Bug 168057
| Summary: | [Navigation Timing] navigationStart and DOM timings in Unix epoch, other timings 0-based for Back-Forward navigations | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Nic Jansma <nic> |
| Component: | Platform | Assignee: | Joseph Pecoraro <joepeck> |
| Status: | NEW | ||
| Severity: | Normal | CC: | cdumez, joepeck, simon.fraser |
| Priority: | P2 | ||
| Version: | Safari 10 | ||
| Hardware: | iPhone / iPad | ||
| OS: | iOS 10 | ||
Nic Jansma
Hi,
We're seeing many cases where on Back-Forward navigations, the NavigationTiming timestamps are incorrect.
It appears that navigationStart is Unix-epoch based, along with the dom* events. However, all other timestamps are 0-based.
It also appears navigationStart happens after the dom* events.
Example 1:
Mozilla/5.0 (iPhone; CPU iPhone OS 10_2 like Mac OS X) AppleWebKit/602.3.12 (KHTML, like Gecko) Version/10.0 Mobile/14C92 Safari/602.1
navigation.type 2 // back-forward
redirectStart 0
redirectEnd 0
navigationStart 1486655091185
fetchStart 0
domainLookupStart 2
domainLookupEnd 27
connectStart 27
connectEnd 50
requestStart 52
responseStart 434
responseEnd 0
domLoading 1486655020306
domInteractive 1486655044330
domContentLoadedEventStart 1486655044330
domContentLoadedEventEnd 1486655044339
domComplete 1486655088205
loadEventStart 0
Example 2:
Mozilla/5.0 (iPhone; CPU iPhone OS 10_2_1 like Mac OS X) AppleWebKit/602.4.6 (KHTML, like Gecko) Version/10.0 Mobile/14D27 Safari/602.1
navigation.type 2 // back-forward
redirectStart 0
redirectEnd 0
navigationStart 1486648661894
fetchStart 0
domainLookupStart 0
domainLookupEnd 0
connectStart 0
connectEnd 0
requestStart 3
responseStart 259
responseEnd 0
domLoading 1486648620150
domInteractive 1486648620956
domContentLoadedEventStart 1486648621014
domContentLoadedEventEnd 1486648621070
domComplete 1486648652486
loadEventStart 0
This doesn't happen all of the time, but in one data set, we're seeing it in 17 out of 254,357 beacons. It doesn't appear to be all Back-Forward Safari beacons, as there are 27,593 other Back-Foward beacons in that dataset whose timestamps look OK.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Joseph Pecoraro
I can look at this.