Julius de Bruijn

This example is quite nice for the throughput due to not opening any long-running transactions. The downside is you must write your logic in T-SQL, losing some of the compile-time...

@esheppa has experimental branch for long-running transaction api in [this pull request](https://github.com/prisma/tiberius/pull/129). Please take a look if interested and maybe get the ball rolling again, if you'd think the api...

Apologies from my side too, I've never set up Kerberos anywhere, and the implementation in Tiberius came from @dwink who might be able to help you with this. Also read...

This doesn't sound bad at all. And a fun thing for write! We're all Prisma having a company offsite next week. I have some presentations to do and fun to...

Hey! Could the direct client methods work for you better? - https://docs.rs/tiberius/0.11.0/tiberius/struct.Client.html#method.query - https://docs.rs/tiberius/0.11.0/tiberius/struct.Client.html#method.execute

https://docs.microsoft.com/en-us/sql/t-sql/data-types/int-bigint-smallint-and-tinyint-transact-sql?view=sql-server-ver15 ``` bigint = i64 int = i32 smallint = i16 tinyint = u8 ```

Sure, it should be an easy PR to do. Do you want to try?

Also any Windows users here are more than welcome to help how we could make using Tiberius easier with named pipes or LSP.

Also what makes this a bit questionable is that the server already caches statements, so creating your own statement object to be cached in the user application is probably not...

You cannot really do any IO on `Drop`, due to Rust not having an async `Drop` mechanism. Tiberius is also not bound to any runtime, meaning triggering a `spawn` (which...