Changelog

0.2.0 (5 September 2014)

Rename a number of methods to be clearer and more consistent with that they actually do:
- EmojiChar.char() → EmojiChar.render()
- EmojiData.find_by_unified() → EmojiData.from_unified()
- EmojiData.find_by_str() → EmojiData.scan()
Don't worry, the old names are still aliased in so you don't have to change anything in your existing code. This change is make things clearer for people new to the library.
Add new .from_short_name() library method for fast keyword lookups.
DEVELOPERS: Internal code cleanup and better comments.
DEVELOPERS: Add benchmark suite for comparing method implementation time across versions of this library.

Add support for Unicode variant encodings, used by MacOSX 10.9 / iOS 7.
- For more info: http://www.unicode.org/L2/L2011/11438-emoji-var.pdf
- By default, EmojiChar.to_s() and .char() will now use the variant encoding.
With adding support for variants, the speed of find_by_str regressed by approximately 20% (because there are more codepoints to match against). To counter this, we switched to a Regex based scan than improves performance of the method by over 250x(!). A complete sorted search against 1000 strings now takes ~2ms where before it would take around a half second.
Import latest version of iamcal/emoji-data.
100% test coverage. 😎

On initialization, create hashmaps to cache lookups for .find_by_unified().

In a quick benchmark in MRI 2.1.1, this reduces the time needed for one million lookups from 13.5s to 0.3s!

This is only for lookup by unified ID for now, since the other find_by_*() methods are actually searches that can return multiple values. I'll look at nested hashmaps for those if there is a pressing performance need later.