April 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: wierd find-file bug 25 years, 7 months

SL Baur

Rajesh Vaidheeswarran <rv(a)fore.com> writes in xemacs-beta(a)xemacs.org: ... > These are my find-file-hooks: > Value: (whitespace-buffer fume-add-menubar-entry p4-find-file-hook vc-find-file-hook font-lock-set-defaults) > whitespace-buffer is from whitespace.el (http://www.dsmit.com/lisp) > p4-find-file-hook is from p4.el (http://www.dsmit.com/lisp) > a M-x p4-edit used to cause a visited file in a buffer to be checked out of > the p4 VC system (this causes the read-only mode to be changed, refreshing > the buffer which causes all the find-file-hooks to be called). Everything > including cursor position was preserved in XEmacs 20.4, but with XEmacs-21.0 > my cursor goes to point 1. > Any ideas? Can you narrow this down a bit, preferably to a sequence that does not use vc or p4? Or at least vc + rcs or cvs?

1 participants
0 comments

Re: crontab bug 25 years, 7 months

SL Baur

nbecker <nbecker(a)fred.net> writes in xemacs-beta(a)xemacs.org: > When I try to exit crontab-edit using C-x C-s I get: > Signaling: (wrong-type-argument stringp nil) > intern(nil [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ~/\.crontab 0]) > vc-file-getprop(nil vc-backend) This looks like a vc problem not a crontab.el bug. It works fine without it.

1 participants
0 comments

Re: Canna is hosed wrt Backspace and Delete keysyms 25 years, 7 months

SL Baur

Glynn Clements <glynn(a)sensei.co.uk> writes in xemacs-beta(a)xemacs.org: > SL Baur wrote: >> What is (event-to-character event) supposed to return when `event' is >> #<keypress-event backspace> or #<keypress-event delete>? > I don't know what it's supposed to return, but event-to-character > appears to return the symbol's 'ascii-character property. These used > to be set to ?\010 and ?\177 respectively. O.K. The broken code in question looks like: (defun canna-functional-insert-command (arg) "Use input character as a key of complex translation input such as\n\ kana-to-kanji translation." (interactive "*p") (let ((ch)) (if (char-or-char-int-p arg) (setq ch last-command-char) (setq ch (event-to-character last-command-event))) (canna:functional-insert-command2 ch arg) )) The problem appears to be with `last-command-char' being nil. > If at all practical, I would suggest that the code should be using > (event-key event) rather than (event-to-character event). O.K. but I have no idea where the second code path is being taken. The way canna.el does input is broken, but it _used_ to work. Sigh.

2 participants
1 comments

Canna is hosed wrt Backspace and Delete keysyms 25 years, 7 months

SL Baur

[21.0.67 and 21.2.13 behave identically.] Hitting either Backspace or Delete in Canna fence mode yields: Signaling: (wrong-type-argument characterp nil) canna-key-proc(nil) canna:functional-insert-command2(nil 1) byte-code("..." [nil ch char-or-char-int-p arg last-command-char event-to-character last-command-event canna:functional-insert-command2] 3) canna-functional-insert-command(1) call-interactively(canna-functional-insert-command) The only way to delete backwards in fence mode is to type an explicit [(control h)]. This is one of the worst of all possible worlds. What is (event-to-character event) supposed to return when `event' is #<keypress-event backspace> or #<keypress-event delete>?

2 participants
1 comments

wierd find-file bug 25 years, 7 months

Rajesh Vaidheeswarran

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 21.0 "20 minutes to Nikko" [Lucid] (sparc-sun-solaris2.7, Mule) of Fri Mar 26 1999 on agastya configured using `configure --with-mule' Please describe exactly what actions triggered the bug and the precise symptoms of the bug: These are my find-file-hooks: Value: (whitespace-buffer fume-add-menubar-entry p4-find-file-hook vc-find-file-hook font-lock-set-defaults) whitespace-buffer is from whitespace.el (http://www.dsmit.com/lisp) p4-find-file-hook is from p4.el (http://www.dsmit.com/lisp) a M-x p4-edit used to cause a visited file in a buffer to be checked out of the p4 VC system (this causes the read-only mode to be changed, refreshing the buffer which causes all the find-file-hooks to be called). Everything including cursor position was preserved in XEmacs 20.4, but with XEmacs-21.0 my cursor goes to point 1. Any ideas? Thanks rv Recent keystrokes: C-M-p C-M-n C-M-p C-M-n C-a C-up C-up C-up C-up C-up C-up down C-M-n C-M-p C-a next prior next C-s d e f u n SPC p 4 - c h e c k RET C-a C-M-n C-M-p down down down down down down down down next prior C-a C-up C-up C-up up C-x b RET C-x k RET M-x b y t e - c o m TAB f TAB RET RET C-x 1 M-x g n u s RET down down RET M-> M-x e m a c s - v TAB RET q q y misc-user misc-user misc-user Recent messages (most recent first): Composing a message...done Wrote /us/rv/Mail/mh/drafts/4 Composing a message... Loading emacsbug...done Loading emacsbug... Saving /us/rv/.newsrc.eld...done Wrote /us/rv/.newsrc.eld Saving /us/rv/.newsrc.eld... Wrote /us/rv/.newsrc Are you sure you want to quit reading news? (y or n) Yes

