| Daniel Dunbar | 7b90be7 | 2009-07-31 07:59:05 +0000 | [diff] [blame^] | 1 | import Util |
| 2 | |
| 3 | class ShLexer: |
| 4 | def __init__(self, data): |
| 5 | self.data = data |
| 6 | self.pos = 0 |
| 7 | self.end = len(data) |
| 8 | |
| 9 | def eat(self): |
| 10 | c = self.data[self.pos] |
| 11 | self.pos += 1 |
| 12 | return c |
| 13 | |
| 14 | def look(self): |
| 15 | return self.data[self.pos] |
| 16 | |
| 17 | def maybe_eat(self, c): |
| 18 | """ |
| 19 | maybe_eat(c) - Consume the character c if it is the next character, |
| 20 | returning True if a character was consumed. """ |
| 21 | if self.data[self.pos] == c: |
| 22 | self.pos += 1 |
| 23 | return True |
| 24 | return False |
| 25 | |
| 26 | def lex_arg(self, c): |
| 27 | if c in "'\"": |
| 28 | str = self.lex_arg_quoted(c) |
| 29 | else: |
| 30 | str = c |
| 31 | while self.pos != self.end: |
| 32 | c = self.look() |
| 33 | if c.isspace() or c in "|><&": |
| 34 | break |
| 35 | elif c == '"': |
| 36 | self.eat() |
| 37 | str += self.lex_arg_quoted('"') |
| 38 | else: |
| 39 | str += self.eat() |
| 40 | return str |
| 41 | |
| 42 | def lex_arg_quoted(self, delim): |
| 43 | str = '' |
| 44 | while self.pos != self.end: |
| 45 | c = self.eat() |
| 46 | if c == delim: |
| 47 | return str |
| 48 | elif c == '\\' and delim == '"': |
| 49 | # Shell escaping is just '\"' to avoid termination, no actual |
| 50 | # escaping. |
| 51 | if self.pos == self.end: |
| 52 | Util.warning("escape at end of quoted argument in: %r" % |
| 53 | self.data) |
| 54 | return str |
| 55 | c = self.eat() |
| 56 | if c != delim: |
| 57 | str += '\\' |
| 58 | str += c |
| 59 | else: |
| 60 | str += c |
| 61 | Util.warning("missing quote character in %r" % self.data) |
| 62 | return str |
| 63 | |
| 64 | def lex_one_token(self): |
| 65 | """ |
| 66 | lex_one_token - Lex a single 'sh' token. """ |
| 67 | |
| 68 | c = self.eat() |
| 69 | if c == ';': |
| 70 | return (c) |
| 71 | if c == '|': |
| 72 | if self.maybe_eat('|'): |
| 73 | return ('||',) |
| 74 | return (c,) |
| 75 | if c == '&': |
| 76 | if self.maybe_eat('&'): |
| 77 | return ('&&',) |
| 78 | if self.maybe_eat('>'): |
| 79 | return ('&>',) |
| 80 | return (c,) |
| 81 | if c == '>': |
| 82 | if self.maybe_eat('&'): |
| 83 | return ('>&',) |
| 84 | if self.maybe_eat('>'): |
| 85 | return ('>>',) |
| 86 | return (c,) |
| 87 | if c == '<': |
| 88 | if self.maybe_eat('&'): |
| 89 | return ('<&',) |
| 90 | if self.maybe_eat('>'): |
| 91 | return ('<<',) |
| 92 | return self.lex_arg(c) |
| 93 | |
| 94 | def lex(self): |
| 95 | while self.pos != self.end: |
| 96 | if self.look().isspace(): |
| 97 | self.eat() |
| 98 | else: |
| 99 | yield self.lex_one_token() |
| 100 | |
| 101 | ### |
| 102 | |
| 103 | import unittest |
| 104 | |
| 105 | class TestShLexer(unittest.TestCase): |
| 106 | def lex(self, str): |
| 107 | return list(ShLexer(str).lex()) |
| 108 | |
| 109 | def testops(self): |
| 110 | self.assertEqual(self.lex('a2>c'), |
| 111 | ['a2', ('>',), 'c']) |
| 112 | self.assertEqual(self.lex('a 2>c'), |
| 113 | ['a', '2', ('>',), 'c']) |
| 114 | |
| 115 | def testquoting(self): |
| 116 | self.assertEqual(self.lex(""" 'a' """), |
| 117 | ['a']) |
| 118 | self.assertEqual(self.lex(""" "hello\\"world" """), |
| 119 | ['hello"world']) |
| 120 | self.assertEqual(self.lex(""" "hello\\'world" """), |
| 121 | ["hello\\'world"]) |
| 122 | self.assertEqual(self.lex(""" he"llo wo"rld """), |
| 123 | ["hello world"]) |
| 124 | |
| 125 | if __name__ == '__main__': |
| 126 | unittest.main() |