| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes inconsistent writing of HistoryItem formData
for the case when formData is not null,
but it gets flattened and saved as empty string.
FormData identifier also gets written in that case,
which creates problem on HistoryItem reading,
when parser does not expect to meet form identifier
after it reads zero as FormData size.
Change-Id: I77cf00b5b2a0f2a7cf4355d1bc1b8d4830ac1afb
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Visit orange.deezer.com, press tab marked "My Music".
Browser crashes every time.
Pressing the button will trigger a re-layout of the page.
This in turn will cause some rendering nodes of type
RenderLayer to be removed from the rendering tree.
When such a node is removed, it is important to also
update certain lists in ancestor nodes that may hold references
to this node. A node that may hold such a reference is
identified as being a "stacking context".
However, in Android, when the symbol ENABLE_COMPOSITED_FIXED_ELEMENTS
is defined, the definition of what is a stacking context
is expanded. In this case, a node that is a stacking context
and holds references to descendants, changes one of the conditions
that form part of Android's expanded stacking context definition.
So, now it is no longer a stacking context, but the reference list
is not deleted/updated. When the descendant node is removed a
search for an ancestral stacking context is made, but it will
not find this node since it is no longer a stacking context.
The solution is to make sure that the list of references is
updated/cleared whenever the node changes a condition that
may cause its status as a stacking context to also change.
Change-Id: If5a7b63715020bc3d23749a7c09003a86d90e28d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Accessing cssrulelist in javascript will create a new
CSSRuleList object and wrapper JS object. The wrapper
JS object will be added into hidden array in parent JS
object which is alive during the whole execution. Thus
memory leak happens (CSSRuleList in webkit, wrapper object
and weak global handle in v8).
Cherry pick 2 patches from upstream solves the problem:
http://trac.webkit.org/changeset/90949
This patch changes hidden reference from array to named
property. So new wrapper will replace old wrappper. But the
memory leak still exists because the CSSRuleList wrapper
will be added into an object group of current document. So
they will still be alive during execution.
http://trac.webkit.org/changeset/91256
This patch avoids to adding CSSRuleList wrapper into document
object group. Combined with the first patch, it can resolve
the memory leak problem.
Change-Id: Icb523db52963726f27b6c02596822cfb6e8d5049
Author: Vitaly Repeshko <vitalyr@chromium.org>
Signed-off-by: Xi Qian <xi.qian@intel.com>
Signed-off-by: Shuo Gao <shuo.gao@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Author-tracking-BZ: 32630
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The “pure color tile” optimization finds the pure-color tiles and
then draws the tiles simply with the color (otherwise the texture).
But the RGBA color pointer type is char, which may overflow when
the value is larger than 127. In such case, the tiles in question
are rendered black, because of the wrong color. This bug is fixed
by defining the pointer as unsigned char.
Change-Id: I5c7214423fa6961e707bde72f0e18c610374745d
Author: Yuyang Du <yuyang.du@intel.com>
Signed-off-by: Yuyang Du <yuyang.du@intel.com>
Signed-off-by: Jack Ren <jack.ren@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>
|
|\
| |
| |
| | |
Change-Id: I9b3379ec26c3a0fb208b8cdd66a0b60088c1b01d
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The Java class WebSettings has accessor methods for User Agent
and Accept-Language strings. These methods ensure that the values
are updated if the Locale has changed. However, WebSettings.cpp
in WebKit reads the private fields in WebSettings.java directly,
and thereby fails to trigger the update.
The correction is to use the appropriate accessor methods from
WebSettings.cpp.
