June 1998 - XEmacs-Beta - XEmacs Mailing List Archives

XEmacs 21.0-beta43 "Spanish" is released 26 years, 5 months

SL Baur

The next edition of XEmacs 21.0 has been uploaded to the ftp site. ftp://ftp.xemacs.org/pub/xemacs/beta/xemacs-21.0/ CVS sources are at http://cvs.xemacs.org/ Installable Lisp packages are in: ftp://ftp.xemacs.org/pub/xemacs/beta/xemacs-21.0/packages/binary-packages/ The CVS tag for this beta is `r21-0b43'. To read about the Spanish Goat please see: http://www.ansi.okstate.edu/breeds/goats/spanish/index.htm * Comments We are officially in feature freeze. Code freeze is rapidly approaching. Beta-42 worked pretty well for me. This contains a lot of fixes and hopefully works well for you too. As we wind down towards release, please prefer testing the tarballs off the FTP site instead of CVS if you can. Note to developers: the incoming patch queue still contains a number of patches including nearly all Lisp package patches. These will be applied and released on Saturday with pre1. Note to Adrian: Your changes to texinfo.texi bombed. Since the texinfo manuals are out of date (they're based on GNU texinfo-3.11) I went ahead and synched them with the latest versions from texinfo-3.12. * Updated packages None. VM has been updated to 6.50, but new binaries won't be uploaded until tomorrow's release. I'm aware it contains MS Windows fixes so I'll try to make it available as early as I can. * Recent Changes Sound files are now located in an external package. Please load the sounds-au (Sun sounds), or sounds-wav (Microsoft sounds) as appropriate for your platform. Sounds and xbms are now supported on Microsoft Windows. *IMPORTANT* xemacs-base has been updated since beta40. Please download and install the new version before starting. The function `kbd' in edmacro.el has been moved into the XEmacs core. *IMPORTANT* pcl-cvs has a critical bug fix due to an XEmacs interface change starting in beta40. If you are using pcl-cvs please upgrade to the new version. We've decided to push forward on Drag'n'Drop and try to get _something_ into the 21.0 release. Hopefully it will be something better than just wildly experimental. Please see the tests in tests/Dnd (droptest.el and dragtest.el) and give them a spin and see well they work or don't work. There's a lot of new MS Windows support in this, including toolbar, processes, and Xpm. I'm happy with the package path/startup path stuff for the first time in a number of betas. Hopefully, we're past the worst of the changes. All path stuffs should work again. The configure option --gung-ho now controls whether minimal tagbits and indexed lrecords are used. Please test the -elc tarball with this beta. I've generated it from a tty-only build for testing. While this should work (and the build now currently assumes it does work), and it seems to work for me, it needs as much testing as possible prior to release. Canna detection should work better. LDAP support should build and run. Multicast support should build and run. The default target of `make' is to build everything. This is almost what `make dist' used to do except that finder-inf.el (the finder database) gets properly bytecompiled. DLL support is now autodetected. The configure option `--with-dlmalloc' has been added so that gmalloc.c can still be used on systems with the Doug Lea new GNU Malloc. The Installation file format has changed slightly -- the version number is now printed in the summary, old copies are not retained, so if you wish old copies you will have to make the copy by hand. The Installation summary is available via the command M-x describe-installation. `XEmacs' can be used for X resources. This feature needs further testing but appears to work. dlmalloc now works with mmap'ping enabled for Linux Libc5 systems. I'm very interested in reports on how XEmacs' virtual memory footprint changes with time with dlmalloc. My own experiences with mmap disabled (only for glibc-2.0 and Mule) show a net win over 20.4. I'm hoping mmap improves things further. The source-packages subdirectory on the FTP site has been abolished in favor of distributing Lisp sources via CVS. The minimum required packages are the following: [Only xemacs-base updated since the last beta, the package version number is relatively unimportant] xemacs-base-1.20-pkg.tar This package contains core functionality not dumped into XEmacs. It includes things like the debugger, comint, ring, etc. mail-lib-1.15-pkg.tar This package contains core functionality common to mail and newsreading. *NOTE:* As of beta39, you must use mule-base-1.16 or later. Note also, as of beta34, overlay.el has been moved to the fsf-compat package. As of beta40, you must have xemacs-base-1.20 or later. If you are building with Mule, you will need to have the mule-base package installed *prior* to dumping XEmacs or you will lose. If you see messages about mule-charset not found, or get a bomb immediately after seeing a message about `loading ... dialog', then you need to stop and install the mule-base package. Other various packages need to be installed prior to dumping XEmacs. They are marked with `dump t' in the package description, see package-get-base.el and package-get.el for an interface. ** ImageMagick has gone away courtesy of Jareth Hein. ** The generalization of the XEmacs internals has continued. Quoting Kyle Jones: If you're feeling maximally gung, add -DUSE_MINIMAL_TAGBITS and -DUSE_INDEXED_LRECORD_IMPLEMENTATION to your compiler flags when you build 20.5. If all is well XEmacs should build and run normally. Update: These options are now enabled by specifying `--gung-ho'. -DUSE_MINIMAL_TAGBITS will be automatically set if using dlmalloc. NOTE: It should *never* be required to use these options for proper operation of XEmacs (with the single exception of -DUSE_MINIMAL_TAGBITS for dlmalloc). A failure _without_ these options set is a bug that needs to be fixed. ** Chinese and Korean Wnn server support has been added to LEIM. Someone familiar to cWnn and kWnn will need to test this out. I haven't heard any reports positive or negative yet. * Futures ** 21.0-pre1 is scheduled for Saturday, June 13. ** The highest priorities are testing and bug fixing. For those not acquainted with recent XEmacsen (things have changed since 20.3/21.0-b11), there are 4 files in the distribution. Each file has been PGP signed. *** xemacs-21.0-b42-21.0-b43.patch.gz This is a patch from the previous release. Please see the file etc/BETA for patching instructions. As of beta13 we're through the worst of the reorganizations and patching should be pretty smooth from here on out. *** xemacs-21.0-b43.tar.gz This is the core distribution without bytecompiled lisp and without processed texinfo manuals. If you are on a fast machine with a possibly slow connection to the FTP site, this is all you need for an XEmacs non-mule. Untar it, run configure, then run `make' and `make install'. *** xemacs-21.0-b43-info.tar.gz This the processed texinfo manuals. If you have makeinfo installed and your machine has any kind of decent speed, it is quicker to run `make info' from the toplevel and rebuild it yourself than it is to download and untar it. You will need a fairly recent version of makeinfo in order to be able process all the info documentation. At least `makeinfo (GNU texinfo 3.11) 1.68' is recommended. Note that the latest version of GNU texinfo is now 3.12. *** xemacs-21.0-b43-elc.tar.gz This is the bytecompiled lisp. It was freshly bytecompiled from scratch prior to distribution. to 21.0 beta43 "Spanish" -- texinfo documentation synched from GNU texinfo-3.12 -- Manual cleanup from Adrian Aichner -- Miscellaneous fixes from Rick Rankin -- D'n'D fixes from Oliver Graf -- startup changes from Stephen Turnbull -- Miscellaneous fixes from Kirill Katsnelson -- PNG Graphics fix from Gunnar Evermann -- Miscellaneous MS windows patches from Andy Piper -- Graphics update from Hrvoje Niksic -- Miscellaneous patches from Hrvoje Niksic -- NEWS updates from Hrvoje Niksic -- Grayscale JPEG fix from Jareth Hein -- Miscellaneous patches from Didier Verna -- Miscellaneous patches from Altrasoft

