Interesting... Thanks Robert for pointing this out! > "To obtain correct results for locale insensitive strings, use toUpperCase(Locale.ENGLISH)"
Actually this is one of the things I tried and did solve it - with toUpperCase(Locale.US) - not exactly Locale.ENGLISH but quite similar I assume - and as you suggest it felt wrong, for wrong reasons... Perhaps I'll change it like this, case insensitivity is a good think when running in various OS's. On Sun, Feb 6, 2011 at 6:55 PM, Robert Muir <rcm...@gmail.com> wrote: > Thanks for catching this Doron. Another option if you want to keep the > case-insensitive feature here would be to use > toUpperCase(Locale.ENGLISH) > > It might look bad, but its actually recommended by the JDK for > locale-insensitive strings: > > http://download.oracle.com/javase/6/docs/api/java/lang/String.html#toUpperCase() > > On Sun, Feb 6, 2011 at 11:43 AM, <dor...@apache.org> wrote: > > Author: doronc > > Date: Sun Feb 6 16:43:54 2011 > > New Revision: 1067699 > > > > URL: http://svn.apache.org/viewvc?rev=1067699&view=rev > > Log: > > LUCENE-1540: Improvements to contrib.benchmark for TREC collections - fix > test failures in some locales due to toUpperCase() > > > > Modified: > > > > lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/java/org/apache/lucene/benchmark/byTask/feeds/TrecDocParser.java > > > > lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/test/org/apache/lucene/benchmark/byTask/feeds/trecdocs.zip > > > > Modified: > lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/java/org/apache/lucene/benchmark/byTask/feeds/TrecDocParser.java > > URL: > http://svn.apache.org/viewvc/lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/java/org/apache/lucene/benchmark/byTask/feeds/TrecDocParser.java?rev=1067699&r1=1067698&r2=1067699&view=diff > > > ============================================================================== > > --- > lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/java/org/apache/lucene/benchmark/byTask/feeds/TrecDocParser.java > (original) > > +++ > lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/java/org/apache/lucene/benchmark/byTask/feeds/TrecDocParser.java > Sun Feb 6 16:43:54 2011 > > @@ -29,7 +29,12 @@ import java.util.Map; > > public abstract class TrecDocParser { > > > > /** Types of trec parse paths, */ > > - public enum ParsePathType { GOV2, FBIS, FT, FR94, LATIMES } > > + public enum ParsePathType { GOV2("gov2"), FBIS("fbis"), FT("ft"), > FR94("fr94"), LATIMES("latimes"); > > + public final String dirName; > > + private ParsePathType(String dirName) { > > + this.dirName = dirName; > > + } > > + } > > > > /** trec parser type used for unknown extensions */ > > public static final ParsePathType DEFAULT_PATH_TYPE = > ParsePathType.GOV2; > > @@ -46,7 +51,7 @@ public abstract class TrecDocParser { > > static final Map<String,ParsePathType> pathName2Type = new > HashMap<String,ParsePathType>(); > > static { > > for (ParsePathType ppt : ParsePathType.values()) { > > - pathName2Type.put(ppt.name(),ppt); > > + pathName2Type.put(ppt.dirName,ppt); > > } > > } > > > > @@ -59,7 +64,7 @@ public abstract class TrecDocParser { > > public static ParsePathType pathType(File f) { > > int pathLength = 0; > > while (f != null && ++pathLength < MAX_PATH_LENGTH) { > > - ParsePathType ppt = pathName2Type.get(f.getName().toUpperCase()); > > + ParsePathType ppt = pathName2Type.get(f.getName()); > > if (ppt!=null) { > > return ppt; > > } > > > > Modified: > lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/test/org/apache/lucene/benchmark/byTask/feeds/trecdocs.zip > > URL: > http://svn.apache.org/viewvc/lucene/dev/branches/branch_3x/lucene/contrib/benchmark/src/test/org/apache/lucene/benchmark/byTask/feeds/trecdocs.zip?rev=1067699&r1=1067698&r2=1067699&view=diff > > > ============================================================================== > > Binary files - no diff available. > > > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org > For additional commands, e-mail: dev-h...@lucene.apache.org > >
