[SVN] Re: [Patches] TM::Executor additions

[Alex Ross]

On Mar 26, 2009, at 4:29 PM, Luke Daley wrote:

>
> On 26/03/2009, at 10:34 PM, Martin Kühl wrote:
>
>> On Thu, Mar 26, 2009 at 00:50, Luke Daley <ld at ldaley.com> wrote:
>>>
>>> On 26/03/2009, at 4:45 AM, Martin Kühl wrote:
>>>
>>>> • Passing the :input and :interactive_input options through to the
>>>> underlying TM::Process.run invokation.
>>>> The Maude bundle uses this to script the maude interpreter from the
>>>> TextMate command, usually to wrap a command around the current
>>>> word/selection.
>>>> • Deactivating the interactive input library if the variable
>>>> TM_INTERACTIVE_INPUT_DISABLED is defined.
>>>> Allan requested this on IRC.
>>>
>>> What was the reason for needing two ways to disable interactive
>>> input?
>>
>> As I understand it, one way works on the command level, where it
>> doesn’t make sense for some commands to receive interactive input,
>> while the other works on the project (or global) level, where
>> unfortunate circumstances make the interactive input library behave
>> badly (this should, to quote Allan, "solve that/those python
>> problems").
>>
>> If you mean that commands could set TM_INTERACTIVE_INPUT_DISABLED
>> instead of :interactive_input to disable it, I guess that’s more of a
>> stylistic question, but think that it should be consistent for
>> Executor.run and Process.run.
>>
>> Does that answer your question?
>
> Yes, fair enough.
>
> What are the problems with it though? Are they bugs in
> tm_interactive_input.dylib?

Yes, tm_interactive_input.dylib is causing some python programs to  
crash mysteriously.  Something to do with flat-namespacing.

—Alex