email: use same strategy as quoted-string to fold comments

- No forced space to indent,
- use almost the same code as the `bare-quoted-string`
block above, handle nested comments and escaped parenthesis
in addition

Author: beledouxdenis

Lib/email/_header_value_parser.py

@@ -101,6 +101,12 @@ def make_quoted_pairs(value):
     return str(value).replace('\\', '\\\\').replace('"', '\\"')
 
 
+def make_parenthesis_pairs(value):
+    """Escape parenthesis and backslash for use within a comment."""
+    return str(value).replace('\\', '\\\\') \
+        .replace('(', '\\(').replace(')', '\\)')
+
+
 def quote_string(value):
     escaped = make_quoted_pairs(value)
     return f'"{escaped}"'
@@ -2941,10 +2947,7 @@ def _refold_parse_tree(parse_tree, *, policy):
                 # the way encoded strings handle continuation lines, we need to
                 # be prepared to encode any whitespace if the next line turns
                 # out to start with an encoded word.
-                line = newline + tstr
-                if line[0] not in WSP:
-                    line = ' ' + line
-                lines.append(line)
+                lines.append(newline + tstr)
 
                 whitespace_accumulator = []
                 for char in lines[-1]:
@@ -2966,6 +2969,13 @@ def _refold_parse_tree(parse_tree, *, policy):
                     [ValueTerminal(make_quoted_pairs(p), 'ptext')
                      for p in newparts] +
                     [ValueTerminal('"', 'ptext')])
+            if part.token_type == 'comment':
+                newparts = (
+                    [ValueTerminal('(', 'ptext')] +
+                    [ValueTerminal(make_parenthesis_pairs(p), 'ptext')
+                     if p.token_type == 'ptext' else p
+                     for p in newparts] +
+                    [ValueTerminal(')', 'ptext')])
             if not part.as_ew_allowed:
                 wrap_as_ew_blocked += 1
                 newparts.append(end_ew_not_allowed)
@@ -2980,10 +2990,7 @@ def _refold_parse_tree(parse_tree, *, policy):
         # We can't figure out how to wrap, it, so give up.
         newline = _steal_trailing_WSP_if_exists(lines)
         if newline or part.startswith_fws():
-            line = newline + tstr
-            if line[0] not in WSP:
-                line = ' ' + line
-            lines.append(line)
+            lines.append(newline + tstr)
         else:
             # We can't fold it onto the next line either...
             lines[-1] += tstr

