Normalize output format handling (-t) to match input (-f)

[alerque]

Fixes #5.

There were a couple logic problems.

1. The results of processing the --to argument were getting clobbered by file name guessing
2. There were two variables being held with the target format, args.to and output_format. They were used in different places and it was possible for them to be out of sync, causing unexpected results.

This normalizes the whole mess on one variable (args.to_format) and handles setting it just like --from instead of ... differently. The only place other variables are used is in passing arguments to panflute, which apparently wants named arguments in a slightly different format.

I have not been able to test this very exhaustively with all the backends and situations that might be affected. Please proceed with caution on this one, and if something needs fixing I'd be happy to do it.

Depending on what order you merge this or #11 might need to be rebased before merging. Feel free to do ether one and I'll update the PR for the other.

[coveralls]

Pull Request Test Coverage Report for Build 42

• 20 of 23 (86.96%) changed or added relevant lines in 2 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-1.9%) to 87.547%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pancritic/main.py	14	17	82.35%

Files with Coverage Reduction	New Missed Lines	%
pancritic/main.py	1	86.73%

Totals
Change from base Build 36:	-1.9%
Covered Lines:	184
Relevant Lines:	205

---

💛 - Coveralls

[alerque]

I'm not familiar with testing frameworks in python and am not sure what, if anything, needs to be done about the coverage issue. Feedback welcome.

[alerque]

Rebased to resolve merge conflict...

[alerque]

Travis failure appears to be related to #13 again. I think this is working as expected.