Could you write a test that recreates this bug? Race conditions are flaky to test, but looking at example code will help explain the issue (which I think I've understood but I'm not 100% sure).

I've encountered this Issue while looking to solve a problem I'm facing on my own app. My use case is that I have a list of preset values that I need to programmatically fill into several atoms of the same family.

Inside the useRecoilCallback, I'm looping through the array given out by the API and using set to fill in the atom value. The problem I was seeing was that only random atoms would get persisted to Local Storage, sometimes with incomplete values. Applying the change submitted by @bartvanremortele also fixed the issue for me, without unintended side-effects from what I could gather so far.

Could you write a test that recreates this bug? Race conditions are flaky to test, but looking at example code will help explain the issue (which I think I've understood but I'm not 100% sure).

I'm sorry, I don't have time to write a test. This bug can be easily reproduced by updating two atoms in the same useRecoilCallback.

#37 does fix this issue, but two tests fail for the async storage, which is to be expected since with the fix onSet does not take a async function anymore. I didn't go too deep but I don't see an easy solution except maybe adding a asyncStorage boolean flag to the config. 'asyncStorage` could default to true to avoid breaking changes... anyways I'm just brainstorming here, what do you think @polemius ?

I have fixed race condition for synchronize storage (like localStorage) in 2.7.0 version. But the problem still exist for async storage.