| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
As the texture generator thread isn't an infinite loop, we could miss
a signaled new request.
Bug:3155078
Change-Id: Ifad1ec356b061eec24399cad4b572866f6d74f06
|
|
|
|
|
|
|
|
|
| |
This CL reverts the rollback of the original CL now that we have
fixed bug #3165953.
This reverts commit 144ccd9c8dba05ffaa0ae598f9b70032050fc20e.
Change-Id: Ibb5254f652ed95aa3b7b71bf24a2578c51e74a70
|
|\ |
|
| |
| |
| |
| |
| | |
bug: 3166223
Change-Id: I5c18b76ac19119c3620636ce5d3b33892e18f523
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| |
| |
| | |
Minor changes to allow check-webkit-style to complete without warning for
FontCacheAndroid.cpp.
Change-Id: Ifaa8634d53374c94251641d6b9daf5f8114af9e9
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | | |
Also fixed indenting (3 spaces?).
Change-Id: If4a137d1af3ec8a597e8b929e5b4f10240f2f054
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
It's safe to pass this across threads since NewRunnableMethod
copies the argument.
Change-Id: I620ebfe64ead7947ac3a6733507865938e04a8df
|
|\ \ \ \
| |_|/ /
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When a page has layers, it causes the main body of the
page to become a layer as well. The scroll position
of the WebView must be applied to the layer to
translate the cursor rings correctly for hit-testing.
Today, applying the scroll position of the WebView
screws up overflow scrolling. Until this gets figured
out, skip using the scroll position in CacheFrame
adjust() and unadjust().
The position of the layer, and all of its parent layers,
should be considered when adjusting the rectangle
bounds contained by that layer. Before, only the
child layer's position was considered.
The clip used by scrollable divs may be initialized
but never set. If it is empty, ignore it, rather than
clipping out all content.
In CacheBuilder, remove a couple of unnecessary
lines -- the CacheInput initialization clears all
members.
In SelectText, reverse the order of the xfer mode
and paint objects to remove a Skia ref count assert.
In CachedInput, make the debug printout current.
Overall, added more debugging output (turned
off by default).
bug:3030370
Change-Id: Ic19c24b3bf33d081a1d0c1f8c06601dcb56ae881
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When we have determined that the form can be autofilled, send
a preview string over to Java so that we can display it in the
drop down box.
Requires a change in frameworks/base:
https://android-git.corp.google.com/g/#change,77127
Change-Id: Ia0dd899d659c6e5710155f33749255058d7c3faf
|
| | | |
| | | |
| | | |
| | | | |
Change-Id: Ic4d4866302645edb6c0a457ef6d0bd1d8f9cf484
|
|\ \ \ \
| | | | |
| | | | |
| | | | | |
rect.""
|
| | | | |
| | | | |
| | | | |
| | | | | |
This reverts commit fc92ca2409a95b539274985812d88016b6b84b7e.
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Adopted the code style change for TextRunWalker class from
chromium/FontLinux.cpp to android/FontAndroid.cpp.
Change-Id: I7f49efcaf5685181980a4ffa3b654d1577e3f9cd
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
On network errors, we were seeing strange "Data connectivity problem"
alert boxes. This is because we were using the HTTP status code instead
of a real error code; and when there's a connection error, the HTTP
status is uninitialised, so the "error code" was 0 (no error).
Bug: 3140283
Change-Id: I0566c29359a0ecb2998573d1a2a77306922c87be
|
|\ \ \ \ \ \
| |_|/ / / /
|/| | | | /
| | |_|_|/
| |/| | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Using existing Skia APIs, FontCache::createFontPlatformData can detect when
the retrieved font is the Skia default and will reject that choice,
allowing WebCore to try the next font on the CSS fallback list. If the
requested family is a generic family, e.g. "serif" or "monotype", the Skia
default will be accepted.
Change-Id: I8007caae96f26cd58ea1cf5bf2da8170e3ce6d9d
Bug: 2720133
|
|\ \ \ \ \
| |_|_|_|/
|/| | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This CL causes the selective repainting of BaseTiles (and their
associated GL textures) using the information provided by Webkit.
If tiles need repainted they are marked with as dirty and the
thread painting and uploading the textures will only operate on
dirty tiles. This change resulted in some significant refactoring
most of which revolved around these changes...
(1) Removed PaintingInfo from the Texture object and instead track
the state of the tile in Tile object.
(2) Removed all pending TileSets for a TiledPage when the page
produces a new set. This ensures that the tiles currently visible
to the user are painted instead of ones that may already be offscreen.
Change-Id: I93845d8e6e7b066e6bab84bcde11be4a6940002f
|
| |_|_|/
|/| | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
First of changes based on willchan's feedback.
Note that this code doesn't really work as intended anyway.
Change-Id: I91d43f1daa78248700738338cf3d24a12496ac19
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | | |
Change-Id: Ide08371e0bc8f74ef80fe25d416623bd03ac6038
|
|\ \ \ \ \
| |/ / / / |
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This also moves some of the android file handling to Java so it is not
duplicated in both places.
