version 0.6.4
=============
- If a file is not found in the $ROT_PATH envirnoment variable, it errors and
  spits out the value of the $ROT_PATH variable, rather than just saying that
  it's not found.
- Fixed the --parse option so that it only says the file parses correctly
  if it did.
- Fixed the default $ROT_PATH so that / is added to it.  This means that 
  you can now specify the full path name to a file on the command line.
  Oops.
- Fixed the --print option.  The header is now broken up into multiple lines,
  instead of one very long line.  This was because I forgot to append a "\n"
  to the text to be printed.  Sigh...
- Got rid of the extraneous "line too long" bug.  This brings the count up
  to three really stupid bugs fixed.  I believe it's time for release...

- Added the $ROT_NOWARN environment variable.  If this is set, rotfl will
  not say anything about warnings and will not generate the warning log
  in /tmp.  Just set $ROT_NOWARN to anything in your shell and warnings
  will get ignored.
- Line numbers are back, with one problem.  If you have a file with several
  records and skip around, any warnings you encounter will be reported as
  being on the wrong line.  Going straight through the file will report
  the correct line numbers.  I know what causes this and am working on it.
- rotfl can now find rotfl formatted files that are not specified with an
  extension.  If you do not provide an extension for the filename, it 
  will first look through $ROT_PATH for that filename, followed by the 
  filename with .rot extension.

- The $ROTPATH envirnoment variable has been renamed $ROT_PATH to maintain
  consistency with some new variables I might be introducing.  Sorry if
  this gets anyone.
- Updated the man page.  
- The documentation for rotfl has been moved to where it should have been
  all along - /usr/local/doc/rotfl (or whatever other prefix you choose.
  If you don't have a /usr/local/doc directory, one will be made for you.
- My email address has (finally) changed to <chris@bangmoney.org>.  Please
  update whatever accordingly.

version 0.6.3
=============
- All documents must now have a title.  If a record does not have a title, a
  warning will be generated.  This is because I am going to stay backwards
  compatible with older versions of rotfl for now.  Expect this to change
  once it hits 1.0.  The default name will still be "untitled".

- Added the E tag for the author's email address.  This is currently not 
  displayed, and is not set to any default.
- Added the M tag.  This is the same as the D tag, only it specifies when
  the record was last modified.  This tag is not required, and will not be
  set to any default if left out.
- Changed around the layout of the header.  I thought it would be better to
  line up the dates, so they are.  I moved the author's name down under the
  title.
- rotfl now makes use of the ROTPATH environment variable.  Just make this
  variable a colon delimited list of directories where you hide your rotfl
  files.  When it starts up, rotfl will automatically search each directory
  in ROTPATH (starting with the current directory) until it finds the file.

- Added the CHANGES file, which details changest that have been made to the
  source.  These changes may not be visible to the user, but are included
  so that anyone developing rotfl-related programs can see what source code
  has been modified.  The casual user does not need to be concerned with this.
- The extension for a rotfl file is now .rot.  Please use this for all your
  text files that are to be processed by rotfl so that you can tell what
  kind of file they are.
- Updated the documentation for the news tags and stuff.

version 0.6.2
=============
- Fixed some of my syntax to work with version of Perl that are older than
  mine (pre 5.005_03).  Thanks to Konrad Roziewski <cr@xox.pl> for this one.
- Issues with the name of the ReadKey module in the install script have been
  resolved.  They were actually resolved the day after I posted version
  0.6.1.  An updated one was immediately posted.  Sorry for making such a
  stupid mistake.
- I just realized that I am an idiot.  All my docs say that numbered lists
  were introduced in version 0.6.0.  But they we're really introduced until
  version 0.6.1.  The docs now reflect that properly.

- Changed the install.pl and uninstall.pl scripts.  install.pl now asks you
  for an installation prefix directory.  This is then automagically written
  into the uninstall.pl script so it knows where to uninstall from.
- Added the o and O tags.  These are the overline tags, and work in exactly
  the same way as the underline tags do.  However, they print a line above
  the text instead of under it.
- Added the -r and --print command line options.  This option just streams
  formatted output to the screen.  There is no opportunity for user
  interaction.  This is so you can send redirect the output to your printer
  and get a hard copy.  It looks rather bad right now.  Ideally, i'd have it
  insert a page feed between each record.
  
- Updated man page.
- Added the WARNING file to describe some upcoming dangers about the V tag.
- Prioritized the TODO file.
- Changed how I do the version number.
- This NEWS file is now grouped by what kind of changes I made.  The first
  group is bug fixes, the second is new features, and the last is miscellaneous
  and non-critical fixes (like this).
  
version 0.6.1
=============
- Problems with error messages getting printed out twice at the end of the
  program has been solved.  This was a stupid error.
- Another stupid error.  The date, author, title, and version is now reset to
  nothing at the start of each record.  I should have been doing this all along.

