July 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: Discrepancy between Installation and configure 26 years, 1 month

SL Baur

wmperry <wmperry(a)aventail.com> writes in xemacs-beta(a)xemacs.org: wmperry> How about `DSO' like apache uses? Dynamic-Shared-Object > karlheg(a)cathcart.sysc.pdx.edu (Karl M. Hegbloom) writes: >> That's not bad. The "shared object" extension of `.so' could become >> `.dso' then, on Linux. Perhaps on 95/98/NT and other Unices as well? > Yup - the extension doesn't matter at all. To make life easy for us here > when we auto-discover modules, we use different extensions on all of our > platforms for various types of shared objects. .cfm .mod .uga all work > just dandy (even on windows :) wmperry> Or go whole hog and spell the #%!＠#%! thing out. :) >> >> I suppose for the `configure' switch, that would be appropriate. It >> makes it readable, like long lisp variable names. > --with-really-spiffy-but-dangerous-dyanmic-loaading-code=yes On closer inspection, this option is also referred to in two other places by the name `modules' via --site-modules and --moduledir. Of the suggestions I like the DSO one the best, but since we're already using the modules name in installed directories [whacks the top of the computer monitor with a beer[1]] let's do both ... :-) 1999-07-26 SL Baur <steve＠miho> * configure.in: Rename --with-shlib to --with-modules for consistency with the other two options that use that name. * configure.usage (--with-modules): Document it. Index: configure.in =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs/configure.in,v retrieving revision 1.111.2.36 diff -u -r1.111.2.36 configure.in --- configure.in 1999/07/22 07:32:39 1.111.2.36 +++ configure.in 1999/07/26 04:55:45 ＠＠ -520,7 +520,7 ＠＠ use_assertions | \ memory_usage_stats | \ with_clash_detection | \ - with_shlib | \ + with_modules | \ no_doc_file ) dnl Make sure the value given was either "yes" or "no". case "$val" in ＠＠ -3616,12 +3616,12 ＠＠ dnl autodetect dll support AC_CHECK_HEADERS(dlfcn.h, [have_dlfcn=yes AC_DEFINE(HAVE_DLFCN_H)]) -test -z "$with_shlib" && test ! -z "$have_dlfcn" && { AC_CHECK_LIB(dl, dlopen, [ AC_DEFINE(HAVE_DLOPEN) DLL_LIB=dl; with_shlib=yes]) } -test -z "$with_shlib" && test ! -z "$have_dlfcn" && { AC_CHECK_LIB(c, _dlopen, [ AC_DEFINE(HAVE_DLOPEN) DLL_LIB=; with_shlib=yes]) } -test -z "$with_shlib" && test ! -z "$have_dlfcn" && { AC_CHECK_LIB(c, dlopen, [ AC_DEFINE(HAVE_DLOPEN) DLL_LIB=; with_shlib=yes]) } -test -z "$with_shlib" && { AC_CHECK_LIB(dld, shl_load, [ AC_DEFINE(HAVE_SHL_LOAD) DLL_LIB=dld; with_shlib=yes]) } -test -z "$with_shlib" && { AC_CHECK_LIB(dld, dld_init, [ AC_DEFINE(HAVE_DLD_INIT) DLL_LIB=dld; with_shlib=yes]) } -if test "$with_shlib" = "yes"; then +test -z "$with_modules" && test ! -z "$have_dlfcn" && { AC_CHECK_LIB(dl, dlopen, [ AC_DEFINE(HAVE_DLOPEN) DLL_LIB=dl; with_modules=yes]) } +test -z "$with_modules" && test ! -z "$have_dlfcn" && { AC_CHECK_LIB(c, _dlopen, [ AC_DEFINE(HAVE_DLOPEN) DLL_LIB=; with_modules=yes]) } +test -z "$with_modules" && test ! -z "$have_dlfcn" && { AC_CHECK_LIB(c, dlopen, [ AC_DEFINE(HAVE_DLOPEN) DLL_LIB=; with_modules=yes]) } +test -z "$with_modules" && { AC_CHECK_LIB(dld, shl_load, [ AC_DEFINE(HAVE_SHL_LOAD) DLL_LIB=dld; with_modules=yes]) } +test -z "$with_modules" && { AC_CHECK_LIB(dld, dld_init, [ AC_DEFINE(HAVE_DLD_INIT) DLL_LIB=dld; with_modules=yes]) } +if test "$with_modules" = "yes"; then XE_SHLIB_STUFF if test "$can_build_shared" = "yes"; then AC_DEFINE(HAVE_SHLIB) ＠＠ -3632,7 +3632,7 ＠＠ AC_CHECK_FUNCS(dlerror _dlerror) else AC_MSG_WARN(disabling shared library support) - with_shlib=no + with_modules=no fi fi ＠＠ -4147,7 +4147,7 ＠＠ athena ) echo " Using Athena dialog boxes." ;; athena3d ) echo " Using Athena-3d dialog boxes." ;; esac -test "$with_shlib" = "yes" && echo " Compiling in DLL support." +test "$with_modules" = "yes" && echo " Compiling in DSO module support." test "$with_clash_detection" = yes && \ echo " Clash detection will use \"$lockdir\" for locking files." echo " movemail will use \"$mail_locking\" for locking mail spool files." ＠＠ -4190,7 +4190,7 ＠＠ ac_output_files="${ac_output_files+$ac_output_files }$file" done ac_output_files="$ac_output_files src/paths.h lib-src/config.values" -if test "$with_shlib" = "yes"; then +if test "$with_modules" = "yes"; then ac_output_files="$ac_output_files lib-src/ellcc.h" fi Index: configure.usage =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs/configure.usage,v retrieving revision 1.17.2.9 diff -u -r1.17.2.9 configure.usage --- configure.usage 1999/07/05 08:20:29 1.17.2.9 +++ configure.usage 1999/07/26 04:55:21 ＠＠ -136,6 +136,8 ＠＠ --mail-locking=TYPE (*) Specify the locking to be used by movemail to prevent concurrent updates of mail spool files. Valid types are `lockf', `flock', and `file'. +--with-modules Compile in experimental support for dynamically + loaded libraries (Dynamic Shared Objects). --with-site-lisp=yes Allow for a site-lisp directory in the XEmacs hierarchy searched before the installation packages. --with-site-modules=no Disable site-modules directory in the XEmacs hierarchy, Footnotes: [1] For those not acquainted with American television, there was a series of beer commercials in the early 90's that featured people at a bar arguing over which sport to watch on T.V. The dispute was resolved by someone hitting the top of the T.V. with a bottle of beer whereby the sports were combined. My favorite was Football/Golf -- sink your putt and race to the tee before you get tackled ... :-)

