v2 format issues and fix local_to_utc zdump tests

5 files changed, 153 insertions, 28 deletions

diff --git a/gen_tests.py b/gen_tests.py
index 14d8512..82140dc 100644
--- a/gen_tests.py
+++ b/gen_tests.py
@@ -23,7 +23,11 @@ def main():
         print 'Collecting zdump(1) output for %s in zdump.out' % (zone,)
         tname = zone.replace(
                 '+', '_plus_').replace('-', '_minus_').replace('/','_')
-        zd_out, zd_in = popen2.popen2('%s -v -c 1800,2038 %s' % (zdump, zone))
+        # We don't yet support v2 format tzfile(5) files, so limit
+        # the daterange we test against - zdump understands v2 format
+        # files and will output historical records we can't cope with
+        # otherwise.
+        zd_out, zd_in = popen2.popen2('%s -v -c 1901,2038 %s' % (zdump, zone))
         zd_in.close()
         # Skip bogus output on 64bit architectures, per Bug #213816
         lines = [
diff --git a/src/pytz/tests/test_tzinfo.py b/src/pytz/tests/test_tzinfo.py
index d79b26f..313a2e2 100644
--- a/src/pytz/tests/test_tzinfo.py
+++ b/src/pytz/tests/test_tzinfo.py
@@ -26,6 +26,22 @@ NOTIME = timedelta(0)
 UTC = pytz.timezone('UTC')
 GMT = pytz.timezone('GMT')
 
+
+def prettydt(dt):
+    """datetime as a string using a known format.
+
+    We don't use strftime as it doesn't handle years earlier than 1900
+    per http://bugs.python.org/issue1777412
+    """
+    if dt.utcoffset() >= timedelta(0):
+        offset = '+%s' % (dt.utcoffset(),)
+    else:
+        offset = '-%s' % (-1 * dt.utcoffset(),)
+    return '%04d-%02d-%02d %02d:%02d:%02d %s %s' % (
+        dt.year, dt.month, dt.day,
+        dt.hour, dt.minute, dt.second,
+        dt.tzname(), offset)
+
 class BasicTest(unittest.TestCase):
 
     def testVersion(self):
@@ -194,13 +210,15 @@ class USEasternDSTStartTestCase(unittest.TestCase):
 
             # Make sure arithmetic crossing DST boundaries ends
             # up in the correct timezone after normalization
+            utc_plus_delta = (utc_dt + delta).astimezone(self.tzinfo)
+            local_plus_delta = self.tzinfo.normalize(dt + delta)
             self.failUnlessEqual(
-                    (utc_dt + delta).astimezone(self.tzinfo).strftime(fmt),
-                    self.tzinfo.normalize(dt + delta).strftime(fmt),
+                    prettydt(utc_plus_delta),
+                    prettydt(local_plus_delta),
                     'Incorrect result for delta==%d days.  Wanted %r. Got %r'%(
                         days,
-                        (utc_dt + delta).astimezone(self.tzinfo).strftime(fmt),
-                        self.tzinfo.normalize(dt + delta).strftime(fmt),
+                        prettydt(utc_plus_delta),
+                        prettydt(local_plus_delta),
                         )
                     )
 
@@ -340,7 +358,6 @@ class VilniusCESTStartTestCase(USEasternDSTStartTestCase):
     # causing the clocks to go backwards for this summer time
     # switchover.
     tzinfo = pytz.timezone('Europe/Vilnius')
-    instant = timedelta(seconds=31)
     transition_time = datetime(1941, 6, 23, 21, 00, 00, tzinfo=UTC)
     before = {
         'tzname': 'MSK',
@@ -354,6 +371,109 @@ class VilniusCESTStartTestCase(USEasternDSTStartTestCase):
         }
 
