[00:07] <[Coke]> https://github.com/raku/raku.org/issues [00:18] *** rir joined [00:23] *** rir left [00:24] *** rir joined [00:24] *** rir left [01:02] *** vrurg_ left [01:02] *** vrurg joined [01:55] *** hulk joined [01:56] *** kylese left [02:15] *** hulk left [02:15] *** kylese joined [02:29] *** nine left [02:29] *** camelia left [02:30] *** nine joined [02:31] *** camelia joined [02:58] *** rir joined [02:58] *** rir left [04:17] *** vrurg_ joined [04:18] *** vrurg left [04:54] *** Sgeo left [06:11] *** kjp left [07:47] *** kjp joined [07:48] *** dakkar joined [07:52] *** tgt joined [07:52] *** tgt left [08:15] landyacht: can provide a minimal example? then we can creare an issue and take a look at it [09:45] landyacht: also, are you running with RAKUDO_RAKUAST=1 or on moar-2026.09? [09:47] PSA: there will not be a Raku Weekly this week [11:27] *** hankache joined [11:38] *** hankache left [11:38] *** hankache joined [11:54] *** dakkar left [11:54] :-( [11:55] *** dakkar joined [12:11] *** guestsa123 joined [12:12] *** guestsa123 left [12:14] *** tgt joined [12:15] *** tgt left [12:42] *** lichtkind joined [12:56] I red an interesting comment about Raku here https://www.reddit.com/r/ProgrammingLanguages/comments/1x7cvdg/agile_programming_languages/ claiming that - in theory - it could be feasable to tweak Raku to have (pure) functions executing on a different computer, perhaps in "the cloud" - seamlessly. "Serverless done right?" is the underlying question. The basis for this claim are the special WHAT and HOW objects in Raku. Just in [12:56] theory: would it be possible for a scheduler to also include "remote compute" when executing Raku programms and not just the local CPU cores? [13:10] *** simcop2387 left [13:13] *** simcop2387 joined [13:24] you probably don't even need scheduler, just a trait [13:25] something like `sub add($a, $b) is remote(remote-http :url("https://my-remote")) { $a + $b }` [13:26] the trait would add a CALL-ME method [13:26] which would then do all the work, I guess [13:26] could be, yes, I have no idea how would the remote receive the code though [13:28] traits can't get RakuAST nodes, can they? [13:29] well, they run at compile time, so technically that should be possible.... [13:29] there's always $*CU [13:29] but then figuring out where you are... could be tricky... [13:29] I guess we would need an API for that [13:45] yes, something that would get RakuAST::Sub from Routine:D [14:13] *** simcop2387 left [14:18] *** simcop2387 joined [14:35] <[Coke]> let's not reinvent CORBA. :) [14:54] Interesting, thank you for your inputs. [15:02] (I keep laughing every time I read $*CU… it seems I’m not mature enough yet…) [15:03] I forgot about CORBA... [15:19] I think the remote feature should be part of a module, but the module itself shouldn't do anything too special [15:21] so it should be easy to send the sub body, maybe some serialized variables [15:22] the execution should be as easy as parsing ast and using serialized scope, arguments etc. for it [15:46] <[Coke]> I had to do a bunch of CORBA work while I was Enron before that imploded, so it's kind of stuck in my brain. :) [16:08] I worked at some point with Enron Sorpus. [16:10] *** stanrifkin joined [16:37] <[Coke]> What is Enron Sorpus? [16:37] <[Coke]> (Corpus?) [16:47] *** dakkar left [17:13] *** stanrifkin left [17:14] *** human-blip left [17:16] *** human-blip joined [19:07] *** zupss is now known as zups [20:58] The texts of Enron emails used for Machine Learning research. [21:04] *** Pixi` joined [21:05] <[Coke]> that's "Corpus", not "Sorpus" [21:05] <[Coke]> Never did check if I was in there, but can't see why I'd be. [21:06] <[Coke]> pronounced like К [21:08] *** Pixi left [21:11] *** Pixi` is now known as Pixi [21:53] *** snonux left [21:55] *** snonux joined [22:09] *** japhb left [22:14] *** japhb joined [22:19] *** Sgeo joined