Michael Albinus writes:
Hi Kai,
>>> I see. It is not a problem of multi-hop per se, but a general problem of
>>> su(do) password prompts in Tramp. Will see what I could do.
>>
>> Ok, yes - but it's only a problem of the `sudo` method. For `su` the
&
"Kai Tetzlaff" writes:
Hi Kai,
>> I see. It is not a problem of multi-hop per se, but a general problem of
>> su(do) password prompts in Tramp. Will see what I could do.
>
> Ok, yes - but it's only a problem of the `sudo` method. For `su` the
> current behaviour is correct.
And perhaps 'doas'
"Kai Tetzlaff" writes:
Hi Kai,
>> Btw, in your example I don't understand why you want to access the
>> remote host as user "notroot", and afterwards to change the user on that
>> host to "notroot", again.
>
> That's not what I want. The `sudo::` part *should* obtain `root`
> privileges.
Kai Tetzlaff writes:
Hi Kai,
>>> with auth-source enabled when SSHing to a server with sudo as 2nd hop
>>> like `notr...@host.example.com|sudo:/etc/passwd` TRAMP apparently tries
>>> to look up the sudo password using `user=root, host=host.example.com,
>>> port=sudo` (at least that is what
[Adding tramp-devel to Cc]
Daniel Martín writes:
> Hi, Michael:
Hi Daniel,
> I have a server with a command line tool that formats code and I'd like
> to run this tool on an Emacs region via TRAMP. I tried to use
> shell-command-on-region, but it seems like it isn't compatible with
> TRAMP,
Kai Tetzlaff writes:
> Hi Michael,
Hi Kai,
> with auth-source enabled when SSHing to a server with sudo as 2nd hop
> like `notr...@host.example.com|sudo:/etc/passwd` TRAMP apparently tries
> to look up the sudo password using `user=root, host=host.example.com,
> port=sudo` (at least that is
bermel writes:
Hi,
> First, the current workaround works fine for me.
Thanks for the feedback, so I'll add it to the next Tramp release.
> Concerning applying the patch: I would need more instructions for that.
> I am a (spacemacs) user (emphasis on user) and I never got really
> familiar
Stephen Berman writes:
Hi Steve,
>> Thanks. The problem is more difficult. Tramp believes that the remote
>> uid/gid is 0 (which is the value of the root location "/"). Because of
>> this, Tramp regards file modes "drwxr-s---" of directory "/b/bermasbp/"
>> as unreadable, because this directory
Stephen Berman writes:
Hi Stephen,
> I applied the patch and rebuilt, cleared all Tramp connections,
> restarted Emacs, activated the VPN tunnel and tried to access the
> remoted directory, but got the same error. I've attached a gzipped
> verbosity level 10 trace.
Thanks. The problem is more
Stephen Berman writes:
> Hi Michael,
Hi Stephen,
> I'm unable to access a remote site via sftp using Tramp. It's the same
> site I've had Tramp problems with several time over the last few years
> -- and each time you've been able to fix the problem, so I'm hoping that
> happens now too! The
bermel writes:
Hi,
> However, I didn't mange to apply the patch. Emacs is installed via snap
> and I learned that it is not easy to modify files within snap as it
> applies a read-only file-format. So I will wait for the next version.
No problem, you can install Tramp from GNU ELPA. The recent
Michael Albinus writes:
Hi Novak,
>> I believe that the text in the echo are is misleading ("Passphrase for
>> /ssh:me@myotherhost:"). A better option might, perhaps, be that
>> original message from *tramp/ssh me@myotherhost*.
>
> The idea is, that Tra
bermel writes:
> Debug buffer:
Thanks. It still uses German locale, though:
--8<---cut here---start->8---
> 13:39:08.302097 tramp-call-process (6) # ‘gio info
> davs://xy...@webdav.smartdrive.web.de/’ nil *tramp/davs
> xy...@webdav.smartdrive.web.de*
>
bermel writes:
Hi,
> Unfortunately, with
> # env LC_ALL=en_US.UTF-8 emacs
> the issue persists (as already tried before).
OK. Please try
--8<---cut here---start->8---
# env LC_ALL=en_US.UTF-8 emacs -Q --eval '(setq tramp-verbose 6)'
bermel writes:
Hi,
> I am not sure whether it is related to the language.
