Window remains white with GPU enabled (macOS)

[magreenblatt]

The 10.8 SDK version for Java is the problem. See https://bitbucket.org/chromiumembedded/ ... t-22035822

[radar]

Thanks - at least it's clear now where the problem is.

Until I figure out away to get around the 10.8 SDK with java: Is it safe to set NSWindowHostsLayersInWindowServer?

[magreenblatt]

Thanks - at least it's clear now where the problem is.

Until I figure out away to get around the 10.8 SDK with java: Is it safe to set NSWindowHostsLayersInWindowServer?

Try it and let us know how it works for you?

[radar]

For a small standalone Java / SWT test application it looks all good.

We're in the process of integrating this into a large Eclipse/RCP-based application. I'll get back when I see how it goes there...

Thanks - at least it's clear now where the problem is.

Until I figure out away to get around the 10.8 SDK with java: Is it safe to set NSWindowHostsLayersInWindowServer?

Try it and let us know how it works for you?

[Czarek]

From code comments in Chromium bug 626733 it shows that in 10.9 SDK and later window hosts layers in window server by default, thus registering `NSWindowHostsLayersInWindowServer` is not required anymore. It was just unnecessary code that was removed since Chromium started targetting 10.9 as minimum OS.

[radar]

To finally report back:
* With main executables that are compiled with 10.9 SDK and later: Everything is fine.
* Earlier: Chromium switch "--disable-remote-core-animation" as workaround.

I suspect however, that above switch might disappear at some point.

What we'll do in our case (Eclipse Equinox launcher 3.8.2, which hasn't even macOS min SDK compiled in): recompile the executable and specify min SDK. Then everything works fine.

[Czarek]

So calling `NSWindowHostsLayersInWindowServer` which you reported earlier that fixed the issue, didn't actually fix it? Looks to me there are two different issues here.