:-) I changed the main lines and compiled the QueryParser.java after that I compiled the entire package, build a new jar file and installed it on my server. (I followed the same procedure I used when I added a new analyzer)
Cheers Chris -----Original Message----- From: Erik Hatcher [mailto:[EMAIL PROTECTED] Sent: 04 July 2005 18:33 To: java-user@lucene.apache.org Subject: Re: free text search with numbers A couple of sanity checks... you changes the main lines, not the comments I hope :) And also, you need to rebuild the parser code by running JavaCC on it. If you're tinkering with a copy of Lucene's source code you can run the Ant target "javacc" and you must have JavaCC installed per the build instructions. Erik On Jul 4, 2005, at 11:38 AM, BOUDOT Christian wrote: > I have found in the QueryParser.jj those lines of comments: > > // OG: to support prefix queries: > // http://nagoya.apache.org/bugzilla/show_bug.cgi?id=12137 > // Change from: > // | <WILDTERM: <_TERM_START_CHAR> > // (<_TERM_CHAR> | ( [ "*", "?" ] ))* > > // To: > // > // | <WILDTERM: (<_TERM_CHAR> | ( [ "*", "?" ] ))* > > > So as indicated I changed my line to the second option but it > didn't solve > the problem I still getting the lexical error. Do I have to do > something > special to this .jj file or is it read when the QueryParser.java > file is > compiled? > > Cheers > Chris > > > -----Original Message----- > From: Erik Hatcher [mailto:[EMAIL PROTECTED] > Sent: 04 July 2005 16:15 > To: java-user@lucene.apache.org > Subject: Re: free text search with numbers > > > On Jul 4, 2005, at 9:02 AM, BOUDOT Christian wrote: > > >> Hi, >> >> I modified the analyzer (it is now vegetarian and won't eat numbers >> anymore >> :-) but I have hit a new problem. The parser won't accept a keyword >> to start >> with a wildcard character. (*/12/2003) Any hints to solve this new >> issue? >> > > This is by-design "issue" with QueryParser (to avoid WildcardQuery's > from being created that run through every term in the index). It is > the way the grammar has been defined. You would need to build your > own QueryParser to change this behavior. The relevant QueryParser > grammar piece is this (from QueryParser.jj): > > | <WILDTERM: <_TERM_START_CHAR> > (<_TERM_CHAR> | ( [ "*", "?" ] ))* > > > Erik > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
