This splits waitingDuration (formerly known as waitingTime in v3) into two:
waitingDuration - which the time until (workerStart or fetchStart). This will mostly be redirect time but I've avoided explicitly calling that as redirectStart/End can't be used and it's easy to see when testing instances where looking at workerStart or fetchStart still leaves a non-0 amount of time that is not redirects.
cacheDuration which is the time from then until domainLookupStart. This will mostly be HTTP cache time, except for service-worker controlled sites when it will be service worker time.
See PerformanceNavigationTiming breakdown for more context:
As part of this I've also moved any delays between connectEnd and requestStart from being attributed to connectionDuration (i.e before the delay) to requestDuration (i.e after the delay). This helps avoiding attributing this delay to connectionDuration for service-worker controlled sites when this should be 0, but without adding an extra breakdown that would not be that useful.
Fixes #395
This splits
waitingDuration(formerly known aswaitingTimein v3) into two:waitingDuration- which the time until (workerStartorfetchStart). This will mostly be redirect time but I've avoided explicitly calling that asredirectStart/Endcan't be used and it's easy to see when testing instances where looking atworkerStartorfetchStartstill leaves a non-0 amount of time that is not redirects.cacheDurationwhich is the time from then untildomainLookupStart. This will mostly be HTTP cache time, except for service-worker controlled sites when it will be service worker time.See PerformanceNavigationTiming breakdown for more context:
As part of this I've also moved any delays between
connectEndandrequestStartfrom being attributed toconnectionDuration(i.e before the delay) torequestDuration(i.e after the delay). This helps avoiding attributing this delay toconnectionDurationfor service-worker controlled sites when this should be 0, but without adding an extra breakdown that would not be that useful.