Add .interleaveOrdered to combine sorted streams (i.e. "merge-sort")

[ivan-klass]

An implementation of "merge-sort" merging for two streams of sorted data.
This might be useful in practice when combining streams from different underlying sources.

Discord conversation

[mpilquist]

Thanks! Haven't looked at this PR yet but it reminded me of:

• https://gist.github.com/johnynek/689199b4ac49364e7c94abef996ae59f
• https://gist.github.com/gatorcse/1f92aa7e52a04d9c91511ca79f73e911

[mpilquist]

Approach looks good to me (though haven't done a line by line review yet). I lean towards coming up with a different name, as merge in fs2 implies concurrency whereas this combinator doesn't do concurrent pulls of the two sources (and I agree it shouldn't). Our deterministic combinators that combine two streams include zip and variants, interleave. Any ideas?

[ivan-klass]

Approach looks good to me (though haven't done a line by line review yet). I lean towards coming up with a different name, as merge in fs2 implies concurrency whereas this combinator doesn't do concurrent pulls of the two sources (and I agree it shouldn't). Our deterministic combinators that combine two streams include zip and variants, interleave. Any ideas?

Totally agree. While "merge" is associated with concurrency, "zip" usually means a product over original element type, so I think "interleave" is a good verb, what about "interleaveSorted"?
Also, could you please suggest a better suite file to place a test? Or should it be a new one? Thanks

[mpilquist]

+1 for interleaveSorted or interleaveOrdered

[mpilquist]

Looks good to me besides the name bike-shedding. Thanks!

[ivan-klass]

@mpilquist @diesalbla I've updated method name and found a better place for test

UPD: Also refactored the implementation to much more readable with ".uncons"

[mpilquist]

This combinator must be implemented with stepLeg and not uncons. See #1678 (comment) for details as to why.

[ivan-klass]

This combinator must be implemented with stepLeg and not uncons. See #1678 (comment) for details as to why.

@mpilquist thank you for the explanation, just reverted the last commit