Age | Commit message (Collapse) | Author |
|
This is a clean Cherrypick of 2 CLs from chromium:
https://codereview.chromium.org/920623002
And https://codereview.chromium.org/922883002/
Original description:
Don't send extra text changed AX events on Android
This fix is small and safe, for merging. The reason this
was happening has to do with contenteditable support
and I'll follow up with a cleaner fix once this is merged.
BUG:19294187
Change-Id: Ifd0de22fdad20a917ca74174c90fce8bec7d87ad
|
|
scheme."
Fixes a crash when WebView apps which have enabled
WebSettings.setAllowUniversalFileAccessFromFiles(true) try to call JS
history functions with URLs that are from a non-file:// origin, which is
used by some Cordova apps.
> BUG=449075
> TEST=content_unittests --gtest_filter=NavigationControllerTest.IsInPageNavigation
>
> Review URL: https://codereview.chromium.org/855883002
>
> Cr-Commit-Position: refs/heads/master@{#313051}
Bug: 19173646
|
|
Clean cherry-pick of chromium
crrev.com/d6552661356cafdc55df5e5d6abf378c9065c04c
Note crbug.com/450396 not an issue in m39.
BUG: 18926944
Original description:
Virtual context is turned on according to FeatureInfo::Workarounds.
Note that this doesn't work on Android WebView until crbug.com/450396
is fixed.
Change-Id: Ib6c997f3929293023288188c4766c1a4b032c407
|
|
When using a prebuilt chromium, the build was failing as the set of V8
makefiles being included is no longer correct with the M39 version of
Chromium. Remove generate_trig_table which no longer exists, and add
v8_libplatform which is a new dependency required to build the host
mksnapshot binary.
Bug: 18675666
Change-Id: Ic00eea284a5b40af41883e000e092ee16d6bfcc7
|
|
Clean cherry-pick of chromium
crrev.com/7476088d404e3b04f2cf805f3e0622283606ea77
BUG: 18685397
Original description:
A small and safe workaround to get some failing webgl conformance tests
passing in android webview.
Change-Id: I56d718f1bd124a9d9aad72c92e0ce57e554c9b57
|
|
|
|
Cherry-pick with conflicts of
crrev.com/81c62e2d9601d38c44f71857dfabc31e2726cd70
BUG: 18706908
Original description:
The problematic sequence of events is as follows:
1. ShouldRequestOnNonUiThread, which posts a closure
(request_draw_gl_closure_) to UI thread
2. That closure gets run on UI thread, and it schedules the invokeFunctor
with Android framework
3. Before the corresponding invokeFunctor actually happens on RT (which
is DrawGL process mode), ShouldRequestOnUiTdread is called on the UI
thread. At this point, pending_non_ui_ is not null, we cancel the
callback, which does nothing, because WebView has already scheduled an
invokeFunctor with the Android framework in Step 2. Then we schedule
another invokeFunctor immediately on the UI thread. So there are 2
invokeFunctors queued in Android framework in this case.
This CL tries keep track of whether or not we've
queued an invokeFunctor in Android framework already.
Change-Id: I0a084d92ea74412c3938645d63170604e6352318
|
|
Cherry pick from https://codereview.chromium.org/794023002
Original description:
After CL: https://codereview.chromium.org/369703002, Android
WebView just has a NULL SSLHostStateDelegate. So it does not remember
any user decisions on SSL errors. This is a regression of behavior from
L (m37).
History:
JB behavior: Larger error codes are assumed to have higher severity. And
if the user has allowed an SSL error with a high severity, the user
won't be prompted for a lower severity SSL error.
K and L behavior: A specific SSL error will be allowed only if the error
bit field is a subset of previously allowed error codes.
trunk behavior for webview (without this patch): We don't remember
user's decision at all.
This CL:
Maintain the same behavior with K and L.
Conflicts:
android_webview/browser/aw_browser_context.h
android_webview/javatests/src/org/chromium/android_webview/test/AwContentsTest.java
android_webview/javatests/src/org/chromium/android_webview/test/TestAwContentsClient.java
BUG: 18665047
Change-Id: I0ecbc0b78e9323ef911cd61679fec5e4a6146489
|
|
|
|
Bug: 18736075
Cherry-pick https://codereview.chromium.org/796003002/.
Clean cherry-pick.
The proxy config service for android does not allow
one to configure an exception for localhost be made.
The implementation for Linux did allow this in:
proxy_config_service_linux.cc
it just seems it was overlooked for android.
