Rewrite XMLInputFormat in terms of chars not bytes. Fix handling of self-closing tags

[srowen]

This should supersede:

• Add test case for self-closing tag inside row #316
• Fix case for self closing tag inside row #321
• [fix] Self-closing tags are not supported as top-level rows #149

To make the fix, it seemed necessary to rewrite the logic in terms of characters, not bytes. Honestly I don't know why the original implementation worked in terms of bytes, even though I might have been part of it.

This does handle self-closing tags, but not nested copies of the same row tag. The latter is either something we need to drop or, will require another rewrite to handle sanely.

[codecov-io]

Codecov Report

Merging #352 into master will not change coverage.
The diff coverage is 0%.

@@          Coverage Diff          @@
##           master   #352   +/-   ##
=====================================
  Coverage       0%     0%           
=====================================
  Files          14     14           
  Lines         737    728    -9     
  Branches       57     66    +9     
=====================================
+ Misses        737    728    -9

Impacted Files	Coverage Δ
...cala/com/databricks/spark/xml/XmlInputFormat.scala	0% <0%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f7db55a...9701345. Read the comment docs.