April 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: XEmacs and its multiuser roots (was Re: change \C-m binding to newline-and-indent) 25 years

Jan Vroonhof

Raymond Toy <toy(a)rtp.ericsson.se> writes: > When a user gets a package, does the package get installed in the > XEmacs package directory or does it go into the users local area? By default is goes to the XEmacs package directory. The idea is that in the future this is changed to the users local area when the user has no write permission on central files. > If the package goes into XEmacs package directory, that could be > really bad if random people update packages whenever they feel like > it and unknowningly hose others who depended on some behavior of the > original version. The package system assumes this is taken care of by normal users not having write permission to the central stuff. I guess that could give problems on Windows 9x. Jan

1 participants
0 comments

Re: Problem with fonts under NT 25 years

Jan Vroonhof

Matt Stupple <matts(a)tibco.com> writes: > As I don't have CVS access (don't you just love firewalls?) I'm stuck > on 21.0.66 for the moment. 21.0.67 is on the FTP site. It is _not_ in the beta directory, but under the xemacs-21.0 release directory. This is because 21.0.67 is released in some sense[1]. Jan Footnotes: [1] This situation is confusing. We did a "release" (champagne and all) in Tsukuba. But it was not widely announced. Moreover the extra ".67" is confusing people (and even XEmacs itself because 21.0.67 itself thinks it is a beta, c.f. splash screen, ebola warnings, debugging code etc). Personally I think this was all a trick by Steve to get at the champagne and I hope there will be a real release (without a patch level) soon.

1 participants
0 comments

Re: XEmacs and its multiuser roots (was Re: change \C-m binding to newline-and-indent) 25 years

Raymond Toy

>>>>> "Holger" == Holger Schauer <schauer(a)coling.uni-freiburg.de> writes: Holger> [packages should be installed by sysadmins] sb> Users have always been encouraged to install their own lisp. Holger> That is something that gives any sysadmin the headaches. I have been Holger> bitten much to often by having old packages lying around, breaking new Holger> clean installs (mailcrypt, VM and bbdb, anybody ?). Holger> What you say is certainly true for private additions and Holger> configurations - but it is a very different story, if all users keep Holger> their own "newest version of package xy, since that bloddy sysad never Holger> installs it" - and the sysadmin in fact is only one day late or has Holger> good reason not to install a package. When a user gets a package, does the package get installed in the XEmacs package directory or does it go into the users local area? (I've never used package-get because I've never gotten around to fixing efs to understand our ftp system here to get through our firewall.) If the package goes into XEmacs package directory, that could be really bad if random people update packages whenever they feel like it and unknowningly hose others who depended on some behavior of the original version. Ray

1 participants
0 comments

[Bug #588] x-win-init-sun-crash, pure copying broken. 25 years

Jan Vroonhof

This bug is a GC bug for sure The simplest way to see the crash is to type this in *scratch* on a sun. (aref (compiled-function-constants (symbol-function 'x-win-init-sun)) 0) and do "C-j". It doesn't crash if you "(aref ... 1)" Important points 1. You need to be on a sun and on X because 'x-win-init-sun needs to be run and a Garbage collect must happen. If you are not on a sun or not on X do: xemacs -nw -vanilla (unintern 'x-remap-keysyms-using-function-key-map) C-j (garbage-collect) C-j (aref (compiled-function-constants (symbol-function 'x-win-init-sun)) 0) C-j 2. Here is the relevant part of x-win-init-sun (defun x-win-init-sun () (defun x-remap-keysyms-using-function-key-map (from-key to-key) ...... (vector (append prefix (list to-key)))))) .... (unintern 'x-remap-keysyms-using-function-key-map) ... byte code for x-win-init-sun: args: nil 0 constant x-remap-keysyms-using-function-key-map 1 constant <compiled-function> args: (from-key to-key) 3. x-win-init-sun is dumped. Theory: (aref (compiled-function-constants (symbol-function 'x-win-init-sun)) 0) is the symbol x-remap-keysyms-using-function-key-map which is uninterned. The constants vector is the only reference. However this vector is in pure space and thus is not swept on GC. Therefore there is no link to the symbol and thus its gets GC. Thus we get boom boom. I am not sure this what Martin actually fixed. It might well be that the bug is just hidden. Jan

1 participants
0 comments

Behaviour of call-process 25 years

Enrico Scholz

Hi, at my system (i686-linux-gnu, glibc2.1, kernel 2.2.5) call-process is not interruptable by "C-g" anymore. E.g. at xemacs-21.2-13 (call-process "sleep" nil nil nil "10") can not be stopped by this keypress (xemacs-20.4 at another machine allows it). That's why some modules, for example the voice-recording in tm-edit, won't be working anymore because they depends on it. Is this a documented behaviour or just a bug? Enrico -- eMail: enrico.scholz(a)wirtschaft.tu-chemnitz.de talk: ensc(a)ultra.csn.tu-chemnitz.de

