aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
21 hoursMerge "Fix the res ID of headline font family." into mainHEADmastermainRam Peri
22 hoursFix the res ID of headline font family.Ang Li
Bug: 343516276 Test: (Tested in git-main) atest -c SystemUIGoogleRNG(2)Tests Change-Id: I54ef8b70436780daf2dda028b13da321584de447
6 daysMerge "Use reflection to operate on BackdropFrameRenderer" into mainTreehugger Robot
6 daysUse reflection to operate on BackdropFrameRendererGoogler
Android frameworks decide to remove this class in V. PiperOrigin-RevId: 636947308 Change-Id: I71cca7753d5e8926c5068ce226ffb8449d34edc4
7 daysMerge "Align several shadows with upstream." into mainTreehugger Robot
7 daysAlign several shadows with upstream.Ang Li
Bug: 341163842 Test: (Tested in git-main) atest -c CtesqueRoboTests atest -c SystemUIGoogleRobo(2)RNGTests Ignore-AOSP-First: The purpose of this CL is to avoid the merge conflict caused from a downstream merge. Merged-In: Iebe80a1a0e9378dcd405c73fa954bc74df13574b Change-Id: I100cf32f73cd96307b1140f0f9740da3a8a175fd
13 daysMerge "Fix the res ID of headline font family." into mainTreehugger Robot
13 daysFix the res ID of headline font family.Ang Li
Bug: 341163842 Test: (Tested in git-main) atest -c SystemUIGoogleRNG(2)Tests Change-Id: Ie99b23e44c3e3edc6c6d66f64dd545ff1d46cec1
13 daysMerge "Revert "Add shadows for two inner classes for natives."" into mainRam Peri
13 daysRevert "Add shadows for two inner classes for natives."Ram Peri
This reverts commit ea6173f2aae78b04c003d3d2f48a4b2fbb2326cc. Reason for revert: V version conflict. Change-Id: Ib204f709b387cd465f7c2e91cc61c3b420de5591
13 daysMerge "Add shadows for two inner classes for natives." into mainRam Peri
13 daysAdd shadows for two inner classes for natives.Ang Li
Due to framework side support of Color and ColorSpace in Ravenwood, we need to add shadows for `Matrix.ExtraNatives` and `ColorSpace.Rgb.Native` to let these native functions be visible to Robolectric. Bug: 341163842 Test: (Tested in git-main) atest -c CtesqueRoboTests atest -c SystemUIGoogleRobo(2)RNGTests Change-Id: I332b4e864e99e08ef1152f2fe818737d5dad23c0
2024-05-16Merge "Makes directory for common picker Robolectric testing files" into mainTreehugger Robot
2024-05-16Makes directory for common picker Robolectric testing filesChris Poultney
Bug: 327689068 Change-Id: Ief3f8aaa52d2bbd428803bf9a83da719254ff1d6 Flag: TEST_ONLY Test: Robolectric tests all pass
2024-05-16Merge "Include all systemui libs in visibility list" into mainTreehugger Robot
2024-05-15Merge "Add Launcher3 to visibility list" into mainTreehugger Robot
2024-05-15Include all systemui libs in visibility listPeter Kalauskas
Bug: 289353932 Bug: 236188940 Flag: NONE Test: m Change-Id: Icb98d4b98079d2c24adee6556be02568c47e7b9d
2024-05-15Add Launcher3 to visibility listUwais Ashraf
Test: atest Launcher3RoboTests Bug: 316551909 Flag: NA Merged-In: If4f2a20f6af95dc6b83d70852077691aec765296 Change-Id: If4f2a20f6af95dc6b83d70852077691aec765296
2024-05-15Merge "Prevent annotation processing/build breakages when AndroidVersions ↵Rex Hoffman
detects mismatches." into main
2024-05-15Prevent annotation processing/build breakages when AndroidVersions detects ↵Rex Hoffman
mismatches. Bug: 340843961 Test: local build with modified AndroidVersions files, new tests in cl. Change-Id: I51657959be1651df73f970b8146d8f50632d2d9e
2024-05-14Merge "Turn on robolectric compile time validtor." into mainTreehugger Robot
2024-05-14Turn on robolectric compile time validtor.Rex Hoffman
Prevents sdk checks from running on released versions of the sdk when validateCompileSdk=true Bug: 333045648 Test: mma, with 'lunch aosp_cf_x86_64_phone-ap1a-eng', failed before edit to SdkStore, succeeded after. Change-Id: I97f719bef24021e84d4d15ba67e435c1cfc7e608
2024-05-14Merge "Revert "Turn on robolectric compile time validtor"" into mainSam Chiu
2024-05-14Revert "Turn on robolectric compile time validtor"Sam Chiu
This reverts commit e3f9546ff1bf10dadab49d71065332b9f89d0462. Reason for revert: To verify broken build. b/340367293 Change-Id: Iabbbe04a24d5c8dc1a03d682c5ae8318fe98a910
