Improve performance of speculative rental vehicle use in reverse search

[tkalvas]

Summary

Do not speculate use of rental networks which are not allowed by the request. They just cause extra states,
which in the end will not reach targets anyway.

Issue

The speculative fanout of extra rental network states can be massive in pathological cases, and serious
in accidental difficult cases.

Unit tests

The existing test StreetEdgeGeofencingTest did not find as many new states as it used to, and this
shows exactly what the fix is about. The expectation of the result was changed accordingly.

I can get better code coverage if this is not rejected outright.

Documentation

No documentation change, no high-level visible behavior state except execution speed.

Changelog

Could consider skip changelog.

Bumping the serialization version id

No.

[codecov]

Codecov Report

Attention: Patch coverage is 90.32258% with 3 lines in your changes missing coverage. Please review.

Project coverage is 69.71%. Comparing base (481c3b8) to head (5b5c68a).
Report is 416 commits behind head on dev-2.x.

Files with missing lines	Patch %	Lines
.../opentripplanner/street/model/edge/StreetEdge.java	90.32%	1 Missing and 2 partials ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##             dev-2.x    #6260      +/-   ##
=============================================
- Coverage      69.73%   69.71%   -0.02%     
- Complexity     17681    17701      +20     
=============================================
  Files           2007     2008       +1     
  Lines          75776    75844      +68     
  Branches        7755     7769      +14     
=============================================
+ Hits           52845    52878      +33     
- Misses         20218    20255      +37     
+ Partials        2713     2711       -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

---

🚨 Try these New Features:

• JS Bundle Analysis - Avoid shipping oversized bundles

[leonardehrenfried]

Can you fix the formatting?

[tkalvas]

Yeah.. didn't even check, but of course the new parameter made the code too wide. Fixed now.

[optionsome]

Why did the order of states change btw?

[tkalvas]

We only know if the unknown network case will be generated after generating the known cases. With recursive streams it was stuffed in the middle, but with an array it's easier (and more performant) to put it in the end. It shouldn't really matter what the order is, but in debugging it can be helpful that it is stable, so using State[] and not Set<State> is understandable.