Improve test coverage

[jsign]

This PR's main intention is to improve our test coverage.

As usually happens, looking at coverage profiles, it forces you to see some code in general, and that triggered some overall code improvements like removing duplicated-ish code, adding some extra validation checks, removing unused methods, etc. I’ll explain them in the comments.

Our pre-PR coverage numbers were pretty good. However, 1% of uncovered code in some files could be bad, tree.go in particular; this is natural due to organic code evolution. I ensured all important branches are covered now and left some in-comment TODOs in some places I’d like to dive into later to cover appropriately (proof generation/verification related).

Here’re the pre and post-coverage numbers (i.e: pre% -> post%):

• Overall coverage: 84.8% → 91.6%
• Per file:
  • config.go (100.0% → 100%)
  • config_ipa.go (74.1% → 88.9%)
  • conversion.go (80.5% → 80.5%)
  • empty.go (61.5% → 92.3%)
  • encoding.go (93.8% → 95.8%)
  • hashednode.go (62.5% → 87.5%)
  • ipa.go (73.3% → 92.3%)
  • proof_ipa.go (91.9% → 91.9%)
  • proof_json.go (80.7% → 80.7%)
  • tree.go (87.1% → 94.3%)
  • unknown.go (23.1% → 92.3%)

[gballet]

Nice, I left a few comments but LGTM overall.

[jsign]

Nice, I left a few comments but LGTM overall.

Thanks! I adjusted with some extra commits to many of your comments. I closed all convos except one that I still need to make a decision on; I'll wait for your feedback.

[jsign]

@gballet, all ready.

[gballet]

LGTM thanks