lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Uwe Schindler (JIRA)" <>
Subject [jira] Updated: (LUCENE-1796) Speed up repeated TokenStream init
Date Mon, 10 Aug 2009 19:48:14 GMT


Uwe Schindler updated LUCENE-1796:

    Attachment: LUCENE-1796.patch

New patch that optimizes the iteration over the AttributeImpls using the computed State linked
list. This also adds the default buffer size to KeywordTokenizer, that got lost during the
move to the new API.

To test performance, I reactivated the clearAttributes() call in CharTokenizer.

If this is now all ok, I would like to fix this issue as soon as possible to be able to do
more perf testing with the optimized impls. The big hammer of isMethodOverridden is now removed
and speed came back to the original one (with some small slowdown caused by mixing old and
new TokenFilters together).

> Speed up repeated TokenStream init
> ----------------------------------
>                 Key: LUCENE-1796
>                 URL:
>             Project: Lucene - Java
>          Issue Type: Improvement
>            Reporter: Mark Miller
>            Assignee: Uwe Schindler
>             Fix For: 2.9
>         Attachments: LUCENE-1796.patch, LUCENE-1796.patch, LUCENE-1796.patch, LUCENE-1796.patch
>  by caching isMethodOverridden results

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message