DOMApplicationRegistry: get list of apps lazily

NEW
Assigned to

Status

()

Core
DOM: Apps
P4
normal
5 years ago
a year ago

People

(Reporter: gwagner, Assigned: fabrice)

Tracking

(Blocks: 1 bug, {perf})

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: [c=progress p= u= s=])

Attachments

(1 attachment)

(Reporter)

Description

5 years ago
We spend a little bit more than 0.5 sec getting the list of applications during a high workload scenario. This shows up during profiling incoming calls:

http://people.mozilla.com/~bgirard/cleopatra/#report=128e2cf57e726e038e7d4a05c0e72754db02a44f

http://mxr.mozilla.org/mozilla-central/source/dom/apps/src/AppsServiceChild.jsm#35
(Reporter)

Updated

5 years ago
Blocks: 847592
(Reporter)

Updated

5 years ago
Blocks: 860114
Gregor, should we block on this? It seems it blocks a blocker...
Flags: needinfo?(anygregor)
The fact is, we don't know yet what we need to do to fix bug 847592.  I'm a bit skeptical that this is needed or will help a lot.  But I was really wrong last time I made a similar guess.

Rather than mark a bunch of bugs as tef+ on suspicion they may be needed to fix a blocker, I'd rather that you give someone permission to tef+ whatever is actually needed to fix bug 847592.  I already have permission in bugzilla to do so.  Does that work for you?
(In reply to Justin Lebar [:jlebar] from comment #2)
> The fact is, we don't know yet what we need to do to fix bug 847592.  I'm a
> bit skeptical that this is needed or will help a lot.  But I was really
> wrong last time I made a similar guess.
> 
> Rather than mark a bunch of bugs as tef+ on suspicion they may be needed to
> fix a blocker, I'd rather that you give someone permission to tef+ whatever
> is actually needed to fix bug 847592.  I already have permission in bugzilla
> to do so.  Does that work for you?

Absolutely, just copy me in those bugs
(Assignee)

Comment 4

5 years ago
I don't know either if this fix will help. We will not spend the 0.5s that early, but at some point if we call any of the getXXX() function from AppsService we will need to do this ipc round trip. Only writing the patch will tell us!
(Assignee)

Comment 5

5 years ago
Created attachment 736043 [details] [diff] [review]
patch v1

Gregor, let me know if this helps.
Attachment #736043 - Flags: feedback?(anygregor)
Was it concluded that this change is not necessary to help fix bug 847592?
(In reply to Dietrich Ayala (:dietrich) from comment #6)
> Was it concluded that this change is not necessary to help fix bug 847592?

I don't think we'll be able to reach any conclusions until we establish that bug 847592 is fixed.  This is not high on my list of things to try, but I also think it can't hurt.

Updated

4 years ago
Keywords: perf
Whiteboard: [c= p= u= s=]
Priority: -- → P4
Whiteboard: [c= p= u= s=] → [c=progress p= u= s=]
(Reporter)

Updated

2 years ago
Flags: needinfo?(anygregor)
(Assignee)

Updated

a year ago
Attachment #736043 - Flags: feedback?(anygregor)
You need to log in before you can comment on or make changes to this bug.