7:44 p.m.
This bug report will be sent to the XEmacs Development Team,
not to your local site managers!!
Please write in English, because the XEmacs maintainers do not have
translators to read other languages for them.
In XEmacs 21.4 (patch 4) "Artificial Intelligence (candidate 2)" [Lucid]
(i586-pc-win32) of Fri Jul 20 2001 on D5DC120J
configured using `configure UNKNOWN'
Please describe exactly what actions triggered the bug
and the precise symptoms of the bug:
Crash during normal usage.
Recent keystrokes:
M-x g n u s RET n n n n s M-x r e c - s e s s SPC RET
s p d d d x y C-x k RET M-x r e c a SPC BS o SPC f
i SPC SPC SPC M-BS a SPC SPC RET n n n n C-x o C-x
1 C-x j RET C-x k RET C-x C-b C-x o C-x 1 n d x g n
n n RET g p RET SPC q s M-x r e p SPC o SPC r t - e
m SPC SPC RET
Recent messages (most recent first):
Loading iso-acc...
Loading lazy-lock...done
Loading lazy-lock...
Loading emacsbug...done
Loading emacsbug...
Making completion list...
Making completion list...
Saving c:\Users\AichnerAd\.newsrc.eld...done
Wrote c:\Users\AichnerAd\.newsrc.eld
Saving c:\Users\AichnerAd\.newsrc.eld...
Window procedure called during GC???????
<297 lines, all alike, deleted by Adrian>
Window procedure called during GC???????
Fatal error.
Your files have been auto-saved.
Use `M-x recover-session' to recover them.
If you have access to the PROBLEMS file that came with your
version of XEmacs, please check to see if your crash is described
there, as there may be a workaround available.
Otherwise, please report this bug by running the send-pr
script included with XEmacs, or selecting `Send Bug Report'
from the help menu.
As a last resort send ordinary email to `crashes(a)xemacs.org'.
*MAKE SURE* to include the information in the command
M-x describe-installation.
Lisp backtrace follows:
# bind (print-message-label)
display-error((wrong-type-argument eventp
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
8:10 a.m.
>>>> "APA" == Adrian Aichner
<Adrian.Aichner(a)t-online.de> writes:
Hi Stephen, Ben, Paul,
would following patch fix this problem for 21.4.4 candidate 2 also?
Best regards,
Adrian
From: Paul Krause <pkrause(a)soundbite.com>
Subject: [PATCH] infloop workaround for Windows procedure called during GC
To: xemacs-patches(a)xemacs.org
CC: xemacs-beta(a)xemacs.org
Date: 5 Jun 2001 08:11:05 -0400
Message-ID: <FENWAY95sao0eCdPap400001472(a)fenway.corp.soundbite.com>
APA> This bug report will be sent to the XEmacs Development Team,
APA> not to your local site managers!!
APA> Please write in English, because the XEmacs maintainers do not have
APA> translators to read other languages for them.
APA> In XEmacs 21.4 (patch 4) "Artificial Intelligence (candidate 2)"
[Lucid] (i586-pc-win32) of Fri Jul 20 2001 on D5DC120J
APA> configured using `configure UNKNOWN'
APA> Please describe exactly what actions triggered the bug
APA> and the precise symptoms of the bug:
APA> Crash during normal usage.
APA> Recent keystrokes:
APA> M-x g n u s RET n n n n s M-x r e c - s e s s SPC RET
APA> s p d d d x y C-x k RET M-x r e c a SPC BS o SPC f
APA> i SPC SPC SPC M-BS a SPC SPC RET n n n n C-x o C-x
APA> 1 C-x j RET C-x k RET C-x C-b C-x o C-x 1 n d x g n
APA> n n RET g p RET SPC q s M-x r e p SPC o SPC r t - e
APA> m SPC SPC RET
APA> Recent messages (most recent first):
APA> Loading iso-acc...
APA> Loading lazy-lock...done
APA> Loading lazy-lock...
APA> Loading emacsbug...done
APA> Loading emacsbug...
APA> Making completion list...
APA> Making completion list...
APA> Saving c:\Users\AichnerAd\.newsrc.eld...done
APA> Wrote c:\Users\AichnerAd\.newsrc.eld
APA> Saving c:\Users\AichnerAd\.newsrc.eld...
APA> Window procedure called during GC???????
APA> <297 lines, all alike, deleted by Adrian>
APA> Window procedure called during GC???????
APA> Fatal error.
APA> Your files have been auto-saved.
APA> Use `M-x recover-session' to recover them.
