This is the mail archive of the
guile@sources.redhat.com
mailing list for the Guile project.
Re: Will guile ever have (or care about having) users?
- To: jl_daschbach at pnl dot gov
- Subject: Re: Will guile ever have (or care about having) users?
- From: bernard URBAN <Bernard dot Urban at meteo dot fr>
- Date: 20 Jul 2000 11:17:07 +0200
- Cc: guile at sourceware dot cygnus dot com, Jost Boekemeier <jostobfe at calvados dot zrz dot TU-Berlin dot DE>
- References: <200007191649.MAA04553@wd29688.emsl.pnl.gov>
>>>>> "John" == John Daschbach <d3h486@wd29688.emsl.pnl.gov> writes:
[...]
John> In the current guile (1.4) a scheme function written in C,
John> called from another C function via gh_lookup() will not be
John> replaced by a user redefinition from the repl (guile-user),
John> yet a direct call to this function will be redefined.
This behaviour is obviously annnoying.
Nevertheless, I understand it that way:
If you stick with gh_* without gh_repl, you are extending your
chemical package with guile.
If you use gh_* WITH gh_repl, you are in fact extending guile with
your chemical package (which could then best define its primitives
through a guile module, and its data structures through smobs).
I am not sure this second use of guile was really intended when the
gh_ interface was designed, hence the difficulty you encounter.
And do not forget: the executable guile is only a specific use of
libguile, ant the use of guile can itself be interactive (with repl)
or not (scripts).
I am perfectly happy that guile allows all these possibilities
(I do not know of a libperl), but
you must be aware of what you really want.
Regards.
--
B. Urban