Lib/test/test_email/test__header_value_parser.py

@@ -3298,67 +3298,18 @@ def test_address_list_with_long_unwrapable_comment(self):
         policy = self.policy.clone(max_line_length=40)
         cases = [
             # (to, folded)
-
-            # 1. Unwrappable Comments
-
-            # Entire line is <= 40 characters, 40 characters exactly
-            # No folding
-            ('spy@example.org(loremipsumdolorsitametc)', 'spy@example.org(loremipsumdolorsitametc)\n'),
-            ('(loremipsumdolorsitametc)spy@example.org', '(loremipsumdolorsitametc)spy@example.org\n'),
-            # Entire line is > 40 characters, 41 characters
-            # Folding triggered
-            ('spy@example.org(loremipsumdolorsitametco)','spy@example.org\n (loremipsumdolorsitametco)\n'),
-             ('(loremipsumdolorsitametco)spy@example.org', '(loremipsumdolorsitametco)spy@example.org\n'),
-            # Entire line is > 40 characters, 54 characters, `len(tstr) <= maxlen - len(lines[-1])` is `True`
-            # Folding triggered
-            # Comment part < 40 characthers, 39 characters, `len(tstr) + 1 <= maxlen` is `True`
-            # No attempt to fold the subpart
-            ('spy@example.org(loremipsumdolorsitametconsecteturadip)',
-             'spy@example.org\n'
-             ' (loremipsumdolorsitametconsecteturadip)\n'),
-            ('(loremipsumdolorsitametconsecteturadip)spy@example.org',
-             '(loremipsumdolorsitametconsecteturadip)spy@example.org\n'),
-            # Entire line is > 40 characters, 55 characters, `len(tstr) <= maxlen - len(lines[-1])` is `True`
-            # Folding triggered
-            # Comment part >= 40 characters, 40 characters exactly, `len(tstr) + 1 <= maxlen` is `False`
-            # Attempt to fold the subpart
-            ('spy@example.org(loremipsumdolorsitametconsecteturadipi)',
-             'spy@example.org\n'
-             ' (loremipsumdolorsitametconsecteturadipi)\n'),
-            ('(loremipsumdolorsitametconsecteturadipi)spy@example.org',
-             '(loremipsumdolorsitametconsecteturadipi)spy@example.org\n'),
-
-            # 2. Wrappable comments
-
-            # Entire line is <= 40 characters, 40 characters exactly
-            # No folding
-            ('spy@example.org(loremipsumd olorsitamet)', 'spy@example.org(loremipsumd olorsitamet)\n'),
-            ('(loremipsumd olorsitamet)spy@example.org', '(loremipsumd olorsitamet)spy@example.org\n'),
-            # Entire line is > 40 characters, 41 characters
-            # Folding triggered
-            # Comment part < 40 characters
-            ('spy@example.org(loremipsumd olorsitametc)', 'spy@example.org\n (loremipsumd olorsitametc)\n'),
-            ('(loremipsumd olorsitametc)spy@example.org', '(loremipsumd olorsitametc)spy@example.org\n'),
-            # Entire line is > 40 characters, 56 characters
-            # Folding triggered
-            # Comment part > 40 characters, 41 characters
-            ('spy@example.org(loremipsumd loremipsumdolorsitametconse)', 'spy@example.org(loremipsumd\n loremipsumdolorsitametconse)\n'),
-            ('(loremipsumd loremipsumdolorsitametconse)spy@example.org', '(loremipsumd\n loremipsumdolorsitametconse)spy@example.org\n'),
-            # Entire line is > 40 characters, 70 characters
-            # Folding triggered
-            # Comment part > 40 characters, 55 characters
-            # One word in the comment > 40 characters, 41 characters
-            ('spy@example.org(loremipsumd loremipsumdolorsitametconsecteturadipisci)', 'spy@example.org(loremipsumd\n loremipsumdolorsitametconsecteturadipisci)\n'),
-            ('(loremipsumd loremipsumdolorsitametconsecteturadipisci)spy@example.org', '(loremipsumd\n loremipsumdolorsitametconsecteturadipisci)spy@example.org\n'),
-
-            # 3. Nested comments
-
-            ('spy@example.org((loremipsumdolorsitametconsecteturadi))', 'spy@example.org(\n (loremipsumdolorsitametconsecteturadi))\n'),
-            ('spy@example.org((loremipsumdolorsitametconsecteturadip))', 'spy@example.org(\n (loremipsumdolorsitametconsecteturadip)\n )\n'),
-            ('spy@example.org((loremipsumdolorsitam)(loremipsumdolorsitam))', 'spy@example.org((loremipsumdolorsitam)\n (loremipsumdolorsitam))\n'),
-            ('spy@example.org((loremipsumdolorsitametc)(loremipsumdolorsitametc))', 'spy@example.org(\n (loremipsumdolorsitametc)\n (loremipsumdolorsitametc))\n'),
-            ('spy@example.org(loremipsumdolorsitametc(loremipsumdolorsitametc))', 'spy@example.org(loremipsumdolorsitametc\n (loremipsumdolorsitametc))\n'),
-            ('spy@example.org((loremipsumdolorsitametc)loremipsumdolorsitametc)', 'spy@example.org(\n (loremipsumdolorsitametc)\n loremipsumdolorsitametc)\n'),
+            ('(loremipsumdolorsitametconsecteturadipi)<spy@example.org>',
+             '(loremipsumdolorsitametconsecteturadipi)<spy@example.org>\n'),
+            ('<spy@example.org>(loremipsumdolorsitametconsecteturadipi)',
+             '<spy@example.org>(loremipsumdolorsitametconsecteturadipi)\n'),
+            ('(loremipsum dolorsitametconsecteturadipi)<spy@example.org>',
+             '(loremipsum dolorsitametconsecteturadipi)<spy@example.org>\n'),
+             ('<spy@example.org>(loremipsum dolorsitametconsecteturadipi)',
+             '<spy@example.org>(loremipsum\n dolorsitametconsecteturadipi)\n'),
+            ('(Escaped \\( \\) chars \\\\ in comments stay escaped)<spy@example.org>',
+             '(Escaped \\( \\) chars \\\\ in comments stay\n escaped)<spy@example.org>\n'),
+            ('((loremipsum)(loremipsum)(loremipsum)(loremipsum))<spy@example.org>',
+             '((loremipsum)(loremipsum)(loremipsum)(loremipsum))<spy@example.org>\n'),
         ]
         for (to, folded) in cases:
             with self.subTest(to=to):