10:27 p.m.
I'm doing a small sync with Emacs 21.3 to pick up some new features of
define-minor-mode that they have. I think that part is done. I am now
trying to build the packages to I can check whether my sync messed
anything up. I am compiling with a MULE-enabled XEmacs. I am
encountering two problems.
(1) When compiling the packages with XEmacs 21.4.14 on a Linux box with
1GB RAM and after doing "unlimit stacksize", the make is bombing out
like this:
make[2]: Entering directory `/usr/src/xemacs/packages/xemacs-packages/xslide'
xemacs -no-autoloads -vanilla -batch -eval '(setq stack-trace-on-error t
load-ignore-out-of-date-elc-files t load-show-full-path-in-messages t)' -l
/usr/src/xemacs/packages/package-compile.el -- ispell mail-lib xemacs-base -- -l cus-face
-l regexp-opt -l font-lock -l sendmail -l xslide-data -l xslide-abbrev -l xslide-font -l
xslide-process -l xslide -f batch-byte-compile auto-autoloads.el
Loading /usr/local/lib/xemacs-21.4.14/lisp/auto-autoloads...
Loading /usr/src/xemacs/packages/xemacs-packages/xemacs-base/auto-autoloads...
Loading /usr/src/xemacs/packages/xemacs-packages/mail-lib/auto-autoloads...
Loading /usr/src/xemacs/packages/xemacs-packages/ispell/auto-autoloads... # bind (
stack-trace-on-signal debug-on-signal stack-trace-on-error debug-on-error)
signal(error ("Stack overflow in regexp matcher"))
# bind (data error-symbol)
signal-error(error ("Stack overflow in regexp matcher"))
# bind (error-data)
normal-top-level()
# (condition-case ... . error)
# (catch top-level
...)
Stack overflow in regexp matcher
xemacs exiting
.
make[2]: *** [auto-autoloads.elc] Error 255
make[2]: Leaving directory `/usr/src/xemacs/packages/xemacs-packages/xslide'
make[1]: *** [xslide/bytecompile.target] Error 2
make[1]: Leaving directory `/usr/src/xemacs/packages/xemacs-packages'
make: *** [xemacs-packages/bytecompile.target] Error 2
Trying that part of the compile by hand, I found that the overflow
is occurring while loading xslide-font, on line 191, while doing
this:
(regexp-opt-depth xsl-fo-attribute-symbol-alist-regexp)
where xsl-fo-attribute-symbol-alist-regexp was just produced by
regexp-opt. This does not happen when building with XEmacs
21.5.16.
(2) The build crashes in hyperbole, because hvm.el calls
vm-set-modflag-of, which vm/vm-autoload.el claims is defined in
vm-message.el. It isn't. From:
http://list-archive.xemacs.org/xemacs-patches/200309/msg00011.html
I gather that the call in the hyperbole code should be changed to
vm-set-stuff-flag-of, and that vm-autoload.el should be fixed.
--
Jerry James
http://www.ittc.ku.edu/~james/
11:25 p.m.
Jerry James writes:
(2) The build crashes in hyperbole, because hvm.el calls
vm-set-modflag-of, which vm/vm-autoload.el claims is defined in
vm-message.el. It isn't. From:
http://list-archive.xemacs.org/xemacs-patches/200309/msg00011.html
I gather that the call in the hyperbole code should be changed to
vm-set-stuff-flag-of, and that vm-autoload.el should be fixed.
vm-autoload.el is machine generated. 'make' must not have been
run in a while.
Let's remove VM from the XEmacs package tree. If folks want the
VM package, they can download it from VM's distribution sites.
Is there a way the package index can specify a different download
location for a package?
12:13 a.m.
|--==> "KJ" == Kyle Jones <kyle_jones(a)wonderworks.com> writes:
KJ> Let's remove VM from the XEmacs package tree.
Why would you want to do that?
KJ> Is there a way the package index can specify a different download
KJ> location for a package?
No. The package index doesn't specify download locations. And with
the current implementation of PUI, using multiple download locations in
the one invocation is fraught with danger.
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| XEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------<sryoungs(a)bigpond.net.au>---|
1:15 a.m.
Steve Youngs writes:
|--==> "KJ" == Kyle Jones
<kyle_jones(a)wonderworks.com> writes:
KJ> Let's remove VM from the XEmacs package tree.
Why would you want to do that?
We could get rid of the synchronization problems and extra work
by dropping the XEmacs distribution path. I already create an
XEmacs package for each VM release and VM already has multiple
distribution sites.
KJ> Is there a way the package index can specify a different
download
KJ> location for a package?
No. The package index doesn't specify download locations. And with
the current implementation of PUI, using multiple download locations in
the one invocation is fraught with danger.
Too bad. I rather like the way the FreeBSD ports system works.
The official source tarball is downloaded from its normal place
on the net and then any FreeBSD-specific patches are applied.
Finally the package is built and installed.
2 a.m.
|--==> "KJ" == Kyle Jones <kyle_jones(a)wonderworks.com> writes:
KJ> Steve Youngs writes:
>|--==> "KJ" == Kyle Jones
<kyle_jones(a)wonderworks.com> writes:
>
KJ> Let's remove VM from the XEmacs package tree.
>
>Why would you want to do that?
KJ> We could get rid of the synchronization problems and extra work
KJ> by dropping the XEmacs distribution path. I already create an
KJ> XEmacs package for each VM release and VM already has multiple
KJ> distribution sites.
But what about the packages in the XEmacs tree that have a dependence
on VM? A quick `igrep-find' reveals that BBDB, MailCrypt, TM, and
Hyperbole all need VM.
I'm sure that I could automate much of (if not all of) the extra work
involved in you being able to keep the XEmacs VM in sync with your
master copy. Then having the extra distribution path would be a good
thing, no?
KJ> Is there a way the package index can specify a different download
KJ> location for a package?
>
>No.
KJ> Too bad. I rather like the way the FreeBSD ports system works.
KJ> The official source tarball is downloaded from its normal place
KJ> on the net and then any FreeBSD-specific patches are applied.
KJ> Finally the package is built and installed.
Yeah, the FreeBSD system is pretty cool. The thing about the XEmacs
packages is that they are built _before_ distribution, not after.
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| XEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------<sryoungs(a)bigpond.net.au>---|
12:55 a.m.
On Fri, 2003-11-21 at 00:27, Jerry James wrote:
(1) When compiling the packages with XEmacs 21.4.14 on a Linux box
with
1GB RAM and after doing "unlimit stacksize", the make is bombing out
like this:
[...]
signal(error ("Stack overflow in regexp matcher"))
I saw this too (xslide, 21.4.14, Linux), following Vin's tip cured it:
http://list-archive.xemacs.org/xemacs-beta/200310/msg00480.html
2:34 a.m.
Ville Skyttä <scop(a)xemacs.org> wrote:
I saw this too (xslide, 21.4.14, Linux), following Vin's tip
cured it:
http://list-archive.xemacs.org/xemacs-beta/200310/msg00480.html
Ah, great. Thank you!
--
Jerry James
http://www.ittc.ku.edu/~james/
7483
days inactive
7484
days old
6 comments
4 participants
participants (4)
-
Jerry James -
Kyle Jones -
Steve Youngs -
Ville Skyttä