[Windows] Use an unsized texture format for backing stores on OpenGL ES 2 by krll-kov · Pull Request #192917 · flutter/flutter

krll-kov · 2026-09-17T06:22:58Z

On GPUs limited to D3D11 feature level 10_0, ANGLE only creates an OpenGL ES 2.0 context. CompositorOpenGL::CreateBackingStore passes the sized format (GL_BGRA8_EXT or GL_RGBA8) as the internalformat of glTexImage2D. ES 2.0 requires internalformat to equal format, so the call fails with GL_INVALID_VALUE, the backing store framebuffer is incomplete and the window stays black with no error logged. Flutter 3.44.2 passed the unsized format there, the sized one came in with the MSAA backing store changes.

This change picks the internal format once in Initialize: the unsized general format on an ES 2 context, the sized format otherwise. The three TexImage2D calls use it. The sized format is still reported to the engine through FlutterOpenGLFramebuffer.target, and nothing changes on ES 3 contexts.

Tests:

CreateBackingStoreGLES2Unsized: ES 2 with Skia, glTexImage2D gets GL_BGRA_EXT, the engine still gets GL_BGRA8_EXT.
CreateBackingStoreImpellerGLES2Unsized: ES 2 with Impeller, GL_BGRA_EXT.
CreateBackingStoreGLES3Sized: ES 3 keeps GL_RGBA8.

The two ES 2 tests fail without the change.

Manually verified with a dxcpl Feature level limit of 10_0, where the unpatched engine shows a black window: on 3.47.4 x64 a stock app renders with Skia and Impeller, and so does an app with gradients, shadows, a multiply blend and animation. An animated app runs at 120 FPS at 10_0 with both renderers, the same as the official engine without the limit. With Impeller, content drawn before a BackdropFilter is still lost at 10_0; that is a separate Impeller bug. On main (arm64) the stock app renders with an ES 2 context, a complete framebuffer and no GL errors. Also verified in a production app on 3.47.4.

Fixes #191978

Pre-launch Checklist

…ES 2

gemini-code-assist

Code Review

This pull request updates the Windows OpenGL compositor to support OpenGL ES 2.0 by introducing an internal_format field that dynamically selects between sized and unsized formats based on the detected GLES version. Unit tests are also added to verify this behavior under GLES2 and GLES3 contexts. Feedback suggests adding a defensive null check for the GLES description to prevent potential crashes, and refactoring a static std::vector in the unit tests to comply with the Google C++ Style Guide's rules on trivial destructibility.

gemini-code-assist · 2026-09-17T06:24:02Z

+  const impeller::DescriptionGLES* description = gl_->GetDescription();
+  bool is_gles2 =
+      description->IsES() && description->GetGlVersion().major_version < 3;
+  format_.internal_format =
+      is_gles2 ? format_.general_format : format_.sized_format;


The gl_->GetDescription() call may return nullptr if the OpenGL context or driver description is not available or failed to initialize. Dereferencing it directly without a null check can lead to a crash. Please add a defensive null check before accessing its members.

Suggested change

const impeller::DescriptionGLES* description = gl_->GetDescription();

bool is_gles2 =

description->IsES() && description->GetGlVersion().major_version < 3;

format_.internal_format =

is_gles2 ? format_.general_format : format_.sized_format;

const impeller::DescriptionGLES* description = gl_->GetDescription();

bool is_gles2 = description &&

description->IsES() &&

description->GetGlVersion().major_version < 3;

format_.internal_format =

is_gles2 ? format_.general_format : format_.sized_format;

GetDescription() cannot return null at this point, so I'd rather not add the check.

Initialize bails out when the proc table is invalid, a few lines above:

flutter/engine/src/flutter/shell/platform/windows/compositor_opengl.cc

Lines 270 to 274 in aced6c4

gl_ = std::make_unique<impeller::ProcTableGLES>(resolver_);

if (!gl_->IsValid()) {

gl_.reset();

return false;

}

gl_ = std::make_unique<impeller::ProcTableGLES>(resolver_); if (!gl_->IsValid()) { gl_.reset(); return false; }

ProcTableGLES creates its description first and only becomes valid at the very end of the constructor, after the description passed its own check:

flutter/engine/src/flutter/impeller/renderer/backend/gles/proc_table_gles.cc

Lines 118 to 122 in aced6c4

description_ = std::make_unique<DescriptionGLES>(*this);

if (!description_->IsValid()) {

return;

}

flutter/engine/src/flutter/impeller/renderer/backend/gles/proc_table_gles.cc

Line 192 in aced6c4

is_valid_ = true;

description_ = std::make_unique<DescriptionGLES>(*this); if (!description_->IsValid()) { return; } ... is_valid_ = true;

The existing BGRA check right above the new code already dereferences it the same way:

flutter/engine/src/flutter/shell/platform/windows/compositor_opengl.cc

Line 276 in aced6c4

if (gl_->GetDescription()->HasExtension("GL_EXT_texture_format_BGRA8888")) {

if (gl_->GetDescription()->HasExtension("GL_EXT_texture_format_BGRA8888")) {

gemini-code-assist · 2026-09-17T06:24:02Z

+std::vector<GLint>& TexImage2DInternalFormats() {
+  static std::vector<GLint> internal_formats;
+  return internal_formats;
+}


According to the Google C++ Style Guide, static variables must be trivially destructible. std::vector has a non-trivial destructor, which violates this rule. To comply with the style guide, allocate the vector on the heap and store a pointer to it.

Suggested change

std::vector<GLint>& TexImage2DInternalFormats() {

static std::vector<GLint> internal_formats;

return internal_formats;

}

std::vector<GLint>& TexImage2DInternalFormats() {

static std::vector<GLint>* internal_formats = new std::vector<GLint>();

return *internal_formats;

}

References

According to the Google C++ Style Guide, static variables must be trivially destructible. Class templates like std::vector have non-trivial destructors and are not allowed as static variables. ^(link)

This follows the pattern already used by the Windows embedder tests, where a captureless mock function records into a function-local or file-level static std::vector:

flutter/engine/src/flutter/shell/platform/windows/flutter_windows_view_unittests.cc

Line 61 in aced6c4

static std::vector<int> key_event_logs;

flutter/engine/src/flutter/shell/platform/windows/window_manager_unittests.cc

Line 1212 in aced6c4

static std::vector<UINT> received_messages;

static std::vector<int> key_event_logs; static std::vector<UINT> received_messages;

The mock has to be a plain function pointer for the ProcTableGLES resolver, so it cannot capture state, and the vector is cleared at the start of each test. I kept it consistent with the neighbouring tests

gaaclarke

Looks good to me, thanks for the quick PR. Switching context's and setting up the environment to develop this PR would have taken some time. I just have some cleanup notes.

gaaclarke · 2026-09-17T15:52:44Z

+
+TEST_F(CompositorOpenGLTest, CreateBackingStoreGLES3Sized) {
+  UseHeadlessEngine();
+  TexImage2DInternalFormats().clear();


Let's put this in TearDown() to avoid leaks between tests

gaaclarke · 2026-09-17T15:55:37Z

+  return internal_formats;
+}
+
+void RecordTexImage2D(GLenum,


We have a mechanism to statically assert that this signature is correct, it's probably easiest just to copy it into this file. We should unify these mocks at some point.

flutter/engine/src/flutter/impeller/renderer/backend/gles/test/mock_gles.cc

Line 81 in 9ead586

static_assert(CheckSameSignature<decltype(mockGetString), //

gaaclarke · 2026-09-17T15:56:09Z

 };

+// ANGLE returns these strings for an OpenGL ES 2.0 context
+const unsigned char* MockGetStringGLES2(GLenum name) {


(same comment about static_assert's for these functions)

krll-kov · 2026-09-17T17:01:58Z

Looks good to me, thanks for the quick PR. Switching context's and setting up the environment to develop this PR would have taken some time. I just have some cleanup notes.

done

gaaclarke

LGTM, thanks!

gaaclarke · 2026-09-17T17:11:16Z

@b-luk please be a second reviewer for this.

b-luk

LGTM

auto-submit · 2026-09-17T19:25:53Z

autosubmit label was removed for flutter/flutter/192917, because - The status or check suite Dashboard Checks has failed. Please fix the issues identified (or deflake) before re-applying this label.

gaaclarke · 2026-09-17T19:35:16Z

Some of the CI failures seem apropos:

https://ci.chromium.org/ui/p/flutter/builders/try/Linux_android_emu%20android_hardware_smoke_opengles_tests/4741/overview

+6 -1: should render and match platformViewTextureLayerTest golden [E]

  Bad state: Captured screenshot is blank/empty after 3 attempts.

  test_driver/driver_test.dart 104:9                 main.templateTest
  ===== asynchronous gap ===========================
  test_driver/driver_test.dart 144:5                 main.<fn>
  ===== asynchronous gap ===========================
  package:test_api/src/backend/declarer.dart 253:15  Declarer.test.<fn>.<fn>
  ===== asynchronous gap ===========================
  package:test_api/src/backend/declarer.dart 250:11  Declarer.test.<fn>
  ===== asynchronous gap ===========================
  package:test_api/src/backend/invoker.dart 318:9    Invoker._waitForOutstandingCallbacks.<fn>


I/PlatformViewsChannel( 5487): Using legacy platform view rendering strategy.
I/PlatformViewsController( 5487): Using hybrid composition for platform view: 1
Captured screenshot is blank/empty (attempt 1/3)
Captured screenshot is blank/empty (attempt 2/3)
Captured screenshot is blank/empty (attempt 3/3)

I'll hit rerun to make sure we aren't experiencing a flake. @krll-kov let me know if you want me to take a stab at diagnosing it deeper. There isn't much in the logs to go on.

krll-kov · 2026-09-17T19:46:22Z

, b> Some of the CI failures seem apropos:

https://ci.chromium.org/ui/p/flutter/builders/try/Linux_android_emu%20android_hardware_smoke_opengles_tests/4741/overview

+6 -1: should render and match platformViewTextureLayerTest golden [E]

  Bad state: Captured screenshot is blank/empty after 3 attempts.

  test_driver/driver_test.dart 104:9                 main.templateTest
  ===== asynchronous gap ===========================
  test_driver/driver_test.dart 144:5                 main.<fn>
  ===== asynchronous gap ===========================
  package:test_api/src/backend/declarer.dart 253:15  Declarer.test.<fn>.<fn>
  ===== asynchronous gap ===========================
  package:test_api/src/backend/declarer.dart 250:11  Declarer.test.<fn>
  ===== asynchronous gap ===========================
  package:test_api/src/backend/invoker.dart 318:9    Invoker._waitForOutstandingCallbacks.<fn>


I/PlatformViewsChannel( 5487): Using legacy platform view rendering strategy.
I/PlatformViewsController( 5487): Using hybrid composition for platform view: 1
Captured screenshot is blank/empty (attempt 1/3)
Captured screenshot is blank/empty (attempt 2/3)
Captured screenshot is blank/empty (attempt 3/3)

I'll hit rerun to make sure we aren't experiencing a flake. @krll-kov let me know if you want me to take a stab at diagnosing it deeper. There isn't much in the logs to go on.

Pr only touches windows files in shell/platform/windows, i'm not sure how that could affect android, but i'll try to execute that test locally

gaaclarke · 2026-09-17T19:59:06Z

Pr only touches windows files in shell/platform/windows, i'm not sure how that could affect android, but i'll try to execute that test locally

Ah yea, I saw OpenGLES and Android and thought we triggered something. You're right. I'll keep poking at CI to try to get this past.

krll-kov · 2026-09-17T20:03:51Z

Ah yea, I saw OpenGLES and Android and thought we triggered something. You're right. I'll keep poking at CI to try to get this past.

just checked that shell/platform/windows/BUILD.gn starts with assert(is_win), so none of the changed files are compiled into the Android embedder at all. Looks like a capture flake or something that is not related to this pr

eduardohr-muniz · 2026-10-02T15:32:04Z

Hi! Thanks for this fix, and for #193306.

Could both be considered for a cherry-pick to the 3.47 stable branch?

Impact: We ship a Windows point-of-sale app (MSIX, Microsoft Store) built with Flutter 3.47.1. Several of our customers run it on older hardware, and on those machines the window stays completely black. The app is running fine underneath: network calls succeed, navigation happens, logs are written. Nothing is drawn, so the app is unusable for these customers.

Affected machine:

CPU: Intel Core i5 650 (Clarkdale, 2010)
GPU: Intel HD Graphics (Ironlake), driver 8.15.10.2900 (2012), WDDM 1.1
Max Direct3D feature level: 10_0
OS: Windows 10 Pro 64-bit (build 19045)

Intel publishes no newer driver for this GPU on Windows 10, so the customer cannot fix it on their side.

Why both PRs: this machine is at feature level 10_0, so it needs this PR to get past the black window (#191978), and #193306 so that Impeller does not erase content drawn before a BackdropFilter (#192918).

Current state: I checked 3.47.6 (latest stable) and 3.49.0-0.2.pre (latest beta), and neither contains this PR. Our only option today is to pin our Windows builds to 3.44.x, which also holds back the rest of the app.

This PR is small and only touches the Windows embedder, so the risk to other platforms looks low. Happy to test a candidate build on the affected hardware if that helps.

Thanks!

…level 10_0 (flutter#194114) Adds a parameter (only used in tests) to `egl::Manager` to cap ANGLE's D3D11 feature level. Adds `compositor_opengl_angle_unittests.cc`, based on [`compositor_opengl_unittests.cc`](https://github.com/flutter/flutter/blob/97db29fcaa2aa757c8d86f06187b5268bf4eed4a/engine/src/flutter/shell/platform/windows/compositor_opengl_unittests.cc). Unlike the existing tests, the new tests exercise the compositor with real ANGLE instead of mocks. The new tests exercise the compositor's backing store creation and presenting. The tests run both at uncapped feature level and capped to 10_0. This emulates the case where flutter#191978 resulted in a black window. Verified locally that when flutter#192917 is reverted, the new tests fail when running at feature level 10_0 and pass when running without a feature level cap. Part of flutter#192899: - This PR exercises the OpenGL compositor and swapchain presenting codepath on 10_0, verifying that the backing store framebuffer is valid (GL_FRAMEBUFFER_COMPLETE) with no GL errors. This catches issues that occur at the compositor level, like flutter#191978. - This would not catch issues where the rendering behavior of Impeller is broken due to taking different code paths on 10_0 (e.g. flutter#192918 resulting in different BackdropFilter behavior because offscreen MSAA is unavailable with 10_0). - I intend to follow up with a PR that extends our existing [`engine_integration_golden_test.dart`](https://github.com/flutter/flutter/blob/38785ef8b19e2a1ee1303167074ec98f95f7cb5b/dev/integration_tests/engine_integration_golden_test/integration_test/engine_integration_golden_test.dart) devicelab golden test, which currently [runs on Windows](https://github.com/flutter/flutter/blob/38785ef8b19e2a1ee1303167074ec98f95f7cb5b/dev/devicelab/lib/tasks/integration_tests.dart#L242-L248), to also run with feature level 10_0. These golden tests would catch issues where our Impeller rendering is broken. Goldens do not catch compositor level issues because the golden functionality reads its images from Impeller, skipping the compositor portion of the rendering pipeline. - This PR, together with the anticipated follow up PR, will give us basic smoke testing coverage on 10_0 of the rendering pipeline from Flutter app code, through Impeller rendering, to presenting on the swapchain. ## Pre-launch Checklist - [x] I read the [Contributor Guide] and followed the process outlined there for submitting PRs. - [x] I read the [AI contribution guidelines] and understand my responsibilities, or I am not using AI tools. - [x] I read the [Tree Hygiene] wiki page, which explains my responsibilities. - [x] I read and followed the [Flutter Style Guide], including [Features we expect every widget to implement]. - [x] I signed the [CLA]. - [x] I listed at least one issue that this PR fixes in the description above. - [x] I updated/added relevant in-code documentation (doc comments with `///`). - [x] If this PR introduces a new feature or capability, I created and linked a website documentation issue or PR in [flutter/website] (or verified none is needed). - [x] I added new tests to check the change I am making, or this PR is [test-exempt]. - [x] I followed the [breaking change policy] and added [Data Driven Fixes] where supported. - [x] All existing and new tests are passing. If you need help, consider asking for advice on the #hackers-new channel on [Discord]. If this change needs to override an active code freeze, provide a comment explaining why. The code freeze workflow can be overridden by code reviewers. See pinned issues for any active code freezes with guidance. **Note**: The Flutter team is currently trialing the use of [Gemini Code Assist for GitHub](https://developers.google.com/gemini-code-assist/docs/review-github-code). Comments from the `gemini-code-assist` bot should not be taken as authoritative feedback from the Flutter team. If you find its comments useful you can update your code accordingly, but if you are unsure or disagree with the feedback, please feel free to wait for a Flutter team member's review for guidance on which automated comments should be addressed.  [Contributor Guide]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#overview [AI contribution guidelines]: https://github.com/flutter/flutter/blob/main/CONTRIBUTING.ai.md [Tree Hygiene]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md [test-exempt]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#tests [Flutter Style Guide]: https://github.com/flutter/flutter/blob/main/docs/contributing/Style-guide-for-Flutter-repo.md [Features we expect every widget to implement]: https://github.com/flutter/flutter/blob/main/docs/contributing/Style-guide-for-Flutter-repo.md#features-we-expect-every-widget-to-implement [CLA]: https://cla.developers.google.com/ [flutter/tests]: https://github.com/flutter/tests [flutter/website]: https://github.com/flutter/website [breaking change policy]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#handling-breaking-changes [Discord]: https://github.com/flutter/flutter/blob/main/docs/contributing/Chat.md [Data Driven Fixes]: https://github.com/flutter/flutter/blob/main/docs/contributing/Data-driven-Fixes.md

gaabrielricci · 2026-10-22T14:02:15Z

Confirming that #192917 ("[Windows] Use an unsized texture format for backing stores on OpenGL ES 2") fixes this for us.

We built our Windows app with 3.50.0-0.1.pre (framework 07502c7, engine b890cdd), which includes that commit, and tested it with 3 users. The window renders normally for
all three; no more black screen.

Before this, 3.47.x opened black on the affected machine (Intel HD Graphics 0x8086:0x0042, 2012 driver, WDDM 1.1), while 3.44.2 worked. That matches the regression in the
OpenGL compositor's backing store format that the PR reverts to unsized.

Thanks for the fix. Is there a plan to cherry-pick it to a 3.47.x stable hotfix? Until then we're shipping from beta.

…level 10_0 (flutter#194114) Adds a parameter (only used in tests) to `egl::Manager` to cap ANGLE's D3D11 feature level. Adds `compositor_opengl_angle_unittests.cc`, based on [`compositor_opengl_unittests.cc`](https://github.com/flutter/flutter/blob/97db29fcaa2aa757c8d86f06187b5268bf4eed4a/engine/src/flutter/shell/platform/windows/compositor_opengl_unittests.cc). Unlike the existing tests, the new tests exercise the compositor with real ANGLE instead of mocks. The new tests exercise the compositor's backing store creation and presenting. The tests run both at uncapped feature level and capped to 10_0. This emulates the case where flutter#191978 resulted in a black window. Verified locally that when flutter#192917 is reverted, the new tests fail when running at feature level 10_0 and pass when running without a feature level cap. Part of flutter#192899: - This PR exercises the OpenGL compositor and swapchain presenting codepath on 10_0, verifying that the backing store framebuffer is valid (GL_FRAMEBUFFER_COMPLETE) with no GL errors. This catches issues that occur at the compositor level, like flutter#191978. - This would not catch issues where the rendering behavior of Impeller is broken due to taking different code paths on 10_0 (e.g. flutter#192918 resulting in different BackdropFilter behavior because offscreen MSAA is unavailable with 10_0). - I intend to follow up with a PR that extends our existing [`engine_integration_golden_test.dart`](https://github.com/flutter/flutter/blob/38785ef8b19e2a1ee1303167074ec98f95f7cb5b/dev/integration_tests/engine_integration_golden_test/integration_test/engine_integration_golden_test.dart) devicelab golden test, which currently [runs on Windows](https://github.com/flutter/flutter/blob/38785ef8b19e2a1ee1303167074ec98f95f7cb5b/dev/devicelab/lib/tasks/integration_tests.dart#L242-L248), to also run with feature level 10_0. These golden tests would catch issues where our Impeller rendering is broken. Goldens do not catch compositor level issues because the golden functionality reads its images from Impeller, skipping the compositor portion of the rendering pipeline. - This PR, together with the anticipated follow up PR, will give us basic smoke testing coverage on 10_0 of the rendering pipeline from Flutter app code, through Impeller rendering, to presenting on the swapchain. ## Pre-launch Checklist - [x] I read the [Contributor Guide] and followed the process outlined there for submitting PRs. - [x] I read the [AI contribution guidelines] and understand my responsibilities, or I am not using AI tools. - [x] I read the [Tree Hygiene] wiki page, which explains my responsibilities. - [x] I read and followed the [Flutter Style Guide], including [Features we expect every widget to implement]. - [x] I signed the [CLA]. - [x] I listed at least one issue that this PR fixes in the description above. - [x] I updated/added relevant in-code documentation (doc comments with `///`). - [x] If this PR introduces a new feature or capability, I created and linked a website documentation issue or PR in [flutter/website] (or verified none is needed). - [x] I added new tests to check the change I am making, or this PR is [test-exempt]. - [x] I followed the [breaking change policy] and added [Data Driven Fixes] where supported. - [x] All existing and new tests are passing. If you need help, consider asking for advice on the #hackers-new channel on [Discord]. If this change needs to override an active code freeze, provide a comment explaining why. The code freeze workflow can be overridden by code reviewers. See pinned issues for any active code freezes with guidance. **Note**: The Flutter team is currently trialing the use of [Gemini Code Assist for GitHub](https://developers.google.com/gemini-code-assist/docs/review-github-code). Comments from the `gemini-code-assist` bot should not be taken as authoritative feedback from the Flutter team. If you find its comments useful you can update your code accordingly, but if you are unsure or disagree with the feedback, please feel free to wait for a Flutter team member's review for guidance on which automated comments should be addressed.  [Contributor Guide]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#overview [AI contribution guidelines]: https://github.com/flutter/flutter/blob/main/CONTRIBUTING.ai.md [Tree Hygiene]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md [test-exempt]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#tests [Flutter Style Guide]: https://github.com/flutter/flutter/blob/main/docs/contributing/Style-guide-for-Flutter-repo.md [Features we expect every widget to implement]: https://github.com/flutter/flutter/blob/main/docs/contributing/Style-guide-for-Flutter-repo.md#features-we-expect-every-widget-to-implement [CLA]: https://cla.developers.google.com/ [flutter/tests]: https://github.com/flutter/tests [flutter/website]: https://github.com/flutter/website [breaking change policy]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#handling-breaking-changes [Discord]: https://github.com/flutter/flutter/blob/main/docs/contributing/Chat.md [Data Driven Fixes]: https://github.com/flutter/flutter/blob/main/docs/contributing/Data-driven-Fixes.md

ricardordr · 2026-11-05T15:03:40Z

Dears, is it possible to cherrypick this fix to 3.47 versions?