1 participants
0 comments

Re: patch very preliminary support for /dev/pts/ 26 years, 1 month

SL Baur

Neal Becker <nbecker(a)fred.net> writes in xemacs-beta(a)xemacs.org: > Here is a very quick patch that seems to work to support the new > /dev/pts on linux glibc2.1. > Obviously, better integration is needed. I suppose HAVE_GETPT needs > to be added to autoconf. ... > +#ifdef LINUX > +#define HAVE_GETPT > +#endif Um, it's going to have be a tad smarter than this to get put in. Doesn't pts support require glibc 2.1 *and* a 2.2 kernel *and* being configured in?

2 participants
1 comments

patch very preliminary support for /dev/pts/ 26 years, 1 month

Neal D. Becker

Here is a very quick patch that seems to work to support the new /dev/pts on linux glibc2.1. Obviously, better integration is needed. I suppose HAVE_GETPT needs to be added to autoconf. retrieving revision 1.11.2.5 diff -u -r1.11.2.5 process-unix.c --- process-unix.c 1998/12/29 10:54:27 1.11.2.5 +++ process-unix.c 1999/07/26 02:39:07 ＠＠ -216,6 +216,23 ＠＠ int fd; int c; +#ifdef LINUX +#define HAVE_GETPT +#endif + +#ifdef HAVE_GETPT +#define PTY_ITERATION +#define PTY_OPEN \ + { \ + int e; \ + fd = getpt(); \ + if (fd < 0) return fd; \ + if ((e = grantpt (fd)) < 0) return e; \ + if ((e = unlockpt (fd)) < 0) return e; \ + } +#define PTY_TTY_NAME_SPRINTF ptsname_r (fd, pty_name, sizeof (pty_name)); +#endif + #ifdef PTY_ITERATION PTY_ITERATION #else ＠＠ -265,7 +282,7 ＠＠ if (access (pty_name, 6) != 0) { close (fd); -#if !defined(IRIS) && !defined(__sgi) +#if !defined(IRIS) && !defined(__sgi) && !defined (HAVE_GETPT) continue; #else return -1;

1 participants
0 comments

Re: GNU Emacs' rect.el 26 years, 1 month

Didier Verna

Kyle Jones <kyle_jones(a)wonderworks.com> writes: > Sounds like total crock to me. Rectangles are hopelessly broken > across variable glyph sizes. I don't see how blowing away > wide characters is a fix, unless you consider a lathe to be an > acceptable solution for putting a square peg into a round hole. In genuine english words, this pretty much summarizes my feeling too :-) -- / / _ _ Didier Verna http://www.inf.enst.fr/~verna/ - / / - / / /_/ / ENST, INFRES C201.1 mailto:verna＠inf.enst.fr /_/ / /_/ / /__ / 46 rue Barrault Tel. +33 (1) 45 81 73 46 75013 Paris, France Fax. +33 (1) 45 81 31 19

1 participants
0 comments

Re: GNU Emacs' rect.el 26 years, 1 month

SL Baur

Hrvoje Niksic <hniksic(a)srce.hr> writes in xemacs-beta(a)xemacs.org: > I've never really grokked the concept of "multicolumn characters". We > already support proportional fonts. Shouldn't Japanese be treated the > same as anything else? It's not the same thing. Under normal settings, Japanese characters display twice as wide as Latin characters. This means that using a computation like (- (point) (point-at-bol)) to determine indentation is broken.

