Updates for Safari 26.4 beta

[Elchi3]

The https://collector.openwebdocs.org project (v10.17.3) found new features shipping in Safari 26.4 beta (21624.1.11.11.3) on macOS 26.4 beta (25E5207k) which was released last night. Currently, the collector covers about 90% of BCD, so the following list might not be exhaustive. Also, if a feature is in Safari TP only, it is not considered here.

This PR contains 88 feature updates or additions and they can be verified with the release notes or with the comments in this PR. Verified features are marked with (*).

• api.Element.currentCSSZoom (*)
• api.FontFaceSet.FontFaceSet (*)
• api.PerformanceResourceTiming.deliveryType (*)
• api.PerformanceResourceTiming.finalResponseHeadersStart (*)
• api.PerformanceResourceTiming.firstInterimResponseStart (*)
• api.ReadableByteStreamController (*)
• api.ReadableByteStreamController.byobRequest (*)
• api.ReadableByteStreamController.close (*)
• api.ReadableByteStreamController.desiredSize (*)
• api.ReadableByteStreamController.enqueue (*)
• api.ReadableByteStreamController.error (*)
• api.ReadableStream.@@asyncIterator (*)
• api.ReadableStreamBYOBReader (*)
• api.ReadableStreamBYOBReader.ReadableStreamBYOBReader (*)
• api.ReadableStreamBYOBReader.cancel (*)
• api.ReadableStreamBYOBReader.closed (*)
• api.ReadableStreamBYOBReader.read (*)
• api.ReadableStreamBYOBReader.releaseLock (*)
• api.ReadableStreamBYOBRequest (*)
• api.ReadableStreamBYOBRequest.respond (*)
• api.ReadableStreamBYOBRequest.respondWithNewView (*)
• api.ReadableStreamBYOBRequest.view (*)
• api.Response.body.readable_byte_stream (*)
• api.WebTransport (*)
• api.WebTransport.WebTransport (*)
• api.WebTransport.WebTransport.options_allowPooling_parameter (*)
• api.WebTransport.WebTransport.options_congestionControl_parameter (*)
• api.WebTransport.WebTransport.options_requireUnreliable_parameter (*)
• api.WebTransport.WebTransport.options_serverCertificateHashes_parameter (*)
• api.WebTransport.anticipatedConcurrentIncomingBidirectionalStreams (*)
• api.WebTransport.anticipatedConcurrentIncomingUnidirectionalStreams (*)
• api.WebTransport.byob_readers (*)
• api.WebTransport.close (*)
• api.WebTransport.closed (*)
• api.WebTransport.congestionControl (*)
• api.WebTransport.createBidirectionalStream (*)
• api.WebTransport.createBidirectionalStream.options_sendOrder_parameter (*)
• api.WebTransport.createSendGroup (*)
• api.WebTransport.createUnidirectionalStream (*)
• api.WebTransport.createUnidirectionalStream.byob_readers (*)
• api.WebTransport.createUnidirectionalStream.options_sendOrder_parameter (*)
• api.WebTransport.datagrams (*)
• api.WebTransport.draining (*)
• api.WebTransport.getStats (*)
• api.WebTransport.incomingBidirectionalStreams (*)
• api.WebTransport.incomingUnidirectionalStreams (*)
• api.WebTransport.protocol (*)
• api.WebTransport.ready (*)
• api.WebTransport.reliability (*)
• api.WebTransport.supportsReliableOnly_static (*)
• api.WebTransportBidirectionalStream (*)
• api.WebTransportBidirectionalStream.readable (*)
• api.WebTransportBidirectionalStream.readable.returns_WebTransportReceiveStream (*)
• api.WebTransportBidirectionalStream.writable (*)
• api.WebTransportBidirectionalStream.writable.returns_WebTransportSendStream (*)
• api.WebTransportDatagramDuplexStream (*)
• api.WebTransportDatagramDuplexStream.byob_readers (*)
• api.WebTransportDatagramDuplexStream.createWritable (*)
• api.WebTransportDatagramDuplexStream.incomingHighWaterMark (*)
• api.WebTransportDatagramDuplexStream.incomingMaxAge (*)
• api.WebTransportDatagramDuplexStream.maxDatagramSize (*)
• api.WebTransportDatagramDuplexStream.outgoingHighWaterMark (*)
• api.WebTransportDatagramDuplexStream.outgoingMaxAge (*)
• api.WebTransportDatagramDuplexStream.readable (*)
• api.WebTransportDatagramsWritable (*)
• api.WebTransportDatagramsWritable.sendGroup (*)
• api.WebTransportDatagramsWritable.sendOrder (*)
• api.WebTransportError (*)
• api.WebTransportError.WebTransportError (*)
• api.WebTransportError.source (*)
• api.WebTransportError.streamErrorCode (*)
• api.WebTransportReceiveStream (*)
• api.WebTransportReceiveStream.getStats (*)
• api.WebTransportSendGroup (*)
• api.WebTransportSendGroup.getStats (*)
• api.WebTransportSendStream (*)
• api.WebTransportSendStream.getStats (*)
• api.WebTransportSendStream.getWriter (*)
• api.WebTransportSendStream.sendGroup (*)
• api.WebTransportSendStream.sendOrder (*)
• css.properties.display.grid-lanes (*)
• css.properties.display.inline-grid-lanes (*)
• css.properties.flow-tolerance (*)
• css.properties.flow-tolerance.infinite (*)
• css.properties.flow-tolerance.normal (*)
• css.properties.font-size.math (*)
• css.properties.math-depth (*)
• css.properties.math-depth.auto-add (*)
• javascript.builtins.Iterator.concat (*)

[github-actions]

Tip: Review these changes grouped by change (recommended for most PRs), or grouped by feature (for large PRs).

