Found another bug in HTMLPaser. If input is HTML "CANON<br/>Model", CANONMODEL will be indexed instead of two words CANON and MODEL.
Thanks, Kevin ----- Original Message ---- From: Cheng Zhang <[email protected]> To: [email protected] Sent: Saturday, January 3, 2009 5:08:44 PM Subject: Re: search results It turns out that the org.apache.jackrabbit.extractor.HTMLParser eats all digits. in method filterAndJoin, all non-letters are removed. Does anybody has any idea why we do so? imo, index "hf100" makes more sense than indexing "hf". Or is there anyway I can configure to use my HTMLParser instead of the default? best, kevin ----- Original Message ---- From: Cheng Zhang <[email protected]> To: [email protected] Sent: Saturday, January 3, 2009 3:02:51 PM Subject: search results Hi, I have a html file as below stored in the repository. <html><body>Manufacture: CANON<br/> Model: HF100<br/> Title: Canon VIXIA hf100 Flash Memory High Definition Camcorder with 12x Optical Image Stabilized Zoom<br/> </body></html> However, if I search for 'hf100', it returns nothing. Any suggestion? Thanks a lot, Kevin
