Hi there...
This might be a longish and confusing message, the latter as I've got
a terrible cold and my head feels like a melon - I beg your pardon in
advance.
>>>>> In article <m2u34n5tsq.fsf(a)altair.xemacs.org>, SL Baur
>>>>> <steve(a)xemacs.org> writes:
> NOTE: The tarballs are named `b48', but this is not a beta. It is a
> qualified[1] release candidate. Build it as highly optimized as you
> dare and make it feel pain and see how it goes.
I understand this is a bit like the moment during a wedding when the
priest tells the bride's lover in the crowd to step forward now or to
shut up forever. I guess I'd rather come forward and be told to sit
down again than not to tell this tale.
Prologue...
I was too busy to follow XEmacs development lately - I did little more
than to fetch the latest releases as you announced them and to build and
use them on my system. A few weeks ago (around b45/b46) I got a "new"
machine, a UDB (that's a cheapish 1995 Digital Alpha in a pizza box) and
decided to install Red Hat LiNUX 5.1 on it. I already had one machine
running a SuSE/Jurix/selfmade-LiNUX mix and my GF's system freshly setup
using the latest SuSE-LiNUX. You can see why I would want to maintain
only one kind of distribution, so all machines migrated to RH5.1.
Setup...
I often work at the UDB/X-Terminal ("din4844") in the living room, with
the programs running on the main machine ("bong") in another room (I log
in on bong using the xdm-chooser, so even the windowmanager runs on the
remote machine).
The Problem...
During the process of the migration I made some mistakes and unwise
decisions,[1] so I didn't pay a lot of attention to the core dumps I
would get occasionally from XEmacs when sending or reading mail with
Gnus. I didn't consider my system stable and other stuff was broken,
too.
Now my system seems to be stable, but it still breaks. I've seen it
twice with 21.0-b47, the first time just after sending a mail in Gnus,
the second time just after reading two or three mails in xemacs-beta
when I entered the group to check whether this is a "known problem".
I started XEmacs both times from the WindowMaker dock, so I didn't see any
output after the crash (I remember that output has the instructions on how
to get the info for a qualified bug report). I'm back to starting XEmacs
from the shell now and I'll spend the next days reading and writing as much
mail as possible to reproduce the problem.
Epilogue...
Please tell me that glibc-2.0.7 is "experimental shit" and that the
breakage is not related to XEmacs at all.
Footnotes:
[1] When bong migrated to RH, I must've made some mistake setting it up.
As a consequence, the libnss failed for some programs, e.g. BitchX and
XEmacs. There's no way I would live without "my" XEmacs. When I was
working on din4844 and started XEmacs it would fail on the hostname
lookup for din4844 (similar for BitchX). Setting DISPLAY to the IP#
instead of the hostname worked fine.
I should add that I run a local bind not only as a caching nameserver,
but also with a fake .bongnet zone - I know this wouldn't be strictly
necessary for the three to five machines in this network, but, hey, I
like to understand how this network shit works and setting it up in
this "sandbox" seems like a good way finding out.
Well, in that situation I did the single most stupid thing I could
think of (of course I considered it extremely clever), which was to
get glibc-2.0.94 and install it just over glibc-2.0.7. Of course
this involved ripping out the present gcc, egcs and some other shit
and replacing it with the latest releases. Of course nothing of
that was available as RPM, so I just got the tar.gzs, removed the
old packages and installed the freshly compiled stuff right in
/usr, "where it belongs".
I learned an awful lot in the process. I ended up with a setup where
BitchX and XEmacs worked beautifully. Nothing else did. I could not
compile most of the SRPMS anymore, let alone install i386-rpms. I
forced the reinstallation of the RH5.1-rpms right over my self-built
stuff. Some stuff worked, some didn't.
So I installed a fresh RH5.1 in a chroot-environment. Than I copied
that to my main system, and this time I made sure no files of previous
installations were left over. Now I think I have a stable system - and
I can compile everything which doesn't depend on some "bleedin' edge"
compiler or library.
--
Regards, Marc
<URL:http://www.saar.de/bong/signatur.html>
Linux is like a Vorlon. It is incredibly powerful, gives terse, cryptic answers
and has a lot of things going on in the background.
#!/bin/sh
# This is with RH-packages gcc-2.7.2.3-11 and glibc-2.0.7-13
./configure i386-bong-linux-gnu \
--debug=no \
--with-x11 \
--error-checking=none \
--memory-usage-stats=no \
--cflags="-g -O4 -malign-jumps=2 -malign-loops=2 -malign-functions=2 -finline-functions" \
--with-dialogs=athena3d \
--with-scrollbars=athena3d \
--with-menubars=lucid \
--with-sound=native \
--with-xim=no \
--with-mule=no \
--with-canna=no \
--with-offix=no \
--with-xface=yes \
--gung-ho \
2>&1 | tee result.configure
make clean
make all-elc 2>&1 | tee result.make # Do you want
make finder 2>&1 | tee result.make.finder # these files, too?
uname -a: Linux bong 2.0.34 #2 Wed Jul 8 13:23:23 /etc/localtime 1998 i586 unknown
./configure 'i386-bong-linux' '--debug=no' '--with-x11' '--error-checking=none' '--memory-usage-stats=no' '--cflags=-g -O4 -malign-jumps=2 -malign-loops=2 -malign-functions=2 -finline-functions' '--with-dialogs=athena3d' '--with-scrollbars=athena3d' '--with-menubars=lucid' '--with-sound=native' '--with-xim=no' '--with-mule=no' '--with-canna=no' '--with-offix=no' '--with-xface=yes' '--gung-ho'
XEmacs 21.0 "Zhong Wei" configured for `i386-bong-linux'.
Where should the build process find the source code? /fs/usr/local/lib/xemacs-21.0-b47
What installation prefix should install use? /usr/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 -O4 -malign-jumps=2 -malign-loops=2 -malign-functions=2 -finline-functions
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 X-Face message headers.
Compiling in support for JPEG image conversion.
Compiling in support for PNG image conversion.
Compiling in support for TIFF image conversion.
Compiling in native sound support.
Compiling in support for Berkeley DB.
Compiling in support for GNU DBM.
Compiling in support for ncurses.
Compiling in support for GPM (General Purpose Mouse).
Compiling in support for proper session-management.
Using Lucid menubars.
Using Athena-3d scrollbars.
Using Athena-3d dialog boxes.
Compiling in DLL support.
movemail will use "dot-locking" for locking mail spool files.
Using Lisp_Objects with minimal tagbits.
Using indexed lrecord implementation.
