April 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: 21.0.67 autosave on crash broken? 25 years

Jan Vroonhof

Marat Boshernitsan <maratb(a)cs.berkeley.edu> writes: > It appears that 21.0.67 does not autosave modified buffers on a crash, > despite the crash message saying so. Indeed. The crash message is wrong. The idea being that this is in the fatal error handler. If the fatal error handler was something like SIGSEGV or SIGBUS you could be overwriting your precious autosave files with corrupted data. The best way to fix this would probably to have 2 different signal handlers depending on how fatal the fatal error is. Things like SIGHUP can then still cause autosaving. However nobody did that yet. Jan

1 participants
0 comments

Re: "Russian" Mule crash in 21.0.67 25 years

Jan Vroonhof

Olivier Galibert <galibert(a)pobox.com> writes: > On Fri, Apr 23, 1999 at 12:28:06AM +0100, Gunnar Evermann wrote: > > (decode-coding-string " \207 " 'koi8-r) C-j > > koi8, ccl, *cry* > > Unless someone fixes this since then, I'll try to have a look next > week. There is a big Mule patch by Tomohiko in the queue. It has a " Also decode characters not in koi8 correctly" or equivalent in the Changelog. I am hoping (I haven't checked) that will fix it. Jan

1 participants
0 comments

Re: new 21.2? 25 years

Jan Vroonhof

Colin Rafferty <craffert(a)ms.com> writes: > For example, imagine I had a string at file scope defined like this: > > const char * const DocString = "....."; > > If it is only used in the C part, would it get marked? No, Only lisp strings have properties. I think the marking can thus be avoided by proclaiming that read-only strings only can have a "nil" plist. Jan

1 participants
0 comments

any motif experts out there? 25 years

Gunnar Evermann

In lwlib/xlwmenu.c we call XmStringLtoRCreate at lines 443 (see below) and also at 814. Is it possible that those calls fail? Should we check? I have a report of a crash somewhere around there. Alas I didn't even find a man page for this function. any comments? #ifdef NEED_MOTIF newstring = XmStringLtoRCreate (newchars, XmFONTLIST_DEFAULT_TAG); XmStringExtent (mw->menu.font_list, newstring, &width, &height); XmStringFree (newstring); XtFree (chars); return width; #else Gunnar -- Gunnar Evermann Speech, Vision & Robotics Group Engineering Department Cambridge University

2 participants
1 comments

Re: "Russian" Mule crash in 21.0.67 25 years

Gunnar Evermann

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > xemacs-21.0.67 -batch -vanilla -l crashkoi.el > > With crashkoi.el > > (with-current-buffer (get-buffer-create "crash") > (erase-buffer) > (insert " \207 ") > (decode-coding-region (point-min) (point-max) 'koi8-r) > (goto-char (point-max)) > (re-search-backward "^-") > ) try: xemacs -nw -vanilla (decode-coding-string " \207 " 'koi8-r) C-j also looks wrong. The cursor doesn't move, the window is cleared and at the beginnin of the first line " ~ " appears. Scrolling, C-l and pressing RET all behave very weird as well this is emacs-version "21.2 \"Demeter\" XEmacs Lucid (beta13)" Gunnar -- Gunnar Evermann Speech, Vision & Robotics Group Engineering Department Cambridge University

