[PATCH] D90543: [Syntax] Start to move trivial Node class definitions to TableGen. NFC

Wed Nov 4 09:25:27 PST 2020

eduucaldas added a comment.

> Compared to Dmitri's prototype, Nodes.td looks more like a class hierarchy and
> less like a grammar. (E.g. variants list the Alternatives parent rather than
> vice versa).

> e.g. we may introduce abstract bases like "loop" that the grammar doesn't care about in order to model is-a concepts that might make refactorings more expressive. This is less natural in a grammar-like idiom.

Do you have a concrete example of such abstract base -- loop is in the grammar <https://eel.is/c++draft/stmt.iter#nt:iteration-statement> ? And in the case such an example appear, in my mind we would just change "our grammar" to have an alternative for this "loop" abstract base.

> e.g. we're likely to have to model some alternatives as variants and others as class hierarchies, the choice will probably be based on natural is-a relationships.

I agree, alternatives and the two ways to model them are a tricky subject...

> it reduces the cognitive load of switching from editing *.td to working with code that uses the generated classes

I think we should consider reading prior to editing.
A grammar is how we represent syntax, and syntax trees model syntax, as such I think we should rather consider the cognitive load of switching between the grammar and the definition of syntax trees.

I have a lot to improve in my writing skills, so I don't know if I could express all my thoughts. We could have a call if you want ;)

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D90543/new/

https://reviews.llvm.org/D90543