1 participants
0 comments

Re: [TPD] First patch 25 years, 7 months

Olivier Galibert

[Redirected to xemacs-beta] On Mon, Apr 05, 1999 at 09:30:52AM -0700, Kyle Jones wrote: > Didier Verna writes: > > Kyle Jones <kyle_jones(a)wonderworks.com> writes: > > > > > I was hoping we could add big changes like the new dumper with > > > #ifdefs so the dumper could be paritally implemented and broken > > > and normal testing/debugging could continue. This doesn't look > > > like the diretion you're heading. I hope it's not too late to > > > change your mind. > > > > I don't know how much work it would involve to #ifdef the new dumper > > code, but in general, for really big changes like this (and that are supposed > > to replace former stuff), how about opening a new CVS branch with a special > > tag[1]? It avoids the problems of making the new code coexist with the old > > one, and postpone the integration process until merge time. > > Someone has to maintain both trees, and I seem to remember that > being a serious deterrent to forking the tree. I think it is a > good idea nonetheless, assuming we have maintainers for both > trees. That's the main interest of prcs. You can merge the modifications of the main tree easily, while with cvs.... Note that I don't advocate prcs yet, I know we can't use it. > Olivier, could you outline how your dumper is going to work? I > remember doubting that it was actually portable, but we argued a > long time ago and I don't trust my memory. I think you're refering to Ben's way. Next patch will add the ability to move lrecords around in memory. I see two main ways to do that, I'm still thinking about them (content description vs. moving method). I'll probably add packing to the current gc in order to test the system. The gc modifications should be ifdef-able. The data or code added to lrecords won't, but it won't be used if the gc is ifdef'ed out. The dumping is quite similar to yours: - gather everything that has to be dumped - sort between readonly/writeable - create a file with all the data in it (using the moving code) Mine is a binary file, yours C source. Reloading is done by: - mmap the file if you can, else malloc and read() - fix the adresses if needed (using the moving code too) - hook up in the global variables Your linked-in C code avoid the two first points, but, I think, has other limitations, mostly by requiring a C compiler. I'd like to see, in the future, partial dumps done for some packages (aka, fast package loader ?) even for people using binary releases. Tradeoffs. Also, there will be the possibility to change the base address in the dump file so that fixing the addresses isn't needed. The idea being that there is no real reason why the address returned by mmap() should change between different invocations of xemacs. Note that modern VMs use sharing and copy-on-write on mmap'ed files. OG.

1 participants
0 comments

Re: Crash in XEmacs 21.0 "20 minutes to Nikko" with gnuserv/gnuclient 25 years, 7 months

Jan Vroonhof

Colin Rafferty <craffert(a)ms.com> writes: > > P.S. Aren't all the linux consoles really one console in the XEmacs > > sense anyway? > > That would surprise me, to say the least. I was just musing... > Each particular console (or any other tty device) is considered its > own device. You can create different frames on that console, but they > are logically stacked on top of each other, and can only be switched > among using `C-x 5 o'. I mean that on the linux console there is really only one keyboard and mouse stream, i.e. one XEmacs console. Each one "linux console" would then be another device on this same console. I.e. finally a case where the XEmacs console/device separation would actually make some sense. These problems are probably caused by the schizofrenic nature of the linux console input. Each virtual console has its own keyboard input stream but it seems there is only 1 mouse stream. Jan

1 participants
0 comments

Re: Crash in XEmacs 21.0 "20 minutes to Nikko" with gnuserv/gnuclient 25 years, 7 months

Colin Rafferty

Jan Vroonhof writes: > P.S. Aren't all the linux consoles really one console in the XEmacs > sense anyway? That would surprise me, to say the least. Each particular console (or any other tty device) is considered its own device. You can create different frames on that console, but they are logically stacked on top of each other, and can only be switched among using `C-x 5 o'. -- Colin

1 participants
0 comments

crontab bug 25 years, 7 months

Neal D. Becker

When I try to exit crontab-edit using C-x C-s I get: Signaling: (wrong-type-argument stringp nil) intern(nil [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ~/\.crontab 0]) vc-file-getprop(nil vc-backend) vc-after-save() basic-save-buffer() save-buffer(1) call-interactively(save-buffer)

1 participants
0 comments

Re: Comparison between integer and character is constant nil (?2 and 0) 25 years, 7 months

SL Baur

Stephen J Turnbull <turnbull(a)sk.tsukuba.ac.jp> writes in xemacs-beta(a)xemacs.org: >>>>>> "Kyle" == Kyle Jones <kyle_jones(a)wonderworks.com> writes: Kyle> If we're getting more bogus warnings than real warnings then Kyle> that is the signal that the ebola-check code should be Kyle> removed. > This is not true for me. Of the two Ebola warnings I remember seeing > in the last 6 months, the one in AUC-TeX is due to use of > `previous-char' instead of `char-before' (a distant cousin of Ebola), I thought the consensus on replacement of `previous-char' with `char-before' was to _not_ do it, since `previous-char' isn't broken any more.

3 participants
2 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