miri: Add separate initializer for SOFT_ASSERTIONS

[do-not-use-parker-timmerman]

Motivation

This PR fixes a recognized bug.

• Fixes MaterializeInc/database-issues#5316

Recently we began running Miri in CI, which allows us to catch some tricky bugs. Unfortunately Miri doesn't support global constructors (see rust-lang/miri#450) which we use when initializing the global SOFT_ASSERTIONS flag. This causes any test that uses soft assertions to fail.

This PR adds a new Miri specific initializer for SOFT_ASSERTIONS which is always true. A second approach we could take would be to use once_cell::Lazy instead of ctor to initialize SOFT_ASSERTIONS, but AFAIK that would introduce a small performance regression, as it would result in SOFT_ASSERTIONS getting initialized on first use, as opposed to when the program is loaded.

Checklist

• This PR has adequate test coverage / QA involvement has been duly considered.
• This PR has an associated up-to-date design doc, is a design doc (template), or is sufficiently small to not require a design.
• This PR evolves an existing $T ⇔ Proto$T mapping (possibly in a backwards-incompatible way) and therefore is tagged with a T-proto label.
• If this PR will require changes to cloud orchestration, there is a companion cloud PR to account for those changes that is tagged with the release-blocker label (example).
• This PR includes the following user-facing behavior changes:

[maddyblue]

Neat!