Re: [patch] sawfish.wm.util.prompt

[matthew <matth love gmail com>]

Jeremy Hankins wrote:
> Matthew Love <matth love gmail com> writes:
>
> > Apparently, I forgot to update prompt-extras, which rename-window
> > calls, to the new definition of those functions from prompt.
> >
> > Updated now and all in working order :)
>
> First, it looks like prompt-for-file and prompt-for-directory need to be
> updated too -- if `existing' is false then `prompt-validation-fun' is
> nil.
>
> But I also have a concern about the original patch.  If I understand
> what's going on, instead of testing to see if there is, e.g., a
> validation/completion/whatever function provided, it actually calls it
> to see if it returns non-nil.  But there are times when calling the
> function may be expensive.  A common example might be accessing the
> filesystem, though caching may handle that case adequately, I don't
> know.  But I also have stuff that completes on menu entries, which could
> conceivably involve calling multiple other functions to expand submenus
> -- some of which (if I ever get it working right) may even involve
> communicating with other apps to fill in the menu.  Is there a way to
> silence the compiler warnings that doesn't involve extra calls to the
> prompt-*-fun functions?
>
> Sorry I didn't say something when the patch first passed through the
> list.

Hm, yes, this is all true, and I see what you mean, this is certainly 
not the best way to handle these functions in those modules...

The original code itself was fine really, other than the compilation 
warnings generated by rep.  I can't think of a solution for this at the 
moment, but will look into it more when I have a system in front of me :)

Cheers,

Matthew