Bump Babylon.js to 9.21.2 and fix i_data slot overlap by bkaradzic-microsoft · Pull Request #1852 · BabylonJS/BabylonNative

bkaradzic-microsoft · 2026-08-25T23:25:26Z

Bumps Babylon.js 9.15.0 -> 9.21.2, and fixes the i_data slot collision that bump exposes.

The bug

AssignBuiltInInstanceSlots numbered the built-in per-instance attributes a draw did not record starting at zero -- the same range the draw-time caller slots use -- and threw when the two collided.

OpenGL/Metal route built-ins through the caller map, so a shader declaring world0-3 and previousWorld0-3 whose first draw records only world0-3 always collides (_renderWithThinInstances creates the previousWorld buffer after the first draw). That is Thin instances + render self motion blur failing on the four Ubuntu legs. D3D keeps built-ins out of that map, so it only collides once a draw mixes a generic instanced attribute with unrecorded built-ins.

Fix: give the built-ins the lowest slots not already claimed by a caller-supplied location, and size the instance data buffer from the recorded attributes plus the built-ins the draw left out, so those land in the buffer's zero padding.

Introduced by #1839, latent on master only because the lockfile pinned 9.15.0.

Supporting commits

Restore the original basic type when narrowing widened uniforms is a shader-compiler fix required by 9.21.2. The rest enable 9 newly-passing tests, downlevel the bundles to ES5 for Chakra (typescript moves to dependencies because the postinstall hook needs it under --omit=dev), and retarget the prepass SSAO / SSR2 exclusions.

Those eight were already globally excluded on master for a Mesa/LLVM crash on the Ubuntu runner; they are now excludedGraphicsApis: ["OpenGL"] so D3D11 keeps covering them. The sweep only runs on Linux/GL and Win32/D3D11, so this enables them nowhere untested.

Testing

	before	after
Linux/OpenGL sweep	288/289	289/289
Win32 D3D11 sweep	306/306	314/314

Copilot

Pull request overview

Warning

Copilot couldn't run its full agentic review because it didn't start before the timeout. Make sure your repository has a runner available, or add a copilot-code-review.yml file specifying one with the runs-on attribute. See the docs for more details.

Fixes i_data slot collisions between caller-supplied instance attributes and shader built-in instance attributes when a draw records only a subset of built-ins (notably affecting GL/Metal thin instances + motion blur).

Changes:

Update built-in instance slot assignment to choose the lowest unclaimed i_data slots rather than starting from 0.
Expand instance data buffer coverage to include recorded instance attributes plus unrecorded built-ins so D3D11 input layouts remain valid.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
Plugins/ShaderCompiler/Source/ShaderCompilerTraversers.cpp	Assigns missing built-in instance attributes to truly free i_data slots, avoiding overlap with caller-packed slots.
Plugins/NativeEngine/Source/NativeEngine.cpp	Computes required i_data span using recorded instance attributes plus missing built-ins to avoid D3D11 layout rejection.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

Clustered lighting, the FrameGraph tests and several prepass/SSAO tests construct scenes against engine APIs added after 9.15.0 and throw before any shader is compiled, so they cannot be validated on the pinned version.

Copilot

Pull request overview

Copilot reviewed 4 out of 5 changed files in this pull request and generated 2 comments.

Files not reviewed (1)

Apps/package-lock.json: Generated file

bgfx exposes every non-sampler uniform as a float vec4, so ChangeUniformTypes retypes each loose scalar/vector uniform to vec4 and inserts a shape conversion at every use site to narrow it back. addShapeConversion only reconciles vector *shape*, not basic type. Given a vec4 source and an int target it builds an EOpConstructInt aggregate whose operand is still a float and calls it done, so the AST claims "int" while holding a float. glslang faithfully lowers that to SPIR-V in which an integer operation consumes a float, and SPIRV-Cross renders it verbatim. For ssao2's blur pass, which declares `uniform int samples` and loops `for (int i = -samples; i < samples; i += 2)`, the generated ESSL 300 was: uniform vec4 samples; ... mediump int _42 = -samples.x; for (mediump int i = _42; i < samples.x; i += 2) GLES rejects both lines, bgfx raises Fatal::InvalidShader, and the validation run aborts. HLSL happened to tolerate it because it converts implicitly, which is why this only ever showed up on the OpenGL and OpenGLES backends. Shape the value to float first and then ask glslang for a real conversion node, which produces the expected `int(...)` and correctly typed SPIR-V. Validated on Windows: D3D11 is unchanged at 351/351, and on an ANGLE/GLES build eleven tests go from Fatal::InvalidShader to passing - the whole prepass SSAO family (287, 289, 293, 296, 299, 302, 304, 305, 306) plus GUI Slate (176) and GUI Near Menu (177). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 60c2ec68-6de1-445d-9fc9-b699db737eae

