[
https://issues.apache.org/jira/browse/TIKA-528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13004699#comment-13004699
]
Jukka Zitting commented on TIKA-528:
------------------------------------
FTR, the HTMLSchema class turned out *not* to be thread-safe in some
circumstances. See TIKA-599 for the problem and the fix.
> Reuse tagsoup HtmlSchema instance across HtmlParsers (performance improvement)
> ------------------------------------------------------------------------------
>
> Key: TIKA-528
> URL: https://issues.apache.org/jira/browse/TIKA-528
> Project: Tika
> Issue Type: Improvement
> Components: parser
> Reporter: Bruno Dumon
> Assignee: Ken Krugler
> Fix For: 0.8
>
> Attachments: htmlparser-reuse-htmlschema-patch.txt
>
>
> While parsing a set of small HTML files (email messages), I noticed using a
> profiler that about a third of the time was being spent in the construction
> of tagsoup's HTMLSchema class.
> Since this is (or appears to be to me) simply a data structure, it is thread
> safe and can be reused. Fortunately this can be done easily, as shown in the
> patch I will attach to this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
