Resgrid IC 2.25

Resgrid IC 2.25

Channel production · build 100025 · run 25 · commit 0e1aaea

Platform	Path	Details
android	native build	37bb3713a34c — binaries attached
ios	native build	2efcfd56a6d6 — binaries attached

Docker: resgridllc/ic:2.25, ghcr.io/resgrid/ic:2.25

Verify downloads with sha256sum -c SHA256SUMS.

Pull request #28: RO-T102 Fixes for Dispatch, pending operation handling, outbox

Summary

• Restrict status hold-to-confirm behavior to native Responder and Unit apps.
• Improve outbox handling for idempotent writes and pending operations.
• Prevent post-delivery refreshes from blocking subsequent writes.
• Clean up stale “Pending sync” markers when their outbox events no longer exist.
• Improve Dispatch unit status refresh behavior and call-file upload deduplication.

Details

Status hold-to-confirm

• Added the status.holdToConfirm capability to all app variants.
• Enabled it only for Responder and Unit variants.
• Dispatch, IC, BigBoard, web, and Electron builds now set statuses immediately without a hold.
• The department’s StatusHoldToConfirm setting is honored only when the app capability and native-platform requirements are met.
• Added manifest, configuration, provider, and behavior tests for the new capability.

Idempotent writes and retry handling

• Added defineIdempotentDataHandler for writes that Core deduplicates using the outbox event ID.
• HTTP 409 responses for these writes are treated as in-progress operations and remain queued instead of being marked rejected.
• Idempotent operations receive an extended retry budget so Core’s five-minute processing claim can expire before the outbox gives up.
• Applied this behavior to call creation, call notes, call-file uploads, personnel status/staffing updates, and unit status updates.
• Call-file uploads now include the outbox event ID as an optional IdempotencyKey, preventing duplicate attachments when an upload is retried.

Outbox delivery and refresh behavior

• Added a separate refresh phase for post-delivery reads.
• Required local acknowledgement work still completes before the next write is sent.
• Screen-refresh reads now run asynchronously after delivery and no longer block the outbox drain.
• Updated call details, messages, check-ins, operations, routes, personnel status, and unit status handlers to use the new refresh flow.
• Added per-failure retry budgets so handlers can use a different limit for specific errors such as HTTP 409.

Pending operation marker reconciliation

• Added centralized reconciliation for queued-operation markers across calls, call details, check-ins, messages, personnel status, unit status, and work orders.
• Markers are removed when their corresponding outbox event is no longer present, including delivered, discarded, expired, or dead-lettered events.
• Reconciliation runs after data initialization and whenever the outbox changes.
• Reconciliation failures are logged while preserving the existing markers.

Dispatch improvements

• Multi-unit status updates no longer wait for status-list reads between each unit submission.
• Dispatch now refreshes unit rows and map markers once after selected-unit status updates are sent.
• Call detail reads support bypassing in-flight cached requests so a save followed by a refresh uses the latest server data.
• Notes and files are refreshed after acknowledgement without delaying subsequent outbox operations.

Validation

• Added coverage for:
  • Idempotent retries and HTTP 409 handling.
  • Extended retry behavior while Core holds a processing claim.
  • Duplicate-safe call-file uploads.
  • Non-blocking post-delivery refreshes.
  • Stale queued-marker cleanup and reconciliation failures.
  • Multi-unit Dispatch status updates.
  • Variant-specific status hold-to-confirm behavior.

Changes

• RO-T101 PR #28 Fixes (2db1efb)
• RO-T102 Fixes for Dispatch, pending operation handling, outbox (7b82e12)