Open rniwa opened 10 months ago
The consensus here seems to be investigating adding HTTP header for caching so that browsers would consistently cache or not cache.
I'd say this is out of v3 at this point. Perhaps v3.1?
Yeah, sounds good to look into this in a later version. v3.0 is not worse than v2.1 in that respect, so technically not a regression.
See https://webkit.slack.com/archives/C05QJJZQZRQ/p1698264053662939.
Apparently the network speed difference can affect Speedometer 2.0 score. We should make sure to either verify that this doesn't happen on Speedometer 3.0 or mitigate it some way e.g. prime network cache prior to start measurements.