[Windows] Use an unsized texture format for backing stores on OpenGL …

b045f34

…ES 2

krll-kov requested a review from a team as a code owner September 17, 2026 06:23

github-actions Bot added engine flutter/engine related. See also e: labels. platform-windows Building on or for Windows specifically a: desktop Running on desktop team-windows Owned by the Windows platform team labels Sep 17, 2026

krll-kov mentioned this pull request Sep 17, 2026

[Windows] Regression in 3.47.0: black window on GPUs limited to D3D11 feature level 9_3 — compositor passes sized internal formats on an ES 2.0 context #191978

Closed

gemini-code-assist Bot reviewed Sep 17, 2026

View reviewed changes

krll-kov mentioned this pull request Sep 17, 2026

[Impeller] [Windows] BackdropFilter erases content drawn before it when offscreen MSAA is unavailable (OpenGL ES 2.0, D3D11 feature level 10_0) #192918

Closed

gaaclarke added the CICD Run CI/CD label Sep 17, 2026

gaaclarke self-requested a review September 17, 2026 15:32

gaaclarke requested changes Sep 17, 2026

View reviewed changes

Update compositor_opengl_unittests.cc

9c3b566

flutter-dashboard Bot removed the CICD Run CI/CD label Sep 17, 2026

krll-kov requested a review from gaaclarke September 17, 2026 17:02

gaaclarke added the CICD Run CI/CD label Sep 17, 2026

gaaclarke approved these changes Sep 17, 2026

View reviewed changes

gaaclarke requested a review from b-luk September 17, 2026 17:10

b-luk approved these changes Sep 17, 2026

View reviewed changes

gaaclarke added the autosubmit Merge PR when tree becomes green via auto submit App label Sep 17, 2026

auto-submit Bot removed the autosubmit Merge PR when tree becomes green via auto submit App label Sep 17, 2026

gaaclarke added the autosubmit Merge PR when tree becomes green via auto submit App label Sep 17, 2026

auto-submit Bot added this pull request to the merge queue Sep 17, 2026

Merged via the queue into flutter:master with commit 94306f7 Sep 17, 2026
35 of 36 checks passed

flutter-dashboard Bot removed the autosubmit Merge PR when tree becomes green via auto submit App label Sep 17, 2026

krll-kov mentioned this pull request Sep 20, 2026

[Impeller][GLES] Link programs in parallel on ANGLE with GL_KHR_parallel_shader_compile #193066

Closed

7 tasks done

This was referenced Oct 9, 2026

[windows] Add tests that render at D3D11 feature level 10_0 #194100

Closed

[windows] Add OpenGL compositor tests against ANGLE at D3D11 feature level 10_0 #194114

Merged

	gl_ = std::make_unique<impeller::ProcTableGLES>(resolver_);
	if (!gl_->IsValid()) {
	gl_.reset();
	return false;
	}

	description_ = std::make_unique<DescriptionGLES>(*this);

	if (!description_->IsValid()) {
	return;
	}

Conversation

krll-kov commented Sep 17, 2026

Pre-launch Checklist

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist Bot Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

krll-kov Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

krll-kov Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

gaaclarke left a comment

Choose a reason for hiding this comment

Uh oh!

gaaclarke Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

gaaclarke Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

gaaclarke Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

krll-kov commented Sep 17, 2026

Uh oh!

gaaclarke left a comment

Choose a reason for hiding this comment

Uh oh!

gaaclarke commented Sep 17, 2026

Uh oh!

b-luk left a comment

Choose a reason for hiding this comment

Uh oh!

auto-submit Bot commented Sep 17, 2026

Uh oh!

gaaclarke commented Sep 17, 2026

Uh oh!

krll-kov commented Sep 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gaaclarke commented Sep 17, 2026

Uh oh!

krll-kov commented Sep 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

eduardohr-muniz commented Oct 2, 2026

Uh oh!

gaabrielricci commented Oct 22, 2026

Uh oh!

ricardordr commented Nov 5, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

krll-kov commented Sep 17, 2026 •

edited

Loading

krll-kov commented Sep 17, 2026 •

edited

Loading