Re: Ellipsis in the _pattern_ Taylor Campbell 12 Oct 2003 13:54 UTC

On Sunday, Oct 12, 2003, at 00:42 US/Eastern, bear wrote:

> You know, if it doesn't make sense to you it probably means the thing
> I'm
> thinking about is a nonissue anyway.  As I said, I haven't eaten and
> breathed
> enough macrology to fully understand the intent of ::: in the first
> place.
> I just have the impression that if we actually need it,

Well, we might not.  I'm not sure, but it seems useful to me when
processing
SYNTAX-RULES macros with SYNTAX-RULES macros (e.g., to implement
SYNTAX-RULES).

>                                                         then something
> might
> break down if we want to go to meta-macros or meta-meta macros and
> don't
> have a corresponding "nested" form or the ability to produce/match it
> in
> the inputs to our higher-order macros.

Perhaps it would be good if I specify further: ::: in the template
literally generates the identifier ':::', because there's nothing
special it would do,
until the macro gets expanded into another macro whose pattern contains
:::.

> Still, the reason I doubt its actual utility is that I don't think
> anything
> prevents ...  from matching a "sequence" of length one whose only
> member is
> another ellipsis.

Yes, ... might match a sequence containing ellipsis, but still I want
to be
able to literally match ellipsis: I want to be _sure_ of the fact that
what I
am matching is ellipsis.

> Maybe I should get embarassed about my ignorance here and shut up.

Better yet, write some amazingly complex CPS macros.  (No cheating by
using
Andre van Tonder's monadic CPS macros!)  You'll learn quite a lot of
macrology
by doing so, and also several pitfalls of SYNTAX-RULES.

> 				Bear