02:45 <mange> Is there an easy way to split an array using a predicate, without losing anything? Like (1..10).to_a.split { it % 4 == 0 } but producing [[1, 2, 3], [4, 5, 6, 7], [8, 9, 10]]?

02:45 <mange> I typed the question, but then did some more reading and came up with (1..10).to_a.chunk_while { |before, after| after % 4 != 0 }.to_a, which is fine, but slightly wordier than I'd like.

05:13 <gr33n7007h> mange: slice_before { it % 4 == 0 }

05:15 <mange> Oh ho ho! That's perfect. Thank you very much!

05:16 <gr33n7007h> np

09:16 <tsujp> I know there are caveats with Ractors at the moment (which appear to be vastly improved in upcoming 3.5) but I am really tempted to use them for (relatively) simple parallel execution of shell commands for a big speedup

09:16 <tsujp> threads are nice but thread won't ever go multicore is my understanding, it's only single core concurrency whereas ractors will go multicore (right)

14:21 <havenwood> tsujp: It looks like Matz wants to go straight to Ruby 4.0 this Christmas.

14:22 <havenwood> tsujp: Yup, pretty much. The GVL prevents many but not all parallel cores for threads. The GVL is released for I/O, etc, and can be released by C-ext.

14:49 <leah2> havenwood: he tweeted that on april 1st :>

16:59 <havenwood> leah2: https://bugs.ruby-lang.org/issues/20205#note-75

16:59 <havenwood> leah2: This was July 10, but granted it's not phrased 100% clear.

17:03 <leah2> hmm

17:03 <leah2> what justifies 4.0 tho?

23:24 <havenwood> leah2: Set becoming a C implementation is the top thing that comes to my mind.

23:25 <havenwood> Major changes to Ractor, but it's experimental.

23:26 <havenwood> SortedSet behavior changes in a breaking way, the C API for Set, and such.