[antonylavelle]

Description

This gets Flash 10 display-list 3D (rotationX / rotationY / z) drawing and hit-testing close enough that Microboats plays correctly. That Armor Games racer is the biggest example of Flash 2.5D I know of: a 2D camera look-at, then the whole world pitched with rotationX, plus standing trackside props (rotationX = 90).

What changed:

• Compose local matrix3D as Flash Euler T * Rz * Ry * Rx * S, and concatenate that through the transform stack.
• Scale world z with the viewport (HiDPI / ShowAll) so perspective uses stage-pixel focal length instead of looking zoomed out.
• Skip 2D AABB render culling when a transform uses 3D, so pitched tiles are not dropped while they are still on screen.
• Keep DisplayObject.rotation in [-180, 180] for 2D, but unwrapped once matrix3D is active, so heading lerps (Microboats boats / chase cam) do not reverse at ±180°.
• Union the perspective-projected silhouette into world_bounds / hitTestObject, so Microboats' onScreen (hitTestObject vs a stage rectangle) still shows obstacles that sit off the 2D stage after camera pitch.
• Setting rotationZ now enters display-list 3D the same way rotationX does.

Assisted-by: Cursor.

Testing

• Played Microboats (SWF10, 600×440) in the macOS desktop player: camera pitch, trackside 3D, heading through 180°, and in-track obstacles.
• Added tests/tests/swfs/avm2/displayobject_3d (Asc). It traces 2D vs 3D rotation wrapping, Euler matrix3D order, and hitTestObject for a clip at y = 500 after rotationX = -60.
• Caveat: output.txt is from Ruffle, not yet captured from the Flash Player debugger. That should be re-run in FP and replaced if any line differs (especially 3D rotation wrapping and hitTestObject).

Checklist

• I, a human, have self-reviewed this PR and fully understand the changes within.
• I have made or updated tests where possible.
• All of my commits are properly scoped, compile successfully, and pass all tests.
• This PR does not make sense to split up into smaller PRs.
• An LLM was involved in the authoring of this code.

[Lord-McSweeney]

This PR does not make sense to split up into smaller PRs.

This includes a lot of changes, please split it into smaller PRs. e.g. one to support hook up rotationX/rotationY/rotationZ to Matrix3D, one to support rendering Matrix3D, etc.

[antonylavelle]

Ok, I'll have to revisit it later to split and simplify each one into something smaller. Nearly all of this came from my seperate project on getting KOLM's 3D working in Ruffle, Microboats was just a huge test case for it so I thought it would be more useful.

…

On Wed, Nov 4, 2026 at 9:57 PM Lord-McSweeney ***@***.***> wrote: *Lord-McSweeney* left a comment (ruffle-rs/ruffle#25101) <#25101 (comment)> This PR does not make sense to split up into smaller PRs. This includes a lot of changes, please split it into smaller PRs. e.g. one to support hook up rotationX/rotationY/rotationZ to Matrix3D, one to support rendering Matrix3D, etc. — Reply to this email directly, view it on GitHub <#25101?email_source=notifications&email_token=ACHIUG2E4OZNF5BTUSFKPQ35XOQ5NA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTMNRUGMYTINRWGY3KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-6643146666>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACHIUG7HWTTROV5NNWVNAAD5XOQ5NAVCNFSNUABFKJSXA33TNF2G64TZHMYTQMZUHAZTSOBYHNEXG43VMU5TMMRRGIYDMNRXGIZ2C5QC> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/ACHIUGYUJRLZKPHBYP4FDTT5XOQ5NA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTMNRUGMYTINRWGY3KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KUZTPN52GK4S7NFXXG> and Android <https://github.com/notifications/mobile/android/ACHIUG633G4HPEANN47ZA3T5XOQ5NA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTMNRUGMYTINRWGY3KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>. Download it today! You are receiving this because you authored the thread.Message ID: ***@***.***>