Re: [Gnu-arch-users] Nit

[Miles Bader]

On Sat, Oct 18, 2003 at 07:34:01PM -0400, Colin Walters wrote:
> True...if the extension "interface" isn't really just an embedded
> programming language.  That's all I am arguing against.  

I'm not sure it makes much difference, compared to an
`extension-language-agnostic interface.'

If, say, Tom sticks a small scheme interpreter in tla* and says `OK that's
the extension language,' then all the tla internal hooks would say
`do_scheme_call (blah, blah, blah)' instead of `do_extension_language_call
(blah, blah, blah)', and if you want to hook up your python script or
something, you need a scheme->python interface layer.

If the embedded scheme interpreter has a sufficiently powerful alien
interface, then you could probably write this interface layer in scheme,
which is arguably an _advantage_.  If it doesn't -- which seems likely for a
small extension lanugage -- then you've probably got to write your
scheme->python interface layer in C, translating scheme's internal
representations to python's.  I'd say that it's likely that this process
would be essentially the same as writing a agnostic->python interface,
especially if the scheme interpreter uses C-friendly internal representations
(which is commonly the case for small embeddable languages).

IOW, the amount of effort required to interface python scripts seems more or
less the same.

* Note: I'm not advocating this or saying it's likely.

-Miles
-- 
I'm beginning to think that life is just one long Yoko Ono album; no rhyme
or reason, just a lot of incoherent shrieks and then it's over.  --Ian Wolff