Redo computation of keywords

Our approach relied too much on pecularities of the grammar file which
made this fragile. This patch rewrites keyword extraction to instead
inspect the generated parser which should be more stable against subtle
formatting changes in the grammar (but might still fail if tree-sitter
changes their parser generator).

crates/tree-sitter-zeek/Cargo.toml

@@ -11,6 +11,6 @@ tree-sitter = "0.24.4"
 
 [build-dependencies]
 cc = "1.1.37"
-regex = "1.11.1"
+regex = { version = "1.11.1", default-features = false }
 tree-sitter = "0.24.4"
 tree-sitter-generate = "0.24.4"

crates/tree-sitter-zeek/build.rs

@@ -8,19 +8,17 @@ use std::{
     path::{Path, PathBuf},
 };
 
-fn generate_keywords(grammar: &Path) {
-    let file = File::open(grammar).unwrap();
+fn generate_keywords(parser_c: &Path) {
+    let file = File::open(parser_c).unwrap();
     let mut buf_reader = BufReader::new(file);
     let mut contents = String::new();
     buf_reader.read_to_string(&mut contents).unwrap();
 
     let mut set: HashSet<String> = HashSet::new();
-    let re = Regex::new(r"'(@?&?[a-z]+-?_?[a-z]+)'").unwrap();
-    for cap in re.captures_iter(&contents) {
-        if cap[1].eq("zeek") || cap[1].eq("extras") {
-            continue;
-        }
-        set.insert(cap[1].to_string());
+    let re = Regex::new(r#"\[anon_sym_.*\] = "(.*)""#).unwrap();
+
+    for cs in re.captures_iter(&contents) {
+        set.insert(cs[1].replace("\\?", "?"));
     }
 
     let out_dir = env::var_os("OUT_DIR").unwrap();
@@ -62,11 +60,13 @@ fn main() {
     // Compile tree-sitter C output.
     let src_dir = out_dir.join("src");
 
+    let parser_c = src_dir.join("parser.c");
+
     Build::new()
-        .file(src_dir.join("parser.c"))
+        .file(&parser_c)
         .include(out_dir.join("src"))
         .warnings(false)
         .compile("tree-sitter-zeek");
 
-    generate_keywords(&grammar);
+    generate_keywords(&parser_c);
 }

crates/tree-sitter-zeek/src/lib.rs

@@ -55,4 +55,4 @@ pub const HIGHLIGHT_QUERY: &str = include_str!("../vendor/tree-sitter-zeek/queri
 /// The content of the [`node-types.json`][] file for this grammar.
 ///
 /// [`node-types.json`]: https://tree-sitter.github.io/tree-sitter/using-parsers#static-node-types
-pub const ZEEK_NODE_TYPES: &str = include_str!("../vendor/tree-sitter-zeek/src/node-types.json");
+pub const ZEEK_NODE_TYPES: &str = include_str!(concat!(env!("OUT_DIR"), "/src/node-types.json"));

src/snapshots/zeek_language_server__complete__test__keyword.snap

@@ -1,8 +1,31 @@
 ---
 source: src/complete.rs
 expression: result
+snapshot_kind: text
 ---
 [
+    CompletionItem {
+        label: "F",
+        label_details: None,
+        kind: Some(
+            Keyword,
+        ),
+        detail: None,
+        documentation: None,
+        deprecated: None,
+        preselect: None,
+        sort_text: None,
+        filter_text: None,
+        insert_text: None,
+        insert_text_format: None,
+        insert_text_mode: None,
+        text_edit: None,
+        additional_text_edits: None,
+        command: None,
+        commit_characters: None,
+        data: None,
+        tags: None,
+    },
     CompletionItem {
         label: "fallthrough",
         label_details: None,