We use gpu for yuv color conversion now. There are some precision problems in gpu path. Mark them fails-if and update the fuzzy-test value.

MozReview-Commit-ID: 6tunOCj3OHB

Bug 1366502 - Update the thread model for RegisterExternalImage(), UnregisterExternalImage() and GetRenderTexture() call. v2. r=sotaro

If we call UnregisterExternalImage() at non-render-thread and decrease the RenderTextureHost's ref-count to zero, the RenderTextureHost will be released in non-render-thread.
That will cause some problems if we use some thread-specific functions in ~RenderTextureHost().
This patch uses a message loop in UnregisterExternalImage() to resolve this problem.

MozReview-Commit-ID: CDazxGkE1cK

The RenderTextureHost might calls some thread-specific functions(e.g. OpenGL calls) in ~RenderTextureHost(). Add a checking here to prevent this problem.

MozReview-Commit-ID: 62QlavmXOig

MozReview-Commit-ID: 8ODLYyXzHjO

Update for DXGITextureHostD3D11 and DXGIYCbCrTextureHostD3D11.

MozReview-Commit-ID: CQqzHK3jqz7

MozReview-Commit-ID: 5cu8cYoTMxT

WR supports the planar-ycbcr image format. We turn to use the planar-ycbcr image to get rid of the software-ycbcr-to-rgb color format conversion(using libyuv) in gecko.

The BufferTextureHost will use 3 image keys for SurfaceFormat::YUV format.
The RenderBufferTextureHost will also use 3 DataSourceSurfaces to represent the 3 channel data in planar-ycbcr format.

MozReview-Commit-ID: 3mMreSzKnMv

These 2 functions are used for WR.
The GetWRImageKeys() will return the proper image keys according to the textureHost format.
The PushExternalImage() will put all necessary WR commands into DisplayListBuilder for the textureHost rendering.

MozReview-Commit-ID: FwCDxwsCMMy

Bug 1366502 - Move the various of image key and WR command settings from WebRenderCompositableHolder into textureHosts. r=sotaro

It's awful to put all combination of image key and WR command settings in the same place. Make the settings go back to textureHosts.
Each textureHost should implement GetWRImageKeys() and PushExternalImage() function.

MozReview-Commit-ID: 5Bmiuv5HgmX

With the video pipeline, we could call the WR image functions in compositor thread.

MozReview-Commit-ID: LtaihkLnbL3

MozReview-Commit-ID: LcSHVkBbM6G

Bug 1369689 - [1.0] Make page load time logs depend on web progress events instead of throbber activity. r=snorp

MozReview-Commit-ID: HZRKSkpZrWi

UPGRADE_NSPR_RELEASE, DONTBUILD

--HG--
extra : amend_source : 05aeb0bcaaaaab1d793158910126a4d6063b8e22

Bug 1364346 part 3 - Optimize Array.prototype.unshift by taking advantage of shifted elements. r=anba

--HG--
extra : rebase_source : 793e3dbf21745b1f13661f4856c582a86987a493

MozReview-Commit-ID: 1ifOuh90QEK

--HG--
extra : rebase_source : 4609aefd9e92c9f438bb9e955ea6dadc2750fc19

MozReview-Commit-ID: BDyuSdNALvH

Bug 1370155 - Remove fullscreen mode in basic compositor video test to prevent intermittent problem. r=jmaher

MozReview-Commit-ID: IO5vS3uzj6g

--HG--
extra : rebase_source : 8ca77acd02c6cdb9f5cce9e43dcd5428a3ca2b07

servo: Merge #17179 - shrink Rule and store all heap-allocated selector data inline (from bholley:one_selector_allocation); r=emilio

https://bugzilla.mozilla.org/show_bug.cgi?id=1370107

Source-Repo: https://github.com/servo/servo
Source-Revision: 74ea8ce3ed6aa3d7edfe05924f196ccbe57daed6

--HG--
extra : subtree_source : https%3A//hg.mozilla.org/projects/converted-servo-linear
extra : subtree_revision : a8c3994ae54b1775686552f549605a7fc2ba5646

ProfilerStackFrameRAII and ProfilerStackFrameDynamicRAII are very similar; the
latter lets a dynamic string be specified as well (and lacks the
MOZ_GUARD_OBJECT stuff, for no good reason).

This patch does the following.

- Removes ProfilerStackFrameDynamicRAII, and adds a dynamic string to
  ProfilerStackFrameRAII. It also reorders the constructor's arguments to match
  the field ordering of ProfileEntry. There aren't many usage sites so these
  changes don't affect many places.

- With that done, there is only a single callsite for each of
  profiler_call_enter() and profiler_call_exit(), so the patch also inlines and
  removes them.

These annotations aren't doing anything useful. The important thing with
the PseudoStack is that, during pushes, the stack pointer incrementing happens
after the new entry is written, and this is ensured by the stack pointer being
Atomic.

The patch also improves the comments on PseudoStack.

--HG--
extra : rebase_source : 100f8a5e4b750c15fac66175550c4c284a141f16

MozReview-Commit-ID: EiG9icZCZ3x

MozReview-Commit-ID: Cb3YlCdVhQr

--HG--
extra : rebase_source : 3ed4d5a41898ead386f8a75b0d2f7be1899cf767

Bug 1353356 - Add rematerialized frames to the table on JitActivation after rematerialization succeeds. (r=jandem)

There are a bunch of invalid Cargo.toml files that some python stuff uses for
testing, but these cause errors out of cargo when we point to one of the SM
source tarballs as a git dependency. Easiest workaround is to not include any
non-SpiderMonkey Cargo.toml files in the source tarballs.

MozReview-Commit-ID: 1wLiG2Sq3g

Bug 1367487: Change bundle name (CFBundleName) of content process on OSX/macOS to differentiate it from the Firefox bundle name. r=mstange,ted

Bug 1370025 P4 Use FiringId validity in TimeoutManager::Timeouts::Insert() and ::ResetTimersForThrottleReduction(). r=ehsan