revision 1.152
date: 1999/10/06 14:25:01; author: grossjoh; state: Exp; lines: +6 -4
rcp-handle-make-directory: Check for mkdir errors.
rcp-set-auto-save: Don't turn on auto-save if buffer file name is nil.
ftp://ls6-ftp.cs.uni-dortmund.de/pub/src/emacs/rcp.el
kai
--
Life is hard and then
The good news is: I can reproduce the problem with the scp method not
working. Opening a non-existing file works fine, but after typing a
few letters, saving it doesn't work.
The code in question is in rcp-handle-write-region:
(save-excursion (set-buffer rcpbuf) (erase-buffer))
Francesco Potorti` [EMAIL PROTECTED] writes:
I always write /r:fly:todo and it works already. What's the matter?
You could have abbrevs a for user1@host1 and b for user2@host2. Maybe
even for specific files on the remote host?
kai
--
I like BOTH kinds of music.
Thanks to Daniel Pittman, this version probably works better with
XEmacs and EFS. This means that the comment that you have to set
file-name-handler-alist to nil before loading rcp.el is obsolete.
Those of you using XEmacs and EFS, could you please check if this
version does it? If it doesn't
I have now applied this patch. Will release a new version after
applying more patches...
kai
--
Life is hard and then you die.
,-
| revision 1.174
| date: 1999/10/23 15:05:24; author: kai; state: Exp; lines: +43 -1
| New function rcp-open-connection-nrlogin. This one starts an
| interactive shell using the rlogin program. It groks both password
| prompts and password-less logins.
`-
I'm going to redo the
Daniel Pittman [EMAIL PROTECTED] writes:
This patch will break things unless you also apply my patch from message
[EMAIL PROTECTED] - I screwed up in the function
'rcp-handle-file-name-directory' and call 'rcp-make-rcp-file-name' with
the user and host parameters swapped. :(
Yes. I
Kai Großjohann [EMAIL PROTECTED] writes:
Maybe a workable workaround is to send "echo hello", then wait a few
seconds for output, then send another newline. Hm. Lemme think.
I've got an idea. David, could you go to rcp-open-connection-rsh and
spot the gross hack and increase
David Barr [EMAIL PROTECTED] writes:
[...] but my directory is not in the standard search path, and the
way that rcp.el starts "ksh" (which it needs in order to grok ~
characters) my .profile doesn't get sourced, so I can't change my
PATH to include the mimencode program. Any ideas?
Ah,
I tried to do a few bug fixes since 1.143. RCS log enclosed.
Everybody: I think that it might be possible that this perhaps may
even work with ssh2... I tried the `sm2' method. What happens with
other methods that explicitly refer to ssh2?
Matthias: does rcp-open-connection-rlogin now work?
Daniel Pittman [EMAIL PROTECTED] writes:
It already does EFS (or ange-ftp) file name tracking by setting
'comint-file-name-prefix' to the right user/machine syntax.
Sadly, this is done *after* any user-definable hooks are run. If he
wants, Francesco could
'M-x eval-expression (setq
Harry Putnam [EMAIL PROTECTED] writes:
The machine is about 1800 Miles away and I never have a really good
connection to it but still, after several experiments it seems ange-ftp
is consistently faster by something like a factor or 2.
Well, maybe it is the encryption used by ssh. If the
Joe Stoy [EMAIL PROTECTED] writes:
... appears not to be used at all.
You're right! Thank you very much, I have now put a new version on
the server with this fix in it.
kai
--
This gubblick contains many nonsklarkish English flutzpahs,
but the overall pluggandisp can be glorked from
Gerd Bavendiek [EMAIL PROTECTED] writes:
debug rcp/scp root@gigi
---
$ stty -onlcr -echo
$ set +o history
[[INCOMPLETE!]]#
/
rcp.el was waiting for a specific output from the shell; in this case
the shell prompt I think. The `incomplete' means that rcp.el
Thanks to Paul Stevenson [EMAIL PROTECTED] for this feature
wish.
ftp://ls6-ftp.cs.uni-dortmund.de/pub/src/emacs/rcp.el
kai
--
Life is hard and then you die.
