Matthieu MOY <[EMAIL PROTECTED]> writes:

> Quoting Mark Triggs <[EMAIL PROTECTED]>:
>
>> Currently my tla-run-arch is just defined as:
>> 
>>   (defun tla-run-arch (arglist run-asynchron &rest keys)
>>     (message "tla-run-arch: %S" arglist)
>>     (if run-asynchron
>>         (apply 'tla-run-asynchronously "tla" arglist keys)
>>       (apply 'tla-run-synchronously "tla" arglist keys))) 
>> 
>> So it's really just a wrapper around those two functions.  They're
>> implemented quite differently, but they behave fairly similarly (they
>> both take :finish, :error and :killed arguments, for instance).  Is
>> this the sort of thing you mean, or would it be better to have the
>> caller choose which function they want?
>
> Yes, but then, you can just remove tla-run-arch, and let the user call
> tla-run-asynchronously or tla-run-synchronously directly.
>
> Keeping the same interface has the advantage that changing the synchronousness
> of a process is easier (Just a 'a' to add or remove), but it doesn't seem to 
> be
> mandatory to me.

Sure, I don't disagree.  Actually, the reason I kept a single function
called `tla-run-arch' was just to have something resembling
backwards-compatibility, but if we're going to break things we might as
well do it properly :o) I'm happy to go with just having two separate
functions if you are.

Cheers,

Mark

-- 
Mark Triggs
<[EMAIL PROTECTED]>

Reply via email to