Wrap transaction creation keys

[devrandom]

And make them public. This allows calls to InMemoryChannelKeys.

The wrapping of TxCreationKeys makes it obvious to signer implementers that the pre-derived keys are a local cache and should not be trusted in a validating signer.

[codecov]

Codecov Report

Merging #664 into master will increase coverage by 0.82%.
The diff coverage is 80.64%.

@@            Coverage Diff             @@
##           master     #664      +/-   ##
==========================================
+ Coverage   91.30%   92.12%   +0.82%     
==========================================
  Files          35       35              
  Lines       21400    23242    +1842     
==========================================
+ Hits        19539    21412    +1873     
+ Misses       1861     1830      -31     

Impacted Files	Coverage Δ
lightning/src/ln/chan_utils.rs	96.25% <66.66%> (-0.93%)	⬇️
lightning/src/chain/keysinterface.rs	95.04% <100.00%> (+0.01%)	⬆️
lightning/src/ln/channel.rs	89.47% <100.00%> (+2.82%)	⬆️
lightning/src/ln/functional_tests.rs	97.63% <100.00%> (+0.48%)	⬆️
lightning/src/util/enforcing_trait_impls.rs	100.00% <100.00%> (ø)
lightning/src/ln/channelmanager.rs	89.11% <0.00%> (+3.86%)	⬆️

---

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 d735a24...8058202. Read the comment docs.

[TheBlueMatt]

The two new commits basically LGTM. Would be nice to add a new sentence to the old TxCreationKeys docs to indicate that its the "trusted" version, and then can just make the full key set pub.

[devrandom]

doc and pub change done

[devrandom]

I made this PR independent of #662 and incorporated the TxCreationKeys constructor visibility change here.

[TheBlueMatt]

Oops, got trigger-happy with a "nit" commit. Oh well, we'll survive.