strongly typed AST

[UniqueFool]

I actually renamed the other question, but I was looking at the verbosity levels because I was really trying to get a dump of the strongly-typed ASTs to see how everything hangs together, could you maybe provide a pointer how to do that ?
I realize that there are different kinds of ASTs depending on the stage of the solver - I am mainly interested in the BFS implementation that traverses the AST to compare candidate trees.

In particular, I would like to learn how new objects/types are represented at the AST level.

Thank you

[UniqueFool]

This is admittedly a little offtopic, but how difficult do you think would it be to use inductive for creating a different syntax tree format for a subset/dialect of ECMAScript and how would you approach doing something like that ?
Or rather, would you instead suggest manually converting the generated JavaScript code by post-processing the generated code ?

Actually, I am currently doing the latter using a sed script (which works well enough), so this is more of a hypothetical question, mainly inspired by looking at your template/macro system and the way you described it would work internally, i.e. wondering if it could be directly told to use a different format for certain expressions of the AST

In the context of JavaScript/ECMAScript6, the point could be using default arguments or making use of other new features (think for/let loops), but also the other idea of using named function arguments. So this isn't totally unrelated to the current use-case.

Speaking more generally: how difficult would it be to literally use the macro/template system in conjunction with the type system to make use of more recent/experimental JavaScript/node features ?

Thank you

[omphalos]

I think this is probably doable in one of two ways. Either use something like babel-register with a babel plugin that supports eval. I'm not sure if one exists for arbitrary strings. It's possible to write one.

Another option: eval in the codebase is handled by the evalUtil file. The important use of this is in Specification. I could expose an option that lets the consumer override the evalUtil call with a custom implementation if they want. At that point it would be up to the consumer to pass a function that calls into the babel parser API (or their parser of choice).

[UniqueFool]

thanks, that should suffice as a pointer so that I can have a look.