8:13 a.m.
Vin Shelton <acs(a)alumni.princeton.edu> writes:
Michael -
Hi Vin,
Thanks for the update. Unfortunately, as you suspected, there's
an
issue building the package repo. I'm using latest 21.4 from hg, and
I'm getting a failure in ediff-vers.el. This compiled fine before
your changes; I'm guessing it has something to do with autoloads.
The actual error is:
byte-compile-report-error((file-error "Cannot open load file"
"executable"))
which seems kinda weird. I'm investigating now, but in case someone
has some ideas, I'd be happy for any pointers.
In (the new) tramp-compat.el, there is the following code snippet:
--8<---------------cut here---------------start------------->8---
(eval-and-compile
;; Some packages must be required for XEmacs, because we compile
;; with -no-autoloads.
(when (featurep 'xemacs)
(require 'cus-edit)
(require 'env)
(require 'executable)
(require 'outline)
(require 'passwd)
(require 'pp)
(require 'regexp-opt))
(require 'advice)
(require 'custom)
(require 'format-spec)
(require 'shell)
(require 'tramp-loaddefs)
...
--8<---------------cut here---------------end--------------->8---
I have added the "(when (featurep 'xemacs) ..." form, because there were
compilation errors otherwise in my local environment. But I'm open to
any enhancement.
Thanks,
Vin
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:54 p.m.
New subject: Tramp modified-p behaviour
Hi, Michael Albinus -
(Michael Kifer - I'm copying you because I had to add sh-script to the
list of ediff required packages.)
On Mon, Jan 21, 2013 at 3:13 AM, Michael Albinus <michael.albinus(a)gmx.de> wrote:
Vin Shelton <acs(a)alumni.princeton.edu> writes:
> The actual error is:
>
> byte-compile-report-error((file-error "Cannot open load file"
"executable"))
>
> which seems kinda weird. I'm investigating now, but in case someone
> has some ideas, I'd be happy for any pointers.
In (the new) tramp-compat.el, there is the following code snippet:
--8<---------------cut here---------------start------------->8---
(eval-and-compile
;; Some packages must be required for XEmacs, because we compile
;; with -no-autoloads.
(when (featurep 'xemacs)
(require 'cus-edit)
(require 'env)
(require 'executable)
(require 'outline)
(require 'passwd)
(require 'pp)
(require 'regexp-opt))
(require 'advice)
(require 'custom)
(require 'format-spec)
(require 'shell)
(require 'tramp-loaddefs)
...
--8<---------------cut here---------------end--------------->8---
I have added the "(when (featurep 'xemacs) ..." form, because there were
compilation errors otherwise in my local environment. But I'm open to
any enhancement.
OK. Thanks for the tip. I added sh-script to the requires lines of
ediff and pcl-cvs and I could compile those packages.
I can now build all the packages except tramp:
xemacs -no-autoloads -vanilla -batch -eval '(setq stack-trace-on-error
t load-always-display-messages t load-ignore-out-of-date-elc-files t
load-show-full-path-in-messages t)' -eval '(setq load-path (list
lisp-directory))' -l
/home/opt/build/xemacs-packages-2013-01-20/package-compile.el -- tramp
xemacs-base vc efs dired mail-lib gnus ediff sh-script edebug -- -f
batch-byte-compile lisp/tramp-gvfs.el
Loading /opt/xemacs/lib/xemacs-21.4.22/lisp/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/edebug/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/sh-script/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/ediff/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/gnus/lisp/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/mail-lib/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/dired/auto-autoloads...
Loading /home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/efs/auto-autoloads...
Loading /home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/vc/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/xemacs-base/auto-autoloads...
Loading
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/tramp/lisp/auto-autoloads...
Compiling
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/tramp/lisp/tramp-gvfs.el...
While compiling toplevel forms in file
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/tramp/lisp/tramp-gvfs.el:
** reference to free variable dbus-get-unique-name
Loading customization dependencies...
Loading customization dependencies...done
!! File error (("Cannot open load file" "dbus"))
backtrace(nil t)
# bind (error-info)
byte-compile-report-error((file-error "Cannot open load file"
"dbus"))
# bind (error-info)
I could not find a dbus module anywhere in the XEmacs sources (not
21.4, 21.5 or packages). I'm pretty sure FSF emacs has dbus support -
is that something you've added to your XEmacs environment?
