May 2000 - XEmacs-Beta - XEmacs Mailing List Archives

(interrup-process ...) broken on NT native 24 years, 6 months

Adrian Aichner

Hello Ben! I can't figure out which change broke (interrupt-process ...) in XEmacs 21.1 and XEmacs 21.2 built natively on NT4SP5. This is the reason why kill-compilation is no longer working in *compile* buffers (needless to say: It used to work to approx. 21.1.7). Perhaps you'll have some ideas. I am copying Gunnar who has also done work in the process area. Here's how far my diagnosis goes: This works: (start-process "sleepy" nil "sleep" "60") (message "before: %S" (get-process "sleepy")) (interrupt-process "sleepy") (message "after: %S" (get-process "sleepy")) This doesn't: (start-process "shell" nil "cmd" "/c" "sleep" "60") (message "before: %S" (get-process "shell")) (interrupt-process "shell" t) (message "after: %S" (get-process "shell")) It would be great if we could fix this for 21.1 even, if not for 21.1.10. Best regards, Adrian

2 participants
1 comments

XEmacs 21.1.10 "Capitol Reef" has been Released 24 years, 6 months

Vin Shelton

XEmacs-21.1.10 "Capitol Reef" is available from ftp.xemacs.org. Get it from: ftp://ftp.xemacs.org/pub/xemacs/xemacs-21.1/xemacs-21.1.10.tar.gz ftp://ftp.xemacs.org/pub/xemacs/xemacs-21.1/xemacs-21.1.10-elc.tar.gz ftp://ftp.xemacs.org/pub/xemacs/xemacs-21.1/xemacs-21.1.10-info.tar.gz A patch to upgrade from 21.1.9 is also available: ftp://ftp.xemacs.org/pub/xemacs/xemacs-21.1/xemacs-21.1.9-21.1.10.patch.gz [The patch tarball only includes patches from xemacs-21.1.9.tar.gz to xemacs-21.1.10.tar.gz; you will still need to retrieve the elc and info tarballs. After applying the patch, please run the command 'touch configure' in order to make sure 'configure' is newer than 'configure.in'.] XEmacs-21.1 is the stable branch of XEmacs. As such, only bugfix submissions are allowed into these releases. Here are the relevant URLS for the bugs fixed in 21.1.10: From: Jonathan Harris <jhar(a)tardis.ed.ac.uk> Subject: Stop mswindows leaking DDE strings X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200002/msg00116.html From: Andy Piper <andy(a)xemacs.org> Subject: Re: [comp.emacs.xemacs] Re: Is the latest stable version X-archive-URL: http://www.xemacs.org/list-archives/xemacs-beta/200002/msg00391.html From: Yoshiki Hayashi <t90553(a)m.ecc.u-tokyo.ac.jp> Subject: Reset detection state when autodetecting X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/199912/msg00160.html From: "Kirill 'Big K' Katsnelson" <kkm(a)dtmx.com> Subject: [PATCH] Have (write-file) ask for coding system X-archive-URL: http://www.xemacs.org/list-archives/xemacs-review/200002/msg00276.html From: Oscar Figueiredo <Oscar.Figueiredo(a)di.epfl.ch> Subject: [21.1]: LDAP fixes X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200002/msg00191.html From: Sandra Wambold <wambold(a)xemacs.org> Subject: FAQ patch: XEmacs won't start on Windows X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00006.html From: SL Baur <steve(a)beopen.com> Subject: Fixes for with-string-as-buffer-content and with-output-to-string X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00007.html From: Martin Buchholz <martin(a)xemacs.org> Subject: configure.in: Unconditionally define SHELL for Autoconf 2.14.1 X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200002/msg00176.html From: Jan Vroonhof <vroonhof(a)math.ethz.ch> Subject: Fix 21.1.x crash X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00011.html From: Jan Vroonhof <vroonhof(a)math.ethz.ch> Subject: *_end_of_line, handle line_cache failure modes X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00012.html From: Vin Shelton <acs(a)xemacs.org> Subject: Remove 'load .emacs' button from all menubars X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00014.html From: Jonathan Harris <jhar(a)tardis.ed.ac.uk> Subject: Run handlers even if local machine doesn't have symlinks X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00023.html From: Yoshiki Hayashi <yoshiki(a)xemacs.org> Subject: Terminal coding system fix X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00000.html From: sperber(a)informatik.uni-tuebingen.de (Michael Sperber [Mr. Preprocessor]) Subject: PROBLEMS update for AIX X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00044.html From: Reini Urban <rurban(a)x-ray.at> Subject: nt/PROBLEMS patch X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200003/msg00058.html From: Martin Buchholz <martin(a)xemacs.org> Subject: Fix horrible data corruption bug X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200004/msg00078.html From: Martin Buchholz <martin(a)xemacs.org> Subject: 21.1: Backport of big-string mule crash fix from 21.2 X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200004/msg00082.html From: Jan Vroonhof <vroonhof(a)math.ethz.ch> Subject: Re: assertion failed; looks like a bug X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200004/msg00029.html From: SL Baur <steve(a)turbolinux.co.jp> From: toshihiro(a)turbolinux.co.jp Subject: Backport to 21.1 of Hrvoje Niksic patch of November 29, 1998 Subject: XEmacs debug 2 X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200004/msg00067.html From: Mike Alexander <mta(a)arbortext.com> Subject: Small windows with scroll-step set X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200004/msg00081.html From: "Mike Woolley" <mike(a)ariel.co.uk> Subject: RE: Patch to fix recurrent non-starting problem in NT XEmacs 21.1.x X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200005/msg00006.html From: Vin Shelton <acs(a)xemacs.org> Subject: hangs on long shell input lines on DEC OSF 4.0 X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200004/msg00057.html From: Jan Vroonhof <vroonhof(a)math.ethz.ch> Subject: Re: NEED_MOTIF in configure (PR#1129) X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200002/msg00073.html From: Gunnar Evermann <ge204(a)eng.cam.ac.uk> Subject: Re: <approved> Re: NEED_MOTIF in configure (PR#1129) X-archive-URL: http://www.xemacs.org/list-archives/xemacs-patches/200002/msg00133.html The 21.1.x series of releases is being named for US National Parks. Capitol Reef is located in Utah. For more information see: http://www.nps.gov/care/. Vin Shelton

