🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | log inspection situation still under development | For MoarVM see #moarvm
Set by lizmat on 22 May 2021.
00:06 reportable6 left 00:30 linkable6 joined 02:15 unicodable6 left, squashable6_ left, statisfiable6_ left, quotable6 left, releasable6 left, committable6 left, greppable6 left, linkable6 left, sourceable6_ left, bloatable6 left, shareable6_ left, bisectable6_ left, coverable6 left, nativecallable6_ left, notable6 left, benchable6 left 02:16 quotable6 joined, nativecallable6 joined, sourceable6 joined, squashable6 joined, statisfiable6 joined 02:17 bloatable6 joined, greppable6 joined 02:18 coverable6 joined, releasable6 joined 02:23 ggoebel left 02:29 evalable6 joined 03:09 reportable6 joined 03:16 notable6 joined, bisectable6 joined 03:17 shareable6 joined 03:34 frost joined 04:17 linkable6 joined, benchable6 joined, unicodable6 joined 04:58 [Coke] left 05:17 committable6 joined 06:07 reportable6 left 06:10 [Coke] joined 06:47 ggoebel joined 06:52 ggoebel left, ggoebel joined
Geth rakudo: fb749d99ec | (Elizabeth Mattijsen)++ | 3 files
Give CompUnit::PrecompilationUnit::File its own file

There is no reason to keep this inside CompUnit::Precompilation::Store::File
nine lizmat: EVALing CURS is an interesting idea! 07:48
lizmat: I see CompUnit::PrecompilationUnit::File as an implementation detail of CompUnit::PrecompilationStore::File. 07:51
lizmat well, perhaps it is now, but maybe not so in the future ? 07:52
nine github.com/rakudo/rakudo/blob/fb74...onUnit.pm6 could use some splitting though.
Why would that change?
The public interface is CompUnit::PrecompilationUnit
Though now that I look at it, I have to admit that it's not all that PrecompilationStore::File specific. It could be used in any store that stores precompiled modules in files. 07:54
lizmat right
08:35 sena_kun joined 10:00 evalable6 left, linkable6 left 10:09 reportable6 joined 10:28 TempIRCLogger left 10:29 TempIRCLogger joined 10:33 TempIRCLogger left, TempIRCLogger joined
MasterDuke nine: is gist.github.com/MasterDuke17/0c5b3...ff5b4fcfcd (that's in `optimize_call()`) the sort of thing you're thinking of? 10:36
`MVM_SPESH_BLOCKING=1 CALL_NO_WANT=1 valgrind --tool=callgrind raku -e 'my Int $r = (^100).pick; multi foo (Int $i) { $i + $r }; my int $a; $a = foo 3 for ^100_000; say now - INIT now; say $a'` reports ~2m instructions fewer 10:38
11:00 evalable6 joined 11:03 linkable6 joined
MasterDuke also making the signature `int` now works with the literal value (and correctly doesn't work if the literal value is > max int value) 11:20
12:08 reportable6 left
patrickb jdv: Compiled tarballs are done. 12:19
12:21 discord-raku-bot left, discord-raku-bot joined
lizmat notable6 weekly 12:27
notable6: weekly
notable6 lizmat, No notes for “weekly”
12:41 Altai-man joined
dogbert11 lizmat: managed to create a Moarvm PR in case you need more material :) 12:47
13:02 [Coke] left
lizmat dogbert11: thanks :-) 13:02
MasterDuke nine: however, i just remembered that i stopped whatever that other thing i was working on was because jnthn pointed out Actions/Optimizer/etc are getting dramatically changed in rakuast. so this might be another thing to punt until after that lands 13:03
13:07 Altai-man left, [Coke] joined 13:12 [Coke]_ joined 13:13 [Coke] left 13:22 frost left 13:26 [Coke]_ is now known as [Coke] 13:56 Xliff_ left 14:11 reportable6 joined
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2022/03/21/2022-...easomatic/ 14:56
[Tux] Rakudo v2022.03-3-gfb749d99e (v6.d) on MoarVM 2022.03-1-gefff4cd31
csv-ip5xs0.804 - 0.852
csv-ip5xs-205.181 - 5.618
csv-parser3.739 - 4.147
csv-test-xs-200.399 - 0.458
test6.520 - 6.540
test-t1.457 - 1.701
test-t --race0.980 - 1.065
test-t-2021.220 - 21.482
test-t-20 --race6.389 - 6.982
I think it is safe to ignore the 1.7. Likely something kicked in in the background, like a cron-job of some sort 16:01
nine Oh, so maybe we should rally on RakuAST 16:49
MasterDuke i haven't been following it very closely. any lhf available? 16:54
nine No idea. Compiling it for the first time just now 17:03
MasterDuke: there's a src/Raku/ast/README.md in the branch with an "If you want to help..." section 17:17
MasterDuke ugh, doesn't cleanly rebase onto master for me 17:19
nine I'd have been surprised if it did considering the intrusiveness of the branch 17:20
MasterDuke i thought i saw that someone else had rebased it a couple days ago 17:21
yeah, on the 14th
oh, at least that what i assume it was 17:22
did you just gen nqp and moarvm? 17:28
oh, once i did things correctly it did rebase onto master 17:31
17:57 nine left 17:58 nine joined 18:00 ilogger2 joined 18:02 lucs_ is now known as lucs 18:05 camelia left, nine left 18:06 nine joined 18:09 reportable6 left 18:14 camelia joined
Geth rakudo: ca09e6fcc3 | (Elizabeth Mattijsen)++ | lib/CompUnit/Repository/Staging.rakumod
Add .deploy method to CUR::Staging

This should allow tests to install modules to be tested into a Staging directory, run tests, and then simply call .deploy to copy the binary files into the right location (rather than needing to re-install and re-precompile again).
lizmat ugexe nine ^^
18:43 linkable6 left 19:03 melezhik joined
melezhik hi raku devs. I am thinking about distributed test systems for selected Raku modules (zef test) against Rakudo daily commits, to catch Rakudo bugs against community modules, difference with Blin - this is distributed system would be run on cluster - in parallel for many modules so we could imagine significantly short time, I am not sure if this is 19:41
a problem with Blin right now, just testing a water here ...
nine melezhik: that more or less already exists: build.opensuse.org/project/show/ho...rakudo-git 19:44
sena_kun melezhik, Blin can split work over all the cores for a single machine, and with nowadays monsters with 12 or 16 cores it takes about an hour to test everything.
in the bright future of tens of thousands of modules that would be great, but maybe not in the short term 19:45
nine Re-reading your message, I dare say the exact thing you're describing exists
MasterDuke `Unknown compilation input 'optimize'`, which is an NQP error, when running `RAKUDO_RAKUAST=1 ./rakudo-m --ll-exception -I lib/ t/01-sanity/99-test-basic.t` 19:47
with not a single mention of t/01-sanity/99-test-basic.t anywhere in the backtrace... 19:48
nine MasterDuke: in !produce-reader? That's what I'm currently working on
MasterDuke yeah. cool, i'll move on
19:49 melezhik left
MasterDuke oh, lots of them are doing that. well, hopefully when you solve that a bunch of stuff will get unlocked 19:50
even t/12-rakuast/literals.t doesn't run? `===SORRY!=== 20:06
Could not find MONKEY-SEE-NO-EVAL in:
nine Haha, that's the one I picked and went with :)
That MONKEY-SEE-NO-EVAL is easy to get around
even the optimize issue is. Getting !produce-reader to actually produce a reader though is harder 20:07
drakonis i find this might be a better place to ask than #moarvm 20:50
given that rakudo does not have TCO, how would it be added to it? 20:51
20:54 melezhik joined 21:01 melezhik left 21:10 reportable6 joined
SmokeMachine melezhik: could it be a k8s cluster with containers/pods running each module’s tests? 21:21
21:47 melezhik joined 22:32 melezhik left
MasterDuke ok, multi-hour distraction over (why is buying a house so hard?)...so optimize and MONKEY-SEE-NO-EVAL can be gotten around? you mean you fixes those problems or they aren't really problems? 22:36
nine Just rather trivial to stub/fix. What I've gotten so far: gist.github.com/niner/a73b112bcee3...0a03c1bc55 22:39
22:46 linkable6 joined
japhb imagines needing to learn secret handshakes before being allowed to buy a house: "Oh come on! This handshake has 537 moves and takes half an hour! Can't I just sign a stack of documents instead?" 22:47
Geth rakudo/master: 5 commits pushed by (Daniel Green)++, MasterDuke17++ 23:31
23:44 melezhik joined 23:56 melezhik left