Enable cobalt browser tests for media* tests - #1
Open
linxinan-yt wants to merge 48 commits into
Open
Conversation
🤖 Gemini Suggested Commit Message💡 Pro Tips for a Better Commit Message:
|
Bug: 447659892
Unit test jobs in CI occasionally time out, even when tests are configured to retry. The existing job timeout of 2100 seconds does not provide sufficient time for a second attempt to complete after an initial test timeout of 1800 seconds. Increase the default job timeout to 2700 seconds to give a subsequent test attempts a fighting chance. Issue: 470590965
The android.text.TextUtils class was being used in CobaltActivity without a corresponding import statement. This change adds the necessary import to ensure proper compilation and functionality. It was removed in youtube#8504 but used in youtube#8485, as PRs were merged out of order, resulting in missing this import. Issue: 471083158 Issue: 418842688
Patch by @abhijeetk and myself. Now that the tvOS app entry point is no longer in Starboard, make Cobalt start the Starboard machinery from main.mm together with the `//content` machinery. Doing so requires a few changes and cleanups: * Make: the `QueueApplication` dependency internal to `application_darwin.mm` so that its header can be used without any Starboard implementation-specific includes. * Make `starboard_group`'s dependency on `starboard` (and consequently `starboard_platform`) public just like it already is on Android. Both Android and tvOS need to reference platform-specific files from Cobalt, so they need to be able to depend on "starboard_platform" even if indirectly. Bug: 432823234 Bug: 456692793 --------- Co-authored-by: fayev-g <fayev@google.com>
Introduce NullMarked and related nullability annotations to the Cobalt Android shell classes. This change aligns the null-safety practices with Chromium upstream's Content shell, improving static analysis and preventing potential null pointer exceptions. It involves adding `Nullable` to fields and parameters, and utilizing assertNonNull and assumeNonNull for explicit null checks. Issue: 471076830 Issue: 418842688 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6472891
Currently, our nl_langinfo test data is mostly based off of glibc specifications. Most of the glibc data in regards to locales has not been updated since the early 2000s due to the preservation of legacy systems. This makes the data incorrect when comparing them to data received from the more recently updated ICU library. This PR changes the test data to better reflect the data received from our ICU library. **This change is currently no-op** as these tests are not currently ran. Follow-up PRs will use this test data to test our implementations. Bug: 466160361
youtube#8506) These changes ensure WidevineTimer follows the same portable pthread patterns established in commits 9d29311 and 42a3596, and correctly handles platform-specific pthread_setname_np() semantics on Apple platforms. Bug: 432507888 Bug: 438955200 Bug: 471365322
Introduce a mechanism to expose Cobalt-specific base::feature flags to the Android Java codebase. This involves new build targets, JNI interfaces, and Java classes to query feature status from C++. This code loosely follows this Chromium guide as an example. For ease of use, we have made it so that all STARBOARD_FEATURES are now exposed to our Java code. This makes adding a new feature for Cobalt Java code simple, as a developer only needs to define a STARBOARD_FEATURE for the java feature they want to add. A new method was added to Starboard's FeatureList class (FeatureList::IsEnabledByName)to be able to retrieve an SbFeature by name. To get the Python script java_cpp_features.py to work with our STARBOARD_FEATURE macro, this PR: - Edits the Python script to support STARBOARD_FEATURES - Places the starboard/extension/feature_config.h file inside a GN source_set, allowing the script to read feature_config.h This PR is extremely similar to youtube#8377, but was made into its own PR due to JNI differences between m114 and m138. Bug: 447651764
This should only run on `.github` files and statically check to see if github actions files for syntax and common mistakes. Fixed: 473609586
Refactor command-line argument logging to improve testability and clarity. - Extracted the command-line formatting logic into a new helper function 'FormatCommandLineSwitchesForDebugging' in 'cobalt/browser/command_line_debugging_test_support.cc'. - Updated 'LogCommandLineForDebugging' to use this new formatting function. - Added a dedicated unit test 'cobalt_command_line_debugging_unittest.cc' to verify the correctness of the command-line formatting, ensuring order-independent assertions. - Updated 'BUILD.gn' files to correctly integrate the new test and its dependencies. Test: Manually verified flag output to console. Test: `out/android-x86_devel/bin/run_cobalt_unittests --gtest_filter=CommandLineDebuggingTest.* ` Bug: 473056330
…lity list (youtube#8531) This fixes `gn gen` after youtube#8283. These targets are generated automatically by other GN actions and need to be allowed to depend on `for_content_tests`. Bug: 456692793 Bug: 463395742
…build (youtube#8532) PR youtube#8442 started calling `SbSystemGetErrorString()` in shared code, so an implementation is needed for tvOS as well. Bug: 469772944 Bug: 456692793
…ng-format when the naming is not the same (youtube#8295) Bug: 466129151 Follow up from youtube#8241, where we update clang-format to the m138 version. Google C++ style guide and clang-format enforce that an impl file's "relevant header" is at the top, but the cobalt codebase contains impl names that don't exactly match the header name and so clang-format moves those headers in the includes order. After offline discussion, exclude those headers from clang-format for now. Gemini recommends that we consolidate impls or split headers to satisfy formatting and style rules.
…ube#8454) Refactors the navigation browser tests and related utilities. This includes updating dependencies, modifying test logic to align with new APIs, and merging the duplicate test switches. This is one of several changes to navigation_browsertests as part of the effort to enable browser tests on the main branch for Cobalt. Note: This is branched on top of youtube#8452 as part of a chain. Bug: 459831767
…youtube#8455) This change enables the navigation_browsertest and addresses several issues preventing cobalt_browsertests from building and running correctly. Key updates include integrating StarboardBridge into the Android test application, unifying TestShell with Shell in browser test utilities, and streamlining the Android manifest for test builds. These changes are part of the ongoing effort to ensure browser tests function reliably on the main branch. Note: This is branched on top of youtube#8454 as part of a chain. Bug: 459831767
Bug: 474079809 (cherry picked from commit ae7c6b8)
Integrate the cobalt_browsertests target into the Android test suite for both android-arm and android-arm64 architectures. This enables browser tests to be executed on Android devices, significantly broadening test coverage for Cobalt on the Android platform. It includes a new Python script to simulate test execution and cleanup in Cobalt's CI, improving local integration testing efforts. Although initially all tests are filtered as failing, this lays the necessary groundwork for future stabilization and enabling. This is one of several changes to navigation_browsertests as part of the effort to enable browser tests on the main branch for Cobalt. Note: This is branched on top of youtube#8455 as part of a chain. Also note: This is essentially a cherry-pick of youtube#8328, which merged these same changes to the 26.android branch. Test: vpython3 cobalt/testing/browser_tests/integration/run_android_tests_test.py Test: python3 cobalt/testing/browser_tests/integration/run_android_tests.py --apk-path out/android-x86_qa/cobalt_browsertests_apk/cobalt_browsertests-debug.apk --build-dir out/android-x86_qa --test-list test_list.txt --android-serial localhost:27359 Bug: 459831767
…#8557) Extract fake_graphics_context_provider into its own static_library target under starboard/testing. This encapsulates the build logic for this testing utility. Previously, its source files were included directly in multiple test targets. Moving it to a dedicated target improves build system modularity and reusability, allowing consumers to simply depend on the library. Bug: 468356861
Bug: 474079809 Cherry pick of youtube#8545.