1 participants
0 comments

Re: Did someone here get news ... 26 years, 5 months

Bob Weiner

>>>>> "sb" == SL Baur <steve(a)vmailer.xemacs.org> writes: sb> Bob Weiner looked into it about a year ago and didn't find a publisher. Yes, Altrasoft did pursue this and found the upfront costs to be about $15-20k (just on the publishing end). We couldn't find any source of significant copy orders so there was no point in pursuing it. There were about four people as I recall who wanted single copies. We have a lot of other interesting stuff going on though. If you really want to help fund XEmacs development and you know of anyone in a significant company whose management has dictated that freed software like XEmacs must be certified to be Y2000-compliant in order to continue to be used at the site, please let them know that Altrasoft can provide such certification and can be reached at 408-243-3300 or www.altrasoft.com. (Please, no discussion of whether or not XEmacs is Y2K compliant; this is a separate issue from the corporate need to have a company take responsibility and provide them with a certification). Regards, Bob

1 participants
0 comments

x-init.el and `x-initialize-keyboard' 26 years, 5 months

David Bakhash

okay, so I mentioned this earlier, but just wanted to make sure... It seems as though the function `x-initialize-keyboard' uses `load' instead of require: (defun x-initialize-keyboard () "Perform X-Server-specific initializations. Don't call this." ;; <some more comments> (let ((vendor (x-server-vendor))) (cond ((or (string-match "Sun Microsystems" vendor) ;; MIT losingly fails to tell us what hardware the X server ;; is managing, so assume all MIT displays are Suns... HA HA! (string-equal "MIT X Consortium" vendor) (string-equal "X Consortium" vendor)) ;; Ok, we think this could be a Sun keyboard. Load the Sun code. (load "x-win-sun")) ((string-match "XFree86" vendor) ;; Those XFree86 people do some weird keysym stuff, too. (load "x-win-xfree86"))))) Note that these things are loaded. I guess this is as safeguard to deal with the situation when someone opens up an X frame on different kinds of computers? I don't know. All I do know is that it's not as simple as putting (require 'x-win-sun) instead of the load line, because those features arn't provided. So they get re-loaded every time (a waste of time, especially on slow computers). Don't know what the best fix is. Personally, I've added a defvar and fixed it on my own, but it's not so nice looking: (defvar keyboard-initialized-p nil "Non-nil iff the `x-initialize-keyboard' code has been loaded.") (defadvice x-initialize-keyboard (around no-load activate) (unless keyboard-initialized-p (setq keyboard-initialized-p t) ad-do-it)) dave

