test(scripts): dispatch a script of our own instead of hunting for someone else's

[giokur]

My previous fix for this was wrong

Read a script's execution history and facets has aborted on every nightly on record. In #2220 I
moved it into the post-mingo phase on the theory that it simply ran before anything dispatched a
script.

The 2026-09-18 nightly disproved that cleanly. The case ran in post-mingo, on a freshly registered
tenant, after mingo had created and run scripts successfully at 02:26:17 and 02:29:22 — and it
aborted anyway.

The actual cause

scriptWithHistory() walks the script list and wants one that still exists and has executions:

for (Script candidate : ScriptApi.listScripts()) {
    if (ScriptApi.getExecutionFilters(candidate.getId()).getFilteredCount() > 0) return candidate;
}
return null;

MingoDeviceTest.@AfterEach deletes every script it creates (scriptsToClean → deleteScript). So
the scripts mingo dispatches are gone before the phase ends, and no candidate ever survives. Ordering
was never the problem, and no amount of reordering could have fixed it.

Both causes — "runs before its data exists" and "its data is cleaned up by whoever made it" — produce
the identical abort message. That is how the wrong reading stuck for two days.

The fix

Stop depending on found state. The case now creates a script, dispatches it to an ONLINE Windows
device, polls until the execution lands, and then runs the existing history and facet assertions
against its own script. The class already owned and cleaned up one script for the schedule half; this
adds a second and deletes it the same way.

That needed a dispatch binding the library did not have, so this also adds ScriptApi.runScript over
the product's existing runScript(input: RunScriptInput!): DispatchResponse!. The unused
RunScriptRequest DTO is left alone — different, older shape, nothing calls it.

The move to post-mingo was still the enabling step

Dispatching needs an enrolled ONLINE machine, and one only exists after the device and assistant
phases. In functional this fix would not be possible at all. The class javadoc and the tag comment
both asserted the old, wrong reason and are corrected to say this.

Still skips in one case, on purpose

If the tenant has no ONLINE Windows device the case still skips. That is an environment precondition
rather than missing found state, and asserting a page of zero rows against zero facets would pass
hollowly.

Module compiles.

🤖 Generated with Claude Code

[github-actions]

🦩 Flamingo Code Review

No findings on the current head.

Mode: advisory

---

Need another pass? Commits pushed after this review are not reviewed automatically.

• Review the new commits — the commits added since this review
• Review the whole diff again — ignoring what was already reviewed

Prefer typing? Comment @flamingo-review, or @flamingo-review full. To review every push on this pull request, add the flamingo-review-always label.

React 👍/👎 on inline comments to teach the reviewer.

Started 2026-09-18 06:10 UTC · updated 2026-09-18 06:11 UTC · workflow run