Gerd Bavendiek [EMAIL PROTECTED] writes:
Please tell me which detailed info is useful.
Set rcp-debug-buffer to t, set debug-on-error to t, repeat error, send
contents of *Backtrace* buffer, of *rcp/debug...* buffer, of *rcp...*
buffer.
Thanks.
kai
--
A large number of young women don't trust
Daniel Pittman [EMAIL PROTECTED] writes:
I think you are right. It should definitely be preserving the newline in
the argument.
I have already submitted this as a bug report, and RMS has replied
with the (very sensible) question: What should shell-quote-argument
_do_ in order to preserve the
Pete Forman [EMAIL PROTECTED] writes:
The current version of rcp.el is not working on XEmacs 20.x or 21.1.x
owing to a problem in XEmacs. I raised a problem report a couple of
weeks ago. Rcp is meant to wait about 30s for a connection to be
established but a fault in the itimers means that
"Tom Roche" [EMAIL PROTECTED] writes:
That would be great, except it'd also be nice to post a .info file for
those of us without the conversion utility. (I gotta setup Cygwin,
after the semester gets going), and HTML for the rest of the world.
And perhaps you should setup a CVS server
"DE WEERD, Mario" [EMAIL PROTECTED] writes:
I think that there should be some more verification of a succesfull
write. When something goes wrong, I sometimes end up with a file
with nothing in it. This is due to the fact that the 'uuencode'
could not be executed for example.
I have now
Tim Carroll has found and fixed a bug where "unset PS1 PS2 PS3" didn't
work in all shells. rcp.el now uses "PS1=''; PS2=''; PS3=''" instead.
Does it work?
ftp://ls6-ftp.cs.uni-dortmund.de/pub/src/emacs/rcp.el
Since it is so short: patch attached.
kai
--
I like BOTH kinds of music.
---
DE WEERD, Mario [EMAIL PROTECTED] writes:
I have created a method that allows me to transfer files under a
compressed format.
Unfortunately the '~'-expansion does not work when calling these
commands and I am obliged to change the '/bin/sh' in the 'rcp.el' file
by '/bin/ksh'.
Then it
Charlie Zender [EMAIL PROTECTED] writes:
Here is my process table, showing a ksh started by rcp weeks ago
which is never being used for anything:
UIDPID PPID CSTIME TTY TIME CMD
zender 97640 97658 0 Dec 03 ? 0:00 tcsh -c /bin/sh
zender
"DE WEERD, Mario" [EMAIL PROTECTED] writes:
I still have problems with my "~user/bin/decode" references since the ~
does not get expanded on my /bin/sh systems.
Hm. rcp.el tries to find a shell which groks tilde expansion, so I'm
not sure where the problem might be. Hm.
After the
Greg Stark [EMAIL PROTECTED] writes:
[EMAIL PROTECTED] (Kai Großjohann) writes:
Note the strange `D'. Both of the above were on a FreeBSD system.
Uh, is this repeatable?
Yes, it happens every time that I try this.
Report it as a *serious* bug to the FreeBSD team?
Ok, will do
"DE WEERD, Mario" [EMAIL PROTECTED] writes:
I think that the local command is executed using the /bin/sh.
The /bin/sh does not grok ~ expansion, so ... .
Indeed. Hm. Have you got a suggestion what to do about this? It
does not work to use /bin/ksh or /usr/bin/ksh on all systems, as this
"DE WEERD, Mario" [EMAIL PROTECTED] writes:
I may have found something.
Consider the following:
An argument can be grouped if quoted simply by using "" or ''.
In order to test, I have copied the rcp-shell-quote-argument function
and I modified it.
In case the argument has got newlines, I
"DE WEERD, Mario" [EMAIL PROTECTED] writes:
Why not do the same thing for the local shell as for the remote shell.
Look (only once) which shell works for ~ expansion and put this in a
variable.