Instead of adding in the ability to configure
localhost as an exclusion, it makes more sense
to just force the exception; it seems silly to
send traffic for localhost through a proxy.
This patch addresses the following bug in android:
https://code.google.com/p/android/issues/detail?id=37031
Change-Id: I3b654d45e9f534961cae9147374023ef597aa927
|
|
BUG: 18725105
Change-Id: I384fde4c9b36c948e995841469b1061da763584d
(cherry picked from commit 5ea7c0fdf34dbf084fc7a0f6c540de3a2e0a24f3)
|
|
background thread
Same as the upsteam patch, with purely syntactic adaptations of "override"
keywords usage to the legacy style used in M39.
Bug: 18520475
Original description:
commit a4b0d4cc0b9780aecd3e23a94293b2e3002baf7e
Author: mnaganov <mnaganov@chromium.org>
Date: Tue Dec 09 13:30:08 2014
[Android] Java Bridge: handle requests from Java Script on the background thread
This allows the page code that calls into injected objects methods to run even
if the browser UI thread is blocked (in WebView, the browser UI thread is the main
app thread, and sometimes people use synchronization primitives there that wait
on some action performed by the injected object, see the bug for an example).
In this CL, we make the minimal required amount of changes in order to simplify
cherry-picking into older branches. A cleanup CL (or CLs) will follow.
High-level description of changes:
1. We intercept messages from renderers on the IO thread using BrowserMessageFilter
and handle Java Bridge messages entirely on the background thread. This implies
the following changes:
2. GinJavaBridgeDispatcherHost has become RefCountedThreadSafe
(via BrowserMessageFilter)
3. We have to use route IDs of RenderFrameHosts instead of pointers to them, as
we can't access RFHs from IO or background threads.
4. Objects registry is now accessed from UI and background threads, and we have to use
locking. We can't anymore restrict access to the registry to a single thread.
This also means we can't anymore implement the registry using IDMap, as it is
explicitly non-thread-safe.
5. A lot of code was removed, as we now serve synchronous requests from renderers
without inter-thread hopping!
6. Of course, a new regression test has been added that now passes.
BUG=438255
Review URL: https://codereview.chromium.org/772123002
Change-Id: I9e517024878f0b95a32f95524de4f8017ba8dc21
Cr-Commit-Position: refs/heads/master@{#307460}
|
|
Required for the main patch fixing b/18520475.
These changes are purely in the syntax -- retrieve the underlying raw pointer
directly, instead of relying on overridden operators. This is to match
the current code in M40 and trunk.
Bug: 18520475
|
|
interfaces removal
Required for the main patch fixing b/18520475.
Bug: 18520475
Original description:
commit 4d4182c4af0a11a1ad5acae88f9abe2b76853fe1
Author: mnaganov <mnaganov@chromium.org>
Date: Mon Dec 01 11:59:14 2014
[Android] Fix a subtle issue in Java Bridge regarding interfaces removal
Update JavaBridgeBasicsTest.testRemovalNotReflectedUntilReload to actually
call Java GC after removing the interface, and verify that the Java object
is still callable from the page side. Fix the code to make the test pass.
BUG=437761
Review URL: https://codereview.chromium.org/767453003
Cr-Commit-Position: refs/heads/master@{#306176}
Change-Id: If12d53e3aad721e5822d8e12c85f1240e1aac5e6
|
|
This commit was generated by merge_to_master.py.
Change-Id: I57fcd19e820d8f60c46f407b04ebb2daeb20203d
|
|
This commit was generated by merge_to_master.py.
Change-Id: I6695b44b7c48580907460fb1a183c166e6ab505f
|
|
Disable NOTIMPLEMENTED() as the messages generated are not actionable by
app developers.
Bug: 18291756
Change-Id: I8d52c69874fe2818ac58235d9441936798dc60d6
(cherry picked from commit bf063fe8d7408cd65ec84b146e4254baa2aa7055)
|
|
This commit was generated by merge_to_master.py.
Change-Id: If5c41c8901c327e0abeb154e756e609eec2d8c0d
|
|
This commit was generated by merge_to_master.py.
Change-Id: I20d82af796d2bd2787e3300a2b35ea1b00216a26
|
|
|
|
This is a clean cherry pick from:
https://crrev.com/06091eaf4543692caf7a0bcaf5c3a550dfca7078
We need to do this because in the Android WebView pop up flow, the
native WebContents object could possibly have received RenderViewReady
IPC message before the CVC is constructed and starts to observe the
WebContents.
