Open
Bug 1189851
Opened 10 years ago
Updated 3 years ago
Collect concurrent cache read/revalidation telemetry
Categories
(Core :: Networking: HTTP, defect, P5)
Core
Networking: HTTP
Tracking
()
NEW
Tracking | Status | |
---|---|---|
firefox42 | --- | affected |
People
(Reporter: mayhemer, Unassigned)
Details
(Whiteboard: [necko-would-take])
Scenario:
- channel 1 fetches and caches a resumable/re-validable response
- channel 2 requests the same response while channel 1 still writes the cache
- channel 2 can do a concurrent cache read
- but, channel 2 cannot do it from reasons described at [1]
- so, channel 2 abandons concurrent cache read and waits for channel 1 to deliver a complete content to the cache first
This is suboptimal. One solution is to ignore conditional headers and let the second request simply fully from the cache w/o revalidation.
I don't know if this is a spec violation or not. Also I want to see how often it happens we get a 200 on a conditional request so close to a previous request on the same URL. If sufficiently low, I think we could do this optimization and allow concurrent read in this IMO pretty often happening scenario.
Other solution is to allow two requests during nsHttpChannel lifetime. That's not a really small change, but would allow us to do this optimization w/o any spec violations.
http://hg.mozilla.org/mozilla-central/annotate/e9389ca320ff/netwerk/protocol/http/nsHttpChannel.cpp#l3291
Updated•9 years ago
|
Whiteboard: [necko-would-take]
![]() |
Reporter | |
Updated•9 years ago
|
Assignee: honzab.moz → nobody
Status: ASSIGNED → NEW
Comment 1•8 years ago
|
||
Bulk change to priority: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258
Priority: -- → P5
Updated•3 years ago
|
Severity: normal → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•