- Lists can now be numbered.  This is done by specifying the list's marker
  as a hash symbol (#).  You can nest numbered list, and each list will follow
  its own numbering system.  A hash can be used as the list marker by using
  \# as the marker.  I did this to make a system similar to how c and perl
  make you escape certain characters to use them literally.

  The numbered list is available only in records designed for versions >=
  0.6.1.  Older documents will still have a # as the marker, instead of the
  numbering.  It has now become important to use the version tag.  The default
  version for a record is the version number of the current version of rotfl.
  

version 0.6.0
=============
- Added the -p and --parse command line options.  Both of these only check the
  grammar of the input file.  If there are any warnings or errors, those
  messages are output as normal.  The actual formatted output is suppressed.
- I added a table of contents.  This is a listing of all the records in the
  file.  You press a number and can skip to that specific article.  You can
  jump around all you want.  Press 't' at the prompt to get to it.
- The prompt between screens of text now does now.  You can press various
  keys and do stuff.  Check HOWTO to see what you can do.

- Updated the man page accordingly
- New example file to show off the table of contents
- The record number and line number is no longer displayed on the prompt.  This
  is because I cannot do that with the current code and how the table of
  contents works.  I believe that the record/line numbers will be coming back 
  in the next version (0.6.1?)
- Articles without a title will be shown as "untitled" in the table of 
  contents.

version 0.5.10
==============
NOTE: This version was unreleased because it was all expreimental and stuff.
      I am getting the 0.6 release ready, and this version was broken all over
      the place while I decided how to go about doing a few things.  Also,
      more code cleanup stuff.

- Regular text that is in the header or in between records no longer causes
  a warning.

- Added a few command line options: -h, --help, -v, --version
- We are now using the Term::ReadKey library so that I can get more control over
  the keyboard.  Because of this, rotfl will determine your screen size and 
  output a screenful of text that fits your screen's size.
- The V tag is now enabled.  This is placed in the header and is used for you
  to tell which version of rotfl that the record was written for.  It is not
  currently displayed anywhere on the screen.  There's really no use for the 
  V tag yet, but there might be later.

- Code cleanup stuff: I reversed all the one-line loops, including while, for,
  and foreach.  I sure hope this doesn't confuse me later.  Also, I remove a
  lot of the stupid for loops and replaced them with the .. operator...it just
  makes things cleaner.

version 0.5.9
=============
NOTE: This version was unreleased because I was just doing some general code
      cleanup stuff.  I didn't do anything really deserving of a public 
      release...but I did enough to bump up the version number.  By the way,
      we are ramping up to the 0.6 release.

- Error file now named only using the file's name, not the path to it.
- An unrecognized tag has been moved from an error down to a warning, which 
  means that rotfl will no longer die if it sees a tag it doesn't understand.

- Change from using substr() to get tags to the faster (and cooler) unpack().
- Removed a close() and open() on the input file by replacing them with seek().
- General code clean-up stuff.

version 0.5.8
=============
- Error log now appears in /tmp by default.  This is because you may have some
  rotfl-formatted files laying around in places with tight permissions, and 
  you can't view them because it can't create an error log.

- Now deals with SIGINT.
- The error log is now only created if you have warnings.  Warning messages 
  are placed into an array as they occur.  When the program exits, the size
  of that array is checked, and any warnings are written to the log.

version 0.5.7
=============
NOTE: first publically released version.

- Fixed up error checking in the preprocessor.  It now makes sure the header
  and record tags are not nested as it counts up everything else.  This means
  that there shouldn't be any of those errors as the text is being displayed.
  This also means that I should be able to get rid of some of the later error
  checking, but I will probably leave it in for now.

- Back to using list markers.  A marker can be defined immeadiately after the 
  opening tag of a list.  This marker gets stuck on the front of every item in
  the list.  If no tag is given, the default is an 'o'.  These markers are 
  local only to the current list.  So, you can have a different marker for
  every list and repeat them and so forth.  Wahoo.

version 0.5
===========
- Changed how tags work.  They no longer are simply .%.  There's now an open
  form (.[) and a close form (.]).  This allows for cool things like nesting.
- Allow nesting of lists.
- Lists can no longer specify their marker.  This is because we could be nesting
  them, and we don't know what marker to come back to.  This could probably be
  fixed rather easily by making some sort of "marker stack".  In fact, that's
  probably the next thing I will work on...after a little testing.

before version 0.5
==================
- Didn't keep a changelog
- Wrote everything except what's in the changelog