Fix method to check size of file size in dexterity (thanks @hvelarde)…

transmogrify/wordpress/blueprints/fetchattachment.py

@@ -1,6 +1,7 @@
 # -*- coding: utf-8 -*-
 from collective.transmogrifier.interfaces import ISection
 from collective.transmogrifier.interfaces import ISectionBlueprint
+from plone.dexterity.interfaces import IDexterityContent
 from requests.exceptions import ConnectionError
 from requests.exceptions import RequestException
 from transmogrify.wordpress.logger import logger
@@ -69,8 +70,12 @@ def __iter__(self):
                 # content-length header could be missing if remote web
                 # server is misconfigured for some mime types
                 size = int(r.headers.get('content-length', 0))
+                if IDexterityContent.providedBy(obj):
+                    objsize = obj.get_size()
+                else:  # Archetypes
+                    objsize = obj.size()
 
-                if size == obj.size():  # already downloaded it
+                if size == objsize:  # already downloaded it
                     yield item
                     continue
 
@@ -83,7 +88,7 @@ def __iter__(self):
 
             if r.status_code != 200:  # log error and skip item
                 fetch_errors.append(url)
-                msg = u'Error {0} when fetching {1}'.format(r.status_code, url)
+                msg = 'Error {0} when fetching {1}'.format(r.status_code, url)
                 logger.warn(msg)
                 yield item
                 continue

transmogrify/wordpress/blueprints/moveattachment.py

@@ -4,6 +4,7 @@
 from collective.transmogrifier.utils import defaultMatcher
 from collective.transmogrifier.utils import traverse
 from plone import api
+from plone.dexterity.interfaces import IDexterityContent
 from zope.interface import classProvides
 from zope.interface import implements
 
@@ -47,7 +48,12 @@ def __iter__(self):
             if obj.portal_type not in ('Image', 'File'):
                 continue
 
-            references = obj.getBackReferences()
+            if IDexterityContent.providedBy(obj):
+                from Products.Archetypes.interfaces import IReferenceable
+                adapted = IReferenceable(obj)
+                references = adapted.getBackReferences()
+            else:  # Archetypes
+                references = obj.getBackReferences()
             if len(references) != 1:
                 continue
 

transmogrify/wordpress/blueprints/resolveuid.py

@@ -7,7 +7,6 @@
 from plone.app.linkintegrity.handlers import modifiedDexterity
 from plone.app.textfield.value import RichTextValue
 from plone.dexterity.interfaces import IDexterityContent
-from Products.Archetypes.interfaces import IBaseObject
 from transmogrify.wordpress.logger import logger
 from transmogrify.wordpress.utils import fix_id
 from urlparse import urlparse
@@ -74,10 +73,10 @@ def __iter__(self):
 
             # XXX: this seems to be very expensive
             # Update linkintegrity references
-            if IBaseObject.providedBy(obj):
-                modifiedArchetype(obj, event=None)
-            elif IDexterityContent.providedBy(obj):
+            if IDexterityContent.providedBy(obj):
                 modifiedDexterity(obj, event=None)
+            else:  # Archetypes
+                modifiedArchetype(obj, event=None)
 
     def resolve_uid(self, x):
         """Parse HTML and update with URLs pointing to Plone objects.

transmogrify/wordpress/utils.py

@@ -1,7 +1,9 @@
 # -*- coding: utf-8 -*-
 from HTMLParser import HTMLParser
+from plone.i18n.normalizer.interfaces import IIDNormalizer
 from Products.CMFPlone.utils import safe_unicode
 from urllib import unquote_plus
+from zope.component import getUtility
 
 
 class MLStripper(HTMLParser):
@@ -33,5 +35,6 @@ def fix_id(item_id):
     """Unquote plus and remove accents from ids."""
     item_id = unquote_plus(item_id)
     item_id = safe_unicode(item_id)
-    item_id = item_id.encode('ascii', 'ignore')
+    id_normalizer = getUtility(IIDNormalizer)
+    item_id = id_normalizer.normalize(item_id)
     return item_id