[Grace-core] Some comments from Tijs on parsing Grace

[Michael Homer]

On Wed, Jul 27, 2016 at 11:48 PM, James Noble <kjx at ecs.vuw.ac.nz> wrote:
>
>> On 27/07/2016, at 23:01PM, Tijs van der Storm <storm at cwi.nl> wrote:
>>
>> Cool! But wouldn't this be a post-fix operator? For true circumfix, the inside must be the receiver, no?
>
> I think they're resolved only in lexical scope: there's no obvious receiver, especailly for [ ]
Yes, it treats circumfix calls as implicit-receiver requests to the
method "circumfix[]" (or "circumfix[{$[]$}]", ...), which will be
found in a surrounding object.

There were once postcircumfix methods also, but they were removed to
make way for [ ... ] used as a singleton argument. Circumfix is really
only still there to let me introduce support for [ ... ] with the
minimum required change.

It's true that arguably by analogy with "prefix" these aren't really
circumfix, but "infix" seems left out at that point.
-Michael