ZCatalog manage_beforeDelete bug
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Silva |
Fix Released
|
High
|
Unassigned | ||
Zope 2 |
Invalid
|
Low
|
Unassigned |
Bug Description
Almost there, a few points:
- If events are registered for base classes, of course they get triggered for
subclasses too. This had the interesting side effect that publications would
never show up in the SMI, because the 'added' event for both SilvaObject and
Folder called add_ordered_id, and that method is implemented in a particularly
stupid way: it calls refresh_ordered_ids directly, which adds the id if it's not
there, and removes it when it is. Hence, calling add_ordered_id twice will
result in removing the id again. I'll fix that.
- A tougher problem is that unindex_object() no longer seems to work for Silva
Root objects. I think the implementation is in zope itself, so I'll have to look
there for clues. This is what's causing some tests to fail.
Changed in silva: | |
assignee: | thisfred → aaltepet |
Changed in silva: | |
assignee: | thisfred → aaltepet |
Changed in silva: | |
assignee: | aaltepet → nobody |
Changed in silva: | |
status: | Fix Committed → Fix Released |
Changed in zope2: | |
importance: | Undecided → Low |
status: | New → Confirmed |
corollary of the first point:
some code duplication could be removed, so we have a few less events now.