Re: [Gnu-arch-users] Extension language

[Tom Lord]

    > From: Colin Walters <address@hidden>

    > I think you're missing my point.  The best way is for tla/libarch to be
    > flexible enough to allow for many languages to easily wrap it.

It already is, plus or minus a few details of input and output formats.


    > If someone decides to write a FICL wrapper, then good.  But no language,
    > be it FICL, Python, Perl, Tom's language, or anything else should enjoy
    > some sort of privileged status.

Nah.

"Wrapping" tla, from _any_ langauge, into _any_ half-way decent
candidate for an extension language, starting from nothing more than
the existing code, is perhaps a few hours or couple of days of work.

And conversely, any language for which it would not be a few hours or
couple of days work is not even a half-way decent candidate.

It's not the "wrapping" that's tricky: it's building up "stuff" around
it to make a high-level programming environment that strikes a chord.
And that "stuff" is likely to be highly language specific.

I suspect that you're talking about "librification" of tla so that the
extension language and tla can all be in the same process.  It's
_perhaps_ an interesting direction to go but my first answer to that
idea is "nah, don't bother".  A (sub-)process is a pretty handy
abstraction -- just about what you want for the typical tla
subcommand.

-t