© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
73416ba
(Get the code!)
Recent work on the "Offload Queue" has apparently limited to 200 the number of events processed by syncdaemon.
Pending operations above this number are stored on disk, and they are only retrieved from disk when older operations are completed.
This change has cut on the memory usage, but for status aggregator (the code that shows the notification bubbles) this looks like the operations have not been queued yet.
After a quick look I came with two possible solutions:
1) modify the messages to say "more that 100" if more than 100 operations are scheduled.
2) let the status aggregator take a peek at the operations before storing them on disk.
Option 2 seems like a much bigger change, since we'll need to make sure that all in-memory metadata used by the aggregator is up to date at the point where it looks at the operations. Also, we would need to make sure that the aggregator holds *no* references to pending operations, so they can be safely serialized.
Anyway, I'm sure we can come up with a better ways to fix this.