Wildcard query only accepts 1 or more characters for prefixes (not 0 or more) [LUCENE-26] about lucene

[asfimport]

Lee Mallabone (migrated from JIRA)

Created an attachment (id=1465)
Fixes the bug by providing a correct, commented replacement for the 'wildcardEquals' method

[asfimport]

Otis Gospodnetic (migrated from JIRA)

Hm, I can't duplicate the behaviour you describe.
I have an index with a few travel-related entries.
When I search for travel I get 1 match containing "Land a hot travel deal ...."
When I search for travel* I get 2 matches. The first one is the same as the one
above, and the second one contains "Travelocity.com" in the indexed field.

Doesn't the fact that a search for travel* matches entries that contain only the
term "travel", and no other terms with "travel" as their base, confirm that
wildcards work properly, matching zero or more characers, even when used at the
end of the query string?

[asfimport]

Lee Mallabone (migrated from JIRA)

Is that definitely using the WildcardQuery class? I know the query parser
defaults to using PrefixQuery for wildcards at the end of the term. In
PrefixQuery, the bug does not seem to be present. However, I'm able to search
for words in my indexes, get results, then search again using WildcardQuery with
a * character at the end, and get zero results.

[asfimport]

Otis Gospodnetic (migrated from JIRA)

Could you please provide a class that is a simple self-enclosing test case
that demonstrates this?

[asfimport]

Lee Mallabone (migrated from JIRA)

Created an attachment (id=1488)
Simple piece of code that demonstrates an incorrect search result when doing a wildcard search.

[asfimport]

Otis Gospodnetic (migrated from JIRA)

Thank you, I applied the patch (well, pasted it, as I couldn't get patch to apply the
diff).

[asfimport]

ASF subversion and git services (migrated from JIRA)

Commit 1521851 from @ovalhub in branch 'pylucene/trunk'
https://svn.apache.org/r1521851

fixed bug PYLUCENE-26 (Martin Anon)