I've checked further. There is an error in Tramp. Internally, it tries
to call the "gio info *" command with an English locale, but it doesn't
work as expected.
I've fixed this, and my tests for this case run
Michael Albinus writes:
Hi Kai,
> Thanks for the report! This is an obvious error, I will fix it. First I
> need to check whether similar errors happened when
> tramp-compat-string-search was added to Tramp last summer. I'll keep you
> informed.
There's nothing else
bermel writes:
> Hello,
Hi,
> a davs connection is not working. I use it for org-mobile.
>
> The command (file-directory-p davs-directory) returns nil.
>
> System: Xubuntu 20.04 Emacs 27.2. I can access the directory via
> thunar.
> Under Ubuntu 18.04 - Emacs 26.3 (but with the same emacs
Novak Boškov writes:
> Hi Michael,
Hi Novak,
> I believe that the text in the echo are is misleading ("Passphrase for
> /ssh:me@myotherhost:"). A better option might, perhaps, be that
> original message from *tramp/ssh me@myotherhost*.
The idea is, that Tramp says "password" when a password
Kai Tetzlaff writes:
Hi Kai,
> The documentation of `tramp-connection-properties` claims:
>
> List of static connection properties.
> Every entry has the form (REGEXP PROPERTY VALUE). The regexp
> matches remote file names. ...
>
> However, the actual implementation of
Novak Boškov writes:
Hi Novak,
> I'm trying to do two-hop ssh over Tramp.
>
> `/ssh:me@myhost|ssh:me@myotherhost:/`
>
> And I'm getting
>
> `Passphrase for me@myotherhost:`
>
> I don't expect to be asked any passphrases as my public key from
> `myhost` is present in
Malcolm Cook writes:
Hi Malcolm,
> I find in my environment that I must
>
> `(setq tramp-use-ssh-controlmaster-options nil)`
>
> in order for vanilla configuration of tramp to support locally display
> of remote X11
>
> Details:
>
> From my home Windows 10 laptop with Xming running I use
Steffen Rick writes:
> Hi,
Hi Steffen,
> I have an odd case where I'm unable to connect to my Freebsd server
> from my Debian desktop. Locally on the Debian box /ssh:: works
> fine. Same goes for the FreeBSD server making a local /ssh::
> connection. .emacs file only contains
>
> (setq
Hello,
The Tramp team is happy to announce the release of Tramp 2.5.2.
Tramp is a remote file editing package for Emacs. Tramp uses different
methods like ssh or scp to access files on remote hosts as if they were
local files. Access to the remote file system for editing files,
version
Reza Housseini writes:
Hi Reza,
> that was perhaps a bit too quick, I meant to say that I made a typo
> and the error still persists when I use the correct syntax: cd
> /ssh:msc@s1-mb-com-02|sudo:s1-mb-com-02:~ gives me
> sudo:s1-mb-com-02:~: command not found
Reproducible here. It looks,
Reza Housseini writes:
> Hello
Hi Reza,
> I am entering a remote shell in eshell via cd /ssh:msc@s1-mb-com-02:~.
> If I then want to have sudo access I have to chain the sudo command as
> follows:
> cd /ssh:msc@s1-mb-com-02:sudo:s1-mb-com-02:~. But I always get
> sudo:s1-mb-com-02:~: command
Reza Housseini writes:
> Hi Michael
Hi Reza,
> Thank you for your fast reply, that did the trick!
Good to know.
> For my own understanding: is it generally better for tramp to connect
> via sh instead of bash or zsh?
Not necessarily. /bin/sh can be a link to any shell, like /bin/bash.
More
Reza Housseini writes:
> Hello
Hi Reza,
> I have a problem with my ssh connection via tramp, I always hang wit
> "Waiting for prompts from remote shell..." and have to cancel the
> request. I have tried to connect via the exec command displayed in the
> tramp-verbose log and everything just
Hi,
the upcoming Tramp 2.6 (included in Emacs 29) will change the signature
of functions, which will be passed as argument of
`tramp-find-foreign-file-name-handler'. Packages, which uses this
mechanism, shall upgrade. See
Daniel Kraus writes:
> Hi!
Hi,
> It's like you said, when you "cd /" you're no longer on the remote server
> but on your local machine.