These were excluded against Babylon.js 9.15.0 and now pass. Verified on Win32 D3D11 both per-test in isolation and in a single sequential process covering indices 56-719 (ran=256 passed=256 failed=0), since several of these exclusions describe order-dependent behaviour that a per-test sweep cannot reproduce. 137 Volumetric Light Scattering Post Process with Morph Targets exact 287 Prepass SSAO + particles exact 289 Prepass SSAO + instanced bones 0.012% 299 Prepass SSAO + GUI 1.064% 302 Prepass SSAO + highlight layer 0.018% 304 Prepass SSAO + on/off post-process exact 305 Prepass SSAO + thin instances 0.003% 306 Prepass SSAO + depth renderer 0.044% 363 Screen Space Reflections 2 1.315% Three carried backend-specific exclusion reasons that cannot be reproduced on a D3D11 host, so CI is the arbiter for them: 137 "Pixel comparison fails on Linux (large diff)" 287 "SSAO2 blur post-process shader fails to compile on desktop GL" 299 OpenGL "mediump float" compile failure in PrePassRenderer, plus an order-dependent state leak that produced a ~6000 px diff right at the 2.5% threshold; it now measures 1.064% in sequential order. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 60c2ec68-6de1-445d-9fc9-b699db737eae

With the uniform basic-type fix in place the ssao2 blur shader compiles on OpenGL, so the Ubuntu jobs now get past shader compilation for this test and reach the particle shader instead, where llvmpipe's JIT aborts the process: BGFX attr a_tangent: 2 BGFX instance data i_data0: 4 ... LLVM ERROR: Cannot emit physreg copy instruction --- BN: ABORT --- SIGABRT raised. All four Ubuntu jobs stop at exactly the same place, immediately after linking the particle program (diffuseSampler / textureMask / i_data0-3), so this is not the SSAO shader and not something we emit differently - it is a Mesa software rasterizer codegen bug on the CI runner. The test itself passes on D3D11 and on an ANGLE/GLES build. The other eight tests enabled by the Babylon.js 9.21.2 bump stay enabled. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 60c2ec68-6de1-445d-9fc9-b699db737eae

With the uniform basic-type fix in place these shaders compile on OpenGL, but the Ubuntu runner's llvmpipe then aborts the process while JIT-compiling them: LLVM ERROR: Cannot emit physreg copy instruction --- BN: ABORT --- SIGABRT raised. The abort moved from 287 to 289 once 287 was skipped, so it follows the shader family rather than any one scene. All four Ubuntu jobs stop at byte-identical positions, and this is a register allocator failure inside Mesa's software rasterizer, not something Babylon Native can emit its way around. The family already had a history of this: 288, 290, 291 and 292 were excluded on master with "Test crashes or hangs on Babylon Native". These tests do pass on D3D11 and, with this PR's shader fixes, on an ANGLE/GLES build, which is recorded in each reason so the exclusion is understood as CI-runner-specific rather than a statement about the renderer. Tests 137 and 363 remain enabled. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 60c2ec68-6de1-445d-9fc9-b699db737eae