1 participants
0 comments

RE: (interrup-process ...) broken on NT native 24 years, 6 months

Charlie Fineman

I'm running NT4 SP6. I ran through your examples and my results match yours (visa vie shell vs non-shell) btw, I used both tcsh and cmd and got the same results. I also launched sleep from perl (start-process "shell" null "perl" "t.pl") And the interupt killed the perl and the sleep processes so it seems to be some interaction with the shell (but this is definately an extrapolation :-) -----Original Message----- From: Adrian Aichner [mailto:aichner＠ecf.teradyne.com] Sent: Saturday, May 06, 2000 1:00 PM To: Charlie Fineman Cc: 'Gunnar Evermann'; Adrian Aichner; Ben Wing; XEmacs Beta List; XEmacs NT List Subject: Re: (interrup-process ...) broken on NT native >>>>> "Charlie" == Charlie Fineman <fineman(a)arzoon.com> writes: Charlie> Here are some data points from 21.2 build32 Charlie, what is yourr OS type and version? I have NT4 SP5. Following is what I experience on my system: Charlie> - interrupt-process has no effect Well, this isn't strictly true. It works fine for processes not run under a shell ("cmd" in my case). If you evaluate these manually one-by-one (C-x C-e (eval-last-sexp)), you should not see the sleep process in task manager. (start-process "sleepy" nil "sleep" "60") (message "before: %S" (get-process "sleepy")) (interrupt-process "sleepy") (message "after: %S" (get-process "sleepy")) What does not work is interrupting a process runnning under a "cmd" shell: (start-process "shell" nil "cmd" "/c" "sleep" "5") (message "before: %S" (get-process "shell")) (interrupt-process "shell" t) (message "after: %S" (get-process "shell")) Neither "cmd" nor "sleep" get interrupted. Charlie> - kill-process prints the exit message in the process Charlie> buffer and the process no longer shows up in Charlie> list-processes but the NT process still shows up in the Charlie> Task Manager Not for me. Processes not running as sub-processes of a shell are really killed. Charlie> - delete-process has the same result as kill-process (no Charlie> suprise since the doc seems to indicate that it calls Charlie> kill-process) See above. Charlie> - killing the process from the Task Manager seems to work Charlie> fine (i.e. emacs does notice it) Yep, same here. Charlie> -----Original Message----- Charlie> From: Gunnar Evermann [mailto:ge204＠eng.cam.ac.uk] Charlie> Sent: Saturday, May 06, 2000 10:52 AM Charlie> To: Adrian Aichner Charlie> Cc: Ben Wing; XEmacs Beta List; XEmacs NT List Charlie> Subject: Re: (interrup-process ...) broken on NT native Charlie> Have you tried using Fdelete_process instead of Finterrupt_process? Charlie> random other idea: Can you kill the subprocess from outside xemacs and Charlie> see whether we at least notice that? Charlie> Gunnar

