--- Begin Message ---
Subject: |
24.4.50; (process-file) erroneously raises its buffer when running with TRAMP |
Date: |
Fri, 20 Jun 2014 01:09:28 -0700 |
Hi.
The documentation is a bit unclear, so I'm not 100% sure this is a bug;
it's definitely a surprising behavior, though.
I have a bit of elisp to create a temporary buffer and to run a process,
sending its output to this buffer:
(let ((output-buffer (get-buffer-create "*test-buf*")))
(with-current-buffer output-buffer
(erase-buffer)
(let ((default-directory "/tmp"))
(process-file "whoami" nil output-buffer t))))
Note that I do not ask for this buffer to be raised. On my machine
(Debian/sid amd64) this indeed does not raise the *test-buf* buffer, and
I do not even see it if I don't explicitly switch to it. This is good.
If I change the directory from "/tmp" to any TRAMP path (for instance
"/sudo::/tmp") then this elisp DOES raise *test-buf*. This difference
between normal and TRAMP behavior sounds like a bug to me.
Note that I have (process-file ... ... ... t). Changing this to nil
resolves the issue. The documentation says
Fourth arg DISPLAY non-nil means redisplay buffer as output is
inserted.
I don't know if "redisplay" includes "raise", but I do think the
behavior should be the same, TRAMP or not.
Thanks
--- End Message ---
--- Begin Message ---
Subject: |
Re: bug#17815: 24.4.50; (process-file) erroneously raises its buffer when running with TRAMP |
Date: |
Sun, 22 Jun 2014 11:28:21 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/24.4.50 (gnu/linux) |
Stefan Monnier <address@hidden> writes:
>>>> @Stefan: This is no regression, I could reproduce it even with Emacs
>>>> 23.4. Therefore, I will fix it in the trunk. Please tell me if you
>>>> believe it shall go into emacs-24.
>>> Show me the patch (when it's ready), so I can see whether it looks
>>> safe enough.
>> That's what I've committed to the Tramp repository:
>
> Looks safe enough for emacs-24, thanks.
I've committed the lisp files to the emacs-24 branch as 117284, closing
the bug. tramp-tests.el will be committed to the trunk, next time
emacs-24 has been merged there.
> And in trunk, could you try and reduce the code-duplication between
> tramp-sh.el and tramp-adb.el?
Well, all handlers I could factor out for several backends, live in
tramp.el as `tramp-handle-...'. `tramp-adb-handle-process-file' and
`tramp-sh-handle-process-file' contain subtle differences, it will be
harder to refactor them.
> Stefan
Best regards, Michael.
--- End Message ---