Repository navigation
Update tree-sitter and grammars to version 0.20 #674
Description
Activity
Based on my issues adding
tree-sitter-c-sharpI thought I would do a little look at the feasibility of rolling forward to 0.20.- tree-sitter-java (version = "0.19.0" | tree-sitter = "0.19" | PR)
- tree-sitter-typescript (version = "0.20.0" | tree-sitter = ">= 0.19, < 0.21")
- tree-sitter-javascript (version = "0.20.0" | tree-sitter = ">= 0.19, < 0.21")
- tree-sitter-python (version = "0.19.0" | tree-sitter = ">= 0.19, < 0.21" | PR)
- tree-sitter-rust (version = "0.20.1" | tree-sitter = "0.20" )
For c-sharp there is a PR mentioned in #789 by the maintainer.
The PRs are being ignored by the maintainers, we could use forks and specify dependencies as git repositories (https://doc.rust-lang.org/cargo/reference/specifying-dependencies.html#specifying-dependencies-from-git-repositories).
For example, for tree-sitter-python, we could specifytree-sitter-python = { git = "https://github.com/Luni-4/tree-sitter-python", branch = "update-0.20" }.- linked a pull request that will close this issueUpdate all dependencies to use tree-sitter v0.20 #812
on Mar 31, 2022 @marco-c I was looking at #812 and was wondering if we have a final action plan here?
That PR looks like it is uncovering some good bugs uncovered by the upgrade (or some that were there already?).I did notice Python points to its tree-sitter master and the other to a fork from @BrainBuzzer that is a work in progress.
In the merge, this would point to something more stable than master though.I do want to ask though, what does 0.20 give us? Were there important security fixes, perf, features, that the project benefited from? Couldn't we settle on 0.19 until the situation improves with dependencies? This only makes sense if we think the situation CAN be improved. If it isn't something any of us can influence, then the forking strategy is the way to go.
