Giacomo Boffi <giacomo.boffi(a)polimi.it> writes:
> XEmacs Release Engineer writes:
> > You can be at the bleeding edge, and use the latest CVS xemacs-21.2
> > sources. They are not guaranteed to even compile.
> >
> > cvs checkout -r release-21.2
>
> They are not guaranteed to even be retrieved...
> % cvs -z3 checkout -r release-21.2 xemacs
the tag in Martin´s mail should have been ´release-21-2´
Gunnar
User-visible release Engineering changes for this release:
This announcement contains the ChangeLog entries for the release.
The latest released beta now has the additional CVS tag `r21-2-latest'.
We now have 3 levels of stability available via CVS:
You can be at the bleeding edge, and use the latest CVS xemacs-21.2
sources. They are not guaranteed to even compile.
cvs checkout -r release-21.2
You can get the latest released beta, available via tarball or,
equivalently, via CVS.
ftp://…
[View More]ftp.xemacs.org/pub/xemacs/beta/xemacs-21.2/xemacs-21.2.24.*
cvs checkout -r r21-2-24
cvs checkout -r r21-2-latest
Or you can use xemacs-21.1. xemacs-21.1 is on the trunk of CVS,
although that is very likely to change soon.
ftp://ftp.xemacs.org/pub/xemacs/xemacs-21.1/xemacs-21.1.8.*
cvs checkout -A
Brief summary of changes to 21.2.24 "Hecate"
-- (coding-system-list) deals properly with coding system aliases, Shenghuo ZHU
-- configure support for ESD sound rewritte, Martin Buchholz
-- directory separator fix from Mike Alexander
-- Windows process support cleanup, Adrian Aichner
-- NT now encapsulates fstat to get correct file mod time, Adrian Aichner
-------- ChangeLog entries from xemacs-21.2.24/ChangeLog -------
1999-12-13 Martin Buchholz <martin(a)xemacs.org>
* configure.in:
* configure.usage:
- Autodetect NAS. Change Docs accordingly.
* configure.in: Warn if configure --with-xpm --without-x11.
1999-12-09 Martin Buchholz <martin(a)xemacs.org>
* configure.in: Clean up sound support.
- variable `old_nas' was used but never set.
- change `with_esd' to `with_esd_sound' for consistency.
- Don't trust the output of `esd-config --libs`; test it.
- Add `esd-config --cflags` to c_switch_site.
- Die if ESD sound requested, but not available.
- ESD is not dependent on X, therefore use LIBS, not libs_x.
1999-12-07 Martin Buchholz <martin(a)xemacs.org>
* configure.in (--with-sound): Variable with_esd was mispelled.
-------- ChangeLog entries from xemacs-21.2.24/lib-src/ChangeLog -------
1999-12-13 Martin Buchholz <martin(a)xemacs.org>
* etags.c: Fix warning: `_GNU_SOURCE' redefined.
-------- ChangeLog entries from xemacs-21.2.24/lisp/ChangeLog -------
1999-12-12 Gunnar Evermann <ge204(a)eng.cam.ac.uk>
* about.el (about-hackers): Update my email address.
-------- ChangeLog entries from xemacs-21.2.24/lwlib/ChangeLog -------
1999-12-14 Andy Piper <andy(a)xemacs.org>
* xlwtabs.c (TabsResize): reset need_layout so that we don't go
into infloop death.
1999-12-14 Andy Piper <andy(a)xemacs.org>
* xlwtabs.c (TabsSetValues): re-allocate GCs if font has changed.
1999-12-13 Andy Piper <andy(a)xemacs.org>
* xlwtabs.c (TabsResize): We need to expose the tabs after
clearing the window they are in.
1999-12-08 Andy Piper <andy(a)xemacs.org>
* xlwtabs.c: sync with Tabs 2.1.
1999-12-07 Andy Piper <andy(a)xemacs.org>
* lwlib-Xlw.c (lw_lucid_widget_p): make sure we pick up the
clip-window as well.
-------- ChangeLog entries from xemacs-21.2.24/man/ChangeLog -------
1999-12-07 Gunnar Evermann <ge204(a)eng.cam.ac.uk>
* xemacs/startup.texi (Startup Paths): fix typo: EMACSPACKAGEPATH
instead of PACKAGEPATH
From Marcus Harnisch <harnisch(a)mikrom.de>
-------- ChangeLog entries from xemacs-21.2.24/nt/ChangeLog -------
1999-11-28 Adrian Aichner <aichner(a)ecf.teradyne.com>
* xemacs.mak (depend): Only update `depend' if there were changes.
Use "perl ./make-src-depend" instead of "mkdepend".
-------- ChangeLog entries from xemacs-21.2.24/src/ChangeLog -------
1999-12-13 Martin Buchholz <martin(a)xemacs.org>
* sound.c (vars_of_sound): Provide esd-sound if HAVE_ESD_SOUND
* nas.c: Fix compiler warnings. Ansify. C++ compilation support.
1999-12-13 Kazuyuki IENAGA <ienaga(a)jsys.co.jp>
* input-method-xlib.c: Make sure src/ does not use Athena, except
indirectly through lwlib.
1999-12-13 Gunnar Evermann <ge204(a)eng.cam.ac.uk>
* dbxrc (pobj): Add opaque_ptr, remove opaque_list.
1999-12-10 Shenghuo ZHU <zsh(a)cs.rochester.edu>
* file-coding.c (add_coding_system_to_list_mapper):
- (coding-system-list) should list aliases correctly.
