnonsense.com> wrote:
>This is going to sound strange, no doubt, but I recently upgraded my
>laptop (Dell Latitude D600) to Fedora Core 2 (Linux). Mostly it's been
>great, with a few things that didn't work under Fedora 1 now working.
>
>The funny thing is, if I start SLIME from within Emacs, using SBCL as my
>Lisp, it crashes the laptop so badly that I have no option but to hit the
>power switch.
FWIW, I saw a similar thing with FC2 test 3. Starting SBCL from X
reliably froze the whole machine, starting it from a console worked
fine. Downgrading to kernel 2.4 from FC1 fixed the problem for me.
--
Juho Snellman
William Bland <····@abstractnonsense.com> writes:
> Any ideas?
What communication style (see *communication-style*) are you using?
The default for SBCL is :spawn, which uses multiple threads. Perhaps
SBCLs use of threads confuses the kernel.
SLIME also supports signal driven I/O, i.e. a Lisp function is invoked
every time the socket to Emacs becomes readable. SBCL does some
tricky things to implement such signal handlers: a signal handler,
written in C, inserts a so called "fake frame" above the interrupted
function. The job of the fake frame is to call the Lisp signal handler
when the C signal handler returns. Perhaps this scheme doesn't work on
Fedora.
You could try to set *communication-style* to :fd-handler; this uses
the SERVE-EVENT mechanism (i.e. select) and should be more robust, but
less convenient, than the other styles.
Helmut.
Well, if it were me I'd uninstall and reinstall slime.
See if you can clear out those config files completely when uninstalling.
Check your .emacs file, and set that exactly the way the SLIME doc tells
you to. Unfortunately common-lisp.net is down right now, but here's the
hook from my .emacs:
(setq inferior-lisp-program "/usr/bin/lisp")
(add-to-list 'load-path "/usr/share/emacs/site-lisp/slime")
(require 'slime)
(add-hook 'lisp-mode-hook (lambda () (slime-mode t)))
(add-hook 'inferior-lisp-mode-hook (lambda () (inferior-slime-mode t)))
(add-hook 'slime-mode-hook (lambda () (slime-autodoc-mode t))) ; Make info about the function appear in the minib
uffer
Put it after the load-path which is at the very top.
Hope this helps, and good luck. Slime is definetely worth the effort, it's
really nice.
- Jeremy L. Rayman
EnlightenOS
On Sat, 29 May 2004 05:10:57 +0000, William Bland wrote:
> This is going to sound strange, no doubt, but I recently upgraded my
> laptop (Dell Latitude D600) to Fedora Core 2 (Linux). Mostly it's been(setq inferior-lisp-program "/usr/bin/lisp")
> great, with a few things that didn't work under Fedora 1 now working.
>
> The funny thing is, if I start SLIME from within Emacs, using SBCL as my
> Lisp, it crashes the laptop so badly that I have no option but to hit the
> power switch. I guess something in the chain of SLIME, SBCL, Emacs,
> XOrg, is tickling a bug in the kernel. I'd love to provide a good bug
> report to someone who can fix it, but so far starting SLIME is the only
> way I've found to trigger the bug. Starting SBCL from a terminal doesn't
> trigger it, and neither does just starting Lisp as an inferior process
> from Emacs.
>
> Any ideas?
>
> Cheers,
> Bill.
Sage wrote:
>
> Well, if it were me I'd uninstall and reinstall slime.
>
> See if you can clear out those config files completely when
> uninstalling.
>
> Hope this helps, and good luck. Slime is definetely worth the effort,
> it's really nice.
>
Slime *crashes his computer*. I don't think a mere reinstall will help.
There is a good chance that it's caused by FC's imperfect romance with
unstable, experimental technologies such as NPTL; I'd suggest recompiling
SBCL/whatever without support for threads. It's the most likely culprit.
This being a laptop, I'd suggest going with Gentoo anyway - it seems to be
the best distribution to get stable, up-to-date software these days, and
it dfinitely has the best support for laptops of any I've ever tried.
Being able to say "USE=-threads emerge sbcl" helps too, although I haven't
needed to.
