Fix chroma lexer mapping

[wxiaoguang]

Fix some edge cases for ".hcl" and ".v" files, and add more tests

[silverwind]

Comment written by Claude (AI assistant)

I reviewed this PR. The changes look correct and well-structured. A few observations:

What the PR does well:

1. The new conflictingAliasLangMap cleanly mirrors the existing conflictingExtLangMap pattern — one handles file extension conflicts, the other handles language name/alias conflicts. This is a consistent design.

2. The extracted processLexerNameAliases function adds proper conflict detection with panics on unexpected duplicates. This is important because without it, the last lexer registered for a given alias silently wins, which is non-deterministic and could change between Chroma versions.

3. The panic-based validation at init time (both for unexpected conflicts and for the final consistency check) is the right approach — it catches misconfigurations immediately at startup rather than producing subtly wrong highlighting at runtime.

4. The test coverage is thorough. The split into casesWithContent (testing the full detectChromaLexerWithAnalyze pipeline) and casesNameLang (testing detectChromaLexerByFileName directly, including the byLang return value) makes the different detection layers easy to reason about. The .hcl case is particularly well-tested since it exercises the full conflict resolution path: ext map → Terraform, then enry → "HCL", then alias map → HCL lexer.

5. The rename from setMapWithLowerKey to setFileNameMapWithLowerKey improves clarity since there's now a separate path for name/alias processing.

LGTM

[silverwind]

Wow, first time I see Claude not complaining about anything 😆.