| Summary | Driver: $_count cache gets dirty on modifications |
| Queue | Turba |
| Queue Version | Git master |
| Type | Enhancement |
| State | Feedback |
| Priority | 1. Low |
| Owners | |
| Requester | thomas.jarosch (at) intra2net (dot) com |
| Created | 07/29/2013 (4307 days ago) |
| Due | |
| Updated | 07/30/2013 (4306 days ago) |
| Assigned | 07/29/2013 (4307 days ago) |
| Resolved | |
| Milestone | |
| Patch | No |
(-> not exported by the API). Could we also just remove / deprecate
it?
and is used by View/Browse.php only. Does not seem to be exported by the API.
I can take a look at this, though it's really low priority.
Let's get Kolab and ActiveSync playing nicely together first :)
called after a modification takes place in the same request.
However, these are the types of things that can later cause a
headache trying to track down if this behavior ever changes. So if
you provide a patch, I agree it should be committed.
(-> not exported by the API). Could we also just remove / deprecate it?
The real counting stuff in done in Turba_List.
Type ⇒ Enhancement
State ⇒ Feedback
Priority ⇒ 1. Low
State ⇒ Feedback
after a modification takes place in the same request.
However, these are the types of things that can later cause a headache
trying to track down if this behavior ever changes. So if you provide
a patch, I agree it should be committed.
Priority ⇒ 1. Low
Type ⇒ Bug
Summary ⇒ Driver: $_count cache gets dirty on modifcations
Queue ⇒ Turba
Milestone ⇒
Patch ⇒ No
State ⇒ Unconfirmed
while reading through the Turba driver code, I've noticed
there is an internal cache for the total count of contacts: $_count.
A generic count() implementation is in lib/Driver.php and a more specific
one for SQL in lib/Driver/Sql.php.
IMHO we should reset (=unset) the count cache in
lib/Driver/Sql.php:_add(), _delete() and _deleteAll().
Is the owner allowed to change in _save()? Then _modify() would be
affected, too.
Same goes for the generic count function.
Cheers,
Thomas