Vitali Lovich

As I understand it, typically the Linux kernel policy is that documentation or not, changes that break "correct" observable user-space behavior aren't allowed of course what is "correct" is a...

Should the error handling for the short write be in `Source`, `DmaFile` or somewhere else? I was looking at `uring.rs` and maybe it's in `process_one_event` but the generic `process_one_event` method...

So just so I'm making sure I'm understanding properly. Change the return type so that Source returns `Result` and then `DmaFile` would be responsible for detecting short reads? I think...

I had a mild concern that the message would be different (hence why I initially chose 8 kib - the probability of you seeing it is minimized unless you're really...

I think I've addressed the comments. 1. Extracted out constant to a standalone variable. 2. Mimic the error response to be similar to what it would look like if FL...

Are the pre block/post block annotations needed for epoll too? Or is epoll understood properly and this only applies to condvar and futexes? I have a complicated multiprocess system (processes...

Took another look and it turns out that I didn't actually implement the new `_coz_` C API so they'd always be no-ops. Fixed it up and added comments as best...

Fair point although my experience has been that @mrbbot usually knows what I'm talking about. Specifically passing the ID is talking about doing something like: ``` await mf.getDurableObjectStorage(durableObject.id) ``` where...

Right, but then there's the inverse problem where miniflare's DurableObjectNamespace isn't possible to pass into code that accepts a workers-types DurableObjectNamespace. Hence this error-prone dance.

Noticing this for more than just date. For example, if I grab an object like ``` { foo: { bar: {}, car: [], } } ``` from the DO storage,...