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

Update the dedicated unit test.

3 files changed, 17 insertions, 7 deletions

diff --git a/ChangeLog b/ChangeLog
index f1860b8..83734b2 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,6 +1,9 @@
 ChangeLog for logilab.common
 ============================
 
+	-- 
+    * texutils: apply_units raise ValueError if string isn'nt valid (closes #88808)
+
 2011-10-28  --  0.57.1
     * daemon: change $HOME after dropping privileges (closes #81297)
 
diff --git a/test/unittest_textutils.py b/test/unittest_textutils.py
index e8b46b1..db18d0d 100644
--- a/test/unittest_textutils.py
+++ b/test/unittest_textutils.py
@@ -184,6 +184,12 @@ class UnitsTC(TestCase):
         result = tu.apply_units('1 000 KB', self.units)
         self.assertEqual(result, 1000 * self.units['kb'])
 
+    def test_unit_wrong_input(self):
+        self.assertRaises(ValueError, tu.apply_units, '', self.units)
+        self.assertRaises(ValueError, tu.apply_units, 'wrong input', self.units)
+        self.assertRaises(ValueError, tu.apply_units, 'wrong13 input', self.units)
+        self.assertRaises(ValueError, tu.apply_units, 'wrong input42', self.units)
+
 RGX = re.compile('abcd')
 class PrettyMatchTC(TestCase):
 
@@ -246,9 +252,8 @@ class UnormalizeTC(TestCase):
                          'ab _ cd')
 
     def test_unormalize_backward_compat(self):
-        self.assertRaises(ValueError, tu.unormalize, u"\u8000",
-                          ignorenonascii=False)
-        self.assertEqual(tu.unormalize(u"\u8000", ignorenonascii=True), u'')
+        self.assertRaises(ValueError, tu.unormalize, u"\u8000")
+        self.assertEqual(tu.unormalize(u"\u8000", substitute=''), u'')
 
 
 class ModuleDocTest(DocTest):
diff --git a/textutils.py b/textutils.py
index bdeed41..f55c004 100644
--- a/textutils.py
+++ b/textutils.py
@@ -313,6 +313,8 @@ _BLANK_RE = re.compile(_BLANK_URE)
 __VALUE_URE = r'-?(([0-9]+\.[0-9]*)|((0x?)?[0-9]+))'
 __UNITS_URE = r'[a-zA-Z]+'
 _VALUE_RE = re.compile(r'(?P<value>%s)(?P<unit>%s)?'%(__VALUE_URE, __UNITS_URE))
+_VALIDATION_RE = re.compile(r'^((%s)(%s))*(%s)?$' % (__VALUE_URE, __UNITS_URE,
+                                                    __VALUE_URE))
 
 BYTE_UNITS = {
     "b": 1,
@@ -352,12 +354,12 @@ def apply_units(string, units, inter=None, final=float, blank_reg=_BLANK_RE,
     """
     if inter is None:
         inter = final
-    string = _BLANK_RE.sub('', string)
+    fstring = _BLANK_RE.sub('', string)
+    if not (fstring and _VALIDATION_RE.match(fstring)):
+        raise ValueError("Invalid unit string: %r." % string)
     values = []
-    for match in value_reg.finditer(string):
+    for match in value_reg.finditer(fstring):
         dic = match.groupdict()
-        #import sys
-        #print >> sys.stderr, dic
         lit, unit = dic["value"], dic.get("unit")
         value = inter(lit)
         if unit is not None: