Update of /cvsroot/tmda/tmda/TMDA
In directory usw-pr-cvs1:/tmp/cvs-serv16187/TMDA
Modified Files:
ChangeLog FilterParser.py
Log Message:
Bugfix. Filter file rule lines with 'headers' or 'body' sources are
not split correctly if the second field of the line (match) contains
whitespace. This makes perfectly valid regular expressions fail to
match.
The 'headers' and 'body' sources now use parentheses as delimiters,
rather than whitespace.
Reported by Alex Chaffee in <20020107082800.A13221@xxxxxxxxxx> on
tmda-users. Patch contributed by Tim Legant.
Index: ChangeLog
===================================================================
RCS file: /cvsroot/tmda/tmda/TMDA/ChangeLog,v
retrieving revision 1.128
retrieving revision 1.129
diff -u -r1.128 -r1.129
--- ChangeLog 2002/01/22 18:53:56 1.128
+++ ChangeLog 2002/01/24 17:01:42 1.129
@@ -1,3 +1,29 @@
+2002-01-24 Tim Legant <tim@xxxxxxxxxxx>
+
+ * FilterParser.py (FilterParser): Removed 'whitespace' and
+ source_match regular expressions and added 'most_sources',
+ 'hdrbody_sources' and 'matches' regular expressions. Modified
+ 'tag_action' regular expression.
+
+ * (FilterParser.__readrule): Changed whitespace collapse to simply
+ substitue a space for every tab.
+
+ * (FilterParser.__parserule): Rewrote to handle parenthesized
+ regular expressions with embedded spaces (for the 'headers' and
+ 'body' sources).
+
+ * (FilterParser.__adderror): Fixed massive bug. Should have been
+ '__exception', not 'exception'. When 'checking' was True, this
+ resulted in a NameError exception.
+
+ * (FilterParser.__buildactions): Fixed to handle multiple spaces
+ between headers and actions.
+
+ * (FilterParser.firstmatch): Created 'line' return value by
+ calling new module-level function, '_rulestr'.
+
+ * (_rulestr): New function.
+
2002-01-19 Tim Legant <tim@xxxxxxxxxxx>
* FilterParser.py: Rewrite to support multiline,
Index: FilterParser.py
===================================================================
RCS file: /cvsroot/tmda/tmda/TMDA/FilterParser.py,v
retrieving revision 1.13
retrieving revision 1.14
diff -u -r1.13 -r1.14
--- FilterParser.py 2002/01/22 18:53:56 1.13
+++ FilterParser.py 2002/01/24 17:01:42 1.14
@@ -39,16 +39,24 @@
class FilterParser:
bol_comment = re.compile(r'\s*#')
- whitespace = re.compile(r'\s+')
-
- source_match = re.compile(r"""
- (?:( (?:(?:to|from)(?:-(?:file|cdb|dbm|mailman\.\S+))?)
- | (?:(?:body|headers)(?:-file)?)
+ most_sources = re.compile(r"""
+ ( (?:(?:to|from)(?:-(?:file|cdb|dbm|mailman\.\S+))?)
| size )
+ \ (?# NOTE: preceding character must be an actual space)
+ """, re.VERBOSE | re.IGNORECASE)
+
+ hdrbody_sources = re.compile(r"""
+ ( (?:body|headers)(?:-file)?)
