Consider ManagedSourceBuffer in MSE support detection

[niklaskorz]

This PR will...

Consider ManagedSourceBuffer as well when detecting MSE support.

Why is this Pull Request needed?

On iOS Safari, isMSESupported(), and thus isSupported(), will currently always return false as the prototype checks on SourceBuffer fail, because self.SourceBuffer (and self.WebKitSourceBuffer) does not exist on iOS Safari.

Are there any points in the code the reviewer needs to double check?

The change is rather simple, so no.

Resolves issues:

No issue has been reported in this regard, but #6161 might be related.

Checklist

• changes have been done against master branch, and PR does not conflict
• new unit / functional tests have been added (whenever applicable) => there were no tests for isSupported() to begin with
• API or design changes are documented in API.md => no API changes (getSourceBuffer is not exported)

[robwalch]

Hi @niklaskorz,

On iOS Safari, isMSESupported(), and thus isSupported(), will currently always return false

In which versions of iOS are you able to reproduce this issue?

I've been running HLS.js on iOS 17.1-18+ without this issue.

the prototype checks on SourceBuffer fail, because self.SourceBuffer (and self.WebKitSourceBuffer) does not exist on iOS Safari.

While correct, there are other platforms that do not expose SourceBuffer globally, therefor isMSESupported() and isSupported() still pass when getSourceBuffer() returns undefined:

hls.js/src/is-supported.ts

Lines 20 to 24 in 3e77cb5

	// if SourceBuffer is exposed ensure its API is valid
	// Older browsers do not expose SourceBuffer globally so checking SourceBuffer.prototype is impossible
	const sourceBuffer = getSourceBuffer();
	return (
	!sourceBuffer ||

[niklaskorz]

@robwalch I now realize I indeed misunderstood the purpose of the SourceBuffer check. It makes sense that needing to check for these functions on Managed MSE is not required in the first place, as the only (current) implementation of Managed MSE (WebKit) is feature-complete.

I'll double check on Monday how I came upon this in the first place, sorry for the noise!

[robwalch]

@niklaskorz no worries. The changes make sense. I had to debug by connecting an iPhone to understand why it was working.

I think we can take this change. I'll want to run them on the device too before giving a formal review.

[robwalch]

As for #6161, the simulator is another story. I don't use iOS simulators for media testing Problems with playback in simulators should be reported to Apple via Feedback Assistant. Maybe there's a change we should make (in addition to this one) that would help developers know that they should test on a real device (or a real bug that needs to be fixes for hls.js? - let me know if you identify any).

[robwalch]

The only place getSourceBuffer is called is in this module and it never has the preferManagedSourceBuffer argument passed in. I understand this is to align with getMediaSource, but the argument is not used and thus unnecessary.

[robwalch]

Could this be simplified to the following?

Suggested change

	const msb =
	(preferManagedSourceBuffer || !self.SourceBuffer) &&
	((self as any).ManagedSourceBuffer as undefined | typeof self.SourceBuffer);
	return msb || self.SourceBuffer || (self as any).WebKitSourceBuffer;
	return self.SourceBuffer || (self as any).ManagedSourceBuffer || (self as any).WebKitSourceBuffer;