julia-vscode / juliaworkspaces.jl Goto Github PK

At the moment JuliaSyntax produces SyntaxNode for us, and we use that to detect test items and friends. The question, though, is whether that is the right node type for us? My understanding is that it doesn't fully round-trip, which seems like something we need.

I guess one alternative would be to parse into CSTParser.EXPR instead, i.e. reuse our old data structure but use the parsing logic from JuliaSyntax. Not clear to me whether that is a good idea? Would probably be useful if those that have worked more with CSTParser chimed in on that question. Things that to me look not super ideal are: 1) I'd prefer it to be an immutable, 2) I dislike the meta field, I think we should move to a world where any state from the linter/semantic analysis is stored outside of the syntax tree, 3) the parent field also doesn't strike me as ideal, as it means that one can never reuse a child-subtree in a new tree, at least if we move to immutable, 4) having val as String is probably also not ideal? Doesn't that mean a lot of allocations that one could potentially avoid by just using say a SubString that is based on the full original doc, or something like that? 5) I think one of the design goals of EXPR was that it is somewhat similar to Expr with the idea that at some point it might become the official parser for Julia. I think that goal is no longer relevant, so I'm wondering whether one would actually design the data structure differently today if that goal is no longer?

So maybe we should create a new type?

Another consideration is how this plays with the JuliaFormatter types, my understanding is that it uses yet another node type internally, somehow? Maybe there is an opportunity to design one node type that can power both JuliaWorkspace and JuliaFormatter?

I think these just got copied in from LS.jl, but are not defined in this context.

LSInvalidFile
LSOffsetError
LSPositionToOffsetException

This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!