APA> If you have access to the PROBLEMS file that came with your
APA> version of XEmacs, please check to see if your crash is described
APA> there, as there may be a workaround available.
APA> Otherwise, please report this bug by running the send-pr
APA> script included with XEmacs, or selecting `Send Bug Report'
APA> from the help menu.
APA> As a last resort send ordinary email to `crashes(a)xemacs.org'.
APA> *MAKE SURE* to include the information in the command
APA> M-x describe-installation.
APA> Lisp backtrace follows:
APA> # bind (print-message-label)
APA> display-error((wrong-type-argument eventp
APA> --
APA> Adrian Aichner
APA> mailto:adrian＠xemacs.org
APA> http://www.xemacs.org/
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
8:59 a.m.
>>>> "APA" == Adrian Aichner
<Adrian.Aichner(a)t-online.de> writes:
APA> would following patch fix this problem for 21.4.4 candidate 2
APA> also?
It will make the infloop and crash go away; it will probably leave you
with a unstable XEmacs. Ben sez:
it was never safe; i just added the printf recently to flush out
the problem.
patch forthcoming.
and later (along with the patch)
the gc-in-window-procedure fix is probably a serious bug too, but
it less obviously leads to crashes. it probably leads to some
subtle weird problems, but you can pass on it if you want.
So I did pass (ie, postpone), without looking closely at that point.
Um, Ben, as I understand it asserts get defined away to "DO_NOTHING;"
in a --debug=no XEmacs. Is assert the right thing to do here?
I have a patch for the IPv6 stupidity in hand, and a limited amount of
time to fix any showstoppers before I leave on vacation. So I'm going
to roll out 21.4.4 within a few hours.
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
9:37 a.m.
>>>> "Stephen" == Stephen J Turnbull
<turnbull(a)sk.tsukuba.ac.jp> writes:
>>>> "APA" == Adrian Aichner
<Adrian.Aichner(a)t-online.de> writes:
APA> would following patch fix
this problem for 21.4.4 candidate 2
APA> also?
Stephen> It will make the infloop and crash go away; it will
Stephen> probably leave you with a unstable XEmacs. Ben sez:
Stephen> it was never safe; i just added the printf recently
Stephen> to flush out the problem.
Stephen> patch forthcoming.
Stephen> and later (along with the patch)
Stephen> the gc-in-window-procedure fix is probably a serious
Stephen> bug too, but it less obviously leads to crashes. it
Stephen> probably leads to some subtle weird problems, but you
Stephen> can pass on it if you want.
Stephen> So I did pass (ie, postpone), without looking closely at
Stephen> that point.
I did look at old threads about this yesterday, but I wasn't able to
determine whether a good fix for this exists.
What should be done?
Stephen> Um, Ben, as I understand it asserts get defined away to
Stephen> "DO_NOTHING;" in a --debug=no XEmacs. Is assert the
That's correct.
Stephen> right thing to do here?
Stephen> I have a patch for the IPv6 stupidity in hand, and a
Stephen> limited amount of time to fix any showstoppers before I
Stephen> leave on vacation. So I'm going to roll out 21.4.4
Have a good one!
Stephen> within a few hours.
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
10:05 a.m.
>>>> "APA" == Adrian Aichner
<adrian(a)xemacs.org> writes:
Ben> the gc-in-window-procedure fix is probably a serious
Ben> bug too, but it less obviously leads to crashes. it
Ben> probably leads to some subtle weird problems, but you
Ben> can pass on it if you want.
Stephen> So I did pass (ie, postpone), without looking closely at
Stephen> that point.
APA> I did look at old threads about this yesterday, but I wasn't
APA> able to determine whether a good fix for this exists.
Nothing I've seen.
APA> What should be done?
Tell you what. I'll make Paul's change so that XEmacs bitches no
matter what, but without inflooping (I assume). And I'll add Ben's
assert and you can build an XEmacs with --debug=yes. That will give
Ben and Paul the maximum amount of debug information to work with,
since you seem to be hitting this every few hours.
This should be fairly safe, since the alternative of leaving it as is
means a crash anyway. Think about it second or two and see if I'm
missing something. I really don't want to roll another XEmacs this
week (or rather, I'd like to make my second release this week be
21.5.2 not 21.4.5 ;-)
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
10:25 a.m.