Change-Id: Id1aa1f4de38d2a73cad9660afc35944c663dbd4a
|
|\ \ \
| | | |
| | | |
| | | |
| | | | |
* commit '8ac8b3f2e5f239eb5068c9a5ac17dae28cceec8b':
Make viewport dirtyness calculations recursive
|
| |\ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
If a layer marks the viewport dirty, it should mark all of its children as well,
since they likely have also moved.
bug:6547356
Change-Id: I0db2704f35e7c8a78108d6ed8372201a7f025994
|
|\ \ \ \ \
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
surfaces
* commit 'c6d9134e49f290332ca788d10104749a37e9a727':
Don't allow layers that draw via drawGL to merge with other surfaces
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
bug:6694807
They don't paint onto surfacebackings, and draw afterward, so ordering issues
occur. Ideally, they should still be allowed to merge onto surfaces and not
allow other painting layers to join the surface afterward, but this shouldn't be
wasteful in practice.
Change-Id: I7a8764bc2b8e531e718602c6114b027a9dbce42d
|
|\ \ \ \
| |/ / /
|/| / /
| |/ /
| | |
| | |
| | | |
setSizeScreenWidthAndScale"
* commit '903d2ec97b2ac9cf4f50fc3d7f9e503611bce470':
Don't use deleted Node object in setSizeScreenWidthAndScale
|
| |/
| |
| |
| |
| |
| |
| |
| | |
The call to forceLayout may cause the Node object to be
deleted. Avoid crash by check that the Node object is
still valid when forceLayout has returned.
Change-Id: Ieb74a59ca7f2e457b3b17d01bcb02cd566f84ef2
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Bug: 6212665
This is a cherry-pick from master branch. sha:
0ea6dad65eb77f9347a59704b1f8a019cfc9e01f
(cherry picked from commit 57633fd9f863f5e1e4ce8fab0eb05ddb535e1be6)
Change-Id: I9c2fa81aa3e16d39ed79d171ae09b1f99d9e7082
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Bug: 6270583 - Vertical glyph doesn't work after switching from ARM gcc4.4 to gcc4.6.
Fixed 3 problems:
1. substituteWithVerticalGlyphs() does not check the error code for HB_GSUB_Select_Script() call.
In gcc4.4, the scriptIndex is default to 0 but not in gcc4.6.
2. Removed the redundant checks for isCJKIdeograph() calls.
3. Convert text to vertical forms if substituteWithVerticalGlyphs() returns an error.
Change-Id: I3b0778affa3e13b608e60190266880c1f47f905a
|
| |\
| | |
| | |
| | | |
Change-Id: I11384be619b8147cc0a335ee134eb6e91a495398
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Issue: 26518
Fixes a race condition when initializing TilesManager by
removing the unnecessary waitForGenerator call entirely
and giving TexturesGenerator a reference to the TilesManager in
its ctor, removing all calls to TilesManager::instance from a non-UI thread
Change-Id: Ib1b60825275ef24f4733562ed1f6ecace0594c88
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Bug 6667146
Framework Change: I7fc107b02f73f4c54e8fe66484038ea9d7b69d61
Change-Id: I95ffcfa6fa47baad60b1af55f84bee35dfcc7972
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Bug: 6634077
Software was accidentally calling the inner updatePositions instead
of the outer updatePositions which the hardware path calls. Move
the inner updatePositions to be protected instead of public to
guard against this in the future as well.
Change-Id: Ib509f151ce64bf390383eaa0c32a2bae89428e89
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
An issue occurs on Stingray where calling glTexSubImage2d in order to blit an
update to a texture fails silently if the texture has already been updated this
frame.
With this change, we now draw the tile off-screen to work around the issue just
before the second blit is required. This way, the second blit succeeds.
bug:6630040
Change-Id: I07213954669b0f3d3f5b87a748a591e532b097b4
|
|/ / / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
bug:6579681
Regression from https://android-git.corp.google.com/g/#/c/197452/
Change-Id: If185faec6bb2d014c304afbb8ff752ebfc6e4b51
|
|\ \ \ \
| |/ / /
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Bug: 6637033
It seems that somehow these macros were forgotten to be removed
while dropping http stack, it is likely a merge/rebase issue.
