Vesa Karvonen

As a side note, the removal of the snapshot mechanism wasn't just about optimization. Here is a puzzle about the snapshot mechanism. The code below uses the previous 0.4.1 version...

Note that providing a snapshot operation can be easier for some other lock-free queue approaches. The ["lazy" queue](https://github.com/ocaml-multicore/saturn/pull/145), for example, makes it particularly easy. The approach using [two (essentially immutable)...

Yes, I have those locally and they also seem to be installed on the CI machine. ``` ... ∗ conf-libev 4-12 [required by dream] ... ∗ dream 1.0.0~alpha7 [required by...

I noticed that the dining philosophers test seems to hang on my machine frequently. It sometimes passes quickly, but other times it doesn't seem to terminate at all with CPU...

Thanks for the feedback! > * [line 256: ](https://github.com/ocaml-multicore/picos/blob/7eac8be13bce4bce0841b54964ac3612d3f0f777/lib/picos/picos.mli#L256) the `and` and the end of the line makes it somehow unclear to me. I can't decide if the end of...

`Picos_io` now has a workaround (attempt) for this issue: #269 Let’s close this issue after it has been fixed in the unix library.

Looking at the `Mutex` API, the `protect` argument of [`use_rw`](https://ocaml-multicore.github.io/eio/eio/Eio/Mutex/index.html#val-use_rw) seems unnecessary. ```diff -Mutex.use_rw ~protect:false mutex thunk +Mutex.use_rw mutex thunk ``` ```diff -Mutex.use_rw ~protect:true mutex thunk +Mutex.use_rw mutex (Cancel.protect thunk)...

> https://bugs.launchpad.net/ubuntu/+source/glibc/+bug/1899800/comments/5 says: > > > This bug was fixed in the package glibc - 2.32-0ubuntu5 IIUC, that applies only to glibc on Ubuntu (?) and the version of glibc...

It is an interesting situation. I [read through the thread here](https://sourceware.org/bugzilla/show_bug.cgi?id=25847) and, IIUC, there is a mention that [the "mitigation" patch used in Ubuntu still has issues](https://sourceware.org/bugzilla/show_bug.cgi?id=25847#c45). So, to put...

I was asked to comment about whether there might be some reason to prefer the more complex initial design with two type parameters representing an offset into a block of...