Chrome 8 added speak CSS property

[queengooborg]

This PR updates and corrects version values for Chromium (Chrome, Opera, Samsung Internet, WebView Android) for the speak CSS property. The data comes from the mdn-bcd-collector project (v10.13.3).

Check out the collector's guide on how to review this PR.

Tests Used: https://mdn-bcd-collector.gooborg.com/tests/css/properties/speak

[github-actions]

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

[caugner]

While we're here: Safari added this property between 4.0.5 and 5.1.5 and removed it in 12.1 11.1 (possibly renamed to speak-as).

Originally implemented via WebKit/WebKit@3b7c2c2 in WebKit 534.10 before Safari 5.1 (534.48).

Renamed via WebKit/WebKit@1f2d229 in WebKit 605.1.19 before Safari 11.1 (605.1.33).

This last change looks like it removed support for speak, which supported these values: normal, none, spell-out, digits, literal-punctuation, no-punctuation.

This explains why only the collector test for the speak property passes in Safari 10.1 and Safari 11.1, but the tests for the values don't.

So here's what I would suggest:

1. Add a partial_implementation entry for Safari added 5.1 removed 11.1, pointing to the speak-as property.
2. Add an entry to the css.properties.speak_as with alternative_name: "speak".

[queengooborg]

Looking at the bug filed for Chrome about speak, it sounds like it implemented speak as an alternative name for speak-as, just like what Safari used to do. Maybe what we should do is remove the data for the speak property since WebKit only implemented it as an alias to speak-as?

[caugner]

I'm not sure I fully understand. Are you suggesting to merge speak data into speak-as data?

[queengooborg]

That's correct -- it's a strange scenario since the speak property is its own property in the spec, but since no browser ever implemented speak as anything but an alias for speak-as...!

[caugner]

After looking at it again, I agree insofar as we should move the Chrome/Safari data over with alternative_name, but maybe let's keep this feature for now, given that there is an impl_url for Firefox. For Chrome, we can use https://issues.chromium.org/issues/40813740 as the impl_url.