Mourits de Beer

We've managed to replace the previous module-level lock (single-instance solution - linked in the OP), with a Valkey lock (multi-instance solution) - The "AsyncRedisLock" logic - https://github.com/didx-xyz/acapy/pull/17/files - Our new...

My recommendation for a longer-term ACA-Py solution: - define a generic class, something like "AcapyLock", which can be initialised with some key (e.g. "cred_rev_id_lock") - define `__aenter__` and `__aexit__` to...

It should be possible. Redis/Valkey has the benefit of writing values with expiry times, and returning False if the same key is being set again. That makes a Redis lock...

A note from a colleague: > PostgreSQL has a [locking](https://www.postgresql.org/docs/16/explicit-locking.html) object thingy > Looks like someone made a [SQLAlchemy based package](https://pypi.org/project/sqlalchemy-dlock/) for using it for distributed locking in Postgres. >...

I think first is to just get it working with Askar. Which means Postgres - and the other DB backends (is it just Postgres and SQLite?) What I thought about...

Related PRs: - #3782 - #3784 - #3793

Regex patterns can be computationally expensive ... difficult to read / maintain ... so my proposal would be to simplify as much as possible, and deduplicate the implementation, so the...

The `did:sov` identifier pattern (`[123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz]`) can theoretically be simplified: `[1-9A-HJ-NP-Za-km-z]` - because that uses character ranges, and more clearly omits the characters that should be ignored: `0, O, l, I`....

@dbluhm Maybe I can implement this in `pydid`? i.e. define the regex patterns for sov and generic methods in [validation.py](https://github.com/Indicio-tech/pydid/blob/main/pydid/validation.py), then import and use in ACA validators

I'm thinking there's two parts to this problem: - which did methods are supported in the backend, - and how did method validation is enforced in the request/response schemas. If...