Hm. Yes, that might be an approach. I'll put it on the todo list.
kai
--
A preposition is not
Could you please M-x load-library RET rcp.el RET (`.el' is important),
then M-x toggle-debug-on-error RET, then repeat the error and mail me
the backtrace? Thanks,
kai
--
A preposition is not a good thing to end a sentence with.
"Tom Roche" [EMAIL PROTECTED] writes:
Kai Großjohann [EMAIL PROTECTED] 1/9/00 7:13:48 AM
You say that you _want_ to use scp, but that does not mean that you
_need_ to use scp. Suppose you type some file name, and Emacs opens
a connection to the remote host using ssh, th
Dale R Worley [EMAIL PROTECTED] writes:
Well, you should be able to depend on (emacs-version) over the long
haul. But definitely, you should be able to reduce it to a small
experssion that can be clipped out if the user has an Emacs-19 that he
wants to test it with.
To reduce what to a
Some of you had problems -- rcp.el would say things like `couldn't set
remote shell prompt' or `remote shell didn't come up' or something
like this.
I have now done work in this area -- could you please check whether it
still fails?
Thanks a lot,
kai
--
Beware of flying birch trees.
Dale R Worley [EMAIL PROTECTED] writes:
I'm runing Red Hat Linux 6.1. /bin/sh is an alias for /bin/bash.
Bash has the property that if it thinks it's non-interactive, it won't
print a prompt, no matter what you set PS? to.
You can now frob the method parameter rcp-remote-sh. For example,
Dale R Worley [EMAIL PROTECTED] writes:
And I notice that while the 'stty ...' commands are logged into the
debug buffer, the initial 'exec /bin/sh' is *not*. That could make
some problems hard to debug.
I hope that command is now logged, too.
kai
--
Beware of flying birch trees.
Dale R Worley [EMAIL PROTECTED] writes:
Seeing as how I'm reading this through a Linux console (24x80,
monochrome, monofont), I'd have to say "No." Let me mock up something
for you...
Well, considering that it was meant for me to read, as in `could you
turn on debugging and send me the
Gerd Bavendiek [EMAIL PROTECTED] writes:
Symbol's function definition is void: process-coding-system
On subsequent tries xemacs simply hangs until I hit CTRL-g.
When setting up the shell buffer, rcp.el looks to see if there are CR
characters in the buffer. If there are any, it assumes that
[EMAIL PROTECTED] (Kai Großjohann) writes:
Thanks a lot, I included this in the distribution. Have you signed
papers with the FSF for assigning the copyright to them?
I have been assured that papers aren't necessary in this case. So
just disregard that question...
kai
--
Beware of flying
[EMAIL PROTECTED] (Mark A. Hershberger) writes:
I'm using a chained connection (multiu-ssh-telnet) and have ffap
loaded so that it is used for `C-x C-f' and not for `C-u C-x C-f'.
When point is not sitting on a blank line and I type `C-x C-f filename
RET', I get back the following error:
The last time we discussed using URLs as file names, one of you
pointed out how broken the `ftp:' URL syntax is -- people have to type
%3F (or what was it) to open a file in the root directory (rather than
relative to the home directory).
This put me off quite a bit, so I almost decided not to
Daniel Pittman [EMAIL PROTECTED] writes:
For quite a while now I have had a real problem with the performance of
inline methods while using rcp.el. This finally got to me enough to go
and do some testing and all.
The problem seems to be that when using a pty connection to an
asynchronous
Hello XEmacs or NT users,
are you doing something special to make rcp.el work, such as replacing
`:' with `;' in the file name syntax, or is rcp.el now magically
working for XEmacs, too?
I think I ought to try it again on XEmacs. Hm.
kai
--
Beware of flying birch trees.
[EMAIL PROTECTED] (Mark A. Hershberger) writes:
Is v. 1.292 supposed to have the multiu fix?
No, I added it in 1.295.
kai
--
Beware of flying birch trees.
Francisco Solsona [EMAIL PROTECTED] writes:
To be honest, right now I don't have time to try other methods (ssh,
etc.), but rcp works out of the box for me, using the above
configuration.
