Override config location with environment variable

[ashcrow]

KEYLIME_CONFIG can be set to override the default config file location.

Note:
I think we should eventually rework the configuration system so that we load the configuration once and then reload when we receive the proper signal. My reasoning is that the current set up means that if the file changes while the agent is running then the result could be old and new variables in use based on when they have been read. I've started to make structs for the keylime config. If folks agree I'll slowly but surely chip away at that no matter if this merges or not 😄

[lkatalin]

This is going to be a great improvement; thank you for working on this!

[ashcrow]

Hrm the failure looks like code I didn't touch ...

114 | / pub(crate) fn run_revocation_actions(
115 | |     json: Value,
116 | | ) -> Result<Vec<Result<Output>>> {
117 | |     #[cfg(not(test))]
...   |
152 | |     Ok(outputs)
153 | | }
    | |_^

I'm willing to look at it deeper if required though.

[lkatalin]

Hrm the failure looks like code I didn't touch ...

114 | / pub(crate) fn run_revocation_actions(
115 | |     json: Value,
116 | | ) -> Result<Vec<Result<Output>>> {
117 | |     #[cfg(not(test))]
...   |
152 | |     Ok(outputs)
153 | | }
    | |_^

I'm willing to look at it deeper if required though.

Looks like the failure is related to the revocation scripts (not sure why clippy didn't catch this on the earlier PR). I'm making a fix for this and then you can rebase on that. The changes LGTM!

[lkatalin]

I went down a bit of a rabbit hole, but #162 should help once it is merged.

[ashcrow]

Rebased on #162

[ashcrow]

Woo! All green!

[lkatalin]

Looks great : )

[ashcrow]

Rebased

[ashcrow]

Merging per discussion.