LtUnlintNonExistingDefect falsely warns for range +unlint that covers only part of the defects

Problem

LtUnlintNonExistingDefect produces a false "doesn't make sense" warning when a line-range +unlint covers only part of the defects of a lint.

Root cause

src/main/java/org/eolang/lints/DefectMissing.java:50 uses allMatch:

missing = !lines.stream().allMatch(new UnlintInRange(unlint));

For +unlint foo:10-20 with foo defects on lines [5, 15], allMatch returns false → the whole unlint is reported as "doesn't make sense", even though line 15 is legitimately suppressed by the range.

Minimal example

EO program with +unlint ascii-only:1-10 where the ascii-only lint reports defects both inside and outside the range (e.g. two non-ASCII comments on lines 5 and 15).

Expected: no warning — line 5 is suppressed by the range, line 15 is not and is reported by the lint itself.

Actual: unlint-non-existing-defect warns "Suppressing 'ascii-only:1-10' doesn't make sense, because there are no defects with it" (well — a defect exists, on line 15, but the warning fires anyway).

Expected behavior

A range unlint makes sense if any defect of that lint falls into the range → use anyMatch (the same semantics LtUnlint uses via removeIf(inRange), see src/main/java/org/eolang/lints/LtUnlint.java:63).

Related

• #1232/#950 (line-specific +unlint handling)
• The pack unlint/allows-unlint-for-defects-in-the-line-range.yaml only covers the all-inside case, so the partial case is untested.

Start in src/main/java/org/eolang/lints/DefectMissing.java:50 and compare its matching behavior with src/main/java/org/eolang/lints/LtUnlint.java:63. Add a regression case to the pack, alongside unlint/allows-unlint-for-defects-in-the-line-range.yaml, covering defects both inside and outside the range. Run the relevant linter tests and confirm that only the valid suppression is accepted without an unlint-non-existing-defect warning.