Closed Kyuuhachi closed 2 months ago
I think this is all right as currently implemented, with .span()=>
required if the desired span is the span of an AST node.
In my own macros I prefer avoiding Spanned
, so adding a new implicit application of it would be counterproductive.
In almost all cases I find myself using the span of an existing AST node, so needing to write
.span()=>
everywhere is a lot of visual noise. And sinceSpan
isimpl Spanned
, I believe this would be backward compatible (though there could very well be some sort of special case I have not thought of). Would this be a reasonable change, or is there some reason it is not so?