Re: [CASL] concrete syntax - mixfix, imports

["Bjarke Wedemeijer" <wedemeij@wins.uva.nl>]

Dear all,

Peter wrote the following:

>> There is a related issue that I'd like to check here: should the
> mixfix analysis take account of the particular number of argument
> sorts in non-mixfix declarations or not?  E.g., consider
>   ops f:s*s->s; f__:s->s; c:s 
> the question is whether f(c) has a unique mixfix analyis or not.  My
> personal opinion is that the non-mixfix declaration f:s*s->s should
> have the same effect as if one had declared f(__,__):s*s->s (note that
> parentheses and commas are not allowed in IDs, so one can't actually
> write such a mixfix declaration).  Thus f(c) is unambiguously analysed
> as an application of f__.  (I believe that Bjarke's prototype mixfix
> analyser implements my preference, but I haven't checked that yet.)
>
In my mixfix parser f__ has higher preference than the single f since
it has one placeholders instead of none,

cheers
bjarke