Revision history for Perl extension Spidy.
0.21 Fri Apr 15 2005
        - Added support for multiple 'graphics builders'.
          - abstracted the Image Magick command line support 
          - added support for the perl Imager module
          - added support for the PerlMagick Image::Magick module
        - fixed some really bad logic for database (hierarchy) inserts
          
0.20 Tue Apr 23 2003
        - minor tweaks dealing with command line arg parsing
        - added flags to Spidy to be able to control the parsing
          a bit more (expand patterns, autogroups, and names).
        - abstracted output_name paramater to a subroutine so it can
          deal with custom sizes better.
        - fixed minor issue with spaces if output file names.
        - dont create html in database mode
        - ConfigWriter saves global (custom size) information
        - Total rewrite of databse mode - working with Mysql/PostgreSql'
        - fixed minor isolate bug
        - capture errors from Image::Size (file not found)
        - added rfe #670733: rewrote mod_perl modules Spidy::Registry
          ... now working
        - working for basic CGI mode now
        - fixed bug #658855: if DBD library does not exist the Spidy
          module will fail to load.
        - fixed bug #670729: column count was not getting set

0.12 Sun Nov  3 2002
        - added Spidy::ConfigWriter.  This will allow you to convert a
          Spidy::Tree structure back to xml.
        - Moved all GetOptions commands from spider to Spidy.pm so 
          other tools can use the GetOptions implicitely.
        - Many, many changes so Spidy could be used in Santeri Paavolainen's
          gui editor 'spiddit'.  Most of the changes were superficial, 
          just added a few new interface functions into the Spidy backend.
        - Works with Windows!  Thanks to DJ Montoya for getting this working!

0.11 Sun Oct  6 2002
        - added --watermark option (along with accompanying --gravity
          option) to add watermark text to an image.  Added watermark
          boolean flag to sizes, to see if any watermark should be
          applied.  Also the output image must be a png or gif image.
          Most of this was contributed by Santeri Paavolainen
        - The Graphics system calls now use system(@cmd) instead of
          system($cmd) to prevent shell expansion.  Also from Santeri.
        - fixed minor bug with autogroups and outputImagePathType of
          relative.
        - added --outputImageType to force all images to be gif,jpg,png etc.
          The default if nothing.  So the output images will be the 
          same type as the input images.
        - fixed minor issue with quoted strings in the config files
        - now allows for comments in the config file (with leading '#'
          characters)
        - update the test scripts to test the watermark, gravity
          and outputImageType options.
        - added spider.cgi to the dist.  You can use spider in cgi mode
          where it will not create any html images.  You will still 
          have to create the images before the cgi mode will work properly.
          This allows for dynamic web content.  You can modify the template
          files to include what every you want.  It is also very easy to
          overload the newly added Spidy::CGI::Group/Image classes to design
          a program for many odd purposes.
        - fixed an bug with the reported "Image" counts on the group pages.
        - modified the tests to check for correct image counts.

0.10 Tue Aug  6 2002
        - added rfe 589551.  The --config option will now load in options
          for spider to run from the config file specified.  If the --config
          option is not used, but there is a 'config' file in the current
          directory, spider will load that.  If you use 'spider --config
          foobar --sizes small' then sizes small will override any values
          already in config.
        - bug where the 'small' size was excluded with the --sizes option,
          and the index.html was getting linked to non-existant
          small_index.html.  This was reported by Andre Torrez.
        - fixed but with t/000loader.t where DBI and Apache::Constants 
          needed to be installed.  Since those modules are not working
          currently, I have commented out those tests.
        - fixed minor bug with the icons creation while using CVS. CVS 
          files were getting copied into the icons directories and 
          tests were failing because of the extra files.  It now skips
          CVS related files.
        - add customizable sizes.  If adding sizes add a <gallery>
          tag around the entire document.  Then withing a <gallery>
          tag define <size> or <group> tags as needed. The attributes
          to <size> are padding, max_height, max_width, quality.
        - fixed bug with icons not getting created when the 
          bg_color changes in the xml document.
        - added 'original' image size.  It is not default, so you
          need to add it to the --sizes list to include it.
        - updated test
        - updated documentation

