[FIX] website: make filter_duplicate work

itertools' groupby function expects the iterable to be sorted as
hinted to in the documentation:

"Generally, the iterable needs to already be sorted on the same key
function."

When running groupby on an unsorted iterable non-adjacent duplicates
will remain:

>>> from itertools import groupby
>>> [e[0] for e in groupby([1, 2, 1])]
[1, 2, 1]

Because of this filter_duplicate would occassionally return
duplicates. To resolve this sort on the key field. Afterwards sort the
result on the usual inherit order: (priority, id).

1 files changed, 2 insertions, 2 deletions

diff --git a/addons/website/models/ir_ui_view.py b/addons/website/models/ir_ui_view.py
index c3c9de8016d..9a793b84fe6 100644
--- a/addons/website/models/ir_ui_view.py
+++ b/addons/website/models/ir_ui_view.py
@@ -55,9 +55,9 @@ class View(models.Model):
     def filter_duplicate(self):
         """ Filter current recordset only keeping the most suitable view per distinct key """
         filtered = self.env['ir.ui.view']
-        for dummy, group in groupby(self, key=lambda record: record.key):
+        for dummy, group in groupby(self.sorted('key'), key=lambda record: record.key):
             filtered += sorted(group, key=lambda record: record._sort_suitability_key())[0]
-        return filtered
+        return filtered.sorted(key=lambda view: (view.priority, view.id))
 
     @api.model
     def _view_obj(self, view_id):