From f07e6973fedb2fed4ff7f0e1a9d166bea7bedcab Mon Sep 17 00:00:00 2001
From: Nik Revenco <154856872+NikitaRevenco@users.noreply.github.com>
Date: Wed, 14 May 2025 19:41:18 +0100
Subject: [PATCH] fix: do not consider languages that do not have comment
 tokens

---
 Cargo.lock                |  4 ----
 Cargo.toml                |  2 +-
 helix-core/src/comment.rs | 32 ++++++++++++++++++++------------
 helix-core/src/syntax.rs  |  6 +++++-
 4 files changed, 26 insertions(+), 18 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 3a52e24df..214bc1ee4 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -2801,8 +2801,6 @@ checksum = "bfb942dfe1d8e29a7ee7fcbde5bd2b9a25fb89aa70caea2eba3bee836ff41076"
 [[package]]
 name = "tree-house"
 version = "0.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "803311306ba3279e87699f7fa16ea18fbcc8889d0ff0c20dc0652317f8b58117"
 dependencies = [
  "arc-swap",
  "hashbrown 0.15.2",
@@ -2818,8 +2816,6 @@ dependencies = [
 [[package]]
 name = "tree-house-bindings"
 version = "0.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d2f6894df414648c56f1f5b129830447140ff1017867773694ba882d093aa140"
 dependencies = [
  "cc",
  "libloading",
diff --git a/Cargo.toml b/Cargo.toml
index 1daf1b03c..113946c4f 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -37,7 +37,7 @@ package.helix-tui.opt-level = 2
 package.helix-term.opt-level = 2
 
 [workspace.dependencies]
-tree-house = { version = "0.1.0", default-features = false }
+tree-house = { path = "../../projects/tree-house/highlighter", default-features = false }
 nucleo = "0.5.0"
 slotmap = "1.0.7"
 thiserror = "2.0"
diff --git a/helix-core/src/comment.rs b/helix-core/src/comment.rs
index ee5751473..0c0c374aa 100644
--- a/helix-core/src/comment.rs
+++ b/helix-core/src/comment.rs
@@ -54,19 +54,27 @@ pub fn get_injected_tokens(
 ) -> (Option<Vec<String>>, Option<Vec<BlockCommentToken>>) {
     // Find the injection with the most tightly encompassing range.
     syntax
-        .map(|syntax| {
-            let config = loader
-                .language(
-                    syntax
-                        .layer(syntax.layer_for_byte_range(start, end))
-                        .language,
-                )
-                .config();
+        .and_then(|syntax| {
+            syntax
+                .layers_for_byte_range(start, end)
+                .into_iter()
+                .rev()
+                .find_map(|layer| {
+                    let lang_config = loader.language(syntax.layer(layer).language).config();
 
-            (
-                config.comment_tokens.clone(),
-                config.block_comment_tokens.clone(),
-            )
+                    let has_any_comment_tokens = lang_config.comment_tokens.is_some()
+                        || lang_config.block_comment_tokens.is_some();
+
+                    // if the language does not have any comment tokens, it does not make
+                    // any sense to consider it.
+                    //
+                    // This includes languages such as comment, jsdoc and regex: These
+                    // languages are injected and never found in files by themselves
+                    has_any_comment_tokens.then_some((
+                        lang_config.comment_tokens.clone(),
+                        lang_config.block_comment_tokens.clone(),
+                    ))
+                })
         })
         .unwrap_or_default()
 }
diff --git a/helix-core/src/syntax.rs b/helix-core/src/syntax.rs
index e232ee69b..da408cd82 100644
--- a/helix-core/src/syntax.rs
+++ b/helix-core/src/syntax.rs
@@ -412,7 +412,7 @@ impl FileTypeGlobMatcher {
 
 #[derive(Debug)]
 pub struct Syntax {
-    inner: tree_house::Syntax,
+    pub inner: tree_house::Syntax,
 }
 
 const PARSE_TIMEOUT: Duration = Duration::from_millis(500); // half a second is pretty generous
@@ -450,6 +450,10 @@ impl Syntax {
         self.inner.layer_for_byte_range(start, end)
     }
 
+    pub fn layers_for_byte_range(&self, start: u32, end: u32) -> Vec<Layer> {
+        self.inner.layers_for_byte_range(start, end)
+    }
+
     pub fn root_language(&self) -> Language {
         self.layer(self.root_layer()).language
     }