Changes for version 1.51 (Apr 26, 2006)
-------------------------
  - Patch to t/format_price.t supplied by Cosimo Streppone to allow it
    to work with non-US locales

Changes for version 1.5 (Feb 14, 2005)
------------------------
  - Require Perl 5.8 - was 5.003 before
  - Require POSIX.pm - was optional before
  - If C<THOUSANDS_SEP> is set to the empty string, format_number will
    not insert any separators.
  - Replace "use vars" with "our"
  - Support scientific notation - skip formatting
  - Change format_bytes to support options (precision, unit, and base)
    instead of precision number

Changes for version 1.45 (Aug 8, 2002)
------------------------
  - No code change; only change is to test suite.
  - Replace all tab characters with spaces

Changes for version 1.44 (Dec 11, 2001)
------------------------
  - Add support for GIGA_SUFFIX.  Can't do TERA because of integer overflow.
  - Detect decimal part correctly in Perl 5.6 and up

Changes for version 1.43 (Dec 10, 2001)
------------------------

  - Fix bug in format_price that caused an extra space to be inserted
    in some situations, and no space in others, between the currency
    string and the number.
  - Add support for gigabytes (formerly only kilo and mega) to
    format_bytes.  (Thanks to Oeyvind Flaam <flaam@flaam.priv.no>)

Changes for version 1.42 (May 17, 2000)
------------------------

  - Fixed bug in format_picture that was causing subsequent calls to
    return the previous result and the current one concatenated
    together.  This is actually due to a bug in Perl, but a workaround
    was implemented.  To avoid this problem in your own code, don't
    use code like this: "my VARIABLE = EXPRESSION if CONDITION"
    because if the CONDITION is false, the previous value of VARIABLE
    will still be there, even if it had gone out of scope!  This bug
    was pointed out by Buddy Burden <buddy@barefoot.net> on 21 Mar
    1999, Frank Tobin <ftobin@uiuc.edu> on 11 Jun 1999, Mark
    William Rostron <mrostron@us.oracle.com> on 12 Dec 1999, and Ben
    Tilly <ben_tilly@hotmail.com> on 16 May 2000.  Fix suggested by
    Buddy Burden, Mark William Rostron, and Ben Tilly.
  - Fixed bug in round() that was causing e.g. round(1.005, 2) to return
    1 instead of 1.01.  This is due to the vagaries of floating point
    arithmetic.  I changed .5 to .5000001 in the rounding formula and
    added a check for this in the test suite.  That seems to fix it
    but please let me know if you experience any weird results.  This
    bug was reported by Kentaro Takatomi <kentaro.takatomi@msdw.com>
    on 25 May 1999.
  - Fixed several points in the code where 'use of uninitialized
    value' warnings were generated.  Reported by Frank Tobin
    <ftobin@uiuc.edu> on 11 Jun 1999.
  - In format_picture, replaced the use of $_ with a locally scoped
    variable ($char).  This should reduce interference between this
    routine and any use of $_ by the calling program.  Bug report and
    suggested fix by Mark William Rostron <mrostron@us.oracle.com> on
    29 Dec 1999.

Changes for version 1.41 (Nov 30, 1998)
------------------------

  - Fix 'undefined value' warnings in format_picture
  - Fix bug in format_picture to show decimal point if and only if it
    is present in the picture string (bug reported by Michael Scoltock
    <scoltock@ictp.trieste.it>)
  - Fix bug in format_price that would suppress currency symbol if no
    decimal digits (bug reported by Michael Scoltock
    <scoltock@ictp.trieste.it>)

Changes for version 1.40 (Nov 27, 1998)
------------------------

  - Add new function format_bytes() to display '1K' for '1024', etc.
    (suggested by Curt Tilmes <curt@ltpmail.gsfc.nasa.gov>)
  - Fix bugs in round() and format_picture() when dealing with negative
    numbers.  NOTE: format_picture() now includes at least one extra
    space in its output for positive numbers.  This may present a minor
    incompatibility issue in some obscure cases.
  - Minor documentation updates.

Changes for version 1.30 (Oct 31, 1998)
------------------------

  - Fix several major bugs, including the POSIX-not-required issue, as
    well as better support for negative values.
  - Add setting for default size of decimal area so you don't have to
    specify the number of digits if it's other than 2.
  - Add setting and function to display negative numbers in different
    formats.
  - Add setting to force trailing 0's in the decimal portion to fill out
    the number of digits required.
  - Fix problem where 0's inside format_picture()'s output were being
    replaced with spaces.
  - Add 0 before decimal point if number is between -1 and 1.
  - Fix errors with supposedly-optional POSIX
  - From: Yair Lenga <yair@bpa1.sbi.com>
    Suggested adding controls to pad out result of format_number with
    zeroes as well as the ability to present negative numbers with
    parentheses.  Provided a patch which was useful in constructing a
    solution.
  - Add test for defined() from patch by IMAZU Hideyo <himazu@ms.com>
  - From: "Joseph Norton" <namaste1_@hotmail.com>
    Bugs in format_number(): processed negative numbers incorrectly,
    and lacked the zero before the decimal point if number is between
    0 and 1.

Changes for version 1.20 (Aug 31, 1998)
------------------------

  - Make POSIX no longer required (users on MicroSoft systems sometimes
    don't have it); uses US defaults if no POSIX.pm found.
  - Add new optional third argument to format_number() to add trailing
    zeroes.

Changes for version 1.12
------------------------

  - Remove 'use of undefined value' warnings when used with perl -w

Changes for version 1.10 (Nov 21, 1997)
------------------------

  - Add Locale support to set the default values for $THOUSANDS_SEP,
    $DECIMAL_SEP, and $CURRENCY based on the user's locale.
  - Add object oriented interface.
  - NOTE: Some aspects of the API have changed.  Please reread the
    documentation carefully if you used 1.0 previously!!!

Initial version 1.00
--------------------

  - Limited release beta version.