Fixes for issue #445: Wraps unexpected NullPointerException

[arthurscchan]

This PR provides a possible fix for issue #445 by adding a conditional check to ensure _cleanedTextValue is not null even if the input in invalid.

[cowtowncoder]

@arthurscchan In the meantime, one process thing: have we gotten a CLA from you? (I am not 100% sure; please LMK if you have).
If not, it'd be from here:

https://github.com/FasterXML/jackson/blob/master/contributor-agreement.pdf

and is usually easiest to print, fill & sign, scan/photo, email to cla at fasterxml dot com.
This only needs to be done once before merging the first contribution; after that it's good for any and all contributions.

Looking forward to merging tests, fixes -- thank you for contributing these!

[cowtowncoder]

@arthurscchan Thank you for providing the test! I was able to change the resolution slightly, figuring out that test triggers buffering and thereby the reason for missing _cleanedTextValue -- problems is that code was not using previously decoded int/long value but assuming buffered number String.