2 participants
1 comments

Re: XEmacs crashes on empty menus 25 years

Gunnar Evermann

Tudor Hulubei <tudor(a)cs.unh.edu> writes: > Click on Options, then on Crash. You will get an item saying "No > menu". Try going to that item with the keyboard (type right-arrow) > instead of the mouse. XEmacs will crash. For these "No menu" entries the accel field didn't get initialised properly. 1999-04-23 Gunnar Evermann <ge204(a)eng.cam.ac.uk> * menubar-x.c (pre_activate_callback): set accelerator field in "No menu" entries to nil. Avoid crash in command_builder_operate_menu_accelerator Index: src/menubar-x.c =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs/src/menubar-x.c,v retrieving revision 1.18.2.4 diff -u -r1.18.2.4 menubar-x.c --- menubar-x.c 1998/12/18 23:24:01 1.18.2.4 +++ menubar-x.c 1999/04/23 08:33:14 ＠＠ -465,10 +465,12 ＠＠ wv = xmalloc_widget_value (); wv->type = CASCADE_TYPE; wv->next = NULL; + wv->accel = LISP_TO_VOID (Qnil); wv->contents = xmalloc_widget_value (); wv->contents->type = TEXT_TYPE; wv->contents->name = (char *) "No menu"; wv->contents->next = NULL; + wv->contents->accel = LISP_TO_VOID (Qnil); } assert (wv && wv->type == CASCADE_TYPE && wv->contents); replace_widget_value_tree (hack_wv, wv->contents); -- Gunnar Evermann Speech, Vision & Robotics Group Engineering Department Cambridge University

1 participants
0 comments

21.0.67 autosave on crash broken? 25 years

Marat Boshernitsan

It appears that 21.0.67 does not autosave modified buffers on a crash, despite the crash message saying so. I can reliably reproduce this on both Linux and Solaris by doing "kill -HUP" on running xemacs with a buffer that has been modified since last save/autosave. What gives? Marat.

1 participants
0 comments

Re: core dumped 25 years

Gunnar Evermann

Hi Nick, do you still have the core file and executable around? The upper part (above frame 11) is uniteressting. During GC we can't print a Lisp bactrace. "Nick V. Pakoulin" <npak(a)kazbek.ispras.ru> writes: > #11 <signal handler called> > #12 0x807e1de in garbage_collect_1 () at alloc.c:3160 > #13 0x80a6681 in Ffuncall (nargs=1, args=0xbfffe9bc) at eval.c:3135 > #14 0x80a7389 in run_hook_with_args_in_buffer (buf=0x8876610, nargs=1, > args=0xbfffe9bc, cond=RUN_HOOKS_TO_COMPLETION) at eval.c:3671 somehow the line number in frame 12 seems wrong. Could you go to that frame and look what statement was executed and what kind of signal was caught? Gunnar -- Gunnar Evermann Speech, Vision & Robotics Group Engineering Department Cambridge University

1 participants
0 comments

XEmacs/Windows NT installer 25 years

A.J. Rossini

Who can I talk with about getting a clue as to how the windows nt xemacs installer is constructed/set up? (a friend on a related GNU-licensed project, with lots of optional packages, was curious, after being "wow, this is fantastic!"'d by the whole thing). best, -tony -- A.J. Rossini Research Assistant Professor of Biostatistics Center for AIDS Research UW Biostatistics 206-720-4282 (4209=fax) 206-543-1044 (xxxx=fax) rossini(a)u.washington.edu rossini(a)biostat.washington.edu http://www.biostat.washington.edu/~rossini/

2 participants
1 comments

Re: The patch to end all redisplay problems. 25 years

Gunnar Evermann

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > 1999-04-02 Jan Vroonhof <vroonhof(a)math.ethz.ch> > > * xemacs-20/src/window.c (struct window_config): Removed frame > size members. > (window_config_equal): No longer compare frame sizes. > (Fset_window_configuration): Resize old top window to fit in the > current frame directly, no longer use a fake frame resize. > (Fcurrent_window_configuration): No longer save frame size I have been running XEmacs with this patch for a couple of weeks and just noticed, that resizing of minibuffers is broken (I think it worked before). xemacs -nw -vanilla M-x resize-minibuffer C-x C-f enter a really long name, so that the minibuffer is resized to two lines (everything OK so far) RET minibuffer stays two lines high and only the bottom line is used. This also happens under X, with the added benefit that you get funny redisplay glitches if you drag the modeline after this... I think I have backed out all my other temporary redisplay fiddling, but I am not really sure. Gunnar -- Gunnar Evermann Speech, Vision & Robotics Group Engineering Department Cambridge University

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 April 1999