@unotest/mobile 0.9.1 → 0.10.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (2) hide show
  1. package/CHANGELOG.md +31 -15
  2. package/package.json +5 -5
package/CHANGELOG.md CHANGED
@@ -4,6 +4,22 @@ All notable changes to `@unotest/mobile` will be documented in this file.
4
4
 
5
5
  The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
6
6
 
7
+ ## [0.10.0] - 2026-08-06
8
+
9
+ ### Patch Changes
10
+
11
+ - Updated dependencies [7ed750c]
12
+ - Updated dependencies [574841f]
13
+ - Updated dependencies [337a31c]
14
+ - @unotest/dsl@0.10.0
15
+ - @unotest/protocol@0.10.0
16
+ - @unotest/core@0.10.0
17
+ - @unotest/viewer@0.10.0
18
+
19
+ ## [0.9.2] - 2026-06-16
20
+
21
+ Released in lockstep with the rest of the ecosystem. No functional changes.
22
+
7
23
  ## [0.9.1] - 2026-06-14
8
24
 
9
25
  ### Changed
@@ -62,8 +78,8 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
62
78
  all running on the same Mac), the iOS keyboard layer couldn't keep
63
79
  up and randomly dropped 1-3 characters from the middle of the string
64
80
  — observed live as `petr@volkov.io` → `pr@volkov.io`, `Qwerty34##`
65
- → `Qwery34##`. P4 / S8 had only addressed the unrelated *wrong-
66
- keyboard-layout* failure (Cyrillic keyboard eating Latin chars
81
+ → `Qwery34##`. P4 / S8 had only addressed the unrelated _wrong-
82
+ keyboard-layout_ failure (Cyrillic keyboard eating Latin chars
67
83
  wholesale); this race is a separate failure mode that survived.
68
84
  - **Fix 1 — `ConnectHardwareKeyboard` pinned OFF on install.**
69
85
  The `pinEnglishKeyboard` step (S8) now also writes
@@ -113,7 +129,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
113
129
  `npx @unotest/mobile@latest install ...`. The stale global frequently
114
130
  had the pre-0.1.4 strict env schema, so the MCP server crashed at
115
131
  startup with `Invalid environment. ... INVITE_DEEPLINK_PREFIX:
116
- Required, API_BASE_URL: Required, ...`. Claude Code then surfaced
132
+ Required, API_BASE_URL: Required, ...`. Claude Code then surfaced
117
133
  `-32000 Connection closed` and the agent fell back to draft-mode
118
134
  scenario authoring — defeating the entire MCP-driven flow the
119
135
  install just bootstrapped.
@@ -173,7 +189,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
173
189
  `result.detectedPermissions`.
174
190
  - **Pre-grant on install.** When the resolved permissions list is
175
191
  non-empty, every service is granted with `simctl privacy <udid> grant
176
- <service> <bundleId>` immediately after install. Pre-empts SpringBoard
192
+ <service> <bundleId>` immediately after install. Pre-empts SpringBoard
177
193
  permission dialogs (location, motion, photos, …) that would otherwise
178
194
  block scenarios on first launch — those dialogs live in SpringBoard,
179
195
  not in the app's a11y tree.
@@ -213,7 +229,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
213
229
  - **New `alert:` section in `mode: "outline"`.** When a native
214
230
  `UIAlertController` is on screen its title, body, and button labels
215
231
  appear as the first section of the outline, plus a `_meta.alert_active:
216
- true` flag. Previously the alert lived in SpringBoard (outside the
232
+ true` flag. Previously the alert lived in SpringBoard (outside the
217
233
  app's a11y tree) and the agent kept tapping the underlying button
218
234
  through the modal — usually the wrong one, sometimes destructive.
219
235
  - **Grammar contract.** The `alert:` section and `_meta.alert_active`
@@ -267,7 +283,7 @@ After the fixes, `run_test` on a freshly-generated exploration returned
267
283
  is unchanged — keychain reset only fires when the caller explicitly
268
284
  opts into a fresh state.
269
285
  - Keychain reset is invoked **unconditionally** in the `app_install
270
- { clean: true }` flow even when `simctl uninstall` reports "app not
286
+ { clean: true }` flow even when `simctl uninstall` reports "app not
271
287
  installed" — prior-run tokens may still linger, that's the whole
272
288
  reason for B5.
273
289
  - MCP descriptions for `app_install`'s `clean` parameter and
@@ -283,7 +299,7 @@ After the fixes, `run_test` on a freshly-generated exploration returned
283
299
  one tool — `explore_step { action: "tap" | "type" | … }`. The single
284
300
  rule: passing `explorationId` records the call into an exploration
285
301
  session; omitting it runs the action ad-hoc. There is no `record:
286
- false` override.
302
+ false` override.
287
303
  - **Alerts (`accept_alert`, `dismiss_alert`) appear here for the first
288
304
  time as `explore_step` actions** — not as separate tools. The
289
305
  Driver+DSL alert layer landed in P0; this release wires the MCP
@@ -384,12 +400,12 @@ unchanged.
384
400
  on UIAlertController, last on action sheet).
385
401
  - `readAlert()` — return title + body of the active alert as a string,
386
402
  or throw `NoAlertPresentError` if none is on screen.
387
- Before this, scenarios had no way to dismiss SpringBoard alerts: the
388
- resolver-driven `tap(getByTestId(...))` couldn't reach buttons that
389
- live outside the app process — `ordinal` selectors appeared to match
390
- but the tap landed on the wrong element. Pre-granting permissions via
391
- `xcrun simctl privacy` was the only workaround. MCP exposure (as
392
- `explore_step` actions) lands in a follow-up.
403
+ Before this, scenarios had no way to dismiss SpringBoard alerts: the
404
+ resolver-driven `tap(getByTestId(...))` couldn't reach buttons that
405
+ live outside the app process — `ordinal` selectors appeared to match
406
+ but the tap landed on the wrong element. Pre-granting permissions via
407
+ `xcrun simctl privacy` was the only workaround. MCP exposure (as
408
+ `explore_step` actions) lands in a follow-up.
393
409
 
