[engine] Honor targetWidth/targetHeight for animated images by mohamedalsamrai · Pull Request #194615 · flutter/flutter

mohamedalsamrai · 2026-10-17T01:30:58Z

Description

ui.instantiateImageCodec, ui.instantiateImageCodecWithSize and
ImageDescriptor.instantiateCodec ignore targetWidth/targetHeight for images
with more than one frame. ImageDescriptor::instantiateCodec only forwards the
target size to SingleFrameCodec; MultiFrameCodec always decodes, composites
and uploads every frame at the intrinsic size of the image. As a result,
cacheWidth/cacheHeight and ResizeImage have no effect on animated GIF,
WebP and APNG images, which is a common source of high memory usage (for
example, grids of animated stickers).

This PR forwards the target size to MultiFrameCodec:

Frames are still decoded and composited at the intrinsic size. Frame rects,
blending and disposal are defined in the coordinate space of the full canvas,
and lastRequiredFrame_ is reused as the backdrop of later frames, so this
part is unchanged.
Only the finished frame is resized before it is uploaded, using the same
sampling as single-frame images (SkFilterMode::kLinear, no mipmaps, as in
ResizeRasterImage and ResizeOnCpu).
All three upload paths (Skia, Impeller on iOS, Impeller elsewhere) upload the
resized frame, so the resulting ui.Image has the requested size.
No Dart API changes are needed: the framework already passes the target size.
Like single-frame images, frames are never uploaded larger than the maximum texture size: the Skia path passes limitToMaxTextureSize, and the Impeller path clamps the target size with ImageDecoderImpeller::ClampToMaxTextureSize, which this PR factors out of DecompressTexture.

Memory example: a 1000×1000 animated GIF displayed with cacheWidth: 200 now
uploads frames of about 160 KB instead of about 4 MB each (plus mipmaps on the
Skia backend). The CPU-side compositing buffers (one frame and the cached
backdrop) keep the intrinsic size, exactly as today.

Alternatives considered

Both alternatives were evaluated with an emulation of
MultiFrameCodec::State::GetNextFrameImage on top of Skia's real GIF (Wuffs)
and WebP codecs (skia-python, Skia m144), over 35 files (17 Skia/engine
fixtures, 6 synthetic GIFs covering every disposal method, 12 adversarial GIFs),
9 scales and 2 loops each (14,742 frames).

Decode frames directly at a reduced size with
ImageGenerator::GetScaledDimensions(), as SingleFrameCodec does, keeping
the single cached backdrop. Skia's GIF and WebP codecs support arbitrary
downscaling (SkScalingCodec), but with a single backdrop the output
diverges from Skia's own reduced-size decode whenever a frame's required
frame is older than the cached frame and a kRestoreBGColor clear was
applied at the reduced size: the cleared rect is rounded out, while frame
draws cover pixels by their centers. 812 of 14,742 frames differed,
including 6 frames of Skia's randPixelsAnim.gif (1px transparent slivers).
Same as 1, with a dependency-aware cache of backdrops (derived from
FrameInfo::required_frame). This matched Skia's reference for all frames,
but heart.webp (an existing engine fixture) needs 15 backdrops cached at
the same time, and the bookkeeping is considerably more complex. This could
be a follow-up optimization where reduced-size decoding saves significant
CPU memory (WebP).
The chosen approach is exact by construction: every delivered frame is today's
full-size frame, resized. In the same emulation, the current full-size
algorithm matched Skia's reference decode for all 1,638 frames, and the
half-size output equalled the rounded 2×2 average of the full-size frames
(maximum channel difference 0) on every even-sized fixture.

Tests

Added to testing/dart/codec_test.dart:

Animated gif frames are resized to targetWidth and targetHeight:
test640x479.gif decoded at 160×120 for two loops.
Animated webp frames are resized when only targetWidth is given:
heart.webp with targetWidth: 100 yields 100×100 frames.
instantiateImageCodecWithSize resizes animated images: the reproduction
from Resizing parameters are not supported for multi-frame images in the engine. #137668.

./testing/run_tests.py --type=dart --dart-filter=codec_test.dart --variant <host_debug_unopt variant>

Out of scope

Web (CanvasKit/Skwasm) still ignores targetWidth/targetHeight for
multi-frame images (see implement targetWidth and targetHeight flutter-team-archive/engine#38028).
TargetPixelFormat is still not forwarded to MultiFrameCodec.

Related issues

Fixes #90804
Fixes #137668

Pre-launch Checklist

I read the [Contributor Guide] and followed the process outlined there for submitting PRs.
I read the [Tree Hygiene] wiki page, which explains my responsibilities.
I read and followed the [Flutter Style Guide], including [Features we expect every widget to implement].
I signed the [CLA].
I listed at least one issue that this PR fixes in the description above.
I updated/added relevant documentation (doc comments with ///).
I added new tests to check the change I am making, or this PR is [test-exempt].
I followed the [breaking change policy] and added [Data Driven Fixes] where supported.
All existing and new tests are passing.
[Contributor Guide]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#overview
[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/
[breaking change policy]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#handling-breaking-changes
[Data Driven Fixes]: https://github.com/flutter/flutter/blob/main/docs/contributing/Data-driven-Fixes.md

Testing note

I could not build the engine locally (not enough disk
space), so I have not run the engine tests myself and am relying on
presubmit CI. The C++ changes are formatted with clang-format, and
codec_test.dart passes dart format and dart analyze.

gemini-code-assist

Code Review

This pull request implements resizing for animated (multi-frame) images (such as GIFs, WebPs, and APNGs) to the requested target width and height. Frames are decoded and composited at their intrinsic size to maintain correct blending and disposal behavior, and only the final composited frame is resized to the target size. This change updates MultiFrameCodec and its internal State to accept a target size, introduces a ResizeFrame helper using bilinear filtering, and adapts both Skia and Impeller rendering paths to use the resized output. Comprehensive unit tests have been added to verify the resizing behavior and correctness against downscaled reference frames. I have no feedback to provide.

b-luk

The presubmit checks show a Linux customer_testing failure. That should be resolved if you merge from master.

b-luk · 2026-10-23T23:11:57Z

                           : generator_->GetPlayCount() - 1),
-      is_impeller_enabled_(UIDartState::Current()->IsImpellerEnabled()) {}
+      is_impeller_enabled_(UIDartState::Current()->IsImpellerEnabled()),
+      target_size_(target_size.isEmpty() ? generator_->GetInfo().dimensions()


When does target_size.isEmpty() happen? It's the default parameter value for the MultiFrameCodec constructor, but I don't think the MultiFrameCodec constructor is ever called with the default value.

When called from Dart, I think it always uses non-zero values for width and height.

Can we get rid of the constructor's default value, and get rid of this isEmpty fallback here?

Only when a caller passes a size with a 0 dimension. The production caller, ImageDescriptor::instantiateCodec, always passes the size computed in Dart, which is positive except in an integer-division corner case with extreme aspect ratios (e.g. targetWidth: 50 for a 1000x10 image gives targetHeight 0), where the single-frame path can't produce an image either (ResizeRasterImage rejects empty dimensions). The default was only used by two tests (image_decoder_unittests.cc and shell_io_manager_unittests.cc). Removed the default and the fallback; the tests now pass the GIF's size.

b-luk · 2026-10-23T23:20:12Z

+  // See https://github.com/flutter/flutter/issues/90804.
+  SkBitmap output = bitmap;
+  if (output.dimensions() != target_size_) {
+    std::optional<SkBitmap> resized = ResizeFrame(bitmap, target_size_);


With the single-frame codec, resizing gets capped to max texture size:

Impeller here

flutter/engine/src/flutter/lib/ui/painting/image_decoder_impeller.cc

Lines 365 to 380 in 0590b2b

SkISize target_size =

SkISize::Make(options.target_width, options.target_height);

// Shrink both sides by the same factor so the aspect ratio is kept, but

// never below one pixel.

if (options.target_width > max_texture_size.width ||

options.target_height > max_texture_size.height) {

const double scale = std::min(static_cast<double>(max_texture_size.width) /

std::max(options.target_width, 1u),

static_cast<double>(max_texture_size.height) /

std::max(options.target_height, 1u));

target_size = SkISize::Make(

std::max(1,

static_cast<int>(std::lround(options.target_width * scale))),

std::max(1,

static_cast<int>(std::lround(options.target_height * scale))));

}

and Skia here

flutter/engine/src/flutter/lib/ui/painting/image_decoder_skia.cc

Lines 205 to 211 in 0590b2b

sk_sp<SkImage> texture_image =

SkImages::CrossContextTextureFromPixmap(

context.get(), // context

pixmap, // pixmap

true, // buildMips,

true // limitToMaxTextureSize

);

Should we do the same for the multi-frame case?

Yes, done. On Skia, CrossContextTextureFromPixmap now gets limitToMaxTextureSize = true, as in UploadRasterImage. On Impeller, I moved the clamping out of DecompressTexture into ImageDecoderImpeller::ClampToMaxTextureSize() without changing its behavior (the existing DecompressTexture max-texture-size checks still cover it, plus a small direct test), and MultiFrameCodec clamps the target size with it before resizing.

b-luk · 2026-10-23T23:22:09Z

+  // `cacheWidth`/`cacheHeight` or `ResizeImage` in the framework).
+  //
+  // See https://github.com/flutter/flutter/issues/90804.
+  SkBitmap output = bitmap;


Could we reuse/reassign the existing bitmap and info variables instead of creating new output and output_info variables?

Done: bitmap and info are reassigned after resizing. lastRequiredFrame_ keeps its reference to the full-size pixels, and the upload code is back to what's on master apart from limitToMaxTextureSize.

mohamedalsamrai · 2026-10-24T00:04:48Z

Thanks for the review! Rebased onto the latest master, which picks up the customer_testing fix from #194968, and addressed the comments.

b-luk · 2026-10-24T00:29:53Z

Thanks for the review! Rebased onto the latest master, which picks up the customer_testing fix from #194968, and addressed the comments.

This PR is still based on the same original commit. Did you push to github after rebasing?

ImageDescriptor::instantiateCodec only forwarded the target size to SingleFrameCodec, so animated GIF, WebP and APNG frames were always decoded and uploaded at their intrinsic size, and cacheWidth, cacheHeight and ResizeImage had no effect on them. MultiFrameCodec now receives the target size. Frames are still composited at the intrinsic size, because frame rects, blending and disposal are defined in canvas coordinates and lastRequiredFrame_ is reused as a backdrop. Only the finished frame is resized, with the same sampling as single-frame images, before it is uploaded.

…re size, reuse bitmap/info

mohamedalsamrai · 2026-10-24T00:38:59Z

Sorry @b-luk , my mistake the review changes were pushed, but the rebase hadn't gone through. It's now rebased onto the latest master.

gaaclarke · 2026-10-26T23:16:32Z

+  ///        ratio, so that it fits within |max_texture_size|. Each side stays
+  ///        at least one pixel. A size that already fits is returned as is.
+  static SkISize ClampToMaxTextureSize(const SkISize& target_size,
+                                       const impeller::ISize& max_texture_size);


I'd just copy the ISize instead of passing it by ref. Since it small it should be faster. I don't think the linter will complain.

Done, SkISize and impeller::ISize are now passed by value

gaaclarke · 2026-10-26T23:17:50Z

+// image_decoder_skia.cc and `ResizeOnCpu` in image_decoder_impeller.cc), so
+// `targetWidth` and `targetHeight` behave the same way for every image.
+static std::optional<SkBitmap> ResizeFrame(const SkBitmap& frame,
+                                           const SkISize& target_size) {


gaaclarke · 2026-10-26T23:19:01Z

+      std::ostringstream ostr;
+      ostr << "Could not resize frame " << nextFrameIndex_ << " to "
+           << target_size.width() << "x" << target_size.height();
+      std::string decode_error = ostr.str();


use absl::StrCat please

Done for target_size. I kept frame as a const reference since copying an SkBitmap bumps its pixel ref count

gaaclarke · 2026-10-26T23:20:06Z

+                  resourceContext.get(),  // context
+                  pixmap,                 // pixmap
+                  true,                   // buildMips
+                  true                    // limitToMaxTextureSize


please use proper argument comments: /*buildMips=*/true (the names will be validated by the linter)

gaaclarke · 2026-10-26T23:22:51Z

+    codec.dispose();
+  });
+
+  test('Resized animated frames match the downscaled full-size frames', () async {


This is overkill, let's remove it.

gaaclarke · 2026-10-26T23:24:22Z

+  /// @brief Creates a codec for an animated (multi-frame) image.
+  ///
+  /// @param[in] generator    The decoder for the encoded image.
+  /// @param[in] target_size  The size at which frames are delivered to Dart,


mention the filtering that is happening when resizing

gaaclarke

lgtm, thanks!

…-target-size

mohamedalsamrai · 2026-10-27T15:48:37Z

@b-luk can u add CICD label again ?

auto-submit · 2026-10-27T18:18:12Z

autosubmit label was removed for flutter/flutter/194615, because - The status or check suite Google testing has failed. Please fix the issues identified (or deflake) before re-applying this label.

mohamedalsamrai · 2026-10-27T18:25:54Z

Google testing is red and stayed red on the rerun, so it doesn't look
like a flake. I can't open go/frob_pr/194615 — @gaaclarke could you check
what's failing when you get a chance? Happy to fix anything on my side.

b-luk · 2026-10-27T18:38:04Z

The Google testing failure is likely due to the internal Google tests using an engine build based on an older commit compared to your PR. A roll is in progress to update the internal Google engine version. Once that completes, I'll re-run the tests for your PR and hopefully the problem will be resolved.

gaaclarke · 2026-10-27T19:59:06Z

Oops, sorry @b-luk I just bypassed the failures since they weren't apropos. I don't think this could affect google anyways since their testing infra is on skia.

b-luk · 2026-10-27T20:05:30Z

Oops, sorry @b-luk I just bypassed the failures since they weren't apropos. I don't think this could affect google anyways since their testing infra is on skia.

Seems fine to me. I suggested waiting for google's next engine roll and then rerunning just in case, because I didn't exhaustively inspect all the failures and there isn't an urgency to get this PR in immediately. But I think it's highly unlikely that this PR breaks google tests, so bypassing is fine with me.

mohamedalsamrai · 2026-10-27T22:03:51Z

Well done guys
thank you

flutter/flutter@d4c8577...0c2473c 2026-10-28 [email protected] Roll Packages from a8549ae to 13d38e0 (9 revisions) (flutter/flutter#195488) 2026-10-28 [email protected] Use int.oneBitCount in PersistentHashMap and scope 53-bit caps to JS (flutter/flutter#195005) 2026-10-28 [email protected] Roll pub packages (flutter/flutter#195483) 2026-10-28 [email protected] [iPhone Duo] Use the correct UIScreen to report the display size and update when fold/unfold (flutter/flutter#195199) 2026-10-28 [email protected] [fuchsia] Fix portable_ui_test.h includes for newer Fuchsia SDKs (flutter/flutter#195420) 2026-10-28 [email protected] Change default pipeline options to match production draw calls (flutter/flutter#194597) 2026-10-28 [email protected] Roll Fuchsia Test Scripts from 2T18KC7-LDQtmQiLP... to xl1LWBuppzu-BnOzV... (flutter/flutter#195450) 2026-10-27 [email protected] [flutter_tools] Inject dependencies into Darwin build targets, native assets, and Xcode build settings (flutter/flutter#194052) 2026-10-27 [email protected] Revert: Reland square corners fix (flutter/flutter#195436) 2026-10-27 [email protected] [flutter_tools] Add DeviceService app launch and flutter run support for tool extensions (flutter/flutter#194589) 2026-10-27 [email protected] [engine] Honor targetWidth/targetHeight for animated images (flutter/flutter#194615) 2026-10-27 [email protected] [Android][a11y] Set visibleToUser=false for hidden semantics nodes (flutter/flutter#194411) 2026-10-27 [email protected] Fix repo deletion test flake (flutter/flutter#195349) 2026-10-27 [email protected] Roll Packages from 23a05b6 to a8549ae (10 revisions) (flutter/flutter#195434) 2026-10-27 [email protected] [web] Capture pointers to the Flutter view so drags survive platform view iframes (flutter/flutter#195187) 2026-10-27 [email protected] [web] Add opt-in raw samples, measured sample count and warm-up phase hook to the web benchmark recorder (flutter/flutter#195118) 2026-10-27 [email protected] Link migration guide in deprecation messages (flutter/flutter#195361) 2026-10-27 [email protected] Pin basic_material_app_macos__compile to macOS 26 (flutter/flutter#195429) 2026-10-27 [email protected] Add benchmark for warmed pipeline usage (flutter/flutter#195001) 2026-10-27 [email protected] [flutter_tools] Add device log streaming support to DeviceService and ExtensionBackedDevice (flutter/flutter#194964) 2026-10-27 [email protected] Roll Dart SDK from 1d16c04dc346 to 65fc979759ca (4 revisions) (flutter/flutter#195383) If this roll has caused a breakage, revert this CL and set the roller to dry run mode using the controls here: https://autoroll.skia.org/r/flutter-packages Please CC [email protected],[email protected] on the revert to ensure that a human is aware of the problem. To file a bug in Packages: https://github.com/flutter/flutter/issues/new/choose To report a problem with the AutoRoller itself, please file a bug: https://issues.skia.org/issues/new?component=1389291&template=1850622 Documentation for the AutoRoller is here: https://skia.googlesource.com/buildbot/+doc/main/autoroll/README.md

