Hi Alex,
Thanks for the easily reproducible test case. The German stemmer has been fixed, a new snapshot is available [1].
Greetings to the other end of the lake, Christian
[1] https://files.basex.org/releases/latest/
On Thu, Jun 17, 2021 at 6:21 PM Alexander Witzigmann alexander.witzigmann@tanner.de wrote:
hi,
if you call ft:normalize using a string containing "*" ft:normalize("a*",map { 'stemming': 'true', 'language':'de' })
this results in
Improper use? Potential bug? Your feedback is welcome: Contact: basex-talk@mailman.uni-konstanz.de Version: BaseX 9.5.2 Java: Oracle Corporation, 13.0.2 OS: Windows 10, amd64 Stack Trace: java.lang.ArrayIndexOutOfBoundsException: Index -1 out of bounds for length 8 at org.basex.util.TokenBuilder.get(TokenBuilder.java:191) at org.basex.util.ft.GermanStemmer.resub(GermanStemmer.java:150) at org.basex.util.ft.GermanStemmer.stem(GermanStemmer.java:42) at org.basex.util.ft.Stemmer.nextToken(Stemmer.java:99) at org.basex.util.ft.FTLexer.nextToken(FTLexer.java:156) at org.basex.query.func.ft.FtTokenize.tokens(FtTokenize.java:49) at org.basex.query.func.ft.FtNormalize.item(FtNormalize.java:23) ....
anything I missed?
Alex