Change-Id: I37b14488dbfd786e3e0deab9775f42f21d2205e7
|
|\ \ \ \
| | | | |
| | | | |
| | | | | |
into jb-dev
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Bug: 6635039
This makes sense if you are scrolling the layer with an actual mouse,
but all it does for us is result in some weird CSS :hover style changes
and unnecessary painting - disable it.
Change-Id: I08907c5f536c94a00d72fecf72411a1765a4a067
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The available tile textures pool is not updated while we try to prepare image.
This can lead to getting obsolete tile textures.
bug:6613009
Change-Id: I0170f0cfba197c7a7ed48866028b30075417af30
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Bug: 6569073
This is part a cleanup of destroy to eliminate the gap
between finalize() and destroy() by allowing both
nativeDestroy and nativeStopGL to be called statically
Change-Id: I6418cfe64ca03c58d2a395a239739e1577c54dea
|
|\ \ \ \ \ \
| | |/ / / /
| |/| | | | |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In the slow draw path, we are drawing the small background image one by one.
Now in the fast path, we can use the GL shader to repeat the iamges.
bug:6579681
Change-Id: I8ab413bdd073cb867002b77b36d4a1361290ecb2
|
|\ \ \ \ \ \
| |_|_|/ / /
|/| | | | | |
|
| | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Bug: 6628376
m_scrollLimits has the correct max scroll range, even for iframes
Change-Id: Id6f23e23652e38fe6df8d6540977dc0ab7e8d332
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Bug: 6508448
Cherry pick from master, sha:
6c9f55f6f57bd4da36053c76a1cfce8a15a00a97
This is to make sure webkit usage of scoped refptrs are compatible
with chromium changes, see bug for more details.
Change-Id: Ib8aa3e29409bfe6fd94ef56376bd2ff2297ec868
|
| |/ / / /
|/| | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Bug: 6613097
The bug here is that the clipping rects it is calculating are
wrong, which results in the hit test failing. However, as we
now actually scroll our overflow layers (which we didn't at the
time this was added), this code is no longer necessary anyway.
WebKit knows how to do hit tests correctly, just let it.
This is a partial revert of I0da2d8db
Change-Id: I7b42f36dd199d431508da934878fc0f56e76559c
|
|\ \ \ \ \ |
|
| | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
With no knowledge of the actual SkTypeface used to render text before
render-time, the best we can do is turn fake bold on for all text and
simply ignore it at render-time if the SkTypeface's style is bold.
Also the complex FontPlatformData hashing mechanism needed to account
for fake styles, as the versions of FontPlatformData with fake styles
enabled are otherwise indistinguishable from the normal style version
Bug: 6522642
Change-Id: Ic0e9f1bbd8cae9fdd3a6d1d015bb9224c8be545c
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
bug:6605853
Change-Id: I65583e61197452ac70c5adf5f967182f0fd7578a
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Fixes an issue where tiles outside the new bounds weren't invalidated, as the
layer shrank in one dimension.
bug:6609350
Change-Id: I1f6715670654c9e6211dbec71875aa2ea11cd8f8
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Base surface can expand dynamically (see BaseLayerAndroid::updatePosition) to
fill the viewport, so disable clipping on it.
bug:6593707
Change-Id: I3afafb28fbe95d3162c76f8e835932577ded65ce
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Bug 6594983
Change-Id: I8adf636345e5d5284e372e28176c31c7e937334d
|
|/ / / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Cherry pick of WebKit r119158.
Note the layout test is tweaked slighly from upstream to make it
happy in our older webkit snapshot.
See: http://trac.webkit.org/changeset/119158
Bug: 6578213
Change-Id: Id6cc23938b5139cf0416e3d4c6e7ba227b1b6a0c
|
|\ \ \ \ \ |
|
| | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | | |
bug:6593154
Change-Id: Ic9438f758c7c67d0a2e0579dbc53dcdf8dd109cd
|
|/ / / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Also isolates layer translation fudging so as to not affect child transforms.
bug:6585341
Change-Id: If416c6fad90ee37fbe3438e01706b34e683165b7
|