It feels good to hear this. I had begun to think that it wasn't
possible at all to use rcp.el on XEmacs
[EMAIL PROTECTED] (Mark A. Hershberger) writes:
Signaling: (wrong-type-argument stringp t)
file-name-absolute-p(t)
file-truename("/r@scp:[EMAIL PROTECTED]:/usr/local" (95)
((("/r@scp:[EMAIL PROTECTED]:/usr/" . "/r@scp:[EMAIL PROTECTED]:/usr/"
It turned out that
So, I have probably alienated a number of you by either asking you
whether the old bug still persists, or by telling you that I wouldn't
bother with that behavior again.
And the whole thing only got me down to 89 ticked messages from 100 or
so. Sigh.
kai
--
Beware of flying birch trees.
[EMAIL PROTECTED] (Mark A. Hershberger) writes:
Lucky me I follow the mailing list.
Whee. Oops.
Tonight I tried to edit a file using rcp.el:
/usr/local/apache/conf/httpd.conf. Except we didn't partition the
disk right and so have kludged /usr/local to be a symlink to
/home/local. So,
Yuji Yamano [EMAIL PROTECTED] writes:
- Need texinfo 4.0 or later for making the manual.
- If you would like to make the manual with old texinfo,
replace @command with @code first.
I have now documented it. Also, I'll try to include the info file and
the DVI file in the tar ball from now
[EMAIL PROTECTED] (Mark A. Hershberger) writes:
The sudo invocation should look like `sudo -u user prog', but
evidently it doesn't.
For getting sudo to really work right, I still need to do the changes
which remove the need for `cd'. It might work to use `/bin/sh' as the
`command', though.
[EMAIL PROTECTED] (Mark A. Hershberger) writes:
How do I do the equivilent of \000 in a emacs regexp? `\000', `\x00',
etc don't seem to work.
It appears to be possible to use the `-p' option for sudo to change
the password prompt.
You can type C-q C-@ to insert a literal ^@ character into a
Pete Forman [EMAIL PROTECTED] writes:
I've since narrowed it down to itimer, or perhaps the XEmacs core.
I can send a test case to anyone that is interested. Roughly
speaking, a function that is called in a timeout sets a variable. At
the end of that function the variable has the correct
"Daniel Pittman" [EMAIL PROTECTED] writes:
My guess is that the coding system isn't right in the buffer.
But I am trying to frob it! Why does that not work? Could some
XEmacs expert have a look at the relevant section in rcp.el to find
out what might be wrong?
The string to search for is
I have just tried it manually on the following system:
/
| grossjoh@zx3 uname -a
| AIX zx3 3 4 0040529A4C00
\
No problem with embedded newlines. But I'll try the connection, too.
kai
--
Beware of flying birch trees.
I am seeing _very_ strange behavior from AIX /bin/sh:
/
| grossjoh@zx3 uname -a
| AIX zx3 3 4 0040529A4C00
| grossjoh@zx3 /bin/sh
| u@h set +o history
| /bin/sh: history: 0403-010 A specified flag is not valid for this command.
| u@h echo foo
| foo
| u@h exit
| grossjoh@zx3 /bin/sh
| u@h set
Could you also say M-x describe-coding-system RET RET while in the
buffer for the connection?
kai
--
Beware of flying birch trees.
Pete says that colons in rcp-file-format don't work for him, whereas
Francisco and Daniel say they work fine. What can we do to find out
why the colons don't work for Pete?
The XEmacs versions used:
PeteXEmacs 21.1.7
Fransisco 21.1 (patch 8) "Bryce Canyon" XEmacs Lucid
Daniel
Pete Forman [EMAIL PROTECTED] writes:
I'm also getting "Symbol's function definition is void: process-
coding-system". Is this fixed yet?
Yes, it checks if the function is bound and uses `stty -onlcr' if it
isn't.
kai
--
Beware of flying birch trees.