GUI Near Menu was excluded from OpenGL with this recorded reason: OpenGL: BGFX FATAL shader compile error in GUI fragment shader ('=' : cannot convert from 'highp float' to 'bool') That is precisely the bug fixed earlier in this branch: a bool uniform widened to a float vec4 was narrowed back by shape alone, so the AST claimed bool while holding a float. The test now compiles and passes on OpenGL, so the graphics API exclusion is removed and it runs everywhere. Screen Space Reflections 2 is excluded instead. It reaches a multiple-render- target resolve, where bgfx does: const GLenum drawBuffer = GL_COLOR_ATTACHMENT0 + colorIdx; GL_CHECK(glDrawBuffers(1, &drawBuffer) ); GLES requires bufs[i] to be GL_NONE or GL_COLOR_ATTACHMENT0 + i, so resolving any attachment past the first is GL_INVALID_OPERATION. Desktop GL uses glDrawBuffer just above and has no such restriction, but Babylon Native builds the ES path on Linux, so the Ubuntu jobs hit it. The test passes on D3D11. A full sequential run now completes on an ANGLE/GLES build with no assert and no BGFX FATAL: 286 ran, 282 passed. The four remaining reds are the three motion blur tests that BabylonJS#1839 fixes, plus one ANGLE-only pixel difference in MeshDebugPluginMaterial. D3D11 is unchanged at 297/300, red on the same three motion blur tests. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 60c2ec68-6de1-445d-9fc9-b699db737eae

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

Copilot

Pull request overview

Copilot reviewed 4 out of 5 changed files in this pull request and generated 3 comments.

Files not reviewed (1)

Apps/package-lock.json: Generated file

Suppressed comments (6)

Apps/Playground/Scripts/config.json:1858

This global exclusion also skips the test on D3D11 (and every other backend), even though the updated reason says the failure is specific to llvmpipe/OpenGL and that D3D11 passes. Use the existing API-specific exclusion so passing backends retain coverage.

            "excludeFromAutomaticTesting": true,

Apps/Playground/Scripts/config.json:1938

This global exclusion also skips the test on D3D11 (and every other backend), even though the updated reason says the failure is specific to llvmpipe/OpenGL and that D3D11 passes. Use the existing API-specific exclusion so passing backends retain coverage.

            "excludeFromAutomaticTesting": true,

Apps/Playground/Scripts/config.json:1962

This global exclusion also skips the test on D3D11 (and every other backend), even though the updated reason says the failure is specific to llvmpipe/OpenGL and that D3D11 passes. Use the existing API-specific exclusion so passing backends retain coverage.

            "excludeFromAutomaticTesting": true,

Apps/Playground/Scripts/config.json:1978

This global exclusion also skips the test on D3D11 (and every other backend), even though the updated reason says the failure is specific to llvmpipe/OpenGL and that D3D11 passes. Use the existing API-specific exclusion so passing backends retain coverage.

            "excludeFromAutomaticTesting": true,

Apps/Playground/Scripts/config.json:1986

This global exclusion also skips the test on D3D11 (and every other backend), even though the updated reason says the failure is specific to llvmpipe/OpenGL and that D3D11 passes. Use the existing API-specific exclusion so passing backends retain coverage.

            "excludeFromAutomaticTesting": true,

Apps/Playground/Scripts/config.json:1994

This global exclusion also skips the test on D3D11 (and every other backend), even though the updated reason says the failure is specific to llvmpipe/OpenGL and that D3D11 passes. Use the existing API-specific exclusion so passing backends retain coverage.

