From d5b1e26d708ea064897f528efa2c739d0373903b Mon Sep 17 00:00:00 2001
From: Wilfred Hughes <me@wilfred.me.uk>
Date: Thu, 14 Nov 2024 22:55:00 -0800
Subject: [PATCH] Add a debug helper for syntax tree as DOT

---
 src/main.rs         | 23 +++++++++++++++++++++++
 src/options.rs      | 22 ++++++++++++++++++++++
 src/parse/syntax.rs | 37 +++++++++++++++++++++++++++++++++++++
 3 files changed, 82 insertions(+)

diff --git a/src/main.rs b/src/main.rs
index 170df6897e..271b73c555 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -161,6 +161,29 @@ fn main() {
                 }
             }
         }
+        Mode::DumpSyntaxDotty {
+            path,
+            ignore_comments,
+            language_overrides,
+        } => {
+            let path = Path::new(&path);
+            let bytes = read_or_die(path);
+            let src = String::from_utf8_lossy(&bytes).to_string();
+
+            let language = guess(path, &src, &language_overrides);
+            match language {
+                Some(lang) => {
+                    let ts_lang = tsp::from_language(lang);
+                    let arena = Arena::new();
+                    let ast = tsp::parse(&arena, &src, &ts_lang, ignore_comments);
+                    init_all_info(&ast, &[]);
+                    syntax::print_as_dot(&ast);
+                }
+                None => {
+                    eprintln!("No tree-sitter parser for file: {:?}", path);
+                }
+            }
+        }
         Mode::ListLanguages {
             use_color,
             language_overrides,
diff --git a/src/options.rs b/src/options.rs
index be9d3a055f..b3c6268a79 100644
--- a/src/options.rs
+++ b/src/options.rs
@@ -123,6 +123,15 @@ fn app() -> clap::Command<'static> {
                     "Parse a single file with tree-sitter and display the difftastic syntax tree.",
                 ).help_heading("DEBUG OPTIONS"),
         )
+        .arg(
+            Arg::new("dump-syntax-dotty")
+                .long("dump-syntax-dotty")
+                .takes_value(true)
+                .value_name("PATH")
+                .long_help(
+                    "Parse a single file with tree-sitter and display the difftastic syntax tree, as a dotty graph.",
+                ).help_heading("DEBUG OPTIONS"),
+        )
         .arg(
             Arg::new("dump-ts")
                 .long("dump-ts")
@@ -485,6 +494,11 @@ pub(crate) enum Mode {
         ignore_comments: bool,
         language_overrides: Vec<(LanguageOverride, Vec<glob::Pattern>)>,
     },
+    DumpSyntaxDotty {
+        path: String,
+        ignore_comments: bool,
+        language_overrides: Vec<(LanguageOverride, Vec<glob::Pattern>)>,
+    },
 }
 
 fn common_path_suffix(lhs_path: &Path, rhs_path: &Path) -> Option<String> {
@@ -644,6 +658,14 @@ pub(crate) fn parse_args() -> Mode {
         };
     }
 
+    if let Some(path) = matches.value_of("dump-syntax-dotty") {
+        return Mode::DumpSyntaxDotty {
+            path: path.to_owned(),
+            ignore_comments,
+            language_overrides,
+        };
+    }
+
     if let Some(path) = matches.value_of("dump-ts") {
         return Mode::DumpTreeSitter {
             path: path.to_owned(),
diff --git a/src/parse/syntax.rs b/src/parse/syntax.rs
index 746dcc60eb..47676655c3 100644
--- a/src/parse/syntax.rs
+++ b/src/parse/syntax.rs
@@ -369,6 +369,43 @@ pub(crate) fn init_all_info<'a>(lhs_roots: &[&'a Syntax<'a>], rhs_roots: &[&'a S
     init_next_prev(rhs_roots);
 }
 
+pub(crate) fn print_as_dot<'a>(roots: &[&'a Syntax<'a>]) {
+    println!("digraph {{");
+    print_as_dot_(roots);
+    println!("}}");
+}
+
+fn print_as_dot_<'a>(nodes: &[&'a Syntax<'a>]) {
+    for node in nodes {
+        let label = match node {
+            List {
+                open_content,
+                close_content,
+                ..
+            } => {
+                if open_content != "" {
+                    &format!("[label=\"{open_content}{close_content}\"]")
+                } else {
+                    &"[style=dotted]".to_owned()
+                }
+            }
+            Atom { content, .. } => {
+                let content = content.replace("\"", "\\\"");
+                &format!("[label=\"{content}\"]")
+            }
+        };
+
+        println!("  id{} {};", node.id().get(), label);
+
+        if let List { children, .. } = node {
+            for child in children {
+                println!("  id{} -> id{};", node.id().get(), child.id().get());
+            }
+            print_as_dot_(children);
+        }
+    }
+}
+
 fn init_info<'a>(lhs_roots: &[&'a Syntax<'a>], rhs_roots: &[&'a Syntax<'a>]) {
     let mut id = NonZeroU32::new(1).unwrap();
     init_info_on_side(lhs_roots, &mut id);