January 2000 - XEmacs-Beta - XEmacs Mailing List Archives

Strange lookup auto-autoloads 25 years, 2 months

Nick V. Pakoulin

My xemacs-21.2.27 is built without MULE, lookup version 1.04, warning on startup: (1) (warning/warning) Autoload error in: /usr/local/lib/xemacs/xemacs-packages/lisp/lookup/auto-autoloads: Cannot open load file: mule It's strange, isn't it? Nick.

3 participants
2 comments

more on emacspeak and keeping the mark active: 25 years, 2 months

T. V. Raman

I'd ideally like to get this working without tampering with zmacs-regions. So here goes. Type some text, set the mark. The mark is set correctly as evinced by moving the cursor and executing C-x C-x (I've seen the zmacs-regions documentation and it does say explicitly that c-x c-x is one command that will still work even if the mark becomes inactive...) So going back to the test, with zmacs-regions set to the default value of t type some text; set the mark by executing C-x h --region is hopefully active. now hit C-w --with emacspeak loaded this throws a "mark is not active now" --so here is what emacspeak does to the kill-region command --hopefully this will shed some light on what is deactivating the mark. I'll annotate the advice definition some for those of you who dont hack advice. ;;; around advice means you get called *instead of * the ;;;original defun. ;;; in the advice below, we will check if we were called ;;; interactively --and if so --speak some info about what is ;;;getting nuked by kill-region (defadvice kill-region (around emacspeak pre act) "Indicate region has been killed. Use an auditory icon if possible." (cond ((interactive-p) ;;;so let's produce auditory feedback (let ((count (count-lines (region-beginning) (region-end)))) ;;;call original version of kill-region ad-do-it (emacspeak-auditory-icon 'delete-object ) (message "Killed region containing %s lines" count))) ;;;called programmatically so just default behavior (t ad-do-it)) ;;;return computed value so caller doesn't get confused by ;;;advised version ad-return-value) -- Best Regards, --raman Email: raman(a)cs.cornell.edu WWW: http://cs.cornell.edu/home/raman/ PGP: http://cs.cornell.edu/home/raman/raman.asc

1 participants
0 comments

Re: the culprit is zmacs-regions 25 years, 2 months

Colin Rafferty

T V Raman writes: > found the problem --or at least its cause --(mark) in xemacs returns > nil even if the mark is set. Only when (eq zmacs-regions t), and the mark has been made inactive. I imagine that emacspeak runs commands all the time, and these commands deactivate the mark, so (mark) returns nil. Any of these commands should set `zmacs-region-stays' to `t' so that the region will stay active (or not). > Emacspeak examines the mark value for all kinds of things --especially > to indicate to the user what the > is being deleted etc -- If the region is not active, then C-w will not work, and you won't need to tell the user anything anyway. Alternatively, you can force people to set `zmacs-regions' to `nil' to use emacspeak, but that seems evil. -- Colin

2 participants
1 comments

Re: mark in xemacs with emacspeak loaded 25 years, 2 months

Hrvoje Niksic

"T. V. Raman" <ramantv(a)earthlink.net> writes: > With emacspeak loaded, if I: Hit C-spc to set the mark, cursor over > by even one character Does emacspeak redefine the cursor-movement functions? In XEmacs the zmacs region gets deactivated unless the calling function explicitly chooses to leave it active. This is by design: the only functions that are supposed to leave the mark on are movement functions, C-SPC itself, and such. To specify that the mark should remain on, the interactive specification should start with _. Alternatively, you can set `zmacs-region-stays' to t at the end of the function.

2 participants
1 comments

Re: gnuclient's immediate exit 25 years, 2 months

Jan Vroonhof

Karl Kleinpaste <karl(a)justresearch.com> writes: > - ...and then immediately disconnect and exit. > > Where in this logic path does the possibility of having gnuclient wait > for `C-x 5 0' even exist? I feel I must be missing something obvious. The waiting for the final result from the server is done in disconnect_from_server (in gnuslibc.c). Jan

1 participants
0 comments

Re: Can't build some texinfo files 25 years, 2 months

Mats Lidell

>>>>> I wrote: I> For some texinfo files I get problems with xref. Example from I> ./comm/bbdb. I> ---------------------------------------------------------------------- I> $ make bindist I> (cd texinfo; makeinfo -o bbdb.info bbdb.texinfo) I> bbdb.texinfo:404: `.' or `,' must follow cross reference, not ). I> bbdb.texinfo:1004: `.' or `,' must follow cross reference, not f. I> bbdb.texinfo:1022: `.' or `,' must follow cross reference, not f. I> bbdb.texinfo:1066: `.' or `,' must follow cross reference, not ). I> [...] I> ---------------------------------------------------------------------- Since this didn't attend anybodies attention I RTFM an found that ＠xref is supposed to be like that and that a ＠pxref seems to be just what should be there instead. So I changed the offending xrefs to prefs (in bbdb.texinfo) and at least it did makeinfo happy. (I haven't checked that the cross references are OK.) (Maybe I should point out that I see this problem in a few other texinfo-files as well and it stops med from building all packages from the top.) So what's up. I must have gotten something wrong or nobody is trying to build the packages the way I do. (make -r bindist) Which of course I find utterly, utterly unbelievable. Yours -- %% Mats

1 participants
0 comments

gnuclient's immediate exit 25 years, 2 months

Karl Kleinpaste

I reported in -b26 that gnuclient was exiting immediately after being started, which confused me especially because it departed with status zero, but I saw no exit of that sort. Kyle pointed out that gnuclient's main() actually does "return 0;", which makes sense as far as that goes. So I just walked through gnuclient with gdb, and I simply don't understand the path taken. My short summary of gnuclient's path is: - initialize a bunch of general state/option variables. - check for batch operations and window suppression ("-nw"). - connect to xemacs via connect_to_unix_server(). - get xemacs' pid. - disconnect. - set signal state (in the -nw case). - connect to xemacs again. - ship over a gnuserv-edit-files command. - ...and then immediately disconnect and exit. Where in this logic path does the possibility of having gnuclient wait for `C-x 5 0' even exist? I feel I must be missing something obvious. 21.2.27, Linux RH6.0. --karl

2 participants
1 comments

Re: Drag-N-Drop 25 years, 2 months

Gunnar Evermann

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > Olivier Graf is supposed to be working on Xdnd support (which is the > standard GNOME and KDE agreed on). > > However we haven't heard from in 6 months or so. If you want to work on it > I suggest you try to contact him first. If he is to busy it would be > nice if somebody could take over. speaking of DnD... I propose we disable DnD by default at least for all 21.1.x versions. I think Oliver agreed to do this a long time ago but nobody ever bothered to send the patch. Could one of the configure gurus make a patch for that? thanks! The DragMotionProto crash still hurts lots of people on Solaris and I don't see us fixing it anytime soon unless somebody with access to Sun's Openwin sources comes along... Gunnar

3 participants
3 comments

xemacs 21.2.27 25 years, 2 months

T. V. Raman

There are a couple of weirdnesses I noticed with xemacs 21.2.27 --that are possibly bugs: 1) I built xemacs in /disk/local3/... on my machine and then did a make install to put the thing in /usr/local/lib/... However, the thing still "remembers" where it was built --for instance describe-variable emacs shows that the variable is defined in /disk/local3/... 2) functions find-function and find-function-on-key both fail --this is probably related to (1) 3) when running on a tty (dont know if it is true under X) Go to the scratch buffer, type some text, hit C-x h (mark-whole-buffer) and hit C-w --xemacs says "region is not active now" or some such. For an instant it felt like FSF emacs with transient mark permanently on --I checked and find that xemacs doesn't have a transient-mode variable. (the above is a serious bug --and makes xemacs unusable for me at present) Thanks, --Raman -- Best Regards, --raman Email: raman(a)cs.cornell.edu WWW: http://cs.cornell.edu/home/raman/ PGP: http://cs.cornell.edu/home/raman/raman.asc

3 participants
3 comments

Re: the culprit is zmacs-regions 25 years, 2 months

Hrvoje Niksic

"T. V. Raman" <raman(a)earthlink.net> writes: > found the problem --or at least its cause --(mark) in xemacs returns > nil even if the mark is set. (mark) returns nil if the region is not active. Use (mark t) to find the value of mark regardless of whether the region is active.

1 participants
0 comments

2025

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 January 2000