Refactor: coverage-increments-1-3-testable-seams #199

• Work Mode: full-feature

Problem / Why

The COM/VSTO coverage exemption (#197) made the coverage metric meaningful by scoping the denominator to genuinely-testable code, but the measured production-only rate (71.65%) is still below the 80% floor. The coverage roadmap (artifacts/research/csharp-coverage-roadmap.2026-06-12.md §6) identifies concrete, low-risk unit-test increments on already-testable seams that raise covered code toward the floor. This entry implements Increments 1-3.

Proposed Behavior

Add MSTest unit tests (MSTest + Moq + FluentAssertions) for the testable seams that #197 deliberately preserved in the denominator. No production behavior change is expected; introduce the smallest seam only if strictly required and flag it.

• Increment 1 - ToDoModel (ToDoModel.Test): ToDoLoader.SetAndSave<T> (all overloads, read-only guard, null objectSetter/objectSaver paths); IDList.GetNextToDoID(string) (base case, ID-already-present loop, length-boundary); ProjectEntry (SetProjectId happy/null/malformed, CompareTo cases); BaseChanger remaining uncovered branches.
• Increment 2 - QuickFiler (QuickFiler.Test): keyboard-action value objects (KaChar, KaKey, KaStringAsync, KaCharAsync, KaKeyAsync), KbdActions<> remaining branches, FilerQueue, QfcQueue pure queue-management paths.
• Increment 3 - TaskMaster (TaskMaster.Test): AppStagingFilenames (injected settings stub), AppFileSystemFolderPaths.MatchBestSpecialFolder (pure LINQ), AppQuickFilerSettings remaining pure properties.

Acceptance Criteria

• Increment 1 tests added and passing; ToDoModel covered-line count increases for the named seams.
• Increment 2 tests added and passing; QuickFiler covered-line count increases for the named seams.
• Increment 3 tests added and passing; TaskMaster covered-line count increases for the named seams.
• All tests follow the General + C# Unit Test Policy (MSTest, Moq, FluentAssertions, AAA, deterministic, no temp files, no external dependencies).
• Full C# toolchain green (csharpier, analyzers, nullable, MSTest); no coverage regression on changed lines; new/changed code targets >= 90%.
• Production-only coverage re-measured and recorded; net increase vs the 71.65% post-Refactor: com-vsto-coverage-exemption #197 baseline.
• No exempted (COM/VSTO/WinForms) code is un-exempted or tested via live Outlook/WinForms.

Constraints & Risks

• Tests only target seams Refactor: com-vsto-coverage-exemption #197 left measured; do not add tests that require a live Outlook process or WinForms message loop.
• Prefer no production change; if a minimal injectable seam is unavoidable (e.g., a MyBox/settings wrapper already present in source), use the smallest one and flag it.
• Large test-authoring effort across three assemblies; phase by increment.
• Determinism: no temp files, no mutable global state, no timing/sleep hacks (avoid the flaky-timing pattern tracked in Bug: timeout-task-flaky-timing #191/Flaky tests fail CI on main: non-thread-safe test list and real-solution-file write handle #176).

Test Conditions

• Positive, negative, edge, and error scenarios per the General Unit Test Policy.
• Pure-logic/arithmetic boundaries (IDList base-36, BaseChanger, MatchBestSpecialFolder).
• Queue state transitions (FilerQueue, QfcQueue).

Source

From: docs/features/potential/2026-06-14-coverage-increments-1-3-testable-seams.md