Re: svn commit: r1492995 - /jackrabbit/oak/trunk/oak-lucene/src/main/java/org/apache/jackrabbit/oak/plugins/index/lucene/LuceneIndexEditor.java

2013-06-14 Thread Jukka Zitting
Hi, On Fri, Jun 14, 2013 at 11:22 AM, alexparvule...@apache.org wrote: +private AtomicLong indexedNodes; A normal long should be fine here, as the editor instance isn't being accessed concurrently from multiple threads. BR, Jukka Zitting

Re: svn commit: r1492995 - /jackrabbit/oak/trunk/oak-lucene/src/main/java/org/apache/jackrabbit/oak/plugins/index/lucene/LuceneIndexEditor.java

2013-06-14 Thread Alex Parvulescu
good point, I was thinking about concurrent commits, but that's not how it works. I'll change to a normal long then thanks, alex On Fri, Jun 14, 2013 at 10:28 AM, Jukka Zitting jukka.zitt...@gmail.comwrote: Hi, On Fri, Jun 14, 2013 at 11:22 AM, alexparvule...@apache.org wrote: +

Re: svn commit: r1492995 - /jackrabbit/oak/trunk/oak-lucene/src/main/java/org/apache/jackrabbit/oak/plugins/index/lucene/LuceneIndexEditor.java

2013-06-14 Thread Alex Parvulescu
As it turns out, using a normal Long doesn't work (because of primitive unwrapping we lose the original reference). As per Jukka's suggestions I'm looking into introducing a context class to wrap all the baggage we pass from the 'parent' and this counter as well. On Fri, Jun 14, 2013 at 10:41