Re: [elpa] master 5b011ac 1/2: multishell - note that the tramp bug is on remote|sudo to homedir

Re: [elpa] master 5b011ac 1/2: multishell - note that the tramp bug is on remote|sudo to homedir

[Michael Albinus]

Ken Manheimer <ken.manheimer@gmail.com> writes:

Hi Ken,

> branch: master
> commit 5b011ac2c545be3ca824862f48032c9a2980798b
> Author: Ken Manheimer <ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> Commit: Ken Manheimer <ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>
>     multishell - note that the tramp bug is on remote|sudo to homedir

Could you please write a bug report about this? I would like to check
what's up.

Please start with emacs -Q, and set tramp-verbose to 6. There will be a
Tramp debug buffer, which you might append to the bug report.

Thanks, and best regards, Michael.

Re: [elpa] master 5b011ac 1/2: multishell - note that the tramp bug is on remote|sudo to homedir

[Ken Manheimer]

[-- Attachment #1: Type: text/plain, Size: 2184 bytes --]

Hey, Michael.

When I run emacs with -Q, I'm unable to produce the problem! (It's not
multishell, per se - I'm unable to produce it via multishell or separately
from it, in an emacs started with -Q.) I'll remove the notes in multishell
about it, and work on isolating what in my local setup is at fault.

Interestingly, the problem *does* happen if I start emacs with '-q',
instead of '-Q' (I didn't even think about trying -Q until you suggested
it). This suggests that the problem is in something my my site
configuration, I gather. I'll try to uncover it, but it looks like that's
going to be via a process of elimination. Unfortunately, I don't get a
traceback, even with tramp-verbose=6, when the problem occurs, only
"Selecting deleted buffer", and in *Messages*, "byte-code: Selecting
deleted buffer". Nothing more happens if debug-on-error and debug-on-signal
are set. This all suggests to me that something is swallowing the signal,
yet still returning to top-level. I'd be interested to hear if there's
common strategies for dealing with such a puzzle.

Thanks for following up, and steering me in the right direction. Sorry
about misleading anyone, mistaking the problem as being in tramp rather
than some combination. I'll let you know if there's anything interesting
about what's actually responsible for the problem, when I uncover it, and
suggestions about investigating the "Selecting deleted buffer" thing would
be appreciated.

Ken

On Sat, Jan 23, 2016 at 4:18 AM, Michael Albinus <michael.albinus@gmx.de>
wrote:

> Ken Manheimer <ken.manheimer@gmail.com> writes:
>
> Hi Ken,
>
> > branch: master
> > commit 5b011ac2c545be3ca824862f48032c9a2980798b
> > Author: Ken Manheimer <
> ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> > Commit: Ken Manheimer <
> ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> >
> >     multishell - note that the tramp bug is on remote|sudo to homedir
>
> Could you please write a bug report about this? I would like to check
> what's up.
>
> Please start with emacs -Q, and set tramp-verbose to 6. There will be a
> Tramp debug buffer, which you might append to the bug report.
>
> Thanks, and best regards, Michael.
>

[-- Attachment #2: Type: text/html, Size: 2969 bytes --]

Re: [elpa] master 5b011ac 1/2: multishell - note that the tramp bug is on remote|sudo to homedir

[Ken Manheimer]

[-- Attachment #1: Type: text/plain, Size: 2957 bytes --]

On Sat, Jan 23, 2016 at 5:36 PM, Ken Manheimer <ken.manheimer@gmail.com>
wrote:

> Hey, Michael.
>
> When I run emacs with -Q, I'm unable to produce the problem! (It's not
> multishell, per se - I'm unable to produce it via multishell or separately
> from it, in an emacs started with -Q.) I'll remove the notes in multishell
> about it, and work on isolating what in my local setup is at fault.
>
> Interestingly, the problem *does* happen if I start emacs with '-q',
> instead of '-Q' (I didn't even think about trying -Q until you suggested
> it). This suggests that the problem is in something my my site
> configuration, I gather. I'll try to uncover it, but it looks like that's
> going to be via a process of elimination. Unfortunately, I don't get a
> traceback, even with tramp-verbose=6, when the problem occurs, only
> "Selecting deleted buffer", and in *Messages*, "byte-code: Selecting
> deleted buffer". Nothing more happens if debug-on-error and debug-on-signal
> are set. This all suggests to me that something is swallowing the signal,
> yet still returning to top-level. I'd be interested to hear if there's
> common strategies for dealing with such a puzzle.
>
> Thanks for following up, and steering me in the right direction. Sorry
> about misleading anyone, mistaking the problem as being in tramp rather
> than some combination. I'll let you know if there's anything interesting
> about what's actually responsible for the problem, when I uncover it, and
> suggestions about investigating the "Selecting deleted buffer" thing would
> be appreciated.
>

I was mistaken about being mistaken! (-: / )-:

It does fail in emacs -Q. I just needed to be sure that my recipe does
*not* allow emacs to convert the destination in a way that fills in the
homedir with an explicit path.