Svein Ove Aas schrieb:
> This being a laptop, I'd suggest going with Gentoo anyway - it seems to be
> the best distribution to get stable, up-to-date software these days, and
> it dfinitely has the best support for laptops of any I've ever tried.
I can second that.
--
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/
[mail and news]
William Bland <····@abstractnonsense.com> writes:
> This is going to sound strange, no doubt, but I recently upgraded my
> laptop (Dell Latitude D600) to Fedora Core 2 (Linux). Mostly it's been
> great, with a few things that didn't work under Fedora 1 now working.
>
> The funny thing is, if I start SLIME from within Emacs, using SBCL as my
> Lisp, it crashes the laptop so badly that I have no option but to hit the
> power switch. I guess something in the chain of SLIME, SBCL, Emacs,
> XOrg, is tickling a bug in the kernel. I'd love to provide a good bug
> report to someone who can fix it, but so far starting SLIME is the only
> way I've found to trigger the bug. Starting SBCL from a terminal doesn't
> trigger it, and neither does just starting Lisp as an inferior process
> from Emacs.
What kernel version does Fedora use?
Which version of SBCL are you running and is sb-thread in *features*?
SLIME is ASDF friendly. If you put a symlink to swank.asd in
$SBCL_HOME/lib/site-systems and a symlink to the slime route directory
in $SBCL_HOME/lib/site, then you can use ASDF to load swank from the
terminal:
·····@apostrophe:~/usr/lib/sbcl
$ ls -l site-systems/swank.asd
lrwxrwxrwx 1 david david 35 Apr 17 19:32 site-systems/swank.asd -> /home/david/usr/src/slime/swank.asd
·····@apostrophe:~/usr/lib/sbcl
$ ls -l site/slime
lrwxrwxrwx 1 david david 26 Apr 17 19:35 site/slime -> /home/david/usr/src/slime/
·····@apostrophe:~/usr/lib/sbcl
$ cat sbclrc
(require :asdf)
(require :asdf-install)
·····@apostrophe:~/usr/lib/sbcl
$ sbcl
This is SBCL 0.8.10.60, an implementation of ANSI Common Lisp.
More information about SBCL is available at <http://www.sbcl.org/>.
SBCL is free software, provided as is, with absolutely no warranty.
It is mostly in the public domain; some portions are provided under
BSD-style licenses. See the CREDITS and COPYING files in the
distribution for more information.
; loading system definition from
; #P"/home/david/usr/lib/sbcl/systems/sb-bsd-sockets.asd" into
; #<PACKAGE "ASDF2754">
; registering #<SYSTEM SB-BSD-SOCKETS {9AEB051}> as SB-BSD-SOCKETS
; registering #<SYSTEM SB-BSD-SOCKETS-TESTS {916A8A1}> as SB-BSD-SOCKETS-TESTS
; loading system definition from
; #P"/home/david/usr/lib/sbcl/systems/sb-posix.asd" into #<PACKAGE "ASDF2924">
; registering #<SYSTEM SB-POSIX {9A09C01}> as SB-POSIX
; registering #<SYSTEM SB-POSIX-TESTS {9B2FEC9}> as SB-POSIX-TESTS
* (require :swank)
; compiling file "/home/david/usr/src/slime/swank-loader.lisp" (written 30 APR 2004 01:31:58 PM):
; compiling top level form:
; compiling top level form:
; recognizing DEFUN MAKE-SWANK-PATHNAME
; compiling DEFUN MAKE-SWANK-PATHNAME:
; compiling top level form:
; compiling DEFPARAMETER *SYSDEP-PATHNAMES*:
; compiling top level form:
; compiling top level form:
; recognizing DEFUN FILE-NEWER-P
; compiling DEFUN FILE-NEWER-P:
; compiling top level form:
; recognizing DEFUN BINARY-PATHNAME
; compiling DEFUN BINARY-PATHNAME:
; compiling top level form:
; recognizing DEFUN COMPILE-FILES-IF-NEEDED-SERIALLY
; compiling DEFUN COMPILE-FILES-IF-NEEDED-SERIALLY:
; compiling top level form:
; recognizing DEFUN USER-INIT-FILE
; compiling DEFUN USER-INIT-FILE:
; compiling top level form:
; compiling top level form:
; compiling top level form:
; compiling top level form:
; /home/david/usr/src/slime/swank-loader.fasl written
; compilation finished in 0:00:01
WARNING: discarding unsupported DOCUMENTATION of type SYMBOL for object SYMBOL
WARNING: These Swank interfaces are unimplemented:
(DISASSEMBLE-FRAME LIST-CALLEES
LIST-CALLERS
PROFILE-PACKAGE
RESTART-FRAME
WHO-BINDS
WHO-CALLS
WHO-MACROEXPANDS
WHO-REFERENCES
WHO-SETS
WHO-SPECIALIZES)
("SB-INTROSPECT")
* *features*
(:ASDF :SB-FUTEX
:SB-THREAD
:ANSI-CL
:COMMON-LISP
:SBCL
:UNIX
:SB-DOC
:SB-TEST
:IEEE-FLOATING-POINT
:X86
:LINUX
:GENCGC
:STACK-GROWS-DOWNWARD-NOT-UPWARD
:C-STACK-IS-CONTROL-STACK)
* (quit)
·····@apostrophe:~/usr/lib/sbcl
$ uname -a
Linux apostrophe 2.6.5-686 #1 Tue Apr 13 01:53:00 EDT 2004 i686 GNU/Linux
·····@apostrophe:~/usr/lib/sbcl
$
See what you can do with that.
I'm running SLIME from 5-25-04 on this machine and Emacs 21.3.1 from
Debian/testing. I can at least start slime with no problems.
--
An ideal world is left as an excercise to the reader.
--- Paul Graham, On Lisp 8.1
In article <······························@abstractnonsense.com>,
William Bland <····@abstractnonsense.com> wrote:
> This is going to sound strange, no doubt, but I recently upgraded my
> laptop (Dell Latitude D600) to Fedora Core 2 (Linux). Mostly it's been
> great, with a few things that didn't work under Fedora 1 now working.
>
> The funny thing is, if I start SLIME from within Emacs, using SBCL as my
> Lisp, it crashes the laptop so badly that I have no option but to hit the
> power switch. I guess something in the chain of SLIME, SBCL, Emacs,
> XOrg, is tickling a bug in the kernel. I'd love to provide a good bug
> report to someone who can fix it, but so far starting SLIME is the only
> way I've found to trigger the bug. Starting SBCL from a terminal doesn't
> trigger it, and neither does just starting Lisp as an inferior process
> from Emacs.
>
> Any ideas?
Yes, remove the SLIME from your computer and hope the best.
Good luck!
> Cheers,
> Bill.
Rainer Joswig wrote:
> In article <······························@abstractnonsense.com>,
> William Bland <····@abstractnonsense.com> wrote:
>
>
>>This is going to sound strange, no doubt, but I recently upgraded my
>>laptop (Dell Latitude D600) to Fedora Core 2 (Linux). Mostly it's been
>>great, with a few things that didn't work under Fedora 1 now working.
>>
>>The funny thing is, if I start SLIME from within Emacs, using SBCL as my
>>Lisp, it crashes the laptop so badly that I have no option but to hit the
>>power switch. I guess something in the chain of SLIME, SBCL, Emacs,
>>XOrg, is tickling a bug in the kernel. I'd love to provide a good bug
>>report to someone who can fix it, but so far starting SLIME is the only
>>way I've found to trigger the bug. Starting SBCL from a terminal doesn't
>>trigger it, and neither does just starting Lisp as an inferior process
>>from Emacs.
>>
>>Any ideas?
>
>
> Yes, remove the SLIME from your computer and hope the best.
>
> Good luck!
Boy ("Yob" spelled backward), you sure told him! It is a pity
better-informed people ruined it by confirming his result and suggesting
workarounds.
:)
kenny
--
Home? http://tilton-technology.com
Cells? http://www.common-lisp.net/project/cells/
Cello? http://www.common-lisp.net/project/cello/
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Your Project Here! http://alu.cliki.net/Industry%20Application
William Bland wrote:
> This is going to sound strange, no doubt, but I recently upgraded my
> laptop (Dell Latitude D600) to Fedora Core 2 (Linux). Mostly it's been
> great, with a few things that didn't work under Fedora 1 now working.
