Nom based parser

[khernyo]

This is a 🐛 bug fix.

Fixes #3 by rewriting the parser using the nom library. The new version should handle all errors, but more tests are necessary to make sure. Another benefit is that it's about twice as fast as before.

There is one problem though, error messages are hard to understand. It can be fixed, but it needs some work.

Checklist

• tests pass
• tests and/or benchmarks are included

Semver Changes

Minor level because of the deprecated function Header::from_vec.

[yoshuawuyts]

Heya, thanks for this patch!

To be fair I'm not entirely sure what to do here. I love the fact that the code is faster & fixes the input bug - but I find it really hard to read, and I'm not sure if I could update the code when needed.

Probably the most important bit right now would be to find a fix for the bug that the fuzzer showed. I'm less worried about performance bottlenecks, as this is unlikely to be a bottleneck. Do you reckon there might be a way to fix the byteorder based parser instead?

Hope it makes sense where I'm coming from. Thanks so much again!

[khernyo]

It makes sense. I'm curious, though, which part of it do you find hard to read? Is it primarily convert_nom_result() in header.rs or the parser itself in parsers.rs?

I'll see how hard it is to fix the bug without rewriting.

[khernyo]

Overridden by #6