Give fault reporting a seam a test can observe

[fiskus]

Third code unit of quiltdata/quilt-specs#37 (w-report-seam), after #820 and #822.

Problem

report_anomaly and report_error were associated functions calling the crash SDK directly. They worked only because that SDK keeps a process-global client — which meant they bypassed Telemetry's state entirely and, more to the point, nothing could observe that a path reported a fault.

That matters now rather than in principle. Instrumenting the background engine (next unit) is mostly about failures, and the change already settled that failure counting rides the crash sink rather than the analytics vocabulary — there are no *_failed events. So the one mechanism the telemetry work leans on for failure visibility was the one nothing could assert against.

Approach

Both are methods over a Faults field:

• Live — captures via the SDK, as before.
• DryRun — writes to the developer's console. A real gain on its own: until now a fault in a local build went nowhere at all, because a dev build has no crash client.
• Recorded (cfg(test)) — keeps them in memory so a test can read them back.

Deliberately not a trait, per the d-no-provider-traits decision on that PR: this is a seam for observation, not an abstraction over vendors.

The failed-send path inside track routes through the field too, so the whole fault path is observable rather than most of it.

What it immediately bought

The spec's claim about logout is now checked rather than merely written. A logout that finds nothing stored is still a success — a second click must not raise an error — but it emits no event and reports an anomaly, because the settings page lists hosts from those very directories, so a missing one means the list and the erase disagree about the name.

Two tests, and the second earns its place: that the anomaly is reported, and that a logout which does erase something reports nothing. Without the second, the first would pass on an implementation that reported an anomaly every time.

Verification

just lint exit 0, cargo fmt --check exit 0, 307 quilt-sync tests, workspace green. CI green on the branch.

Five new tests: two on the logout behaviour above, and three on Faults itself — anomalies and errors recorded distinguishably and in order, only the recorder reports (so a test holding a live sink cannot pass by reading someone else's recording), and a local build dry-runs its faults.

Locally checkable via the dry run from #820 — faults now print alongside events:

just start
# telemetry(dry-run) anomaly: …
# telemetry(dry-run) fault: …

Note

One cfg(test) enum variant, which is unusual enough to flag. The alternative was a trait — declined for the reasons on #37 — or leaving fault reporting unobservable, which is the status quo this unit exists to end.

🤖 Generated with Claude Code

Greptile Summary

The PR introduces an instance-owned fault-reporting seam while preserving production Sentry reporting and making development and test behavior observable.

• Adds Faults modes for live capture, development console output, and test-only recording.
• Routes telemetry send failures and authentication-command faults through the new seam.
• Adds tests covering fault ordering, recorder isolation, sink selection, and logout anomaly behavior.

Confidence Score: 5/5

The PR appears safe to merge, with production fault delivery preserved and the new development and test observation paths consistently wired.

All fault-reporting callers use the new instance methods, production still reaches the same Sentry capture APIs, and test recorders are isolated per telemetry instance.

Important Files Changed

Filename	Overview
quilt-sync/src-tauri/src/commands/auth.rs	Migrates logout fault reporting to the telemetry instance and verifies both anomalous and successful logout behavior.
quilt-sync/src-tauri/src/telemetry.rs	Adds the fault sink to telemetry state, routes failed analytics sends through it, and exposes test-only recorded reports.
quilt-sync/src-tauri/src/telemetry/sentry.rs	Implements live, dry-run, and test-recording fault destinations with focused behavior and isolation tests.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    Caller[Telemetry caller] --> Report[Telemetry report_anomaly / report_error]
    Track[Failed analytics send] --> Report
    Report --> Faults{Faults mode}
    Faults -->|Live| Sentry[Sentry SDK]
    Faults -->|DryRun| Console[Developer console]
    Faults -->|Recorded in tests| Memory[In-memory fault log]
    Memory --> Assertions[Test assertions]

Loading

_{Reviews (1): Last reviewed commit: "Give fault reporting a seam a test can o..." | Re-trigger Greptile}