macOS native app backend: follow-ups (guarantees, evidence, persistent helper, session backend)
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 15/100
- Issue 类型
- 功能
- 描述清晰度
- 基本清楚
- 活跃度
- 活跃
- 技术栈
- electron, macos, swift, typescript
- 领域
- desktop-dev, testing-qa, tooling
调研方向
Start with item 1: read ADR 0031, then trace resolvePressTarget → smallestElement in BackgroundInteraction.swift and the shared guards behind interaction-guarantees.ts. Fixture evidence lands in the macOS host lane in macos.yml (item 3); helper.ts and main.swift are to be split before the item 6 transport work, and MACOS_CLICK_* ↔ MouseClickSchedule.swift is the item 11 pin. Done for the umbrella means items 1–3 merged, each as its own PR naming this issue, with 4–12 done or split out.
由索引模型根据 Issue 内容生成。
描述
Purpose
Follow-ups for the opt-in macOS native app backend (AGENT_DEVICE_MACOS_APP_BACKEND=native): #3188 (helper app surface), #3189 (routing, admission facts, ADR 0031), #3195 (ghost cursor). The backend works: accessibility actions, no XCTest, the app may stay in the background. These items make it provably correct, evidence-backed, then faster and broader.
The plan was reviewed by Claude and by Codex (codex exec, read-only). The order below is the reconciled one: correctness and evidence come before persistence.
Work items (in order)
-
Guarantee classification and fallback correctness (M, no deps).
- Problem: when the hit test names no control,
resolvePressTarget→smallestElement(BackgroundInteraction.swift) may act on a different element than the snapshot node the shared guards evaluated. That contradicts ADR 0031's "ADR 0011 guarantees unchanged". Repeated presses also stop early and report fewer clicks without a typed outcome. - Required: add a row for this path to
interaction-guarantees.ts, and either bound the fallback to the target node's own subtree or report the acted element (role, label, frame, window) so the daemon can verify it. Partial outcomes are typed, never replayed automatically. ADR 0031 is updated.
- Problem: when the hit test names no control,
-
Snapshot quality and typed AX failures (S–M).
- Problem: the Chromium warning is free text, and it disappears once the flag is already on without re-checking population. AX attribute errors and permission problems become empty trees.
- Required: emit a typed
SnapshotQualityVerdict(as iOS does), plus typed accessibility and permission failures.
-
Native fixture lane (M, needs 1–2).
- Required: a fixture app (AppKit window plus WKWebView) in the
macos.ymlhost lane, run withAGENT_DEVICE_MACOS_APP_BACKEND=nativeset explicitly. - Done when the lane asserts, through production routing and with the app in the background: resulting app state, foreground and real pointer unchanged, typed refusals, and the delivery mechanism per action.
- Electron coverage is separate: WKWebView does not prove Electron.
- Required: a fixture app (AppKit window plus WKWebView) in the
-
Capture-frame contract (M). Snapshots and screenshots carry explicit frame metadata: window identity, origin, scale and display. Native snapshots traverse several windows while the screenshot captures one; XCTest subtracts the window origin at dispatch.
-
Keys, menu and value actions (S–M, needs 3). Separate increments, each with fixture evidence:
- key combos via process key events (today's evidence covers only text, Return and Tab);
AXShowMenuwhere a control supports it;- increment/decrement for sliders and steppers.
Not general double-click via
AXOpen, and not Cmd+[ as a universalback. -
Persistent helper transport (L, needs 3).
- Measure first, separately: process start, per-invocation source fingerprinting (
helper.ts), AX work and cursor time. - Then a JSON protocol over stdio following ADR 0002: fresh state per command; cancellation, serialization, and EOF/crash cleanup; bounded waits; behavior across TCC grant/revoke and binary replacement; cursor reset/hide and display changes.
- Split
helper.ts(install / transport / clients) andmain.swiftbefore adding behavior.