1 participants
0 comments

Re: Urk! Potential patch to scrollbar.c? 24 years, 6 months

Hrvoje Niksic

Vin Shelton <acs(a)xemacs.org> writes: > >>>>> On 05 May 2000 12:11:20 -0500, wmperry(a)aventail.com (William M. Perry) said: > > BP> Here's a patch against 21.2 CVS of today. The same code can go > BP> into 21.1 trivially. This removes all the ifdef'd code about > BP> window systems in scrollbar.c except for setting the default > BP> scrollbar sides. Maybe that could be a device method? I'm > BP> still trying to figure out the best way to get the nextstep > BP> scrollbars on the left by default... there is some disgusting > BP> voodoo in the theme that I can't really emulate very well. > BP> *sigh* > > I don't see how this fits under the definition of bug fix, so this > won't go into 21.1. It doesn't -- I don't think Bill intended the patch to be applied to 21.1.

1 participants
0 comments

Re: (interrup-process ...) broken on NT native 24 years, 6 months

Adrian Aichner

>>>>> "Charlie" == Charlie Fineman <fineman(a)arzoon.com> writes: Charlie> Here are some data points from 21.2 build32 Charlie, what is yourr OS type and version? I have NT4 SP5. Following is what I experience on my system: Charlie> - interrupt-process has no effect Well, this isn't strictly true. It works fine for processes not run under a shell ("cmd" in my case). If you evaluate these manually one-by-one (C-x C-e (eval-last-sexp)), you should not see the sleep process in task manager. (start-process "sleepy" nil "sleep" "60") (message "before: %S" (get-process "sleepy")) (interrupt-process "sleepy") (message "after: %S" (get-process "sleepy")) What does not work is interrupting a process runnning under a "cmd" shell: (start-process "shell" nil "cmd" "/c" "sleep" "5") (message "before: %S" (get-process "shell")) (interrupt-process "shell" t) (message "after: %S" (get-process "shell")) Neither "cmd" nor "sleep" get interrupted. Charlie> - kill-process prints the exit message in the process Charlie> buffer and the process no longer shows up in Charlie> list-processes but the NT process still shows up in the Charlie> Task Manager Not for me. Processes not running as sub-processes of a shell are really killed. Charlie> - delete-process has the same result as kill-process (no Charlie> suprise since the doc seems to indicate that it calls Charlie> kill-process) See above. Charlie> - killing the process from the Task Manager seems to work Charlie> fine (i.e. emacs does notice it) Yep, same here. Charlie> -----Original Message----- Charlie> From: Gunnar Evermann [mailto:ge204＠eng.cam.ac.uk] Charlie> Sent: Saturday, May 06, 2000 10:52 AM Charlie> To: Adrian Aichner Charlie> Cc: Ben Wing; XEmacs Beta List; XEmacs NT List Charlie> Subject: Re: (interrup-process ...) broken on NT native Charlie> Have you tried using Fdelete_process instead of Finterrupt_process? Charlie> random other idea: Can you kill the subprocess from outside xemacs and Charlie> see whether we at least notice that? Charlie> Gunnar

1 participants
0 comments

RE: (interrup-process ...) broken on NT native 24 years, 6 months

Charlie Fineman

Here are some data points from 21.2 build32 - interrupt-process has no effect - kill-process prints the exit message in the process buffer and the process no longer shows up in list-processes but the NT process still shows up in the Task Manager - delete-process has the same result as kill-process (no suprise since the doc seems to indicate that it calls kill-process) - killing the process from the Task Manager seems to work fine (i.e. emacs does notice it) -----Original Message----- From: Gunnar Evermann [mailto:ge204＠eng.cam.ac.uk] Sent: Saturday, May 06, 2000 10:52 AM To: Adrian Aichner Cc: Ben Wing; XEmacs Beta List; XEmacs NT List Subject: Re: (interrup-process ...) broken on NT native Have you tried using Fdelete_process instead of Finterrupt_process? random other idea: Can you kill the subprocess from outside xemacs and see whether we at least notice that? Gunnar

