XEmacs -- Emacs: The Next Generation
     Searching XEmacs
Quick Links About XEmacs Getting XEmacs Customizing XEmacs Troubleshooting XEmacs Developing XEmacs


This page has not been updated in a very long time. An initial update is forthcoming soon (ca. 2001-12-03), and further updates will occur regularly thereafter.

Last text update: 2001-11-29
Last status update: 2001-07-08

Bug Status: Open

Currently open bugs in the release-21-2 branch (pre-21.4), categories in rough order of priority. Within category, oldest report is first. Some attempt is made to collect apparently similar bug reports together. Please report any classification errors to the Release Manager.

Bug titles are generally taken from report Subject headers, and include the approximate date in yyyy-mm-dd format. Better summary descriptions will be added to the original title (not replace it).

Titles are followed by status, platforms affected, version information (including CVS checkout date and time when available), developers addressing the problem, a summary description, and mailing list references. Where mailing list references are not hyperlinks, that's because I got to it before MHonArc did.

Bug status values include Reported, Replied (meaning that a developer has acknowledged the bug), Identified (meaning that the code causing the bug has been localized), Analysis (meaning that a developer has said he will fix the bug), (Patch) Pending (meaning a developer has reported a fix), and Fixed (meaning the fix has been committed and tested).


A showstopper is a critical bug that justifies postponing release until a fix is available or judged infeasible, usually one that manifests cross-platform.

Yeah, ``the management'' simply moved this whole section into Critical. Them bugs just don't show up often enough to stop the show; we'll fix them as soon as possible. That's why we call this gamma, and continue to maintain 21.1. The XEmacs developers all trust their data to 21.4, though. Therefore we offer you the choice by making the public release.


Typical critical bugs include failure to build, failure of the built executable to start, crashes, program errors potentially resulting in data loss, and security exploits.

Xlib async reply

Status: Analysis
Versions: beta45
Developers: Michael Sperber

Argl, this one's back.  I've now seen it 4 times within 1 hour, always
when starting a subprocess:

Xlib: unexpected async reply (sequence 0x3b71)!
Xlib: sequence lost (0x1043b > 0x3b71) in reply type 0x17!
Xlib: sequence lost (0x10421 > 0x3b71) in reply type 0x17!
Xlib: sequence lost (0x10000 > 0x3b71) in reply type 0x0!

xemacs: X Error of failed request:  0
  Major opcode of failed request:  113 (X_KillClient)
  Serial number of failed request:  0
  Current serial number in output stream:  15217

It was gone for a long time before beta45.

I'll try reverting some recent patches as soon as I can.	


CRASH: problem in event code (2001-03-26)

Status: Reported
Platforms: Windows
Developers: Ben Wing

[...] this creates very long entries in the Tools->Grep submenu, which lists the ten most recent grep commands. Too long, in fact -- you get errors reported when you try to invoke an entry on the Grep submenu. Recently, I tried this by accident, and got a crash!

command_event_queue was indeed empty, but then the call to event_stream_event_pending_p (1) puts an event onto the command_event_queue. Now I don't understand how the misc-user event that was already invoked and processed is getting seen again, but clearly something is amiss in this code. since this twisty windy code was reworked a lot for windows and i don't yet understand it, could some windows expert help?


crash when playing with LDAP (2001-02-24)

Status: Reported
Versions: In XEmacs 21.2 (beta45) "Thelxepeia" [Lucid] (i686-pc-linux) of Fri Feb = 23 2001 on pip configured using `configure --prefix=3D/lap/xemacs/21.2 --with-dialogs=3D= athena --with-widgets=3Dathena --with-athena=3D3d --pdump'

When playing with LDAP I got a crash. It is repeatable.

In the *scratch*-buffer eval:

(load-library "ldap")
(ldap-search-entries "torkel")
xemacs: error.c:221: ldap_parse_result: Assertion `r !=3D ((void *)0)' fa=


crash while scrolling (2001-02-25)

Status: Reported
Platforms: Windows
Developers: Ben Wing (reporter)

error_check_window(long 108) line 264 + 17 bytes
real_window_internal(long 108, window_mirror * 0x0000006e, window_mirror *
0x01f91d80) line 406 + 9 bytes
real_window(window_mirror * 0x01f91d80, int 1) line 489 + 30 bytes
mswindows_handle_scrollbar_event(HWND__ * 0x00370132, int 0, int 0) line 184 +
14 bytes
mswindows_wnd_proc(HWND__ * 0x00170112, unsigned int 277, unsigned int 0, long
3604786) line 2727 + 26 bytes


Patch Fumbles

Patches that people seem to have lost track of....

