Gabriel Scherer
Gabriel Scherer
> Surely repeating the same I/O or synchronization code cannot be better than just waiting for the other to complete. I am sure of the contrary. I have an example,...
@chambart in a follow-up commit I have revised the implementation to follow your recommendation to never drop exceptions. Exceptions raised by the thunk or by `discard` are always propagated to...
Using `split_from_parent` means that the value will be computed on the main domain on the first `Domain.spawn` call if it has not been called before. This is probably fine, but...
My feeling is that busy-waiting is not a reasonable default behavior for a forcing function, but repeating the computation is. When @chambart or @lthls write that repeating the computation is...
(On the other hand, I do see that there is a marked opposition to my proposal of allowing ourselves to repeat computations -- which is the very point of this...
I discussed this again with @lthls today. We don't completely agree on the design, but we sort of agree that simply exposing a Mutex-blocking version would also probably be fine...
350 PRs later, #14393 proposes a `Lazy.Concurrent_blocking` module where thunks are protected by a mutex during forcing.
@lpw25 this PR is waiting for your response to @garrigue's feedback. Do you think that you will have some time to help get it to the finish line? If not,...
@lpw25 gentle ping; how do you think we could get this PR to the finish line?
The intent of the guidelines is not to discourage the use of AI tools to contribute to the compiler distribution. I tried to write it in a way that is...