Alexandru Nedelcu comments

Results 173 comments of


                                            Alexandru Nedelcu

[4.x Design] Add blocking method to Scheduler

In fairness `Scheduler` is already a pair of thread-pools — because we need to support scheduling, coupled with thread-pool implementations that don't. `ForkJoinPool` doesn't do scheduling, but we add an...

[4.x Design] Add blocking method to Scheduler

I think that giving the ability to safely trigger blocking operations would make `Scheduler` much more valuable.

Observable.debounce is not consistent with ReactiveX

`debounce` only emits the last emitted element after the source has gone quiet for the specified duration. In the sample it only emits something in case the source hasn't emitted...

Observable.debounce is not consistent with ReactiveX

So the suppressing of all events until the source is quiet is definitely the right behavior. Actually the question is what happens with `onComplete`. Does it emit the last event,...

ConcurrentChannel with bounded buffer can deadlock with short-lived consumers

Yikes 😔

TaskApp will block when throw fatal errors

This is weird, as `TaskApp` does use `IOApp` under the hood. But it is to be investigated. We are short on resources currently, so I've assigned #924 to the next...

Bump cats-effect to 3.0.0-RC2; port catnap+tests to CE3

This is pretty awesome, thanks @oleg-py, will do a review soon. I think we should switch to the 4.x branch, although there might be a need to cherry-pick PRs from...

Bump cats-effect to 3.0.0-RC2; port catnap+tests to CE3

I changed the base to `series/4.x`. Might need a merge with it.

Blocking code is not always properly handled

Thanks for the analysis @ghik, needs to be investigated.

Blocking code is not always properly handled

@ghik thanks. I merged the PR. Will plan for a new 3.x release.