Read the documentation of the tokenstream API. its been this way since 2.9 I changed the code intentionally to throw this exception so people would be forced to fix their code.
On Thu, Nov 8, 2012 at 8:20 AM, Spyros Kapnissis <ska...@yahoo.com> wrote: > Hello, > > Noticed the following issue during our recent code migration to LUCENE_40. > The test below will fail with an ArrayIndexOutOfBoundsException -1. It will > pass only if tokenizer.reset() is called before incrementing the tokens. > > @Test > public void whitespaceTokTest() throws IOException { > > String text = "a b c d"; > Tokenizer tokenizer = new WhitespaceTokenizer(Version.LUCENE_40, new > StringReader(text)); > List<String> tokens = new ArrayList<String>(); > while (tokenizer.incrementToken()) { > tokens.add(tokenizer.getAttribute(CharTermAttribute.class).toString()); > } > assertEquals(tokens, Arrays.asList(new String[]{"a","b","c","d"})); > } > > This used to work, at least until LUCENE_33. Is this a bug, or am I missing > something? > > Thank you, > Spyros --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org