fix(layout): apply the webcam layout only to clips that have a camera

[EtienneLescot]

Fixes #248.

The rule, in one line: the layout preset is global — one panel for the whole timeline — but the camera is per clip, so the preset must only apply to the clips that actually have one.

The Layout panel already got the first half of that right on release/v1.9.0: LayoutPane disables every control unless hasAnyClipWithCamera(assets, clips), so it's available as soon as one clip has a camera and greyed out when none does. That part needed no change. What was missing is the second half — the preset was being forced onto the clips that have no camera. Two independent bugs, one in each renderer.

1. The export drew the screen inside the PiP box

LiveParams::has_webcam already existed for this, but only live.rs derived it, so the live preview was right and every export was wrong:

• an export posts its LiveParams once for the whole timeline (compositor-view-napi), keeping the true default;
• ExportDialog sends the screen path as webcamPath for a camera-less clip, purely so the decoder has something valid to open.

So the box was drawn with the screen recording decoding behind it — the recording duplicated into its own corner. And because the Layout panel is (correctly) greyed out with no camera, the reporter had no way to switch the preset off. Hence "I had to set up my phone as a webcam to unlock the panel".

webcam_is_real moves from live.rs to frame_geometry.rs, next to the field it decides, and walk_composited_timeline — the walk shared by MP4 and GIF across all three backends — rebinds it per clip. A targeted set_has_webcam rather than a per-clip set_live_params, which would clobber whatever the caller already posted.

2. The block presets kept reshaping the screen — preview and export

Gating the camera's draw is not enough, and this is the half that actually shows. dual-frame and vertical-stack size the screen off the block: they reserve the camera's half of the frame. A camera-less clip therefore kept its screen squeezed into that half with nothing beside it, which no draw-time gate can undo. Under picture-in-picture it's invisible — the screen stays full-frame there and only a thumbnail is added.

layoutByClip already carried a resolved layout per visible clip; it simply never asked whether that clip had a camera. So a camera-less clip now lays out as if the preset were "no-webcam" — which computeCompositeLayout already implements (full-frame screen, no webcam rect). The predicate matches the webcamPath sent with the clip exactly, so the layout and the decoder cannot disagree. It is deliberately not hasAnyClipWithCamera, which gates the panel and ignores visible on purpose so the panel stays reachable to un-hide a camera.

PreviewCanvas had the same hole, and its own comment named it: it hid the webcam slot for a camera-less clip but left the screen geometry alone.

Verification

Export, real path (run_composited_multi → h264_amf), camera path equal to the screen path — the production case exactly: before, the screen recording sits in a rounded PiP box bottom-right; after, nothing is drawn.

poc-d3d --cfg C8 --export --fixture <dir> --webcam <dir>/screen.mp4 --out out/

The bench gains that --webcam override because the no-camera case is not different content but an identical path — my first attempt used two files with identical pixels and measured a 0 px diff, because webcam_is_real compares paths and was right to say "yes, that's a camera".

Layout, 5 new regression tests, all confirmed failing on the parent commit: one per preset for "only the clip that has a camera gets a webcam rect", and one per block preset for "the camera-less clip gets its full frame back". The existing webcamRect test asserted a PiP rect for an asset with no camera — that premise was the bug, so its fixture gains the camera its preset presupposes and goes on testing the px→fraction conversion it describes.

Confirmed by hand in the app on a mixed timeline (one imported clip with no camera + one recording with one) under side-by-side.

1622 JS tests pass (6 pre-existing failures in webm-seek-index.test.ts, unrelated — they fail identically on the parent commit); 122 compositor unit tests pass; tsc --noEmit, biome, and clippy clean on the changed lines.

Deliberately out of scope

• The ticket also asks for no-webcam to be the default preset when no camera is connected. Cosmetic now that both renderers are gated — the preset no longer changes anything for a camera-less clip — but the panel still reads misleadingly. Happy to add it.
• ExportDialog ignores cameraTrack.visible where sceneDescription honours it. Unreachable today: no UI path writes visible: false, and a recording made without a camera gets cameraTrack: null.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: c1f6d85f-d134-4569-81e9-4d3440d79f3b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}