            "excludeFromAutomaticTesting": true,

`Win32_x64_D3D11` and `Win32_x64_D3D11_Sanitizers` have hit the one-hour job timeout on every run of this PR. Both are the jobs that build the default JavaScript engine, Chakra; the V8, Hermes, JSI and QuickJS jobs on the same matrix all pass. The hang is reproducible locally with a Chakra build and is deterministic: `Playground --headless --test-index=23` never returns, spinning one core and growing the heap by ~7 MB/s until the runner gives up. Test 23 is "Glow layer and LODs". The scene itself renders fine - what never completes is `Scene.executeWhenReady`, because `EffectLayer.isReady` stays false forever: frame 300 scene.isReady=false layer.isReady(subMesh)=false _shadersLoaded=false isLayerReady=false while the same probe on V8 flips everything to true by frame 50. The one link that never settles is `ThinGlowLayer._importShadersAsync()`. The cause is a Chakra bug, exposed by a Babylon.js code-generation change. `super.x` inside an arrow function nested in a class method resolves to the *derived* class's own method instead of the base: class A { foo() { return "BASE"; } } class B extends A { foo() { const s = Object.create(null, { foo: { get: () => super.foo } }); return s.foo.call(this); // V8: "BASE" Chakra: recurses } } TypeScript emits exactly that `Object.create(null, { get: () => super.x })` helper for a `super` call inside an `async` method, and Babylon.js started shipping it in the UMD bundle in 9.16.0 - which is precisely where this PR's bump to 9.21.2 crosses. Called synchronously it dies with "Out of stack space"; called from a promise chain, as `_importShadersAsync` is, each level is a fresh microtask, so it recurses forever without ever overflowing the stack, never settles, and burns CPU and memory - the exact signature seen on the runner. The repo already has the remedy. `scripts/downlevelNativeScripts.mjs` transpiles the bundles to ES5 for this very reason ("Babylon Native's Chakra engine consumes ES5-level script"); TypeScript's ES5 emit rewrites `super.x` to `_super.prototype.x` and drops the arrow entirely. It was only ever wired into `getNightly`, so builds that take Babylon.js from npm - which is every normal build - ran the un-downleveled ES2015 bundle. Running it from `postinstall` closes that gap for `npm install`, `npm ci`, CI and local builds alike, and leaves the nightly path alone (`getNightly.js` still downlevels the files it refills from the CDN). Validated on Windows/D3D11, Debug, tests 0-52 and 56-719 (53-55 crash locally in Debug regardless of this change): | engine | before | after | |---|---|---| | Chakra | hangs at test 23 | **297/300** | | V8 | 297/300 | **297/300** | Byte-identical results on V8, and Chakra now matches it. The three remaining failures are the motion-blur trio that BabylonJS#1839 fixes. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 60c2ec68-6de1-445d-9fc9-b699db737eae

Copilot

Pull request overview

Copilot reviewed 4 out of 5 changed files in this pull request and generated 1 comment.

Files not reviewed (1)

Apps/package-lock.json: Generated file

bghgary

[Reviewed by Copilot on behalf of @bghgary]

LGTM after the open maxInstanceData thread is addressed.

Copilot

Pull request overview

Copilot reviewed 5 out of 6 changed files in this pull request and generated 2 comments.

Files not reviewed (1)

Apps/package-lock.json: Generated file

bghgary · 2026-08-26T23:28:31Z

[Commented by Copilot on behalf of @bghgary]

The implementation now carries the compiler-assigned built-in slot map into runtime and packs recorded attributes into their exact slots. The PR description still says it chooses the lowest unclaimed slots and relies on zero padding, which describes the superseded implementation. Please update the Fix paragraph before merge.

…ce attributes AssignBuiltInInstanceSlots numbered the built-ins the draw did not record from zero, the same range the draw-time caller slots use, and threw when the two collided. On OpenGL/Metal the built-ins are routed through the caller map, so a shader declaring world0-3 and previousWorld0-3 whose first draw records only world0-3 always collided -- this is the "Thin instances + render self motion blur" failure on the Linux CI legs since Babylon.js 9.21.2. Assign the lowest free slots instead, and size the instance data buffer from the recorded attribute count plus the built-ins the draw left out, so those land in the buffer's zero padding on both backends. Verified: Linux/OpenGL validation sweep 289/289 (was 288/289), Win32 D3D11 306/306. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 88569c10-a7ff-4373-9a58-afa9c68b8c09

The reasons on these eight say the failure is specific to the OpenGL CI runner and that D3D11 passes, but the global excludeFromAutomaticTesting flag skipped every backend. Narrow them to excludedGraphicsApis: ["OpenGL"]; the sweep runs only on Linux/GL and Win32/D3D11, so this restores D3D11 coverage without enabling them anywhere untested. Win32 D3D11 sweep: 314/314 (was 306/306). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 88569c10-a7ff-4373-9a58-afa9c68b8c09

Bring in i_data slot-overlap fix (BabylonJS#1852) and JsRuntimeHost pin (BabylonJS#1854 content via shotgun fork). Keep shotgun JsRuntimeHost/SPIRV-Cross forks, fragCoord Y flip, GPU instance repack, compute shader path, and playground enables. Resolve config by preserving shotgun enablements and taking upstream OpenGL CI exclusions where applicable. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 60c2ec68-6de1-445d-9fc9-b699db737eae

Copilot AI lite review requested due to automatic review settings August 25, 2026 23:25

Copilot AI reviewed Aug 25, 2026

View reviewed changes

Comment thread Plugins/ShaderCompiler/Source/ShaderCompilerTraversers.cpp Outdated

