Data-Maker - Changes

0.26 (11.08.2010) - Philip Garrett <philgarr@gmail.com>

Fixed a bug in Data::Maker::Field where fields were being cached globally, causing independent maker objects to clobber each other.

0.25 (05.17.2010) - John Ingram <john@funnycow.com>

Updated `t/benchmark.pl` -- apparently Data::Maker::Field::DateTime does not work without start and end dates anymore. Maybe I'll look into that.
Also found that Makefile.PL had the wrong version in 0.24. Fixed that.

0.24 (05.15.2010) - John Ingram <john@funnycow.com>

Data::Maker::Field::Number did some pretty dumb things. Much more straightforward now.
Also changed the type of the `precision` attribute of Data::Maker::Field::Number to PositiveInt

0.23 (05.05.2010) - John Ingram <john@funnycow.com>

Fixed a bug in Data::Maker::Field in which a formatting atom of 0 was ignored... in fact, I think everything after a zero would also be ignored. (because returning zero is no better than returning nothing)
Modified Data::Maker::Field::DateTime so that the `start` and `end` parameters can be either a year (as it was originally written) or an actual DateTime object, as I should have written it in the first place.
Added MooseX::Aliases as a dependency

0.22 (05.03.2010) - John Ingram <john@funnycow.com>

Once again, I neglected to update the version number in Makefile.PL. Added that to my checklist.

0.21 (05.02.2010) - John Ingram <john@funnycow.com>

Removed Data::Maker::Field::Currency from this distribution and created a standalone distribution for it.

0.20 (05.02.2010) - John Ingram <john@funnycow.com>

Significant (and mostly-accidental) performance improvement with Field object caching
Removed everything having to do with the "object_cache", due to the above performance improvement. Object caching is now done using the Moose method modifier "around".
Finally noticed that I have not updated the version field in Makefile.PL since version 0.01, so I updated that.
The header() method assumed that there was a `label` attribute. Now it uses `name` if there is no `label`
Added boolean attribute `allow_unknown` to the Gender class. If false, a random selection between M and F will be made instead of returning a gender of U
Updated the benchmarks in Maker.pm, due to performance improvement mentioned above
Added benchmarking script to the distribution (t/benchmark.pl)
Added Data::Maker::Field::Number (thanks to Adam Corum for giving me a much more efficient way to do numeric ranges)
Added Data::Maker::Field::IP
Added Data::Maker::Field::Password

0.19 (04.04.2010) - John Ingram <john@funnycow.com>

0.18 (04.03.2010) - John Ingram <john@funnycow.com>

Slightly-modified the first example in the Maker.pm POD, to match the same example in Tutorial.pod (thanks to David N. Blank-Edelman for pointing out the discrepancy)

0.17 (03.14.2010) - John Ingram <john@funnycow.com>

Added default delimiter of "\t", so that the delimited() method will return a tab-delimited record by default
Added more detailed documentation to Data::Maker::Field::Initials, to point out the two different ways to use that class.
There was already a "formatted" attribute defined for any field, which is a code reference that is to be applied to the value after it is generated but before it is returned. But it was not being used in the code, so I fixed that.

0.16 (03.07.2010) - John Ingram <john@funnycow.com>

Fixed an oversight in the benchmarking documentation (forgot to change the 100 to 200)
Modified the seeding code to make sure seeding never happens more than once
Added Data::Maker::Field::MultiSet
Added Data::Maker::Field::Initials

0.15 (03.07.2010) - John Ingram <john@funnycow.com>

This was a botched release. I accidentally released a copy of trunk when my changes had been made to a feature branch

0.14 (02.06.2010) - John Ingram <john@funnycow.com>

Brought back SSN field, which is a good demonstration of creating a Format subclass
Fixed some version numbers because I have not been paying enough attention to them lately

0.13 (02.06.2010) - John Ingram <john@funnycow.com>

Improved performance by more than 100% by changing the dynamic field methods to use AUTOLOAD instead of using dynamic Moose attributes. We'll see if there's a downside to this.

0.12 (02.05.2010) - John Ingram <john@funnycow.com>

Added some notes to the POD; one about performance and one acknowledging the existence of another module called Data::Faker
Added reset() method to Data::Maker, which sets generated() to 0
Added `from_name` attribute to Data::Maker::Field::Person::Gender, mainly for benchmarking purposes so I could benchmark the Gender field by itself. If provided, it will use that name instead of getting the name from the field referenced by `from_field`.
Tweaked the Maker.pm POD a little, adding a section on performance, including benchmark results.

0.11 (02.05.2010) - John Ingram <john@funnycow.com>

Two modules required by only one Field class each were not declared in Makefile.PL. I added them, though I need to consider making each Data::Maker::Field subclass that requires other CPAN modules having its own distribution and not being required for Data::Maker itself.
Added the Set and Lorem classes to the Data::Maker man page, because all classes are listed there -- which I forgot. I need a checklist! :)
Got rid of all signs of Data::Maker::Field::Person::SSN. I still want to have it as an example of how you can easily add your own Format subclass, but I obviously something since that original idea and I have to give it fresh thought.
Added Data::Maker::Field::Person, which right now is nothing more than a way to import all of the Person classes.
Added a `before` trigger to record_count to reset the `generated` to zero whenever `record_count` is changed. This might come back to bite me, but the only time it really makes sense to change the `record_count` is when you want to start over.
Added the Data::Maker->random class method for when you just need a really simple random pick. I added this specifically for hierarchical data, so you can have a maker object called in between `next_record` calls of another maker object that creates a random number of child objects. It made sense to use $maker2->record_count( Data::Maker->random ).

0.10 (02.04.2010) - John Ingram <john@funnycow.com>

0.09 (02.03.2010) - John Ingram <john@funnycow.com>

Now requiring that a "code" attribute be sent to Data::Maker::Field::Code
Removed erroneous import of MooseX::AttributeHelpers -- thanks to Andreas Koenig (RT #54237)
Removed lib/Maker.pm, which should have not been in the distribution.

0.08 (01.31.2010) - John Ingram <john@funnycow.com>

I keep forgetting to update $VERSION in the files when the version changes. Just updated them

0.07 (01.31.2010) - John Ingram <john@funnycow.com>

0.06 (01.31.2010) - John Ingram <john@funnycow.com>

0.05 (01.31.2010) - John Ingram <john@funnycow.com>

Changed $VERSION in the code to match the actual version (forgot to do that in 0.04)
Modified the description in the POD
Documented the individual Data::Maker::Field subclasses a little in the main Data::Maker page
Removed some Data::Maker::Field subclasses that should never have been in the distribution
Attributed my employer, ICA, for its support of the Perl community

0.04 (01.31.2010) - John Ingram <john@funnycow.com>

0.03 (01.31.2010) - John Ingram <john@funnycow.com>