From 2225ac4446083c806db5a5f28e4bdc445d9ed011 Mon Sep 17 00:00:00 2001
From: Ole Herman Schumacher Elgesem <ole@northern.tech>
Date: Tue, 21 Apr 2026 22:43:47 +0200
Subject: [PATCH] Added new dev subcommand for printing syntax trees

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: Ole Herman Schumacher Elgesem <ole@northern.tech>
---
 CLAUDE.md                       | 11 +++++++++
 src/cfengine_cli/dev.py         |  3 +++
 src/cfengine_cli/main.py        |  2 ++
 src/cfengine_cli/syntax_tree.py | 42 +++++++++++++++++++++++++++++++++
 4 files changed, 58 insertions(+)
 create mode 100644 src/cfengine_cli/syntax_tree.py

diff --git a/CLAUDE.md b/CLAUDE.md
index 9aa03b0..c674094 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -29,6 +29,17 @@ When fixing issues, these are usually the files to look at:
 - The implementation of `cfengine format` is in `src/cfengine_cli/format.py`.
 - The implementation of `cfengine lint` is in `src/cfengine_cli/lint.py`.
 
+## Syntax trees
+
+When working on the formatter or the linter, it is often useful to look at the syntax tree of the policy file.
+There is a `dev` subcommand for this:
+
+```bash
+uv run cfengine dev syntax-tree tests/lint/001_hello_world.cf
+```
+
+The command above prints the syntax tree for `tests/lint/001_hello_world.cf` to the terminal (standard output).
+
 ## Test suites
 
 As mentioned above, the `make check` command runs all the tests.
diff --git a/src/cfengine_cli/dev.py b/src/cfengine_cli/dev.py
index c082785..b242815 100644
--- a/src/cfengine_cli/dev.py
+++ b/src/cfengine_cli/dev.py
@@ -8,6 +8,7 @@
     print_release_dependency_tables,
 )
 from cfengine_cli.docs import update_docs, check_docs
+from cfengine_cli.syntax_tree import syntax_tree
 
 
 def generate_release_information_command(
@@ -72,6 +73,8 @@ def dispatch_dev_subcommand(subcommand, args) -> int:
         return format_docs(args.files)
     if subcommand == "lint-docs":
         return lint_docs()
+    if subcommand == "syntax-tree":
+        return syntax_tree(args.file)
     if subcommand == "generate-release-information":
         return generate_release_information(
             args.omit_download, args.check_against_git, args.minimum_version
diff --git a/src/cfengine_cli/main.py b/src/cfengine_cli/main.py
index d805e32..39a7100 100644
--- a/src/cfengine_cli/main.py
+++ b/src/cfengine_cli/main.py
@@ -97,6 +97,8 @@ def _get_arg_parser():
     parser.add_argument("files", nargs="*")
     parser = dev_subparsers.add_parser("lint-docs")
     parser.add_argument("files", nargs="*")
+    parser = dev_subparsers.add_parser("syntax-tree")
+    parser.add_argument("file", help="CFEngine policy file to print syntax tree for")
     parser = dev_subparsers.add_parser("generate-release-information")
 
     parser.add_argument(
diff --git a/src/cfengine_cli/syntax_tree.py b/src/cfengine_cli/syntax_tree.py
new file mode 100644
index 0000000..4017e66
--- /dev/null
+++ b/src/cfengine_cli/syntax_tree.py
@@ -0,0 +1,42 @@
+"""Print tree-sitter syntax tree for a .cf file."""
+
+import tree_sitter_cfengine as tscfengine
+from tree_sitter import Language, Parser
+
+_LANGUAGE = Language(tscfengine.language())
+_PARSER = Parser(_LANGUAGE)
+
+
+def format_sexp(sexp: str) -> str:
+    """Format an S-expression with indentation."""
+    out = []
+    indent = 0
+    i = 0
+    while i < len(sexp):
+        c = sexp[i]
+        if c == "(":
+            if out and out[-1] != "\n":
+                out.append("\n")
+            out.append("  " * indent)
+            out.append("(")
+            indent += 1
+            i += 1
+        elif c == ")":
+            indent -= 1
+            out.append(")")
+            i += 1
+        elif c == " " and i + 1 < len(sexp) and sexp[i + 1] == "(":
+            i += 1  # skip space before '(', the '(' handler adds newline
+        else:
+            out.append(c)
+            i += 1
+    out.append("\n")
+    return "".join(out)
+
+
+def syntax_tree(path: str) -> int:
+    with open(path, "rb") as f:
+        data = f.read()
+    tree = _PARSER.parse(data)
+    print(format_sexp(str(tree.root_node)), end="")
+    return 0