quoted-strings: Fix only-when-needed on corner cases

Change implementation of `required: only-when-needed`, because
maintaining a list of `START_TOKENS` and just looking at the first
character of string values has proven to be partially broken.

Cf. discussion at
https://github.com/adrienverge/yamllint/pull/246#issuecomment-612354097.

Fixes https://github.com/adrienverge/yamllint/issues/242 and
https://github.com/adrienverge/yamllint/pull/244.

tests/rules/test_quoted_strings.py

@@ -316,3 +316,44 @@ class QuotedTestCase(RuleTestCase):
                    '  "word 1\\\n'            # fails
                    '   word 2"\n',
                    conf, problem1=(12, 3))
+
+    def test_only_when_needed_corner_cases(self):
+        conf = 'quoted-strings: {required: only-when-needed}\n'
+
+        self.check('---\n'
+                   '- ""\n'
+                   '- "- item"\n'
+                   '- "key: value"\n'
+                   '- "%H:%M:%S"\n'
+                   '- "%wheel ALL=(ALL) NOPASSWD: ALL"\n'
+                   '- \'"quoted"\'\n'
+                   '- "\'foo\' == \'bar\'"\n'
+                   '- "\'Mac\' in ansible_facts.product_name"\n',
+                   conf)
+        self.check('---\n'
+                   'k1: ""\n'
+                   'k2: "- item"\n'
+                   'k3: "key: value"\n'
+                   'k4: "%H:%M:%S"\n'
+                   'k5: "%wheel ALL=(ALL) NOPASSWD: ALL"\n'
+                   'k6: \'"quoted"\'\n'
+                   'k7: "\'foo\' == \'bar\'"\n'
+                   'k8: "\'Mac\' in ansible_facts.product_name"\n',
+                   conf)
+
+        self.check('---\n'
+                   '- ---\n'
+                   '- "---"\n'                     # fails
+                   '- ----------\n'
+                   '- "----------"\n'              # fails
+                   '- :wq\n'
+                   '- ":wq"\n',                    # fails
+                   conf, problem1=(3, 3), problem2=(5, 3), problem3=(7, 3))
+        self.check('---\n'
+                   'k1: ---\n'
+                   'k2: "---"\n'                   # fails
+                   'k3: ----------\n'
+                   'k4: "----------"\n'            # fails
+                   'k5: :wq\n'
+                   'k6: ":wq"\n',                  # fails
+                   conf, problem1=(3, 5), problem2=(5, 5), problem3=(7, 5))

yamllint/rules/quoted_strings.py

@@ -77,16 +77,30 @@ DEFAULT = {'quote-type': 'any',
            'required': True}
 
 DEFAULT_SCALAR_TAG = u'tag:yaml.org,2002:str'
-START_TOKENS = {'#', '*', '!', '?', '@', '`', '&',
-                ',', '-', '{', '}', '[', ']', ':'}
 
 
-def quote_match(quote_type, token_style):
+def _quote_match(quote_type, token_style):
     return ((quote_type == 'any') or
             (quote_type == 'single' and token_style == "'") or
             (quote_type == 'double' and token_style == '"'))
 
 
+def _quotes_are_needed(string):
+    loader = yaml.BaseLoader('key: ' + string)
+    # Remove the 5 first tokens corresponding to 'key: ' (StreamStartToken,
+    # BlockMappingStartToken, KeyToken, ScalarToken(value=key), ValueToken)
+    for _ in range(5):
+        loader.get_token()
+    try:
+        a, b = loader.get_token(), loader.get_token()
+        if (isinstance(a, yaml.ScalarToken) and a.style is None and
+                isinstance(b, yaml.BlockEndToken)):
+            return False
+        return True
+    except yaml.scanner.ScannerError:
+        return True
+
+
 def check(conf, token, prev, next, nextnext, context):
     if not (isinstance(token, yaml.tokens.ScalarToken) and
             isinstance(prev, (yaml.BlockEntryToken, yaml.FlowEntryToken,
@@ -121,25 +135,25 @@ def check(conf, token, prev, next, nextnext, context):
     if required is True:
 
         # Quotes are mandatory and need to match config
-        if token.style is None or not quote_match(quote_type, token.style):
+        if token.style is None or not _quote_match(quote_type, token.style):
             msg = "string value is not quoted with %s quotes" % (quote_type)
 
     elif required is False:
 
         # Quotes are not mandatory but when used need to match config
-        if token.style and not quote_match(quote_type, token.style):
+        if token.style and not _quote_match(quote_type, token.style):
             msg = "string value is not quoted with %s quotes" % (quote_type)
 
     elif not token.plain:
 
         # Quotes are disallowed when not needed
-        if (tag == DEFAULT_SCALAR_TAG and token.value
+        if (tag == DEFAULT_SCALAR_TAG and token.value and
-                and token.value[0] not in START_TOKENS):
+                not _quotes_are_needed(token.value)):
             msg = "string value is redundantly quoted with %s quotes" % (
                 quote_type)
 
         # But when used need to match config
-        elif token.style and not quote_match(quote_type, token.style):
+        elif token.style and not _quote_match(quote_type, token.style):
             msg = "string value is not quoted with %s quotes" % (quote_type)
 
     if msg is not None: