You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 0d4b58c
Browse filesBrowse the repository at this point in the historyBrowse files
Copy file name to clipboardExpand all lines: docs/evidence/android-automation-performance-2026-10-19.md
+2Lines changed: 2 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -4,6 +4,8 @@ The comparison starts from `eef1a3f76d7216da006ff59c975620c8b9dd7182`.
4
4
The reference experiment is [tester-army/e2e#1234](https://github.com/tester-army/e2e/pull/1234), inspected at `2593e45025edc52b2c381e809b41cb63d5878e2b`.
5
5
Its measurements are author-reported. Its direct engine sets text through `ACTION_SET_TEXT`, injects input through `UiAutomation`, and captures screenshots inside its instrumentation process.
6
6
7
+
The follow-up [original-app comparison](android-e2e-parity-2026-10-19.md) runs the exact 42 Android cases and operation benchmark. Both agent-device variants passed all 42 cases; the direct engine passed 40 and failed two. The base/head suite times were 327.95/321.25 seconds. The fixture primitive gains below do not establish the reference experiment's total-suite gain.
8
+
7
9
## Reproduce
8
10
9
11
Use a dedicated Pixel 7 API 36 emulator and the current `examples/test-app` Release APK.
0 commit comments