Comment thread Plugins/NativeEngine/Source/NativeEngine.cpp Outdated

Comment thread Plugins/NativeEngine/Source/NativeEngine.cpp Outdated

Copilot started reviewing on behalf of bkaradzic-microsoft August 25, 2026 23:44 View session

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch 2 times, most recently from 35e0059 to eadcb87 Compare August 25, 2026 23:58

Bump Babylon.js from 9.15.0 to 9.21.2

54ea2f0

Clustered lighting, the FrameGraph tests and several prepass/SSAO tests construct scenes against engine APIs added after 9.15.0 and throw before any shader is compiled, so they cannot be validated on the pinned version.

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch from eadcb87 to d09fac1 Compare August 26, 2026 00:20

bkaradzic-microsoft changed the title ~~Fix i_data slot overlap when a draw records only some built-in instance attributes~~ Bump Babylon.js to 9.21.2 and fix i_data slot overlap Aug 26, 2026

bkaradzic-microsoft mentioned this pull request Aug 26, 2026

Correct gl_FragCoord Y orientation on D3D, Metal and Vulkan #1840

Open

bkaradzic-microsoft requested a balanced review from Copilot August 26, 2026 00:50

Copilot AI reviewed Aug 26, 2026

View reviewed changes

Comment thread Plugins/ShaderCompiler/Source/ShaderCompilerTraversers.cpp

Comment thread Plugins/ShaderCompiler/Source/ShaderCompilerTraversers.cpp Outdated

Copilot started reviewing on behalf of bkaradzic-microsoft August 26, 2026 01:09 View session

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch 2 times, most recently from 60742d1 to 4379a7d Compare August 26, 2026 14:23

bkaradzic and others added 5 commits August 26, 2026 07:59

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch from 4379a7d to dc3c8f2 Compare August 26, 2026 15:00

bkaradzic-microsoft requested review from CedricGuillemet and bghgary and a balanced review from Copilot August 26, 2026 15:17

Copilot AI reviewed Aug 26, 2026

bkaradzic-microsoft requested a balanced review from Copilot August 26, 2026 17:17

Copilot started reviewing on behalf of bkaradzic-microsoft August 26, 2026 17:18 View session

Copilot AI reviewed Aug 26, 2026

View reviewed changes

Comment thread Apps/Playground/Scripts/config.json Outdated

Comment thread Apps/Playground/Scripts/config.json Outdated

Comment thread Apps/package.json

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch from dc3c8f2 to 0864f3d Compare August 26, 2026 18:31

bkaradzic-microsoft requested a balanced review from Copilot August 26, 2026 18:32

Copilot started reviewing on behalf of bkaradzic-microsoft August 26, 2026 18:41 View session

Copilot AI reviewed Aug 26, 2026

View reviewed changes

Comment thread Plugins/NativeEngine/Source/NativeEngine.cpp Outdated

bghgary approved these changes Aug 26, 2026

View reviewed changes

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch from 0864f3d to 934d42c Compare August 26, 2026 21:42

bkaradzic-microsoft requested a balanced review from Copilot August 26, 2026 21:43

Copilot started reviewing on behalf of bkaradzic-microsoft August 26, 2026 21:47 View session

Copilot AI reviewed Aug 26, 2026

View reviewed changes

Comment thread Plugins/ShaderCompiler/Source/ShaderCompilerTraversers.cpp Outdated

Comment thread Apps/Playground/Scripts/config.json

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch from 934d42c to a6f4866 Compare August 26, 2026 23:00

bkaradzic-microsoft enabled auto-merge (squash) August 26, 2026 23:13

bkaradzic and others added 2 commits August 26, 2026 17:19

bkaradzic-microsoft force-pushed the fix/i-data-slot-overlap branch from a6f4866 to c722274 Compare August 27, 2026 00:19

bkaradzic-microsoft merged commit 6e9afae into BabylonJS:master Aug 27, 2026
34 checks passed

Uh oh!

Conversation

bkaradzic-microsoft commented Aug 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

The bug

Supporting commits

Testing

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

bghgary left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

bghgary commented Aug 26, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

bkaradzic-microsoft commented Aug 25, 2026 •

edited

Loading