diff --git a/openerp/addons/base/ir/ir_attachment.py b/openerp/addons/base/ir/ir_attachment.py
index dbfd1a82ced..f6bb00fed5f 100644
--- a/openerp/addons/base/ir/ir_attachment.py
+++ b/openerp/addons/base/ir/ir_attachment.py
@@ -242,6 +242,8 @@ class ir_attachment(osv.osv):
         # performed in batch as much as possible.
         ima = self.pool.get('ir.model.access')
         for model, targets in model_attachments.iteritems():
+            if model not in self.pool:
+                continue
             if not ima.check(cr, uid, model, 'read', False):
                 # remove all corresponding attachment ids
                 for attach_id in itertools.chain(*targets.values()):
diff --git a/openerp/addons/base/res/res_partner_report_address.xsl b/openerp/addons/base/res/res_partner_report_address.xsl
index ca212f2d2b0..83fb709e356 100644
--- a/openerp/addons/base/res/res_partner_report_address.xsl
+++ b/openerp/addons/base/res/res_partner_report_address.xsl
@@ -1,13 +1,13 @@
 <?xml version="1.0" encoding="utf-8"?>
 <xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:fo="http://www.w3.org/1999/XSL/Format">
-	<xsl:variable name="initial_bottom_pos">20.5</xsl:variable>
-	<xsl:variable name="initial_left_pos">1</xsl:variable>
-	<xsl:variable name="height_increment">6.5</xsl:variable>
-	<xsl:variable name="width_increment">10</xsl:variable>
-	<xsl:variable name="frame_height">5.5cm</xsl:variable>
-	<xsl:variable name="frame_width">10cm</xsl:variable>
-	<xsl:variable name="number_columns">2</xsl:variable>
-	<xsl:variable name="max_frames">8</xsl:variable>
+	<xsl:variable name="initial_bottom_pos">24.5</xsl:variable>
+	<xsl:variable name="initial_left_pos">0.5</xsl:variable>
+	<xsl:variable name="height_increment">3.5</xsl:variable>
+	<xsl:variable name="width_increment">7</xsl:variable>
+	<xsl:variable name="frame_height">3.7cm</xsl:variable>
+	<xsl:variable name="frame_width">7cm</xsl:variable>
+	<xsl:variable name="number_columns">3</xsl:variable>
+	<xsl:variable name="max_frames">24</xsl:variable>
 
 	<xsl:template match="/">
 		<xsl:apply-templates select="addresses"/>
diff --git a/openerp/addons/base/res/res_users_view.xml b/openerp/addons/base/res/res_users_view.xml
index ed52de5eba2..9c84d4326f4 100644
--- a/openerp/addons/base/res/res_users_view.xml
+++ b/openerp/addons/base/res/res_users_view.xml
@@ -274,7 +274,7 @@
             <field eval="18" name="priority"/>
             <field name="arch" type="xml">
                 <form string="Users" version="7.0">
-                    <field name="image" widget='image' class="oe_right oe_avatar" options='{"preview_image": "image_small"}'/>
+                    <field name="image" readonly="0" widget='image' class="oe_right oe_avatar" options='{"preview_image": "image_small"}'/>
                     <h1>
                         <field name="name" readonly="1" class="oe_inline"/>
                     </h1>
diff --git a/openerp/osv/fields.py b/openerp/osv/fields.py
index 4ae57881273..d9ecaa69b64 100644
--- a/openerp/osv/fields.py
+++ b/openerp/osv/fields.py
@@ -601,8 +601,10 @@ class one2many(_column):
                 else:
                     cr.execute('update '+_table+' set '+self._fields_id+'=null where id=%s', (act[1],))
             elif act[0] == 4:
-                # Must use write() to recompute parent_store structure if needed
-                obj.write(cr, user, [act[1]], {self._fields_id:id}, context=context or {})
+                cr.execute("select 1 from {0} where id=%s and {1}=%s".format(_table, self._fields_id), (act[1], id))
+                if not cr.fetchone():
+                    # Must use write() to recompute parent_store structure if needed and check access rules
+                    obj.write(cr, user, [act[1]], {self._fields_id:id}, context=context or {})
             elif act[0] == 5:
                 reverse_rel = obj._all_columns.get(self._fields_id)
                 assert reverse_rel, 'Trying to unlink the content of a o2m but the pointed model does not have a m2o'
diff --git a/openerp/osv/orm.py b/openerp/osv/orm.py
index 71053a02cae..610d0a5d503 100644
--- a/openerp/osv/orm.py
+++ b/openerp/osv/orm.py
@@ -3681,11 +3681,6 @@ class BaseModel(object):
         if fields_to_read is None:
             fields_to_read = self._columns.keys()
 
-        # Construct a clause for the security rules.
-        # 'tables' hold the list of tables necessary for the SELECT including the ir.rule clauses,
-        # or will at least contain self._table.
-        rule_clause, rule_params, tables = self.pool.get('ir.rule').domain_get(cr, user, self._name, 'read', context=context)
-
         # all inherited fields + all non inherited fields for which the attribute whose name is in load is True
         fields_pre = [f for f in fields_to_read if
                            f == self.CONCURRENCY_CHECK_FIELD
@@ -3706,6 +3701,11 @@ class BaseModel(object):
                     return 'length(%s) as "%s"' % (f_qual, f)
                 return f_qual
 
+            # Construct a clause for the security rules.
+            # 'tables' hold the list of tables necessary for the SELECT including the ir.rule clauses,
+            # or will at least contain self._table.
+            rule_clause, rule_params, tables = self.pool.get('ir.rule').domain_get(cr, user, self._name, 'read', context=context)
+
             fields_pre2 = map(convert_field, fields_pre)
             order_by = self._parent_order or self._order
             select_fields = ','.join(fields_pre2 + ['%s.id' % self._table])
@@ -3720,6 +3720,7 @@ class BaseModel(object):
                 self._check_record_rules_result_count(cr, user, sub_ids, result_ids, 'read', context=context)
                 res.extend(results)
         else:
+            self.check_access_rule(cr, user, ids, 'read', context=context)
             res = map(lambda x: {'id': x}, ids)
 
         if context.get('lang'):
diff --git a/openerp/tools/convert.py b/openerp/tools/convert.py
index 47f75e43fe8..6868b4fd897 100644
--- a/openerp/tools/convert.py
+++ b/openerp/tools/convert.py
@@ -952,7 +952,7 @@ def convert_csv_import(cr, module, fname, csvcontent, idref=None, mode='init',
     result, rows, warning_msg, dummy = registry[model].import_data(cr, uid, fields, datas,mode, module, noupdate, filename=fname_partial)
     if result < 0:
         # Report failed import and abort module install
-        raise Exception(_('Module loading failed: file %s/%s could not be processed:\n %s') % (module, fname, warning_msg))
+        raise Exception(_('Module loading %s failed: file %s could not be processed:\n %s') % (module, fname, warning_msg))
     if config.get('import_partial'):
         data = pickle.load(file(config.get('import_partial')))
         data[fname_partial] = 0