Request: allow adjusting the screen recording positioning, placing it off-center in the final video

[boweiliu]

I made my LLM agent take a shot at this: #1959

[linear]

CAP-745

[tembo]

The linked PR #1959 looks like a solid approach to this feature. For anyone reviewing or iterating on it, here's some relevant codebase context:

Key files for screen positioning logic:

• crates/rendering/src/lib.rs - The display_layout function (previously display_offset) calculates where the screen recording appears within the output frame. The PR refactors this to return both offset and size, which is the right pattern since offset becomes asymmetric once positioning is allowed.
• crates/rendering/src/coord.rs - Coordinate space transformations that depend on display_offset. The PR correctly updates to_frame_space to use the new layout calculation.
• crates/project/src/configuration.rs - The BackgroundConfiguration struct where position_offset_x and position_offset_y are added alongside existing padding.

Implementation note from the PR's own review:
The current math uses (offset / 100.0) * room where room is the available space after fitting the video. This means the effective slider range is [-50, +50] to move from edge to edge - values beyond that clamp to no-ops. If the intended UX is for [-100, +100] to represent the full range, the divisor should be 200.0 instead. Worth confirming the desired behavior before merging.

Existing patterns to follow:
The PR correctly follows the existing pattern for background settings - adding fields to BackgroundConfiguration with defaults, then wiring them through the rendering pipeline and exposing them in both ConfigSidebar.tsx (video editor) and BackgroundSettingsPopover.tsx (screenshot editor).