Hello,
just upgraded to version 2.2.12 and suddenly getting this error in
messages buffer:
byte-code: `/bin/sh -l -c 'echo f3741e22b8b80b555d0f8445582a1250
\"$PATH\"'' returns with error
And the explanation is
$ /bin/sh -l -c 'echo f3741e22b8b80b555d0f8445582a1250 \"$PATH\"'
-l: bad option(s)
> And the explanation is
> $ /bin/sh -l -c 'echo f3741e22b8b80b555d0f8445582a1250 \"$PATH\"'
> -l: bad option(s)
>
> (remote host is an AIX host BTW)
>
This is coming from "tramp-get-remote-path" of "tramp-sh.el"
AIX (and many others) "/bin/sh" doesn't supports "-l" option.
>> AIX (and many others) "/bin/sh" doesn't supports "-l" option.
>
> Thanks for the bug report.
>
> I doubt "*many* others". This patch is already part of Tramp 2.2.11, and
> it is also synchronized with Emacs' git trunk since last December. No
> fault report so far.
By *many* I mean all Unixes wh
Emacs 26.0.91 (pretest version)
Tramp 2.3.3.26.1 (bundled with Emacs 26)
I now have to explicitly a method in remote path names:
"/:@".
"/@" fails with this kind of message:
Setting current directory: No such file or directory, /@:/
--
Gilles
___
Tr
> Yes, this is intended. Read etc/NEWS of Emacs 26.1:
Thanks for pointing me in the right direction,
BTW I think that the setting of "comint-file-name-prefix" in rlogin.el
have to be fixed accordingly:
replace:
(setq comint-file-name-prefix
(concat "/" rlogin-
> This was already done:
I've compiled Emacs using archive emacs-26.0.91.tar.xz from
alpha.gnu.org, dated 22 Jan 2018, the rlogin.el included don't include
those fixes.
___
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.org/mailman/listi
To make tramp works on solaris I think it is needed to backslash the "^"
character in the sed commend after the "uname" command, because solaris's
/bin/sh interprets unquoted "^" as a pipe.
*** tramp.elWed Sep 6 15:45:13 2006
--- tramp.el.oriSun Sep 3 14:12:18 2006
***
> I'm not confident whether this could have side effects on other
> platforms. What about quoting the expression (does it work for you?):
It works, but I'm afraid that in this version, the sed command has non
effect on the "uname" output:
1st version gives:
"SunOS 5.9"
2nd one give :
SunOS 5.9
Wagemans, Peter wrote:
>
>
>> vec "(uname -sr) 2>/dev/null | sed -e 's/^\\|$/\"/g'")))
> Doesn't work. It seems that there are sed versions that don't
> understand the regular expression.
>
>
You're right, by saying "it works", I only wanted to say that the sed
command is now correctly pa
Nelson E. Ingersoll-2 wrote:
>
> Fails with OpenVMS FTP file open
> --text follows this line--
>
>
A long time ago I've written the following lisp code attached to this
message to overcome this problem. The only drawback is that you have to use
"!" character instead of ":" in the drive part
Nelson E. Ingersoll-2 wrote:
>
>Thank you for the modification. I assume that the mod is to be used
> instead of "removing" Tramp. To allow Tramp to handle VMS filenames when
> in
> FTP mode. Right?
>
Exactly. It is a kind of "patch" to the ange-ftp code that tramps calls
behind the sce
To
make tilde expansion work it has been necessary for me to modify the
"tramp-handle-expand-file-name" function like this
Replaced:
(setq uname
(with-connection-property v "home-directory"
(tramp-send-command v (format "cd %s; pwd" uname))
(with-current-buffer (tramp-get-buff
Using GNU Emacs 22.1.1 (sparc-sun-solaris2.9, X toolkit)
Using tramp version 2.1.10
I'm pretty sure this worked with older versions:
Inside a dired buffer of a remote directory over tramp, using any function
on a file (like chmod (M), chgrp (G), etc) gives:
apply: Symbol's function definition
Very strange things happening with my tramp/emacs combo...