2024-05-14Merge "Turn on robolectric compile time validtor" into mainTreehugger Robot
2024-05-13Turn on robolectric compile time validtorRex Hoffman
Bug: 333045648 Test: mma Change-Id: Iae6fbbf1b4a13e900c8b3efa19ce96e26ba8eeb0
2024-05-09Merge "Fix the font configuration resource ID." into mainTreehugger Robot
2024-05-09Fix the font configuration resource ID.Ang Li
Bug: 332407470 Test: (in git-main) atest -c SystemUIGoogleRobo(2)RNGTests Change-Id: I1bb8c6ac6b7cac4e272f59041a82c273147c077d
2024-05-08Merge "Merge branch 'upstream-google' into some_merge" into mainAng Li
2024-05-08Merge branch 'upstream-google' into some_mergeAng Li
Bug: 339459903 Test: mma Merged-In: I7ee2de434aac767aa009bb0367580618690cb786 Change-Id: I7ee2de434aac767aa009bb0367580618690cb786
2024-05-08Fix for ShadowSpeechRecognizer on Android VGoogler
With Android V, the internal implementation for SpeechRecognizer has changed. Prepare ShadowSpeechRecognizer to be extended in the future to support V's changed implementations. PiperOrigin-RevId: 631880010
2024-05-08Handle the `NullPointerException` when looking up resource value of a ↵Googler
`DynamicRefTable` when the data type is invalid. When the data type is invalid, it would make sense to return the corresponding bad code instead of throw an exception. PiperOrigin-RevId: 631864822
2024-05-08Allow the configs parameter for eglChooseConfig to be null in shadowSamuel Freilich
Most of the time, the output parameters for EGL methods are expected to be non-null. But the `configs` parameter of `eglChooseConfig` is one of the exceptions. When `configs` is not null, `eglChooseConfig` fills that with the first `configSize` configs and fills `numConfig` with the number of returned configs (no more than `configsSize`). But when `configs` is_ `null` it fills `numConfigs` with the number of _matching_ configs, ignoring `configsSize`: https://registry.khronos.org/EGL/sdk/docs/man/html/eglChooseConfig.xhtml The Android code wrapping this API checks that `attrib_list` and `num_configs` output arrays are not null, but allows `configs` to be null (see frameworks/base/core/jni/android_opengl_EGL14.cpp). PiperOrigin-RevId: 631820937
2024-05-07Bump nativeruntime-dist-compat to 1.0.10Michael Hoisie
This fixes inflating widgets like ProgressBar when native graphics is used in certain locales, and fixes HW rendering on Mac. See: https://android.googlesource.com/platform/frameworks/base/+log/refs/heads/and roid12-hostruntime-dev for relevant changes. Fixes #8478 Fixes #9044 PiperOrigin-RevId: 631527808
2024-05-07Rewrites ShadowCaptioningManager so it uses more real code.Julia Sullivan
Before, this would fake out some, but not all of the functionality of ShadowCaptioningManager. This lead to a disjointed Shadow. This now has users update the Settings.Secure as they would on real android and only shadows functions that call into AccessibilityManager since that calls into a service. PiperOrigin-RevId: 631491210
2024-05-07Add integration test for Kotlin Flow + Shadow Bluetooth APIsMichael Hoisie
This came up in a discussion about controlling coroutines in Robolectric tests. PiperOrigin-RevId: 631456525
2024-05-07Reuse HardwareRenderer objects when taking HW screenshotsMichael Hoisie
Previously, new HardwareRenderer objects were created each time a HW screenshot was captured. The problem was that if they were GC'd, `HardwareRenderer.nDeleteProxy` was called, which also freed a lot of associated native objects. This caused problems if the same Views were HW rendered again, as the native objects were corrupted. PiperOrigin-RevId: 631452676
2024-05-07Update ShadowInCallService.setPhone to use @Constructor correct.Julia Sullivan
Before, this would use inCallService as the realObject, but instead it should be just using the reflector with no real object added to it. PiperOrigin-RevId: 631442992
2024-05-07Add shadow method for TelephonyManager.rebootModemGoogler
Also add a public shadow API to keep track of the reboot count. PiperOrigin-RevId: 631439597
