December 1999 - XEmacs-Beta - XEmacs Mailing List Archives

LaTeX font-lock problem 24 years, 6 months

nicolls＠dip.ee.uct.ac.za

Full_Name: Fred Nicolls OS: FreeBSD 3.3-STABLE Version: xemacs-21.1.7 mule: no severity: Aggravating Submission from: (NULL) (196.7.0.164) Quite often it seems to happen (at least for me) that the LaTeX mode font-lock loses sync between math and non-math. The primary cause appears to be when fairly complex inline math expressions are split across line boundaries. The following file excerpt exhibits this problem: --- \documentclass[a4paper,10pt]{article} \usepackage{a4wide} \usepackage{amsmath} \begin{document} Basilevsky~\cite{Basilevsky83}, for example, provides such a matrix, which is closely related to the DFT matrix. A second approach, suggested by Johnson~\cite{Johnson94}, is to require conjugate symmetry in the random draws of ${\bf y}$, so $y_k = (y_{2 M + 2 - k})^*$. In that case the transformed data ${\bf z}={\bf F}^\dagger {\bf y}$ will be real, as required. The procedure outlined here may be extended to the generation of 2-D random samples without any further conceptual complications. In this case the covariance of the required data will be doubly block Toeplitz, which can be embedded into a doubly circulant matrix in much the same way as for the 1-D case. The 2-D DFT can be used to efficiently diagonalise the resulting covariance, and samples generated accordingly. \end{document} --- The problem can be reproduced by performing the following steps: 1. Create a file on disk (sg.tex in my case) containing the above code snippet. 2. Load the file using "xemacs -vanilla sg.tex". 3. Enable font-locking "M-x font-lock-mode" 4. The buffer fontifies correctly. However, if anything in the second paragraph is changed, the line switches to fontification as if in math mode. For example, adding a space somewhere in the centre of the third line of the second paragraph switches the entire line to math fontification. The problem also occurs when using Peter Galbraith's font-latex.el package. This seems to suggest a more general font-locking bug. The problem does not occur with GNU emacs. --=-=-=--

1 participants
0 comments

gcc 2.95.2 brokenness 24 years, 6 months

Yoshiki Hayashi

It seems gcc 2.95.2 does wrong optimization. (emacs-version) "XEmacs 21.2 (beta22) \"Mercedes\" [Lucid] (i686-pc-linux, Mule) of Sat Dec 4 1999 on buck" gcc -v Reading specs from /usr/lib/gcc-lib/i386-linux/2.95.2/specs gcc version 2.95.2 19991109 (Debian GNU/Linux) When I evaluate (decode-buffer 1), XEmacs 21.2 crashes. When I pass arguments other than integers or floats, it doesn't crash. XEmacs 21.1 is fine. Lisp backtrace follows: decode-buffer(1) eval((decode-buffer 1)) # bind (expr) eval-interactive((decode-buffer 1)) # bind (standard-output opoint ignore-quotes eval-last-sexp-arg-internal) eval-last-sexp(nil) # bind (command-debug-status) call-interactively(eval-last-sexp) # (condition-case ... . error) # (catch top-level ...) First part of stacktrace is: #0 0x403e25d1 in kill () from /lib/libc.so.6 #1 0x808e074 in fatal_error_signal (sig=11) at emacs.c:276 #2 0x403e2508 in sigaction () from /lib/libc.so.6 #3 0x806261b in Fdecode_buffer (buffer=3) at buffer.c:372 #4 0x8092c8d in Feval (form=139875680) at eval.c:2975 #5 0x80936a7 in Ffuncall (nargs=2, args=0xbffff388) at eval.c:3178 Crash happens in decode_buffer at buffer.c L357 if (allow_string && STRINGP (buffer)) return XBUFFER (get_buffer (buffer, 1)); It doesn't matter allow_string is 0 or 1. It always crashes when checking the if statement. So I bet this is a compiler bug. Can anyone reproduce this? If someone does, I'll report this to gcc development team. The only way for me to workaround this problem is to add -no-gcse to cflags. ./configure --prefix=/opt/local '--cflags=-g -O3 -Wall -Wno-switch -fno-gcse' seems OK. uname -a: Linux buck 2.2.13 #2 Thu Nov 25 20:06:18 JST 1999 i686 unknown ./configure '--prefix=/opt/local' '--with-mule' XEmacs 21.2-b22 "Mercedes" configured for `i686-pc-linux'. Where should the build process find the source code? /src/xemacs What installation prefix should install use? /opt/local What operating system and machine description files should XEmacs use? `s/linux.h' and `m/intel386.h' What compiler should XEmacs be built with? gcc -g -O3 -Wall -Wno-switch Should XEmacs use the GNU version of malloc? yes (Using Doug Lea's new malloc from the GNU C Library.) Should XEmacs use the relocating allocator for buffers? yes What window system should XEmacs use? x11 Where do we find X Windows header files? /usr/X11R6/include Where do we find X Windows libraries? /usr/X11R6/lib Compiling in support for XAUTH. Compiling in support for XPM images. Compiling in support for PNG image handling. Compiling in support for (builtin) GIF image handling. Compiling in support for JPEG image handling. Compiling in support for TIFF image handling. Compiling in support for X-Face message headers. Compiling in native sound support. Compiling in support for Enlightened Sound Daemon. Compiling in support for Berkeley DB. Compiling in support for ncurses. Compiling in Mule (multi-lingual) support. Compiling in XIM (X11R5+ I18N input method) support. Using raw Xlib to provide XIM support. Compiling in support for proper WM_COMMAND handling. Using Lucid menubars. Using Lucid scrollbars. Using Motif native widgets. Using Motif dialog boxes. Compiling in DSO module support. movemail will use "dot-locking" for locking mail spool files. Compiling in extra code for debugging. WARNING: --------------------------------------------------------- WARNING: Compiling in support for runtime error checking. WARNING: XEmacs will run noticeably more slowly as a result. WARNING: Error checking is on by default for XEmacs beta releases. WARNING: --------------------------------------------------------- -- Yoshiki Hayashi

