fix(#950): keep line-zero defects with line-specific unlint

[VasilevNStas]

Fixes #950.

Problem

In LtUnlint.defects() the loop that re-adds non-suppressed defects uses the guard line != 0 && defect.line() == line. When a lint produces a defect at line 0 (e.g. mandatory-package reports at line 0 when the whole +package meta is absent) and the EO program also contains a line-specific +unlint mandatory-package:5, the defect at line 0 is silently discarded: global is set, line 0 stays in problematic but is skipped by line != 0, and the fallback if (!added.get() && !global) is also skipped because global is true.

Solution

Drop the line != 0 guard in the re-add loop and compare defect.line() == line only. Line 0 is an intentional production case (lineno.xsl and empty-object.xsl handle it explicitly) and is only meant to be suppressed by a global +unlint lint-name, never by a line-specific +unlint lint-name:N.

Changes

• src/main/java/org/eolang/lints/LtUnlint.java — re-add defects on line 0 when their line stays problematic.
• src/test/resources/org/eolang/lints/unlint-mandatory-package-line.eo — new fixture: no +package, +unlint mandatory-package:5.
• src/test/java/org/eolang/lints/LtUnlintTest.java — new test keepsLineZeroDefectWithLineSpecificUnlint (failed before the fix).

Tests

• LtUnlintTest — 12 tests green (incl. new one).
• Regression LtByXslTest, SourceTest, PkMonoTest, MonoLintsTest, LtUnlintNonExistingDefectTest — 443 tests green after mvn clean.
• mvn -Pqulice install — clean.

[VasilevNStas]

@yegor256 @volodya-lombrozo please review

[volodya-lombrozo]

@VasilevNStas To be honest, I don't clearly understand the purpose of these changed. Can u provide actual and expected behaviour in the issue description please? Also, please, pay attention that the issue you are trying to solved weren't accepted by maintainers.

[VasilevNStas]

Replying to the review summary: I added the actual/expected explanation in the thread on unlint-mandatory-package-line.eo. In short: before this change, +unlint <name>:0 could never suppress the line-0 defect (e.g. from mandatory-package), because of the line != 0 guard; the fix removes that guard so line 0 behaves like any other line. This was found during a code audit, not requested in the ticket. I understand the ticket was not formally accepted — if the line != 0 guard is intentional, please say so and I will close the PR.

[yegor256]

@VasilevNStas my suggestion: don't try to fix issues that are not yet labeled as "bugs" or "enhancement". You mostly waste your and our time by doing this.

[VasilevNStas]

Understood, thank you for the clear guidance. The PR came from a code audit where the line-0 suppression looked like an inconsistency, not from a labeled issue. I will keep the PR open as-is for reference, but going forward I will only work on issues that are explicitly labeled as bug or enhancement, and won't spend time on unlabeled ones. Appreciate the directness.