DI Maximilian Renkin [EMAIL PROTECTED] writes:
mxrenkin@kastanie:~ # Initializing remote shell
$ exec /bin/sh
# Waiting 30s for remote `/bin/sh' to come up...
exec /bin/sh^M
mxrenkin@kastanie:~ # Setting up remote shell environment
exec /bin/sh^M
mxrenkin@kastanie:~ stty -echo^M
#
Okay. I'm quite confused now. It appears that rcp.el is correctly
frobbing the coding system -- before frobbing the coding system, ^M
appeared in the buffer, and after frobbing the coding system, it was
foo-dos.
Hm. You may be seeing the timeout bug. I think I should be doing
something about
Pete Forman [EMAIL PROTECTED] writes:
(if (fboundp 'efs-file-handler-function) ; test for a standard autoload
(require 'efs))
How about putting that into rcp.el?
Sounds good. Yes, this is the right thing to do. In a moment...
done.
If one of you XEmacs users would like to test this
Pete Forman [EMAIL PROTECTED] writes:
So none of these profiles grok tilde expansion but not test -nt. That
means that you should be reasonably confident that rcp-handle-file-
newer-than-file-p will not barf.
Good. For good measure, I have documented the whole issue in
"Stefan Monnier" [EMAIL PROTECTED] writes:
The two bugs below: [...] are probably "obsolete" now that
set-file-modes is implemented.
Right. I have removed the todo items.
PS: oh, and file operations that are not implemented yet but whose
default version is obviously wrong should not use
"Daniel Pittman" [EMAIL PROTECTED] writes:
My guess is that the coding system isn't right in the buffer.
Could it be that this is caused by the timeout bug? In other words,
has the bug gone away with the recent timeout fix?
Max, could you please try this, too?
Many thanks,
kai
--
Beware of
Thanks, done.
kai
--
Beware of flying birch trees.
"Edward J. Sabol" [EMAIL PROTECTED] writes:
However, it still doesn't work on Digital Unix 4.0. :-(
Hm. I have now made rcp.el check if `test FOO -nt BAR' works. If it
doesn't, rcp-handle-file-newer-than-file-p barfs.
Is this okay? Maybe I should try the `find' thing on systems where
"Stefan Monnier" [EMAIL PROTECTED] writes:
PS: I wish someone had time to re-implement rcp-handle-file-attributes
using Perl rather than ls, so as to get mtime and friends.
I'll put the following on the todo list:
* Try to find a perl and set a local variable.
* Use perl for various
Joe Stoy [EMAIL PROTECTED] writes:
To fix the simple cases, in rcp-make-rcp-file-name change the line that
says
(cons ?p path)
to
(cons ?p (or path ""))
but I haven't worked out the corresponding change (if any) to the
multi-hop
New version. Tries to work around the XEmacs with-timeout bug. (You
still need with-timeout for XEmacs 20, or the fsf-compat package for
XEmacs 21!)
ftp://ls6-ftp.cs.uni-dortmund.de/pub/src/emacs/rcp.tar.gz
I have stopped distributing the rcp.el file alone, it is causing too
much trouble.
Pete Forman [EMAIL PROTECTED] writes:
Mea culpa, but a comment by the require statement would have helped.
Thanks. The comment will show up whenever I generate the next tar
ball. It's in CVS already.
kai
--
Beware of flying birch trees.
Pete Forman [EMAIL PROTECTED] writes:
Wrong number of arguments: format, 0, "Couldn't find remote `%s' prompt."
Starting remote shell `/bin/ksh' for tilde expansion...
(The first indicates a minor bug in the error message code.)
Thanks, I fixed it.
The debug buffer is
# Opening
"Edward J. Sabol" [EMAIL PROTECTED] writes:
However, it still doesn't work on Digital Unix 4.0. :-(
How would I find that out? I did a quick test and it appears that
maybe the following mechanism works:
test / -nt / 2/dev/null ; echo $?
The above prints 1 if `-nt' is known, and a different
[EMAIL PROTECTED] (Lothar Brendel) writes:
LB The code in rcp-time-diff "snarfed from Emacs 21 time-date.el"
LB sucks. It produces a list (of two integers) which is returned
LB and compared to an integer in e.g. rcp-wait-for-regexp. Since we just need
LB the difference expressed in seconds,
"Edward J. Sabol" [EMAIL PROTECTED] writes:
[...] I guess we don't really have to worry about '-nt' working
after all.
