Re: PATCH: scheme-modules.texi

[Marius Vollmer]

Aaron VanDevender <address@hidden> writes:

> I noticed the documentation for load-extension was somewhat
> incomplete (non-existant) so I fleshed it out a bit. Someone
> might want to check it for accueracy before commiting it.

Thanks!  I see some problems for your text, however: we already have a
reference documentation for 'load-extension', it's docstring.  It has
not been included in the manual yet...  Second, I find your example
for scm_register_extension not very straightforward.  It's purpose is
not to overwrite a extension initialization function.  It might be
used that way as a kludge, but that was not the motivation for
providing it.

An extension is expected to be loaded only once via 'load-extension'.
When it wants to protect against being initialized twice, it can
simply use a global 'already_initialized' flag.

The problem you describe wuld be solved more elegantly by having the
(science electron) module simply use the (math bessel) module.  It
does not need to access the extension directly.

The purpose of scm_register_extension to allow 'load-extension' to
work even when dynamic linking is not available or not wanted.

So I'm afraid we can't really use your documentation... thanks a lot
anyway!

-- 
GPG: D5D4E405 - 2F9B BCCC 8527 692A 04E3  331E FAF8 226A D5D4 E405