Revision history for Perl extension HiLiter.

0.01  Wed Jun 30 15:11:29 2004
    * - original version; created by h2xs 1.22 with options
    * -X -n HiLiter -b 5.6.1

0.01 - 0.04 testing

0.05 Fri Jul  9 12:48:34 CDT 2004
    * - public release via CPAN

* 0.06
    * use Text::ParseWords instead of original clumsy regexps in prep_queries()
    * add support for 8211 (ndash) and 8212 (mdash) entities
    * tweeked StartBound and EndBound to not match within a word
    * fixed doc to reflect that debugging prints on STDOUT, not STDERR

* 0.07
    * made HTML::Parser optional to allow for more flexibility with using methods
    * added perldoc for previously undocumented methods
    * corrected perldoc for Queries() to refer to metanames as second param
    * updated SWISH::API example to avoid using HTML::Parser
    * added unicode entity -> ascii equivs for better DocBook support
    * (NOTE: this expands the ndash/mdash feature from 0.06)
    * misc cleanup

* 0.08
    * fixed bug in SWISH::API example with ParsedWords and updated Queries()
      perldoc to reflect the change.
    * removed dependency on HTML::Entities by hardcoding all relevant entities.
      (HTML::Entities does a 'require HTML::Parser' which made the parser=>0 feature break.)

* 0.09
    * added Print feature to new() to allow Run() to return highlighted text instead
      of automatically printing in a streaming fashion. Set Print=>0 to turn off print().
    * Run() now returns highlighted text if Print=>0.
    * changed parser=>0 to Parser=>0.
    * the ParsedWords bug reported in 0.08 was really with my example in get_snippet().
      so rather than blame someone else's code, I fixed mine... :)
    * fixed bug with count of real HTML matches that was most evident with running hilite()
    * added test2.t test to test the Parser=>0 feature


* 0.10
    * fixed prep_queries() perldoc head
    * Queries() now returns hash ref of q => regexp
    * fixed SWISH::API example to use new Queries()
    * fixed Queries() perldoc
    * added StopWords note to prep_queries()
    * fixed regexp that caused make test to fail in perl < 5.8.1 (thanks to m@perlmeister.com)
    * added note to hilite() perldoc to always use Inline()

* 0.11
    * separated debugging into 3 levels for increasing verbosity (1-3).
    * changed default colors to lighter pastels.
    * misc perldoc fixes.
    * internal object key 'query_array' preserves query order.
    * fixed default HTML::Parser handler to buffer/print according to HiLiter object param.
    * added support for TextFilter and TagFilter.
    * renamed $debug to $Debug.
    * added 'debug' param in new() call.
    * fixed bugs with default StartBound and EndBound.
    * fixed bugs with &#NN; numeric entities in $White_Space.
    * added support to let phrases match over all non-WordChars, not just whitespace.
    * Queries() now returns either hash ref or array depending on context.
    * Queries() may now take either an array ref or a scalar text string.
    * Various small tweeks to better support the SWISHE param.
    * Queries() now keeps query array in same order as original.
    * moved all Changes here and out of .pm file
    * moved examples to their own directory and out of POD
    * renamed all private routines to start with _
    * added the 'nohiliter' attribute option, to prevent highlighting within marked tagsets
    * general cleanup and optimization
    * build_regexp() now returns an array ref of two regexp: for HTML and plain text.
    * build_regexp() now uses Ignore Chars from SWISH::API object
    * prep_queries() now supports SWISH Fuzzy Mode via SWISH::API::Fuzzy (version 0.03 or newer)
    * changed new() SWISHE param to SWISH -- but either will work
    * 
* 0.12
    * bug fix: plaintext() not called if text contains entities
    * mpeters@plusthree.com contributed the HiClass feature.
    * added filter example to lightfile.cgi

* 0.13  10 Nov 2005

* 0.14  26 Sep 2009
    * rewrite to use Search::Tools. At the same time considered replacing HTML::Parser with XML::LibXML for
      speed reasons, but when comparing the RT queues for both, it became obvious that HTML::Parser was a much
      safer route. That, and I couldn't get tests in XML::LibXML to pass against libxml2 2.7.
    * The API has changed. Read the SYNOPSIS.
    * since Search::Tools normalizes everything to UTF-8, the output of HTML::HiLiter will always be UTF-8.
      As a convenience, if the HiLiter encounters a http-equiv meta charset tag of anything other than
      ascii or utf-8, a new meta tagset will be inserted in its place indicating utf-8 encoding. If you
      really do not want to display UTF-8, you'll need to convert back to your desired encoding,
      using something like the Encode module.