def _ParseCharClass(self):
# type: () -> List[glob_part_t]
"""
Returns:
a CharClass if the parse suceeds, or a Literal if fails. In the latter
case, we also append a warning.
"""
first_token = glob_part.Literal(self.token_type, self.token_val)
balance = 1 # We already saw a [
tokens = [] # type: List[Tuple[Id_t, str]]
# NOTE: There is a special rule where []] and [[] are valid globs. Also
# [^[] and sometimes [^]], although that one is ambiguous!
# And [[:space:]] and [[.class.]] has to be taken into account too. I'm
# punting on this now because the rule isn't clear and consistent between
# shells.
while True:
self._Next()
if self.token_type == Id.Eol_Tok:
# TODO: location info
self.warnings.append(
'Malformed character class; treating as literal')
parts = [first_token] # type: List[glob_part_t]
for (id_, s) in tokens:
parts.append(glob_part.Literal(id_, s))
return parts
if self.token_type == Id.Glob_LBracket:
balance += 1
elif self.token_type == Id.Glob_RBracket:
balance -= 1
if balance == 0:
break
tokens.append(
(self.token_type, self.token_val)) # Don't append the last ]
negated = False
if len(tokens):
id1, _ = tokens[0]
# NOTE: Both ! and ^ work for negation in globs
# https://www.gnu.org/software/bash/manual/html_node/Pattern-Matching.html#Pattern-Matching
# TODO: Warn about the one that's not recommended?
if id1 in (Id.Glob_Bang, Id.Glob_Caret):
negated = True
tokens = tokens[1:]
strs = [s for _, s in tokens]
return [glob_part.CharClass(negated, strs)]
def Parse(self):
# type: () -> Tuple[List[glob_part_t], List[str]]
"""
Returns:
regex string (or None if it's not a glob)
A list of warnings about the syntax
"""
parts = [] # type: List[glob_part_t]
while True:
self._Next()
id_ = self.token_type
s = self.token_val
#util.log('%s %r', self.token_type, self.token_val)
if id_ == Id.Eol_Tok:
break
if id_ in (Id.Glob_Star, Id.Glob_QMark):
parts.append(glob_part.Operator(id_))
elif id_ == Id.Glob_LBracket:
# Could return a Literal or a CharClass
parts.extend(self._ParseCharClass())
else: # Glob_{Bang,Caret,CleanLiterals,OtherLiteral,RBracket,EscapedChar,
# BadBackslash}
parts.append(glob_part.Literal(id_, s))
# Also check for warnings. TODO: location info.
if id_ == Id.Glob_RBracket:
self.warnings.append('Got unescaped right bracket')
if id_ == Id.Glob_BadBackslash:
self.warnings.append('Got unescaped trailing backslash')
return parts, self.warnings