1999-12-07 Andy Piper <andy(a)xemacs.org>
* fileio.c (vars_of_fileio): directory separator fix from Mike Alexander.
* windowsnt.h (ditto):
1999-11-27 Adrian Aichner <aichner(a)ecf.teradyne.com>
* sysfile.h: Encapsulate `fstat' for Windows NT just like stat to
get consistent file modification times.
* sysdep.h: Declare `wait_for_termination' to use process handle,
not PID, on Windows NT native.
* sysdep.c (wait_for_termination): Use process handle, not PID, on
Windows NT native. Set exit code correctly on Windows NT.
(sys_subshell): Use process handle, not PID,
on Windows NT native.
(sys_fstat): Add ENCAPSULATE_FSTAT for Windows NT.
* process-nt.c (nt_create_process): Remove negative pid hack for
Windows 95.
(nt_send_process): Declare vol_proc volatile.
(get_internet_address): warn_when_safe if IP address cannot be
found.
(nt_open_network_stream): warn_when_safe if network stream fails
to open.
* ntproc.c (create_child): Remove negative pid hack for Windows
95. Remove incorrect and unnecessary USE_UNION_TYPE handling.
* nt.c (fstat): Encapsulate for Windows NT just like stat to get
consistent file modification times.
* callproc.c (call_process_cleanup): Use process handle, not PID,
on Windows NT native.
(Fcall_process_internal): Ditto. Close fd_error in parent. Don't
check for negative pid on Windows NT/9[58].
[View Less]
tor(a)spacetec.no (Tor Arntsen) writes:
> With the naming convention used it can indeed be confusing and it would
> probably help with having `branch' in the name. Unfortunately it isn't
> really possible to change the branch name after the fact :-(
But as the idea is to move the 21.2 "branch" to the head that isn't
really a problem. It just needed remember to create the 21.1 branch
with "branch" in its name.
Jan
P.S. It is funny that although almost the complete free software
…
[View More]community uses CVS as its tools nobody has come in and fixed at least
the simple annoyances (or stepped in and helped accelerate PRCS 2.0
development)
[View Less]
Martin Buchholz <martin(a)xemacs.org> writes:
> I actually have plans for optimizing map* in a bigger way. The basic
> idea is to avoid repeating computations for each function invocation.
You mean actually using that you know that you are invoking the same
function each time and stuff like that?
Jan
Has anybody noticed that this function doesn't work in
21.1 (patch 8) "Bryce Canyon"?
For instance if I type
C-h a test RET
the first entry in the list of "Functions and Macros" is
add-latest-path
If I click mouse-2 on this link, I get the right doc opened up in a
bottom window. But within this doc window, the following of links
does not seem to work.
On Fri, Dec 17, 1999 at 03:20:47AM -0800, Martin Buchholz wrote:
> Hey guys, when you remove something from the sources, you should do a
> grep to make sure you caught all the stragglers.
I did.
*sigh*
OG, looking for a brown paperbag.
nbecker(a)fred.net writes:
> I don't know - how do you all feel? I can understand that
> browse-url-??? is getting too crowded, and maybe some would rather
> not add kfm.
No matter how crowded it was, *removing* features the users depended
on without explanation is not excusable. browse-url-kfm should go
back. If it's crowded, then it should be proclaimed obsolete (with a
warning that warns of the "correct" usage) and removed several years
from now.
Martin Buchholz <martin(a)xemacs.org> writes:
> Here's a new patch, containing ONLY bug fixes, and incorporating
> Hrvoje's suggestions:
Seconded, for now. (Except for the ChangeLog, for which see Jan's
mail.)
> if (LISTP (seq))
> {
> + /* A devious `fn' could either:
> + - insert garbage into the list in front of us, causing XCDR to crash
> + - break the list behind us using (setcdr), causing the remaining elts
> + to lose their GCPRO status.
…
[View More]> +
> + We avoid these problems by making a copy of the list to be traversed.
> +
> + (We could also use EXTERNAL_LIST_LOOP and GCPRO the tail).
I think we should do that -- use EXTERNAL_LIST_LOOP and GCPRO the
tail. XEmacs is already *very* stack-hungry and copies a lot. All
unnecessary copying should be avoided.
> - Bufbyte *p = XSTRING_DATA (seq);
> - for (i = 0; i < leni; i++)
> + /* The string data of `seq' might be relocated during GC. */
Also, someone could change the characters in the [0, i) range, and
relocate string's chars under us by changing the byte length of a
character before us. (See set_string_char() and resize_string().)
[View Less]
>>>>> "Jan" == Jan Vroonhof <vroonhof(a)math.ethz.ch> writes:
Jan> nbecker(a)fred.net writes:
>> 1) Probably noone else is going to figure this out, as compared to the
>> older implementation which put kfm right on the Options menu.
>> Perhaps a FAQ entry would help.
>>
>> 2) You need to redump in order to update the menus after updating the
>> packages.
Jan> Would be better to simple write browse-…
[View More]url-kfm again and sent it to
Jan> Matt Curtin?
Jan> It should never have been removed. Sorry for missing that.
I don't know - how do you all feel? I can understand that
browse-url-??? is getting too crowded, and maybe some would rather not
add kfm. I think kde is so important that kfm does belong on the menu
- but I don't want to force my preferences on everyone else.
[View Less]