# HG changeset patch
# User Paul Boddie <paul@boddie.org.uk>
# Date 1268603628 -3600
# Node ID 59caeb446995a4cf2e7a3ff972db492957932252
# Parent  3b719c63fd172de896991b3c308366ea0efe22ad
Removed time zone abbreviations and improved UTC offset parsing.
Added UTC offset calculation for Olson identifier-derived time zones;
unfortunately, pytz version 2007k seems to return incorrect offsets for
"Europe/Helsinki" (with a datetime of '2010-03-14 22:52:26.311507', it returns
6000 seconds, not 7200 seconds as dateutil is able to).

diff -r 3b719c63fd17 -r 59caeb446995 EventAggregatorSupport.py
--- a/EventAggregatorSupport.py	Sun Mar 14 21:06:15 2010 +0100
+++ b/EventAggregatorSupport.py	Sun Mar 14 22:53:48 2010 +0100
@@ -50,13 +50,17 @@
 month_regexp_str = ur'(?P<year>[0-9]{4})-(?P<month>[0-9]{2})'
 date_regexp_str = ur'(?P<year>[0-9]{4})-(?P<month>[0-9]{2})-(?P<day>[0-9]{2})'
 time_regexp_str = ur'(?P<hour>[0-2][0-9]):(?P<minute>[0-5][0-9])(?::(?P<second>[0-6][0-9]))?'
-timezone_regexp_str = ur'(?P<zone>[A-Z]{3,}|[a-zA-Z]+/[-_a-zA-Z]+|[-+][0-9]{1,4})'
+timezone_olson_str = ur'(?P<olson>[a-zA-Z]+/[-_a-zA-Z]+)'
+timezone_offset_str = ur'(?:UTC)?(?P<sign>[-+])(?P<hours>[0-9]{2})(?::?(?P<minutes>[0-9]{2}))?'
+timezone_regexp_str = ur'(?P<zone>' + timezone_olson_str + '|' + timezone_offset_str + ')'
 datetime_regexp_str = date_regexp_str + ur'(?:\s+' + time_regexp_str + ur'(?:\s+' + timezone_regexp_str + ur')?)?'
 
 month_regexp = re.compile(month_regexp_str, re.UNICODE)
 date_regexp = re.compile(date_regexp_str, re.UNICODE)
 time_regexp = re.compile(time_regexp_str, re.UNICODE)
 datetime_regexp = re.compile(datetime_regexp_str, re.UNICODE)
+timezone_olson_regexp = re.compile(timezone_olson_str, re.UNICODE)
+timezone_offset_regexp = re.compile(timezone_offset_str, re.UNICODE)
 
 verbatim_regexp = re.compile(ur'(?:'
     ur'<<Verbatim\((?P<verbatim>.*?)\)>>'
@@ -1053,25 +1057,31 @@
 
         zone = self.time_zone()
 
-        # Only attempt to return a UTC offset where an explicit offset has been
-        # set.
+        # Attempt to return a UTC offset where an explicit offset has been set.
 
-        if zone and zone[0] in "-+":
-            digits = zone[1:]
-            if zone[0] == "-":
+        match = timezone_offset_regexp.match(zone)
+        if match:
+            if match.group("sign") == "-":
                 sign = -1
             else:
                 sign = 1
 
-            if 1 <= len(digits) <= 2:
-                return int(digits) * sign, 0
-            elif len(digits) == 3:
-                hours = int(digits[:1]) * sign
-                minutes = int(digits[1:]) * sign
-                return hours, minutes
-            elif len(digits) == 4:
-                hours = int(digits[:2]) * sign
-                minutes = int(digits[2:]) * sign
+            hours = int(match.group("hours")) * sign
+            minutes = int(match.group("minutes") or 0) * sign
+            return hours, minutes
+
+        # Attempt to return an offset where an Olson identifier has been given.
+
+        match = timezone_olson_regexp.match(zone)
+        if match:
+            if pytz is not None:
+                tz = pytz.timezone(match.group("olson"))
+
+                # NOTE: Using internal attribute.
+
+                seconds = tz._utcoffset.seconds
+                hours = seconds / 3600
+                minutes = (seconds - hours * 3600) / 60
                 return hours, minutes
 
         return None
@@ -1147,9 +1157,9 @@
     if m:
         groups = list(m.groups())
 
-        # Convert all but the zone to integer or None.
+        # Convert date and time data to integer or None.
 
-        return DateTime(map(int_or_none, groups[:-1]) + groups[-1:])
+        return DateTime(map(int_or_none, groups[:6]) + [m.group("zone")])
     else:
         return None