Ourin Extensions and Behavioral Notes

This document summarizes non-standard or platform-specific behavior in Ourin.

Platform focus

Ourin is a macOS-native baseware. Several behaviors intentionally use AppKit/SwiftUI mechanics over strict SSP parity.

Examples:

• Window control and animation via NSWindow APIs
• Visual effects/state modeled through SwiftUI view models
• Native settings/open flows for browser, mail, dialogs

SakuraScript execution model

Parser and executor are separated:

• SakuraScriptEngine tokenizes
• GhostManager interprets tokens and enqueues playback units

This supports typing effects, deferred commands, and command timing control.

Extended move command handling

move/moveasync supports option-style forms:

• --X, --Y
• --time
• --base, --base-offset, --move-offset
• --option=ignore-sticky-window
• moveasync,cancel (with optional scope)

Positional legacy form is kept for compatibility.

Window group control extensions

• set,zorder accepts broader scope-token parsing (current, all, delimited lists)
• set,sticky-window supports grouped forms (including --group=...|...)

These are practical extensions for multi-character window control on macOS desktops.

Dressup/bind handling

bind supports:

• category form
• repeated tuple form
• disable/unbind semantic values (0, false, off, none, default)

Rendering is still partial; binding state is maintained for future full compositor behavior.

SERIKO coexistence model

Ourin currently keeps:

• legacy AnimationEngine path
• new SerikoParser + SerikoExecutor path

Runtime uses executor first and can fall back to legacy engine when needed.

SAORI host bridge

YAYA plugin operations are bridged to Swift host via JSON host operations:

• saori_load
• saori_unload
• saori_request

This is an Ourin architecture choice for stable Rust/C++/Swift boundary handling.

Known compatibility notes

• Some SakuraScript subcommands are partially implemented (documented in SUPPORTED_SAKURA_SCRIPT.md)
• Full SSP-level behavior parity is still in progress for advanced animation/dressup combinations
• Existing test suite includes unrelated historical failures; feature validation is often done with focused test targets