…194615) ## Description `ui.instantiateImageCodec`, `ui.instantiateImageCodecWithSize` and `ImageDescriptor.instantiateCodec` ignore `targetWidth`/`targetHeight` for images with more than one frame. `ImageDescriptor::instantiateCodec` only forwards the target size to `SingleFrameCodec`; `MultiFrameCodec` always decodes, composites and uploads every frame at the intrinsic size of the image. As a result, `cacheWidth`/`cacheHeight` and `ResizeImage` have no effect on animated GIF, WebP and APNG images, which is a common source of high memory usage (for example, grids of animated stickers). This PR forwards the target size to `MultiFrameCodec`: - Frames are still decoded and composited at the intrinsic size. Frame rects, blending and disposal are defined in the coordinate space of the full canvas, and `lastRequiredFrame_` is reused as the backdrop of later frames, so this part is unchanged. - Only the finished frame is resized before it is uploaded, using the same sampling as single-frame images (`SkFilterMode::kLinear`, no mipmaps, as in `ResizeRasterImage` and `ResizeOnCpu`). - All three upload paths (Skia, Impeller on iOS, Impeller elsewhere) upload the resized frame, so the resulting `ui.Image` has the requested size. No Dart API changes are needed: the framework already passes the target size. Like single-frame images, frames are never uploaded larger than the maximum texture size: the Skia path passes limitToMaxTextureSize, and the Impeller path clamps the target size with ImageDecoderImpeller::ClampToMaxTextureSize, which this PR factors out of DecompressTexture. Memory example: a 1000×1000 animated GIF displayed with `cacheWidth: 200` now uploads frames of about 160 KB instead of about 4 MB each (plus mipmaps on the Skia backend). The CPU-side compositing buffers (one frame and the cached backdrop) keep the intrinsic size, exactly as today. ## Alternatives considered Both alternatives were evaluated with an emulation of `MultiFrameCodec::State::GetNextFrameImage` on top of Skia's real GIF (Wuffs) and WebP codecs (skia-python, Skia m144), over 35 files (17 Skia/engine fixtures, 6 synthetic GIFs covering every disposal method, 12 adversarial GIFs), 9 scales and 2 loops each (14,742 frames). 1. **Decode frames directly at a reduced size** with `ImageGenerator::GetScaledDimensions()`, as `SingleFrameCodec` does, keeping the single cached backdrop. Skia's GIF and WebP codecs support arbitrary downscaling (`SkScalingCodec`), but with a single backdrop the output diverges from Skia's own reduced-size decode whenever a frame's required frame is older than the cached frame and a `kRestoreBGColor` clear was applied at the reduced size: the cleared rect is rounded out, while frame draws cover pixels by their centers. 812 of 14,742 frames differed, including 6 frames of Skia's `randPixelsAnim.gif` (1px transparent slivers). 2. **Same as 1, with a dependency-aware cache of backdrops** (derived from `FrameInfo::required_frame`). This matched Skia's reference for all frames, but `heart.webp` (an existing engine fixture) needs 15 backdrops cached at the same time, and the bookkeeping is considerably more complex. This could be a follow-up optimization where reduced-size decoding saves significant CPU memory (WebP). The chosen approach is exact by construction: every delivered frame is today's full-size frame, resized. In the same emulation, the current full-size algorithm matched Skia's reference decode for all 1,638 frames, and the half-size output equalled the rounded 2×2 average of the full-size frames (maximum channel difference 0) on every even-sized fixture. ## Tests Added to `testing/dart/codec_test.dart`: - `Animated gif frames are resized to targetWidth and targetHeight`: `test640x479.gif` decoded at 160×120 for two loops. - `Animated webp frames are resized when only targetWidth is given`: `heart.webp` with `targetWidth: 100` yields 100×100 frames. - `instantiateImageCodecWithSize resizes animated images`: the reproduction from flutter#137668. ``` ./testing/run_tests.py --type=dart --dart-filter=codec_test.dart --variant <host_debug_unopt variant> ``` ## Out of scope - Web (CanvasKit/Skwasm) still ignores `targetWidth`/`targetHeight` for multi-frame images (see flutter-team-archive/engine#38028). - `TargetPixelFormat` is still not forwarded to `MultiFrameCodec`. ## Related issues Fixes flutter#90804 Fixes flutter#137668 ## Pre-launch Checklist - [x] I read the [Contributor Guide] and followed the process outlined there for submitting PRs. - [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 documentation (doc comments with `///`). - [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. [Contributor Guide]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#overview [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/ [breaking change policy]: https://github.com/flutter/flutter/blob/main/docs/contributing/Tree-hygiene.md#handling-breaking-changes [Data Driven Fixes]: https://github.com/flutter/flutter/blob/main/docs/contributing/Data-driven-Fixes.md ## Testing note I could not build the engine locally (not enough disk space), so I have not run the engine tests myself and am relying on presubmit CI. The C++ changes are formatted with clang-format, and codec_test.dart passes dart format and dart analyze.

github-actions Bot added the engine flutter/engine related. See also e: labels. label Oct 17, 2026

gemini-code-assist Bot reviewed Oct 17, 2026

View reviewed changes

gaaclarke requested a review from b-luk October 20, 2026 20:42

b-luk added the CICD Run CI/CD label Oct 22, 2026

b-luk requested changes Oct 23, 2026

View reviewed changes

flutter-dashboard Bot removed the CICD Run CI/CD label Oct 23, 2026

mohamedalsamrai requested a review from b-luk October 24, 2026 00:07

mohamedalsamrai added 2 commits October 24, 2026 03:35

Address review: no default target size, limit frames to the max textu…

1d3549a

…re size, reuse bitmap/info

mohamedalsamrai force-pushed the multi-frame-codec-target-size branch from f757c0a to 1d3549a Compare October 24, 2026 00:35

