Syntax for absolute qualified identifiers - #1095
Conversation
Add isAbsolute argument to qualified identifiers Update call sites to match
Allow leading dot in qualified identifier
Add syntax tests
Add isAbsolute to ExprIdent Revise uses to match
Add tests for fpp-syntax and fpp-format
Add initial dots to qualified identifiers in state machines
sylvesterkaczmarek
left a comment
sylvesterkaczmarek
left a comment
There was a problem hiding this comment.
One integration concern with landing the syntax phase independently: the semantic visitors currently discard isAbsolute (Unqualified(name, _), ExprIdent(id, _)), so .a is accepted but can silently resolve exactly like relative a. Could absolute nodes be rejected until the semantics step lands, or should syntax and resolution merge atomically to avoid wrong-symbol binding?
|
This is a PR into a PR into a feature branch. The feature branch won't be merged into main until everything is implemented and tested. |
|
In working on the semantics, I realized that in the AST, we should store the |
2 participants
This PR represents the second step of #990.