[caugner]

api.HTMLSelectedContentElement (unsure about this one, could be early exposure?)
html.elements.selectedcontent (unsure about this one, could be early exposure?)

The interface is behind two prefs:

• HTMLEnhancedSelectParsingEnabled - stable since WebKit/WebKit@d51bd85 (624.1.2 before Safari 26.4 beta)
• HTMLEnhancedSelectEnabled - stable since WebKit/WebKit@38528ac (625.1.7 after Safari 26.4 beta) - previously named HTMLEnhancedSelectSelectedContentEnabled, but only testable then

So this looks indeed like accidental exposure, or intentional temporary exposure (like Firefox's early beta).

[caugner]

• api.PerformanceResourceTiming.deliveryType (maybe shipping)

• api.PerformanceResourceTiming.finalResponseHeadersStart (maybe shipping)

• api.PerformanceResourceTiming.firstInterimResponseStart (maybe shipping)

Landed in WebKit 624.1.4 before Safari 26.4 beta, and the commit does not indicate any pref, so I would assume this shipped intentionally?

[caugner]

PS: For some reason, the WebKit repo does not seem to have a safari-*624*-branch branch yet.

[zloirock]

It also seems Iterator.concat is shipped.

[caugner]

It also seems Iterator.concat is shipped.

Does https://collector.openwebdocs.org/tests/javascript/builtins/Iterator/concat pass for you?

[zloirock]

@caugner I see it in core-js compat tests and in console.

[zloirock]

@caugner it's passed too.

[caugner]

@Elchi3 Any idea why your collector test missed this? Or did you accidentally/intentionally not include it in the PR?

[Elchi3]

I intentionally left this out because it is not in the release notes and I haven't confirmed if it is shipping in earlier versions. I just added a collector test for Iterator.concat last week. Will add it here or in a separate PR it is shipping in earlier versions.

[jdatapple]

api.PerformanceResourceTiming.deliveryType

This property exists but as an empty string. Maybe the test should be updated to look for a non-empty string? As it exists in Safari, it doesn't do anything useful.

api.PerformanceResourceTiming.finalResponseHeadersStart
api.PerformanceResourceTiming.firstInterimResponseStart

These are supported and will be included in an update to the release notes.

html.elements.selectedcontent

This should not be accessible in Safari 26.4 beta.

Iterator.concat

This shipped in Safari 18.4. See Safari 18.4 Release Notes

Keyboard Lock API

This is enabled and should be available on macOS, iPadOS, and iOS.

[zloirock]

This shipped in Safari 18.4

Maybe it was implemented in 18.4, but definitely not shipped. It was missed in 26.3 and was shipped in Bun only 3 weeks ago.

[jdatapple]

This shipped in Safari 18.4

Maybe it was implemented in 18.4, but definitely not shipped. It was missed in 26.3 and was shipped in Bun only 3 weeks ago.

This appears to be correct! It was implemented, but did not get enabled until this beta (26.4 beta).

[Elchi3]

api.PerformanceResourceTiming.deliveryType

This property exists but as an empty string. Maybe the test should be updated to look for a non-empty string? As it exists in Safari, it doesn't do anything useful.

Thanks, the empty string behavior is per spec so Safari is not wrong with this initial implementation.

html.elements.selectedcontent
This should not be accessible in Safari 26.4 beta.

Removed from this PR.

This shipped in Safari 18.4

Maybe it was implemented in 18.4, but definitely not shipped. It was missed in 26.3 and was shipped in Bun only 3 weeks ago.

This appears to be correct! It was implemented, but did not get enabled until this beta (26.4 beta).

Updated to set as shipping in 26.4

Keyboard Lock API
This is enabled and should be available on macOS, iPadOS, and iOS.

Thanks, but I still can't confirm this. See also https://bugs.webkit.org/show_bug.cgi?id=306525 and the comment on https://bugs.webkit.org/show_bug.cgi?id=299616.

[caugner]

Thanks, but I still can't confirm this. See also bugs.webkit.org/show_bug.cgi?id=306525 and the comment on bugs.webkit.org/show_bug.cgi?id=299616.

The now-stable FullScreenKeyboardLock pref depends on the FULLSCREEN_API build option (cf. https://bugs.webkit.org/show_bug.cgi?id=265923), which is disabled by default.