Changed tokenizer to accept digits that start with the deciaml separator

[jhudson]

Hello,

I use exp4j for very simple but sometimes long equations where the values can be expressed as .002 or .1, etc. Currently the Tokenizer doesn't recognise these as numbers and throws an UnparsableExpressionException exception.

This push request adds this ability and tests cases for it, I have also added a test suite to test all test cases in one go.

[fasseg]

Hey jhudson!

Thanks for the pull request. Unfortunately I'm planning to release a new Version of exp4j using a new Tokenizer which resides in the branch exp4j-fast. This tokenizer does correctly tokenize decimals without a leading zero.
So I kinda retired the main branch and hope to release a new version over the next couple of days.