AAC: Investigate ChainstateError::InvalidStracksBlocks(String)

[fdefelici]

As reported in this discussion https://github.com/stacks-network/stacks-core/pull/6531/files#r2389233969

append_block wraps ChainstateError variant types as String for the variant ChainstateError::InvalidStracksBlocks(String)

An example: https://github.com/stacks-network/stacks-core/blob/fe868ea6a1bf993793f1e09c3d397b5dde49b93d/stackslib/src/chainstate/nakamoto/mod.rs#L4612-L4618

To be understood if this is the right approach, or if it would be better to expose the native errors.

Potentially this kind of change could be consensus breaking.

[jcnelson]

The inner String is just to be an informative message. I don't think there are any consensus-critical code branches that depend on the message contents.

[jferrant]

The inner String is just to be an informative message. I don't think there are any consensus-critical code branches that depend on the message contents.

I think the bigger concern is that we may have consensus critical paths that match on ChainstateError::InvalidStacksBlock rather than ChainstateError::{Error type that was converted to a string and shoved into InvalidStacksBlock}

I.e. it is converting its own error type into itself but as a different variant and perhaps this is because of a code path that ONLY triggers for that specific converted type rather than the original.

Should probably complete https://github.com/stacks-network/stacks-core/issues/6597 before making any changes here.