allow interpolating vars in `message` for all constraints

[VladimirAlexiev]

(Related to #173, inspired by #158)
https://www.w3.org/TR/shacl/#x5.3.2-mapping-of-solution-bindings-to-result-properties specifies how {?vars} are interpolated into the message, but only for SPARQL constraints.

It would be useful to allow this for all shapes. Why should standard constraint components lack this feature?

What vars can be used in interpolation?

• ?this
• ?value
• Constraint Component parameters as used in POTENTIAL DEFINITION IN SPARQL
  • eg for https://www.w3.org/TR/shacl/#MaxExclusiveConstraintComponent that's ?maxExclusive
  • but not all Constraint Components in the spec mention such parameters
  • I'm sure there is a comprehensive file that defines all Constraint Components and their params?

BTW, is it ok to use {$value} in sh:message, or should it be spelled as {?value}?

[HolgerKnublauch]

Both {$value} and {?value} can be used interchangeably.

[bertvannuffelen]

I support improvements for message handling, so that other languages have also access to the valuable information in their messages.

[VladimirAlexiev]

To support a case in SEMICeu/MLDCAT-AP#19:

sh:minCount, sh:maxCount should bind ?value to the actual number of occurrences found

If that is confusing (because ?value is supposed to be an actual resource found in the graph, not a count), then maybe a new variable ?_count should be introduced.
I used an underscore because the name ?count may well conflict with a variable that the shape author wanted to use for a different purpose