I submitted a bug report. The recipe, for the curious (using "example.com"
for a host to which you have ssh and sudo access):

   - Start emacs with -Q
   - In any buffer (*scratch* is fine):
   M-Esc (setq default-directory "/ssh:example.com|sudo:root@example.com:")
   - In that same buffer, Esc-x shell

I'd love to know if there's a workaround that I could ship with multishell.

Ken

Ken
>
> On Sat, Jan 23, 2016 at 4:18 AM, Michael Albinus <michael.albinus@gmx.de>
> wrote:
>
>> Ken Manheimer <ken.manheimer@gmail.com> writes:
>>
>> Hi Ken,
>>
>> > branch: master
>> > commit 5b011ac2c545be3ca824862f48032c9a2980798b
>> > Author: Ken Manheimer <
>> ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>> > Commit: Ken Manheimer <
>> ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>> >
>> >     multishell - note that the tramp bug is on remote|sudo to homedir
>>
>> Could you please write a bug report about this? I would like to check
>> what's up.
>>
>> Please start with emacs -Q, and set tramp-verbose to 6. There will be a
>> Tramp debug buffer, which you might append to the bug report.
>>
>> Thanks, and best regards, Michael.
>>
>
>

[-- Attachment #2: Type: text/html, Size: 4685 bytes --]

Re: [elpa] master 5b011ac 1/2: multishell - note that the tramp bug is on remote|sudo to homedir

[Ken Manheimer]

[-- Attachment #1: Type: text/plain, Size: 2614 bytes --]

On Sat, Jan 23, 2016 at 9:04 PM, Ken Manheimer <ken.manheimer@gmail.com>
wrote:

> On Sat, Jan 23, 2016 at 5:36 PM, Ken Manheimer <ken.manheimer@gmail.com>
> wrote:
>
>> Hey, Michael.
>>
>> When I run emacs with -Q, I'm unable to produce the problem! (It's not
>> multishell, per se - I'm unable to produce it via multishell or separately
>> from it, in an emacs started with -Q.) I'll remove the notes in multishell
>> about it, and work on isolating what in my local setup is at fault.
>> [...]
>>
>
> I was mistaken about being mistaken! (-: / )-:
>
> It does fail in emacs -Q. I just needed to be sure that my recipe does
> *not* allow emacs to convert the destination in a way that fills in the
> homedir with an explicit path.
>
> I submitted a bug report. The recipe, for the curious (using "example.com"
> for a host to which you have ssh and sudo access):
>
>    - Start emacs with -Q
>    - In any buffer (*scratch* is fine):
>    M-Esc (setq default-directory "/ssh:example.com|sudo:root@example.com
>    :")
>    - In that same buffer, Esc-x shell
>
> I'd love to know if there's a workaround that I could ship with multishell.
>

I actually have a workaround, and the premise of it resides in the special
conditions I included for the repeating the problem. In order to repeat the
problem, you have to specifically use a default-directory that is set to
the remote+sudo+homedir format, and not allow something like `(cd path)' to
first resolve the home directory to an explicit path. So there's my
workaround, for multishell - use `(cd path)', rather than directly setting
default-directory.

Upshot is that I believe that there still is a bug worth unravelling, in
the way that shells start when default-directory is set to
remote+sudo+homedir format, but I can avoid that by using `cd'.

Ken

Ken
>
> Ken
>>
>> On Sat, Jan 23, 2016 at 4:18 AM, Michael Albinus <michael.albinus@gmx.de>
>> wrote:
>>
>>> Ken Manheimer <ken.manheimer@gmail.com> writes:
>>>
>>> Hi Ken,
>>>
>>> > branch: master
>>> > commit 5b011ac2c545be3ca824862f48032c9a2980798b
>>> > Author: Ken Manheimer <
>>> ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>>> > Commit: Ken Manheimer <
>>> ken.manheimer-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>>> >
>>> >     multishell - note that the tramp bug is on remote|sudo to homedir
>>>
>>> Could you please write a bug report about this? I would like to check
>>> what's up.
>>>
>>> Please start with emacs -Q, and set tramp-verbose to 6. There will be a
>>> Tramp debug buffer, which you might append to the bug report.
>>>
>>> Thanks, and best regards, Michael.
>>>
>>
>>
>

[-- Attachment #2: Type: text/html, Size: 4899 bytes --]