Re: Extensible infix operators

[John Hind <john.hind@...>]

Thanks - I clearly should look into this stuff, it looks really powerful.

However at a formal (rather than implementation) level would using an 
unadorned token really cause ambiguities? It would be neater without a 
special syntax and it should surely be possible to disambiguate this 
given that the compiler knows that the token refers to a function (or 
not). I'd kind of like to add "mathematical" convention multiplication 
too i.e. a b is interpreted as a * b.


On 11/29/06, *Fabien* <fleutot+lua@gmail.com 
<mailto:fleutot+lua@gmail.com>> wrote:

   On 11/29/06, *John Hind* <john.hind@zen.co.uk
   <mailto:john.hind@zen.co.uk>> wrote:

       I'd like to be able to offer extensible infix operators so
       you'd be able to write something like:

       c = a func b
       This would be evaluated as:
       c = func(a, b)


   I really like Haskell's solution, which allows to put arbitrary
   functions in infix position, if you put it between backquotes, as
   in: "a `func` b". Unfortunately, it clashes with my syntax for
   algebraic datatypes


On second though, it  doesn't clash. It's actually a straightforward 
extension to implement:

-{ mlp.expr.add_postfix_sequence{ "`", mlp.id <http://mlp.id>, "`", 
mlp.expr,
    prec = 30, assoc = "left", builder = |x| `Call{ x[2], x[1], x[3] } } }

But there's still an issue, it relies on a partially unimplemented 
feature of the grammar generator: precedence and associativity hints 
aren't handled properly by postfix sequences, so the extension above 
will behave as if it had infinite precedence and were right associative. 
But that will be fixed eventually, so this infix call extension will work.