Please let me know what the right course of action is.
Regards,
Vin
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
3:22 p.m.
New subject: Tramp modified-p behaviour
Vin Shelton <acs(a)alumni.princeton.edu> writes:
Hi, Michael Albinus -
Hi Vin,
I can now build all the packages except tramp:
While compiling toplevel forms in file
/home/opt/build/xemacs-packages-2013-01-20/xemacs-packages/tramp/lisp/tramp-gvfs.el:
** reference to free variable dbus-get-unique-name
Loading customization dependencies...
Loading customization dependencies...done
!! File error (("Cannot open load file" "dbus"))
backtrace(nil t)
# bind (error-info)
byte-compile-report-error((file-error "Cannot open load file"
"dbus"))
# bind (error-info)
I could not find a dbus module anywhere in the XEmacs sources (not
21.4, 21.5 or packages). I'm pretty sure FSF emacs has dbus support -
is that something you've added to your XEmacs environment?
Nope, XEmacs has no dbus support yet. I have removed tramp-gvfs from the
Makefile.
Maybe we need to remove also tramp-gw, don't know.
Regards,
Vin
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:35 a.m.
New subject: Tramp modified-p behaviour
Michael Albinus <michael.albinus(a)gmx.de> writes:
Nope, XEmacs has no dbus support yet. I have removed tramp-gvfs from
the
Makefile.
Maybe we need to remove also tramp-gw, don't know.
I have also removed tramp-ftp, tramp-gw and tramp-loaddefs from the
Makefile. However, there might be a problem with tramp-loaddefs.el. In
the package repository, I've copied the generated one from the Tramp
repository. This still contains entries for tramp-ftp, tramp-gvfs and
tramp-gw, which are invalid. Wouldn't it be better to generate
tramp-loaddefs.el in the package repository? How shall I plug into the
package Makefile? In the Tramp repository Makefile, there is
--8<---------------cut here---------------start------------->8---
autoloads: $(builddir)/tramp-loaddefs.el
$(builddir)/tramp-loaddefs.el: $(LISP_FILES)
# XEmacs must be advised to put tramp.el definitions at the beginning.
# We redefine `autoload-trim-file-name', therefore.
$(EM) -l autoload \
--eval "(setq generate-autoload-cookie \";;;###tramp-autoload\")"
\
--eval "(setq generated-autoload-file \
(expand-file-name \"tramp-loaddefs.el\"))" \
--eval "(setq make-backup-files nil)" \
--eval "(unless (string-equal \"$(EMACS_INFO)\" \"emacs\")
\
(defalias 'autoload-trim-file-name \
'file-name-sans-extension))" \
-f batch-update-autoloads .
# XEmacs uses hard coded feature `lisp-autoloads'.
ï¼ sed -e s/lisp-autoloads/tramp-loaddefs/g $ï¼ >$ï¼ .$$$$ && \
mv -f $ï¼ .$$$$ $ï¼
# Emacs 22 does not add the `provide' form to the generated loaddefs.el.
ï¼ grep -q provide $ï¼ || echo "(provide 'tramp-loaddefs)" >>$ï¼
--8<---------------cut here---------------end--------------->8---
> Regards,
> Vin
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:07 p.m.
New subject: Tramp modified-p behaviour
Michael Albinus writes:
tramp-gw, which are invalid. Wouldn't it be better to generate
tramp-loaddefs.el in the package repository?
No. The XEmacs package build process automatically collects autoloads
into a file called auto-autoloads.el. XEmacs shouldn't use
tramp-loaddefs even if it does exist.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
3:15 p.m.
New subject: Tramp modified-p behaviour
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
> tramp-gw, which are invalid. Wouldn't it be better to
generate
> tramp-loaddefs.el in the package repository?
No. The XEmacs package build process automatically collects autoloads
into a file called auto-autoloads.el. XEmacs shouldn't use
tramp-loaddefs even if it does exist.
Unfortunately, it is not so simple. Tramp uses two different autoload
cookies: ";;;###autoload" and ";;;###tramp-autoload". The former one
is
collected into auto-autoloads.el. The latter one must be collected into
a file tramp-loaddefs.el, which is required by the other Tramp files.