2 participants
2 comments

Re: Did someone here get news ... 26 years, 5 months

SL Baur

Didier Verna <verna(a)inf.enst.fr> writes: > ... from the guy that contacted us once, with the plan of writing a > book on XEmacs ? (I don't know why this suddenly came to my mind again). I'm not sure what you're referring to. Kazuyuki IENAGA was approached by a Japanese publisher some time ago, but I believe that fell through. Bob Weiner looked into it about a year ago and didn't find a publisher. Or do you mean the person who expressed interest in writing documentation? That's dead as far as I know.

1 participants
0 comments

Re: Xkb-Extensions and XEmacs 26 years, 5 months

SL Baur

Holger Schauer <schauer(a)coling.uni-freiburg.de> writes: > In german linux groups people quite often report problems with using > Xkb-Extensions of Xfree86 and XEmacs[1]. For me, this is not a real > problem, since I don't use the extensions. But if there is a solution, > I would like to know about it (in order to forward it to the > FAQ-maintainer of the german linux groups. And hey, why isn't it in > the XEmacs FAQ ?) If something isn't in the FAQ, it's either not a FAQ we've seen, or perhaps we don't know the answer either. Doing the XEmacs FAQ is a difficult job[1] and sometimes things get missed. Other than Stefanie Teufel (if I recall correctly), I don't know of anyone who reads the German Linux groups. What exactly is the problem? XKEYBOARD shows up in my server's list of extensions and I haven't had any problem remapping keys. Footnotes: [1] Applause for Christian Nybø, the current XEmacs FAQ maintainer.

1 participants
0 comments

Re: Where is ph.texi and message.texi? 26 years, 5 months

SL Baur

Raymond Toy <toy(a)rtp.ericsson.se> writes: > SL Baur <steve(a)vmailer.xemacs.org> writes: >> tried and failed to get Lars to accept them. He backed them out >> immediately when it was learned that they cause Emacs 19.34 to lose >> when creating info files from the editor. > I take it that we don't have that problem or that we use makeinfo > instead. Right? We use makeinfo and bypass the issue of inter-XEmacs compatibility. I don't think there should be a problem putting the direntry's back in after we get the Gnus 5.6 we'll ship with 21.0.

1 participants
0 comments

Re: 'italic 'bold etc faces not inheriting from default? 26 years, 5 months

Jan Vroonhof

Andy Piper <andyp(a)parallax.co.uk> writes: > This gets every face apart from > > bold > italic > bold-italic > > Which appear to be hard-wired to Courier. Can anyone tell me why this is? Because they are already are derived from default before your .emacs gets loaded. This is probably because they appear in the splash screen. Maybe the use of these fonts in the splash screen should be ditched untill the font model gets improved. Jan

1 participants
0 comments

Re: XEmacs 21.0 and the options menu 26 years, 5 months

Jan Vroonhof

Did > Yup. So people upgrading will believe they lost their options > setting. Not much better :-( If we don't find a better solution, this should > at least be documented in the NEWS file. Of course it should be documented in the NEWS file. In fact it would be nice to have some "help with migrating" buffer pop up the first time after a major upgrade. I just think that thse kinds of things are better to do with the user being consious about it in stead of trying to hide everything. > Wrong. The options menu saving button calls > customize-save-customized. If you load the options from the old file, the > options will actually be changed outside Custom, and thus the save button will > not save them. Then we need a "Customize-save-all-non-defaults" to be called on the "options" group. Then we can put something like that in a migratete.el Jan

1 participants
0 comments

Re: info patch 3 26 years, 5 months

Raymond Toy

Raymond Toy <toy(a)rtp.ericsson.se> writes: > > Here is the last set of direntry patches for the info files. This > should be applied after the other info patches. > Oops. I sent this to the wrong list. My apologies. Ray

1 participants
0 comments

Re: Bug when selecting TTY frames 26 years, 5 months

Hrvoje Niksic

Kyle Jones <kyle_jones(a)wonderworks.com> writes: > focus-frame is still crocked with regard to the frame selection > hooks which means it may have to be commented out for the release. It passed my initial tests. It should be no more broken than `select-frame', because it (among other things) actually *calls* select-frame. Do you have a specific repeat-by procedure? -- Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia --------------------------------+-------------------------------- "Beware of bugs in the above code; I have only proved it correct, not tried it." -- Donald Knuth

2 participants
1 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 June 1998