✨ performance now tracks first-input-delay
#542
Conversation
GoodForOneFare
force-pushed
the
performance-listen-for-first-input-delay
branch
from
cf64073 to
cdf859b
Compare
|
Trying to figure out tests for this, but I'd appreciate comments now :) |
|
|
||
| if (typeof perfMetrics !== 'undefined') { | ||
| perfMetrics.onFirstInputDelay(delay => { | ||
| this.lifecycleEvent({ |
There was a problem hiding this comment.
Is this really a lifecycle event? It's a thing that happens once, so I think so. But it's also the only one that concerns duration instead of start. Should we expose a way for consumers to know which value is significant?
There was a problem hiding this comment.
This definitely does feel a bit weird but I think we are kind of stuck putting it here. Consumers at least have a way of figuring out whether it was some sort of duration event or a point in time by looking at duration, but I agree it's not the cleanest abstraction. IMO, this is fine for the time being.
|
|
||
| if (typeof perfMetrics !== 'undefined') { | ||
| perfMetrics.onFirstInputDelay(delay => { | ||
| this.lifecycleEvent({ |
There was a problem hiding this comment.
This definitely does feel a bit weird but I think we are kind of stuck putting it here. Consumers at least have a way of figuring out whether it was some sort of duration event or a point in time by looking at duration, but I agree it's not the cleanest abstraction. IMO, this is fine for the time being.
GoodForOneFare
force-pushed
the
performance-listen-for-first-input-delay
branch
from
702ba1f to
15cd659
Compare
| @@ -12,4 +16,47 @@ describe('Performance', () => { | |||
| performance.finish(); | |||
| expect(spy).toHaveBeenCalledWith(expect.any(Navigation)); | |||
| }); | |||
|
|
|||
| describe.skip('first-input-delay', () => { | |||
There was a problem hiding this comment.
@lemonmade tests aren't runnable right now, but I think they'd pass once a performance mock is available. Can you sanity check, please?
| performance.on('lifecycleEvent', lifecycleSpy); | ||
|
|
||
| // eslint-disable-next-line typescript/no-non-null-assertion | ||
| perfMetrics.callback!(1000, new Event('click')); |
There was a problem hiding this comment.
Oh interesting, I wonder if we should capture the event type as metadata for the fid event?
There was a problem hiding this comment.
Let's get some results in first (aka, I'd like to ship this now 😛 )
GoodForOneFare
force-pushed
the
performance-listen-for-first-input-delay
branch
from
87916e0 to
6b49b1f
Compare
Basically, this measures the milliseconds between the user's first event, and the browser passing the event on to handlers. See https://github.com/GoogleChromeLabs/first-input-delay for more information. Note: this is only useful when the first-input-delay polyfill is present in the document's head.
GoodForOneFare
force-pushed
the
performance-listen-for-first-input-delay
branch
from
6b49b1f to
88c1467
Compare
Basically, this measures the milliseconds between the user's first event, and the browser passing the event on to handlers. See https://github.com/GoogleChromeLabs/first-input-delay for more information.
Note: this is only useful when the first-input-delay polyfill is present in the document's head.