Refactor: Extract new crate simple-git

[NobodyXu]

binstalk-downloader contains stuff about http(s) before the git code is moved into it and now it becomes http and git.

While git indeed uses http stuff, which is why I decided to put it into binstalk-downloader, it is more than just downloading since it is stateful (can be cached locally and updated) where as http is stateless.

Also binstalk-downloader's codegen time now increases dramatically and it also creates extra dependencies for binstalk-fetchers, delaying its execution.

The git code also don't use anything from binstalk-downloader at all, it makes sense to be an independent crate.

[NobodyXu]

@passcod @Byron Just realize that progress_tracing implementation actually uses tokio, so that part cannot be merged into gix trivally.

[NobodyXu]

On x86_64h-apple-darwin, this indeed improves compilation time a lot:

In other targets where caching is effective, the build time is mainly limited by:

• the order of compilation: cargo decided to compile and run cargo-binstall build-script quite late in the process
• not enough CPU cores

But this is still a win though given that simple-git took significantly longer to compile than binstalk-downloader, even if it contains less code.

[Byron]

I really appreciate the compile-time work you are doing here and since you are determined to do so, if you want to lead this in gitoxide and start earlier than I planned, we can do that.

I'd rather learn to work with the difficulties that come with extensive feature toggling than to pass up on the opportunity. The work to improve compile times has to be put in at some point anyway, and starting early allows for these costs to be incremental, especially if the workflow issues around feature toggles are resolvable (i.e. optimizing compile times shouldn't be a permanent tax on development time).

[NobodyXu]

@passcod Invitation sent!