Patch for bogus cygwin paths in gnuclient (2001-03-16)

Status: Patch Pending
Platforms: Cygwin
Developers: Ben Wing

before the patch, I would see [Cygwin] looking for a file name of: //E\htdocs\file3.php which doesn't exist as a path name. However, with this patch, it translates it to /htdocs/file3.php which is correct!

Ben sez: great! do it!

xemacs-nt NDBBJJECCLCABPILAFECCEPNFHAA.jstuart-devel@neo.rr.com

Installation Problems

Parse Erorrs in setup Win32 version (2001-03-08)

Status: Reported
Platforms: Win32
Versions: Unknown

setup.ini line 1: parse error
setup.ini line 1: unrecognized line in setup.ini headers (do you have the lates setup?)

Parse Erorrs in setup Win32 version

Doesn't Work

Functionality that is expected to work (eg, has been reported to work in similar environments) doesn't.

Gutter vs VM (2000-11-20)

Status: Reported
Platforms: cygwin
Versions: XEmacs 21.2 (beta37) "Pan" [Lucid] (i686-pc-cygwin) of Sat Nov 18 2000 on PASTON

xemacs -q
M-x vm
Get stack trace below. Repeat M-x vm and frame is set up fine.

Is xemacs -unmapped broken? (2001-02-26)

Status: Reported
Platforms: Linux

I believe I've tracked this down. It seems to be a problem with xemacs
-unmapped. The following behavior is repeatable in 21.2.45 on 2 RedHat
7.0 systems without a .emacs or .gnus:

1. xemacs -unmapped -f gnuserv-start &
2. gnuclient
3. M-x gnus
4. quit gnus
5. gnuclient -nw
6. M-x gnus in the tty gnuclient.
7. Enter any group.
8. SPC in the *Summary* buffer (`gnus-summary-next-page') does not
   scroll the *Article* buffer.

If one kills the X gnuclient (delete-frame), SPC starts to work again.
Start another X gnuclient and SPC will stop working.

This problem does not exist in XEmacs 21.1.12. Without the -unmapped
option, again there is no problem.

-debug-init doesn't in xemacs v21.2 beta (2001-03-05)

Status: Replied
Developers: Vin Shelton

bingalls@panix.com (Bruce Ingalls) writes:
> I'm quite happy with the latest xemacs beta.
> Here are some issues:
> 	xemacs -debug-init
> Still doesn't work. I would think that this one would be easy to get
> working- just do a
> 	xemacs -q
> 	(setq debug-on-error 't)
> 	(load-file ".xemacs/init.el")	;or equivalent
> Unfortunately, if I have errors in my .emacs,
> 	xemacs -debug-init
> loads without a peep, whereas plain
> 	xemacs
> complains.

frame-title-format doesn't work in xemacs v21.2 beta (2001-03-05)

Status: Replied
Developers: Vin Shelton

bingalls@panix.com (Bruce Ingalls) writes:
> I'm quite happy with the latest xemacs beta.
> I set the following in my .emacs:
> 	(defconst system-name (system-name))
> 	(setq frame-title-format
> 	  '(" " system-name ":" default-directory " %12b"))
> which used to work fine. Now it is almost blank.


Wrong Computation

Module path initialization (2001-01-05 #### much earlier)

Status: Analysis
Platforms: Unix?
Versions: All?
Developers: Jerry James, Stephen Turnbull

Path-finding code doesn't do sensible things with modules under some combinations of --srcdir, run-in-place, previous installation, etc.

