From 8f5fdfb2d4ea5f5388292d860b7d2b0c9613042d Mon Sep 17 00:00:00 2001
From: Markus Unterwaditzer <markus@unterwaditzer.net>
Date: Sat, 17 May 2014 15:20:46 +0200
Subject: [PATCH] Hash items resilient against formatting

---
 tests/__init__.py           |  2 ++
 tests/utils/test_vobject.py | 13 +++++++++++--
 vdirsyncer/storage/base.py  |  9 +++------
 vdirsyncer/utils/vobject.py | 15 +++++++++++++++
 4 files changed, 31 insertions(+), 8 deletions(-)

diff --git a/tests/__init__.py b/tests/__init__.py
index 67706f2..e013bde 100644
--- a/tests/__init__.py
+++ b/tests/__init__.py
@@ -85,4 +85,6 @@ END:VCALENDAR'''
 
 SIMPLE_TEMPLATE = u'''BEGIN:FOO
 UID:{r}
+X-SOMETHING:{r}
+HAHA:YES
 END:FOO'''
diff --git a/tests/utils/test_vobject.py b/tests/utils/test_vobject.py
index 9fed788..4cddd6a 100644
--- a/tests/utils/test_vobject.py
+++ b/tests/utils/test_vobject.py
@@ -7,9 +7,11 @@
     :license: MIT, see LICENSE for more details.
 '''
 
-from vdirsyncer.utils.vobject import split_collection, join_collection
+from vdirsyncer.utils.vobject import split_collection, join_collection, \
+    hash_item
 
-from .. import normalize_item, SIMPLE_TEMPLATE, BARE_EVENT_TEMPLATE
+from .. import normalize_item, SIMPLE_TEMPLATE, BARE_EVENT_TEMPLATE, \
+    EVENT_TEMPLATE
 
 
 _simple_joined = u'\r\n'.join((
@@ -74,3 +76,10 @@ def test_split_collection_timezones():
     )
 
     assert given == expected
+
+
+def test_hash_item():
+    a = EVENT_TEMPLATE.format(r=1)
+    b = u'\n'.join(line for line in a.splitlines()
+                   if u'PRODID' not in line and u'VERSION' not in line)
+    assert hash_item(a) == hash_item(b)
diff --git a/vdirsyncer/storage/base.py b/vdirsyncer/storage/base.py
index 6b78575..1d71425 100644
--- a/vdirsyncer/storage/base.py
+++ b/vdirsyncer/storage/base.py
@@ -7,7 +7,6 @@
     :license: MIT, see LICENSE for more details.
 '''
 
-import hashlib
 
 from .. import exceptions
 from .. import utils
@@ -35,17 +34,15 @@ class Item(object):
 
     def __init__(self, raw):
         assert isinstance(raw, utils.text_type)
-        raw = raw.splitlines()
 
-        for line in raw:
+        for line in raw.splitlines():
             if line.startswith(u'UID:'):
                 uid = line[4:].strip()
                 if uid:
                     self.uid = uid
 
-        self.raw = u'\n'.join(raw)
-        self.hash = hashlib.sha256(
-            self.raw.strip().encode('utf-8')).hexdigest()
+        self.raw = raw
+        self.hash = utils.vobject.hash_item(raw)
         self.ident = self.uid or self.hash
 
 
diff --git a/vdirsyncer/utils/vobject.py b/vdirsyncer/utils/vobject.py
index 2c6fdea..86a80fc 100644
--- a/vdirsyncer/utils/vobject.py
+++ b/vdirsyncer/utils/vobject.py
@@ -6,12 +6,27 @@
     :copyright: (c) 2014 Markus Unterwaditzer
     :license: MIT, see LICENSE for more details.
 '''
+import hashlib
+
 import icalendar.cal
 import icalendar.parser
 
 from . import text_type, itervalues
 
 
+def hash_item(text):
+    try:
+        lines = to_unicode_lines(icalendar.cal.Component.from_ical(text))
+    except Exception:
+        lines = sorted(text.splitlines())
+
+    hashable = u'\r\n'.join(line.strip() for line in lines
+                            if line.strip() and
+                            u'PRODID' not in line and
+                            u'VERSION' not in line)
+    return hashlib.sha256(hashable.encode('utf-8')).hexdigest()
+
+
 def split_collection(text, inline=(u'VTIMEZONE',),
                      wrap_items_with=(u'VCALENDAR',)):
     '''Emits items in the order they occur in the text.'''