Upgrade to Polecat 5.19.1 (and JasperFx 2.53.0)

[jeremydmiller]

The pin was a range, so the build has been on 5.12.0

<PackageVersion Include="Polecat" Version="[5.12.0,6.0.0)" />

NuGet resolves a version range to its floor, not its ceiling — so this has been restoring Polecat 5.12.0 all along, not "the latest 5.x" that the range reads like. Confirmed against project.assets.json:

Polecat/5.12.0

Seven minor versions have shipped since (5.15.1, 5.16.0, 5.17.0, 5.18.0, 5.19.0, 5.19.1). Pinned explicitly to 5.19.1 so the version moves when someone moves it, rather than sitting at a floor that looks like a ceiling.

JasperFx 2.53.0 comes with it

Polecat 5.19.1 depends on JasperFx / JasperFx.Events 2.53.0; Wolverine pins the family at 2.47.0, which is a downgrade and will not resolve. The whole family moves to 2.53.0 — JasperFx, JasperFx.Events, JasperFx.Events.SourceGenerator, JasperFx.SourceGenerator — which is also current.

JasperFx.Events.SourceGenerator has to move in lockstep with the runtime: a generator emitting dispatchers for a different JasperFx.Events version fails at runtime with No source-generated dispatcher found and no compile error to point at it. That is a real failure mode, not a hypothetical — it is half of what has had CIFisher red on main.

Testing

Suite	Result
PolecatTests	288 / 290 (2 skipped)
wolverine.slnx Release -f net9.0	clean, 0 warnings

PolecatIncidentService.Tests targets net10.0 only, so it is not covered by the pinned-framework run — CIPolecatWorkflow exercises it.

The JasperFx 2.53.0 lift is also verified more broadly on #3991, where it takes MartenTests from 607/610 to 610/610 and FisherTests from 18/27 to 27/27.

Conflicts with #3991

#3991 bumps JasperFx to 2.53.0 too, for Fisher 1.0.2. The two will conflict on those four lines in Directory.Packages.props; whichever merges second rebases. Nothing else overlaps.

🤖 Generated with Claude Code

https://claude.ai/code/session_01HxbBpN6PmRB5CMSSTdGNr3

[jeremydmiller]

Rebuilt on current main now that #3991 has merged.

The only conflict was the two adjacent pins — kept Polecat 5.19.1 from this branch and Fisher 1.0.2 from main. The JasperFx lines merged cleanly because both sides set 2.53.0, so that bump is now shared rather than duplicated and this PR's remaining delta is just the Polecat pin itself.

Re-verified after the merge: PolecatTests 288/290 (2 skipped), full wolverine.slnx Release -f net9.0 clean at 0 warnings. CI re-running.