Problems with mswindows-shell-execute (2001-02-25)

Status: Replied
Platforms: Windows
Developers: Andy Piper

 o XEmacs 21.2  (beta41) "Polyhymnia" [Lucid] (i686-pc-cygwin, Mule) of Wed Jan 17 2001 on VZELL

My cygwin mount table:

Device              Directory           Type         Flags
D:\bin              /usr/bin            system       binmode
D:\lib              /usr/lib            system       binmode
C:                  /dev/c              system       binmode
D:                  /                   system       binmode

Please take a look at the following lisp expressions:

(mswindows-shell-execute nil "/tmp/inst_notes_ias102.html")

  This works

(mswindows-shell-execute nil "/dev/c/tmp/FAQ.html")

  This works

(mswindows-shell-execute nil "file///tmp//inst_notes_ias102.html")

  file not found: "file///tmp//inst_notes_ias102.html"                  (X)

(mswindows-shell-execute nil "file//D:/tmp//inst_notes_ias102.html")

  file not found: "file//D:/tmp//inst_notes_ias102.html"

(mswindows-shell-execute nil "file///D|/tmp//inst_notes_ias102.html")

  file not found: "file///D|/tmp//inst_notes_ias102.html"

(mswindows-shell-execute nil "http://vzell/")

   file not found: "http://vzell/"

Are the file:// and http:// type URL's supposed to work ?

I'm asking because lots of lisp code uses code like the following.
This is from jde:

(browse-url (concat "file://" (jde-convert-cygwin-path jde-help))

In my case I try to use in browse-url,
browse-url-default-windows-browser for it's browse-url-browser-function.

So I end up with URL's of type (X), but this gives the error above.

gnuclient screwage? (2001-03-05)

Status: Analysis
Platforms: All?
Developers: Stephen Turnbull

Launch 'xemacs' and load gnuserv-start from .xession, with no TMPDIR environment variable set. So gnuserv opens /tmp/gsrvdir967/gsrv as the socket to use.

Another program (exmh) then tries to launch 'gnuclient' - but exmh has set TMPDIR to /tmp/valdis. So it tries to open /tmp/valdis/gsrvdir/gsrv and promptly loses.

Font-lock bug in 21.2.45 (2001-03-06)

Status: Reported
Platforms: All

The current version of font-locking font-locks the following file incorrectly.

The first #+:cmu is fontified as if it were a comment, but it's not. This doesn't happen with 21.2.44.

To see this, I started xemacs -vanilla. Load the file. (The buffer is not fontified.) Kill the buffer. Load the file again. Now the buffer is fontified but fontified incorrectly. If you do the same with 21.2.44, #+:cmu and following is fontified correctly as lisp code, not a comment.

Perhaps it's getting confused by the |# in column 0?


MS-Windows XEmacs doesn't see fonts with a number in the font name (2001-03-07)

Status: Reported
Platforms: Windows

I have a couple of fonts with numbers in the font face, like "Monospac821 BT". Because of mswindows-font-regexp, they are skipped over. Changing (fontname "\\([a-zA-Z0-9 ]+\\)") in the mswindows-font-regexp definition hopefully fix it.

Auto-save filenames on Windows (2001-03-08)

Status: Patch Pending
Platforms: Windows native
Versions: XEmacs 21.1.13 native Windows prebuilt binary.
Developers: Ben Wing

The problem is that the intended auto-save file name contains \\ which is illegal in a file name

Auto-save filenames on Windows

Cygwin-XEmacs doesn't understand "C:/path" (2001-03-14)

Status: Reported
Platforms: Cygwin

No matter how much of my work uses Cygwin tools, I still use some things that expect DOS paths, like "C:/path", even with the forward slashes instead of the backslashes. Unfortunately, Cygwin-XEmacs doesn't grok that at all. Completion doesn't work, and "find-file-at-point" gets similarly confused.

Cygwin-XEmacs: get it to understand "C:/path"?

Unexpected menu invocation (2001-03-16)

Status: Patch Pending
Platforms: Windows 2000 native
Versions: 21.2-b45
Developers: Adrian Aichner

from sample.init.el:

  ;; Make Alt+accelerator traverse to the menu in new enough XEmacs
  ;; versions.  Note that this only overrides Meta bindings that would
  ;; actually invoke a menu, and that none of the most common commands
  ;; are overridden.  You can use ESC+key to access the overridden
  ;; ones if necessary.
  (setq menu-accelerator-enabled 'menu-force)

Despite the comment, these commonly used keys now invoke menus instead. Brrrr!

Setting menu-accelerator-enabled back to 'menu-fallback in a running 21.2-b45 XEmacs on native Windows 2000 still activates menus instead of the bound command.


