It was not immediately obvious to me how to construct a correct invocation of mapReduce. Its signature shows it returning an STM (Async a), so it seemed not too far-fetched to try to wait for it in the same STM transaction that created it:
but that fails with the dreaded "thread blocked indefinitely" error. It turns out (after finding an example in one of the tests) that the correct invocation is rather:
With the waitSTM (via wait) in a separate atomic transaction. Some text showing correct usage may also be helpful to others, perhaps with an explanation motivating the need for the separation (but a simple dictate may suffice).
jwiegley
commented
4 years ago
It was not immediately obvious to me how to construct a correct invocation of
mapReduce. Its signature shows it returning anSTM (Async a), so it seemed not too far-fetched to try to wait for it in the same STM transaction that created it:but that fails with the dreaded "thread blocked indefinitely" error. It turns out (after finding an example in one of the tests) that the correct invocation is rather:
With the
waitSTM(viawait) in a separate atomic transaction. Some text showing correct usage may also be helpful to others, perhaps with an explanation motivating the need for the separation (but a simple dictate may suffice).