unified: use a vendored-in copy of tree-sitter-swift

[tausbn]

For ease of iteration on the prototype.

[Copilot]

Pull request overview

This PR vendors the tree-sitter-swift grammar into unified/extractor to make iterating on the unified extractor’s Swift parsing prototype easier, and updates Rust/Bazel wiring to use the in-repo copy instead of the crates.io package.

Changes:

• Add a vendored unified/extractor/tree-sitter-swift crate including generated parser sources, scanner, queries, and build scripts.
• Switch unified/extractor from the registry tree-sitter-swift dependency to a local path dependency and add the new crate to the workspace.
• Update Bazel third-party wiring to remove the crates.io tree-sitter-swift archive and add required deps (cc, tree-sitter-language) for the new local crate.

Show a summary per file

File	Description
unified/extractor/tree-sitter-swift/tree-sitter.json	Adds tree-sitter grammar metadata/config for Swift.
unified/extractor/tree-sitter-swift/src/tree_sitter/parser.h	Vendors tree-sitter parser API header used by generated sources/scanner.
unified/extractor/tree-sitter-swift/src/tree_sitter/array.h	Vendors tree-sitter internal array utilities used by generated sources.
unified/extractor/tree-sitter-swift/src/tree_sitter/alloc.h	Vendors tree-sitter allocator abstraction header.
unified/extractor/tree-sitter-swift/src/scanner.c	Adds Swift external scanner implementation (comments/raw strings/semi handling/etc.).
unified/extractor/tree-sitter-swift/README.md	Vendors upstream README for the grammar.
unified/extractor/tree-sitter-swift/queries/textobjects.scm	Adds textobject queries for Swift.
unified/extractor/tree-sitter-swift/queries/tags.scm	Adds tags queries for symbol definitions.
unified/extractor/tree-sitter-swift/queries/outline.scm	Adds outline queries for structure extraction.
unified/extractor/tree-sitter-swift/queries/locals.scm	Adds locals queries (definitions/scopes).
unified/extractor/tree-sitter-swift/queries/injections.scm	Adds injection queries (regex/comment injections).
unified/extractor/tree-sitter-swift/queries/indents.scm	Adds indentation queries.
unified/extractor/tree-sitter-swift/queries/highlights.scm	Adds syntax highlighting queries.
unified/extractor/tree-sitter-swift/queries/folds.scm	Adds folding queries.
unified/extractor/tree-sitter-swift/package.json	Vendors upstream Node package metadata for the grammar.
unified/extractor/tree-sitter-swift/LICENSE	Adds upstream MIT license for the vendored grammar.
unified/extractor/tree-sitter-swift/grammar.js	Vendors the Swift grammar definition.
unified/extractor/tree-sitter-swift/Cargo.toml	Adds a local Rust crate wrapper for the vendored Swift grammar.
unified/extractor/tree-sitter-swift/BUILD.bazel	Adds Bazel rules to build the vendored grammar as a Rust library.
unified/extractor/tree-sitter-swift/bindings/rust/lib.rs	Provides LanguageFn and embeds node-types/queries; includes basic tests.
unified/extractor/tree-sitter-swift/bindings/rust/build.rs	Builds parser.c + scanner.c via cc during Rust builds.
unified/extractor/tree-sitter-swift/bindings/node/index.js	Vendors Node binding loader.
unified/extractor/tree-sitter-swift/bindings/node/binding.cc	Vendors Node binding implementation exporting the language.
unified/extractor/tree-sitter-swift/binding.gyp	Vendors Node-gyp build configuration for the Node binding.
unified/extractor/Cargo.toml	Switches tree-sitter-swift dependency to local path.
unified/extractor/BUILD.bazel	Adds the new local tree-sitter-swift Bazel target as a dependency.
MODULE.bazel	Adds Bazel module repos for cc and tree-sitter-language; removes crates.io tree-sitter-swift repo.
misc/bazel/3rdparty/tree_sitter_extractors_deps/defs.bzl	Removes vendored crates.io tree-sitter-swift archive; adds mappings for the new local crate + its deps.
misc/bazel/3rdparty/tree_sitter_extractors_deps/BUILD.tree-sitter-swift-0.7.2.bazel	Deletes the autogenerated BUILD file for the removed crates.io tree-sitter-swift dependency.
misc/bazel/3rdparty/tree_sitter_extractors_deps/BUILD.bazel	Adds aliases for cc and tree-sitter-language.
Cargo.toml	Adds the vendored tree-sitter-swift crate as a workspace member.
Cargo.lock	Converts tree-sitter-swift from registry source to a workspace package entry (removes source/checksum).

Copilot's findings

Comments suppressed due to low confidence (1)

unified/extractor/tree-sitter-swift/Cargo.toml:22

• This crate’s Rust tests and doctest example reference the tree_sitter crate (tree_sitter::Parser), but Cargo.toml doesn’t declare a tree-sitter dependency (and it can’t be used transitively). Add an explicit tree-sitter dependency (or at least a dev-dependency) so cargo test/doctests compile.

# When updating these dependencies, run `misc/bazel/3rdparty/update_cargo_deps.sh`
[dependencies]
tree-sitter-language = "0.1"
[build-dependencies]
cc = "1.2"

• Files reviewed: 31/35 changed files
• Comments generated: 1

[asgerf]

This adds 600k lines of code, of which 550k comes from the auto-generated parser.cc and 30k from the generated node-types.json.

I think we should go against tree-sitter conventions and avoid checking it in these generated artifacts, and instead rely on Bazel rules to rebuild when needed. WDYT?

[tausbn]

This adds 600k lines of code, of which 550k comes from the auto-generated parser.cc and 30k from the generated node-types.json.

I think we should go against tree-sitter conventions and avoid checking it in these generated artifacts, and instead rely on Bazel rules to rebuild when needed. WDYT?

Good idea. I'll try to set it up.

[tausbn]

Rerun has been triggered: 2 restarted 🚀

• Swift Integration tests Windows
• Java Checks Linux