Wrong mail address, again :>
--------
Hi Somik,
I thought I'd brief you on how my investigation in the SwingParser was
going. I took your CVS module and managed with some changes to integrate
it into Swing's JEditorPane HTML renderer to make a simple HTML browser.
It soon become apparent however that the renderer requires perfectly
formed HTML. After playing with the idea of trying to fix bad HTML
myself, I realised the enormity of this task and looked for an existing
implementation.
JTidy (http://www.sourceforge.net/projects/jtidy), a port of a C library
(HTML Tidy), is another SourceForge project which performs HTML
validation and pretty-printing. It produces a DOM of the HTML page from
an InputStream from which I performed the relevant callbacks. The result
is a good replacement for Sun's DocumentParser, and it produces a nice
output of what was wrong/fixed during parsing. I am still trying to
determine whether the 174kb it adds on to any project is worth it tho
(and if there are any performance implications).
I haven't checked my code back in since it longer depends on htmlparser
in any way, but I can send it to you if you're interested.
-craig
|
