Hi guys, excuse the delay.
The code you are talking about is most probably my copy/paste error :) I'll update the compareTo method. the code is still experimental (that explains why there are no tests yet for it). this feature is enabled via a system property (so we can test the performance of the patch) See the JR issue that tracks the code changes https://issues.apache.org/jira/browse/JCR-2959 best, alex On Mon, Oct 10, 2011 at 3:19 PM, Julian Reschke <[email protected]>wrote: > On 2011-10-08 09:48, Julian Reschke wrote: > >> On 2011-10-08 05:48, Dave Brosius wrote: >> >>> This comparator seems odd to me, as it never returns (d) but perhaps i'm >>> confused, is it ok? >>> >>> org.apache.jackrabbit.core.**query.lucene.sort.**ValueComparableWrapper >>> >>> public int compareTo(**ValueComparableWrapper o) { >>> final int d = compare(v, o.getValue()); >>> if (d != 0) { >>> if (reversed) { >>> return -d; >>> } >>> return -d; >>> } >>> return 0; >>> } >>> >> >> It certainly looks broken. (If it's intended it would need a comment >> explaining why it is so) >> >> Do we need a test before fixing? >> >> Best regards, Julian >> > > From a quick test it seems that this code is never exercised from the unit > tests. The call hierarchy indicates that this should be used from SQL2... > > Does anybody recall what's going on here? It would be good to have a test > before touching this... > > Best regards, Julian >
