There was a small issue with how we were recording the definition references for declarations. We were counting the whole entry of the AST as its definition. For example, here:
state new_state {
...
}
the definition start was the start of the word "state" and the end was the closing curly bracket. This PR fixes the definition references to be the span of the identifier (in this example, just the new_state).
Benefits:
allows to distinguish between comma-separated declarations (which otherwise would have the same dsym value);
enables the language server to correctly identify the symbols.
There was a small issue with how we were recording the definition references for declarations. We were counting the whole entry of the AST as its definition. For example, here:
the definition start was the start of the word "state" and the end was the closing curly bracket. This PR fixes the definition references to be the span of the identifier (in this example, just the
new_state). Benefits: