| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
Change-Id: I85149bf4405420813d07bee5259f11a192e86a73
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@digia.com>
Reviewed-by: Zoltan Arvai <zarvai@inf.u-szeged.hu>
|
|
|
|
|
|
|
|
| |
Add missing guard to clipboard.h.
Exclude unneeded Windows specific source files from build.
Change-Id: I40fbea35dc1721ddf4e9c184515d57428a7ffe4d
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This delegates the linking step from ninja to qmake so that we can
let qmake decide the destination of the target itself, easing the
deployment and installation logic across platforms.
The module is only deployed as a binary and no header are available
outside of the source tree. This is only to make sure that the
dependence of the QtWebEngine and QtWebEngineWidgets libraries on it
is resolved at runtime exactly the same way as with other Qt modules,
on all platforms.
Ninja still takes care of the compilation and gyp lets qmake know how
and what to link by dumping the list of flags and input files in a
generated .pri file.
This has to be done in a separate .pro file so that we can make sure
that ninja is run inconditionally before make reaches the dependency
check in core_module.pro, ensured by the parent Makefile.
Note 1: This patch removes RPATH hacks that are no longer necessary
Note 2: Other targets like ffmpegsumo are still linked by ninja. The
same logic could be moved to a qmake file but this require some more
work to make sure that some switches (e.g. -stdlib=libc++) are
coordinated between gyp and qmake.
Change-Id: If65968547bde5b9cf732e31e97931c17ae1921a7
Reviewed-by: Zoltan Arvai <zarvai@inf.u-szeged.hu>
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
| |
Replace hash with zeros and add patch to update-patches script.
Change-Id: I9cd7aa557d7951d5b749037c627095e525dbc924
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@digia.com>
|
|
|
|
|
|
|
|
| |
Use these dependencies from environment. No need to include them
in the snapshot.
Change-Id: I81623cc15e8c1e9431c5569985c90a4fbac638e7
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The GTK clipboard integration isn't working for us and we need a way
to get clipboard requests from the renderer into the system clipboard.
This patch adds a Qt implementation to let Qt handle the platform
abstraction, even though the Chromium clipboard direct integration
could work on some platforms.
Change-Id: Ie7af5c6b4c662fc8f4109b2e85e9f594d6d90b0d
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
|
|
| |
Paths are paritally hardcoded for Windows in libvpx.
Change-Id: Ic1c537bcc8b988c9803326709e38f44cb0943ee3
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
| |
DxDiagNode needs build_config.h when using OS_WIN guard.
Change-Id: I15e2b115bf41cd648dafecb362d81f0b8722d485
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
GRIT uses filename strings as keys in its dictionary so generated
filenames must be exactly the same strings as they are defined in
.grd files. GRIT expects forward slashes as path separator and
also .grd files are containing forward slashes.
Similar conversion used in the else section of the outer
if statement in the same function.
Change-Id: I81bbf2e15e7aaefb52176405990d49457fce1589
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
| |
Put it together with the change that enabled the build of event.cc.
Also update the 3rdparty submodule.
Change-Id: I98bea50d053a108e765c890e55f3886fea5f4e93
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
| |
Also add the matching patch files.
Change-Id: I7c9e2ace4070cb1bcbe43286e9d1ad207b389eb6
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Chromium currently triggers the load of an error page synchronously
from the render process when it encounters a load failure.
This has the nasty effect of producing extra loadStarted and
loadFinished signals, the later also emitted as a successful load.
Disable error pages loading until we can implement the error page
extension properly.
Change-Id: Id6aebc6f63bd810b37d89e9297c0b221e8b81448
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Clean the way that we maintain patches by keeping them ordered by
their number and let the shell order them when giving the list to
git am.
Provide a update-patches.sh that maintains the proper command lines
to use and run a cleaning script on the resulting patch to avoid
the SHA1 to be added to the file as it changes every time the patch
is applied.
Change-Id: Ia93cf4b35f5f847a8723a31fc476443864242737
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This follows the model used by the Android WebView's AwRenderViewExt
class.
QtRenderViewObserverHost is attached to the WebContents and
QtRenderViewObserver is attached to the RenderView in the render
process. Both can exchange messages together and allow async commands
to be carried from WebContentsAdapter and the result sent back
through WebContentsAdapterClient.
This patch also adds a renderer subdirectory to start matching the
directory structure of Chromium.
Change-Id: I724ca2fe2a597dcd2a15e8e1a23c4eeba1190703
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
| |
Change-Id: I3bdad02a2ef69d242980e34774a8219ef0c74ea9
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
|
|
|
|
|
|
|
|
| |
Let's make sure that the submodule integration succeeds first.
This reverts commit 6aa4bb60d7d10e50cb562756ef8210f559b35d1e.
Change-Id: I5c9b81654296b3a8a240118f9af222b12dd2ecc4
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
| |
Change-Id: I3cd2ddc401e67c232c4bcc88d3c20252eece6be3
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
NOTE: To build after this you should rerun init-repository.py or run
$> git submodule sync
$> git submodule update
$> git config qtwebengine.chromiumsrcdir src/3rdparty
This makes everything build by adjusting paths.
Other mixed-in changes:
- Rename qtwebengine_src variables in scripts to qtwebengine_root to
avoid confusion.
- Cleanup the release and debug extra targets that were in lib.pro.
This file has also been split into src.pro and core.pro.
Change-Id: Ieee9158a65f526b15244eaca59e779b7069d337e
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A few changes are necessary to allow fetching textures provided by
the render processes through IPC and bound to their respective GL
context in the GPU process and use them in the QtQuick scene graph.
- Remove the plain color test textures.
- Allow setting the QtQuick QOpenGLContext's handle as the share
context for all context set as shared in the Chromium GPU process.
We do this by letting the GpuChannelManager ask us for a
ShareGroup instance responsible for returning a sharing GL context
handle.
- Fetch texture IDs from the MailboxManager used by the GPU process
using the Mailbox given to us in the DelegatedFrameData.
This is the same mechanism used by Chromium to share textures
between "client" GL contexts.
- Keep the QtQuick scene graph threads and Chromium in-process GPU
thread separate. The complicated part of merging those two
rendering pipelines on the same thread is that it would force Qt
to also use only one thread for rendering. For the moment we will
try to synchronize those threads together instead.
- Lock the Qt SG thread while waiting for resource sync points.
Do so by posting a callback to the Chromium GPU thread and wait
until the sync point of every resource has been retired by the
producing contexts.
- Acknowledge the delegated from once QtQuick swapped the GL buffers
instead of right after we added the frame to the scene graph.
This fixes some issues where the textures for the previous frame
would already be released as Chromium was producing the new frame.
There are still a few issues regarding synchronization that have to
be fixed, especially when Qt triggers the rendering of a new frame
while Chromium is starting to produce the next frame.
Note: To enable it we still need to pass the following command switches:
--enable-delegated-renderer --enable-threaded-compositing --in-process-gpu
Change-Id: I2d4f7fac603b1808ed1495a8d689cb48e9ed41b9
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
|
|
|
|
|
|
|
| |
This allows the Chromium GL code to initialize on the display created
by Qt.
Change-Id: I3d21f56c3792d6e99cb94690f56d7dd5c542045a
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
|
|
|
|
| |
Change-Id: Ice94e3cefca317573db3492dd6622d20f4b6e00a
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
|
|
|
|
|
|
|
| |
This seems to be necessary to compile for OSX 10.7,
and this is the OSX version we currently use in the CI system.
Change-Id: I1edcf35baee040328de11801f58ce78ed1f8f6cb
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
| |
It was missing in ac95a438620f3ec7c72ae60396ee26b3778d38b2.
Change-Id: Ib08ea6466835e9d2c55bc7944ae1b758370da30a
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
|
|
|
|
|
|
|
|
| |
This seems to be necessary to compile for OSX 10.7,
and this is the OSX version we currently use in the CI system.
Change-Id: I2363b48b66cd7fda603ff9783934ca44c1656723
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This provides a stub implementation that renders quads and RenderPasses
using QSGNodes in QQuickWebEngineView.
The BackingStore code path is still supported when the delegated renderer
switch is not enabled. To use the new rendering, pass the following switches:
--enable-delegated-renderer --enable-threaded-compositing
This uses debug, plain color textures until we can fetch tiles and
textures from the ResourceProvider across OpenGL context boundaries.
Change-Id: I33ea0738dc5a326ef79c1435f75c044c42e8551f
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The regular workflow now is just to clone the qtwebengine repository
and then execute the following commands.
git submodule init && git submodule update --recursive && qmake && make
This will also clone a submodule called 3rdparty which contains the
chromium and ninja sources without pulling in further submodules
and without a complete history.
Developers that do want to have a complete chromium checkout instead
should not use the above command sequence.
Instead they should just clone qtwebengine, execute the
init-repository.py script, run qmake and make.
The init-repository.py script will then checkout the complete
chromium sources into a subdirectory called 3rdparty_upstream.
The location of these sources will be picked up by qmake automatically.
Change-Id: I0fa4f1d554bdca2e852b6a97aa2e5462d90d8664
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Including Assertions.h leads to __darwin_ssize_t being
defined inside namespace WebKit but not outside that namespace.
This is because certain headers will not be pulled in again
later on in the same compilation unit.
This patch has already been upstreamed,
and can be removed with the next Chromium update.
Change-Id: I302bbf779fcabb909a9ffda53d8b9fdccc87e248
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
| |
Change-Id: I64aa8e01e3d157670f2fd118d1ed25730b61a967
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
|
|
|
| |
- googleurl headers are now in url/.
- dom_storage_namespace.h moved.
Change-Id: I87f08d2426ac0da5b50fcf424ca446fb03f207fd
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
| |
Change-Id: I93ceb4e5d1ccf63cbeaec1fdff4ab1c5c8a57a65
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Remove -Wno-deprecated-register from common.gypi.
This warning is not supported by recent clang versions
shipping with XCode. Therefore replacing it with -Wno-deprecated.
This is a bit a too broad fix, but the best we can do for now.
It seems that the flag is still supported in upstream clang.
Change-Id: I609355484c9b3ae111591a17181f02104bed6b6b
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
|
|
|
|
| |
Change-Id: Ib592d8498cabdafda0dc65cf81de4721f139ffc8
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
WebTouchEvents have a slightly different behavior than QTouchEvent in
that the type of the event is TouchStart for each new point press,
while Qt sends a TouchBegin only for the first point press.
Since we already need to use ui::TouchEvent to be able to use
ui::GestureRecognizer, always do this conversion first to also let
UpdateWebTouchEventAfterDispatch chose the proper event type.
Some of the code from render_widget_host_view_aura.cc was copied into
render_widget_host_view_qt.cpp to fill the needed functionality.
Change-Id: Iab1ca0c449b5256a39b5479ce89b662d4e133935
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With this patch init-repository.py will also abort uncommited
git am/merge/rebase sessions and it will do a git reset --hard
on every submodule.
We are no longer executing a 'git submodule update --recursive'
in the patch-chromium.sh script, since init-repository.py
takes care of checking out the proper revisions.
Change-Id: I6cc50944d888ffce5e3d282500dffe3daffecdf4
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
| |
Change-Id: I9e197aafd3dd342ba965ac27ae3b2543cd377bf3
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
| |
Change-Id: I6d0cf5ada32594d91c680518868055fc03466ac7
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Override base::FILE_EXE in ContentMainDelegateQt::PreSandboxStartup()
with the path to our child process and get rid of the command line
argument and the content::CHILD_PROCESS_EXE override.
This reduces 0001-My-local-fixes.patch to export ContentMainRunner
so the patch can be renamed to better reflect it's purpose.
This hunk is still needed since we need to access ContentMainRunner's
Initialize() to set our ContentMainDelegate.
Change-Id: I808cb114100300a33f8cb1b154c6cde89cbf5d98
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
ContentBrowserClientQt is a singleton which makes it possible
to access it from WebContentsViewQt and removes the need for
patching chromium.
This is similar to how ShellContentBrowserClient is managed
in the content shell.
Change-Id: I67f35520935388888c7230806ad543a58b3211c3
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
QtDeclarative is consciously leaking the namespace QtQml.
Therefore clang would warn about header hygiene.
Unfortunately this cannot be fixed in QtDeclarative,
unless we accept not to be able to use QtQuick1 and QtQuick2
in the same application.
Change-Id: I5ba78eb20f98e2e04d9c190794c83710b7f196e4
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
|
|
|
|
|
|
|
|
|
|
| |
We currently only have a single process binary for Qt.
Therefore we should not choose a different binary based on required privileges.
However, this patch can only be a preliminary solution.
May be we should instead implement our own ChildProcessHost.
Change-Id: I0e76ff44f474842c7a96317a05f9964782f68080
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
| |
Adding several patches to the repository that need
to be applied to the chromium sources in order to allow
building on Mac.
Change-Id: Ie06250a828b3533e2f48563ce374e63fc25d16cf
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The new submodule layout does not need gclient anymore.
Instead it relies on the init-repository.py script and git submodule.
We are still checking for CHROMIUM_SRC_DIR for the time being.
Just in case somebody still wants to use external chromium sources.
In this case it will be necessary to master gclient manually though.
Change-Id: I1633cda873c94fd86433be1f60fd422c7f230cb2
Reviewed-by: Andras Becsi <andras.becsi@digia.com>
|
|
|
|
|
| |
Change-Id: If41e2991be7877060f7739877f756acf2679b688
Reviewed-by: Zeno Albisser <zeno.albisser@digia.com>
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Rename the class to WebEngineContext and keep it as a ref-counted member
of pages instead. Also:
- Change the user-agent product to QtWebEngine
- Don't pass actual command line arguments to Chromium anymore
- Allow attaching to the event loop through Before/AfterRun instead of blocking
|
|
|
|
|
|
| |
Use a gypi files for that with target defaults with our file
exclusion patterns.
Another patch bites the dust !
|
|
|
|
| |
Unfortunately this requires some minor reinterpret_casts.
|
| |
|
| |
|