Closed Bug 385156 Opened 19 years ago Closed 19 years ago

Performance regression on startup (hang)

Categories

(Calendar :: General, defect)

Product:
Calendar
Component:
General
Version:
Sunbird 0.5
Platform:
x86
Windows XP
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

(Not tracked)

Status:
RESOLVED WONTFIX

People

(Reporter: aryx, Unassigned)

Details

http://ftp.mozilla.org/pub/mozilla.org/calendar/sunbird/nightly/2007-06-04-05-mozilla1.8/ needs 18 seconds cpu time to launch and show events and tasks http://ftp.mozilla.org/pub/mozilla.org/calendar/sunbird/nightly/2007-06-08-03-mozilla1.8/ needs 25 seconds cpu time to launch and show events and tasks http://ftp.mozilla.org/pub/mozilla.org/calendar/sunbird/nightly/2007-06-12-05-mozilla1.8/ needs 38 seconds cpu time to launch and show events and tasks Notice: Around 30 tasks and 200 events active. Remote calendar and local ics calendar with around 500 events deactivated - it took minutes to load Sunbird.
Flags: blocking-calendar0.5?
Thanks to archaeopteryx, it shows up that the current todo pane causes this regression. As already mentioned in <https://bugzilla.mozilla.org/show_bug.cgi?id=384726#c6>, the pane's todos are loaded twice. IMO this is not blocking 0.5. WONTFIX for now, we should verify the bug once the agenda/todo pane is reworked.
Status: NEW → RESOLVED
Closed: 19 years ago
Flags: blocking-calendar0.5? → blocking-calendar0.5-
Resolution: --- → WONTFIX
Daniel, Bug 384726 was checked in 2007-06-19 and affects only Lightning. Therefore the performance regression can't be the caused by this. The only checkin between 2007-06-08 and 2007-06-12 was for Bug 382840.
(In reply to comment #2) > Daniel, Bug 384726 was checked in 2007-06-19 and affects only Lightning. > Therefore the performance regression can't be the caused by this. The only > checkin between 2007-06-08 and 2007-06-12 was for Bug 382840. Ok, my comment is incorrect/misleading. I don't say that bug 384726 (Lightning only) fixes the problem. IMO this performance bug is uncovered (again) by the ics/alarm fix, queuing getItems() calls until the ics file is fully loaded. Bug 384726 just showed up that Lightning's the todo list issues getItems() twice. I suspected the same problem applies to Sunbird; archaeopteryx has isolated the problem to ics subscriptions. Investigating Sunbird startup: XXX ICS action: refresh Starting calendar alarm service observer added XXX ICS action: get_items get_items: 10000000000001011 XXX ICS action: get_items get_items: 1000 XXX ICS action: get_items get_items: 111 XXX ICS action: get_items get_items: 10000000000011111 XXX ICS action: get_items get_items: 1000 XXX ICS action: get_items get_items: 111 XXX ICS action: get_items get_items: 1000 XXX ICS action: get_items get_items: 1000 XXX ICS action: get_items get_items: 10000000000001011 => 3x fetching todos: I suspect 1x alarm service (10000000000011111), 2x todo list (111, resources/content/unifinderToDo.js:192).
Flags: blocking-calendar0.5-
You need to log in before you can comment on or make changes to this bug.