Vitali Lovich

Also, how do you want to manage the 1.0 change? Maybe a `KJ_1_0` macro for the in-development branch? While we're at it, perhaps we want to establish a minimum compiler...

Yeah, requiring C++20 for kj-async I think may be a good idea to let us optimize KJ internals with async/await.

@Sarcasm is the `AsyncOutputStream` change important to you?

Gotcha. I don't think we'll move quickly on making kj-async require C++20. My preference would be to see if we can identify hotspots and maybe just have dual implementations until...

I think another 1.0 change might be proper infrastructure for weak references. Currently, there's `kj::atomicAddRefWeak` but the name `kj::atomicTryAddRef` seems like it better describes what's happening because there's no actual...

Another one although this one is perhaps more controversial. `Promise` methods that consume the promise should require an r-value reference to make it more clear that the original variable is...

The Ekam changes are already landed but the user has to do some integration work currently. Basically: ``` EKAM_REMAP_BYPASS_DIR=$(ccache -k cache_dir)/ ``` The trailing slash is important. Arguably maybe Ekam...

My understanding is that this is to try to solve the problem that someone uses a `KJ_DISALLOW_COPY` type & only tests release builds & thus `auto x = NonCopyable()` passes...

At compile time is a bit easier to configure (eg many unit tests so it would require adding some infrastructure to do this globally) but if the preference is that...

Should this control the stack tracing in general or just `backtrace` on Linux? I'm thinking it should disable Windows too, in which case maybe `KJ_USE_STACKTRACE` may be more descriptive of...