run-hook-with-args: Symbol's function definition is void: shell-directory-tracker

[Barry Warsaw]

In current bzr trunk, using Emacs 23.3.1, hit C-c ! to run py-shell. The 2.7.2 shell comes up. At the prompt, type

>>> from datetime import * [RET]

nothing happens, but you get the error in the Summary.

[Andreas Roehler (a-roehler)]

Am 19.10.2011 17:40, schrieb Barry Warsaw:
> Public bug reported:
>
> In current bzr trunk, using Emacs 23.3.1, hit C-c ! to run py-shell.
> The 2.7.2 shell comes up. At the prompt, type
>
>>>> from datetime import * [RET]
>
> nothing happens, but you get the error in the Summary.
>
> ** Affects: python-mode
> Importance: Undecided
> Status: New
>

shell-directory-tracker is a compiled Lisp function in
`shell.el'.

loading shell.el should do it

Cheers,

Andreas

[Barry Warsaw (barry)]

On Oct 19, 2011, at 05:09 PM, Andreas Roehler wrote:

>shell-directory-tracker is a compiled Lisp function in
>`shell.el'.
>
>loading shell.el should do it

Sure, but why should the *Python* buffer care? It's not going to use shell
commands like cd, pushd, and popd.

[Andreas Roehler (a-roehler)]

Am 19.10.2011 19:23, schrieb Barry Warsaw:
> On Oct 19, 2011, at 05:09 PM, Andreas Roehler wrote:
>
>> shell-directory-tracker is a compiled Lisp function in
>> `shell.el'.
>>
>> loading shell.el should do it
>
> Sure, but why should the *Python* buffer care? It's not going to use shell
> commands like cd, pushd, and popd.
>

ipython shell does.

which brings some convenience IMHO

think I'll just add a (require 'shell)

(?)

[Barry Warsaw (barry)]

On Oct 19, 2011, at 06:01 PM, Andreas Roehler wrote:

>ipython shell does.
>
>which brings some convenience IMHO
>
>think I'll just add a (require 'shell)

Maybe just when the ipython shell is used?

After a fresh restart of Emacs, after I open a "normal" shell, the error
message and hang does not seem to happen, since by that time

(featurep 'shell)
-> t

[Andreas Roehler (a-roehler)]

Am 19.10.2011 20:31, schrieb Barry Warsaw:
> On Oct 19, 2011, at 06:01 PM, Andreas Roehler wrote:
>
>> ipython shell does.
>>
>> which brings some convenience IMHO
>>
>> think I'll just add a (require 'shell)
>
> Maybe just when the ipython shell is used?
>
> After a fresh restart of Emacs, after I open a "normal" shell, the error
> message and hang does not seem to happen, since by that time
>
> (featurep 'shell)
> -> t
>

Hmm, IMO 'shell is not such a exotic feature.
May have a look, but think requiring it should not be a matter anyway.

BTW what do you think implementing "cd" and the other basic shell
commands ipython-shell has for all python shells?

[Barry Warsaw (barry)]

On Oct 19, 2011, at 07:27 PM, Andreas Roehler wrote:

>Hmm, IMO 'shell is not such a exotic feature.
>May have a look, but think requiring it should not be a matter anyway.

Probably not.

>BTW what do you think implementing "cd" and the other basic shell
>commands ipython-shell has for all python shells?

Hmm, I think I would keep the buffer interface as a thin layer above the
underlying shell.

[Andreas Roehler (a-roehler)]

Am 20.10.2011 04:19, schrieb Barry Warsaw:
> On Oct 19, 2011, at 07:27 PM, Andreas Roehler wrote:
>
>> Hmm, IMO 'shell is not such a exotic feature.
>> May have a look, but think requiring it should not be a matter anyway.
>
> Probably not.
>
>> BTW what do you think implementing "cd" and the other basic shell
>> commands ipython-shell has for all python shells?
>
> Hmm, I think I would keep the buffer interface as a thin layer above the
> underlying shell.
>

OK, made a separate report lp:878669 for the latter.

Will avoid the bug with a (require 'shell) for the moment.