[LUCENE-190] [PATCH] Demo HTML parser does not properly handle meta tag attributes.

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Component/s: modules/examples

Environment

Operating System: All
Platform: All

• Bugzilla Id: 27423
• Assignee: Lucene Developers
• Reporter: Matt Chaput
• Created: 04/Mar/04 08:05
• Updated: 28/Aug/22 11:15
• Resolved: 27/May/06 01:37

Description

Version 1.3final.

The meta tag parsing in the demo HTML parser
(demo/org/apache/lucene/demo/html/HTMLParser.jj) incorrectly relies on the meta
tag's "name" attribute coming before its "content" attribute. In XML/HTML,
attribute order is supposed to be insignificant.

So, if I have tags:

<meta content="blah" name="blarg" />
<meta content="gluh" name="glarg" />

...the parser will not parse them correctly. (In fact, it will simply fill in
name/content pairs as it encounters attributes in the stream, without regard to
which meta tags the attributes are actually in. So, in the above example, I will
get one meta property of "blarg"="gluh".)

This is a problem because my XSLT happens to result in meta tags with attributes
in the above order.

It may not seem like a big deal since it's in demo code, but because
HTMLParser.jj is many times faster than more heavy-weight solutions, I'd love
for this to be fixed, if possible.