Lucene sorting uses always english locale

[Lassi Immonen]

Lucene sort does not work correctly for different than english locales

This fails:

createDoc("1", "Ä", fi);
createDoc("2", "B", fi);
createDoc("3", "A", fi);
List<Dto> dtos = dao.getAll(fi);
assertEquals("3", dtos.get(0).getId());
assertEquals("2", dtos.get(1).getId() );
assertEquals("1", dtos.get(2).getId());

[Lassi Immonen (lassi-immonen)]

Added new constructor parameter to LuceneSerializer:

public LuceneSerializer(boolean lowerCase, boolean splitTerms, Locale sortLocale)

which is used in sort field generation:

sorts.add(new SortField(toField(path), sortLocale, reverse));

In the future we might want to add support to set sort locale per field, but I think it can wait until a real use case emerges.

[Timo Westkämper (timo-westkamper)]

Released in 2.1.1