Chris Hostetter wrote:
: Would you mind opening a Solr bug to fix LengthFilter, and i'll spin up a
Gah! ... Solr has it's own LengthFilter which is broken, but Solr's
LengthFilterFactory uses the LengthFilter provided by Lucene (which is
also broken) ... so can you open a Lucene bug to fix LengthFilter and a
Solr bug to deprecate LengthFilter and make it wrap Lucene's version :)
I've created SOLR-520 for the deprecation of oas.a.LenghtFilter and
LUCENE-1253 to keep track of the LengthFilter issue.
However I'm quite sure that this issue can happen will all filter that
may remove tokens, like StopFilter, infact you can easily put a
stopfilter after the worddelimiter (instead of the LengthFilter), and
obtain the same error if you put in stopwords.txt the letters 'u', 's',
'a' (or just try with 'S.T.S. and U.K.', as 's' and 't' are already in
stopwords.txt file).
I didn't stress the index, but AFAIK solr work perfectly with that
index, even if checkindex complains over it (I've played a bit in order
to retrieve the docs with this issue with queries, and they always
present in the results).
I will keep an eye on both mailing-lists as I would like to understand
what are the implication of this issue, if any, apart from checkindex
Thanks,
Walter
