Add option to preserve comments when parsing templates

src/jinja2/environment.py

@@ -600,6 +600,7 @@ def parse(
         source: str,
         name: t.Optional[str] = None,
         filename: t.Optional[str] = None,
+        preserve_comments: bool = False,
     ) -> nodes.Template:
         """Parse the sourcecode and return the abstract syntax tree.  This
         tree of nodes is used by the compiler to convert the template into
@@ -610,15 +611,21 @@ def parse(
         this gives you a good overview of the node tree generated.
         """
         try:
-            return self._parse(source, name, filename)
+            return self._parse(source, name, filename, preserve_comments)
         except TemplateSyntaxError:
             self.handle_exception(source=source)
 
     def _parse(
-        self, source: str, name: t.Optional[str], filename: t.Optional[str]
+        self,
+        source: str,
+        name: t.Optional[str],
+        filename: t.Optional[str],
+        preserve_comments: bool = False,
     ) -> nodes.Template:
         """Internal parsing function used by `parse` and `compile`."""
-        return Parser(self, source, name, filename).parse()
+        return Parser(
+            self, source, name, filename, preserve_comments=preserve_comments
+        ).parse()
 
     def lex(
         self,
@@ -663,12 +670,13 @@ def _tokenize(
         name: t.Optional[str],
         filename: t.Optional[str] = None,
         state: t.Optional[str] = None,
+        preserve_comments: bool = False,
     ) -> TokenStream:
         """Called by the parser to do the preprocessing and filtering
         for all the extensions.  Returns a :class:`~jinja2.lexer.TokenStream`.
         """
         source = self.preprocess(source, name, filename)
-        stream = self.lexer.tokenize(source, name, filename, state)
+        stream = self.lexer.tokenize(source, name, filename, state, preserve_comments)
 
         for ext in self.iter_extensions():
             stream = ext.filter_stream(stream)  # type: ignore

src/jinja2/lexer.py

@@ -146,17 +146,22 @@
     f"({'|'.join(re.escape(x) for x in sorted(operators, key=lambda x: -len(x)))})"
 )
 
-ignored_tokens = frozenset(
+comment_tokens = frozenset(
     [
         TOKEN_COMMENT_BEGIN,
         TOKEN_COMMENT,
         TOKEN_COMMENT_END,
-        TOKEN_WHITESPACE,
         TOKEN_LINECOMMENT_BEGIN,
         TOKEN_LINECOMMENT_END,
         TOKEN_LINECOMMENT,
     ]
 )
+ignored_tokens = frozenset(
+    [
+        TOKEN_WHITESPACE,
+        *comment_tokens,
+    ]
+)
 ignore_if_empty = frozenset(
     [TOKEN_WHITESPACE, TOKEN_DATA, TOKEN_COMMENT, TOKEN_LINECOMMENT]
 )
@@ -607,22 +612,30 @@ def tokenize(
         name: t.Optional[str] = None,
         filename: t.Optional[str] = None,
         state: t.Optional[str] = None,
+        preserve_comments: bool = False,
     ) -> TokenStream:
         """Calls tokeniter + tokenize and wraps it in a token stream."""
         stream = self.tokeniter(source, name, filename, state)
-        return TokenStream(self.wrap(stream, name, filename), name, filename)
+        return TokenStream(
+            self.wrap(stream, name, filename, preserve_comments), name, filename
+        )
 
     def wrap(
         self,
         stream: t.Iterable[t.Tuple[int, str, str]],
         name: t.Optional[str] = None,
         filename: t.Optional[str] = None,
+        preserve_comments: bool = False,
     ) -> t.Iterator[Token]:
         """This is called with the stream as returned by `tokenize` and wraps
         every token in a :class:`Token` and converts the value.
         """
+        ignored = ignored_tokens
+        if preserve_comments:
+            ignored -= comment_tokens
+
         for lineno, token, value_str in stream:
-            if token in ignored_tokens:
+            if token in ignored:
                 continue
 
             value: t.Any = value_str
@@ -631,6 +644,10 @@ def wrap(
                 token = TOKEN_BLOCK_BEGIN
             elif token == TOKEN_LINESTATEMENT_END:
                 token = TOKEN_BLOCK_END
+            elif token == TOKEN_LINECOMMENT_BEGIN:
+                token = TOKEN_COMMENT_BEGIN
+            elif token == TOKEN_LINECOMMENT_END:
+                token = TOKEN_COMMENT_END
             # we are not interested in those tokens in the parser
             elif token in (TOKEN_RAW_BEGIN, TOKEN_RAW_END):
                 continue

src/jinja2/nodes.py

@@ -715,6 +715,13 @@ def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
         return self.expr2.as_const(eval_ctx)
 
 
+class Comment(Stmt):
+    """A template comment."""
+
+    fields = ("data",)
+    data: str
+
+
 def args_as_const(
     node: t.Union["_FilterTestCommon", "Call"], eval_ctx: t.Optional[EvalContext]
 ) -> t.Tuple[t.List[t.Any], t.Dict[t.Any, t.Any]]:

src/jinja2/parser.py

@@ -57,9 +57,12 @@ def __init__(
         name: t.Optional[str] = None,
         filename: t.Optional[str] = None,
         state: t.Optional[str] = None,
+        preserve_comments: bool = False,
     ) -> None:
         self.environment = environment
-        self.stream = environment._tokenize(source, name, filename, state)
+        self.stream = environment._tokenize(
+            source, name, filename, state, preserve_comments
+        )
         self.name = name
         self.filename = filename
         self.closed = False
@@ -1025,6 +1028,11 @@ def flush_data() -> None:
                     else:
                         body.append(rv)
                     self.stream.expect("block_end")
+                elif token.type == "comment_begin":
+                    flush_data()
+                    next(self.stream)
+                    body.append(nodes.Comment(next(self.stream).value))
+                    self.stream.expect("comment_end")
                 else:
                     raise AssertionError("internal parsing error")
 

tests/test_lexnparse.py

@@ -314,6 +314,28 @@ def assert_error(code, expected):
         )
         assert_error("{% unknown_tag %}", "Encountered unknown tag 'unknown_tag'.")
 
+    def test_comment_preservation(self, env):
+        ast = env.parse("{# foo #}{{ bar }}", preserve_comments=True)
+        assert len(ast.body) == 2
+        assert isinstance(ast.body[0], nodes.Comment)
+        assert ast.body[0].data == " foo "
+
+        ast = env.parse("{# foo #}{{ bar }}", preserve_comments=False)
+        assert len(ast.body) == 1
+        assert not isinstance(ast.body[0], nodes.Comment)
+
+    def test_line_comment_preservation(self, env):
+        env = Environment(line_comment_prefix="#")
+
+        ast = env.parse("# foo\n{{ bar }}", preserve_comments=True)
+        assert len(ast.body) == 2
+        assert isinstance(ast.body[0], nodes.Comment)
+        assert ast.body[0].data == " foo"
+
+        ast = env.parse("# foo\n{{ bar }}", preserve_comments=False)
+        assert len(ast.body) == 1
+        assert not isinstance(ast.body[0], nodes.Comment)
+
 
 class TestSyntax:
     def test_call(self, env):