Re: srfi requests?

[Per Bothner <per at bothner dot com>]

Chris Dean wrote:
> I guess "all final SRFIs" is too much :-).  At the top of my list is:
>
> - srfi-13

I took a look at srfi-13.  A port isn't trivial.  The main
problem is the dependency on srfi-14, both in the api and
in the implementation.  And unfortunately a Unicode-correct
and efficient implementation of character sets doesn't seem
possible with duplicating the information in the tables that
Java uses internally.  E.g. how do you efficiently do
(char-set-fold k n char-set:letter)?  If I had access to the
Java internals, no problem.  As it is, it's difficult.
(Seems like a topic for a JSR.)

For srfi-13 we can probably manage with a subset of srfi-14.
But it's not a 1-hour job and probably not a 1-day job either.

Then there is performance tweaking.

> - and-let*

srfi-2 isn't a drop in either, since the implementation
uses non-portable define-macro.  I really want a syntax-rules
implementation if possible, and if not, a syntax-case implementation.
Does anyone have one, or want to write one?  I doubt it would be
very difficult, but I haven't looked at it yet.

> - args-fold

Looks straight-forward to incorporate.  We'd want to clean up
the way the option-type is made "privte" - we have better
machanisms in Kawa.
--
	--Per Bothner
per@bothner.com   http://per.bothner.com/