Skip to content

GitLab

T
tangled
Switch branch/tag
  1. 21 May, 2018 33 commits
  3. 20 May, 2018 7 commits
    • Paul Miller's avatar
      WebView: Require charging for variations seed download · 76b89239
      Paul Miller authored 
      This follows the advice of go/power-analysis (internal document). The
download frequency is capped at 1/day anyway, so this shouldn't have
much effect on seed freshness.

BUG=733857

Change-Id: Ifceecf6877542af317ea08128af9bdeb5c6fb991
Reviewed-on: https://chromium-review.googlesource.com/1066801Reviewed-by: default avatarChangwan Ryu <changwan@chromium.org>
Commit-Queue: Paul Miller <paulmiller@chromium.org>
Cr-Commit-Position: refs/heads/master@{#560209}
      76b89239
    • Morten Stenshorne's avatar
      [LayoutNG] Improve support for negative clearance. · 9157c841
      Morten Stenshorne authored 
      When there are adjoining floats that need to be separated from a cleared
block, clearance is predetermined, and we have to force it on the block.
Any large top margin on the child will just be eaten by negative
clearance. There's nothing that can be done to prevent clearance in such
cases. So make sure that we don't try to determine whether to apply
clearance or not based on the clearance offset set in the constraint
space. When clearance has been predetermined (which may have triggered a
BFC offset resolution and what not), refusing to apply it to the child is
a bug.

This aligns the implementation with the "alternative" way of calculating
clearance in the spec [1]. Everyone but Edge does this. I think what
Edge does here is problematic, because it requires us to use a
hypothetical position that was calculated before clearance got applied
(clearance causes margins that would otherwise collapse to be
separated). We'd end up using a hypothetical position not based on the
actual layout situation.[2]

[1] https://www.w3.org/TR/CSS22/visuren.html#flow-control
[2] https://github.com/w3c/csswg-drafts/issues/2608

This fixes one existing test. Added a few new ones as well. Not all of
them failed prior to this CL, but they serve as regression tests for
things I found to lack coverage while working on this.

Cq-Include-Trybots: master.tryserver.chromium.linux:linux_layout_tests_layout_ng
Change-Id: Ia45e9048c75c338477eee4923ff16eea3245bc6a
Reviewed-on: https://chromium-review.googlesource.com/1061470
Commit-Queue: Morten Stenshorne <mstensho@chromium.org>
Reviewed-by: default avatarIan Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#560208}
      9157c841
    • Darren Shen's avatar
      [css-typed-om] Ignore surrounding whitespace in CSSNumericValue.parse. · b3c62041
      Darren Shen authored 
      Fixes an issue where CSSNumericValue.parse considers valid strings
surrounded by whitespace to be invalid.

Bug: 844290
Change-Id: I4c40044631276a611ef9a56aac21ddf09316a550
Reviewed-on: https://chromium-review.googlesource.com/1065532Reviewed-by: default avatarRune Lillesveen <futhark@chromium.org>
Commit-Queue: Darren Shen <shend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#560207}
      b3c62041
    • Ken Rockot's avatar
      Reland: Mojo EDK: Improve internal process handle ownership · 21070080
      Ken Rockot authored 
      Re-reland of r555117 with process handle leak actually fixed.
Totes for reals. Original commit message follows.

Mojo passes around base::ProcessHandle values for various reasons. On
most systems this is fine, but at least on Windows, a ProcessHandle
refers to an owned reference to a system process object, and if not
careful it's possible for a base::ProcessHandle value to inadvertently
change meaning over time.

This CL introduces the concept of a move-only ScopedProcessHandle
within Mojo, which on most platforms is just a base::PlatformHandle.
On Windows, this represents an owned base::ProcessHandle which closes
on destruction and clones correctly using DuplicateHandle rather than
merely copying the raw handle value.

ScopedProcessHandle is used in a few places where process handle
ownership semantics were previously weaker than necessary, or were
correct but implemented ad hoc.

This also updates ScopedPlatformHandle (and supporting code like
Channel::RewriteHandles) such that the |owning_process| field (if not
the current process) is always an owned process handle. This ensures
that when such handles are closed in unsent messages, they can be
safely closed in the target process (from within the source process)
without any risk of raciness against target process termination.

TBR=jcivelli@chromium.org

Bug: 837612
Change-Id: I182e6849fd87fb44e22f6ed21457b17422956613
Reviewed-on: https://chromium-review.googlesource.com/1066999
Commit-Queue: Ken Rockot <rockot@chromium.org>
Reviewed-by: default avatarKen Rockot <rockot@chromium.org>
Cr-Commit-Position: refs/heads/master@{#560206}
      21070080
    • skia-chromium-autoroll's avatar
      Roll src/third_party/skia/ da17f1e7d..5191880cb (2 commits) · 56a0f631
      skia-chromium-autoroll authored 
      https://skia.googlesource.com/skia.git/+log/da17f1e7df1d..5191880cbf3e

$ git log da17f1e7d..5191880cb --date=short --no-merges --format='%ad %ae %s'
2018-05-18 caryclark remove toString
2018-05-18 brianosman Add drawBitmapLattice to SkDebugCanvas

Created with:
  roll-dep src/third_party/skia
BUG=chromium:830651


The AutoRoll server is located here: https://autoroll.skia.org

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
TBR=allanmac@chromium.org

Change-Id: Ic416a02f00901a15956ee78e8f87dd9ee6d8e3cf
Reviewed-on: https://chromium-review.googlesource.com/1067001
Commit-Queue: skia-chromium-autoroll <skia-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Reviewed-by: default avatarskia-chromium-autoroll <skia-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#560205}
      56a0f631
    • Rohit Rao's avatar
      [ios] Provides anchoring information for the Bookmarks context menu. · cceb2b29
      Rohit Rao authored 
      When the UIRefresh UI is enabled, bookmarks is presented using a
UIPresentationController with a custom presentation.  UIAlertController chooses
to draw as a popover in Regular horizontal size classes, so we need to provide a
sourceView and sourceRect.  For now, provide the center of the bottom UIToolbar,
as that roughly corresponds to the "More" button.

BUG=None

Cq-Include-Trybots: luci.chromium.try:ios-simulator-full-configs;master.tryserver.chromium.mac:ios-simulator-cronet
Change-Id: I6e732460eb371158ac9e1754d752b923965f5b0b
Reviewed-on: https://chromium-review.googlesource.com/1064916Reviewed-by: default avatarSergio Collazos <sczs@chromium.org>
Commit-Queue: Rohit Rao <rohitrao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#560204}
      cceb2b29
    • Takuto Ikuta's avatar
      GN: Handle deps always in create_bundle target · 75dfa0ea
      Takuto Ikuta authored 
      Stop to use explicit dependency for indirect dependency in the CL broke
Mac builder.
https://chromium-review.googlesource.com/1041506

This CL uses deps/public_deps in create_bundle target always not to
miss the dependency.

Bug: 843470
Change-Id: Iae66f2a3aec20f67ea2e880bb5fd33d6c2aa2f91
Reviewed-on: https://chromium-review.googlesource.com/1065591Reviewed-by: default avatarSylvain Defresne <sdefresne@chromium.org>
Reviewed-by: default avatarDirk Pranke <dpranke@chromium.org>
Commit-Queue: Takuto Ikuta <tikuta@chromium.org>
Cr-Commit-Position: refs/heads/master@{#560203}
      75dfa0ea