BUG: 18527015
Change-Id: Ibccac065c0607bd2ec45c0a9e152128ea8805b7e
|
|
Cherry pick from:
https://crrev.com/a43a08a85da0a2d2ed78a850fa9f8efe8c0028b9
Original description:
Detecting if we're using synchronous compositor is unreliable because
the synchronous compositor could be created either before or after
RWHVA constructor.
Instead, we can detect if we're using browser compositor in RWHVA. And
this can be done reliably at RWHVA construction time.
Conflicts:
content/test/run_all_unittests.cc,
content/browser/renderer_host/render_widget_host_view_android.cc
BUG: 18377042
Change-Id: I5ad377de0c8651a76eb51776eca7e7ad3fd3e9cd
|
|
clipped." into m39
|
|
Cherry pick from Chromium:
https://codereview.chromium.org/751453005/
Only the fix part from the original CL is cherry picked.
The other changes in the original CL are all tests.
Original description:
An empty clip rect of a render surface does not necessarily mean
the surface is unclippped.
BUG=435797
BUG: 18343246
Change-Id: I56b8f2d557b21aa75619157e611ab4da9bd4688c
|
|
|
|
Cherry pick from Chromium:
https://crrev.com/d1d74981ce7baa010996c82a7b3423d06af466f1
Original Description:
When a paste or cut operation is performed in an editable region, and
the action bar is dismissed, the IME state should be preserved. Ensure
this by always using the editable region bit when clearing the selection
upon action bar dismissal.
BUG=435810
BUG: 18200283
Change-Id: Iefe239341010236ea6833c98ade34499970aae62
|
|
This commit was generated by merge_to_master.py.
Change-Id: Ia17c23f18c98496f73176f545e8331805640ccec
|
|
This commit was generated by merge_to_master.py.
Change-Id: I1bb301dcfe74249d8b66fb60ba577e9095c86b8d
|
|
|
|
Clean cherry-pick of chromium
crrev.com/18abdf1afb8034be81213b23d9e05d34e92bc8d3
BUG: 18386625
Original description:
This is a small webview-specific workaround that affects all drivers.
Will need a better driver/version specific workaround later.
Change-Id: Iec566f9f166397a104badb8b549cc59e6173faac
|
|
* changes:
Update volantis workaround to cover L mr0.1
Cherry-pick: Workaround NVidia eglWaitSyncKHR bug
Cherry-pick: gpu: Use egl client wait if server wait not available
|
|
This patch fixes a regression from K to L. Tapping the back button
while in fullscreen should exit fullscreen instead of the current
activity.
BUG:17985005
Change-Id: I7a5eb4edebb54ce1e7d6a2b578a87dd7e7de9dbf
|
|
|
|
Bug: 18388001
Clean cherry-pick from
https://crrev.com/200b2f027ce820425e37b08ed8b4e5dc84928276
Android platform provides a local proxy that also handles PAC proxy
resolution. The CL below modified enabled chrome network stack to start
using V8 proxy resolution on Android platform:
https://codereview.chromium.org/573013002/
however webview cannot use V8 proxy resolution since Webview is significantly
different from chrome: Webview is single process vs. chrome is multithreaded
and use of V8 isolates creates concurrency problems with Blink's use of V8
isolates.
For now, we want to establish the prior behavior for android webview and
consider making changes to v8 proxy resolver in future.
BUG=432539
Committed: https://crrev.com/200b2f027ce820425e37b08ed8b4e5dc84928276
Cr-Commit-Position: refs/heads/master@{#305166}
Change-Id: I58becc2acf369956df6c458df13a53dac05caff3
|
|
BUG: 18412138
Change-Id: If713c5fb10c2d7317c08f1d85249f621161ed982
(cherry picked from commit 293eed332e8b9da195b229c36299d66f1604fc04)
Conflicts:
gpu/config/gpu_driver_bug_list_json.cc
|
|
Clean cherry-pick of chromium
crrev.com/ba561e7fb76fef1a6dfff772b08bd6dde5c94bc9
BUG: 17690996
Original description:
Workaround is to fallback to client wait.
Change-Id: I8d6b41f5bc94307faecffe1401b660a3a722e24d
|
|
Clean cherry-pick of chromium
crrev.com/bf634aa33aa21ce25ef680bca6305b9ad64acfbc
BUG: 17690996
Original description:
Check for EGL_KHR_wait_sync extension before calling eglWaitSyncKHR. If
extension not present, fallback to EGL_KHR_fence_sync and
eglClientWaitSyncKHR.
Change-Id: Ia3dad04d468ffebb88b549adb06a310f933d9b5b
|
|
|
|
Bug: 18297802
Cherry-pick disable channel id
Android webview does not use channel id since there is no API
to share channel ids and this would break some apps that use
different network stacks. This change is to explicitly
disable it.
