[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Couldn't find exit status opening directory by plink
From: |
poti |
Subject: |
Re: Couldn't find exit status opening directory by plink |
Date: |
Thu, 14 Jun 2007 15:45:04 -0400 |
User-agent: |
Mutt/1.5.13 (2006-08-11) |
On 09:54 Thu 14 Jun , Michael Albinus wrote:
> address@hidden writes:
>
> > Starting Emacs and issuing this command:
> > /plink:address@hidden:/home/me/
> > I get one error the first time, and the one I posted the second time.
>
> According to the debug buffer you've sent, there are problems indeed:
>
> 17:15:09 tramp-send-command-internal (6) # stty -inlcr -echo kill '^U' erase
> '^H'
> 17:15:09 tramp-wait-for-regexp (6) #
> stty: standard input: Invalid argument
> address@hidden ~ $
...
> Could you call the stty command directly in a shell on that host?
When I paste the above command in my shell, through putty or at the
console, the command is accepted with no warning. The commands issued
one at a time appear to work, with -echo leading to a screen in which I
cannot see my typing, and in which pressing return leads to the shell
prompts lining up on a line. stty echo brings all back to normal. Changing
erase visibly makes my delete key no longer work, and I am not sure what
to do to test kill.
> Could you check stty's man page what's wrong with this command?
I am glad to work on it. What outcome does tramp require?
All this works from OSX with the same tramp but older Emacs to the same
Gentoo machine, by the way.
> 17:15:40 tramp-send-command (6) # \stat -c '(("%N") %h %u %g %X.0 %Y.0 %Z.0
> %s "%A" t %i -1)' /home 2>/dev/null; echo tramp_exit_status $?
> 17:15:40 tramp-wait-for-regexp (6) #
> (("`/home'") 8 0 0 -421032914.0 1173248849.0 1173248849.0 4096 "drwxr-xr-x" t
> 2 -1)
>
> ///e7aea0b72c6a0f27266cd78fb6280ba9
> tramp_exit_status 0
> 17:15:40 tramp-send-command-and-check (1) # File error: Couldn't find exit
> status of `\stat -c '(("%N") %h %u %g %X.0 %Y.0 %Z.0 %s "%A" t %i -1)' /home'
>
> Here we see that the prompt did appear before the result of the stat
> command. Therefore, Tramp could not check the return code, and it did
> fail.
>
> I've never seen this. Previous commands in the trace did show the
> right order of exit status and prompt. Maybe we can cure it with a
> proper stty command.
I replicated the same results sending this reply. I am not sure what
function to enter on, and am new to the debugger, but would stepping through
with debug-on-entry give the option of adjusting the options sent with
stty?