To prepare for enabling MissingNullability Metalava check this CL
works on adding missing nullability issues that metalava flags if
we tell it to flag new things since API 29.
This is not a complete CL, mostly addresses public api and
toString/equals for @SystemApi
Exempt-From-Owner-Approval: Large scale nullability clean up
Bug: 124515653
Test: make -j checkapi
Change-Id: I109260842cfc25f06e40694997fcbb4afa02c867
The Javadoc of addJavascriptInterface() mentioned that the injected
Java object will be injected to main frame, but this is not telling
a full picture. The current implementation will inject the Java
object to all the frames actually.
This CL corrected the misleading wording and add a new warning to
explicitly call it out as a security risk.
Bug: 113336656
Test: make ds-docs
Change-Id: Ia79381d1ab38afa963ea7365526749c14e25238c
No change to logic, only docs.
Use class=prettyprint so docs are syntax highlighted. This also fixes a
section which should have been an ordered list but abused <pre> instead.
Bug: 125526827
Test: make -j4 docs
Change-Id: Iac780ace066be35620d5121b1977df404eaf18c1
Update the documentation for WebView.findAddress, as well as the related
functionality in Linkify and TextView, to clarify why the method is
deprecated, why it should not be used, and that it can cause unexpected
exceptions to be thrown from several places on older OS versions.
Fixes: 24676033
Test: m offline-sdk-docs
Change-Id: I45d82b9a4c9cf62d9566898dd21cd2139ad98f37
Most pre-Q devices actually reserved around 130MiB in the 32-bit zygote
for the WebView native library; 100MiB was only the lower limit and the
file size based calculation that older OS versions use would come up
with a larger number. Increase the hardcoded value for Q to 130MiB so
that the behaviour is roughly the same across OS versions.
Fixes: 132081320
Test: manually verified via inspection of process maps
Change-Id: I0f03ac2f118bb0627ea6be229e5064d4590d8afb
A separate CL in this topic adds MimeMap to the @CorePlatformApi and
deletes MimeUtils. This CL inlines calls to MimeUtils to call MimeMap
instead.
Bug: 120171727
Test: Treehugger
Change-Id: Ibad1e62a63342fa7527332d2b156b69517956364
We start WebViewZygote at boot time to ensure the first app to use
WebView doesn't have to wait, but we were not waiting for RELRO creation
to finish before doing so. Move the background task which starts the
zygote to WebViewUpdateServiceImpl and have it wait for the RELRO first
to ensure that we get the memory saving of RELRO sharing whenever
possible, instead of only when the timing happens to work out by chance.
Fixes: 130305037
Test: manual, check logs and relro sharing status after boot
Change-Id: I55c3f80b0db1dc82727b90c70f777618ca77a942
We no longer need to support fallback packages and need to migrate them
back to being enabled (but only once, to allow the user to disable them
manually later). This CL:
1) Removes the logic which enables/disables packages and the test cases
which only cover the enabling/disabling logic.
2) Checks at boot time if the fallback logic is enabled; if so, it
re-enables the fallback package if one exists and then disables the
fallback logic permanently.
3) Amends remaining test cases to use "primary" and "secondary" instead
of primary/fallback, to keep testing that higher priority packages
are more preferred without involving the fallback behaviour.
4) Fixes a bug in the package selection logic that would not have
triggered in the previous setup (as our production configs never had
more than one available-by-default, non-fallback packages) that was
revealed by one of the modified test cases: the boot time logic to
reset the user's setting if their previously chosen provider was no
longer available triggered even if the user had not actually chosen a
provider, causing the automatically selected provider to be
considered a manual user choice, and preventing a higher priority
provider from being automatically chosen if it later becomes
available.
Bug: 126588129
Test: atest WebViewUpdateServiceTest
Change-Id: I532c0faae2427de7132f716725bf62e256975410
The WebView stub is only used in Monochrome configurations and is no
longer required with Trichrome. Remove the special case code used to
support it in the past.
Bug: 129470358
Test: atest CtsWebkitTestCases
Change-Id: Ieb11ae6698f370a1d601459e6253fc7dbe146014
This adds a reference to the View method that controls framework force
dark, which impacts how FORCE_DARK_AUTO is interpreted. It also
documents that FORCE_DARK_AUTO is the default.
Test: documentation only change
Fixes: 128704481
Change-Id: I88e7996362da5b092b269c313e37ef9253b12872
If there are multiple APKs as part of the WebView implementation (due to
splits, shared libraries, or overlays) add them all to the asset path,
instead of just the base APK.
Bug: 124116212
Test: manually tested with modified Trichrome build
Change-Id: I912e741b6de3154f66e5e82c9cf7154c6d39b1a4
This reverts the WebViewFactory changes from commit
527fa8f4de which appear to have changed
the way that the asset path is updated in an incompatible way.
Bug: 124116212
Fixes: 128489090
Test: launch affected applications
Change-Id: Id53dedef4fa55a6b8cea8a60326a4b8acd3422f8
Add all splits and shared libraries used by the WebView implementation
to the application's asset path, instead of just the base APK. Fix the
asset manipulation code path in WebViewFactory to reuse the modern
implementation in WebViewDelegate (using appendLibAssetForMainAssetPath)
instead of relying on addAssetPathAsSharedLibrary.
Add a new variant of appendLibAssetForMainAssetPath which accepts
multiple paths at once, to make it more efficient.
Bug: 124116212
Test: atest CtsWebkitTestCases
Change-Id: Ie55525d78089a4595c5b0b126e4ff1530303afe8
Per API council request, remove setForceDarkMode -> setForceDark to
align with other framework naming.
Bug: 120599879
Test: None
Change-Id: I3f4ed00303d7e4b7d0651a384324a8b265663290
