Improved docs, fixed clippy warnings, fixed cargo fmt

[duesee]

When reading the code I made some minor improvements I wanted to share. Notably, I tried to make the usecase for Bech32 more clear and fix all clippy warnings and "incorrect" formatting.

Have you thought about introducing a CI (e.g., via GitHub actions) which checks against cargo audit, cargo clippy and cargo fmt?

[clarkmoody]

Interesting. I'll circle back after #41 is merged. We'll probably need a rebase then as well.

[sgeisler]

The changes look good, some transformations from index to iterator loops might even make the code faster.

The only problem I don't have a good solution for right now is how to handle the rustfmt::skip attribute in older rust versions (before macros were scoped). Currently it makes our tests fail for rust 1.22. Maybe we can make the attribute conditional and only run rustfmt with more recent rust versions. But I fear the error occurs when parsing the AST, so even disabling the code with cfg won't help. What we could do in that case is removing the attribute using sed when running the 1.22 test since it isn't needed there anyway, but that's an ugly hack.

If we accept rustfmt changes we should definitely add a rustfmt check to our already existing CI pipeline otherwise non-conforming code will be checked in again. Clippy is a bit of a problem since not all fixes are compatible with older rust versions. cargo audit seems most useful as a cron job and not for every PR, we could activate weekly builds in travis and only run it on master.

[clarkmoody]

Needs rebase.

[duesee]

Is there a reason to support Rust 1.22? Maybe it could be edition = '2018' anyway? :-)

[sgeisler]

The rationale for older rust versions is given in rust-bitcoin/rust-bitcoin#338, although I can't remember why it's exactly 1.22. 1.24 seems to be the version we really want to support, but I fear that won't help with our problem.

[duesee]

I don't see any easy way of supporting rustfmt's skip. #[rustfmt::skip] does not work in 1.22.0, Clippy lints on #[cfg_attr(rustfmt, rustfmt_skip)] and I assume #[warn(clippy::deprecated_cfg_attr)] will not work in 1.22 either nor should it be used. I just bypassed the formatting so that not to generate too much hassle with this PR. I don't know if CHARSET_REV needs any extra formatting, but if so, I guess it can be done in a similar fashion.

[duesee]

Regarding cargo audit: having it run on every PR prevents pulling in already vulnerable dependencies -- who knows? :-)

[sgeisler]

The changes already made look good to me. But to preserve the formatting over future PRs we really need to integrate rustfmt into our build pipeline (only for stable I guess). Do you know your way around travis? If not I could do it so we can merge the PR.

[duesee]

I would be glad if you could do it :-)

[sgeisler]

@duesee I'd be glad if you could take a look at the CI tests I added. I think they should check all the changes you did in this PR.

From my side it looks merge ready now. @clarkmoody what do you think?

[sgeisler]

Is there a better way to do this? Apparently cargo fmt doesn't support the --check flag, so I have to use rustfmt directly and specify the files to check.

[sgeisler]

Oh, I just noticed that there are merge conflicts now. I'll try to rebase.

[duesee]

Thank you @sgeisler! The CI should catch everything now. Having cargo audit run on every push might still be a good (and easy to implement) idea. But it could be added at a later time and issue.

[sgeisler]

Oh yeah, I totally forgot about that. Will amend the last commit tomorrow.

[clarkmoody]

This works for me.

@sgeisler We can update the Travis script with something more clever later, if need be.

[sgeisler]

I though about cargo audit again and it doesn't make much sense for us anyway since we don't have any dependencies.

[duesee]

You don't have dependencies now ;-)

[clarkmoody]

I think it should be a long-term goal that this library has no dependencies...