0.09 Thu Aug  1 2002
        - added a bunch of test scripts in the 't' directory.  Removed
          the test.pl script.
        - fixed rfe 553549. unified the way build_html and build_images
          deals with sizes.        
        - fixed bug 587666.  This was introduced while trying to fix
          a bug from another user.
        - Major overhaul of the internals
          * Now groups and images are now longer arrays, but 
            linked lists.  This lets me get at previous and next
            images much easier.
          * The HTML::Template loading is much different.  Previously
            all the data to be displayed in a template had to be present
            in a group or image object.  Now all the template variables
            can be dynamically generated in the set_param_data
            routines in the Spidy::Group or Spidy::Image modules.
          * major simplification to the templates group.tmpl and
            image.tmpl.  A lot of the logic has been pulled out of the
            template and placed in the set_param_data functions where
            it belongs.
          * now data in the xml file will override data presented to
            the spider command line tool.  Previously the spider tool
            overrode data in the xml file.
        - fixed --bundleDownload option
        - Made the spider output more verbose.  The verbosity 
          will be more contollable in future releases.
        - removed hard coded requirement for Spidy to be installed
          in the perl site_perl/<version> directory.  It now installs
          the icons and gets the templates relative to which ever Spidy.pm
          is currently being used.
        - added --parentUrl to make a "Up One Level" link for the
          top level group.

0.08  Mon Jul 22 2002 
        - Fixed 'count' bug where the number of images in a group
          was getting mis-reported.  Bug reported by Richard Clamp.
        - Fixed next image bug where the next image icon was broken
          while using 'reative' outputImagePathType.  This was
          also reported by Richard Clamp.
        - changed the default behavior of outputImagePathType to 'fixed',
          so now all the output images will be placed in the default
          directory of 'images'.  This is now the default because after
          talking to several people, they all used this for each gallery,
          so it makes sence to have the default be this.  If you want your
          images next to the html files, then you now need to specify
          --outputImagePathType 'relative'.
        - made it so multiple comment blocks work, so you can add
          many 'file' comments to a single image by just adding more 
          <comment> blocks with the 'file' attribute.
        - updated documentation
        - fixed bug where parsing failed when xml was inside comment
          tags.
        - made updates to the sample gallery.xml
        - fixed bug 584679.  --sizes was giving an error instead
          or reporting possible size options.

0.07  Mon Jul  1 2002
        - Added support for crosslinking groups.  So now you can click
          'next' and go all the way though a gallery.  To disconnect
          the groups add an isolate="1" argument to the group tag.
        - added autogroup documentation
        - fixed --sizes <list> bug.
        - added 'file' attribute to <comment> tags, so they could be include
          from external sources.  The 'prepend' attribute is also added, 
          so you can specify the contents to be prepended, instead of the
          default of appended.
        - fixed minor html issues in image.tmpl
        - updated documentation for the isolate option mentioned above,
          and added info for the <comment> changes.
        - added inputFilePath option to spider to specify the location
          of the files included for the <comment> tags.

0.06  Tue Jun 25 2002
        - various minor bug fixes
        - Original images that are modified will have output images
          recreated.  Patch provied by Richard Clamp.
        - fixed bugs with 'shortname' inheritance
        - added rfe  #553547. Deprecated <node> keyword.  It is replaced by
          <group>.  <node> is still supported, but creates warnings now.
        - removed support for multiple <copyright> blocks.  Removed 
          Copyright object.
        - added error messages for XML parser to complain about
          multiple <title>, <copyright>, <comment>, etc tags within
          a single <group> or <image> block.
        - removed the mkxml.pl utility.  The functionality has largly
          be replaced with the <autogroup> tag.
        - Updated documentation
        - removed xlarge as a default.  Can still add it in with
          --sizes teeny,small,medium,large,xlarge
        - Fixed several xml parsing issues.  Now create better 
          warning messages and catch several common mistakes.
        - fixed bug #553141.  If user did not give any
          subgroups or images, it would error out. Not the xml
          parser creates an error with appropriate message.
        - fixed bug #553309.  It was not giving correct default
          names to groups.  So there may be many groups with the
          same name.  Now it just numbers the Groups as "Group 1" ...

