Upgrade and swap of dependencies, huge Refactoring

[mfelsche]

• Switch to latest stable rust (1.58.1)
• Switch HTTP framework from rocket to axum as rocket required nightly rust and updating it created conflicts with hubcaps. Also axum is quite new in the tokio ecosystem and is well maintained.
• move from hubcaps to octocrab for github api support, as using hubcaps with more recent versions of tokio and hyper (used for the http framework as transitive dependency) lead to conflicts which made this app fail to compile. (See Dependency conflict when using hubcaps together with hyper 0.14 softprops/hubcaps#299)
• Avoid spawning threads where possible and do everything async. This should be a little easier on resources.
• Fix failing LLVM IR output.
• Add stderr to output.

I know this is a huge PR and it has lots of stuff smooshed together. But it is an improvement in both code quality and recency of dependencies. It should also give a really small performance improvement. I would appreciate feedback from anyone with a Rust background and welcome all questions from anyone without.

The dependency on my fork of octocrab will be removed, once gist update support is merged upstream and released. I will take care of maintaining that dependency.

From local testing i can verify that the steps for installation and deployment are still the same. Nothing changed on the administrative side here.

All existing functionality still exists.

This PR can be easily verified locally with a rustup installation and a personal GITHUB_TOKEN.

[SeanTAllen]

I'm not qualified to review this. Personally, I put faith in you on it @mfelsche. Is there anyone you want to review this?

[SeanTAllen]

@mfelsche if no one weighs in on this by/during sync on the 15th, are you ok with having this merged sans any other eyes?

[mfelsche]

I am totally fine with this being merged without anyone else having a look at this. But then you gotta trust me, that everything is fine with this PR and no one is going to mine bitcoins with this.
I tried to keep the playground the same from a users viewpoint, so running it locally and verifying it does the right thing, should be enough, if noone wants to dive deeper.

[ergl]

I'm not knowledgeable enough about Rust to comment on the code, but I tried this PR locally and everything seems to work fine.

[ergl]

This wasn't covered in the sync call, but I think this should be good to merge. On my quick, unexpert review, everything looked fine. Do you agree @mfelsche ?

[SeanTAllen]

This doesn't work in prod.