greek_romanizer 0.2.1

Configurable Greek-to-Latin romanization: ASCII ELOT 743/ISO 843 style, plus English and German phonetic respellings, from one overridable rule set.

0.2.1

• Shortened the pubspec.yaml description to fit pub.dev's 60-180 character guideline (it was 192 characters, so search engines were truncating it).

0.2.0

Cross-checked the whole engine against the official ELOT 743 / BGN-PCGN reference table letter-by-letter and against native-language sources for the reader presets. Found and fixed three real gaps:

• germanReader's τζ now romanizes as ds, not dz — τζ was hardcoded instead of routing through zLetter like standalone ζ does. German z reads as /ts/, so dz was wrong for a German reader; a native German phrasebook confirms ds. This is a behavior change if you were depending on germanReader's exact τζ output.
• γχ and γξ now route through chiLetter and xLetter instead of a hardcoded ch/x, so overriding either (e.g. chiLetter: 'h') now applies consistently everywhere those letters appear, not just to the standalone letter.
• μπ now collapses to a bare b at word end too, not just word start, matching the official table (word-final μπ essentially never occurs in native Modern Greek vocabulary, but the fix is correct per the source).
• ΐ and ΰ (dieresis and tonos on the same vowel, e.g. ταΐζω, Ταΰγετος) now carry a StressSpan and render with a combining accent when showStress is true, instead of silently losing the stress.

No change to any existing golden-test output — these only affect previously-untested combinations.

0.1.0

• Initial release.
• GreekRomanizer.official() — ASCII, close to ELOT 743 / ISO 843.
• GreekRomanizer.englishReader() / GreekRomanizer.germanReader() — phonetic respellings, ported from the foni app's romanization engine.
• Every letter choice is a field on RomanizationSpec, overridable per preset factory or via RomanizationSpec.copyWith.
• wordOverrides escape hatch for whole-word exceptions.