Damn, and I already did the work to check it :-) But at least I can
skip the work of trying to do it with `find'.
So, if anybody finds that `test foo -nt bar' isn't
"Edward J. Sabol" [EMAIL PROTECTED] writes:
rcp.el Log Message: Michael Sternberg [EMAIL PROTECTED]:
send `mesg n'
I wonder if we might want to send "biff n" as well. Anyone?
Sounds like a good idea. I'll include it.
kai
--
Beware of flying birch trees.
Pete Forman [EMAIL PROTECTED] writes:
The efs.el that ships with XEmacs 21 no longer overloads expand-file-
name. It does still overload load and so things like load-file still
fail.
Hm. I had a cursory glance at the stuff that EFS (from XEmacs
21.1.10) is doing there, and I don't think I
[EMAIL PROTECTED] (Kai Großjohann) writes:
I tried to fix it at the root rather than cure the symptoms.
I might add that I'm not at all sure whether my fix is better than
yours! I think the above sentence might have sounded a bit, err,
inappropriate.
kai
--
Beware of flying birch trees.
[EMAIL PROTECTED] (Kai Großjohann) writes:
Could somebody with XEmacs please test if that workaround works around
the bug in XEmacs? The new version is in CVS. I haven't tested it.
I have now tested the workaround on XEmacs and Emacs, found a bug with
the workaround, fixed the bug, and now
Pete Forman [EMAIL PROTECTED] writes:
The efs.el that ships with XEmacs 21 no longer overloads expand-file-
name. It does still overload load and so things like load-file still
fail.
It is normal for load-file to fail, since the `load' file name handler
operation isn't implemented yet.
Could somebody with XEmacs please test if that workaround works around
the bug in XEmacs? The new version is in CVS. I haven't tested it.
Will come back in a few hours to try it some more.
kai
--
Beware of flying birch trees.
Pete Forman [EMAIL PROTECTED] writes:
It fails on most of these if you force the use of the vanilla Bourne
shell, typically /bin/sh. They all have ksh which works all the time
and should have been found by rcp.el.
Right now, rcp.el checks to see if the Bourne shell groks tilde
expansion,
DI Maximilian Renkin [EMAIL PROTECTED] writes:
Using SuSE 6.4, Gnu Emacs 20.6.1, rcp-1.329 I get another error now. The
Backtrace is:
=
Signaling: (wrong-type-argument number-or-marker-p (0 0))
rcp-wait-for-regexp(#process *rcp/su
Err. Oops. I must have been _very_ tired.
Will try the right mail address now.
kai
--
Beware of flying birch trees.
SSIA. Bin nicht so glücklich damit, kann mich aber im Moment nicht
besser konzentrieren.
Stefan, tust du deinen Teil direkt via CVS da rein?
kai
--
Beware of flying birch trees.
carmen-short.ps
Joe Stoy [EMAIL PROTECTED] writes:
So, anyway, I decided to try your latest version (via CVS), and ran
into another problem. One of my server systems (an NT one, you'll
perhaps not be surprised to hear)
You mean that rcp.el (kinda) works when connecting to an NT system?
Amazing! (Of
Pete Forman [EMAIL PROTECTED] writes:
We should not be looking at return codes to assess test's abilities.
I have now changed rcp.el to send `test / -nt /'. Empty output means
that this operation is supported, anything else means not.
Does this work?
kai
--
Beware of flying birch trees.
DI Maximilian Renkin [EMAIL PROTECTED] writes:
Signaling: (error "Cannot subtract two times")
Which version of Emacs have you got? On XEmacs, it might help to
(require 'itimer).
kai
--
Beware of flying birch trees.
"Stefan Monnier" [EMAIL PROTECTED] writes:
The Irix machines I have access to are really badly setup, but
their `test' does not accept -nt.
Hm. Note that rcp.el might have started ksh on Irix, if their /bin/sh
doesn't grok tilde expansion. Does Irix have ksh?