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4 p.m.
New subject: Tramp modified-p behaviour
Michael Albinus writes:
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
> > tramp-gw, which are invalid. Wouldn't it be better to generate
> > tramp-loaddefs.el in the package repository?
>
> No. The XEmacs package build process automatically collects autoloads
> into a file called auto-autoloads.el. XEmacs shouldn't use
> tramp-loaddefs even if it does exist.
Unfortunately, it is not so simple. Tramp uses two different autoload
cookies: ";;;###autoload" and ";;;###tramp-autoload". The former one
is
collected into auto-autoloads.el. The latter one must be collected into
a file tramp-loaddefs.el, which is required by the other Tramp files.
What is the purpose of this division? Autoloads are not that
expensive if you don't actually call the functions.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4:20 p.m.
New subject: Tramp modified-p behaviour
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
> Unfortunately, it is not so simple. Tramp uses two different
autoload
> cookies: ";;;###autoload" and ";;;###tramp-autoload". The
former one is
> collected into auto-autoloads.el. The latter one must be collected into
> a file tramp-loaddefs.el, which is required by the other Tramp files.
What is the purpose of this division? Autoloads are not that
expensive if you don't actually call the functions.
It's the result of a code refactoring in order to minimize Tramp startup
times, and to hide Tramp as much as possible. Emacs users, who do not
use remote files, will see only tramp.el loaded. This cannot be avoided
completely, because applying "C-x C-f /s TAB" will already load Tramp in
the GNU Emacs case. Even function definitions which check, whether
another module shall be loaded, are pushed into the autoloads.
All other autoload definitions, which could also consist of whole
functions instead of the function definition, are in
tramp-loaddefs.el. Again, this file shall not be loaded for users who
feel disturbed by Tramp.
For sure, one could use a different approach. But is it a problem to
generate the tramp-loaddefs.el file inside the package repository? IIRC,
one could add local rules to the make process.
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
12:33 a.m.
New subject: Tramp modified-p behaviour
Michael Albinus writes:
It's the result of a code refactoring in order to minimize Tramp
startup
times, and to hide Tramp as much as possible.
That's what I thought. But AFAICS, it can't work. Sure, you can
reduce loaded code volume this way, but it doesn't "hide" Tramp; the
same entry points will be visible except perhaps for the most basic
non-users. The only way to hide Tramp is to actually reduce the
number of entry points.
For sure, one could use a different approach. But is it a problem to
generate the tramp-loaddefs.el file inside the package repository? IIRC,
one could add local rules to the make process.
One can always do that kind of thing. Unfortunately, I don't have
time for it for the next few months.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
12:30 p.m.
New subject: Tramp modified-p behaviour
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
> For sure, one could use a different approach. But is it a
problem to
> generate the tramp-loaddefs.el file inside the package repository? IIRC,
> one could add local rules to the make process.
One can always do that kind of thing. Unfortunately, I don't have
time for it for the next few months.
For the time being, I have modified tramp-loaddefs.el manually.
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
3:40 p.m.
New subject: Tramp modified-p behaviour
Michael Albinus writes:
For the time being, I have modified tramp-loaddefs.el manually.
Thanks!
I think that's the best approach for the near term. If you (or
somebody else with a strong interest in Tramp) decide to dive into the
automation, feel free to post here at any sign of insanity in the
Makefile.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:43 a.m.
New subject: Tramp modified-p behaviour
First of all, thanks to you all for the hard work of updating the
XEmacks package. Unfortunately, I have the following issue trying to use
the new tramp package:
--Lisp error: (void-variable tramp-gw-tunnel-method)
tramp-compute-multi-hops(["scpc" nil "myhost" "myfile"
nil])
byte-code("..." [l-method gw l-host d p vec processp process-status (run open)
tramp-get-buffer boundp non-essential nil tramp-file-name-user 0 format "Opening
connection for %s using %s" tramp-file-name-host tramp-file-name-method "Opening
connection for %sï¼ %s using %s" 3 "%s..." (byte-code)])
byte-code("..." [process-name pos tramp-current-connection process-environment
p vec tramp-get-connection-process tramp-get-connection-property "process-name"
nil copy-sequence tramp-get-connection-buffer processp process-status (run open) butlast
append tramp-time-diff current-time 5 suppress (byte-code)])
Thanks,
Nei
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:06 p.m.
