Find out why Nightly build cold start-up is consistently slower than Try build cold start-up
Categories
(Toolkit :: Startup and Profile System, defect)
Tracking
()
People
(Reporter: mconley, Unassigned)
Details
I'm observing this effect when testing various experiments via Try builds on the 2018 reference device using frame recording.
Essentially, I'm noticing a dramatic drop in the "time to settled" (time that the about:home content finishes updating) in builds from Try, when compared to our official Nightly builds.
When I say dramatic drop, I mean on the order of seconds. It's really quite stark.
I suspect there are a number of things that we disable for Try builds that are enabled on Nightly builds. I expect those things to be:
- Crash reporting
- Automatic updating
- BHR?
- Probably other things?
I'd like to understand what Nightly is doing that's adding so many extra seconds of start-up time, because perhaps we can skip or delay it.
|
Reporter | |
Comment 1•7 years ago
|
||
The first thing I'd like to do is verify my suspicion here by doing some explicit frame recordings of two builds, one of our official Nightly's, and a Try shippable build against the same revision, and get a measurement of the difference.
|
|
Reporter | |
Comment 2•7 years ago
|
||
Huh. So, I think perhaps this was less to do with Try / Nightly builds, and more to do with timing; my comparison didn't actually show much difference between the two builds when I tested them (I used July 1st's Nightly, and a try build based on the same revision).
I'm closing this one out as WORKSFORME for now.
Description
•