This CL needs https://android-git.corp.google.com/g/#change,77400
in frameworks/base
Change-Id: I90c1726e6c323a9de3fd64f2e6feef4b64171053
|
|/ / /
| | |
| | |
| | |
| | |
| | | |
Only 3 tests, all passing. See http://b/issue?id=3164771
Change-Id: I2f3b03969a97618f1d4ba9dd91d782e1814914f0
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
1. Reduce the number of calls of normalization function.
The purpose of NFC normalization is to compose base letter and non-spacing
combining marks to be accented character. i.e. Convert A(U+0041) and
acute(U+0301) to A with Acute(U+00C1).
NFC includes three parts - decomposition, canonical sort and composition
for entire string. It is very expensive. TextRunWalker class would like
to convert '\t', '\n' and non-break space(\U+00A0) to space. NFC cannot
achieve this purpose. So, it is not necessary to call normalization for
Font::treatAsSpace().
This improvement saves the NFC nomalization calls for complex script text
has '\t', '\n' or 0xA0 and does not have non-spacing combining marks.
2. FontAndroid.cpp scans text for space and punctuation in 2 separate
loops. Combine these 2 loops to one pass.
Change-Id: I6b1ff20dd1072cf1f03066ce1c87ad39d0040779
|
| |/ /
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Chromium now uses protocol buffers so we need libprotobuf.
See http://src.chromium.org/viewvc/chrome?view=rev&revision=63221
HttpCache takes an extra constructor argument.
See http://src.chromium.org/viewvc/chrome?view=rev&revision=61789
Change-Id: I5e78b0de9a8ff86040b13212455eb37a4ed0e39e
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This fixes a couple of bugs related to layers
and hardware acceleration.
The picture used by select text is now accessed
asynchronously from the UI thread. The UI thread
now unrefs it rather than deleting it outright.
The layer position may not be up to date when
the cursor rings are drawn, so update them if
the cursor ring is in a layer.
bug:3112657
bug:3030370
Change-Id: I99d07175ec30dcba895a26cf1d44bc0364430793
|
|\ \ \ \ |
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Also fixed incorrect keycode constant for numpad parens.
Bug: 2912307
Change-Id: I65d02ca1c131638ea70f333bf06982747f34223c
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Bug:3155066
Change-Id: I3e249050c779d10db75f97b177e09144a52f70b1
|
| |/ /
|/| |
| | |
| | |
| | |
| | | |
This allows us to show the correct background even if the tiles aren't yet painted.
Change-Id: Ia3ef97770aadb5f64490e44f20110ea21578ccd0
|
|\ \ \ |
|
| |/ /
| | |
| | |
| | |
| | | |
bug:3065048
Change-Id: I18ef06a4b9ac3bef077f08efd738958bd5abaafe
|
|/ /
| |
| |
| |
| |
| |
| |
| | |
C++ counterpart to https://android-git.corp.google.com/g/#change,77500
Bug:2975990
Change-Id: I3b8fd27f991d6776059a15eef36d0c9a7b44f9bb
|
|\ \
| | |
| | |
| | | |
from Chromium headers"
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Chromium headers
If the root of the problem is fixed upstream, this warning can be
re-enabled.
Bug: 3143773
Change-Id: Ic419e353e3d0536d3a3c86f9e8e4615d30d9fb25
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
jstring
The JNI spec allows for local objects to be GC'ed before the JNI frame
returns, so this is dangerous.
Also rename the method to WtfStringToJstring() to match the existing
jstringToWtfString() and update call sites to JNIEnv::NewString() to
make use of it.
Change-Id: I84aa25ba1aca2aee346bf6e3d83a58afc3a95219
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
1. Bypass leading white spaces, since webkit does pass them down.
2. Return better hyphenation point, since previously it's more
aggressive and causes display issue in Google books.
issue: 2672163
Change-Id: I8ae47f7c553f533f752d6f7c697cf2fffd421e5b
|
|\ \ \
| | | |
| | | |
| | | |
| | | |
| | | | |
* changes:
CookieManager.getCookie() should return null rather than the empty string
Hook up remaining CookieManager methods
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This is to match the behaviour of the Android Java HTTP stack.
Bug: 3116410
Change-Id: I28fb1392c36754a7c2fd78bc76281db8318a1056
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
hasCookies(), removeExpiredCookie(), removeSessionCookie() and setCookie()
Requires a change to frameworks/base ...
https://android-git.corp.google.com/g/76898
Bug: 3116410
Change-Id: I6b881465f78e090ed29b215785c4018119ccc5d8
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In this CL we just defer the redirect until WebKit has had a chance
to check it, and potentially cancel it.
There's a comment in the old code asking what to do if we're given
a bad request. Looks like this should never happen; it's a DCHECK in
Chrome so I've turned it into an ASSERT here.
We only follow redirects if WebKit does not modify the URL (this is
the same behaviour as Chrome).
Change-Id: I0c8b8cd61c501527a29dda5aca521a7df4a3ccef
|