394
410
  ## [0.3.0] — 2026-05-15
395
411
 
@@ -460,7 +476,7 @@ unchanged.
460
476
 
461
477
  - **Simulator runtime disambiguation.** When multiple simulators share
462
478
  a name across iOS versions, pin to one via `SIM_A_NAME=<name> @
463
- <runtime>` (e.g. `SIM_A_NAME=iPhone 16 @ iOS 17.5`). The runtime
479
+ <runtime>` (e.g. `SIM_A_NAME=iPhone 16 @ iOS 17.5`). The runtime
464
480
  part is matched as a substring of the friendly runtime name, so
465
481
  partial values like `iOS 17` work too. The `Sim "X" not found` error
466
482
  now lists all available sims with their runtime.
@@ -493,7 +509,7 @@ unchanged.
493
509
  persist.
494
510
  - **Pre-launch precondition check.** `WdaDriver.getSession()` now
495
511
  verifies `APP_BUNDLE_ID` is installed on the target sim via `simctl
496
- get_app_container` before starting the WDA session. If missing, throws
512
+ get_app_container` before starting the WDA session. If missing, throws
497
513
  a clear actionable error pointing at `unotest-mobile install`.
498
514
  - `SimctlAdapter.isInstalled(udid, bundleId)` and
499
515
  `SimctlAdapter.erase(udid)` — exposed on the adapter; ios-utils gains
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@unotest/mobile",
3
- "version": "0.9.1",
3
+ "version": "0.10.0",
4
4
  "description": "AI-native E2E testing for iOS React Native apps. MCP server + CLI runner + JS-DSL scenarios.",
5
5
  "license": "MIT",
6
6
  "type": "module",
@@ -37,10 +37,10 @@
37
37
  "sharp": "^0.34.5",
38
38
  "zod": "^3.23.8",
39
39
  "zod-to-json-schema": "^3.23.5",
40
- "@unotest/core": "^0.9.1",
41
- "@unotest/protocol": "^0.9.1",
42
- "@unotest/dsl": "^0.9.1",
43
- "@unotest/viewer": "^0.9.1"
40
+ "@unotest/core": "^0.10.0",
41
+ "@unotest/protocol": "^0.10.0",
42
+ "@unotest/viewer": "^0.10.0",
43
+ "@unotest/dsl": "^0.10.0"
44
44
  },
45
45
  "peerDependencies": {
46
46
  "mysql2": "^3.0.0",