1 participants
0 comments

Re: (interrup-process ...) broken on NT native 24 years, 6 months

Adrian Aichner

>>>>> "Gunnar" == Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: >> SIGINT under NT is implemented by sending CTRL_C_EVENT. >> >> I got lost during debug trying to figure out what's wrong. >> Gunnar> could try backing out patches until it works :-( >> >> Sounds pretty scary without your help. Who knows what interactions >> exist between your patches and Bens' and others'. Gunnar> I think there should be no interactions. My changes are Gunnar> pretty trivial and only consist of some extra Gunnar> CHECK_LIVE_PROCESS calls. I found your patch and it look harmless to me. >> >> It would be great if we could fix this for 21.1 even, if not for >> >> 21.1.10. >> Gunnar> indeed. I think I never intended my changes to go into Gunnar> 21.1, for which I had a simpler version, but this should Gunnar> still work. >> >> Hmmh, can you give me some pointers to problematic portions of your >> patches or point me to the simpler patch on xemacs-patches? Gunnar> http://www.xemacs.org/list-archives/xemacs-patches/199912/msg00066.html Gunnar> In your test cases you have to careful, when you evaluate Gunnar> them. I just evaluated them wrapped in a progn and thought Gunnar> the sleep didn't get killed (on Unix). The real problem Gunnar> was that we never got back to the event-loop and thus Gunnar> status_notify didn't get called before the second Gunnar> message. Put in a (sit-for 1) to be sure. (sit-for 1) didn't work reliably for me either. Sometimes (interrupt-process ...) would report not being able to send signal to process. That's why my test-case is not wrapped in interactive functions. Instead (I should have mentioned this) I C-x C-e (eval-last-sexp) one-by-one. (kill-process ...) instead of (interrupt-process ...) does kill the "cmd /c sleep 60" However, changing (interrupt-process ...) to (kill-process ...) in (defun kill-compilation ...) in compile.el would only cure the symptom, not the ailment. Gunnar> Have you tried using Fdelete_process instead of Gunnar> Finterrupt_process? (delete-process "sleepy") does work too. Killing a compilation in a compilation buffer work too, like so: (delete-process "compilation") Again, changing compile.el would only conseal the breakage in interrupt-process, which used to work, even on NT. Gunnar> random other idea: Can you kill the subprocess from Gunnar> outside xemacs and see whether we at least notice that? Yes, we do. (message "killed from task-manager: %S" (get-process "sleepy")) leads this: "killed from task-manager: nil" Gunnar> Gunnar Currently I suspect (interrupt-process ...) is trying to send CTRL_C_EVENT to the "cmd" shell, not to the sub-process. Also, during debug I see that after d.event = CTRL_C_EVENT; d.event == 0. See the attached send_signal() of 21.1 (latest CVS). I'll investigate if this is correct. Best regards, Adrian static int send_signal (HANDLE h_process, int signo) { HMODULE h_kernel = GetModuleHandle ("kernel32"); DWORD retval; assert (h_kernel != NULL); switch (signo) { case SIGKILL: case SIGTERM: case SIGQUIT: case SIGHUP: { sigkill_data d; d.adr_ExitProcess = GetProcAddress (h_kernel, "ExitProcess"); assert (d.adr_ExitProcess); retval = run_in_other_process (h_process, sigkill_proc, &d, sizeof (d)); break; } case SIGINT: { sigint_data d; d.adr_GenerateConsoleCtrlEvent = GetProcAddress (h_kernel, "GenerateConsoleCtrlEvent"); assert (d.adr_GenerateConsoleCtrlEvent); d.event = CTRL_C_EVENT; retval = run_in_other_process (h_process, sigint_proc, &d, sizeof (d)); break; } default: assert (0); } return (int)retval > 0 ? 1 : 0; } -- Adrian Aichner <adrian(a)xemacs.org>

