Marc Gravell

after some thought, I wonder whether returning something akin to `IDataReader` but backed by the new impl is frankly simpler and more flexible...

Absolutely - and I'm 100% thinking about the API surface from the perspective of what I'd need to do to build a library layer like dapper (or EF) on top...

To confirm, then: using the synchronous API, checking IsPending, and in the pending scenario using asynchronous completion: is equivalent (functionally) to using the asynchronous API in the first instance? (give...

@badrishc I've spiked a switchover to the synchronous API exclusively, with any mention of async *only* in the (rare) `IsPending` scenario; I'll run the numbers when my machine is working

Lots to digest when at PC! Happy to have ignited some discussion. Delete on close: agree useful option, but probably not the default; the other key scenario here is "I...

log options: done `SpanByte` - I have an unsuccessful branch [here](https://github.com/mgravell/FASTERCache/tree/SpanByte); I'm using `SpanByte` keys and values using UTF8 and a buffer (stack or leased, depending on size); write says...

I fixed it by switching to `Memory` instead of `SpanByte`; that's enough for me to make some progress; however... it wasn't the kind of progress I'd hoped for; we now...

I had basically the same - hadn't pushed. I think I've figured out the spanbyte life cycle too, locally - again, not pushed; hopefully tomorrow On Mon, 25 Mar 2024,...

Here we go; definitely has legs! - I switched to `SpanByte` (after figuring out what that means internally) - that gets rid of the *excess* allocations, leaving just the final...

@jodydonetti I've also [thrown it onto NuGet](https://www.nuget.org/packages/FASTERCache/), mostly to stop squatters, but: it is up-to-date