We want to eventually remove support for Direct2D 1.0, let's disable using it by default in this train.

Comment on attachment 8665975 [details] [diff] [review] Disable using Direct2D 1.0 by default Review of attachment 8665975 [details] [diff] [review]: ----------------------------------------------------------------- I like

This adjusts a bunch of test framework expectations.

Comment on attachment 8666761 [details] [diff] [review] Part 2: Disable using Direct2D 1.0 by default v2 Review of attachment 8666761 [details] [diff] [review]: ----------------------------------------------------------------- ::: testing/web-platform/meta/2dcontext/fill-and-stroke-styles/2d.gradient.radial.outside3.html.ini @@ +4,5 @@ > expected: > if not debug and (os == "win") and (version == "5.1.2600") and (processor == "x86") and (bits == 32): FAIL > if debug and (os == "win") and (version == "5.1.2600") and (processor == "x86") and (bits == 32): FAIL > + if not debug and (os == "win") and (version == "6.1.7601") and (processor == "x86") and (bits == 32): FAIL > + if debug and (os == "win") and (version == "6.1.7601") and (processor == "x86") and (bits == 32): FAIL Bad indenting ::: testing/web-platform/meta/2dcontext/path-objects/2d.path.arcTo.shape.curve1.html.ini @@ +3,5 @@ > [arcTo() curves in the right kind of shape] > expected: > if not debug and (os == "win") and (version == "5.1.2600") and (processor == "x86") and (bits == 32): FAIL > if debug and (os == "win") and (version == "5.1.2600") and (processor == "x86") and (bits == 32): FAIL > + if (os == "win") and (version == "6.1.7601") and (processor == "x86") and (bits == 32): FAIL same ::: testing/web-platform/meta/2dcontext/path-objects/2d.path.arcTo.shape.curve2.html.ini @@ +3,5 @@ > [arcTo() curves in the right kind of shape] > expected: > if not debug and (os == "win") and (version == "5.1.2600") and (processor == "x86") and (bits == 32): FAIL > if debug and (os == "win") and (version == "5.1.2600") and (processor == "x86") and (bits == 32): FAIL > + if (os == "win") and (version == "6.1.7601") and (processor == "x86") and (bits == 32): FAIL Indentation looks wrong here.

https://hg.mozilla.org/integration/mozilla-inbound/rev/2ef0bdb65dbf https://hg.mozilla.org/integration/mozilla-inbound/rev/c3ad832a28ec

https://hg.mozilla.org/mozilla-central/rev/2ef0bdb65dbf https://hg.mozilla.org/mozilla-central/rev/c3ad832a28ec

a huge list of improvements/regressions on talos: http://alertmanager.allizom.org:8080/alerts.html?rev=c3ad832a28ecfbf14f1e7c3d2b3e9f90c015b35a&table=1 Bas has confirmed this is expected, can we figure out if the test machines need to be updated at all?

If it's possible to update Windows 7 machines to SP1, we'd be able to start using D2D1.1 there as well, rather than just on Windows 8.

Release Note Request (optional, but appreciated) [Why is this notable]: Font rendering can change and people don't like the result: https://www.reddit.com/r/firefox/comments/4312zk/firefox_44_font_feels_very_anemic/ https://www.reddit.com/r/firefox/comments/4328ur/fonts_too_thin_with_the_new_update/ [Suggested wording]: Support for Direct2D 1.0 has been dropped. This can change the font rendering on systems that don't support Direct2D 1.1 (Windows Vista and Windows 7 without the Platform Update). Windows 7 users can install the Platform Update [https://www.microsoft.com/en-ca/download/details.aspx?id=36805] to re-enable Direct2D support.

Done. Added to Fx44 release notes.