>
> I have this in my config to not "jump out":
>
> (defun eshell/lcd ( directory)
> "Like regular 'cd' but don't jump out of a tramp directory.
> When
"H. Dieter Wilhelm" writes:
> Hello
Hi Dieter,
> I'm on GNU Emacs 28.0.60 (build 1, x86_64-w64-mingw32) of 2021-10-24
> when I open a remote file e.g.
>
> /plinx:Eval:
>
> and do an "M-x eshell" in this dired buffer, I'm getting an eshell in
> the same remote location (~ under a Gnu-Linux
"H. Dieter Wilhelm" writes:
> Hi Michael
Hi Dieter,
> On my working Laptop I wanted to get first ssh from MSYS2 running, so I
> haven't ssh from MS in the path at the moment. Unfortunately (MSYS2)
> ssh isn't working with Tramp (in an MSYS2 terminal ssh works,
> though). It says for emacs-28
"H. Dieter Wilhelm" writes:
> Hi Michael
Hi Dieter,
> Could I run such a lengthy copying process asynchronously with:
>
> (add-to-list 'tramp-connection-properties
> (list (regexp-quote "/scp:user@host:")
> "direct-async-process" t))
>
> ?
Direct
"H. Dieter Wilhelm" writes:
Hi Dieter,
>> Btw, if you have a recent SSH client on your Windows machine, you could
>> also try the Tramp sshx or scpx methods. Some weeks ago, I've made some
>> changes to support them, OpenSSH 7.9 shall be sufficient. See
>>
"H. Dieter Wilhelm" writes:
> Hi Michael
Hi Dieter,
> Right, the archive consists of a few R code files and some html
> documentation with images.
>
>> So it looks like sending a large string to your remote process does not
>> work. Does it help to set tramp-chunksize, like
>>
>> (setq
"H. Dieter Wilhelm" writes:
> Hi Michael
Hi Dieter,
> I'm very sorry, I just found you answer in my SPAM folder!
>
> I'll have a look tomorrow and report the results...
No problem, take your time.
> Dieter
Best regards, Michael.
積丹尼 Dan Jacobson writes:
Hi Dan,
Thanks for the bug report!
> Maybe instead of cryptic:
> /adb::/root:
> (No match)
> total used in directory 0 available 1.6 GiB
>
> Say:
> /adb::/root:
> (No permission to read directory contents)
> total used in directory 0 available 1.6 GiB
>
>
>
"H. Dieter Wilhelm" writes:
> Hello Michael
Hi Dieter,
> here's the recipe
> emacs-28.0.50 -Q
> Only loading tramp-outoloads.el isn't sufficient! One has to do
> (require 'tramp)
> as well.
No, Tramp is autoloaded. (But it doesn't hurt to require it explicitly).
> (setq tramp-verbose 10)
Vivien Kraus writes:
> Hi Michael,
Hi Vivien,
> If you want to avoid extra work on the server, the proposed solutions
> don’t work, so uname -sr should return something useful, or better yet,
> getconf should work. I think I get it.
Indeed. If "getconf PATH" returns the required directories,
ASANO Fumihisa writes:
> Hi.
Hi,
> I cannot use vc-print-log (git) with tramp master (git HEAD).
> Empty log was shown.
> With tramp-2.5.1, It is OK.
>
> This issue happened after 4784087c2ec30966c0ee91681dbca75cee5a986b .
This commit is not wrong, it has just uncovered an existing error.
>
Vivien Kraus writes:
> Hello Michael,
Hi Vivien,
> Is it not possible to just test for the existence of the /run/current-
> system/profile/bin (and .../sbin if needed) directory? I’m not familiar
> with the tramp code base, why should tramp even know it is connecting
> to a guix system server?
Vivien Kraus via writes:
> Dear tramp developers,
Hi Vivien,
> Guix system stores system-installed packages in
>
> "/run/current-system/profile/bin" "/run/current-system/profile/sbin"
>
> Unfortunately, getconf is unable to produce useful results on a Guix
> System server.
>
> In:
>
>
Stephen Gildea writes:
Hi Stephen,
> After 10 seconds, Emacs recovers by re-opening the connection. I like it.
> Most of the time, the recovery is completely smooth. I have to be fast
> (under 10 seconds) to see Emacs get surprised by a closed connection.
