Florian Angeletti

In fact, the idea of removing `Domain.at_each_spawn` from `Format` is quite compelling too since it is better to not expose a function with a single use case. I am adding...

My memory of the last meeting for `dynlink` is that it was probably better to add a mutex to the dynlinking functions to avoid any kind of races on the...

On the subject of `Dynlink`, the current state on trunk is that only the main domain is allowed to call any `Dynlink` functions. Do we want to switch to a...

Concurrent write to the list of`allowed_units` are already dangerous with threads in OCaml 4, aren't they?

Taking in account #11171, #11193, #11225, #11227 and #11228, I am left with the following items for the stdlib audit: -------------------------------------------- # Changes required: ~~- unified uncaught exception handlers mechanism...

Ephemerons weak hash tables are documented as concurrency-unsafe in #11193. Tearing only happens for float arrays on architectures where the word size and double size is mismatched, isn't it? I...

I had forgotten that the `Lazy` documentation was already documented in term of concurrency safety (https://github.com/ocaml/ocaml/blob/trunk/stdlib/lazy.mli#L43). Do we need more documentation or is the current documentation state sufficient?

Summarizing the current state of the audit: we still have to integrate the documentation before the final release, and there one known issue to fix in `Buffer`. Concerning the internal...

Concerning the examples in `Atomic`, I have the impression that it might work better to drop the first example and expand the second example to explain that the use of...

What should we do with this change? Should we consider dropping it and seeing if users are impacted by the ocamlyacc parser shared mutable state?