Michael P. Jung

Results 214 comments of Michael P. Jung

I was thinking of maybe an entirely different approach. Maybe it really is time to add the `CachedStatement` struct as described in a comment here: - https://github.com/deadpool-rs/deadpool/pull/234#issuecomment-1402319677 ---- Recap: ```rust...

Sorry for taking so long to answer. I just had a deeper look at the code and the error message about `()` not implementing `ConnectionLike` points to the `query_async` call....

This is related to OpenFlags: - #214 I think one would want to create two pools for this: - A read-only pool - A read-write pool The latter should probably...

As soon as `redis` 1.0 lands I'll try to make a `deadpool-redis` release asap. I just hope the protocol version is removed from the URL for the 1.0 release: -...

Sorry for the slow response. I just came around writing my reasoning why I haven't merged this, yet. I usually try to follow the Tokio minimum Rust version: https://github.com/tokio-rs/tokio?tab=readme-ov-file#supported-rust-versions -...

I've never thought about using `pool.status()` within the `retain` callback. That is indeed unfortunate and will cause a deadlock as `retain` already locks the pool. In order to make `status()`...

> [@bikeshedder](https://github.com/bikeshedder) I think your suggest code is incorrect, because you want status.size to be updated on every loop. Actually my code with pool.status().size is not corrected either, because status...

> I rebased and addressed your comment in crates/deadpool-postgres/src/lib.rs. Testing for race conditions sounds like a good idea. Does the project have anything in place to do that? (...) tbh....

The only place where the statement cache actually interacts with PostgreSQL are the `prepare` and `prepare_typed` methods. I was wondering if we could either replace those methods that take a...