Strengthen coverage tests to assert UI behavior

[JoshCLWren]

Problem

The recent coverage additions in tests/test_coverage_additions.py increase line coverage, but several tests do not verify the behavior they describe. Examples include:

• Tests that manually recalculate the expected anchor/y values instead of asserting the canvas state produced by _display_image_fast.
• test_open_dialog_applies_color, which does not actually drive the color-entry callback or assert a color change.
• Tests that call methods only to ensure they do not raise, without checking the resulting state or user-visible behavior.
• Production # pragma: no cover markers added around dialog callbacks and cleanup paths without a focused behavioral test.

These tests can pass while the underlying UI behavior is broken and make the 96% coverage target less meaningful.

Fix

• Rewrite coverage-only tests into behavior-focused regression tests.
• Assert canvas item coordinates/anchors, scroll offsets, dialog state, pending quit state, applied colors, cleanup state, and worker scheduling outcomes directly.
• Remove unnecessary # pragma: no cover markers where realistic tests can exercise the code.
• Keep tests deterministic with the existing tk_root fixture.

Verification

• Temporarily mutate the relevant production branch and confirm each regression test fails for the intended reason.
• Run the full coverage command and inspect the missing-line report.
• Do not weaken benchmark tests or mock the I/O paths covered by performance tests.

Done criteria

• New tests assert observable behavior rather than duplicating implementation calculations.
• The tests fail against representative mutations.
• Coverage remains at or above the project threshold without gratuitous exclusions.
• Full lint, type, and test checks pass.