Re: [Chicken-users] Syntactic sugar for regular expressions and URIs

[Arto Bendiken]

On 6/27/07, John Cowan <address@hidden> wrote:

> If you support Perl-style trailing letters, then you should be sure
> to support "m" and "s" also.

Yes, that'd be useful. At present, though, I'm limited by the PCRE
options Chicken's `(regexp)' constructor exposes, and it only has 'i',
'x' and 'u'. I'll see about submitting a patch to the `regex' unit if
it looks like this can facilely be extended.

> > Next, the `uri-literals' egg [3] allows facile use of URIs delimited
> > by '<' and '>', a natural form already familiar from URL use in
> > plain-text media:
>
> I find this troubling, because although <...> are associated with URIs,
> "#<...> has been associated for many years in the Lisp community
> with non-rereadable syntax such as the "#<regex>" in your examples.
> (Common Lisp actually makes this the default; "#<" signals an error.)
> Loading this egg would cause "#<regex>" to be interpreted as a relative
> URL.  I think you should find some alternative syntax.

Yes, I expected this objection :-)

In the context of where I'm using these URI literals, there is no
danger of ambiguity, but in general use I concede that there could
well be.

That said, I did experiment with various alternative URI literal forms
but did not particularly fancy any that I could come up with. The
present one has proven to be most useful despite the overloading of
"#<". Better ideas are welcome, though - that's why I'm posting here.

--
Arto Bendiken | http://bendiken.net/