2024-05-06Adds support for new AdvertisingSet constructor in Android V.Julia Sullivan
In V, an IBluetoothGatt parameter was added as the first argument. Overall this remains functionally the same. PiperOrigin-RevId: 631191573
2024-05-01Merge "Change ID to fix the font related test failures." into mainRam Peri
2024-05-01Change ID to fix the font related test failures.Ang Li
Test: atest -c SystemUIGoogleRobo(2)RNGTests (http://ab/I25700010269749464, http://ab/I09700010269774260) Change-Id: I8c0e131317af25240b3983ebe8b22891c656ff40
2024-04-30Merge "Merge branch 'upstream-google' into merge_upstream" into mainAng Li
2024-04-30Merge branch 'upstream-google' into merge_upstreamAng Li
Bug: 330075029 Test: mma Change-Id: I12d2ce1a7889d3b6e20496639391e205eeddff06
2024-04-30Fix edge-to-edge enforcement check in ShadowCompatibilityMichael Hoisie
Apps that target Android V and above will be required to support edge-to-edge content, where apps will be responsible for drawing content under the status and navigation bars: https://developer.android.com/about/versions/15/behavior-changes-15#edge-to-edge In framework code, this is primarily detected using CompatChanges.isChangeEnabled(ENFORCE_EDGE_TO_EDGE). However, this previously always returned true because CompatChanges.isChangeEnabled returns true by default in Robolectric. In real Android, CompatChanges.isChangeEnabled(ENFORCE_EDGE_TO_EDGE) only returns true when the target SDK is V or above. PiperOrigin-RevId: 629401576
2024-04-29Update ShadowNativeSurface.nativeCreateFromSurfaceTexture for Android VMichael Hoisie
Previously, ShadowNativeSurface.nativeCreateFromSurfaceTexture invoke the JNI binding in U and below, but it would return zero in native code, due to SurfaceTexture not being available on host. This meant that doing the JNI callback was moot. In Android V, the native code of Surface.nativeCreateFromSurfaceTexture is not compiled for host platforms, so '0' has to be returned from the shadow method. To make these consistent, update ShadowNativeSurface.nativeCreateFromSurfaceTexture to return zero for all SDK levels. PiperOrigin-RevId: 629295331
2024-04-26Adjust codec output format update in start to alter pendingOutputFormatMichael Katz
`ShadowMediaCodec.native_start`, if async, simulates a codec updating the output format. The update overwrote the current pendingOutputFormat to the application. The fix is to make those updates to the pendingOutputFormat and send that in the `CB_OUTPUT_FORMAT_CHANGE` event. PiperOrigin-RevId: 628338243
2024-04-26Fix HW rendering in Android VMichael Hoisie
ShadowNativeBaseRecordingCanvas was missing the `callNativeMethodsByDefault` annotation param, which meant that all the native methods were no-ops in Android V. Add some additional tests to capture this issue. PiperOrigin-RevId: 628318882
2024-04-25More unsupported SDK code cleanup.Brett Chabot
Perform more cleanup related to the move to a minimum supported SDK of lollipop (21) missed in last commit (c577257599dc30cf283f8961c286d4ef65e799cb): - Remove methods with a maxSdk of JELLY_BEAN* - Remove RuntimeEnvironment.castNativePtr and associated Number to long conversions - Set default manifest minSdkVersion and targetSdkVersion to 21 - Remove if checks for apiLevel that are now redundant - Remove redundant Config(minSdk = LOLLIPOP) statements PiperOrigin-RevId: 628203572
2024-04-24Drop support for KITKAT SDK and associated dead code cleanup.Brett Chabot
The android ecosystem is moving to support a minimum of minSdkVersion L, and thus Robolectric is following suit and dropping support for running on SDKs < L (21). This change does the following: - removes SDK 19/Kitkat from the list of default SDKs. Tests that attempt to run explicitly on KitKat (e.g Config(sdk = 19) will now fail with 'unsupported sdk' errors. - Remove now redundant methods and logic that was only relevant to SDKs < L, such as removing @Implementation or unit tests with a maxSdk of < L. - Remove redundant minSdk = LOLLIPOP declarations from @Implementation and @Config references. PiperOrigin-RevId: 627785965