1 participants
0 comments

Re: (interrup-process ...) broken on NT native 24 years, 6 months

Gunnar Evermann

Adrian Aichner <aichner(a)ecf.teradyne.com> writes: > >>>>> "Gunnar" == Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: > > Gunnar> could you elaborate what "doesn't work" means? Does it > > Oops, sorry. It does neither kill the shell, nor the sub-process. > A sleep process not run as a sub-process of a "cmd" shell can be > killed fine either by name or object. > > Gunnar> Have you tried running this in a debugger and single > Gunnar> stepping through process_send_signal()? Alternatively you > > Yes. It gave me a bit of a head-ache :-) :-) > SIGINT under NT is implemented by sending CTRL_C_EVENT. > > I got lost during debug trying to figure out what's wrong. > > Gunnar> could try backing out patches until it works :-( > > Sounds pretty scary without your help. Who knows what interactions > exist between your patches and Bens' and others'. I think there should be no interactions. My changes are pretty trivial and only consist of some extra CHECK_LIVE_PROCESS calls. > >> It would be great if we could fix this for 21.1 even, if not for > >> 21.1.10. > > Gunnar> indeed. I think I never intended my changes to go into > Gunnar> 21.1, for which I had a simpler version, but this should > Gunnar> still work. > > Hmmh, can you give me some pointers to problematic portions of your > patches or point me to the simpler patch on xemacs-patches? http://www.xemacs.org/list-archives/xemacs-patches/199912/msg00066.html In your test cases you have to careful, when you evaluate them. I just evaluated them wrapped in a progn and thought the sleep didn't get killed (on Unix). The real problem was that we never got back to the event-loop and thus status_notify didn't get called before the second message. Put in a (sit-for 1) to be sure. Have you tried using Fdelete_process instead of Finterrupt_process? random other idea: Can you kill the subprocess from outside xemacs and see whether we at least notice that? Gunnar

1 participants
0 comments

Re: (interrup-process ...) broken on NT native 24 years, 6 months

Adrian Aichner

>>>>> "Gunnar" == Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: Gunnar> Adrian Aichner <aichner(a)ecf.teradyne.com> writes: >> I can't figure out which change broke (interrupt-process ...) in >> XEmacs 21.1 and XEmacs 21.2 built natively on NT4SP5. >> >> This is the reason why kill-compilation is no longer working in >> *compile* buffers (needless to say: It used to work to >> approx. 21.1.7). >> >> Perhaps you'll have some ideas. I am copying Gunnar who has also done >> work in the process area. Gunnar> OK, thanks. I am afraid I have followed xemacs-beta very ^ NOT? Gunnar> closely in the last months. >> This doesn't: >> (start-process "shell" nil "cmd" "/c" "sleep" "60") >> (message "before: %S" (get-process "shell")) >> (interrupt-process "shell" t) >> (message "after: %S" (get-process "shell")) Gunnar> could you elaborate what "doesn't work" means? Does it Oops, sorry. It does neither kill the shell, nor the sub-process. A sleep process not run as a sub-process of a "cmd" shell can be killed fine either by name or object. Gunnar> make a difference whether you use the name or the process? No. Gunnar> Have you tried running this in a debugger and single Gunnar> stepping through process_send_signal()? Alternatively you Yes. It gave me a bit of a head-ache :-) SIGINT under NT is implemented by sending CTRL_C_EVENT. I got lost during debug trying to figure out what's wrong. Gunnar> could try backing out patches until it works :-( Sounds pretty scary without your help. Who knows what interactions exist between your patches and Bens' and others'. >> It would be great if we could fix this for 21.1 even, if not for >> 21.1.10. Gunnar> indeed. I think I never intended my changes to go into Gunnar> 21.1, for which I had a simpler version, but this should Gunnar> still work. Hmmh, can you give me some pointers to problematic portions of your patches or point me to the simpler patch on xemacs-patches? Gunnar> Gunnar Thanks in advance, Adrian

1 participants
0 comments

Re: (interrup-process ...) broken on NT native 24 years, 6 months

Gunnar Evermann

Adrian Aichner <aichner(a)ecf.teradyne.com> writes: > I can't figure out which change broke (interrupt-process ...) in > XEmacs 21.1 and XEmacs 21.2 built natively on NT4SP5. > > This is the reason why kill-compilation is no longer working in > *compile* buffers (needless to say: It used to work to > approx. 21.1.7). > > Perhaps you'll have some ideas. I am copying Gunnar who has also done > work in the process area. OK, thanks. I am afraid I have followed xemacs-beta very closely in the last months. > This doesn't: > (start-process "shell" nil "cmd" "/c" "sleep" "60") > (message "before: %S" (get-process "shell")) > (interrupt-process "shell" t) > (message "after: %S" (get-process "shell")) could you elaborate what "doesn't work" means? Does it make a difference whether you use the name or the process? Have you tried running this in a debugger and single stepping through process_send_signal()? Alternatively you could try backing out patches until it works :-( > It would be great if we could fix this for 21.1 even, if not for > 21.1.10. indeed. I think I never intended my changes to go into 21.1, for which I had a simpler version, but this should still work. Gunnar

1 participants
0 comments

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

XEmacs-Beta May 2000