[wg-camlp4] Changes to the parsetree

[Alain Frisch]

On 03/25/2013 03:13 PM, jathdr wrote:
> Could we get a summary of the current syntax proposal? I've been
> following along, but I'm a bit lost as to what has been discussed,
> accepted, rejected, etc.

This file describes the current syntax proposal:

http://caml.inria.fr/cgi-bin/viewvc.cgi/ocaml/branches/extension_points/experimental/frisch/extension_points.txt?revision=HEAD&view=markup


> This is probably a bit off topic, but while we're talking about
> interval patterns: I discovered the other day that interval patterns
> only work for chars, while I always assumed they also worked for
> ints. Is there a reason why they don't?

Currently, range patterns for characters are expanded in the parser. 
Doing so for integers as well could lead to huge or-patterns to be 
processed by the rest of the compiler.  Supporting range patterns 
directly in the compiler would add a little bit of extra complexity for 
type-checking (exhaustiveness check) and code generation.  I don't see 
any theoretical problem, though.


-- Alain