1st thing: "list-load-path-shadows" seem ok (tramp 1.10, installed in
site-lisp hiding tramp bundled with emacs):
/opt/emacs-22.1/share/emacs/site-lisp/trampver hides
/opt/emacs-22.1/share/emacs/22.1/lisp/net/trampver
/opt/emacs-22.1/sh
Please forget the "Wrong type argument: listp, ê" thing, I've just fixed it
(FYI, tramp-persistency-file-name had been replaced inadvertently by a
directory)
Unfortunately, the "Symbol's function definition is void:
tramp-handle-call-process" error is still here
--
View this message in c
> tramp-handle-call-process does exist in Tramp 2.1.10.
This next piece of code is from tramp-util.el extracted my local
tramp-2.1.10.tar.gz archive , there is a reference to
tramp-handle-call-process, is this normal?
(defadvice call-process
(around tramp-advice-call-process activate)
Thanks for your help, it worked.
--
Gilles
--
View this message in context:
http://www.nabble.com/chmod%2C-chgrp%2C-etc...-inside-dired-over-tramp-tf4373216.html#a12492379
Sent from the Gnu - Tramp - Dev mailing list archive at Nabble.com.
___
Tram
Hello again,
I would like to signal a small problem when executing remote commands over
tramp (i.e. "shell-command" when "default-directory" matches
"tramp-file-name-regexp"): shell output buffer stays hidden and it contents
are not erased
--
View this message in context:
http://www.nabble.com
Ok, no problem, it was just in case you didn't noticed this issue
Michael Albinus wrote:
>
>
> I guess you are using Tramp 2.1.10, aren't you? Then it looks like it
> has been fixed already. Changelog from August 7th:
>
> * tramp.el (tramp-handle-shell-command): Handle OUTPUT-BUFFER and
> ERR
I have a problem using configure then make: ("*** No rule to make target
`tramp.elc', needed by `all'. Stop.")
typescript of the session:
$ ./configure --prefix /opt/emacs --datadir=/opt/emacs/share/emacs
--with-contrib --infodir=/opt/emacs/share/info
configure: Tramp 2.1.11
checking for gm
I have a problem using configure then make: ("*** No rule to make
target `tramp.elc', needed by `all'. Stop.")
Please forgot my previous message, tar extraction had failed, tramp.el
was missing
*Sorry* for this useless post///
--
Gilles Pion
Si vous n'ete
To be more precise, the emacs grep command over a tramp buffer doesn't
activate a process on the remote host but on the local host instead. It is
the same with the "compile" command
It used to with previous tramp version. (could be a side-effect on my
configuration/environment, not sure...)
--
G
It seems that the grep command is run on the localhost and in the default
home directory:
I do alt-x grep, then I replace grep the command in the minibuffer with
"hostname; pwd"; the *grep* buffers pops out and show this
--
Gilles
--
View this message in context:
http://www.nabble.com/grep-%28a
Michael Albinus wrote:
>
> Unfortunately, an error has been introduced when I've prepared
> tramp-compat.el. Please try the appended patch; it shall fix it.
>
>
It does, thanks!
--
View this message in context:
http://www.nabble.com/grep-%28and-compile%29-over-tramp-seem-to-bee-broken-in--2
Sorry to jump into this thread, but I have the same problem and is not solved
for me
I've downloaded the latest emacs CVS snapshot (23.0.50.1),
I'm using tramp 2.1.11
Files accessed with /scpc2:root@:/path/to/file still revert to
root owner when saved
Sascha Wilde-3 wrote:
>
> Michael Albinu
running, not root as
expected!!!
Gilles Pion wrote:
>
> Sorry to jump into this thread, but I have the same problem and is not
> solved for me
>
> I've downloaded the latest emacs CVS snapshot (23.0.50.1),
> I'm using tramp 2.1.11
>
>
--
Gilles
--
View this messa
Michael Albinus wrote:
>
> Gilles Pion <[EMAIL PROTECTED]> writes:
>
>> I've downloaded the latest emacs CVS snapshot (23.0.50.1),
>
> What do you mean by this? An emacs-snapshot package as offered for
> example for Debian/Ubuntu, or just the files taken
Michael Albinus wrote:
>
>> In fact *very* strange things are happening regarding file ownership when
>> accessed via tramp:
>
> If this still happens with very recent Tramp 2.1.12-pre, please
> provide traces with tramp-verbose of 8. Best would be a message
> generated via `tramp-bug', includi
[tramp version 2.1.12]
I think there is somewhere a too much optimistic assumption regarding short
tracks taken when tramp-local-host-p is true
Probably around "tramp-handle-insert-file-contents" or in one of the
functions it calls
Evidence:
By using "/sudo:@localhost:/" I get en empty buffer
Michael Albinus wrote:
>
>
> Looks like /tmp/U501680 hasn't enough permissions. Is it your local
> temp directory? Does it have write permissions for the world?
