January 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: rssh.el interactions with ange-ftp/efs 25 years, 3 months

Kai Großjohann

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > [...] and have Emacs figure out the best way of accessing based on > user and site configuration. [...] I'm not sure that a table that maps host names (and/or user names) to access methods is the right thing. How about file names for FTP, file names for rsh/rcp, and general remote file names, and the latter means that all available methods are tried until one succeeds? Dunno what one would choose for the file name patterns, though. Right now, we have "/user＠host:/path/to/file" for ange-ftp. Maybe this should become the general syntax and we should use "/ftp:user＠host:/path/to/file" and "/rcp:user＠host:/path/to/file" for the more specific syntaxes? ("/:" could then become an abbreviation for "/verbatim:".) kai -- Just because I'm paranoid doesn't mean they're not out there to get me!

10 participants
26 comments

Fwd: Re: My ISP refused to install Emacs 25 years, 3 months

Olivier Galibert

I wonder... the DOC file is often causing problems on dual mule/non-mule installations. Do we really want it instead than in the executable ? I don't really think it would use more memory, because of the way VMs work. Quite the contrary, actually, given that it is readonly and hence shareable. OG. From: Erik Naggum <erik(a)naggum.no> Newsgroups: comp.emacs,comp.editors Subject: Re: My ISP refused to install Emacs * Erik Naggum | note: I have not optimized for small, quite the contrary, in fact: | my custom Emacs is dumped with all the documentation strings, so | they take up memory, because I didn't like the DOC file access | patterns. * David Kastrup <dak(a)mailhost.neuroinformatik.ruhr-uni-bochum.de> | Ever tried to move the DOC files from the mountable tape you have them | on to disk? funny you should mention that. I tried that first, but then I noticed that NFS was still a significant drag on performance, so I kept the DOC file open, and only closed it when accessing another file for its doc strings, and then I read and cached entire disk blocks, which I think got into the distributed Emacs, and then I decided that this loading form disk business was really the cause of all the problems. better to trust the paging mechanisms than to open files all the time. and it worked just as I had predicted it would: documentation-string-intensive operations like APROPOS complete in non-detectible time, compared to very much detectible time, and all of it being spent on waiting for NFS servers. on local disks, it still causes a momentary burst of completely unnecessary disk I/O -- Emacs does not even remember the doc strings it has previously read from disk. it's not like it's noticeable on my home machine, a dual 400MHz PII with 512M RAM and 40+G of very fast disks, but on the much smaller and slower machines I deploy the software I write on, stuff like unpredictable network performance is now completely avoided. however, it's noticeably faster on the old 50MHz SPARC with fairly slow disks that is machine with the large displays and stuff, too, so I guess only people who have never actually used mountable tapes could be so goddamn stupid as to think that you'd have to go back that far in technology to notice the wrong-headed design decision to optimize for fast disks and little memory, which is really what's behind the decision not to dump with documentation strings, so I think this is tad more significant than you do, obviously, but I hope you have a less idiotic comment to come up with in reply, OK? #:Erik -- SIGTHTBABW: a signal sent from Unix to its programmers at random intervals to make them remember that There Has To Be A Better Way.

1 participants
0 comments

Re: Building a Binary Distribution 25 years, 3 months

Colin Rafferty

David Maslen writes: > I'd like to install it on my laptop, and I don't want to rebuild it > from source there. > Is there a make option to produce a tarball for an easy install in the > /usr/local/ directory on my laptop? I didn't see one. I would do this by mounting the laptop's /usr/local directory onto your desktop's machine, and install it directly. So if you mount laptop:/usr/local onto localhost:/mnt/local, you could use the configure option: --prefix=/mnt/local. Then just `make install. -- Colin

1 participants
0 comments

NT bug with face colors and new frames [was Re: few problems with b62 on NT] 25 years, 3 months

