This still doesn't work quite right, because some common operations like array
indexing lay completely outside the scope/symbol lookup system - they are not
even magics.
This also provides the initial steps towards support for type class "filtered" type inference
fixes an "ordinal type expected" ICE, related to the use of static params
expressions such as Type.field are now recognised by the compiler.
This also fixes a bug, preventing the user-defined to check for the presence of
regular fields in addition to procs
* replace tfAny and tfAll with tyAnd and tyOr
* integrate matchTypeClass into typeRel
* introduce tyBuiltInTypeClass to handle types such as tuple, object, proc, etc
This introduces tyStatic and successfully bootstraps and handles
few simple test cases. Static params within macros are no longer
treated as PNimrodNodes - they are now equivalent to constants
of the designated type.
This commit requires further investigation.
Fixing semExpr(nkCall) once and for all to work with sym-choices and to
allow overloading with immediate templates and macros will probably set
this straight too.
This postpones the semantic pass over the generic's body until
the generic is instantiated. There are several pros and cons for
this method and the capabilities that it enables may still be possible
in the old framework if we teach it a few new trick. Such an attempt
will follow in the next commits.
pros:
1) It allows macros to be expanded during generic instantiation that
will provide the body of the generic. See ``tmacrogenerics``.
2) The instantiation code is dramatically simplified. Dealing with unknown
types in the generic's body pre-pass requires a lot of hacky code and error
silencing in semTypeNode. See ``tgenericshardcases``.
cons:
1) There is a performance penalty of roughly 5% when bootstrapping.
2) Certain errors that used to be detected in the previous pre-pass won't
be detected with the new scheme until instantiation.