Bug Status: Tabled

A bug gets tabled when we decide to release without a fully satisfactory fix to the bug.

Formatted like the open bug reports, with status tabled, and also includes the date tabled and the reason.

Crash (while t(all-hail-xemacs)) (2001-03-02)

Status: Replied
Platforms: Linux
Versions: XEmacs 21.2 (beta44) "Thalia" [Lucid] (i686-pc-linux) of Mon Feb 12 2001 on radioactive.socha.net
Developers: Jan Vroonhof

I eval'ed this expression and XEmacs dropped core.

Nobody else is able to replicate. We'll keep looking. Jan sez:

Does it also happen if you compile without ESD support? To me it very much looks like you are bitten by the problem where libesd sets up timeout for some reason and the SIGARLM ends up in our signal handler.

pdump in latest CVS -- gtk branch ? (2001-03-01)

Status: Patch Pending
Platforms: MSVC, Cygwin, (mingw?)
Versions: 21.2.44 "Thalia" (CVS 2001-03-01)
21.2.45 "Thelxepeia"
Developers: Ben Wing, Stephen Turnbull

Dangling references are kept in elisp_maphash after freeing a resized hashtable; can't load x-iso8859-1.el. Mike Alexander identified the bug.

Here is Ben's patch. I made a comment on xemacs-nt indicating that I thought there was an alternative approach to the one outlined in Ben's comment in the patch, but it turns out that it goes to the same place. :-)

An alternative patch <15025.44473.601020.23227@turnbull.sk.tsukuba.ac.jp> by Martin Buchholz that fixes this bug and similar ones, but is not a complete fix, has been approved.

The best way to deal with the problem is at the moment controversial; my current thinking is to have a fix implemented in 21.5 and backported after review and testing.

pdump in latest CVS -- gtk branch ?


Requests for features, future projects, etc. Normally "closed" by creating a wishlist item or a full-blown task RFC.

RFC: The Great Trunk Move in CVS (2001-02-28)

Status: Analysis
Platforms: n/a
Versions: n/a
Developers: Stephen Turnbull, Ben Wing, Hrvoje Niksic, cast of 000s

Lack of proposal on web site is bug :)


Bug Status: Closed

A bug gets closed when the developers involved consider the solution entirely satisfactory.

Formatted like the open bug reports, with status closed, and also includes the date closed and the reason.


display-time crash (2001-03-12)

Status: Closed
Reason: Fixed by Jan's patch.
Platforms: All?
Developers: Jan Vroonhof

I can get the crash now I run this code

;  (setq g (make-glyph (concat (locate-data-directory "time") "l-0.0.xpm")))
  (setq g (make-glyph "doesnotexist.text"))
  (setq e (make-extent nil nil))
  (setq global-mode-string (list (cons e g)))
  ; (set-extent-begin-glyph (make-extent 10 20) g)
  (make-frame-on-tty nil)
  (hanoi 1)

No crash happens if the glyph is put on a normal extent instead of
in the modeline.

On the display-time crash...

Cannot build Xemacs anymore under cygwin, was [Re: Sound on NT] (2001-03-06)

Status: Closed (2001-03-19)
Reason: Caused by unexec problem, which is fixed. Platforms: Cygwin
Versions: XEmacs 21.2 (beta41) "Polyhymnia" [Lucid] (i686-pc-cygwin, Mule) of Wed Jan 17 2001 on VZELL
Developers: Andy Piper

The difference between my last successfull build of xemacs and now is that I upgraded to cygwin1-20010222 gcc-2.95.2-7

Cannot build Xemacs anymore under cygwin, was [Re: Sound on NT]

Cannot build from CVS under Cygwin (2001-03-06)

Status: Closed (2001-03-19)
Reason: Fixed by defaulting to --relalloc=no.
Platforms: Cygwin
Versions: CVS 2001-03-06
Developers: Andy Piper

temacs hangs in update-elc process; can't build

Re: NT: I need build reports for GTK/XEmacs sources --with-gtk=no
Cannot build from CVS under Cygwin
Re: NT: I need build reports for GTK/XEmacs sources --with-gtk=no (tail of this has most info?)
Infloop in cygwin build ...