New subject: Tramp modified-p behaviour
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
First of all, thanks to you all for the hard work of updating the
XEmacks package. Unfortunately, I have the following issue trying to use
the new tramp package:
--Lisp error: (void-variable tramp-gw-tunnel-method)
tramp-compute-multi-hops(["scpc" nil "myhost" "myfile"
nil])
byte-code("..." [l-method gw l-host d p vec processp process-status
(run open) tramp-get-buffer boundp non-essential nil
tramp-file-name-user 0 format "Opening connection for %s using %s"
tramp-file-name-host tramp-file-name-method "Opening connection for
%sï¼ %s using %s" 3 "%s..." (byte-code)])
byte-code("..." [process-name pos tramp-current-connection
process-environment p vec tramp-get-connection-process
tramp-get-connection-property "process-name" nil copy-sequence
tramp-get-connection-buffer processp process-status (run open) butlast
append tramp-time-diff current-time 5 suppress (byte-code)])
Should be fixed now in the Tramp XEmacs Package Repository.
Thanks,
Nei
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
10:18 a.m.
New subject: Tramp modified-p behaviour
On Di, 2013-01-29 at 15:06 +0100, Michael Albinus wrote:
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
> First of all, thanks to you all for the hard work of updating the
> XEmacks package. Unfortunately, I have the following issue trying to use
> the new tramp package:
Should be fixed now in the Tramp XEmacs Package Repository.
I have updated the package but I still get the error :-(
void is not nil so it will scream, that's my reading. you either need to
(defvar ... nil) the symbols or maybe you want to check boundp instead?
--Lisp error: (void-variable tramp-gw-tunnel-method)
(and tramp-gw-tunnel-method tramp-gw-socks-method (string-match (format
"^\\(%s\\|%s\\)$" tramp-gw-tunnel-method ...
(if (and tramp-gw-tunnel-method tramp-gw-socks-method (string-match ... ...)) (let (...
...) (unless ... ...) ...
(when (and tramp-gw-tunnel-method tramp-gw-socks-method (string-match ... ...)) (let
(... ...) (unless ... ...) ...
(let ((target-alist ...) (hops ...) (item vec) choices proxy) (dolist (proxy ...) (let
... ... ... ...)) (when ...
tramp-compute-multi-hops(["scpc" nil "myhost" "" nil])
Take care and keep up the good work,
Nei
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
12:26 p.m.
New subject: Tramp modified-p behaviour
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
Hi,
> Should be fixed now in the Tramp XEmacs Package Repository.
I have updated the package but I still get the error :-(
How did you update? I have committed the changes to the Mercurial
repository, but there is no new Tramp package yet which could be
downloaded via Tools -> Packages.
This will be announced by Norbert, when available. Or you swith to the
sources from Mercurial.
Nei
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
12:45 p.m.
New subject: Tramp modified-p behaviour
On Mi, 2013-01-30 at 13:26 +0100, Michael Albinus wrote:
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
Hi,
>> Should be fixed now in the Tramp XEmacs Package Repository.
>
> I have updated the package but I still get the error :-(
How did you update? I have committed the changes to the Mercurial
repository, but there is no new Tramp package yet which could be
downloaded via Tools -> Packages.
I used "hg pull -u"
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
6:59 p.m.
New subject: Tramp modified-p behaviour
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
On Mi, 2013-01-30 at 13:26 +0100, Michael Albinus wrote:
> Ailin Nemui <ailin.nemui(a)gmail.com> writes:
>
> Hi,
>
> >> Should be fixed now in the Tramp XEmacs Package Repository.
> >
> > I have updated the package but I still get the error :-(
>
> How did you update? I have committed the changes to the Mercurial
> repository, but there is no new Tramp package yet which could be
> downloaded via Tools -> Packages.
I used "hg pull -u"
You're right, we need to check whether the variables are bound. I've
pushed a respective fix to the repository.
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
9:50 a.m.
New subject: New tramp observations
Just to keep the discussion up, I just ran into this issue with new
tramp: after file changed on remote, XEmacs will ask me to edit(y/n/r),
I choose revert but the question keeps reappearing and revert is not
happening. Any thoughts on this?
