def EvalCharLiteralForRegex(tok):
"""For regex char classes.
Similar logic as below.
"""
id_ = tok.id
value = tok.val
if id_ == Id.Char_OneChar:
c = value[1]
s = _ONE_CHAR[c]
return class_literal_term.ByteSet(s, tok.span_id)
elif id_ == Id.Char_Hex:
s = value[2:]
i = int(s, 16)
return class_literal_term.ByteSet(chr(i), tok.span_id)
elif id_ in (Id.Char_Unicode4, Id.Char_Unicode8):
s = value[2:]
i = int(s, 16)
return class_literal_term.CodePoint(i, tok.span_id)
else:
raise AssertionError
def EvalCharLiteralForRegex(tok):
# type: (Token) -> Optional[class_literal_term_t]
"""For regex char classes.
Similar logic as below.
"""
id_ = tok.id
value = tok.val
if id_ == Id.Char_OneChar:
c = value[1]
s = consts.LookupCharC(c)
return class_literal_term.ByteSet(s, tok.span_id)
elif id_ == Id.Char_Hex:
s = value[2:]
i = int(s, 16)
return class_literal_term.ByteSet(chr(i), tok.span_id)
elif id_ in (Id.Char_Unicode4, Id.Char_Unicode8):
s = value[2:]
i = int(s, 16)
return class_literal_term.CodePoint(i, tok.span_id)
elif id_ == Id.Expr_Name: # [b B] is NOT mutated
return None
else:
raise AssertionError(Id_str(id_))
def _MutateClassLiteral(self, node):
# type: (re_t) -> None
for i, term in enumerate(node.terms):
s = None
if term.tag == class_literal_term_e.SingleQuoted:
s = word_eval.EvalSingleQuoted(term)
spid = term.left.span_id
elif term.tag == class_literal_term_e.DoubleQuoted:
s = self.word_ev.EvalDoubleQuotedToString(term)
spid = term.left.span_id
elif term.tag == class_literal_term_e.BracedVarSub:
s = self.word_ev.EvalBracedVarSubToString(term)
spid = term.spids[0]
elif term.tag == class_literal_term_e.SimpleVarSub:
s = self.word_ev.EvalSimpleVarSubToString(term.token)
spid = term.token.span_id
elif term.tag == class_literal_term_e.CharLiteral:
# What about \0?
# At runtime, ERE should disallow it. But we can also disallow it here.
node.terms[i] = word_compile.EvalCharLiteralForRegex(term.tok)
if s is not None:
# A string like '\x7f\xff' should be presented like
if len(s) > 1:
for c in s:
if ord(c) > 128:
e_die("Express these bytes as character literals to avoid "
"confusing them with encoded characters", span_id=spid)
node.terms[i] = class_literal_term.ByteSet(s, spid)