1 participants
0 comments

Re: change \C-m binding to newline-and-indent 25 years

Tor Arntsen

On Apr 6, 11:39, SL Baur wrote: >The netscape bindings are courtesy of Jamie, I'm fairly sure. .. >Framemaker and its Emacs keybindings predates both Motif and Microsoft >Windows. It predates a stable X11 for that matter -- the first version >I used (1.0) ran under SunView. Those bindings saved my sanity, that's what they did. As far as the Emacs keybindings are concerned I got used to many of them years before I started to use Emacs proper (I didn't before I found Lucid Emacs). I used to work on some bizarre systems where I survived by carrying my own much hacked version of microemacs around, porting it along the way. When I (much later) started to use Unix (only ten years ago) I quite quickly came in touch with Emacs bindings again -- after my initial exposure to csh I thought "there must be something better than this!" and immediately tried out all shells I could get my hands on. And soon I found emacs bindings again. - Tor

1 participants
0 comments

XEmacs and its multiuser roots (was Re: change \C-m binding to newline-and-indent) 25 years

SL Baur

Holger Schauer <schauer(a)coling.uni-freiburg.de> writes in xemacs-beta(a)xemacs.org: >>>>> "sb" == SL Baur schrieb am 02 Apr 1999 01:49:37 -0800: sb> ... I am not opposed to making it a default if it is installed sb> (ie. by having it autoload initialization). > IMHO, this would be another case that shows that XEmacs is starting > to forget its multi-user roots [1]. Is this such a problem? The XEmacs user base has changed dramatically over the last three years in favor of systems that are typically single-user. If XEmacs/(95|98|00|NT) ever catches on, it will only further this trend. I'm not expressing an opinion one way or another, I'm just asking. > What if a sysadmin would install that package on a single user's > request - afterwards all 999 traditional users are stuck with these > silly key bindings ? That's a good point. Consider the suggestion dropped[1]. I've never used XEmacs (or GNU Emacs) in a multiuser environment by the way -- twelve years and counting ... ... > Footnotes: > [1] I whole-heartly believe that the package-interface, as nice as it > is, and as often as I use it, is another such case. No normal user > should install packages, that should be a task left for the system > adminstrator (and yes, I used "should" on purpose). Agreed, but the package fetching UI is very new. There are no doubt plenty of refinements to make. There is also nothing in the model that forces system administrators to allow users to install lisp into the common system directories. There's good reason to disallow it even in single-user systems -- I'm pretty much the only user of my development machine, but the `steve' account doesn't have write access to anything system-wide. Users have always been encouraged to install their own lisp. What hasn't been possible has been for users to install their own lisp and bind them as tightly to Emacs as the things the Emacs maintainers have thrown at them. I see nothing wrong with allowing that. Anyway, granting that "XEmacs is forgotting its multi-user roots" whether or not this is in fact the case, my question is "What difference does it make?" This is not a flame. Footnotes: [1] I plead temporary insanity and a bad case of jet lag.

1 participants
0 comments

Re: change \C-m binding to newline-and-indent 25 years

SL Baur

Tor Arntsen <tor(a)spacetec.no> writes in xemacs-beta(a)xemacs.org: > On Apr 6, 9:48, Holger Schauer wrote: >> Another issue: has anybody noticed the more and more widespread use of >> readline, i.e. emacs key-bindings in other applications ? I don't think it's any more widespread than it has been for years. >> Heck, I can even use C-a, C-k etc. in Netscape, and in a bunch of >> other applications, too. The netscape bindings are courtesy of Jamie, I'm fairly sure. >> Converts from Windows are not the only kind of people we're dealing >> with. > Some of the emacs keybindings even work in Framemaker! Framemaker and its Emacs keybindings predates both Motif and Microsoft Windows. It predates a stable X11 for that matter -- the first version I used (1.0) ran under SunView.

1 participants
0 comments

Re: change \C-m binding to newline-and-indent 25 years

Tor Arntsen

On Apr 6, 9:48, Holger Schauer wrote: >Another issue: has anybody noticed the more and more widespread use of >readline, i.e. emacs key-bindings in other applications ? Heck, I can >even use C-a, C-k etc. in Netscape, and in a bunch of other >applications, too. Converts from Windows are not the only kind of >people we're dealing with. Some of the emacs keybindings even work in Framemaker! - Tor

1 participants
0 comments

Re: change \C-m binding to newline-and-indent 25 years

Holger Schauer

[Sorry, I don't have an opinion on \C-m which hasn't already expressed better by others - I'm against it.] >>>>"sb" == SL Baur schrieb am 02 Apr 1999 01:49:37 -0800: sb> I think changing the [(control m)] binding sucks, but I don't sb> intend to fight against a strong consensus. I despise rebinding sb> [(control c)], [(control x)], and [(control v)] with a passion, sb> however if someone can make Kim Storm's code work with XEmacs, I sb> am not opposed to making it a default if it is installed (ie. by sb> having it autoload initialization). IMHO, this would be another case that shows that XEmacs is starting to forget its multi-user roots [1]. What if a sysadmin would install that package on a single user's request - afterwards all 999 traditional users are stuck with these silly key bindings ? Another issue: has anybody noticed the more and more widespread use of readline, i.e. emacs key-bindings in other applications ? Heck, I can even use C-a, C-k etc. in Netscape, and in a bunch of other applications, too. Converts from Windows are not the only kind of people we're dealing with. Holger Footnotes: [1] I whole-heartly believe that the package-interface, as nice as it is, and as often as I use it, is another such case. No normal user should install packages, that should be a task left for the system adminstrator (and yes, I used "should" on purpose). -- --- http://www.coling.uni-freiburg.de/~schauer/ --- "While it may look like it at a casual glance, XEmacs does not link against every library in Known Space." -- SL Baur on xemacs-beta

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