12:04 p.m.
Just after temacs is created, the building process crashes with this
(no more information, even with debugging options):
EMACSBOOTSTRAPLOADPATH="/usr/local/src/XEmacs-21.2/src/../lisp/:/usr/local/src/xemacs"
EMACSBOOTSTRAPMODULEPATH="/usr/local/src/XEmacs-21.2/src/../modules/:/usr/local/src/xemacs"
./temacs -batch -l /usr/local/src/XEmacs-21.2/src/../lisp/update-elc.el
Fatal error: assertion failed, file /usr/local/src/XEmacs-21.2/src/alloc.c, line 2447,
abort()
make[1]: *** [update-elc.stamp] Aborted
make[1]: Leaving directory `/usr/local/src/xemacs/src'
make: *** [src] Error 2
I'm building with gcc 2.95.2, on a debian potato linux box:
--
/ / _ _ Didier Verna http://www.inf.enst.fr/~verna/
- / / - / / /_/ / EPITA / LRDE mailto:didier＠epita.fr
/_/ / /_/ / /__ / 14-16 rue Voltaire Tel. +33 (1) 44 08 01 77
94276 Kremlin-Bicêtre cedex Fax. +33 (1) 44 08 01 99
1 p.m.
>>>> "dv" == Didier Verna
<didier(a)xemacs.org> writes:
dv>
EMACSBOOTSTRAPLOADPATH="/usr/local/src/XEmacs-21.2/src/../lisp/:/usr/local/src/xemacs"
EMACSBOOTSTRAPMODULEPATH="/usr/local/src/XEmacs-21.2/src/../modules/:/usr/local/src/xemacs"
./temacs -batch -l /usr/local/src/XEmacs-21.2/src/../lisp/update-elc.el
dv> Fatal error: assertion failed, file /usr/local/src/XEmacs-21.2/src/alloc.c, line
2447, abort()
This ones for Olivier.
First, let's have a more informative crash:
Index: src/alloc.c
===================================================================
RCS file: /usr/CVSroot/XEmacs/xemacs/src/alloc.c,v
retrieving revision 1.42.2.51
diff -u -w -r1.42.2.51 alloc.c
--- alloc.c 2000/02/21 23:07:07 1.42.2.51
+++ alloc.c 2000/02/22 12:57:29
＠＠ -2443,8 +2443,8 ＠＠
void
dumpopaque (void *varaddress, size_t size)
{
- if (dumpopaqueidx >= countof (dumpopaquevec))
- abort ();
+ assert (dumpopaqueidx < countof (dumpopaquevec));
+
dumpopaquevec[dumpopaqueidx].data = varaddress;
dumpopaquevec[dumpopaqueidx].size = size;
dumpopaqueidx++;
My guess is:
Didier has a special site-load or whatever to load an unusually large
number of startup elisp. Executing that elisp generates some
opaques (see bytecode.c). This overflows a fixed length array.
First thing I would try (after being able to reproduce the problem,
which I cannot do currently), is to make all the fixed size arrays in
alloc.c into dynarrs.
Martin
1:05 p.m.
On Tue, Feb 22, 2000 at 05:00:31AM -0800, Martin Buchholz wrote:
Didier has a special site-load or whatever to load an unusually
large
number of startup elisp. Executing that elisp generates some
opaques (see bytecode.c). This overflows a fixed length array.
Except that dumpopaque as nothing to do with opaques :-) It is used
only (for now) for the DEFVAR_* pointing to a C interger.
First thing I would try (after being able to reproduce the problem,
which I cannot do currently), is to make all the fixed size arrays in
alloc.c into dynarrs.
Didier, please look at alloc.c and find:
struct dumpopaque_info
{
void *data;
size_t size;
} dumpopaquevec[200];
Increase the 200 to something bigger and tell us whether it changes
anything.
OG.
7:05 a.m.
Martin Buchholz <martin(a)xemacs.org> writes:
My guess is:
Didier has a special site-load or whatever to load an unusually large
number of startup elisp. Executing that elisp generates some
opaques (see bytecode.c). This overflows a fixed length array.
First thing I would try (after being able to reproduce the problem,
which I cannot do currently), is to make all the fixed size arrays in
alloc.c into dynarrs.
Then you guys didn't read my message. :-(
<877lg3dalg.fsf(a)dp50.ecc.u-tokyo.ac.jp>
http://www.xemacs.org/list-archives/xemacs-beta/200002/msg00302.html
You can't build XEmacs with --with-canna=yes right now.
--
Yoshiki Hayashi
7:30 a.m.
>>>> "YH" == Yoshiki Hayashi
<yoshiki(a)xemacs.org> writes:
YH> Then you guys didn't read my message. :-(
YH> <877lg3dalg.fsf(a)dp50.ecc.u-tokyo.ac.jp>
YH> http://www.xemacs.org/list-archives/xemacs-beta/200002/msg00302.html
YH> You can't build XEmacs with --with-canna=yes right now.
This is the same bug and Olivier has already OK'd increasing the size
of the dumpopaquevec array.
Message-ID: <20000222080543.A10718(a)nemesis.ncsl.nist.gov>
Sorry, I don't have the archive ref offhand (typical Tsukuba-dai/SINET
network bogons, can't reach www.xemacs.org), came in this am JST..
--
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."
7:32 a.m.
>>>> "YH" == Yoshiki Hayashi
<yoshiki(a)xemacs.org> writes:
YH> http://www.xemacs.org/list-archives/xemacs-beta/200002/msg00302.html
It's a bit embarrassing to note that w3.el claims this file is Bad
HTML.
Bill?
--
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."
8828
days inactive
8829
days old
5 comments
5 participants
participants (5)
-
Didier Verna -
Martin Buchholz -
Olivier Galibert -
Stephen J. Turnbull -
Yoshiki Hayashi