b-luk added the CICD Run CI/CD label Oct 24, 2026

b-luk previously approved these changes Oct 26, 2026

View reviewed changes

b-luk requested a review from gaaclarke October 26, 2026 22:43

gaaclarke requested changes Oct 26, 2026

View reviewed changes

Address review comments

322b4a3

mohamedalsamrai dismissed b-luk’s stale review via 322b4a3 October 26, 2026 23:50

flutter-dashboard Bot removed the CICD Run CI/CD label Oct 26, 2026

mohamedalsamrai requested a review from gaaclarke October 27, 2026 00:04

gaaclarke approved these changes Oct 27, 2026

View reviewed changes

gaaclarke requested a review from b-luk October 27, 2026 00:33

b-luk added the CICD Run CI/CD label Oct 27, 2026

Merge remote-tracking branch 'upstream/master' into multi-frame-codec…

698bf28

…-target-size

flutter-dashboard Bot removed the CICD Run CI/CD label Oct 27, 2026

gaaclarke added the CICD Run CI/CD label Oct 27, 2026

b-luk approved these changes Oct 27, 2026

View reviewed changes

b-luk added the autosubmit Merge PR when tree becomes green via auto submit App label Oct 27, 2026

auto-submit Bot removed the autosubmit Merge PR when tree becomes green via auto submit App label Oct 27, 2026

gaaclarke added the autosubmit Merge PR when tree becomes green via auto submit App label Oct 27, 2026

auto-submit Bot added this pull request to the merge queue Oct 27, 2026

Merged via the queue into flutter:master with commit 2381790 Oct 27, 2026
36 checks passed

flutter-dashboard Bot removed the autosubmit Merge PR when tree becomes green via auto submit App label Oct 27, 2026

This was referenced Oct 28, 2026

Roll Flutter from d4c857754c0e to 97cd226ec5ee (19 revisions) flutter/packages#13469

Closed

Roll Flutter from d4c857754c0e to 0c2473cd50c8 (21 revisions) flutter/packages#13470

Merged

chunhtai mentioned this pull request Oct 29, 2026

Flutter Notable Commits - 2026 #181433

Open

	SkISize target_size =
	SkISize::Make(options.target_width, options.target_height);
	// Shrink both sides by the same factor so the aspect ratio is kept, but
	// never below one pixel.
	if (options.target_width > max_texture_size.width \|\|
	options.target_height > max_texture_size.height) {
	const double scale = std::min(static_cast<double>(max_texture_size.width) /
	std::max(options.target_width, 1u),
	static_cast<double>(max_texture_size.height) /
	std::max(options.target_height, 1u));
	target_size = SkISize::Make(
	std::max(1,
	static_cast<int>(std::lround(options.target_width * scale))),
	std::max(1,
	static_cast<int>(std::lround(options.target_height * scale))));
	}

	sk_sp<SkImage> texture_image =
	SkImages::CrossContextTextureFromPixmap(
	context.get(), // context
	pixmap, // pixmap
	true, // buildMips,
	true // limitToMaxTextureSize
	);

Conversation

mohamedalsamrai commented Oct 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Alternatives considered

Tests

Out of scope

Related issues

Pre-launch Checklist

Testing note

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

b-luk left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mohamedalsamrai commented Oct 24, 2026

Uh oh!

b-luk commented Oct 24, 2026

Uh oh!

mohamedalsamrai commented Oct 24, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

gaaclarke Oct 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

gaaclarke left a comment

Choose a reason for hiding this comment

Uh oh!

mohamedalsamrai commented Oct 27, 2026

Uh oh!

auto-submit Bot commented Oct 27, 2026

Uh oh!

mohamedalsamrai commented Oct 27, 2026

Uh oh!

b-luk commented Oct 27, 2026

Uh oh!

gaaclarke commented Oct 27, 2026

Uh oh!

b-luk commented Oct 27, 2026

Uh oh!

Uh oh!

mohamedalsamrai commented Oct 27, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

mohamedalsamrai commented Oct 17, 2026 •

edited

Loading

gaaclarke Oct 26, 2026 •

edited

Loading