1 participants
0 comments

Re: pgnus commandeers emacs to check mail 26 years, 1 month

Simon Josefsson

Kyle Jones <kyle_jones(a)wonderworks.com> writes: [threads in emacs] > :) There is still the question for which I still have not heard > a good answer--- Why do this? As this discussion came from the ding list, I take the natural example. Emacs lock up for a long time when Gnus enters a huge folder. If there were support at the language level to start a truly asynchronous thread Gnus could build the summary buffer in background and present it to the user when it's finished. Using callbacks in the "asynchronous" process filters a'la M-x man as has been suggested won't buy us nothing -- the elisp interpreter is still synchronous. I also agree with the opinion that adding thread support at the language level is easier than making applications take advantage of threads. But that's a poor excuse for not implementing threads at the language level. > I agree that we could save some memory with threading vs. multiple > processes, but memory is much cheaper than man hours. We'd have to > overhaul the Lisp system and all the applications to make them > mt-safe, and all without funding? I can't imagine it happening. Neither do I. The only chance I see this happening is a project re-targeting Emacs under Guile or some other language that already have thread support. I think thread support in itself has to little momentum to gather enough interest.

1 participants
0 comments

vc-revert-buffer1 sacrifices fontification preserving minor modes 26 years, 1 month

Adrian Aichner

Both for RCS and CVS I use vc.el to perform version-control from within source buffers directly. Whenever I RCS-checkin or CVS-commit a file using C-x v v (vc-next-action) I loose fontification of that buffer. This also happens for C-x v u (vc-revert-buffer) I have (emacs-version) "XEmacs 21.1 (patch 4) \"Arches\" [Lucid] (i386-pc-win32) of Thu Jul 08 1999 on ZJ75T" and vc 1.18 Version Control for Free systems. I tracked this down to a specific argument `vc-revert-buffer1' supplies in its call to `revert-buffer'. It is true that it preserves minor modes (like footnote-mode) that way, but fontification is lost, although the modeline still says font-lock-mode is on. Only newly entered text is fontified. I tried calling (revert-buffer arg no-confirm nil) instead and this restores the fontification but looses the minor mode. To me this still seems the better alternative. --------------------------------- What is the correct fix for this? --------------------------------- Thanks, Adrian (defun vc-revert-buffer1 (&optional arg no-confirm) ;; Revert buffer, try to keep point and mark where user expects them in spite ;; of changes because of expanded version-control key words. ;; This is quite important since otherwise typeahead won't work as expected. (interactive "P") (widen) (let ((context (vc-buffer-context))) ;; t means don't call normal-mode; that's to preserve various minor modes. (revert-buffer arg no-confirm t) (vc-restore-buffer-context context))) -- NON DISPERDERE IL VETRO NELL'AMBIENTE

1 participants
0 comments

Re: hm--html-menu fix for "fixed" psgml. 26 years, 1 month

Jan Vroonhof

jerry(a)cs.ucsb.edu (Jerry James) writes: > a couple of months ago. I would submit a patch, only I don't know which > file to put the hook in. Good question. psgml-html defines the html-mode command. So that would still be the best place. Jan

1 participants
0 comments

Re: GNU Emacs' rect.el 26 years, 1 month

Hrvoje Niksic

Didier Verna <verna(a)inf.enst.fr> writes: > I'm merging my changes into GNU Emacs' rect.el, and I found a > function they use that we don't have: it's called > `move-to-column-force' and it replaces multicolumn characters I've never really grokked the concept of "multicolumn characters". We already support proportional fonts. Shouldn't Japanese be treated the same as anything else?

1 participants
0 comments

Re: hm--html-menu fix for "fixed" psgml. 26 years, 1 month

Jan Vroonhof

Adrian Aichner <aichner(a)ecf.teradyne.com> writes: > Oops, I wasn't aware that hm--html-minor-mode hooks into PSGML. > > Thanks for the fix. > > broken-out-of-ignorance-ly yours, bad boy! Go add hm--html-minor-mode to your html-mode hook now :-) Jan

2 participants
1 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 July 1999