Allow accessing node parents in visitor function

[yxliang01]

Since the AST format of this library is customized, I feel exporting the function _isASTNode is generally favored as it guarantees that when the format of output of this library changes, this check will/should always be up-to-date.

@federicobond do you agree?

[federicobond]

Hi @yxliang01, can you describe some use cases for this function? I'm not sure I understand well how it would be used.

[yxliang01]

@federicobond Like if you want to write a custom traverser of the AST tree, this function can provide information like whether an attribute of a node is yet another AST node (This is actually my case). While the current definition of this function is very simple, it's really useful :) and I feel it's a good idea to export this to provide some encapsulations.

[federicobond]

I understand your concern. However, I would suggest to just check if the node is an object and has a type key for now, as that should be considered part of the stable interface of this parser. If any other issues come up related to this, I am willing to reconsider that decision.

[federicobond]

I think we can take this opportunity to introduce a better traverser. Please let me know what constraints are driving your custom traverser implementation, as we might be able to do something on that front.

[yxliang01]

@federicobond I basically need the object path of each AST node in the AST tree. So that, I can directly visit the previous node visited node by the object path. I deal with it by adapting the visit myself, so that this information is tracked and passed to the visitor. I think this usage might be specific though.

[federicobond]

That is a pretty common usage. The format of the AST is esprima-compatible, so preprocessing it with something like ast-parents should work well too.

[yxliang01]

I see. Thanks for pointing this. But, this is quite inefficient given that need to traverse the AST for once more and increases memory space by quite much I feel.
So, do you mean that it's a good idea to pass the object path information to visitors by commenting it's a common usage? Or?

[yxliang01]

@federicobond any suggestion for what should be done for this issue? Otherwise, this will be stale :(

[federicobond]

Hey! Sorry, my last few days have been pretty hectic. I think the lib could implement this functionality by passing a second argument to each visit call, with the appropiate parent chain. Would be happy to review a pull request to that end.

As for checking whether something is an AST node, I think you can safely rely on a check for an object with a type property.

[yxliang01]

@federicobond I am also happy to contribute to that :) . Will do it bit latter after I have done my current important tasks.

[yxliang01]

To be sure, @federicobond did you mean passing the parent node object to the visitor? Or, the object path for the whole AST(my need)?

[federicobond]

Yes, passing the full object path makes a lot of sense.