> This works even if the mount point
Michael Albinus writes:
Hi Stephen,
>>> > Two Emacs processes accessing the same remote host can confuse each
>>> other:
>>> >. open an sshfs file in Emacs 1
>>> >. open an sshfs file in Emacs 2 on the same remote host
>>
Stephen Gildea writes:
Hi Stephen,
> I did not mean to imply that the name was poor. Any name more specific
> would probably be too long, e.g., tramp-file-name-remove-non-connection.
>
> I merely intended to ask that the documentation string mention the goal.
> Perhaps:
>
> (defun
Stephen Gildea writes:
Hi Stephen,
> The fix does indeed suppress the error messages. But it does not
> change that Emacs forgets that the new file is a remote file.
Indeed. Fixed in the emacs-28 branch.
Best regards, Michael.
Stephen Gildea writes:
Hi Stephen,
> How about a more descriptive comment for new function tramp-file-name-unify.
> I think the term "unify" is not as specific as it could be. It seems the
> point is that vec objects returned by this function compare 'equal' if they
> refer to the same
Michael Albinus writes:
Hi Stephen,
>> If I open a non-existent file, e.g.,
>> /sshfs:otherhost:/tmp/newfile
>> The following gets logged to *Messages*
>> File is missing: Opening input file No such file or directory
>> /tmp/tramp.sshfs.otherhost/tmp/newf
Henri Salha writes:
> Hello Michael!
Hi Henri,
> Indeed the issue is that the command rclone mount does not work. The
> immediate block I get is that fusermount is not installed, and to do
> this I need to root the tablet, which I refrained from until now.
> If I go forward, I will let you
Stephen Gildea writes:
Hi Stephen,
>> > 3. Unmount if this process did the mount. This is your idea of an
>> > "unmount on cleanup" bit. As with case 2, if this process did not do
>> > the mount, Tramp would have to handle an unexpectedly closed connection.
>> >
>> > 4. Every Emacs
Henri Salha writes:
Hi Henri,
> Hello. I set-up rclone on my android device (Samsung galaxy tab, linux
> access = termux) and successfully connected it to my OneDrive account,
> eg a command (rclone copy onedrive:/path/to/file ~/) works.
Are you saying you are running GNU/Linux on your android
Michael Albinus writes:
Hi Stephen,
> I'll tell you later (today?) when I have pushed something.
Try the appended patch. There's a new user option
`tramp-fuse-unmount-on-cleanup'. When set to non-nil, Tramp shall
unmount the fuse volumes it has mounted itself.
Best regards, Michael.
d
Stephen Gildea writes:
Hi Stephen,
> I see the following ways of dealing with multiple Emacs processes
> sharing an sshfs mount:
>
> 1. Never unmount when closing the connection. This is what Tramp does now.
>
> 2. Always unmount. This might make it hard for Tramp to determine whether
> the
Stephen Gildea writes:
Hi Stephen,
> Using fusermount worked for me. I gave it the same arguments
> tramp-fuse would have used with fusermount3:
>
> fusermount -u /tmp/tramp.sshfs.remotehost
>
> I tested this on Ubuntu 20.04, which has fuse package 2.9.9-3.
Yep, it is just the difference
Stephen Gildea writes:
Hi Stephen,
I'm adding tramp-devel@gnu.org to Cc, likely this is of public interest.
> When I run tramp-cleanup-all-connections or tramp-cleanup-this-connection
> on an sshfs buffer, the sshfs process and mount do not get cleaned up.
There are different scenarii to be
JD Smith writes:
Hi,
> I can confirm that the latest TRAMP devel package fixes the file
> completion slowdown problem on MacOS, without any special
> configuration.
Thanks for the feedback! Tramp 2.5.1.3 will appear next week on GNU ELPA
(unless a disaster happens during final tests, crossing
JD Smith writes:
Hi,
> Yes, there are still discrepancies in my tests (I'm running
> GNU/Linux),
> mainly in tramp-sh-handle-expand-file-name:
>
> I meant only that this darwin-shell-calling behavior hadn’t yet been
> fixed on those newer versions, but I anticipate 2.5.1.3 will fix
JD Smith writes:
Hi,
> In that case it sounds like your testing will more than suffice!
