fix: resolve flaky stops_for_route test by sorting trips

[AhmedAlian7]

Summary

This PR fixes a known flaky test in TestStopsForRouteHandlerEndToEnd where the number of returned stops would fluctuate between 21 and 22 depending on execution order.

closes #336

Changes Made

internal/restapi/stops_for_route_handler.go

• Fix: Added sort.Slice to sort the tripsInGroup slice by Trip.ID before selecting the representative trip.
• Impact: This ensures that the representativeTrip selection is deterministic. The same trip is now always chosen to represent the group, resulting in a consistent list of stops every time.

internal/restapi/stops_for_route_handler_test.go

• Update: Removed the conditional assertion that accepted either 21 or 22 stops.
• Update: Hardcoded strict expectations:
  • Inbound: Expect exactly 21 stops.
  • Outbound: Expect exactly 22 stops.

Verification

Ran the test 10 times locally to ensure flakiness is resolved.

go test -v -count=10 -tags sqlite_fts5 ./internal/restapi/... -run TestStopsForRouteHandlerEndToEnd

[aaronbrethorst]

Hey Ahmed, nice detective work tracking down the root cause of this flaky test — Go's nondeterministic map iteration was indeed the culprit, and sorting tripsInGroup by trip ID before picking the representative is a clean, correct fix. The tightened assertions removing the TODO workaround are a welcome improvement too. Before we can merge this, I need you to make one change:

Important

1. Squash and rebase

The branch has three commits (including a merge commit) and is six commits behind main. Please squash into a single atomic commit and rebase against main. If rebasing feels unfamiliar, here's a short guide that might help: https://www.brethorsting.com/blog/2026/01/git-rebase-for-the-terrified/

Thanks again, and I look forward to merging this change.

[AhmedAlian7]

Hi Aaron, thnx for the feedback and for sharing the article, which was a great help!
I have updated the PR as requested. It’s now ready for your review.