cannot find entry symbol _WinMainCRTStartup (2001-02-28)

Status: Closed (2001-03-19)
Reason: Andy says it's safe. Maybe code should be rewritten to eliminate the warning.
Platforms: cygwin
Versions: 21.2.4x
Developers: Ben Wing, Andy Piper

    What next....cygwin environment attached below

 ./configure --with-x=3Dno --site-includes=3D/usr/local/include            =
 --site-libraries=3D/usr/local/lib --with-dragndrop

 warning: cannot find entry symbol _WinMainCRTStartup; defaulting to 004010=
database.o(.text+0xc08):database.c: undefined reference to `_imp__dbm_first=

Andy sez:
I've had this occur randomly on and off. A full configure and rebuild 
usually cure sit for me

Can't find includes on Cygwin/Win2kSP1 (2001-03-06)

Status: Closed (2001-03-16)
Reason: Cygwin update was incomplete, not XEmacs problem. Platforms: Cygwin/Win2kSP1
Versions: XEmacs 21.2-b45 "Thelxepeia", i686-pc-cygwin
Developers: Andy Piper

John Turner: I've had many successful cygwin builds during the 21.2 dev cycle,
but my last successful cygwin build was b43

fresh CVS checkout after failure with the sandbox I had been using, just
to be sure it wasn't something screwy in my sandbox

INSTALLATION file attached

build fails with:

gcc -c -g -O3  -Demacs -I. -DHAVE_CONFIG_H -Wno-sign-compare -fno-caller-saves
In file included from nt.h:29,
                 from ntplay.c:25:
syswindows.h:58: windows.h: No such file or directory
syswindows.h:73: winspool.h: No such file or directory
syswindows.h:77: mmsystem.h: No such file or directory
syswindows.h:78: shellapi.h: No such file or directory
syswindows.h:79: ddeml.h: No such file or directory
make[1]: *** [ntplay.o] Error 1
make[1]: Leaving directory `/usr/local/src/xemacs-21.2/src'
make: *** [src] Error 2

uname -a: CYGWIN_NT-5.0 ZEPHYR 1.1.8(0.34/3/2) 2001-01-31 10:08 i686 unknown

Andy says the win32api package, recently split out from cygwin proper, needs to be installed.

[Failure] XEmacs 21.2-b45 "Thelxepeia", i686-pc-cygwin

Crash (assertion failed) in 21.2.44

Status: Closed (2001-03-10)
Reason: Bug reporter tested in recent version (21.2.45) and previously 100% reproducible failure did not recur.
Platforms: Cygwin
Versions: XEmacs 21.2 (beta44) "Thalia" [Lucid] (i686-pc-cygwin) of Tue Feb 13 2001 on LUCY
Developers: Stephen Turnbull

	LUCY:~$xemacs & LUCY:~$Fatal error: assertion failed, file /usr/local/src/xemacs-21.2/src/insdel.c, line 1115, ( retval) >= BI_BUF_BEG (buf) && retval <= BI_BUF_Z (buf)
[1]+  Exit 1                  xemacs

This may be related to Mule problems with Matt Tucker's syntax patch, so may have been fixed en passant. Followed up by Stephen Turnbull 2001-03-10.

Crash (assertion failed) in 21.2.44

Autoconfiguration Failures

Failure to detect capabilities or libraries known to be installed. Failure to pick up the right libraries.

Compiling with-x11=no under cygwin? (2001-03-06)

Status: Closed (2001-03-19)
Reason: Andy sez it's not a bug.
Platforms: Cygwin
Versions: XEmacs 21.2 (beta45) "Thelxepeia" [Lucid] (i686-pc-cygwin) of Fri Feb 23 2001 on LUCY
Developers: Andy Piper

Compiling with-x11=no under cygwin?

Doesn't Work

Sound on NT (2001-03-04)

Status: Closed (2001-03-19)
Reason: Andy sez it is not a bug.
Platforms: Cygwin
Versions: XEmacs 21.2 (beta41) "Polyhymnia" [Lucid] (i686-pc-cygwin, Mule) of Wed Jan 17 2001 on VZELL
Developers: Andy Piper

Sound doesn't work. At least one reporter suspects failure in ./configure.

Sound on NT


Conform with <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
Automatically validated by PSGML