Add socket watcher closure if no pending operations #4115
Closed
Conversation
This file contains 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
These two tests failed due to the PR (youtube#3374), and they should be re-enabled once it fixes. b/346868673
We were using a method that caused destruction of the backend image when an ownership transfer is supposed to happen. This causes crashes during conceal events when the images are purged. b/346774396 b/346643423 b/341554972
Bug: In url_request_context.cc, we reset proxy_resolution_service_. However, in http_network_session, we store the original proxy_resolution_service_. From url_request_context.cc, if we delete the orignal proxy_resolution_service and create a new object, http_network_session may still try to access the orignal object. This will trigger the heap-use-after-free error. b/346555722
This PR (youtube#3325) changes SbPlayerGetMaximumNumberOfSamplesPerWrite() to 256 on ATV as Cobalt supports multiple samples per write to SbPlayer (b/227837774), but current testing files for SbPlayerWriteSampleTest.WriteSingleBatch tests are not long enough for the modification. Disable the tests and should be re-enabled once it fixes. b/347097484
The main change in this PR is to update Evergreen's default behavior so that it requests packages containing compressed Cobalt Core binaries. This is to reflect the change in the 2025 Software requirements. Note that this change will also impact devices that were certified before Cobalt 25, after they've completed an update to Cobalt 25, although there is an exception for devices using the Memory Mapped file feature. The --use_compressed_updates flag is replaced with a --use_uncompressed_upddates flag, although it should not be used in production. A few related changes to documentation and tests are also included to reflect that compressed binaries are now the default: not only for updates but also as the system image. b/341064002 Change-Id: I79c052e580cbb28f638a9d09cf7445c0fce8c3a6
Leftover entries not used anymore. b/333270679
…ek/reset (youtube#3468) On certain devices, the |resampler_| in AdaptiveAudioDecoder is utilized to resample the decoded audio output to align with the audio sink's operating frequency. Therefore, after invoking AdaptiveAudioDecoder::Reset(), Cobalt must verify that all audio outputs function correctly. 1. Check media time after Seek() (which invokes AdaptiveAudioDecoder::Reset()). 2. Add SB_DCHECK() to ensure that |resampler_| should not be NULL when the sample rates of audio sink and decoded audio are different. b/345229979 Test-On-Device: true
- Convert URLs to Identifiers - Add Name field everywhere and sync to Chromium naming where applicable - Populate closest_version where it's readily available - Always provide "version" in Git identifier even if it's unknown b/268355217 b/150775273
Changes fsync() to fdatasync(), that does not force OS-wide disk write stalls writing file metadata. This follows upstream code in base/files/file_posix.cc that is the origin of Starboard version. Both Linux and Android ports use fdatasync() Also see: https://crbug.com/40412313 b/231759845 b/323589608
…youtube#3568) Rename AudioBatchedSampleWrite to MaxAudioSamplesPerWrite followed by PR (youtube#3325). b/227837774
- Enforce `name` field to be present - Require `Git` identifier to be provided, with value and version b/268355217 b/150775273
Addressed previously in 09f8bb7 b/325026819
This PR (youtube#3325) changes SbPlayerGetMaximumNumberOfSamplesPerWrite() to 256 on ATV as Cobalt supports multiple samples per write to SbPlayer (b/227837774), but current testing files for SbPlayerWriteSampleTest.WriteSingleBatch tests are not long enough for the modification. 1. Set upper bounds for WriteSingleBatch tests (15 audio samples per write, 60 video samples per write) when SbPlayerGetMaximumNumberOfSamplesPerWrite() > 1. 2. We should revisit the upper bounds when the requirement is changed (b/347728473). b/347097484 Test-On-Device: true
b/302730696 Change-Id: Icdce23666554e47789a9109ea861f5533ced2bf2
git-subtree-dir: third_party/libvpx git-subtree-split: 1759c6ae9316996b9f150c0ce9d0ca78a3d15c02
Currently allocations are done at encoder creation time. Going from threaded to non-threaded would cause a crash. Bug: chromium:1486441 Change-Id: Ie301c2a70847dff2f0daae408fbef1e4d42e73d4 git-subtree-dir: third_party/libvpx/source/libvpx git-subtree-split: 7aaffe2df4c9426ab204a272ca5ca52286ca86d4