Greg Klanderman

(please copy me on replies) >> OK, try >> xemacs -q >> and eval >> (progn >> (set-face-foreground 'default "white") >> (set-face-background 'default "midnightblue") >> (require 'font-lock) >> (set-face-foreground 'font-lock-comment-face "#ff8080")) >> now load a .el file. note the comment color. >> C-x 5 2 >> and note the different comment color. Andy> Andy> The bizarre thing is that if you create additional frames they come up in Andy> the correct color briefly and then switch to the wrong ones. OK try this: (require 'font-lock) => font-lock (face-property 'font-lock-comment-face 'foreground 'all) => ((global (nil . "blue4"))) (set-face-background 'default "red") => "red" (face-property 'font-lock-comment-face 'foreground 'all) => ((global (nil . "blue4"))) (set-face-foreground 'font-lock-comment-face "white") => "white" (face-property 'font-lock-comment-face 'foreground 'all) => ((global (nil . "white"))) (make-frame) => #<mswindows-frame "XEmacs" 0xcdd> (face-property 'font-lock-comment-face 'foreground 'all) => ((#<mswindows-frame "XEmacs" 0xcdd> (nil . "gray80")) (global (nil . "white"))) Creating the frame is adding all these frame-specific color specifiers for the new frame onto the face. WTF!? Greg

1 participants
0 comments

Re: Package update? 25 years, 3 months

Jan Vroonhof

Heiko Muenkel <muenkel(a)tnt.uni-hannover.de> writes: > My package hm--html-menus has the package version 1.09 in the CVS > tree, but the latest version in the package directory is 1.08. What > should I do to get 1.09 in the package directory? Presumably somebody should give the deamon that rebuilds the packages a kick. Which is a good idea since we are in the release-imminent fase once again. Jan

1 participants
0 comments

Package update? 25 years, 3 months

Heiko Münkel

My package hm--html-menus has the package version 1.09 in the CVS tree, but the latest version in the package directory is 1.08. What should I do to get 1.09 in the package directory? Heiko

1 participants
0 comments

XCAR/XCDR as lvalues 25 years, 3 months

Hrvoje Niksic

Many of the recent performance changes included replacing Fsetcar() and Fsetcdr() with XCAR/XCDR on the left side of an assignment. After thinking about it, I must conclude that these changes are dangerous in the general case where we don't know if the cons is writable or not. Fsetcar() contains a call to CHECK_IMPURE, while XCAR does not. This means that functions like Fdelq() can well cause a coredump (instead of a mere error) when called on a pure cons, at least on architectures where purespace is read-only. If we plan to deal with this, I think we need an lvalue version of XCAR and XCDR, e.g.: #define XSETCAR(conscell, newcar) do { CHECK_IMPURE (conscell); XCAR (conscell) = newcar; } while (0) /* And ditto for XSETCDR. */ This macro would allow us to deal correctly with read-only conses, while still avoiding the price of a function call to Fsetcar(). (In fact, on *real* Unixes, we could simply use sigaction() to trap SIGSEGV and have the handler check whether the SEGV was caused by modifying a pure cons, and signal an error in that case. Of course, this would not work on Linux.)

3 participants
2 comments

Re: XEmacs featured in Unix User magazine 25 years, 3 months

Hrvoje Niksic

Kazuyuki IENAGA <ienaga(a)jsys.co.jp> writes: > After roughly I red the story, I had one incompatibility with it. > It says on the cover page that ``XEnacs is, _as its name mentions_, > using X functionality plentifully...''. I don't think they can be blamed for it. Whatever `X' was supposed to stand for, X11 is the first and obvious association.

2 participants
1 comments

XEmacs featured in Unix User magazine 25 years, 3 months

Martin Buchholz

In perhaps the most prominent press coverage in any publication, XEmacs is the cover story in the Japanese magazine Unix User. I bought a hardcopy of this magazine for its sentimental value, although I cannot read it (yet). Check out: http://www.softbank.co.jp/~unixuser/index.html I think the message is something like ``Going with XEmacs'' Martin

3 participants
2 comments

Re: [FIX] 21.2 w3 display image -> SIGSEGV 25 years, 3 months

Norbert Koch

Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: > Norbert Koch <n.koch(a)delta-ii.de> writes: > > > I've tried to display images on <URL:http://www.debis.de/>. Always > > within the third frame (out of four), no matter what order, I get a > > SIGSEGV. > > I've finally tracked this down (after starting to make a local mirror > of the debis site with wget :-) Jeez, what an effort :-) [...] > Actually the above code looks pretty obviously broken. (no offence, > Jareth) :-) But then again this particular gif is probably broken as > well (I don't know the specs at all). Somebody who *does* know the > specs should check this patch. [...] Thanks for the patch Gunnar, I'll apply it next time I've got some spare minutes, maybe this weekend :-) Cheers, norbert.

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 January 1999