Fix bug in textutils.apply_units, raise an Exception if string does not match (closes #88808).

1  ChangeLog for logilab.common
2  ============================
3 
4 +	-- 
5 +    * texutils: apply_units raise ValueError if string isn'nt valid (closes #88808)
6 +
7  2011-10-28  --  0.57.1
8      * daemon: change $HOME after dropping privileges (closes #81297)
9 
10      * compat: method_type for py3k use instance of the class to have a
11        real instance method (closes: #79268)

12 
13      def test_unit_with_blank(self):
14          result = tu.apply_units('1 000 KB', self.units)
15          self.assertEqual(result, 1000 * self.units['kb'])
16 
17 +    def test_unit_wrong_input(self):
18 +        self.assertRaises(ValueError, tu.apply_units, '', self.units)
19 +        self.assertRaises(ValueError, tu.apply_units, 'wrong input', self.units)
20 +        self.assertRaises(ValueError, tu.apply_units, 'wrong13 input', self.units)
21 +        self.assertRaises(ValueError, tu.apply_units, 'wrong input42', self.units)
22 +
23  RGX = re.compile('abcd')
24  class PrettyMatchTC(TestCase):
25 
26      def test_known(self):
27          string = 'hiuherabcdef'

28      def test_unormalize_substitute(self):
29          self.assertEqual(tu.unormalize(u'ab \u8000 cd', substitute='_'),
30                           'ab _ cd')
31 
32      def test_unormalize_backward_compat(self):
33 -        self.assertRaises(ValueError, tu.unormalize, u"\u8000",
34 -                          ignorenonascii=False)
35 -        self.assertEqual(tu.unormalize(u"\u8000", ignorenonascii=True), u'')
36 +        self.assertRaises(ValueError, tu.unormalize, u"\u8000")
37 +        self.assertEqual(tu.unormalize(u"\u8000", substitute=''), u'')
38 
39 
40  class ModuleDocTest(DocTest):
41      """test doc test in this module"""
42      module = tu

43  _BLANK_URE = r'(\s|,)+'
44  _BLANK_RE = re.compile(_BLANK_URE)
45  __VALUE_URE = r'-?(([0-9]+\.[0-9]*)|((0x?)?[0-9]+))'
46  __UNITS_URE = r'[a-zA-Z]+'
47  _VALUE_RE = re.compile(r'(?P<value>%s)(?P<unit>%s)?'%(__VALUE_URE, __UNITS_URE))
48 +_VALIDATION_RE = re.compile(r'^((%s)(%s))*(%s)?$' % (__VALUE_URE, __UNITS_URE,
49 +                                                    __VALUE_URE))
50 
51  BYTE_UNITS = {
52      "b": 1,
53      "kb": 1024,
54      "mb": 1024 ** 2,

55      :type value_reg: regexp with "value" and optional "unit" group
56      :param value_reg: match a value and it's unit into the
57      """
58      if inter is None:
59          inter = final
60 -    string = _BLANK_RE.sub('', string)
61 +    fstring = _BLANK_RE.sub('', string)
62 +    if not (fstring and _VALIDATION_RE.match(fstring)):
63 +        raise ValueError("Invalid unit string: %r." % string)
64      values = []
65 -    for match in value_reg.finditer(string):
66 +    for match in value_reg.finditer(fstring):
67          dic = match.groupdict()
68 -        #import sys
69 -        #print >> sys.stderr, dic
70          lit, unit = dic["value"], dic.get("unit")
71          value = inter(lit)
72          if unit is not None:
73              try:
74                  value *= units[unit.lower()]