Implement CustomLogPipes for Spin Logging/IO

[danbugs]

That's exactly right. If you have your WasiFiles already, you can call ModuleIoRedirects::new directly, and pass that to prepare_component. (At least, you'll be able to if that PR is accepted.)

WAGI does this in a very simple form because it needs different behaviour for stdin, and a slightly different interaction with the buffer for stdout. It's really valuable to know that this is more broadly useful as a public customisation point though!

To provide an update on this from #438 (and maybe close that issue soon), it turns out that I couldn't really make the current architecture use our own pipes — As previously implied, the meat and potatoes of logging is done in execute:
https://github.com/fermyon/spin/blob/c3ff7d90fe115daa3fa85ce3e626c738e7baef76/crates/http/src/spin.rs#L19-L59

It's there that capture_io_to_memory, prepare_component (that receives redirects), and, subsequently, redirect_to_mem_buffer are called.

Considering our builder is set w/ with_engine(), as per:
https://github.com/fermyon/spin/blob/1fbfb23b4e96a580d6d4f4b6d617e916974493f1/crates/engine/src/lib.rs#L103-L119

I tried emulating the work done in execute in the build stage similar to how the store does it:
https://github.com/fermyon/spin/blob/0f51ad0b86b798ae18cb4f211f0c3fcfbbe9a156/crates/engine/src/lib.rs#L304-L316

... but, nothing came of it as execute would eventually run and override/never access my RuntimeContext WASI context data.

All and all, this PR allows users to set CustomLogPipes when creating their ExecutionContextConfiguration — these CustomLogPipes are later accessed by execute and passed in to capture_io_to_memory (which has a side-effect on prepare_component) providing a good and needed override of Spin's default settings.

[Mossaka]

@lann @itowlson @radu-matei for review~

[danbugs]

Note: I had to do a force push to fix commits that weren't signed-off~ I was under the impression that having a git config --global set w/ user.email and user.name made it so that -s in git commit wasn't necessary 🤷*

Edit: Initially I did a lazy git rebase HEAD~10 --sign-off and that signed-off some commits that weren't mine, fixed that later!

[lann]

@danbugs Thanks for the PR! We're currently working on the 0.2 release so might not get to this until tomorrow or Monday.

[danbugs]

⚠️I think my rebases made this PR a bit messy, so I remade it here at #482 for ease of review~

cc: @Mossaka , @itowlson , and @lann