Byte or UTF-8 Char as Unit Release

Changes made with this release:

• Use bytes or UTF-8 Characters are the basis length unit
• Allow non-contiguous columns to be parsed from underlying file
• Better malformed line handling
• Remove column based parsing
• General speedup, cleanup, coverage improvements.

Enhanced row calculations and beginning .7 support

Streaming data into a data frame a field at a time appears slow. In an effort to speed this up I am working to pull a whole column at at time from the source file the first step in this process was proper detection of the number of valid rows in the file and whether any malformed rows are present. This work has been completed with this release.

The first steps to .7 support have also been added in this release. I hope to keep working to get travis nightlies working with the next release.

Malformed file and CI fix

Fixes a malformed file as well as removes malformed tags from repository to fix CI issues.