2 participants
2 comments

Re: Gnus broken with current CVS XEmacs 21.1.9 version 24 years, 6 months

Andreas Jaeger

>>>>> Yoshiki Hayashi writes: Yoshiki> Andreas Jaeger <aj(a)suse.de> writes: >> This change in the 21.1.9 CVS version breaks Gnus: >> 1999-12-03 Yoshiki Hayashi <t90553(a)mail.ecc.u-tokyo.ac.jp> >> >> * lisp/mule-misc.el (coding-system-get): Imported from 21.2. >> (coding-system-put): Ditto. >> (coding-category): Imported from 21.2 and modified for 21.2. Yoshiki> Then, Gnus wasn't running fine on XEmacs 21.2 for some Yoshiki> months? According to the ChangeLog XEmacs 21.2 has Yoshiki> coding-system-get after 1999-06-28. No, it should have worked for 21.2 since lisp/mule/mule-charset-el has: (defalias 'get-charset-property 'get) Yoshiki> [...] >> Two problems with the code: >> >> The first one is XEmacs: >> XEmacs now has coding-system-get - but it doesn't have >> get-charset-property. If we have coding-system-get should we also >> have get-charset-property? Yoshiki> Yes, it would be nice if we have them. Yoshiki> But then again, this is MULE API problem. Some codes are Yoshiki> synched with FSF and some codes are not. We really really Yoshiki> really need to define API and stick with them. Until that Yoshiki> happens, application needs to have workaround. This case, Yoshiki> Gnus. Should we move those to 21.2 also? 1999-05-26 SL Baur <steve(a)gneiss.etl.go.jp> * mule/mule-charset.el (charset-after): New function. (charset-direction): Synch with Mule, update docstring. (get-charset-property): New function. (put-charset-property): New function. (charset-plist): New function. >> The other one is Gnus: Lars, can you change this to check properly >> XEmacs (sorry, I don't have a patch)? Yoshiki> According to the info manual, Gnus supports XEmacs 20.4 and Yoshiki> up. I think (featurep 'xemacs) would be sufficient. Andreas -- Andreas Jaeger SuSE Labs aj(a)suse.de private aj(a)arthur.rhein-neckar.de

2 participants
1 comments

possible bug in efs 24 years, 6 months

Jeff Miller

in efs.el, function efs-abort-recursive-edit-and-then i get the following error while byte-compiling.... While compiling efs-abort-recursive-edit-and-then in file /cvs/xemacs/packages/xemacs-packages/libs/efs/efs.el: ** set called with 1 arg, but requires 2 i think it's this line: (setq efs-abort-recursive-edit-data) i guess it defaults to nil in the above statement? jeff

1 participants
0 comments

Problems with portable dumping 24 years, 6 months

SL Baur

I don't know how this should be treated. -1. I can now run a --pdump XEmacs and load my .emacs file. Cool! 0. I would like to see this made a default as soon as possible. 1. `make clean' does not remove the xemacs.dmp file. 2. Rebuilding temacs does not force the xemacs.dmp file to be regenerated. 2.5. src/Makefile assumes `xemacs' will be built and this causes building to stop after dumping. 3. When an out-of-date xemacs.dmp file is loaded into a freshly built temacs, I get a coredump during load. If item 3 is reasonable behavior, then the .dmp file and temacs files are going to need stamps of some sort so that an out-of-date or incorrect .dmp file can be detected. I wouldn't recommend relying on filesystem timestamps.