I also thought that the AIX
Francesco Potorti` [EMAIL PROTECTED] writes:
test / -nt / 2/dev/null ; echo $?
I think not.
Why do you think that this is a problem? Hm.
kai
--
Beware of flying birch trees.
Thanks to Lothar who points out that rcp.el transfers 10k to open an
empty file. That's quite a bit of stuff. I had never thought it
would be this much.
So, there seems to be a need for caching. But what data should we be
caching, and for how long?
One idea which crossed my mind is to use
DI Maximilian Renkin [EMAIL PROTECTED] writes:
mxrenkin@kastanie:~ # Initializing remote shell
$ exec /bin/sh
# Waiting 30s for remote `/bin/sh' to come up...
exec /bin/sh^M
mxrenkin@kastanie:~ # Setting up remote shell environment
exec /bin/sh^M
mxrenkin@kastanie:~ stty -echo^M
Matt Swift [EMAIL PROTECTED] writes:
Using rsync method and dired, I'm getting an error when I visit a file
with RET:
Signaling: (wrong-type-argument stringp nil)
file-symlink-p(nil)
Strange. Hm. I can't explain where that might come from.
Hm. vc-find-file-hook uses
[EMAIL PROTECTED] writes:
Ah. That must be it! Its using emacs-mule-dos. Note that if I do a M-x
shell and then ssh into the remote server manually it instead uses
emacs-mule-unix for input to process I/O.
Okay. Good that we have found the symptom, now. But why does Emacs
set the coding
Matt Swift [EMAIL PROTECTED] writes:
Using rsync method and dired, I'm getting an error when I visit a file
with RET:
I have now been able to reproduce the error, and I think I have fixed
it now. It was a missing `save-excursion' in the new implementation
of `rcp-handle-file-directory-p' --
I have tried to make rcp-handle-file-newer-than-file-p work even on
systems where `test' does not grok `-nt'. If you have access to such
a system, please test.
Note that rcp.el might be executing ksh or bash on the remote end, and
there the `test' builtin usually groks `-nt'.
Thanks,
kai
--
Dave Love [EMAIL PROTECTED] writes:
Rubbish. Please don't peddle this story;
Sorry.
kai
--
Beware of flying birch trees.
Pete Forman [EMAIL PROTECTED] writes:
I have yet to upgrade to XEmacs 21.1.10 but I've just pulled efs.el
from cvs.xemacs.org. The overloading code that is giving me grief is
still the same.
Did you (require 'rcp) before (require 'efs)? Or vice versa? You
should be doing (require 'efs)
"Stefan Monnier" [EMAIL PROTECTED] writes:
I don't know about XEmacs, but in Emacs, `when' has been standard at
least since Emacs-20. And since I believe that CL is dumped into
XEmacs [...]
Pete?
Since you complained about it and you're using XEmacs...
kai
--
Beware of flying birch
Pete Forman [EMAIL PROTECTED] writes:
Maybe. rcp.el has stopped working as of 1.339. 1.329 does work. The
debug buffers are very similar, apart from the different ways of
setting up the remote PATH.
Signaling: (wrong-type-argument stringp nil)
signal(wrong-type-argument (stringp
Pete Forman [EMAIL PROTECTED] writes:
Also, for the mimencode stuff to work there ought to be a (require
'base64) with a suitable wrapper to ignore failure.
Hm. Hmmm... I think it might be good not to require base64 always,
since people might wish to use rcp.el without base64 encoding.
But
"Stefan Monnier" [EMAIL PROTECTED] writes:
But since (let ((file (symbol-value 'file))) does not strike me as an obvious
piece of code, it seems that a short comment is in order.
Yes, I have now documented the places in question.
kai
--
Beware of flying birch trees.
"Daniel Pittman" [EMAIL PROTECTED] writes:
Hrm. In that case, since it seems to be biting me, I might put in some
effort to learning how MULE works and implementing it properly for RCP.
Assuming you don't object, of course.
I would be very happy if you were to help with that. Thanks!
kai
101 - 200 of 527 matches
Mail list logo