I have no macOS machine, neither for development, nor for remote
access. So your tests will be welcome :-)
Anyway, I have provided a snapshot package. In order to access it, eval
--8<---cut
JD Smith writes:
Hi,
> I’ll give 2.5.1.3 a test; let me know if you have a release candidate
> I can try.
GNU ELPA allows also to release snapshots, I could prepare something for
you.
Disclaimer: I have an ansible script which runs Tramp tests for more
than 100 different config variants,
JD Smith writes:
Hi,
> OK I believe I’ve cracked it.
Great!
> And, at long last, HERE we see the problem, in
> `tramp-compat-temporary-file-directory':
>
> (eval (car (get 'temporary-file-directory 'standard-value)) t))
>
> Why is this an issue? That standard value of
>
JD Smith writes:
Hi,
> In the meantime I’ve searched down many rabbit holes and zeroed in on
> tramp’s file handlers. I am using:
>
> (require 'tramp-sh)
> (defun my/test ()
> (interactive)
> (progn
> (tramp-cleanup-all-connections)
> (elp-instrument-list
> '(file-remote-p
>
JD Smith writes:
Hi,
> Thanks very much for this much improved testing code (which also saves
> the tedium of slowly retyping!). I altered it slightly to include a
> few more function, and target my local wired server:
>
> (progn
> (require 'tramp-sh)
> (tramp-cleanup-all-connections)
>
Michael Albinus writes:
Hi,
>> Let me think over night, what we could do for checking (yes, I
>> often
>> have my best ideas in bed, when I cannot sleep :-)
>>
>> (I hate to wish someone a poor sleep… so: ) Enjoy your weekend, and
>> thanks a
JD Smith writes:
Hi,
> Let me think over night, what we could do for checking (yes, I
> often
> have my best ideas in bed, when I cannot sleep :-)
>
> (I hate to wish someone a poor sleep… so: ) Enjoy your weekend, and
> thanks again!
Thanks, more food for next nights :-)
If you
JD Smith writes:
Hi,
> I appreciate your thoughts and testing, and am glad you could
> reproduce the problem. Using exactly the same “plain TAB-based”
> testing setup you proposed with my local server on a wired connection,
> I can still certainly see a difference starting from *scratch* vs. a
JD Smith writes:
Hi,
> I’m trying to debug a very strange issue I’ve encountered when using
> completion UI’s like vertico, selectrum, and fido to navigate remote
> file paths with tramp (over SSH). It also occurs (I believe) during
> normal tab-driven file completion, but isn’t as obvious
Martin Joerg writes:
> Hi,
Hi Martin,
> it looks like commit 5d614887 introduced an issue with the rsync method which
> occurs unless tramp-use-ssh-controlmaster-options is t and
> tramp-ssh-controlmaster-options is non-empty.
>
> When opening a file larger than tramp-copy-size-limit using the
Fredrik Salomonsson writes:
> Hi,
Hi Fredrik,
> I'm trying to use Per-Connection Local Variables [0] to setup a profile
> for machines I connect to at work. I followed the steps described in the
> TRAMP manual [1] and the emacs manual [0].
>
> I didn't get it to work if I tried with a regexp,
Michael Albinus writes:
Hi Sam,
> Could you try to rerun your test with "emacs -Q"?
PS: If you set tramp-verbose to 10, we'll get a Lisp backtrace as bonus
in case of problems. So you might better call
emacs -Q --eval '(setq tramp-verbose 10)'
Best regards, Michael.
Steingold writes:
Hi Sam,
> when I try to access a remote file on `fence`, the connection to
> `dev-dsk-sdsg` is reopened despite me having called
> tramp-cleanup-all-connections.
> dev-dsk-sdsg is not available because I am not on the VPN at the moment.
>
> this is a recent development, I was
Hello,
The Tramp team is happy to announce the release of Tramp 2.5.1. This is
the development version, covering new features.
Tramp is a remote file editing package for Emacs. Tramp uses different
methods like ssh or scp to access files on remote hosts as if they were
local files. Access to
Steingold writes:
Hi Sam,
> When I edit a remote file using tramp, `compile` works when called directly.
> But when called from `elpy-check`, it asks for credentials (login
> id/password).
