Activities are not queued (correctly) during a project snychronization.
1. During the compression part of the archive file (this is the time where every user is blocked), no requests are send that inform the other side that queuing can now be started.
2. Current mechanism only checks EditorActivities. This is wrong, all activities that operates on resources (have an SPath object) must be queued.
3. A project is internally registered before the archive has received and is decompressed, thus the queuing algorithm will not work for almost 100% of the time.
4. The queuing is not properly synchronized, it is possible that some latest received activities can be executed before the queue is flushed.
Results of that broken functionality are:
#ID: 3516639
#ID: 3512804
#ID: 3458952
#ID: 3455405
Priority 6 -- not necessary for the next release.
In my opinion this should be fixed for the next release in august as this is causing many malfunctions which can only be recovered by the watchdog if this component is able to detect the issues, which is not always the case.
Agreed. As soon as the July release has been rolled-out, I will re-prioritize the open tasks, including this one.
Planned for Release 12.7.27
306198670159835e5547298ff703285b1f6e9e28, ad211fedfbaa113214316e5bf8dbc793afeedbf7 along with other commits that affects the OPN and IPN.
Needs to be verified.