🦋 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.
japhb wonders how much of that is *not* repeatedly boxing and unboxing the first arg to allo 00:00
timo spesh log ought to answer that, right?
00:02 reportable6 left 00:04 reportable6 joined 00:26 squashable6 left 00:38 squashable6 joined 00:46 squashable6 left 00:49 squashable6 joined 01:49 greppable6 left, evalable6 left, quotable6 left, tellable6 left, squashable6 left, bloatable6_ left, releasable6 left, coverable6 left, notable6 left, reportable6 left, linkable6 left, nativecallable6 left, bisectable6_ left, statisfiable6 left, shareable6_ left, unicodable6_ left, committable6_ left, benchable6 left, sourceable6_ left, evalable6 joined, shareable6 joined 01:50 quotable6 joined, nativecallable6 joined 01:51 reportable6 joined, statisfiable6 joined 01:52 committable6 joined 02:49 greppable6 joined, bloatable6 joined, linkable6 joined 02:50 unicodable6 joined, tellable6 joined, bisectable6 joined 02:51 sourceable6 joined, benchable6 joined 02:52 notable6 joined, squashable6 joined 03:49 releasable6 joined 03:51 coverable6 joined 04:17 squashable6 left 04:19 squashable6 joined 04:20 squashable6 left 04:22 squashable6 joined 04:55 squashable6 left 05:57 squashable6 joined 06:02 reportable6 left 07:44 moon-child left, moon-child joined
lizmat PSA: the Rakudo Weekly News will be published on Tuesday this week due to circumstances 08:57
09:03 reportable6 joined 10:09 |Tux| joined, Tux__ joined 10:37 sena_kun joined
|Tux| Rakudo v2021.09-231-g21271aa5f (v6.d) on MoarVM 2021.09-604-gdfb583788
csv-ip5xs1.350 - 1.371
csv-ip5xs-2014.881 - 15.691
csv-parser5.188 - 5.203
csv-test-xs-200.371 - 0.375
test7.393 - 7.413
test-t1.699 - 1.705
test-t --race0.958 - 0.990
test-t-2024.539 - 24.734
test-t-20 --race7.760 - 7.901
11:23 Tux__ left
lizmat < 1.7 as I sorta predicted! :-) 11:24
|Tux| I accidentally ran it on the PC at work, and I was shocked till I realized it was the wrong box 11:58
1.699 is very much ok
(for now)
12:03 reportable6 left 12:06 reportable6 joined 15:31 Kaipi left
ugexe codesections: i dont know how true it is, but a guess about your supply blocks vs Supply question would be that the block form could ensure all the setup is finished for the various whenevers (similar to react) 15:35
16:04 Kaiepi joined 17:17 jdv left 17:18 jdv joined
[Coke] releasable6: status 17:18
releasable6 [Coke], Next release in ≈18 days and ≈1 hour. 4 blockers. Changelog for this release was not started yet
[Coke], Details: gist.github.com/d1ada4cc33fa65931b...cfb0c92b79
17:26 sena_kun left 17:51 linkable6 left 17:53 linkable6 joined 18:01 vrurg_ joined 18:02 reportable6 left, vrurg left
MasterDuke hm, i can't seem to build rakudo with the jvm backend 18:04
it dies while compiling blib/Perl6/Optimizer.jar with `getcodeobj can only be used with a CodeRef` 18:05
ugexe i think its been failing since new disp
MasterDuke i thought the problem was i was using the graalvm jdk, but i switched (i believe) to openjdk 11 and get the same thing
ugexe dev.azure.com/Rakudo/rakudo/_build...nitionId=1
dev.azure.com/Rakudo/rakudo/_build...f&l=38 18:06
MasterDuke oh, somehow i completely missed that 18:07
Geth nqp/new-disp-nativecall: 6790b4528e | (Stefan Seifert)++ | 4 files
API for asking whether the compiler supports a certain nqp op

This can be used to conditionally compile backend specific code in modules like NativeCall
rakudo/new-disp-nativecall: 875ff5139c | (Stefan Seifert)++ | 2 files
Add a hook for custom dispatchers on callables

This lets callable objects have a say in which dispatcher is used when they are invoked. This new mechanism replaces the special handling for wrapped routines in the standard dispatchers.
rakudo/new-disp-nativecall: a1cd39263e | (Stefan Seifert)++ | src/core.c/Compiler.pm6
API for asking whether the compiler supports a certain nqp op

This can be used to conditionally compile backend specific code in modules like NativeCall
rakudo/new-disp-nativecall: db107dd02c | (Stefan Seifert)++ | lib/NativeCall.rakumod
Fix NativeCall performance regression on new-disp