>>>> "Stephen" == Stephen J Turnbull
<turnbull(a)sk.tsukuba.ac.jp> writes:
Stephen> Tell you what. I'll make Paul's change so that XEmacs
Stephen> bitches no matter what, but without inflooping (I
Stephen> assume). And I'll add Ben's assert and you can build an
Stephen> XEmacs with --debug=yes. That will give Ben and Paul the
Stephen> maximum amount of debug information to work with, since
Stephen> you seem to be hitting this every few hours.
I still think this is fairly safe (depends on whether the user
considers an infloop and crash now or possible subtle corruption later
worse), but it's not clear that you personally want to do --debug=yes
for this problem. The reason is that I took a closer look at Ben's
patch and realized that Ben has already identified a couple of places
where this problem might arise. So you may just be letting yourself
in for asserts he already knows about. I'm not going to apply the
rest of the patch until there's been some experience with it in 21.5,
unless Ben really advocates it.
Ben?
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
11:03 a.m.
>>>> "Stephen" == Stephen J Turnbull
<turnbull(a)sk.tsukuba.ac.jp> writes:
>>>> "Stephen" == Stephen J Turnbull
<turnbull(a)sk.tsukuba.ac.jp> writes:
Stephen> Tell you what. I'll
make Paul's change so that XEmacs
Stephen> bitches no matter what, but without inflooping (I
Stephen> assume). And I'll add Ben's assert and you can build an
Stephen> XEmacs with --debug=yes. That will give Ben and Paul the
Stephen> maximum amount of debug information to work with, since
Stephen> you seem to be hitting this every few hours.
Stephen> I still think this is fairly safe (depends on whether the
Stephen> user considers an infloop and crash now or possible
Stephen> subtle corruption later worse), but it's not clear that
Stephen> you personally want to do --debug=yes for this problem.
I am willing to build with --debug=yes.
Stephen> The reason is that I took a closer look at Ben's patch
Stephen> and realized that Ben has already identified a couple of
Stephen> places where this problem might arise. So you may just
Stephen> be letting yourself in for asserts he already knows
Stephen> about. I'm not going to apply the rest of the patch
Stephen> until there's been some experience with it in 21.5,
Stephen> unless Ben really advocates it.
I have been using latest CVS 21.5 exclusively until a few days ago
when 21.4.4 candidate 1/2 came along.
I haven't had regular crashes there.
Adrian
Stephen> Ben?
Stephen> --
Stephen> University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573
JAPAN
Stephen> Institute of Policy and Planning Sciences Tel/fax: +81 (298)
53-5091
Stephen> _________________ _________________ _________________
_________________
Stephen> What are those straight lines for? "XEmacs rules."
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
4:52 p.m.
At 04:59 PM 7/23/01 +0900, Stephen J. Turnbull wrote:
I have a patch for the IPv6 stupidity in hand, and a limited amount
of
time to fix any showstoppers before I leave on vacation. So I'm going
to roll out 21.4.4 within a few hours.
BTW commenting out the code it canonicalize_host_name() improved my
situation but did not fix it - XEmacs without packages then came up in 20
seconds rather than 2 minutes. Definining a fully qualified hostname vastly
improved things.
andy
10:09 a.m.
>>>> "Andy" == Andy Piper <andyp(a)bea.com>
writes:
Andy> At 04:59 PM 7/23/01 +0900, Stephen J. Turnbull wrote:
> client does that already, and it should be reasonably easy to
> configure PPP that way too.
Andy> This is no good to me - I have a fixed net connection (no
Andy> dialup) - until I take my laptop on BART. So I have to have
Andy> the nameserver configured.
That's exactly my situation. The reference to DHCP is where you can
steal the code from. The solution is perfectly general, the only
problem is that the fastest way I've found to check for connectivity
is to wait 5 seconds for ping -c 1 to time out. This is particularly
annoying on the train.
> I have a patch for the IPv6 stupidity in hand, and a limited
> amount of time to fix any showstoppers before I leave on
> vacation. So I'm going to roll out 21.4.4 within a few hours.
Andy> BTW commenting out the code it canonicalize_host_name()
Andy> improved my situation but did not fix it - XEmacs without
Andy> packages then came up in 20 seconds rather than 2
Andy> minutes. Definining a fully qualified hostname vastly
Andy> improved things.
Did you not comment out the code in sysdep.c (init_system_name)?
I don't think an fqdn would help there, unless ... ah, what is in your
/etc/hosts.conf (or nsswitch.conf)? Is DNS files, dns? And you added
the fqdn to the /etc/hosts file?
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."
8688
days inactive
8690
days old
9 comments
4 participants
participants (4)
-
Adrian Aichner -
Adrian Aichner -
Andy Piper -
Stephen J. Turnbull