480212 - nsUpdateService::_start takes 500ms

Status: RESOLVED DUPLICATE of bug 311965

(Toolkit :: Application Update, defect)

Description

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

On CE6, calling _start (in response to profile-after-change) takes 500ms even when there are no updates or anything else pending.  It really shouldn't take even 1/10th that long in that case...

Comment 1

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Note that this is maybe sorta related to bug 311965, but compiling nsUpdateService.js is quite fast; it's executing _start that takes all the time.

Comment 2

[Brian Crowder]

taras:  Have you profiled this at all?

Comment 3

[(dormant account)]

(In reply to comment #2)
> taras:  Have you profiled this at all?

yes and after discussions with Mossop we filed 471219, but nobody has taken that up yet. It should save us a fair bit of time.

Comment 4

[Benjamin Smedberg]

blocking- per crowder

Comment 5

[Robert Strong (they/them - no direct email)]

Vlad, could you check how long this is taking now? Taras' logs on Maemo show the following changes for nsUpdateService.js.

from 7/21
nsUpdateService.js : 109+2+28+36+25 = 200ms

from 10/2
nsUpdateService.js : 138+2+6+9+33   = 188ms

from 10/8 with patch from Bug 512651 which landed on trunk today
nsUpdateService.js : 78+2+7+2+30    = 119

So, it looks like there have been steady improvements and the _start doesn't take hardly any time for me on WinCE though that is without the js_Execute time.

Also, I'd like to dupe this over to Bug 311965 for the remaining work unless you'd prefer not to.