fix(3.3.1:24): ask which indexer to use, and never run without one configured

[MattDHill]

What was broken

store.json.indexer has no default, and the select that sets it declared default: 'fulcrum' — so the form rendered as though Fulcrum were already chosen. Closing it without saving persisted nothing, and the task that would have forced the choice was gated on kind === 'install', so an install that updated into the feature was never asked at all.

With the choice unmade, setupDependencies declared no indexer and watchHosts omitted the ELECTRUM section — while MEMPOOL.BACKEND stayed pinned to z.literal('electrum'). Upstream's own ELECTRUM defaults then applied: 127.0.0.1:3306 with TLS on, which is this package's own MariaDB. The backend reconnected to it once a second, indefinitely:

ERR: Electrum error: {"reason":"wrong version number","code":"ERR_SSL_WRONG_VERSION_NUMBER"}
INFO: Disconnected from Electrum Server at 127.0.0.1:3306
[Warning] Aborted connection 578380 to db: 'unconnected' user: 'unauthenticated'

Both health checks stayed green throughout. Address lookups were silently dead, and nothing surfaced it.

The fix

• actions/selectIndexer.ts — offers none and preselects nothing, so a choice has to be made rather than inherited.
• init/taskSelectIndexer.ts — the critical task is raised whenever the choice is unmade, which reaches installs already in this state on their next start.
• init/watchHosts.ts — writes MEMPOOL.BACKEND alongside ELECTRUM: electrum when an address resolved, none otherwise. The backend is never left on upstream's defaults.
• file-models/mempool-config.json.ts — BACKEND relaxed from z.literal('electrum') to z.enum(['none','electrum']).
• file-models/store.json.ts — indexer gains 'none': absent means the choice is unmade, 'none' means declined.

No migration. watchHosts repairs BACKEND on every init and the task reaches every affected install, so both halves heal on the next start.

default: null needs the SDK signature widened — Start9Labs/start-technologies#3871. The as any goes when that ships.

Verified on demo.local

Upgrading 3.3.1:23 → 3.3.1:24 on a box in the broken state (ELECTRUM: {}, BACKEND: "electrum", 578,380 aborted MariaDB connections):

Check	Result
Task on an update, not just install	mempool:select-indexer present, active: true, critical
Blocks startup	Invalid Request: Service mempool has an unresolved critical task
BACKEND while unmade	flipped to none; POLL_RATE_MS: 8000 intact, so the merge doesn't clobber siblings
Form renders unselected	"default": null, "value": {} over the wire on SDK 2.0.9
Selecting Fulcrum	ELECTRUM {HOST: 10.0.3.1, PORT: 51537, TLS_ENABLED: false}, BACKEND: electrum
Dependencies	currentDependencies gained fulcrum
Task cleared	gone from the task list
Service	running, api and webui both success
Connection	INFO: Connected to Electrum Server at 10.0.3.1:51537 (["Fulcrum 2.1.2","1.4"])
Error loop	0 Electrum errors and 0 aborted MariaDB connections since the upgrade

Not exercised: an address lookup through the published port — it wasn't reachable from the test shell. The Electrum handshake above is what the lookup depends on.