Match run report deltas by module type

[thomhurst]

Summary

• build the results-printer delta lookup from stable module type identifiers
• remove positional Zip coupling between pipeline modules and report entries
• verify reversed report ordering preserves each module's delta

Validation

• SpectreResultsPrinterTests: 5/5
• core Release build: 0 warnings, 0 errors

Closes #3749

[claude]

Code review

Reviewed the change against CLAUDE.md and checked for bugs/logic issues in the diff.

Summary: SpectreResultsPrinter now builds the Δ previous lookup by joining on ModuleRunReport.ModuleTypeName (via ModuleTypeIdentifier.Get) instead of Zip-ing pipelineSummary.Modules with pipelineSummary.RunReport.Modules by list position. This removes the implicit invariant that the two lists stay positionally aligned (called out in #3749), and mirrors the same type-name join already used internally in PipelineRunReportFactory.previousByType.

I traced this against PipelineRunReportFactory.Create/CreateModuleReport to confirm:

• RunReport.Modules entries are always tagged with the same ModuleTypeIdentifier.Get(...) identity that AddModuleRow now looks up with, so the join key is consistent on both sides.
• Duplicate module types (same type instantiated more than once in a pipeline) collapse to a single dictionary entry both before and after this change (the factory itself only ever computes a delta for types where group.Count() == 1), so there's no behavior regression there.
• The new test (ModulesTable_MatchesDurationDeltasByModuleType) exercises exactly the scenario the fix targets — RunReport.Modules in a different order than pipelineSummary.Modules — and asserts each module's delta still lands on the correct row.

No bugs or CLAUDE.md violations found. This is a small, well-scoped, well-tested defensive fix.

[thomhurst]

Refreshed onto current main in f7a2d36 after the required Ubuntu pipeline hit the known core-test host failure (no failed-test summary; exit 2/134 after ~16 minutes). Local validation: SpectreResultsPrinterTests 5/5; core Release build 0 warnings/errors.

[claude]

Code review

Re-checked this PR's current state. The only change since the previous review is a merge from main (f7a2d36) that pulls in unrelated files (docs, RunReportService, ModuleTester, packaging) — the two files this PR actually changes, src/ModularPipelines/Helpers/SpectreResultsPrinter.cs and test/ModularPipelines.UnitTests/Helpers/SpectreResultsPrinterTests.cs, are byte-identical to the version already reviewed at 38b7c19.

No new code to review. The previous review stands: the join-by-ModuleTypeName fix is correct, well-tested, and no bugs or CLAUDE.md violations were found.