feat: add stop words / repository context support for C/C++

[Squadrick]

Requires:

• Add TAGS_QUERY to rust bindings tree-sitter/tree-sitter-c#175
• Add TAGS_QUERY to rust bindings tree-sitter/tree-sitter-cpp#233

[Squadrick]

@wsxiaoys The PRs to exposing TAGS_QUERY (above) have been merged, but there wasn't been a new release tag with those commits. I've raised a request for new tags with the maintainers of those repos. In the meanwhile, I can use cargo's rev property to pick the code, but imo, it's a bit ugly.

So instead of:

tree-sitter-c = "0.20.3"
tree-sitter-cpp = "0.20.3"

we'll have:

tree-sitter-c = { git = "https://github.com/tree-sitter/tree-sitter-c/", rev = "4add3bb" }
tree-sitter-cpp = { git = "https://github.com/tree-sitter/tree-sitter-cpp", rev = "d153fe1" }

What do you think? Go with rev or wait for release tags?

[wsxiaoys]

Go with rev is ok - but you might also considering simply embeds the query scm in https://github.com/TabbyML/tabby/tree/main/crates/tabby-scheduler/queries

Like what we did for go / tsx

[Squadrick]

@wsxiaoys Since the queries were not modified, I went with the rev approach. If you would prefer that we stick to having the queries in-repo, let me know and I'll make that changes.

Squashed and pushed the final commit, can you please take a look?

[wsxiaoys]

Could you add a few verification case (through screenshot) in treesitter's query playground" https://tree-sitter.github.io/tree-sitter/playground, like what we did in #813 (comment)

[Squadrick]

@wsxiaoys Sorry about the delay, here's the verification for C++:

Source code:

struct A {
	int a;
    void method();
};
class B {
	void method();
};
void B::method() {}
union C {
  A a; B b;
};
void printTest(int a, std::string b) {}
typedef struct Person {
	std::string s;
} p;

Query:

(struct_specifier name: (type_identifier) @name body:(_)) @definition.class
(union_specifier name: (type_identifier) @name) @definition.class
(function_declarator declarator: (identifier) @name) @definition.function
(function_declarator declarator: (field_identifier) @name) @definition.function
(function_declarator declarator: (qualified_identifier scope: (namespace_identifier) @local.scope name: (identifier) @name)) @definition.method
(type_definition declarator: (type_identifier) @name) @definition.type
(enum_specifier name: (type_identifier) @name) @definition.type
(class_specifier name: (type_identifier) @name) @definition.class

Image:

[Squadrick]

For C:

struct A {
	int a;
    void method();
};
void printTest(int a, std::string b) {}
typedef struct Person {
	std::string s;
} p;

Query:

(struct_specifier name: (type_identifier) @name body:(_)) @definition.class
(declaration type: (union_specifier name: (type_identifier) @name)) @definition.class
(function_declarator declarator: (identifier) @name) @definition.function
(type_definition declarator: (type_identifier) @name) @definition.type
(enum_specifier name: (type_identifier) @name) @definition.type

Image:

[wsxiaoys]

Otherwise LGTM, thanks for the effort!

[Squadrick]

@wsxiaoys Done, should be good to merge now. Also squashed the commits since I had to change the author to my personal email. Should be good to go now. Thanks for the review.