>
>
Yes: I'm using a private TMP directory, chmoded to 700, for security
purposes (emacs variable "temporary-file-directory")
Mic
Michael Albinus wrote:
>
> OK, here is a real fix: could you, please, test?
>
Well done!! It works
-
Gilles
--
View this message in context:
http://www.nabble.com/-sudo%3A%3Cuser%3E%40172.0.0.1%3A-%3Cfile%3E-works%2C--sudo%3A%3Cuser%3E%40localhost%3A-%3Cfile%3E-doesn%27t-tp15019015p15038
Hello,
To make methods pscp and psftp works under win32 I had to make the following
fix:
replaced in each of the two methods:
(tramp-copy-args(("-P") ("%p") ("-sftp") ("-p" "%k")))
by
(tramp-copy-args(("-P" "%p") ("-sftp") ("-p" "%k")))
(using
I have a problem with last tramp version (2.1.16) when editing files whose
"tramp-local-host-p" predicate is true (like "/@localhost:/ has changed since visited or saved. Save anyway? (y or n) "
I suspect a problem around the shortcut taken inside tramp-handle-write-region
around the comment
"Sh
Michael Albinus gmx.de> writes:
>
> Have you checked the latest version from Tramp's CVS? Lasat days I have
> applied some changes in tramp-do-copy-or-rename-file-directly, which
> could be also good for your problem.
Just fetched the last CVS and the problem is still there:
Steps to reproduce:
Hello,
I have a problem with double parenthesis without embedded space inside remote
command "tramp-do-file-attributes-with-stat"
To be able to use tramp on an exotic remote environment (Opsware "global shell")
I've been forced to create a specific tramp method using /bin/ksh for
"tramp-remote-sh
ives:
ange-ftp-run-real-handler is an alias for `tramp-run-real-handler',
which is not defined. Please make a bug report.
Tramp-version: 2.2.6-24.3
GNU Emacs 24.3.1 (i686-pc-cygwin)
--
Gilles Pion
___
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/tramp-devel
==> note the inalid alignement of "f2" line
Not sure if this really related to tramp, but it's occurs only in dired/tramp
buffers (ange-ftp through tramp ones are ok)
--
Gilles Pion
___
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/tramp-devel
is not really a big deal ...
--
Gilles Pion
___
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/tramp-devel
tramp debug buffer contents
https://www.box.com/s/joyrwdjd6h9lt7wp0z81
--
Gilles Pion
___
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/tramp-devel
Ref: <5ksqr81a2f4fceu0hfr1dgrct0j6qo3...@4ax.com> de Gilles Pion
>Hello,
>
>I'm having the same problem as the one described here:
>
>http://lists.gnu.org/archive/html/tramp-devel/2010-10/msg00010.html
>
>Tramp-version: 2.2.6-24.3
>Emacs-version GNU Emacs 24.
Ref: <972rr81hrmgc4fnkgkms9rdn4luih7s...@4ax.com> de Gilles Pion
>
>It seem there is a race condition inside "tramp-send-command-and-check"
>
>Introducing a short delay (worked with "(sleep-for 0 100)") seem to fix the
>problem:
>
> (with-
Ref: <878v29ksoq@gmx.de> de Michael Albinus
>What happens on your NAS, if you suppress the busybox blacklisting in
>tramp-sh.el, like this (line numbers might vary):
The problem completely disappears as soon I've replaced the lines according to
your patch.
Reproducible at
emote-shell" "sh"))
But evaluating that forms fails with the following error:
"Symbol's value as variable is void: tramp-connection-properties"
What am I doing wrong?
--
Gilles Pion
___
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/tramp-devel
2013/8/12 Michael Albinus
> `tramp-connection-properties' has been added in Tramp 2.2.7. Which
> version do you run?
>
I was using 2.2.6, after having upgraded it's now ok
--
Gilles
___
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.o
44 matches
Mail list logo