> (https://github.com/jorgenschaefer/elpy/blob/master/elpy.el#L1700)
> I load `tramp-sh` as source, set
Kevin Broadey writes:
Hi Kevin,
> I can understand tramp wanting control of the path for remote file
> operations. I agree with that. But when tramp chooses to inveigle
> itself into running “make” remotely, then surely it knows it isn’t
> trying to do a remote file operation and ought to do
Kevin Broadey writes:
Hi Kevin,
> Yep, I googled ‘tramp-own-remote-path’ and found the bit in the massive
> manual about how to change this behaviour. Thanks for the tip.
Why searching the web? 'C-h i g (tramp) RET' gives you the Tramp manual.
> But why should I have to go searching to get
Joseph Padgett writes:
> Hello,
Hi Joseph,
> I posted about this issue on StackExchange and someone suggested
> contacting this email address.
That was me :-)
> The problem I am having occurs when I try to save with privileges. I
> use sudo within emacs to open dired, and from there I open
Kevin Broadey writes:
Hi Kevin,
> The only way I've been able to get Tramp to Do The Right Thing and
> leave my remote PATH the way I want it to be is to put this in my .
> emacs file...
>
> ;; tramp is overzealous about setting the remote PATH,
> ;; contrary to .profile, .bash_profile
Jon Schewe writes:
Hi Jon,
> In the end the problem was not with the compiliation, but with the
> explicit call to load-library.
Good. Btw, does this fix also your initial problem?
Best regards, Michael.
Jon Schewe writes:
Hi Jon,
>> I see that my load-path has ~/.emacs.d/elpa/tramp-2.5.0.4 in my load
>> path before /usr/share/emacs/26.3/lisp/net, so that should make the
>> new
>> version load rather than the old version.
>>
>
>> However when I run emacs without "--no-init" I find that tramp-
Jon Schewe writes:
Hi Jon,
> I tried installing tramp-2.5.0.4 from GNU ELPA and it installed into my
> home directory, but doesn't seem to want to load, I keep getting the
> version of tramp that is installed with Emacs.
I guess it is installed at "~/.emacs.d/elpa/tramp-2.5.0.4". I suppose,
Jon Schewe writes:
Hi Jon,
> Tramp appears to be confused about what directories belong to each
> host.
>
> Emacs : GNU Emacs 26.3 (build 2, x86_64-pc-linux-gnu, GTK+ Version 3.24.14)
> of 2020-03-26, modified by Debian
> Package: tramp (2.3.5.26.3)
Yes, this is a known error:
>
Jim Porter writes:
Hi Jim,
> There's something unusual happening with how shell escaping works on
> MS Windows... pscp works fine for me on Tramp 2.5.0.4, but fails on
> quoted files in the development build. See the table at the end for
> the results; for consistency with my last set of tests,
Jim Porter writes:
Hi Jim,
>> Short update. Since I cannot find a file name quoting which satisfies
>> both pscp and MS Windows scp, I've decided to add a new user option
>> tramp-use-ms-dos-quote-argument. It's default value is nil, but users
>> can change it to non-nil. In that case, Tramp
Jim Porter writes:
Hi Jim,
>> > I'll see if I can get the automated test suites running today or tomorrow.
>>
>> That would be great. I had to stop for today (there's also other
>> business, surprisingly), but I will continue tomorrow.
Short update. Since I cannot find a file name quoting
[I've added the tramp-devel ML to CC, for the archives. Pls keep]
Daniel Mendler writes:
> Guten Morgen!
Guten Morgen, Daniel.
> I am working on a set of completion packages which try to work closely
> with the facilities provided by Emacs proper.
>
> * https://github.com/minad/vertico
> *
Jim Porter writes:
Hi Jim,
> I'll see if I can get the automated test suites running today or tomorrow.
That would be great. I had to stop for today (there's also other
business, surprisingly), but I will continue tomorrow.
> - Jim
Best regards,
Jim Porter writes:
Hi Jim,
> Ok, I've tested in all the combinations I can, with the results in
> the table below. "tramp-dev" is the latest Git revision, whereas
> "tramp-2.5.0.4" is the version on GNU ELPA. I tested against the Win32
> versions of OpenSSH (both v7.7 and v8.1) as well as the
Jim Porter writes:
Hi Jim,
>> Interestingly, there's no such problem when using plink, which uses the
>> same command sequence. That lets me think it is an ssh client problem,
>> buffering the output or whatever. But I might be wrong.