+ \ (?# NOTE: preceding character must be an actual space)
+ """, re.VERBOSE | re.IGNORECASE)
+
+ matches = re.compile(r"""
+ (?: \( ( (?: \\\) | [^)] )+ ) \)
+ | ( \S+ ) )
\ (?# NOTE: preceding character must be an actual space)
- )? ( \S+ )""", re.VERBOSE | re.IGNORECASE)
+ """, re.VERBOSE)
- tag_action = re.compile(r'([A-Za-z][-\w]+) (\S+)')
+ tag_action = re.compile(r'([A-Za-z][-\w]+)\s+(\S+)')
in_action =
re.compile(r'(bounce|reject|drop|exit|stop|ok|accept|deliver|confirm)',
re.IGNORECASE)
@@ -127,8 +135,8 @@
# comment at beginning of line, with or without leading whitespace
if self.bol_comment.match(original_line):
continue
- # collapse any sequence of whitespace to a single space
- line = self.whitespace.sub(' ', original_line)
+ # substitute space characters for tab characters
+ line = string.replace(original_line, '\t', ' ')
# lose end-of-line comments and trailing whitespace
line = string.split(line, ' #')[0]
line = string.rstrip(line)
@@ -162,7 +170,9 @@
with three fields. The three fields are:
source - string: to*, from*, body, headers, size, default
- match - string: the email address to be matched against
+ match - string: the email address to be matched against, a
+ filename or a regular expression enclosed within
+ parentheses
actions - dictionary: a dictionary with a key of 'action' and
a value that is a tuple. The value tuple contains
the 'cookie' type and the 'cookie' option. Ex:
@@ -176,22 +186,27 @@
"""
rule = None
# first, get the source and the match
- mo = self.source_match.match(rule_line)
+ mo = self.most_sources.match(rule_line)
+ if not mo:
+ mo = self.hdrbody_sources.match(rule_line)
if not mo:
self.__adderror(self.__rule_lineno, rule_line)
else:
- (source, match) = mo.groups()
- if not source:
- # missing source!
- self.__adderror(self.__rule_lineno, rule_line)
- else:
- action_line = rule_line[mo.end()+1:]
+ source = mo.group(1)
+ match_line = string.lstrip(rule_line[mo.end():])
+ mo = self.matches.match(match_line)
+ if not mo:
+ # missing match
+ self.__adderror(self.__rule_lineno, match_line)
+ else:
+ match = mo.group(1) or mo.group(2)
+ action_line = string.lstrip(match_line[mo.end():])
actions = self.__buildactions(action_line, rule_line)
- if actions:
- rule = (source, match, actions)
- else:
- # missing action!
- self.__adderror(self.__rule_lineno, action_line)
+ if actions:
+ rule = (source, match, actions)
+ else:
+ # missing action!
+ self.__adderror(self.__rule_lineno, action_line)
return rule
@@ -205,9 +220,9 @@
# raised at the end of the file and will contain a
# list of all bogus lines.
if self.checking:
- if not self.exception:
- self.exception = ParsingError(self.filename)
- self.exception.append(lineno, `line`)
+ if not self.__exception:
+ self.__exception = ParsingError(self.filename)
+ self.__exception.append(lineno, `line`)
def __buildactions(self, action_line, rule_line=None):
@@ -217,7 +232,7 @@
"""
actions = None
if action_line[:len('tag ')] == 'tag ':
- action_line = action_line[len('tag '):]
+ action_line = string.lstrip(action_line[len('tag '):])
while len(action_line) > 0:
mo = self.tag_action.match(action_line)
if not mo:
@@ -233,7 +248,7 @@
else:
# malformed action
self.__adderror(self.__rule_lineno, action)
- action_line = action_line[mo.end()+1:]
+ action_line = string.lstrip(action_line[mo.end()+1:])
else:
mo = self.in_action.match(action_line)
if mo:
@@ -401,11 +416,20 @@
if found_match:
break
if found_match:
- line = str(source) + ' ' + str(match) + ' ' + _actionstr(actions)
+ line = _rulestr(source, match, actions)
else:
actions = None
return actions, line
+
+def _rulestr(source, match, actions):
+ """
+ Build string from source, match and actions.
+ """
+ if source in ('headers', 'body'):
+ match = '(' + str(match) + ')'
+ line = str(source) + ' ' + match + ' ' + _actionstr(actions)
+ return line
def _actionstr(actions):
"""
_______________________________________________
tmda-cvs mailing list
http://libertine.org/lists/listinfo/tmda-cvs
|
