hey Guys, just realized some weird behavior with S...

# arrow-contributors

[Marc]

hey Guys, just realized some weird behavior with STM and Atomic objects, not sure if I’m doing something wrong but decided to make a quick test and it’s not passing, here is the code 🧵

[Marc]

concurrencytest.kt.cpp

concurrencytest.kt.cpp

[Marc]

I’d expect that both assertion are true (I’m testing this with KMP targeting jvm, btw)
cc @Alejandro Serrano Mena @raulraja @simon.vergauwen

[Marc]

the outcome of the test is as follows:

1) expected:<100000> but was:<542345>
2) expected:<100000> but was:<542345>

(3rd assertion succeeds as expected)

[Marc]

perhaps AtomicInt is not compatible with STM and I need to use TVar instead?

[Alejandro Serrano.Mena]

indeed, `Atomic`s are not tracked by STM, so it's allowed to perform concurrent transactions when using those