>
> The funny thing is, if I start SLIME from within Emacs, using SBCL as my
> Lisp, it crashes the laptop so badly that I have no option but to hit
> the
> power switch. I guess something in the chain of SLIME, SBCL, Emacs,
> XOrg, is tickling a bug in the kernel. I'd love to provide a good bug
> report to someone who can fix it, but so far starting SLIME is the only
> way I've found to trigger the bug. Starting SBCL from a terminal
> doesn't trigger it, and neither does just starting Lisp as an inferior
> process from Emacs.
>
> Any ideas?
>
Fedora Core seems to have a lot of problems, for a lot of people. For
instance, when I tried FC2 on my desktop computer I got the exact same
problem as you do.
Switching to Gentoo solved it, but I imagine anything other than FC would
do in a pinch.
Svein Ove Aas wrote:
> Fedora Core seems to have a lot of problems, for a lot of people.
All these stories are reinforcing my intention to avoid Fedora.
Paul
On Sat, 29 May 2004 06:22:42 -0500, Paul F. Dietz wrote:
> Svein Ove Aas wrote:
>
>> Fedora Core seems to have a lot of problems, for a lot of people.
>
> All these stories are reinforcing my intention to avoid Fedora.
>
> Paul
Thanks to all those who responded quickly and with some great information
and suggestions. If FC2 hadn't just given me one more reason to switch, I
might have tried to fix the problem, but it did:
I have a digital camera, for which I had manually added an entry in fstab,
mounting it on /mnt/usbstorage. The first time I plugged the camera in,
that mount point vanished from fstab, and a new one was set up, for
/mnt/camera. Hmm, ok, not entirely happy with my OS second guessing me
but ok, I guess it is a kind of reasonable thing to do. Unfortunately the
second time I plugged the camera in, the /mnt/camera entry vanished from
fstab and wasn't replaced with anything at all!
Switching to Gentoo asap!
Thanks again for all the replies.
Cheers,
Bill.
--
Dr. William Bland.
It would not be too unfair to any language to refer to Java as a
stripped down Lisp or Smalltalk with a C syntax. (Ken Anderson).
Svein Ove Aas <··············@brage.info> writes:
> William Bland wrote:
>
> > Switching to Gentoo asap!
> >
>
> Note that Gentoo also has the option of using NPTL. You should probably
> try to avoid it.
Isn't NPTL standard in 2.6.x kernels now? sb-futex depends on it,
although it is supposed to quietly fall back to kernel mutexes or
whatever if NPTL isn't there.
--
An ideal world is left as an excercise to the reader.
--- Paul Graham, On Lisp 8.1
David Steuber wrote:
> Svein Ove Aas <··············@brage.info> writes:
>
>> William Bland wrote:
>>
>> > Switching to Gentoo asap!
>> >
>>
>> Note that Gentoo also has the option of using NPTL. You should probably
>> try to avoid it.
>
> Isn't NPTL standard in 2.6.x kernels now? sb-futex depends on it,
> although it is supposed to quietly fall back to kernel mutexes or
> whatever if NPTL isn't there.
>
NPTL support in kernel is, yes.
The experimental code I was referring to is in glibc, and SBCL presumably
doesn't use it; it's a version of pthreads that uses NPTL instead of
processes.
You have to do extra work to enable it, though; just avoid the nptl USE
flag and you'll be fine.
Svein Ove Aas <··············@brage.info> writes:
> NPTL support in kernel is, yes.
> The experimental code I was referring to is in glibc, and SBCL presumably
> doesn't use it; it's a version of pthreads that uses NPTL instead of
> processes.
I see.
As I understand it, SBCL does not use pthreads at all. It goes
directly to the kernel for native thread support on Linux x86 if you
add sb-thread in customize-target-features.lisp. With a 2.6.x kernel,
you can also add sb-futex as well.
--
An ideal world is left as an excercise to the reader.
--- Paul Graham, On Lisp 8.1