Controlling an external program with Python

[Paul Duffin]

Donn Cave wrote:
> 
> Quoth claird at starbase.neosoft.com (Cameron Laird):
> | In article <395B32D1.B926FDB3 at hursley.ibm.com>,
> | Paul Duffin  <pduffin at hursley.ibm.com> wrote:
> ...
> |> This sort of arrangement is only possible when people who are interested in
> |> Expect functionality can transcend any language *issues*, by accepting that
> |> the language is basically irrelevant apart from the effect in may have on
> |> the required functionality and work together.
> ...
> | I believe I've heard Don express surprise at Expect's
> | career.  He thought people would steal the ideas for
> | other (languages') libraries.  What he least expected
> | (sic), if I understand it correctly, was that the whole
> | world would be content to invoke Expect as the standalone
> | executable which we know utterly dominates its use.
> 
> Well, I'm surprised he's surprised!  Expect-Tcl works out
> so well that it more or less cancels out the ``wrong language''
> problem.  Mainly because Tcl (as I suppose Lisp/Scheme) can
> support an "expect" statement in a way that doesn't fit in
> Python (that I can think of - this is not a comment on the
> existing Python expects, with which I am not familiar.)
> 
> And then I have looked at the source.  I'm sure it has no peer
> when it comes to comprehensive cross-platform pseudotty support,
> but a lot of barnacles come with that.  I would steal a younger
> horse than this.
> 

You do like mixing metaphors, first it is nautical, then it is criminal,
anyone would think you were a pirate ;-).

The trouble is that any abstraction which does not deal with all those
barnacles will have limited use. There are many people out there using
Expect on older platforms who would not be pleased to lose support.