Diego Alonso

@SystemFw So, the benchmarks that I run with the new version may have been different, on the translation of CE2 to CE3. https://github.com/typelevel/fs2/pull/2296. It does not seem as if that...

I have retried the benchmarks today, on my local laptop, with current `main` branch: ``` Benchmark (streamSize) Mode Cnt Score Error Units broadcastTo1 10 thrpt 5 1610.128 ± 981.404 ops/s...

@SystemFw In that case, we may want to correct the current documentation of the `broadcastThrough` method. > The pipes are all run concurrently with each other, but note > that...

If there is no need to constrain the progress, which is to say if there is no need to coordinate the progress of the pipes, then perhaps it would be...

Opened #2621 to, if not correct the behaviour and add the new variant, at least make the documentation a bit less inaccurate. @mn98 would that be more informative?

What would a more specialised implementation consist of? A bounded queue to insert the items that are executed and extract them in the order they come out? Looking at the...

@balthz The `Stream.emits` method generates a stream with a single chunk, that has all the elements of the given sequence. Cancellation can only happen between chunks. So, in your example,...

So, as mentioned above, the problem could be that `merge` throttles each of the input streams to only process one chunk from each side at a time. Note that both...

So, it may be useful for this bug to know the context in which this bug happened. Namely, the JDK version (8, 11, 16, or 17), the Java Virtual Machine...

> Could also use this in the Applicative instance, I think? Added.