BUG=433427
internal bug b/18297802
Review URL: https://codereview.chromium.org/731593003
Change-Id: Iefd0f61c922227d562667f4a53f0bb35b03b444a
Cr-Commit-Position: refs/heads/master@{#304316}
|
|
is detached" into m39
|
|
Clean cherry-pick of chromium
crrev.com/d51117a99a4e023e791134185c6db12b28028a1d
BUG: 18283959
Original description:
Chromium shares a message loop with Android on the browser UI thread.
This can cause problems when the associated Looper has a sync barrier,
preventing posted Chromium tasks from being dispatched until the
barrier is removed. Make this sharing more fair by marking all Chromium
Message tasks as asynchronous, avoiding stalls when there is a sync
barrier.
Note: This change was originally landed in crrev.com/512333002, but was
reverted to gather more data about sync barrier impact and investigate
alternative solutions. The investigation determiend that this approach
is fine for the near future, though it will likely evolve in time.
Change-Id: I9879a9c1bdad0dcec12c2bbee62356a5b12d9596
|
|
tracing (patchset #6 id:100001 of https://codereview.chromium.org/575103002/)
Cherry-pick of chromium
crrev.com/adefc40cb9652a2ea046bbdbef223d881f4a7463
Conflicts:
base/android/java/src/org/chromium/base/SystemMessageHandler.java
BUG: 18283959
Original description:
Reason for revert:
Barrier detection is no longer necessary, see crrev.com/512333002.
Original issue's description:
> [Android] Experimental sync barrier detection for tracing
>
> When an Android View is invalidated outside of frame dispatch
> (animation/input/draw), a sync barrier may be inserted into the shared
> UI thread message loop. This effectively blocks dispatch of all Chrome
> tasks posted to the browser UI thread for an entire frame (or more if
> the View is continually invalidated).
>
> There are currently no easy or even automated ways to avoid this
> untimely invalidation, and judicious coding can only take us so far. As
> an intermediate assist in debugging the issue, use reflection and a
> crude form of MessageQueue inspection to trace known cases where the
> MessageQueue is blocked by a sync barrier.
>
> Note that this detection is not perfect, neither is it exact, but it's a
> solid proxy for informing traces about such pipeline stalls.
>
> BUG=407133
>
> Committed: https://crrev.com/85bceaa6e05ee1bf42ac220ddcbb50ca268c1ea4
> Cr-Commit-Position: refs/heads/master@{#296249}
Change-Id: Ie92ab895b864bf466ac092a1a1b7fdc80a8fd3ce
|
|
detached
Bug: 18373995
Original Description:
PopupWindows are used to render text handles in WebView. As these popups
are drawn outside of the web contents View hierarchy, it's possible that
their visibility becomes unsynchronized with the WebView itself. For
example, after the WebView has been detached from the window but
before it has been properly disposed of. Address this by overriding
handle visibility when the (content) View is detached, restoring it
only after re-attachment.
BUG=430859
(internal b/18066784)
Change-Id: I43b5ff63368384c3ff30ceece7ddd86aa0a4b738
Committed: https://crrev.com/0563c2785d0b1e62793ffb0f9dc3ac178559aeaa
Cr-Commit-Position: refs/heads/master@{#303987}
|
|
|
|
|
|
Bug: 18200283
Cherry-pick https://codereview.chromium.org/713493003.
Clear the selected text when hiding the handles and the desired
behavior is to "unselect all on dismiss".
BUG=430859
internal b/18200283
Review URL: https://codereview.chromium.org/713493003
Change-Id: Ie8d751774c6341054f0c1b8d4cca91ba6877e373
Cr-Commit-Position: refs/heads/master@{#303946}
|
|
Cherry-pick of chromium
crrev.com/2c8b61a4db617427448947ff740ba002da1a9644
Conflicts:
cc/resources/picture_pile_unittest.cc
BUG: 17965814
Original description:
Previously, there was a hack/optimization where when shrinking a pile,
there would be no invalidations. Unfortunately, this hack's time has
come due. When zooming out, recorded pixels outside of the pile's newly
shrunken bounds can get used, causing pixels from old frames to get
incorrectly sampled, causing bogus hairlines.
To fix this, issue invalidations for previously exposed content in the
same way that invalidations are currently issues for newly exposed
content.
Change-Id: I3287e7e2261328a37781bf58da97ac4b7e60cb73
|
|
|
|
|
|
|