ignore UIDs in http storage

CHANGELOG.rst

@@ -19,6 +19,8 @@ Version 0.13.0
   console output is always unambigous. See :gh:`459`.
 - Custom commands can now be used for conflict resolution during sync. See
   :gh:`127`.
+- :storage:`http` now completely ignores UIDs. This avoids a lot of unnecessary
+  down- and uploads.
 
 Version 0.12.1
 ==============

tests/storage/test_http.py

@@ -54,7 +54,7 @@ def test_list(monkeypatch):
 
     for href, etag in s.list():
         item, etag2 = s.get(href)
-        assert item.uid is None
+        assert item.uid is not None
         assert etag2 == etag
         found_items[normalize_item(item)] = href
 
@@ -65,7 +65,7 @@ def test_list(monkeypatch):
 
     for href, etag in s.list():
         item, etag2 = s.get(href)
-        assert item.uid is None
+        assert item.uid is not None
         assert etag2 == etag
         assert found_items[normalize_item(item)] == href
 

tests/storage/test_http_with_singlefile.py

@@ -24,6 +24,7 @@ class CombinedStorage(Storage):
         self.url = url
         self.path = path
         self._reader = vdirsyncer.storage.http.HttpStorage(url=url)
+        self._reader._ignore_uids = False
         self._writer = SingleFileStorage(path=path)
 
     def list(self, *a, **kw):

vdirsyncer/repair.py

@@ -17,6 +17,8 @@ def repair_storage(storage):
         logger.info(u'[{}/{}] Processing {}'
                     .format(i, len(all_hrefs), href))
 
+        new_item = item
+
         changed = False
         if item.parsed is None:
             logger.warning('Item {} can\'t be parsed, skipping.'
@@ -25,15 +27,14 @@ def repair_storage(storage):
 
         if not item.uid:
             logger.warning('No UID, assigning random one.')
-            changed = change_uid(item, generate_href()) or changed
+            new_item = item.with_uid(generate_href())
         elif item.uid in seen_uids:
             logger.warning('Duplicate UID, assigning random one.')
-            changed = change_uid(item, generate_href()) or changed
+            new_item = item.with_uid(generate_href())
         elif not href_safe(item.uid) or not href_safe(basename(href)):
             logger.warning('UID or href is unsafe, assigning random UID.')
-            changed = change_uid(item, generate_href(item.uid)) or changed
+            new_item = item.with_uid(generate_href(item.uid))
 
-        new_item = Item(u'\r\n'.join(item.parsed.dump_lines()))
         if not new_item.uid:
             logger.error('Item {!r} is malformed beyond repair. '
                          'This is a serverside bug.'
@@ -42,7 +43,7 @@ def repair_storage(storage):
             continue
 
         seen_uids.add(new_item.uid)
-        if changed:
+        if new_item.raw != item.raw:
             try:
                 if new_item.uid != item.uid:
                     storage.upload(new_item)
@@ -53,15 +54,3 @@ def repair_storage(storage):
                 logger.exception('Server rejected new item.')
 
 
-def change_uid(item, new_uid):
-    stack = [item.parsed]
-    changed = False
-    while stack:
-        component = stack.pop()
-        stack.extend(component.subcomponents)
-
-        if component.name in ('VEVENT', 'VTODO', 'VJOURNAL', 'VCARD'):
-            component['UID'] = new_uid
-            changed = True
-
-    return changed

vdirsyncer/storage/http.py

@@ -95,6 +95,16 @@ HTTP_STORAGE_PARAMETERS = '''
 class HttpStorage(Storage):
     __doc__ = '''
     Use a simple ``.ics`` file (or similar) from the web.
+    ``webcal://``-calendars are supposed to be used with this, but you have to
+    replace ``webcal://`` with ``http://``, or better, ``https://``.
+
+    Too many WebCAL providers generate UIDs of all ``VEVENT``-components
+    on-the-fly, i.e. all UIDs change every time the calendar is downloaded.
+    This leads many synchronization programs to believe that all events have
+    been deleted and new ones created, and accordingly causes a lot of
+    unnecessary uploads and deletions on the other side. Vdirsyncer completely
+    ignores UIDs coming from :storage:`http` and will replace them with a hash
+    of the normalized item content.
 
     :param url: URL to the ``.ics`` file.
     ''' + HTTP_STORAGE_PARAMETERS + '''
@@ -121,6 +131,9 @@ class HttpStorage(Storage):
     _repr_attributes = ('username', 'url')
     _items = None
 
+    # Required for tests.
+    _ignore_uids = True
+
     def __init__(self, url, username='', password='', verify=True, auth=None,
                  useragent=USERAGENT, verify_fingerprint=None, auth_cert=None,
                  **kwargs):
@@ -152,8 +165,10 @@ class HttpStorage(Storage):
 
         for item in split_collection(r.text):
             item = Item(item)
-            etag = item.hash
+            if self._ignore_uids:
-            self._items[item.ident] = item, etag
+                item = item.with_uid(item.hash)
+
+            self._items[item.ident] = item, item.hash
 
         return ((href, etag) for href, (item, etag) in self._items.items())
 

vdirsyncer/utils/vobject.py

@@ -40,6 +40,20 @@ class Item(object):
         assert isinstance(raw, str)
         self._raw = raw
 
+    def with_uid(self, new_uid):
+        parsed = _Component.parse(self.raw)
+        stack = [parsed]
+        while stack:
+            component = stack.pop()
+            stack.extend(component.subcomponents)
+
+            if component.name in ('VEVENT', 'VTODO', 'VJOURNAL', 'VCARD'):
+                del component['UID']
+                if new_uid:
+                    component['UID'] = new_uid
+
+        return Item('\r\n'.join(parsed.dump_lines()))
+
     @cached_property
     def raw(self):
         '''Raw content of the item, as unicode string.
@@ -79,8 +93,9 @@ class Item(object):
         # 2. The status file would contain really sensitive information.
         return self.uid or self.hash
 
-    @cached_property
+    @property
     def parsed(self):
+        '''Don't cache because the rv is mutable.'''
         try:
             return _Component.parse(self.raw)
         except Exception: