This is entirely dead code now, right?

(In reply to Matt Woodrow (:mattwoodrow) from comment #0) > Created attachment 8397702 [details] [diff] [review] > remove-deprecated-canvas > > This is entirely dead code now, right? New textures aren't the default on windows. There isn't a lot of work left to get there but I keep going back to b2g stuff.

Why don't we just make them the default? We're not shipping it, and we won't ever ship it with old-textures, why keep them around? At the very least we should take this and make <canvas> use new-textures because this code is blocking some clean up work in BasicLayers.

Last time I proposed to do that, it wasn't possible because we wanted the metro guys to be able to have green try pushes with OMTC on. Now it's a different story and I am all for making them the default. There isn't a lot of remaining tests to fix, see this try push which has some of my local fixes: https://tbpl.mozilla.org/?tree=Try&rev=f94d7b64f756 There is mostly one clipping issue and some fuzzing to add to one or two reftests. It wouldn't take a lot of time but realistically, b2g priorities keep delaying this. Honestly, even with a green try push, OMTC on windows won't be shippable as long as we don't have a proper implementation for canvas layers (unless we want to make all of our cool game demos 6 times slowere). Milan, what do you think? Is it okay to turn on new textures now and fix the remaining reftests later? If we do that we can say adios to almost all of the deprecated classes (except IncrementalContentClient/Host on mac).

Works for me. Also, you two can continue the conversation live in Taipei next week, so the timing is decent for it, only week 3 of the train.