The trick of replacing the native routine's $!do attribute has never worked perfectly as closure cloning could lead to the generic code getting executed. For unknown reason that got a lot worse on new-disp.
Get around that by replacing the fragile $!do replacement with a ... (7 more lines)
MasterDuke got a little farther in the build, but now `Error while compiling op dispatch, no registered operation handler`, which seems like it might not be a quick/easy fix 18:24
nine MasterDuke: so we're missing an #?if moar somewhere? 18:31
MasterDuke i was hoping it would be that simple, but all the nqp::dispatch in rakudo are inside an `#?if moar` 18:32
at least, all the literal ones in the source
same with nqp it looks like 18:33
18:42 vrurg_ is now known as vrurg
MasterDuke or inside something like `if nqp::getcomp('Raku').backend.name eq 'moar'` 18:54
bartolin MasterDuke: I've noticed the broken JVM build too, but wasn't able to invest a lot of time. ooc, what did you do to avoid the original error? 18:58
MasterDuke bartolin: gist.github.com/MasterDuke17/4f9cc...b8d7def040 18:59
bartolin Ah, I see, thanks! 19:00
19:00 patrickb joined
MasterDuke np 19:01
the two args when it gets a dispatch are 'boot-syscall' and 'bind-will-resume-on-failure' 19:03
github.com/rakudo/rakudo/blob/mast...9082-L9099 19:04
bartolin that was introduced with commit github.com/rakudo/rakudo/commit/de24aac1b8. Maybe the old code will still work for the JVM? 19:07
MasterDuke trying now 19:08
Could not instantiate role '<anon|1>': 19:10
CodeRef representation does not support attributes
bartolin :( 19:11
MasterDuke was jvm building right before the new-disp merge? 19:12
19:12 discord-raku-bot left
MasterDuke it's trying to compose an 'is default' mixin 19:13
19:14 discord-raku-bot joined
bartolin my last build was from 2021-09-19 (and that was still fine) 19:14
[Coke] ... I don't think anyone tested JVM on new-disp branch, did we? uhoh 19:27
Is broken JVM a release blocker?
(Ideally it should be, but I think practically it's not) 19:28
19:28 childlikempress joined, moon-child left
bartolin I'd also think it is not. btw, my impression was that most changes were made in a way that they only affected MoarVM. so maybe it's possible to resurrect the jvm backend without too much trouble. 19:30
19:37 Xliff_ left
bartolin MasterDuke: it looks like the initial error "getcodeobj can only be used with a CodeRef" also goes away if one changes back the first adjustment from github.com/Raku/nqp/commit/1ee4398beb -- instead of touching src/vm/jvm/Perl6/Ops.nqp. But there is still the problem with "Could not instantiate role '<anon|1>'" 19:44
MasterDuke hm 19:45
Geth nqp: MasterDuke17++ created pull request #740:
Make JVM tests pass after new-disp merge
MasterDuke oh, we're not going to see anything different in CI because the NQP tests are only run if the rakudo build succeeds 20:00
well, locally `make m-test j-test` now passes for nqp 20:01
bartolin MasterDuke++ 20:04
MasterDuke do you think you'll be able to work on getting rakudo building? 20:05
20:05 reportable6 joined 20:09 monochromec joined 20:10 monochromec left
japhb lizmat: logs.liz.nl/raku-dev/2021-10-04.html#11:23 seems oddly formatted and out of order compared to what I see in irssi in this channel. In particular, |Tux|'s comments from *after* your comment about < 1.7 have been moved to before the test results and split into words across the table. 20:11
bartolin MasterDuke: I'll probably not have much time before the weekend. but I'll try to figure out what goes wrong. (It's quite possible that I'm not able to fix the build, though ...) 20:14
MasterDuke no worries, this was just kind of a yak shaving exercise, so if you can work on it i suspect you'll be quicker to the fix than i 20:19
Geth nqp: 09d7c6917d | (Daniel Green)++ | 2 files
Make JVM tests pass after new-disp merge

The JVM doesn't implement nqp::dispatch, so move any tests with that to the MoarVM tests directory.
The JVM also hasn't implemented hllize via dispatchers, so just skip that test on the JVM.
nqp: 1ecd755b84 | MasterDuke17++ (committed using GitHub Web editor) | 2 files
Merge pull request #740 from MasterDuke17/make_jvm_tests_pass_after_new-disp_merge
20:37 patrickb left 20:51 childlikempress is now known as moon-child
Geth roast: MasterDuke17++ created pull request #758:
Simplify the code used to test destructors
lizmat japhb: thanks for the report! 21:25
will fix it in the next version :-)
japhb :-) 21:40
Geth roast: 7d8f48c0e8 | (Daniel Green)++ | S12-construction/destruction.t
Simplify the code used to test destructors

Sometimes under heavy load we'd hit the timeout instead of a GC triggering the destructor. So instead, just loop until we hit it, but add in a call to explicitly request a GC (which wasn't available when this test was written).
roast: fa5827ac28 | (Vadim Belman)++ (committed using GitHub Web editor) | S12-construction/destruction.t
Merge pull request #758 from MasterDuke17/simplify_destruction_test

Simplify the code used to test destructors
21:53 evalable6 left, linkable6 left 21:55 linkable6 joined 21:56 evalable6 joined 22:12 |Tux| left
japhb lizmat: Created github.com/Raku/problem-solving/issues/302 to start the Blob/CArray/array reconciliation discussion 22:45
23:35 evalable6 left, linkable6 left