Rule-parameter mappings

[szschaler]

Allow rule parameters to be explicitly mapped and woven. Resolves #22.

Tasks

• [x] Introduce mapping syntax
  • [x] Basic syntax
  • [x] Scoping
• [x] Include in formatter
• [x] Enhance morphism conversion to cover rule parameters
  • [x] For actual rules
  • [x] For virtual rules
• [x] Enhance morphism checks to cover rule-parameter mappings
  • [x] Validity of parameter mappings
  • [x] All parameters mapped
  • [x] Parameter types are compatible
  • [x] Parameter kinds are compatible
  • [x] Consistency with node mappings where the parameters are parameter nodes
  • [x] Validity of attribute mappings where expressions use the rule parameters
• [x] Enhance auto-completion to cover rule parameters
• [x] Enhance weaving to cover rule parameters
  • [x] In cases where attributes are mapped both ways
  • [x] In cases where the attribute is introduced via the weaving (so there isn't a parameter mapping), but may still require renaming in the expression.
• [x] Clarify how we need to handle interface_of here I would suggest that a parameter is considered part of the interface if either its type is an EClass and is in the interface or there are slots who are in the interface and whose expressions reference the parameter.
• [x] Update documentation

[szschaler]

@narudocap can you have a play with this and provide feedback via the Github review mechanism?

[narudocap]

I've created a simple example (pls v12) but doesn't seem to work, could you check whether it's a problem in the implementation or in the example? Perhaps it is not what I should be testing. A description is in the README file. https://github.com/gts-morpher/pls/tree/master/pls12

[szschaler]

There were some typos in the assemble rule. I've just committed a fix. Now there are no problems and parameters seem to be woven correctly.

[narudocap]

Here is an additional test case with different combinations of parameters: https://github.com/gts-morpher/pls/tree/master/pls13

[szschaler]

This one is interesting as it exposes a remaining issue with name uniqueness. I think this is related to #41 .