GNU bug report logs - #22145
25.0.50; tramp and auto-revert tests hang on OS X

Previous Next

Package: emacs;

Reported by: Anders Lindgren <andlind <at> gmail.com>

Date: Fri, 11 Dec 2015 19:22:01 UTC

Severity: normal

Found in version 25.0.50

Done: Michael Albinus <michael.albinus <at> gmx.de>

Bug is archived. No further changes may be made.

Full log


Message #50 received at 22145 <at> debbugs.gnu.org (full text, mbox):

From: Anders Lindgren <andlind <at> gmail.com>
To: Michael Albinus <michael.albinus <at> gmx.de>
Cc: 22145 <at> debbugs.gnu.org
Subject: Re: bug#22145: 25.0.50; tramp and auto-revert tests hang on OS X
Date: Fri, 18 Dec 2015 20:55:42 +0100
[Message part 1 (text/plain, inline)]
The cases I found didn't occur when running under `ert', because then I
could not manage to break into the debugger.

However, today I tested with `debug-on-quit' to t and `ert-debug-on-error'.
When I press C-g I go into the debugger and it looks like it's executing
the `test' commands you suggested earlier. I've attached the backtrace (for
what's it worth).

Anyway, it works with simple filenames like "xyz", so it is somehow
connected to the non-ascii file names. By the way, are you aware that OS X
denormalizes file names when stored in the file system, could that be
related to this in any way?

    -- Anders

On Fri, Dec 18, 2015 at 9:12 AM, Michael Albinus <michael.albinus <at> gmx.de>
wrote:

> Anders Lindgren <andlind <at> gmail.com> writes:
>
> > Hi,
>
> Hi Anders,
>
> > I tried to see what actually happened in the test cases.
> >
> > I seem to have boiled things down to the following line, when
> > evaluated, tramp hangs. When enabling debug-on-quit, I get the
> > attached backtrace after hitting C-g.
> >
> > (write-region "test" nil
> >
> "/mock:anders-lindgrens-macbook.local:/private/var/folders/qq/qqGPPGoBHOGgJs3oBoKTHU+++TI/-Tmp-/tramp-test1900-Fy/test")
> >
> >
> > The directory exists and is empty.
> >
> > It looks like the code tries to run an external command "mock" (which
> > doesn't exits, as far as I know).
>
> Nope. "mock" is not a command. It is a fake (mock) Tramp method I create
> on-the-fly when running tramp-test.el. This is because I cannot use any
> real Tramp method, because usually they require user authentication,
> which is a no-go for tests running in batch mode.
>
> According to your backtrace, you have interrupted Tramp at the following
> place:
>
> > Debugger entered--Lisp error: (quit)
> >   accept-process-output(#<process *tramp/mock
> anders-lindgrens-macbook.local*> 1 nil t)
> >   tramp-accept-process-output(#<process *tramp/mock
> anders-lindgrens-macbook.local*> 1)
> >   tramp-wait-for-regexp(#<process *tramp/mock
> anders-lindgrens-macbook.local*> nil
> "\\(^\\|\\)[^#$\n]*///9132fe7dc760b6d1120df52ca7cc56ca#\\$?$")
> >   tramp-wait-for-output(#<process *tramp/mock
> anders-lindgrens-macbook.local*>)
> >   tramp-send-command(["mock" nil "anders-lindgrens-macbook.local"
> "/private/var/folders/qq/qqGPPGoBHOGgJs3oBoKTHU+++TI/-Tmp-/tramp-test1900-Fy/test"
> nil] "echo \\\"`getconf PATH 2>/dev/null`\\\" 2>/dev/null; echo
> tramp_exit_status $?")
>
> This is the command "echo \\\"`getconf PATH 2>/dev/null`\\\"
> 2>/dev/null; echo tramp_exit_status $?", which shouldn't fail. At least
> it didn't fail so far.
>
> Could you send the full Tramp traces? Maybe my last change has damaged
> something else.
>
> > -- Anders
>
> Best regards, Michael.
>
[Message part 2 (text/html, inline)]
[tramp-bt4.txt (text/plain, attachment)]

This bug report was last modified 9 years and 47 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.