mpyw

@abonander I have an additional question regarding the SQLite refactoring approach. Would it be better to keep the existing `transaction_depth` field within `LockedSqliteHandle` and add a synchronized field to `SqliteConnection`,...

@abonander Strictly speaking, it seems more correct to use `RwLock` rather than `AtomicUsize` and to keep it locked during operations. However, since `RwLock` requires handling errors, I chose this approach...

@abonander If we were to replace the implementation with `parking_lot::RwLock`, it would look like this: ```diff diff --git a/sqlx-sqlite/src/connection/worker.rs b/sqlx-sqlite/src/connection/worker.rs index 6dda814a..d36b1df2 100644 --- a/sqlx-sqlite/src/connection/worker.rs +++ b/sqlx-sqlite/src/connection/worker.rs @@ -34,14 +34,14...

@niklaskorz Please check the issue above. Thank you 🙇

I find it counterintuitive that calling with `Alias::new("schema_name.enum_name")` results in `"schema_name"."enum_name"`. The identifier should be fully determined at the alias definition stage and should not be modified. To address this,...

Looks to be reverted in `0.32.0-rc.2` https://github.com/SeaQL/sea-query/commit/6265311ea0599496fc3954554d89d6f7002afd10#diff-47f0d53e0e29c3ecf9dd0171849266fa1329f552eb74a2819292650d7177d095R47

The trait method signature would be: ```rust /// Returns the current transaction depth. /// /// Transaction depth indicates the level of nested transactions: /// - Level 0: No active transaction....

Trait splitting? ```rust pub trait TransactionDepth { /// Returns the current transaction depth synchronously. /// /// Transaction depth indicates the level of nested transactions: /// - Level 0: No active...

@parkourben99 @parallels999 Any updates? I'm going to abandon https://github.com/mpyw/compoships-eager-limit. I'm looking forward to this PR get merged!

@topclaudy @parallels999 We'll appreciate if you review this in your free time 😃