[release] Publish reviewed teasers from release notes

[mattleibow]

Description

This separates irreversible release publication from reviewed customer-facing teaser prose.

release-publish now creates and publishes a marked GitHub-generated-notes body unchanged, then recoverably dispatches release notes. It no longer authors, classifies, approves, or edits a synchronous teaser.

Release notes record every exact published tag/build in data.json shipments[]. The line's single self-contained prose.json holds cumulative website prose and immutable exact-tag release_teasers; the task manifest separates those work axes and preserves unlisted historical teasers. Empty stable deltas render deterministic no-change prose without AI.

After the reviewed notes PR merges, a zero-AI workflow semantically diffs exact teaser entries and updates only eligible marked releases. Non-canceling concurrency, full-batch preflight, body SHA/ETag checks, and idempotent reruns prevent partial or stale updates. GitHub-generated notes remain byte-for-byte frozen, manual content outside markers is preserved, and unmarked legacy releases are skipped.

Related issues

None.

Required skia PR

None.

Areas affected

• Managed API (binding/)
• Native / C API (externals/skia/src/c, include/c)
• Generated P/Invoke bindings
• Native dependency or Skia update
• Views & integrations
• Rendering output / visual behavior
• Performance
• Tests
• Build, packaging, or CI
• Documentation or samples

Changes

None — release automation and documentation only; no public API or runtime behavior changes.

Testing

• 29 release-publish tests
• 17 exact-shipment/prose lifecycle tests
• 23 deterministic updater tests
• 72-page offline release-notes render
• Prose JSON Schema fixture validation
• Python compilation and both modified skill validations
• gh aw compile --strict, workflow source validation, and actionlint
• Traditional updater workflow YAML parse
• git diff --check and clean-tree verification

No runtime, native, or rendering output changed, so platform and golden-image testing is not applicable.

Checklist

• Tests added or updated (if omitted, explain why above)
• Changes above lists all public API and behavioral changes (or "None.")
• New/changed public API docs issue — N/A, no API changes
• Native companion PR and regenerated bindings — N/A, no native changes

[github-actions]

📦 Try the packages from this PR

Warning

Do not run these scripts without first reviewing the code in this PR.

Step 1 — Download the packages

bash / macOS / Linux:

curl -fsSL https://raw.githubusercontent.com/mono/SkiaSharp/main/scripts/get-skiasharp-pr.sh | bash -s -- 4722

PowerShell / Windows:

iex "& { $(irm https://raw.githubusercontent.com/mono/SkiaSharp/main/scripts/get-skiasharp-pr.ps1) } 4722"

Step 2 — Add the local NuGet source

dotnet nuget add source ~/.skiasharp/hives/pr-4722/packages --name skiasharp-pr-4722

More options

Option	Description
--successful-only / -SuccessfulOnly	Only use successful builds
--force / -Force	Overwrite previously downloaded packages
--list / -List	List available artifacts without downloading
--build-id ID / -BuildId ID	Download from a specific build

Or download manually from Azure Pipelines — look for the nuget artifact on the build for this PR.

Remove the source when you're done:

dotnet nuget remove source skiasharp-pr-4722

[github-actions]

📊 SkiaSharp benchmarks — PR #4722

⭐ this PR (full source build) vs 🌙 nightly · Linux · Windows · macOS

Informational only — this never blocks the PR. 🟢 faster / less allocation · 🔴 slower / more allocation; moves under 5% are hidden as noise.

⏱️ Times are raw BenchmarkDotNet means, and the ⭐ PR and baseline legs run on separate CI runners, so microbenchmarks can swing run-to-run — treat small time deltas as noise. Allocations are deterministic and the reliable signal. The interactive perf-dashboard (linked below) applies smoothing for the trend view.

Highlights

⏱️ Time — 🔴 39 slower · 🟢 30 faster

• 🔴 MatrixMapPointsBenchmark.MapVector(Points: 256) · macOS · 2.43 µs → 3.37 µs (+39%)
• 🔴 MatrixMapPointsBenchmark.MapPoints(Points: 256) · macOS · 135.9 ns → 184.9 ns (+36%)
• 🔴 MatrixMapPointsBenchmark.MapRect(Points: 256) · macOS · 3.05 µs → 4.08 µs (+34%)
• 🔴 MatrixMapPointsBenchmark.MapPoints(Points: 4096) · macOS · 1.52 µs → 2.02 µs (+33%)
• 🟢 MatrixMapPointsBenchmark.MapPoints(Points: 256) · Windows · 182.7 ns → 135.5 ns (-26%)
• …and 64 more (see details below)

Full per-OS benchmark deltas

Linux

⏱️ Time (vs 🌙 nightly 4.152.0-nightly.10)

Benchmark	baseline	this PR	Δ
RasterImageLifecycleBenchmark.CreateDataWithReleaseProc(Count: 256)	100.53 µs	108.83 µs	🔴 +8%
MatrixMapPointsBenchmark.MapPoints(Points: 4096)	2.62 µs	2.79 µs	🔴 +7%
ColorMathBenchmark.UnPreMultiply(Colors: 4096)	11.44 µs	12.11 µs	🔴 +6%
ColorMathBenchmark.ToColor(Colors: 4096)	14.99 µs	14.14 µs	🟢 -6%
RasterImageLifecycleBenchmark.CreateRasterImage(Count: 256)	257.65 µs	271.54 µs	🔴 +5%
MatrixMapPointsBenchmark.MapPoints(Points: 256)	209.2 ns	219.8 ns	🔴 +5%

Windows

⏱️ Time (vs 🌙 nightly 4.152.0-nightly.10)

Benchmark	baseline	this PR	Δ
MatrixMapPointsBenchmark.MapPoints(Points: 256)	182.7 ns	135.5 ns	🟢 -26%
RasterImageLifecycleBenchmark.CreateDataWithReleaseProc(Count: 256)	96.45 µs	121.29 µs	🔴 +26%
LargeImageScaleBenchmark.UpscaleOpaque(Size: 2048)	3.25 ms	2.50 ms	🟢 -23%
ColorMathBenchmark.ToColorF(Colors: 4096)	14.14 µs	10.94 µs	🟢 -23%
LargeImageScaleBenchmark.UpscaleCrossfade(Size: 2048)	11.24 ms	8.70 ms	🟢 -23%
SceneRenderBenchmark.RenderFrame(Complexity: 4)	10.38 ms	8.11 ms	🟢 -22%
LargeImageScaleBenchmark.UpscaleCrossfade(Size: 1024)	2.84 ms	2.22 ms	🟢 -22%
LargeImageScaleBenchmark.UpscaleOpaque(Size: 1024)	822.95 µs	645.22 µs	🟢 -22%
MatrixMapPointsBenchmark.MapPoint(Points: 4096)	86.67 µs	68.55 µs	🟢 -21%
ColorMathBenchmark.UnPreMultiply(Colors: 4096)	12.79 µs	10.12 µs	🟢 -21%
SceneRenderBenchmark.RenderFrame(Complexity: 1)	2.15 ms	1.71 ms	🟢 -21%
MatrixMapPointsBenchmark.MapPoint(Points: 256)	5.42 µs	4.29 µs	🟢 -21%
MatrixOpsBenchmark.Invert(Count: 4096)	148.24 µs	117.82 µs	🟢 -21%
MatrixMapPointsBenchmark.MapRect(Points: 256)	8.53 µs	6.81 µs	🟢 -20%
BitmapDrawBenchmark.DrawScaledTiles(Tiles: 64)	361.89 µs	290.50 µs	🟢 -20%
RuntimeEffectShaderBenchmark.DrawFrame	1.22 ms	977.03 µs	🟢 -20%
MatrixMapPointsBenchmark.MapVector(Points: 256)	8.88 µs	7.25 µs	🟢 -18%
MatrixMapPointsBenchmark.MapPoints(Points: 4096)	2.10 µs	1.74 µs	🟢 -17%
ColorMathBenchmark.PreMultiply(Colors: 4096)	14.74 µs	12.27 µs	🟢 -17%
BitmapDrawBenchmark.DrawScaledTiles(Tiles: 256)	1.34 ms	1.12 ms	🟢 -16%
BitmapDrawBenchmark.DrawUnscaledTiles(Tiles: 256)	969.18 µs	811.22 µs	🟢 -16%
MatrixMapPointsBenchmark.MapVector(Points: 4096)	137.55 µs	115.40 µs	🟢 -16%
MatrixMapPointsBenchmark.MapRect(Points: 4096)	128.91 µs	109.03 µs	🟢 -15%
MatrixOpsBenchmark.Concat(Count: 4096)	256.01 µs	216.77 µs	🟢 -15%
CanvasDrawBenchmark.Draw(Shapes: 64)	2.65 ms	2.25 ms	🟢 -15%
…and 6 more

macOS

⏱️ Time (vs 🌙 nightly 4.152.0-nightly.10)

Benchmark	baseline	this PR	Δ
MatrixMapPointsBenchmark.MapVector(Points: 256)	2.43 µs	3.37 µs	🔴 +39%
MatrixMapPointsBenchmark.MapPoints(Points: 256)	135.9 ns	184.9 ns	🔴 +36%
MatrixMapPointsBenchmark.MapRect(Points: 256)	3.05 µs	4.08 µs	🔴 +34%
MatrixMapPointsBenchmark.MapPoints(Points: 4096)	1.52 µs	2.02 µs	🔴 +33%
RasterImageLifecycleBenchmark.CreateRasterImage(Count: 256)	141.56 µs	177.08 µs	🔴 +25%
BitmapDrawBenchmark.DrawUnscaledTiles(Tiles: 64)	148.63 µs	185.34 µs	🔴 +25%
PathBoundsBenchmark.TightBounds(Points: 1024)	8.07 µs	10.05 µs	🔴 +25%
BitmapDrawBenchmark.DrawUnscaledTiles(Tiles: 256)	543.43 µs	669.88 µs	🔴 +23%
BitmapDrawBenchmark.DrawScaledTiles(Tiles: 64)	195.71 µs	240.57 µs	🔴 +23%
ColorParseBenchmark.Parse(Iterations: 1000)	63.10 µs	77.00 µs	🔴 +22%
MatrixMapPointsBenchmark.MapRect(Points: 4096)	49.43 µs	59.64 µs	🔴 +21%
MatrixMapPointsBenchmark.MapRadius(Points: 256)	6.17 µs	7.42 µs	🔴 +20%
ColorMathBenchmark.ToColorF(Colors: 4096)	4.92 µs	5.89 µs	🔴 +20%
MatrixMapPointsBenchmark.MapVector(Points: 4096)	39.10 µs	46.52 µs	🔴 +19%
BitmapDrawBenchmark.DrawScaledTiles(Tiles: 256)	664.73 µs	790.70 µs	🔴 +19%
RasterImageLifecycleBenchmark.CreateDataWithReleaseProc(Count: 256)	66.42 µs	78.94 µs	🔴 +19%
ColorMathBenchmark.ToColor(Colors: 4096)	18.93 µs	22.37 µs	🔴 +18%
LargeImageScaleBenchmark.UpscaleOpaque(Size: 2048)	2.69 ms	3.17 ms	🔴 +18%
MatrixMapPointsBenchmark.MapPoint(Points: 4096)	18.14 µs	21.28 µs	🔴 +17%
CanvasDrawBenchmark.Draw(Shapes: 512)	11.08 ms	13.00 ms	🔴 +17%
MatrixOpsBenchmark.Invert(Count: 4096)	65.06 µs	75.88 µs	🔴 +17%
CanvasDrawBenchmark.Draw(Shapes: 64)	1.39 ms	1.61 ms	🔴 +16%
ColorMathBenchmark.PreMultiply(Colors: 4096)	8.48 µs	9.61 µs	🔴 +13%
PathBoundsBenchmark.TightBounds(Points: 64)	850.3 ns	961.5 ns	🔴 +13%
MatrixOpsBenchmark.Concat(Count: 4096)	111.66 µs	125.53 µs	🔴 +12%
…and 7 more

📈 Full interactive perf-dashboard & run details →

[github-actions]

📦 Artifact size report

Packages from this PR (build 1551049) vs the latest nightly baseline 4.152.0-nightly.17 (observed 2026-08-12).

Total .nupkg size: 569.8 MB → 569.2 MB (−625.9 KB, -0.1%)

Packages

⚠️ marks growth over 500.0 KB or 2%. Changes under 50.0 KB are treated as noise.

Package	baseline	this PR	Δ	Δ%
SkiaSharp.NativeAssets.WinUI	109.1 MB	108.4 MB	🟢 −784.0 KB	-0.7%
SkiaSharp.NativeAssets.WebAssembly	69.7 MB	69.7 MB	🔴 +78.9 KB	+0.1%

+38 package(s) unchanged (< 50.0 KB).

Per-file changes

SkiaSharp.NativeAssets.WinUI

File	Size
runtimes/win-arm64/native/libGLESv2.pdb	64.5 MB → 64.4 MB (🟢 −96.0 KB)
runtimes/win-x64/native/libGLESv2.pdb	66.6 MB → 66.5 MB (🟢 −80.0 KB)
runtimes/win-x86/native/libGLESv2.pdb	67.2 MB → 67.2 MB (🟢 −48.0 KB)

_{Informational only — this never blocks the PR. Native binaries are labelled by os/arch.}

[github-actions]

📖 Documentation Preview

The documentation for this PR has been deployed and is available at:

🔗 View Staging Site
🔗 View Staging Docs
🔗 View Staging Gallery (Blazor)
🔗 View Staging Gallery (Uno Platform)
🔗 View Staging SkiaFiddle

This preview will be updated automatically when you push new commits to this PR.

---

This comment is automatically updated by the documentation staging workflow.