>
> Indeed, it's hard to say for sure what's going wrong,
Jim Porter writes:
Hi Jim,
> I tested out Win32-OpenSSH v8.1, and the quoting issue hasn't been
> resolved. In fact, it's somewhat worse. I found that you can escape
> the spaces like so:
>
> scp "server:~/'foo bar.txt'" dest.txt (1)
>
> However, newer versions of OpenSSH have strict
Jim Porter writes:
Hi Jim,
> Is it possible that there's an issue with how Tramp runs the process?
> In the debug log for Tramp, I see the following commands (shortened
> for readability):
>
> ssh-e none -t -t server /bin/sh && exit || exit
>
> cd /home/jim/src/tramp/info/ && exec
Michael Albinus writes:
Hi Jim,
>> This is an odd issue, and I'm not sure what's happening yet: sometimes
>> when calling `M-x rgrep' or other similar commands, I don't see all
>> the results I'd expect. I'm consistently able to reproduce the bug but
>> it only appea
Jim Porter writes:
Hi Jim,
> This is an odd issue, and I'm not sure what's happening yet: sometimes
> when calling `M-x rgrep' or other similar commands, I don't see all
> the results I'd expect. I'm consistently able to reproduce the bug but
> it only appears rarely, so it usually requires
André A. Gomes writes:
> Hi!
Hi André,
> I investigated a bit, and I don't understand why the following
>
> (tramp-compat-process-running-p "gvfsd-fuse")
>
> returns nil for me, since from M-x proced I see:
>
> aadcg 835 0.0 0.2 Apr 23 00:17
>
Hendrik Tews writes:
> Hi Michael,
Hi Hendrik,
>> a locking mechanism shall be added by the caller.
>
> There are two callers in the above example, the library which
> calls file-attributes or delete-file and the user, who does file
> name completion during find-file. I don't think the user
Hendrik Tews writes:
> Hi,
Hi Hendrik,
>> Could you pls report back whether this helps?
>
> yes, the included version of tramp-get-remote-stat fixes the
> problem. I did
> - (require 'tramp-sh)
> - evaled the new tramp-get-remote-stat defun
> - checked the let expression and also independently
"Felix E. Klee" writes:
Hi Felix,
>> Would you mind to check, whether it fixes the compatibility issue in
>> your setup?
>
> It worked flawlessly.
Thanks for the feedback. I've pushed it to the repositories.
Best regards, Michael.
Michael Albinus writes:
Hi Felix,
>>> Good. However, I need to find out what happens with compiled Tramp
>>> files. Hopefully, I can reproduce the problem locally.
>>
>> Let me know if I you need help reproducing the issue.
>
> Thanks for the offer. If nee
Jim Porter writes:
Hi Jim,
>> Doesn't work. Do you have an idea, which kind of quoting for the remote
>> target I need?
>
> I haven't spent too much time looking at the scp method under MS
> Windows, but I know there have been some changes to Win32-OpenSSH that
> might affect things here. In
Hendrik Tews writes:
> Hi,
Hi Hendrik,
>> That's not possible. All synchronous remote operations use the *same*
>> process. file-attributes, and the operations serving file name
>> completion, are synchronous operations.
>
> OK, but this implies that sentinels, process filters and timers
>
Jim Porter writes:
Hi Jim,
>> That's possible, yes. But besides the POSIX shell, Tramp needs on the
>> remote side also some POSIX tools, like find, grep, stat, you name it.
>>
>> Might be worth to try it, but since I don't run a (remote) MS Windows
>> machine, I have no chance to test.
>
> Git
Hendrik Tews writes:
> Hi,
Hi Hendrik,
> in the application that I outlined in my previous email "optimize
> tramp for many start-file-process'es in parallel", Tramp
> functionality is invoked asynchronously from within a process
> sentinel. The functionality has been purposely moved into
>
Jim Porter writes:
Hi Jim,
>> Accessing an ssh server on MS Windows via Tramp will fail in general,
>> because Tramp expects a POSIX shell on the server side. If I read the
>> docs correctly, the ssh server on MS Windows opens a powershell.
>
> You can actually set the default ssh shell on MS
201 - 300 of 1903 matches
Mail list logo
