Better key output

[rigelrozanski]

Better key output:

gaiacli keys show foo

NAME:	ADDRESS:					PUBKEY:
foo	A954D44A8C703C3367AA90043D9FA1817E4D36B4	1624DE6220C1F063AEEF411280F0C30AB0F2325555E5FA6A4D66E2752282797DDA3AA3B270

gaiacli keys show foo --output=json

{
	"name": "foo",
	"address": "A954D44A8C703C3367AA90043D9FA1817E4D36B4",
	"pub_key": "1624DE6220C1F063AEEF411280F0C30AB0F2325555E5FA6A4D66E2752282797DDA3AA3B270"
}

• Update docs
• Updated all code comments where relevant
• Updated CHANGELOG.md

[codecov]

Codecov Report

Merging #886 into develop will not change coverage.
The diff coverage is 0%.

@@           Coverage Diff            @@
##           develop     #886   +/-   ##
========================================
  Coverage    67.01%   67.01%           
========================================
  Files           68       68           
  Lines         3614     3614           
========================================
  Hits          2422     2422           
  Misses        1008     1008           
  Partials       184      184

[cwgoes]

Tested ACK

I like the output flag, we should make that standard on any command that should be both human & machine readable - which should be all commands?

If tabular outputs are common, we could use a library - e.g. https://github.com/olekukonko/tablewriter - instead of manual '\t' characters.

[rigelrozanski]

@cwgoes thanks! Side comment - unless a library is doing something otherwise pretty unique and important its better for use to not add an extra import and add some more code internally - the big risk is that malicious code gets added to another library and is somehow able to infiltrate / introduce security holes in the cosmos codebase without anyone noticing - I think something as basic as the use of \t shouldn't require an import

[cwgoes]

Could be mitigated by version pinning libraries (we'll never need to update the tabulation library) - but agreed in general.

[jaekwon]

We're supposed to migrate to Bech32Cosmos. #263