Closed
Bug 1351521
Opened 8 years ago
Closed 7 years ago
performance.getEntries() doesn't list entries for resources intercepted by a service worker in e10s mode
Categories
(Core :: DOM: Service Workers, defect, P3)
Core
DOM: Service Workers
Tracking
()
RESOLVED
FIXED
mozilla58
Tracking | Status | |
---|---|---|
firefox58 | --- | fixed |
People
(Reporter: catalinb, Assigned: bkelly)
References
(Blocks 2 open bugs)
Details
Attachments
(2 files, 1 obsolete file)
(deleted),
patch
|
asuth
:
review+
|
Details | Diff | Splinter Review |
(deleted),
patch
|
asuth
:
review+
|
Details | Diff | Splinter Review |
register a service worker with onfetch = function(event) { event.respondWith(fetch(event.request)); } and then load a page with a bunch of resources.
performance.getEntriesByType("resource"); should include PerformanceResourceTiming for these resources, but in e10s mode it doesn't.
Assignee | ||
Updated•8 years ago
|
Blocks: dt-service-worker
Assignee | ||
Updated•8 years ago
|
Blocks: ServiceWorkers-compat
Reporter | ||
Comment 1•8 years ago
|
||
The problem is that InterceptStreamListener::OnStopRequest skips the step where we add the performance entry in nsHttpChannelChild::OnStopRequest. However, I think it's not worth fixing this at this time and we should just wait for :asuth to land parent interception.
Depends on: 1231222
Updated•8 years ago
|
Priority: -- → P3
Assignee | ||
Comment 3•7 years ago
|
||
Andrew, given where you are on parent interception, do you still think we should wait on this? Some partner sites would like to use it.
Flags: needinfo?(bugmail)
Comment 4•7 years ago
|
||
Especially given how we may have to hold off on landing various things because of 57 stability concerns, I see implementing this separately making sense.
Flags: needinfo?(bugmail)
Assignee | ||
Comment 5•7 years ago
|
||
I'm going to take this so I can add a test. I'm worried that I'm regressing performance timeline stuff in bug 1391693.
Assignee | ||
Comment 6•7 years ago
|
||
I did most of the hard bits in the dependent bugs, so this is a small fix now.
Assignee | ||
Comment 7•7 years ago
|
||
Assignee | ||
Comment 8•7 years ago
|
||
Andrew, this patch does a few things:
1. Synthesize a responseStart time in the intercepted OnStartRequest()
2. Synthesize a responseEnd time in the intercepted OnStopRequest()
3. Move the Performance::AddEntry() call to DoPreOnStopRequest() which we already call when synthesizing.
Attachment #8915344 -
Attachment is obsolete: true
Attachment #8915705 -
Flags: review?(bugmail)
Assignee | ||
Updated•7 years ago
|
Attachment #8915345 -
Flags: review?(bugmail)
Updated•7 years ago
|
Attachment #8915705 -
Flags: review?(bugmail) → review+
Updated•7 years ago
|
Attachment #8915345 -
Flags: review?(bugmail) → review+
Pushed by bkelly@mozilla.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/047092e0d447
P1 Add intercepted HttpChannelChild channels to the performance timeline. r=asuth
https://hg.mozilla.org/integration/mozilla-inbound/rev/0d8fa7a51ad0
P2 Update resource-timing.https.html WPT test expectations. r=asuth
Comment 10•7 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/047092e0d447
https://hg.mozilla.org/mozilla-central/rev/0d8fa7a51ad0
Status: ASSIGNED → RESOLVED
Closed: 7 years ago
status-firefox58:
--- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla58
You need to log in
before you can comment on or make changes to this bug.
Description
•