jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Marcel Reutegger (JIRA)" <j...@apache.org>
Subject [jira] Resolved: (JCR-189) QueryHandler should use lucene Input-/OutputStream implementations
Date Wed, 17 Aug 2005 14:01:55 GMT
     [ http://issues.apache.org/jira/browse/JCR-189?page=all ]
Marcel Reutegger resolved JCR-189:

    Fix Version: 1.0
     Resolution: Fixed

The changes for this improvement required some changes in the configuration:

- The FileSystem element in SearchIndex is now optional. Whether the configuration must have
a FileSystem element depends on the QueryHandler implementation. If the QueryHandler does
not use a FileSystem, it can be omitted. Which is now the case for the default Jackrabbit
QueryHandler (org.apache.jackrabbit.core.query.lucene.SearchIndex).
- When using the default Jackrabbit QueryHandler the parameter 'path' is now mandatory. See
sample config src/config/repository.xml

I tried to adapt all config files accordingly (also those under contrib). Hope I didn't forget

svn revision: 233174

> QueryHandler should use lucene Input-/OutputStream implementations
> ------------------------------------------------------------------
>          Key: JCR-189
>          URL: http://issues.apache.org/jira/browse/JCR-189
>      Project: Jackrabbit
>         Type: Improvement
>   Components: query
>  Environment: svn revision: 233161
>     Reporter: Marcel Reutegger
>     Assignee: Marcel Reutegger
>     Priority: Minor
>      Fix For: 1.0

> Currently the QueryHandler uses a jackrabbit specific implementation of the lucene Directory
interface to make use of the jackrabbit FileSystem abstraction. Lucene operations on the file
system however requires quite often random access on the index files. With the current FileSystem
interface / abstraction random access is not possible on a FileSystemResource, therefore it
is simulated by re-aquiring the InputStream and then seeking to the desired position. This
it not efficient at all.
> With respect to performance any other use than file based index storage does not make
sense with lucene. Hence, the current abstraction using FileSystem should be dropped in favour
of direct file access.

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators:
For more information on JIRA, see:

View raw message