3 participants
2 comments

Configure needs to be picker about Enheavied Sound Demon 24 years, 6 months

Stephen J. Turnbull

I have ESD installed (more disk space than time to fight with dselect) but don't use it and don't want it in my XEmacs particularly. However, configure picks up the esd-config and assumes that I can build esd-enabled stuff. Since I don't have the development package installed (or something), configure panics with "CC can no longer build executables" because ld can't find -lesd. -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 _________________ _________________ _________________ _________________ What are those straight lines for? "XEmacs rules."

1 participants
0 comments

Fixing up binary files 24 years, 6 months

Martin Buchholz

I noticed that the nt/*.ico files were not checked into CVS as binary files, as they should be. Therefore, I did: (martin＠lasker) ~/x/ws/dev/nt $ cvs-rw admin -kb lisp.ico RCS file: /usr/CVSroot/XEmacs/xemacs/nt/lisp.ico,v done (martin＠lasker) ~/x/ws/dev/nt $ cvs-rw admin -kb file.ico RCS file: /usr/CVSroot/XEmacs/xemacs/nt/file.ico,v done (martin＠lasker) ~/x/ws/dev/nt $ cvs-rw admin -kb xemacs.ico RCS file: /usr/CVSroot/XEmacs/xemacs/nt/xemacs.ico,v This should be completely transparent to everyone, unless you actually modify these files. In the unlikely event that this is the case you should first make sure the CVS client doesn't get confused. Do: # Make sure you have no files modified in nt first rm -rf nt; cvs update before making changes. Martin

1 participants
0 comments

Problem with `configure --with-sound' 24 years, 6 months

Daniel Pittman

In the CVS version of about 0100 on 08 Dec 99 (GMT), there seems to be a bug with the configure script. The documentation in `configure.usage' suggests that the argument `--with-sound=native,nas,noesd' should prevent support for esd being built in. This does not work, sadly, and continues to compile in esd support. The argument `--with-sound=native,nas' also compiles in esd support. This is a right pain for me because I get SIGABRT if I use sound with an XEmacs that has esd support build in -- probably because I don't have the daemon running or anything. If I get some time I might have a glance at it myself, but I don't think I will in the next while... Daniel -- If a voice inside tells you that you are not a painter, then by all means paint! And that voice will be silenced... -- Vincent Van Gogh

1 participants
0 comments

Re: Problems with portable dumping 24 years, 6 months

SL Baur

Olivier Galibert <galibert(a)pobox.com> writes in xemacs-beta(a)xemacs.org: > What I plan to do after I'm finished whipping our DSP cards into > submission: > - compute a md5 (or sum, or whatever) of the Installation file at > configure time > - slap it into config.h and Makefile This won't necessarily work. The Installation file need not necessarily change as the temacs image changes, especially for developers. I first got the core dumps after doing a `cvs update' which even with rerunning configure won't change the contents of the Installation file. How sensitive is the .dmp file to the environment? For example, if I build and pdump an XEmacs that has staticly linked libraries, then relink a temacs with shared libraries, will that still work? > - rename the xemacs.dmp file to xemacs-[md5].dmp (or > <whatever>-[md5].<whatever>, I'm not religious about that) O.K. > - add a flag to [tx]emacs to force it to ignore .dmp (aka, forced > dump) O.K. > This way: > - we have filesystem timestamps for the normal builds > - we can have -mule, -nomule, -optimized, etc... versions without conflict Good. >> 0. I would like to see this made a default as soon as possible. > There is a bug in the filling code in mule mode I have been unable to > trace (for my level of lisp expertise, the kinsoku processing is > simply unreadable). My face customizations in .emacs are coming out wonky, but I haven't checked a non-pdumped 21.2 recently. > I know of no other bug, but I'm not yet using it > for my day-to-day work. O.K. Jareth owns the kinsoku code and you should check with him on this when he returns from vacation, hopefully some time this week.

2 participants
1 comments

Re: Release Engineering initiatives for the XEmacs project 24 years, 6 months

Vin Shelton

Martin, Please don't forget about the 21.1 line of development while you are doing this improvement work. In particular, please let me know what your plans are wrt the CVS archive changes. I'm also interested in your build scripts - at some point I think they should be checked in to the archive. Thanks, vin

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 December 1999