 
+class LondonHistoryStartTestCase(USEasternDSTStartTestCase):
+    # The first known timezone transition in London was in 1847 when
+    # clocks where synchronized to GMT. However, we currently only
+    # understand v1 format tzfile(5) files which does handle years
+    # this far in the past, so our earliest known transition is in
+    # 1916.
+    tzinfo = pytz.timezone('Europe/London')
+    # transition_time = datetime(1847, 12, 1, 1, 15, 00, tzinfo=UTC)
+    # before = {
+    #     'tzname': 'LMT',
+    #     'utcoffset': timedelta(minutes=-75),
+    #     'dst': timedelta(0),
+    #     }
+    # after = {
+    #     'tzname': 'GMT',
+    #     'utcoffset': timedelta(0),
+    #     'dst': timedelta(0),
+    #     }
+    transition_time = datetime(1916, 5, 21, 2, 00, 00, tzinfo=UTC)
+    before = {
+        'tzname': 'GMT',
+        'utcoffset': timedelta(0),
+        'dst': timedelta(0),
+        }
+    after = {
+        'tzname': 'BST',
+        'utcoffset': timedelta(hours=1),
+        'dst': timedelta(hours=1),
+        }
+
+
+class LondonHistoryEndTestCase(USEasternDSTStartTestCase):
+    # Timezone switchovers are projected into the future, even
+    # though no official statements exist or could be believed even
+    # if they did exist. We currently only check the last known
+    # transition in 2037, as we are still using v1 format tzfile(5)
+    # files.
+    tzinfo = pytz.timezone('Europe/London')
+    # transition_time = datetime(2499, 10, 25, 1, 0, 0, tzinfo=UTC)
+    transition_time = datetime(2037, 10, 25, 1, 0, 0, tzinfo=UTC)
+    before = {
+        'tzname': 'BST',
+        'utcoffset': timedelta(hours=1),
+        'dst': timedelta(hours=1),
+        }
+    after = {
+        'tzname': 'GMT',
+        'utcoffset': timedelta(0),
+        'dst': timedelta(0),
+        }
+
+
+class NoumeaHistoryStartTestCase(USEasternDSTStartTestCase):
+    # Noumea adopted a whole hour offset in 1912. Previously
+    # it was 11 hours, 5 minutes and 48 seconds off UTC. However,
+    # due to limitations of the Python datetime library, we need
+    # to round that to 11 hours 6 minutes.
+    tzinfo = pytz.timezone('Pacific/Noumea')
+    transition_time = datetime(1912, 1, 12, 12, 54, 12, tzinfo=UTC)
+    before = {
+        'tzname': 'LMT',
+        'utcoffset': timedelta(hours=11, minutes=6),
+        'dst': timedelta(0),
+        }
+    after = {
+        'tzname': 'NCT',
+        'utcoffset': timedelta(hours=11),
+        'dst': timedelta(0),
+        }
+
+
+class NoumeaDSTEndTestCase(USEasternDSTStartTestCase):
+    # Noumea dropped DST in 1997.
+    tzinfo = pytz.timezone('Pacific/Noumea')
+    transition_time = datetime(1997, 3, 1, 15, 00, 00, tzinfo=UTC)
+    before = {
+        'tzname': 'NCST',
+        'utcoffset': timedelta(hours=12),
+        'dst': timedelta(hours=1),
+        }
+    after = {
+        'tzname': 'NCT',
+        'utcoffset': timedelta(hours=11),
+        'dst': timedelta(0),
+        }
+
+
+class NoumeaNoMoreDSTTestCase(NoumeaDSTEndTestCase):
+    # Noumea dropped DST in 1997. Here we test that it stops occuring.
+    tzinfo = pytz.timezone('Pacific/Noumea')
+    transition_time = datetime(2037, 3, 1, 15, 00, 00, tzinfo=UTC)
+    before = {
+        'tzname': 'NCT',
+        'utcoffset': timedelta(hours=11),
+        'dst': timedelta(0),
+        }
+    after = {
+        'tzname': 'NCT',
+        'utcoffset': timedelta(hours=11),
+        'dst': timedelta(0),
+        }
+
+
 class ReferenceUSEasternDSTStartTestCase(USEasternDSTStartTestCase):
     tzinfo = reference.Eastern
     def test_arithmetic(self):
diff --git a/src/pytz/tzfile.py b/src/pytz/tzfile.py
index b5f818f..ea8032a 100644
--- a/src/pytz/tzfile.py
+++ b/src/pytz/tzfile.py
@@ -12,12 +12,12 @@ from pytz.tzinfo import memorized_datetime, memorized_timedelta
 
 
 def build_tzinfo(zone, fp):
-    head_fmt = '>4s 16x 6l'
+    head_fmt = '>4s c 15x 6l'
     head_size = calcsize(head_fmt)
-    (magic,ttisgmtcnt,ttisstdcnt,leapcnt,
-     timecnt,typecnt,charcnt) =  unpack(head_fmt, fp.read(head_size))
+    (magic, format, ttisgmtcnt, ttisstdcnt,leapcnt, timecnt,
+        typecnt, charcnt) =  unpack(head_fmt, fp.read(head_size))
 
-    # Make sure it is a tzinfo(5) file
+    # Make sure it is a tzfile(5) file
     assert magic == 'TZif'
 
     # Read out the transition times, localtime indices and ttinfo structures.
diff --git a/src/pytz/tzinfo.py b/src/pytz/tzinfo.py
index 55afcb2..ed035a3 100644
--- a/src/pytz/tzinfo.py
+++ b/src/pytz/tzinfo.py
@@ -68,13 +68,13 @@ class BaseTzInfo(tzinfo):
 class StaticTzInfo(BaseTzInfo):
     '''A timezone that has a constant offset from UTC
 
-    These timezones are rare, as most regions have changed their
-    offset from UTC at some point in their history
+    These timezones are rare, as most locations have changed their
+    offset at some point in their history
     '''
     def fromutc(self, dt):
         '''See datetime.tzinfo.fromutc'''
         return (dt + self._utcoffset).replace(tzinfo=self)
-    
+
     def utcoffset(self,dt):
         '''See datetime.tzinfo.utcoffset'''
         return self._utcoffset
@@ -110,11 +110,10 @@ class StaticTzInfo(BaseTzInfo):
 
 class DstTzInfo(BaseTzInfo):
     '''A timezone that has a variable offset from UTC
-   
-    The offset might change if daylight savings time comes into effect,
-    or at a point in history when the region decides to change their 
-    timezone definition. 
 