…espected (youtube#8474) After running tests in PR#8455, the only test failure was `NavigationBrowserTest.DeprecateUnloadOptOutFlagRespected`. This disables that test to make all currently enabled browsertests green on the main branch. This is one of several changes to navigation_browsertests as part of the effort to enable browser tests on the main branch for Cobalt. Note: This is branched on top of youtube#8455 as part of a chain. Test: `out/android-x86_qa/bin/run_cobalt_browsertests -vvvvv -f *NavigationBrowserTest*` Bug: 459831767
This change modifies the GitHub Actions workflows for internal and on-device tests. It adds the current Git branch name as a label to the test invocation command for Sponge. Bug: 474168791
…utube#8319) This is part of the required changes to make SBPlayer-rendered videos show on the screen together with the rest of the web contents. The idea here is to make hole punching to work by having the videos handled by SbPlayer be shown as underlays: the UIViews with videos are added before the web contents view and are thus supposed to be shown "under" it (this is done in youtube#8350). On the Viz side, this is done by treating video hole quads (which are already added by Starboard media code in `//media`) specially: when found, these quads disable the default quads-to-CALayers code path that normally renders each quad as a separate CALayer (with other CALayers underneath, managed and described by `CARendererLayerTree` and [1]). [1] https://docs.google.com/document/d/1DtSN9zzvCF44_FQPM7ie01UxGHagQ66zfF5L9HnigQY/edit?usp=sharing By doing so: - `CALayerOverlayProcessor::PutForcedOverlayContentIntoUnderlays()` and `CALayerOverlayProcessor::PutQuadInSeparateOverlay()` are always called, and we use `PutQuadInSeparateOverlay()` to replace the video hole quad with a transparent quad just like the existing code in `OverlayStrategyUnderlayStarboard`. - The entire page is received by `CARendererLayerTree` as a single CALayer (like in C25), which includes the transparent quad created above, effectively making the video UIViews poke through the web contents. Bug: 459366150
Reverts youtube#8457 Reasoning: On the infra side, it looks like the client isn't correctly identifying that the entire suite is filtered. Bug: 459831767
Enable cobalt_browsertests on Linux with starboard. 1. add cobalt_browser_tests_main.cc as the entrance for browser tests on starboard 2. remove ShellTestPlatformDelegate as it is not used in cobalt_browsertests 3. tune the expectations to adjust cobalt's single process mode in ContentMainRunnerImplBrowserTest as a demo Test: ``` out/linux-x64x11_devel/cobalt_browsertests --single-process-tests --gtest_filter=*ContentMainRunnerImplBrowserTest.* --no-sandbox --single-process --no-zygote --ozone-platform=starboard ... [ OK ] ContentMainRunnerImplBrowserTest.StartupSequence (50 ms) [----------] 1 test from ContentMainRunnerImplBrowserTest (50 ms total) [----------] Global test environment tear-down [==========] 1 test from 1 test suite ran. (50 ms total) [ PASSED ] 1 test. ``` Issue: 433354983
If an `activeNetworkCheck` is intentionally interrupted or canceled (ie. `managementFuture.cancel(true)`), it throws an `InterruptedException` or `CancellationException` and briefly raises a network error dialog. This PR catches those exceptions to not raise a network error dialog as an interruption indicates that another `activeNetworkCheck` was started, not that there is not a valid internet signal. A `TimeoutException` and any other exception type will still call `handleFailure()` and raise a network error dialog. Bug: 473635416
linxinan-yt
pushed a commit
that referenced
this pull request
Jul 8, 2026
…:ThreadPool (youtube#11184) CobaltPerformance memory measurement APIs (MeasureAvailableCpuMemory, MeasureUsedCpuMemory, MeasureUsedSwapMemory, MeasureReservedVirtualMemory) were executing synchronous system metrics queries and file reads on the Renderer thread sequence, causing UI thread stalls of up to 2.37 ms per invocation (11.04 ms total during trace analysis). This change wraps the system queries using base::ThreadPool::PostTaskAndReplyWithResult, ensuring system file reads run on a background thread pool without blocking the main Renderer message loop. ### Quantitative Performance Metrics (Optimization #1) 1. Invocation Frequency & Call Count: • MeasureUsedSwapMemory : Called 3 times during the trace period. • MeasureUsedCpuMemory : Called 2 times. • MeasureAvailableCpuMemory : Called 1 time. • MeasureReservedVirtualMemory : Called 1 time. • Total synchronous calls on Renderer thread sequence: 7 calls. 2. Main Thread Blocking Duration (CPU / IO Stall Savings): • MeasureUsedSwapMemory : 7.11 ms total (avg 2.37 ms per call). • MeasureUsedCpuMemory : 1.91 ms total (avg 0.96 ms per call). • MeasureAvailableCpuMemory : 1.24 ms total (1.24 ms per call). • MeasureReservedVirtualMemory : 0.78 ms total (0.78 ms per call). • Total direct Renderer thread blocking time saved: 11.04 ms (over 7 invocations). 3. Impact Summary for CL Description: • Before Change: Every call to window.performance memory metrics / H5VCC settings triggered synchronous system calls ( /proc file reads & system stats) on the Renderer message loop, blocking JavaScript execution by up to 2.37 ms per invocation (total 11.04 ms stall during the trace). • After Change: Memory metrics are queried off-thread via base::ThreadPool , reducing Chrome_InProcRendererThread blocking time to 0.00 ms for memory measurement API calls. Bug: 530222962
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
13 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Enable cobalt browser tests for media* tests
Issue: 433354983