rmuir: shouldn't this have a Jira to track the fix & record it in CHANGES.txt ?
: Date: Thu, 10 Apr 2014 21:30:53 -0000 : From: [email protected] : Reply-To: [email protected] : To: [email protected] : Subject: svn commit: r1586473 - : /lucene/dev/trunk/lucene/analysis/icu/src/java/org/apache/lucene/analysis/ : icu/ICUNormalizer2CharFilter.java : : Author: rmuir : Date: Thu Apr 10 21:30:53 2014 : New Revision: 1586473 : : URL: http://svn.apache.org/r1586473 : Log: : fix bug in buffering logic of this charfilter : : Modified: : lucene/dev/trunk/lucene/analysis/icu/src/java/org/apache/lucene/analysis/icu/ICUNormalizer2CharFilter.java : : Modified: lucene/dev/trunk/lucene/analysis/icu/src/java/org/apache/lucene/analysis/icu/ICUNormalizer2CharFilter.java : URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/analysis/icu/src/java/org/apache/lucene/analysis/icu/ICUNormalizer2CharFilter.java?rev=1586473&r1=1586472&r2=1586473&view=diff : ============================================================================== : --- lucene/dev/trunk/lucene/analysis/icu/src/java/org/apache/lucene/analysis/icu/ICUNormalizer2CharFilter.java (original) : +++ lucene/dev/trunk/lucene/analysis/icu/src/java/org/apache/lucene/analysis/icu/ICUNormalizer2CharFilter.java Thu Apr 10 21:30:53 2014 : @@ -104,7 +104,9 @@ public final class ICUNormalizer2CharFil : : // if checkedInputBoundary was at the end of a buffer, we need to check that char again : checkedInputBoundary = Math.max(checkedInputBoundary - 1, 0); : - if (normalizer.isInert(tmpBuffer[len - 1]) && !Character.isHighSurrogate(tmpBuffer[len-1])) { : + // this loop depends on 'isInert' (changes under normalization) but looks only at characters. : + // so we treat all surrogates as non-inert for simplicity : + if (normalizer.isInert(tmpBuffer[len - 1]) && !Character.isSurrogate(tmpBuffer[len-1])) { : return len; : } else return len + readInputToBuffer(); : } : : : -Hoss http://www.lucidworks.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
