This page has not been updated in a very long time. An initial update is forthcoming soon (ca. 2001年12月03日), and further updates will occur regularly thereafter.
Last text update: 2001年11月29日
Last status update: 2001年07月08日
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.
Status:
Analysis
Platforms:
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.
Status:
Reported
Platforms:
Windows
Versions:
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?
Status:
Reported
Platforms:
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'
Developers:
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= iled.
ydyelwoxa26.fsf@pip.hpc2n.umu.se
Status:
Reported
Platforms:
Windows
Versions:
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
3A98E16F.42A03E25@666.com
Patches that people seem to have lost track of....
Status:
Patch Pending
Platforms:
Cygwin
Versions:
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
patch
Status:
Reported
Platforms:
Win32
Versions:
Unknown
Developers:
setup.ini line 1: parse error
setup.ini line 1: unrecognized line in setup.ini headers (do you have the
lates setup?)
Functionality that is expected to work (eg, has been reported to work in similar environments) doesn't.
Status:
Reported
Platforms:
cygwin
Versions:
XEmacs 21.2 (beta37) "Pan" [Lucid] (i686-pc-cygwin) of Sat Nov 18 2000 on PASTON
Developers:
xemacs -q
M-x vm
Get stack trace below. Repeat M-x vm and frame is set up fine.
14872.56606.62000.806209@paston.beasys.com
Status:
Reported
Platforms:
Linux
Versions:
Developers:
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.
m3itlwho4l.fsf@fermat.mts.jhu.edu
Status:
Replied
Platforms:
Versions:
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.
m2itlokxdk.fsf@zion.rcn.com
Status:
Replied
Platforms:
Versions:
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.
m2itlokxdk.fsf@zion.rcn.com
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.
ps7l4950de.fsf@diannao.ittc.ukans.edu
Status:
Replied
Platforms:
Windows
Versions:
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.
kvr90mokqx.fsf@vzell.de.oracle.com
Status:
Analysis
Platforms:
All?
Versions:
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.
200103052236.f25Ma6305816@foo-bar-baz.cc.vt.edu
Status:
Reported
Platforms:
All
Versions:
Developers:
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?
Status:
Reported
Platforms:
Windows
Versions:
Developers:
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.
puftoj6l.fsf@isoware.de
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
Status:
Reported
Platforms:
Cygwin
Versions:
Developers:
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"?
uy1ys3q8c0.fsf@tcsi.com
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.
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.
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.
20010302144036.21117.qmail@socha.net
Status:
Patch Pending
Platforms:
MSVC, Cygwin, (mingw?)
Versions:
21.2.44 "Thalia" (CVS 2001年03月01日)
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.
Requests for features, future projects, etc. Normally "closed" by creating a wishlist item or a full-blown task RFC.
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 :)
15004.64873.55077.815461@turnbull.sk.tsukuba.ac.jp
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.
Status:
Closed
Reason:
Fixed by Jan's patch.
Platforms:
All?
Versions:
Developers:
Jan Vroonhof
I can get the crash now I run this code (progn ; (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) (redisplay-frame) (make-frame-on-tty nil) (hanoi 1) (garbage-collect)) No crash happens if the glyph is put on a normal extent instead of in the modeline.
On the display-time crash...
g3g0gjb3u7.fsf@bart.isltd.insignia.com
Status:
Closed (2001年03月19日)
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]
Status:
Closed (2001年03月19日)
Reason:
Fixed by defaulting to --relalloc=no.
Platforms:
Cygwin
Versions:
CVS 2001年03月06日
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 ...
Status:
Closed (2001年03月19日)
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 Jari ./configure --with-x=3Dno --site-includes=3D/usr/local/include = --site-libraries=3D/usr/local/lib --with-dragndrop /usr/lib/gcc-lib/i686-pc-cygwin/2.95.2-6/../../../../i686-pc-cygwin/bin/ld:= warning: cannot find entry symbol _WinMainCRTStartup; defaulting to 004010= 00 database.o(.text+0xc08):database.c: undefined reference to `_imp__dbm_first= key' Andy sez: I've had this occur randomly on and off. A full configure and rebuild usually cure sit for me
wkvgpvbuc1.fsf@blue.sea.net
Status:
Closed (2001年03月16日)
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 ntplay.c 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年01月31日 10:08 i686 unknown
Andy says the win32api package, recently split out from cygwin proper, needs to be installed.
Status:
Closed (2001年03月10日)
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年03月10日.
Failure to detect capabilities or libraries known to be installed. Failure to pick up the right libraries.
Status:
Closed (2001年03月19日)
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
Status:
Closed (2001年03月19日)
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.
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