From 28d86733712d60e620162e2a7ac3ee433e351153 Mon Sep 17 00:00:00 2001
From: Wilfred Hughes <me@wilfred.me.uk>
Date: Sat, 11 May 2024 16:09:25 -0700
Subject: [PATCH] Treat tree-sitter nodes highlighted as comments as atoms

Fixes #711
---
 CHANGELOG.md                    |  7 +++++++
 sample_files/compare.expected   |  2 +-
 sample_files/elm_1.elm          |  2 ++
 sample_files/elm_2.elm          |  2 ++
 src/parse/tree_sitter_parser.rs | 16 +++++++---------
 5 files changed, 19 insertions(+), 10 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index dc8ef066a2..62885967b8 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,5 +1,12 @@
 ## 0.59 (unreleased)
 
+### Parsing
+
+Difftastic now uses tree-sitter comment highlighing as a hint that
+nodes should be treated as atoms. This ensures comments are treated
+more consistently across languages. This fixes cases in Elm where
+comment differences were ignored, and may improve other languages too.
+
 ## 0.58 (released 11th May 2024)
 
 ### Parsing
diff --git a/sample_files/compare.expected b/sample_files/compare.expected
index 6c59c357ea..e80f8076b9 100644
--- a/sample_files/compare.expected
+++ b/sample_files/compare.expected
@@ -56,7 +56,7 @@ sample_files/elisp_contiguous_1.el sample_files/elisp_contiguous_2.el
 beaf7d6c5136d3db7a36ff49a131b251  -
 
 sample_files/elm_1.elm sample_files/elm_2.elm
-54d73b12de1571cfb9b27d3b2a7f4f62  -
+33b71893107538cff574276f2837adbb  -
 
 sample_files/elvish_1.elv sample_files/elvish_2.elv
 f80b47646e7dd2bd3a49393d00657465  -
diff --git a/sample_files/elm_1.elm b/sample_files/elm_1.elm
index 60d77a6df1..b8c855bc16 100644
--- a/sample_files/elm_1.elm
+++ b/sample_files/elm_1.elm
@@ -1,6 +1,8 @@
 module Main exposing (blue, green, list, x, y, z)
 
 
+{- bar
+ -}
 list : List Int
 list =
     [ 1, 2, 3 ]
diff --git a/sample_files/elm_2.elm b/sample_files/elm_2.elm
index fba7f65bb5..fdc04a117b 100644
--- a/sample_files/elm_2.elm
+++ b/sample_files/elm_2.elm
@@ -6,6 +6,8 @@ module Main exposing
     )
 
 
+{- foo
+ -}
 list : List Int
 list =
     [ 1
diff --git a/src/parse/tree_sitter_parser.rs b/src/parse/tree_sitter_parser.rs
index 28e21125df..0a4b2829ba 100644
--- a/src/parse/tree_sitter_parser.rs
+++ b/src/parse/tree_sitter_parser.rs
@@ -507,7 +507,6 @@ pub(crate) fn from_language(language: guess::Language) -> TreeSitterConfig {
                 atom_nodes: vec![
                     "doctype",
                     "quoted_attribute_value",
-                    "comment",
                     "raw_text",
                     "tag_name",
                     "text",
@@ -956,10 +955,7 @@ pub(crate) fn from_language(language: guess::Language) -> TreeSitterConfig {
             let language = unsafe { tree_sitter_scala() };
             TreeSitterConfig {
                 language,
-                // TODO: probably all comments should be treated as atoms
-                atom_nodes: vec!["string", "template_string", "comment", "block_comment"]
-                    .into_iter()
-                    .collect(),
+                atom_nodes: vec!["string", "template_string"].into_iter().collect(),
                 delimiter_tokens: vec![("{", "}"), ("(", ")"), ("[", "]")],
                 highlight_query: ts::Query::new(
                     language,
@@ -973,9 +969,8 @@ pub(crate) fn from_language(language: guess::Language) -> TreeSitterConfig {
             let language = unsafe { tree_sitter_scheme() };
             TreeSitterConfig {
                 language,
-                atom_nodes: vec!["block_comment", "comment", "string"]
-                    .into_iter()
-                    .collect(),
+                //
+                atom_nodes: vec!["string"].into_iter().collect(),
                 delimiter_tokens: vec![("{", "}"), ("(", ")"), ("[", "]")],
                 highlight_query: ts::Query::new(
                     language,
@@ -1615,9 +1610,12 @@ fn syntax_from_cursor<'a>(
         *error_count += 1;
     }
 
-    if config.atom_nodes.contains(node.kind()) {
+    if config.atom_nodes.contains(node.kind()) || highlights.comment_ids.contains(&node.id()) {
         // Treat nodes like string literals as atoms, regardless
         // of whether they have children.
+        //
+        // Also, if this node is highlighted as a comment, treat it as
+        // an atom unconditionally.
         atom_from_cursor(arena, src, nl_pos, cursor, highlights, ignore_comments)
     } else if node.child_count() > 0 {
         Some(list_from_cursor(