- Measure first, separately: process start, per-invocation source fingerprinting (
-
Element-identity dispatch (L, needs 1 and 6). Act on cached accessibility handles only with helper, app and snapshot-generation tokens, ref-epoch authorization, live validation and bounded retention. Raw
x ycommands keep hit-testing with window ownership. -
Session-scoped backend (L, independent design).
open --backend native|xctest, with the env var only as a default.- Admission and binding share one immutable execution context: admission inspects only
devicetoday, and bindings rebuild device facts. - Define sessionless
prepare, reopen and recovery semantics.
-
Native recording (L). A ScreenCaptureKit window stream as its own lifecycle, replacing the runner refusal.
-
Helper tree presentation (M). Collapse only proven structural wrapper groups (Chromium), keeping raw topology, selector meaning and clip ownership. Use a reproducible fixture, not ad-hoc counts.
-
Parity cleanup (S).
- Pin the click-schedule constants (
MACOS_CLICK_*↔MouseClickSchedule.swift) with a fixture. - Keep scroll travel on the shared golden table rather than fetching a frame and sending pixels.
- Pin the click-schedule constants (
-
Framework matrix and default decision (L, last). Cover AppKit, SwiftUI, Catalyst, Electron and WebKit, including unsupported actions and sparse apps. Native stays opt-in until this evidence exists.
-
Pointer events with window fields (M, needs 1 and 3).
- Evidence (2026-10-05, measurements in the comments): public
CGEventPostToPidwith the target window number in event field 51, plus field 58 or a window-local location, delivers click, double-click and secondary click to a background AppKit, SwiftUI and WKWebView fixture: 54 of 54, no activation, frontmost app unchanged. A bare post, the ADR 0031 measurement, delivers nothing. Posting through SkyLight adds nothing. Drags and WKWebView button clicks need the private focus-without-raise activation. - Required: an event path for double-click, secondary click and points with no accessibility action. Success needs an observed change; otherwise the outcome is a typed unverified result. The path gets its own row in
interaction-guarantees.tsand fixture-lane evidence (item 3), including Electron and Catalyst. Fields 51 and 58 are undocumented, so the fixture pins them and a macOS change fails the lane, not users. - Not: SkyLight posting or focus-without-raise without a separate decision.
- Evidence (2026-10-05, measurements in the comments): public
Non-goals
- A single
{facts, interactor, snapshotRoute}facade: the parts have different lifetimes. - Turning
frontmost-app/menubarpresses into accessibility-only presses before their surface guarantees are defined. Synthetic menu extras need pointer delivery.
Completion
Each item lands as its own PR that names this issue. The umbrella closes when items 1–3 have landed, and items 4–13 are either done or split into their own issues.
- 主要语言
- TypeScript
- 星标
- 4.9k
- 派生
- 328
- 平均合并
- 11 小时 51 分钟
- 30 天内合并 PR
- 535
环境准备
- 没有 Dockerfile 或 Docker Compose 文件
- 没有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
callstack/agent-device 的其他 Issue
-
settings clear-app-state fails with ENOENT … scandir '(null)' for iOS system apps such as Settings未关闭
难度 2/5 1-3 小时 新手友好度 80/100
callstack/agent-device#3305 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 76/100
callstack/agent-device#3296 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 68/100
callstack/agent-device#1869 ·
维护者通常 1 天内回复
-
needs-triage refactor
难度 5/5 一周以上 新手友好度 35/100
callstack/agent-device#3311 ·
维护者通常 1 天内回复
-
refactor(provider-webdriver): ship AWS Device Farm as the `@agent-device/aws-device-farm` plugin未关闭needs-triage refactor
难度 4/5 3-5 天 新手友好度 54/100
callstack/agent-device#3310 ·
维护者通常 1 天内回复
查看 callstack/agent-device 的全部 Issue
相似的 Issue
-
难度 2/5 1-3 小时 新手友好度 85/100
farbenmeer/tapi#531 ·
-
难度 2/5 1-3 小时 新手友好度 75/100
naver/egjs-flicking#971 ·
-
难度 1/5 1 小时以内 新手友好度 85/100
维护者通常 1 天内回复
-
Tenant
难度 2/5 1-3 小时 新手友好度 66/100
MTES-MCT/Dossier-Facile-Frontend#2061 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 85/100
backnotprop/plannotator#1784 ·
维护者通常 1 天内回复