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:16 a.m.
New subject: New tramp observations
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
Just to keep the discussion up, I just ran into this issue with new
tramp: after file changed on remote, XEmacs will ask me to edit(y/n/r),
I choose revert but the question keeps reappearing and revert is not
happening. Any thoughts on this?
Do you mean the file is changed outside XEmacs?
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:23 a.m.
New subject: New tramp observations
On Do, 2013-01-31 at 12:16 +0100, Michael Albinus wrote:
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
> Just to keep the discussion up, I just ran into this issue with new
> tramp: after file changed on remote, XEmacs will ask me to edit(y/n/r),
> I choose revert but the question keeps reappearing and revert is not
> happening. Any thoughts on this?
Do you mean the file is changed outside XEmacs?
yep, it is changed on the remote server with another editor (happens to
be also xemacs but shouldn't matter ;) )
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
11:50 a.m.
New subject: New tramp observations
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
> Do you mean the file is changed outside XEmacs?
yep, it is changed on the remote server with another editor (happens to
be also xemacs but shouldn't matter ;) )
Well, one of the new features in Tramp is caching of file
attributes. You should have seen the performance improvement :-)
The dark side are stale caches. You can control Tramp's caching
behaviour by the variable `remote-file-name-inhibit-cache'. This is a
GNU Emacs variable, which is acknowledged by Tramp. Maybe it shall also
be introduced in XEmacs 21.5. Its definition is
--8<---------------cut here---------------start------------->8---
remote-file-name-inhibit-cache is a variable defined in `files.el'.
Its value is 10
Documentation:
Whether to use the remote file-name cache for read access.
When `nil', never expire cached values (caution)
When `t', never use the cache (safe, but may be slow)
A number means use cached values for that amount of seconds since
caching.
The attributes of remote files are cached for better performance.
If they are changed outside of Emacs's control, the cached values
become invalid, and must be reread. If you are sure that nothing
other than Emacs changes the files, you can set this variable to `nil'.
If a remote file is checked regularly, it might be a good idea to
let-bind this variable to a value less than the interval between
consecutive checks. For example:
(defun display-time-file-nonempty-p (file)
(let ((remote-file-name-inhibit-cache (- display-time-interval 5)))
(and (file-exists-p file)
(< 0 (nth 7 (file-attributes (file-chase-links file)))))))
--8<---------------cut here---------------end--------------->8---
In your case, either set this variable to `t', or to an appropriate
number you feel able to deal with. GNU Emacs' default of 10 might be a
good compromise.
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
12:17 p.m.
New subject: New tramp observations
On Do, 2013-01-31 at 12:50 +0100, Michael Albinus wrote:
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
>> Do you mean the file is changed outside XEmacs?
>
> yep, it is changed on the remote server with another editor (happens to
> be also xemacs but shouldn't matter ;) )
Well, one of the new features in Tramp is caching of file
attributes. You should have seen the performance improvement :-)
Maybe we are misunderstanding each other. Thanks for the pointers to the
cache. Let me explain my issue again:
Edit file / save it / on xemacs "A"
Edit file outside of xemacs
try to resume editing of file IN xemacs
xemacs will detect that the file has changed, and ask you what to do:
you have choice to edit the current file ("y"), abort editing ("n")
and
revert the file to its state on disk ("r"). Maybe there is also option
to do a diff between them
if I know that I have no local changes in this xemacs buffer, and I know
that I changed the file outside xemacs, I would want to type on
"r" (revert, means it will load a fresh copy of the file but keep point
etc)
this used to work for local and tramp files. Now, it gets stuck in an
endless loop on tramp files. When I type "r", it asks me again (y/n/r).
Only y or n can get me out there.
I would expect "r" to revert the file (fetch the changed copy via tramp)
and let me resume editing in the buffer. Which is the previous behaviour
and the behaviour for local files.
Does that make anything clearer? Btw "r" should probably be doing
something similar to M-x revert-buffer
Nei
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
12:37 p.m.
New subject: New tramp observations
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
Hi,
this used to work for local and tramp files. Now, it gets stuck in
an
endless loop on tramp files. When I type "r", it asks me again (y/n/r).