+    The offset might change if daylight savings time comes into effect,
+    or at a point in history when the region decides to change their
+    timezone definition.
     '''
     # Overridden in subclass
     _utc_transition_times = None # Sorted list of DST transition times in UTC
@@ -180,7 +179,6 @@ class DstTzInfo(BaseTzInfo):
         >>> before = eastern.normalize(before)
         >>> before.strftime(fmt)
         '2002-10-27 01:50:00 EDT (-0400)'
-
         '''
         if dt.tzinfo is None:
             raise ValueError, 'Naive time - no tzinfo set'
@@ -194,13 +192,13 @@ class DstTzInfo(BaseTzInfo):
 
     def localize(self, dt, is_dst=False):
         '''Convert naive time to local time.
-        
+
         This method should be used to construct localtimes, rather
         than passing a tzinfo argument to a datetime constructor.
 
         is_dst is used to determine the correct timezone in the ambigous
         period at the end of daylight savings time.
-        
+
         >>> from pytz import timezone
         >>> fmt = '%Y-%m-%d %H:%M:%S %Z (%z)'
         >>> amdam = timezone('Europe/Amsterdam')
@@ -223,7 +221,7 @@ class DstTzInfo(BaseTzInfo):
         AmbiguousTimeError: 2004-10-31 02:00:00
 
         is_dst defaults to False
-        
+
         >>> amdam.localize(dt) == amdam.localize(dt, False)
         True
 
@@ -324,7 +322,7 @@ class DstTzInfo(BaseTzInfo):
                     )
         filtered_possible_loc_dt.sort(mycmp)
         return filtered_possible_loc_dt[0]
-        
+
     def utcoffset(self, dt):
         '''See datetime.tzinfo.utcoffset'''
         return self._utcoffset
@@ -388,7 +386,7 @@ class NonExistentTimeError(InvalidTimeError):
 
 def unpickler(zone, utcoffset=None, dstoffset=None, tzname=None):
     """Factory function for unpickling pytz tzinfo instances.
-    
+
     This is shared for both StaticTzInfo and DstTzInfo instances, because
     database changes could cause a zones implementation to switch between
     these two base classes and we can't break pickles on a pytz version
diff --git a/test_zdump.py b/test_zdump.py
index 6e6299c..9e5b440 100644
--- a/test_zdump.py
+++ b/test_zdump.py
@@ -19,7 +19,6 @@ class ZdumpTestCase(unittest.TestCase):
             loc_dt.replace(tzinfo=None))
 
     def local_to_utc_check(self, zone, utc_dt, loc_dt, loc_tzname, is_dst):
-        loc_tz = pytz.timezone(zone)
         self.failUnlessEqual(
             loc_dt.astimezone(pytz.utc).replace(tzinfo=None),
             utc_dt.replace(tzinfo=None))
@@ -43,6 +42,8 @@ def test_suite():
         else:
             raise RuntimeError, 'Dud line %r' % (line,)
 
+        is_dst = bool(int(is_dst))
+
         if zone != last_zone:
             classname = zone.replace(
                     '+', '_plus_').replace('-', '_minus_').replace('/','_')
@@ -88,19 +89,21 @@ def test_suite():
             is_dst=is_dst):
             self.utc_to_local_check(zone, utc_dt, loc_dt, tzname, is_dst)
         test_utc_to_local.__name__ = test_name
-        #test_utc_to_local.__doc__ = line
         setattr(test_class, test_name, test_utc_to_local)
 
         if not skip_local:
             test_name = 'test_local_to_utc_%04d_%02d_%02d_%02d_%02d_%02d' % (
-                utc_dt.year, utc_dt.month, utc_dt.day,
-                utc_dt.hour, utc_dt.minute, utc_dt.second)
+                loc_dt.year, loc_dt.month, loc_dt.day,
+                loc_dt.hour, loc_dt.minute, loc_dt.second)
+            if is_dst:
+                test_name += '_dst'
+            else:
+                test_name += '_nodst'
             def test_local_to_utc(
                 self, zone=zone, utc_dt=utc_dt, loc_dt=loc_dt, tzname=tzname,
                 is_dst=is_dst):
                 self.local_to_utc_check(zone, utc_dt, loc_dt, tzname, is_dst)
             test_local_to_utc.__name__ = test_name
-            #test_local_to_utc.__doc__ = line
             setattr(test_class, test_name, test_local_to_utc)