Simplify MetaPlaylist Manifest loading by relying on the same code than other transports

[peaBerberian]

I was PoCing protocol-detection when initially fetching the Manifest, so application developers wouldn't e.g. have to tell to us that they want to play "dash".

While doing that, I noticed that all transports had close to the same logic for Manifest fetching. Only differences I've seen are:

• DASH may add integrity checks if the checkManifestIntegrity option is set

• DASH may ask the browser to obtain the response as an ArrayBuffer (and not as a default JS string) if it is probable that we will rely on the WASM MPD parser

• the local transport throws if no custom manifestLoader is declared.

As such, those differences look minor and straightforward enough so the corresponding logic can be factorized into e.g. a single parameterized createManifestLoader function.

This was already done for DASH and Smooth, but I don't know why we didn't use it for the experimental MetaPlaylist feature. This commit fixes this.

We could also use it for the experimental local transport in the future, though I'm unsure of how we could generically make sense of its peculiar rule (no real loading).

[github-actions]

✅ Automated performance checks have passed on commit d79d32d8fb20590eaf727c1aa599a20e026c66cb with the base branch dev.

Details

Performance tests 1st run output

No significative change in performance for tests:

Name	Mean	Median
loading	20.88ms -> 21.02ms (-0.146ms, z: 0.37250)	29.40ms -> 29.25ms
seeking	20.09ms -> 21.39ms (-1.299ms, z: 0.47761)	11.25ms -> 11.25ms
audio-track-reload	27.65ms -> 27.47ms (0.174ms, z: 1.81145)	40.35ms -> 40.20ms
cold loading multithread	49.46ms -> 48.02ms (1.438ms, z: 12.04496)	72.00ms -> 70.95ms
seeking multithread	91.10ms -> 81.67ms (9.436ms, z: 2.21622)	9.95ms -> 9.90ms
audio-track-reload multithread	27.12ms -> 27.03ms (0.087ms, z: 0.65589)	39.90ms -> 39.90ms
hot loading multithread	16.81ms -> 16.67ms (0.149ms, z: 3.19044)	24.45ms -> 24.30ms

[github-actions]

✅ Automated performance checks have passed on commit c2a6715ce71ae549d30d5c6a3686ec6daf1af37c with the base branch dev.

Details

Performance tests 1st run output

No significative change in performance for tests:

Name	Mean	Median
loading	20.30ms -> 21.18ms (-0.883ms, z: 1.70016)	28.35ms -> 28.35ms
seeking	20.59ms -> 21.33ms (-0.734ms, z: 0.34640)	11.10ms -> 11.10ms
audio-track-reload	26.91ms -> 27.07ms (-0.161ms, z: 1.09428)	39.30ms -> 39.45ms
cold loading multithread	48.09ms -> 46.79ms (1.300ms, z: 13.26542)	69.75ms -> 68.55ms
seeking multithread	23.80ms -> 29.82ms (-6.016ms, z: 0.31048)	10.20ms -> 10.20ms
audio-track-reload multithread	25.99ms -> 25.83ms (0.168ms, z: 1.77648)	38.10ms -> 37.95ms
hot loading multithread	15.52ms -> 15.36ms (0.160ms, z: 4.48271)	22.50ms -> 22.20ms

[github-actions]

✅ Automated performance checks have passed on commit 63565fc62055d795301440032ee18e5c9837ab1c with the base branch dev.

Details

Performance tests 1st run output

No significative change in performance for tests:

Name	Mean	Median
loading	20.74ms -> 20.75ms (-0.016ms, z: 0.74139)	29.25ms -> 29.15ms
seeking	11.27ms -> 13.31ms (-2.046ms, z: 0.34811)	12.15ms -> 12.00ms
audio-track-reload	28.24ms -> 28.39ms (-0.151ms, z: 1.39787)	41.25ms -> 41.40ms
cold loading multithread	48.52ms -> 47.18ms (1.340ms, z: 11.67807)	70.50ms -> 69.30ms
seeking multithread	45.39ms -> 52.08ms (-6.689ms, z: 0.54190)	11.10ms -> 10.95ms
audio-track-reload multithread	27.49ms -> 27.25ms (0.234ms, z: 2.86490)	40.50ms -> 40.20ms
hot loading multithread	15.56ms -> 15.47ms (0.095ms, z: 4.33632)	22.80ms -> 22.65ms