Only y or n can get me out there.
I would expect "r" to revert the file (fetch the changed copy via tramp)
and let me resume editing in the buffer. Which is the previous behaviour
and the behaviour for local files.
Does that make anything clearer? Btw "r" should probably be doing
something similar to M-x revert-buffer
I Have no idea what XEmacs does. I have just tried to explain what has
changed in Tramp. And yes, I believe I understand your scenario.
Have you checked, whether setting `remote-file-name-inhibit-cache' to
`t' makes a difference?
Nei
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:51 p.m.
New subject: New tramp observations
On Do, 2013-01-31 at 13:37 +0100, Michael Albinus wrote:
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
Hi,
> this used to work for local and tramp files. Now, it gets stuck in an
> endless loop on tramp files. When I type "r", it asks me again (y/n/r).
> Only y or n can get me out there.
>
> I would expect "r" to revert the file (fetch the changed copy via tramp)
> and let me resume editing in the buffer. Which is the previous behaviour
> and the behaviour for local files.
Have you checked, whether setting `remote-file-name-inhibit-cache' to
`t' makes a difference?
I have checked it and it does not seem to make any difference. Still
gets into the endless "r" loop with remote-file-name-inhibit-cache t
Nei
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4:57 p.m.
New subject: New tramp observations
Ailin Nemui <ailin.nemui(a)gmail.com> writes:
Hi,
I have checked it and it does not seem to make any difference. Still
gets into the endless "r" loop with remote-file-name-inhibit-cache t
I've debugged a while, and maybe I've trapped the error. Could you,
please, check whether my recent patch fixes the problem? I've committed
it to the repository.
There shall also no need to set `remote-file-name-inhibit-cache' for
this special scenario. However, as said earlier, Emacs' default value of
10 might be useful in general, when you expect files being changed
outside XEmacs.
Nei
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4:49 a.m.
New subject: Tramp modified-p behaviour
Michael Albinus wrote:
For sure, one could use a different approach. But is it a problem to
generate the tramp-loaddefs.el file inside the package repository? IIRC,
one could add local rules to the make process.
MH-E, too, uses a loaddefs file to "hide" internal interfaces. It's
something I'll need to deal with for MH-E 8.3, but I haven't gotten to
it yet. My tentative plan is this:
- include upstream's Makefile in the XEmacs package tree (probably call
it Makefile.upstream)
- change xemacs-packages/mh-e/Makefile to include a rule for
lisp/mh-loaddefs.el. The rule will do "make -f Makefile.upstream
loaddefs".
- tweak makefiles as needed
mike
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4:18 p.m.
New subject: Tramp modified-p behaviour
Mike Kupfer <mike.kupfer(a)xemacs.org> writes:
MH-E, too, uses a loaddefs file to "hide" internal
interfaces. It's
something I'll need to deal with for MH-E 8.3, but I haven't gotten to
it yet. My tentative plan is this:
- include upstream's Makefile in the XEmacs package tree (probably call
it Makefile.upstream)
- change xemacs-packages/mh-e/Makefile to include a rule for
lisp/mh-loaddefs.el. The rule will do "make -f Makefile.upstream
loaddefs".
- tweak makefiles as needed
Nice idea. I will check, how does it work for Tramp.
mike
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:08 p.m.
New subject: Tramp modified-p behaviour
Michael Albinus <michael.albinus(a)gmx.de> writes:
> MH-E, too, uses a loaddefs file to "hide" internal
interfaces. It's
> something I'll need to deal with for MH-E 8.3, but I haven't gotten to
> it yet. My tentative plan is this:
>
> - include upstream's Makefile in the XEmacs package tree (probably call
> it Makefile.upstream)
> - change xemacs-packages/mh-e/Makefile to include a rule for
> lisp/mh-loaddefs.el. The rule will do "make -f Makefile.upstream
> loaddefs".
> - tweak makefiles as needed
Nice idea. I will check, how does it work for Tramp.
I've committed respective changes to the package repository.
> mike
Best regards, Michael.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4099
days inactive
4112
days old
27 comments
5 participants
participants (5)
-
Ailin Nemui -
Michael Albinus -
Mike Kupfer -
Stephen J. Turnbull -
Vin Shelton