0.05  Fri May 24 2002  
        - Changed --debug to --verbose.  It seems more logical to me.
        - Now all "comment", "copyright", "link" and "title" blocks
          automatically wrap the data with <![CDATA[...]]> so you 
          no longer need <html>...</html> if there is xmlish data 
          inside the tags.  As a result <html>...</html> is depricated.
        - fixed bug where undefined node comments where being written
          like "HASH(0x844ddf4)"
        - added --outputImagePathType = <fixed|relative> so from the command
          tool spider, you can force all the images to be put in a single
          fixed diretory.  The default if relative.
        - fixed bug #553122, dealing with upper case tags and paramaters.
        - fixed bug #553143 dealing with bg and fg color setting from spider
        - added rfe #553150.  This adds the <autogroup size="x"> xml tag.  It
          will automatically chop up the node into nodes of size "x".
        - fixed bug #553550 where --bgcolor <color> was not getting turned
          into the hex equivalent of the color name
        - fixed bug #553125  The image order of a patter: named image 
          was being set by the directory readdir order (alphabetical)
          and ignoring the order specified in the regex.
        - added rfe #553304  The name attribute is required in <image>
          blocks and now gives an appropriate xml parse error with
          context.
        - added rfe #553126  The ability to use variables is not in the
          xml.  If you use the <var name="foo">text</var> tags, then 
          you can use &foo; later in <comment> and other text blocks.
        - added rfe #553301  Now the xml parser gives an error message
          with context if there is a node and an image at the same level.
        - fixed bug #553148: the td width and height were getting set wrong
        - fixed bug #553315: the icons moved down when there was a mouse over
        - fixed bug #553144: clicking on mouse over images gave an error
        - fixed bug #553121: made current image size a non-link in the bottom
          navigation part of the page
        - added rfe #553313: added <shortTitle> tag.  This is to be used as
          and abbreviated title where space is critical, like the names of nodes
          and the limited space for the site hierarchy menu on the bottom right
          side of each page.
        - added rfe #553548: title and shortTitle can be a paramater to a <node>
          or <image> tag instead of having to specify a <title> or <shortTitle>
          block
        - fixed bug #553145: fixed issue with the newlines for comment and other
          tags getting removed by XML::Parser.

0.04    Thr May 2 2002
        - fixed bug #483651, the html_path of output_images was getting
          set incorrectly.  Reported by Jud Dagnall
        - fixed --sizes issue with spider and Tree.pm
        - fixed max height problem for images getting created with
          different heights.
        - added support for mouse over images.  Just specifiy 
          <image name="..." mouseOver="..."/> and the image will
          automaticaly swap images when the mouse is over the image.

0.03  Thr Aug 16 2001
        - Database create/drop/load work for PostgreSql, Mysql and Sybase.
        - Added an Apache mod_perl interface for speed.  This is still
          pretty kludgy.
        - Made it possible to use Spider Eyeballs from CGI scripts so 
          you dont have to create static html files.  This will work
          by loading the data everytime from the xml file (can be very
          slow) or by using a database.
        - yanked out all the autoloading rountines.  Now the objects
          are tied hashes for developement to allow for name checking
          and argument validation, but this can be turned off to allow for 
          direct hash look ups since tied variables are very very slow.
        - added support to page colors.
        - added brief documention to 'spider' describing color and datbase 
          options.
        - added a cheesy script (mkxml.pl) to automatically create 'quick 
          and dirty' xml files from a a large batch of images.  It is a kludge,
          and this functionality should be wrapped into 'spider'

0.02  Sun Apr 15 2001
        - documented the spider application
          add new features to spider

0.01  Sun Feb 25 2001
        - 100% rewrite of application, started over from scratch
          except user interface.  Added XML configs and HTML::Template
          templates. Cory Bennett
	- original version; created by h2xs 1.19

0.00  ~ 1997 - 1999
        - first version and user interface created by Julian Cash
