spacedandy-agents-simplex
a429a5c6f1
The iOS and desktop smoke tests looked for the note "All results on this page are already in your chats", which the app no longer shows. As nothing marks the end of a search that adds no row, they wait out the request (it times out after 10 s), retry once if it failed, and check the listed chat, no result row for it, no Open and no note. The test plans, prompts and READMEs say the same, and New chat's results as one group with the matching contacts first.
App tests on GitHub's runners
Three workflows build the apps from a branch's own code and test them, only when asked: ios-sim (the iOS app in a simulator on a macOS runner), desktop-linux (the desktop app on a Linux runner) and android-emu (the Android app in an emulator on a Linux runner). Each runs a scripted smoke test and, on request, a Claude session that follows a test plan, and a test backend (relay, directory, names server, peers). Details: iOS, desktop, Android.
Starting a run
Pushes start nothing: a run is started by hand, on any branch, with inputs that are off by default and combine.
| Input | iOS | Desktop | Android |
|---|---|---|---|
| no input | scripted test | scripted test | scripted test |
claude |
scripted test, then Claude | scripted test, then Claude | scripted test, then Claude |
backend |
also the test backend | also the test backend | also the test backend |
release |
— | also the latest release's onboarding, to tell a branch's failure from the runner's | also the latest release's onboarding, to tell a branch's failure from the runner's |
In the Actions tab: ios-sim, desktop-linux or android-emu → Run workflow, choose the branch and tick the inputs. With the GitHub CLI, each input is a field:
gh workflow run ios-sim.yml --ref <branch>
gh workflow run ios-sim.yml --ref <branch> -f claude=true
gh workflow run desktop-linux.yml --ref <branch> -f claude=true -f backend=true -f release=true
gh workflow run android-emu.yml --ref <branch> -f claude=true
A newer run of a workflow on the same branch cancels the older one.
How long a run takes
| Run | iOS | Desktop | Android |
|---|---|---|---|
| Scripted test, builds reused | about 25 min, plus the wait for a macOS runner, which the team's builds share (minutes to an hour) | about 10 min | about 15 min |
| The core rebuilt, after Haskell changes | 30–45 min more | about 25 min more | about 27 min more |
| With the test backend | 20–45 min more when it is rebuilt, then about 10 min to start | about 30 min more when it is rebuilt, then a few minutes to start | as desktop, whose backend it shares |
| With Claude on the smoke plan | 10–15 min more | about 4 min more | about 5 min more |
The Claude job
It needs the GitHub environment app-sim-claude with the secret CLAUDE_CODE_OAUTH_TOKEN, a Claude subscription token. The environment's branch rule alone decides which branches may use it: on a branch it doesn't allow, a run that asks for Claude fails at the Claude job's environment check, and the scripted test runs as usual. Each run spends the subscription's usage limits; the job summary gives its tokens and cost estimated at API prices.