Bonus: Don't install commit-msg hook in external repo b/150775273 --------- Co-authored-by: Kaido Kert <[email protected]>
…#3599) Reverts youtube#2773 May have broken several on-device tests and internal builds b/302730696
b/299639708 Change-Id: Ida75ca986b88ae444684321ae609e1aae5588bf5 --------- Co-authored-by: Kaido Kert <[email protected]>
Issue pointed out here https://cobalt-review.git.corp.google.com/c/cobalt/+/6370 b/348226910 Co-authored-by: Devin Cai <[email protected]>
b/302715109
Issue: 201789468 Issue: 305733809 Reviewed-on: youtube#3602
For old compiler configs, don't use Angle, as system GLES2 interface works fine. Up to date Angle code requires much more modern C++ compilers. b/323209852
Issue: 268355217 Reviewed-on: youtube#3490
b/348014682 Change-Id: Id5039cba694ef992f28f9a7efe4192cd45c76803
The clock ids and the errno error codes are not standard between platforms. b/302335657 Test-On-Device: true Change-Id: I5c5de77703c2b8b8281ceef9d14b1902958ab9bd
Adding additional Media Telemetry to cover SbPlayer DRM functions. The rough range of each function: 1. SbDrmCreateSystem(): 26.307ms to 94.221ms b/303328238
Adding additional Media Telemetry to cover SbPlayer functions. The rough range of each function: 1. SbPlayerGetPreferredOutputMode(): 1.5us to 13us 2. SbPlayerSeek(): 6us to 318us 3. SbPlayerWriteEndOfStream()/audio: 13us to 6.3ms 4. SbPlayerWriteEndOfStream()/video: 3us to 2.8ms 5. SbPlayerSetBounds(): 133us to 26.95ms 6. SbPlayerSetPlaybackRate(): 6us to 11.17ms 7. SbPlayerSetVolume(): 7us to 17.89ms 8. SbPlayerGetInfo(): 1us to 6.8ms 9. SbPlayerGetAudioConfiguration(): 2us to 22.7ms 10. SbPlayerGetCurrentFrame(): 5us to 61.7ms b/303328238
b/302715109
b/150410605
Those codecs are not required by Technical Requirements, disabling by default. Making it configurable can be done as follow-up. b/362356408
b/358385783 Change-Id: I0a9dd1b45d692d71dc681a196f349bcf53ab72fa
Specifically the navigation yaml that links to SB modules. b/328290929
This is no longer necessary b/364421874
Removes reverse dependency on Cobalt from test targets b/364421874
Removes reverse dependency from Starboard to Cobalt b/364421874
Removes reverse dependency from Starboard to Cobalt b/364421874
This removes a reverse dependency from Starboard to Cobalt. b/364421874 b/278599663
Enhanced the `SunnyDay` test in `PosixSocketResolveTest` to include checks for the `ai_addrlen` field. The test now verifies that `ai_addrlen` matches the expected size for `sockaddr_in` (IPv4) and sockaddr_in6 (IPv6) based on the `ai_family` value. This ensures that `getaddrinfo` correctly sets the length of the returned socket address structures. b/349469318 Signed-off-by: Mykola Solianko <[email protected]>
b/360216121 Removes windows based builds and SB15, SB16 code --------- Co-authored-by: Madhura Jayaraman <[email protected]> Co-authored-by: Hao <[email protected]> Co-authored-by: Jelle Foks <[email protected]> Co-authored-by: Oscar Vestlie <[email protected]> Co-authored-by: Yijia Zhang <[email protected]> Co-authored-by: Kaido Kert <[email protected]> Co-authored-by: Colin Liang <[email protected]> Co-authored-by: Colin Liang <[email protected]> Co-authored-by: xiaomings <[email protected]>
In preparation for cross-thread MediaSource objects, a small amount of state and associated setters are being added to the MediaSourceAttachment interface. This will allow for a push model, instead of a pull model, for state such as recent media timestamps, and the error status of the HTMLMediaElement. Usage of these setters is gated through the `"MediaElement.EnableUsingMediaSourceAttachmentMethods"` H5VCC flag. This flag needs to be coordinated between a `HTMLMediaElement`, the attached `MediaSource`, and all `SourceBuffers` used. As such, the value of the flag when the `HTMLMediaElement` was created is used to coordinate the state of the rest of the related objects. This prevents a flag mismatch, such as if the H5VCC value is updated after the `HTMLMediaElement` has been created but before the `MediaSource` is created. This is based off of the following Chromium changes: * https://chromium-review.googlesource.com/c/chromium/src/+/2391934 * https://chromium-review.googlesource.com/c/chromium/src/+/2401808 b/338425449
This was used for standar pointer migration, and has no references left b/366061817
MSoliankoLuxoft
approved these changes
Sep 17, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
b/348241402