You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Marshall Schor (JIRA)" <de...@uima.apache.org> on 2015/06/27 22:17:04 UTC
[jira] [Commented] (UIMA-4487) Position lost on FSIterator copy
[ https://issues.apache.org/jira/browse/UIMA-4487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14604339#comment-14604339 ]
Marshall Schor commented on UIMA-4487:
--------------------------------------
I was able to reproduce this, and see there are several problems. I've got to investigate a bit more.
> Position lost on FSIterator copy
> --------------------------------
>
> Key: UIMA-4487
> URL: https://issues.apache.org/jira/browse/UIMA-4487
> Project: UIMA
> Issue Type: Bug
> Components: Core Java Framework
> Reporter: Richard Eckart de Castilho
> Fix For: 2.8.0SDK
>
>
> Apparently, in some cases, a FSIterator becomes invalid/looses its position after a copy(). Unfortunately, I haven't been able to quickly create a plain UIMA test case (I tried reproducing the effect in GetAllIndexedTest, but couldn't.
> So the following snipped is works (fails) when pasted into the uimaFIT CasUtilTest after upgrading uimaFIT to 2.8.0 RC 1:
> {noformat}
> @Test
> public void lala() throws Exception {
> String text = "Rot wood cheeses dew?";
> tokenBuilder.buildTokens(jCas, text);
> CAS cas = jCas.getCas();
> // Option 1: remove the ".copy()"
> FSIterator<FeatureStructure> fses = cas.getIndexRepository()
> .getAllIndexedFS(cas.getTypeSystem().getTopType()).copy();
> // Option 2: explicitly moveToFirst
> // fses.moveToFirst();
>
> int i = 0;
> while (fses.hasNext()) {
> fses.next();
> i++;
> }
>
> assertEquals(6, i);
> }
> {noformat}
> Maybe you have some idea or can highlight another UIMA core test case where this could be conveniently tried out. (Unfortunately, setting up a test CAS in plain UIMA is a bit painful/verbose)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)