Prevent Any Storage Errors from Clobbering aZKS Entry

[dillongeorge]

No description provided.

[dillongeorge]

Current

Prior to this patch, AKD would incorrectly assume that any error during initialization of
a Directory should result in an aZKS being created and stored in the underlying database
implementation. The problem here is that this behavior happened on every error, even those which
do not definitively indicate that the read for an aZKS passed but a record was not found.

For example, if a connection issue occurs when connecting to the storage layer, the current implementation
could wrongly overwrite an existing aZKS. In such a scenario, the original read for the aZKS fails
during Directory initialization, but a subsequent write succeeds. This can result in clobbered records
and unexpected behavior.

Proposed/New

With this change, we update the initialization logic for the Directory type to only create an aZKS if
the returned Err indicates that the read was successful, but an aZKS was not found. In all other instances,
we propagate the Err back to the caller instead of swallowing it and wrongly creating an aZKS.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.64%. Comparing base (0094631) to head (a1848e6).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #432      +/-   ##
==========================================
+ Coverage   88.57%   88.64%   +0.07%     
==========================================
  Files          39       39              
  Lines        9077     9109      +32     
==========================================
+ Hits         8040     8075      +35     
+ Misses       1037     1034       -3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.