🦋 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.
tbrowder lizmat: i just published new module Date::YearDay with yr code, thnx! maybe put in Raku later. i need it now... 00:13
i see David Adler passed, so sorry, prayers for his friends. his name is very familiar...was he at YAPC Orlando in '16? 00:16
Yes! nice guy 00:19
good lecture...
white camel award! 00:20
Xliff Hello 07:24
Getting this, now:
const_iX NYI
This is following a segfault on a precompile.
bartolin nine, lizmat: Could it be that github.com/rakudo/rakudo/commit/ae8fec579a also shouldn't have been a single (squashed) commit? (I've noticed the discussion about the new disp nativecall branch for MoarVM.) 07:44
Xliff \o 07:48
Here's a question: Are there any differences between JIT for AMD and Intel? I'm noticing stark differences in behavior between my Intel system and my AMD one, particularly after new-dispatch. 07:49
nine bartolin: yes it shouldn't have. lizmat++ reverted it and did a proper merge 07:55
bartolin also for the rakudo merge? 07:56
nine Oh, it happened there as well?
bartolin I think so, yes
nine Ah yes, then we need to fix it there as well 07:57
bartolin oki. hopefully still easy-ish to fix.
but on a related note: I noticed that a 'gmake install' blowed up on the JVM backend. (no registered operand handler for "nqp::dispatch('boot-syscall', ..."). It looks like that could be fixed by moving this line github.com/rakudo/rakudo/blob/5a6d...t.raku#L10 to the moar-specific key additions a bit below. Could that have some negative consequences for MoarVM? 08:00
nine Oh, yes, indeed. Will fix that 08:41
Geth rakudo/master: 14 commits pushed by (Stefan Seifert)++
review: github.com/rakudo/rakudo/compare/5...7fb4ab2ea5
09:00
nine That's the revert + branch merge with an amended last commit fixing the JVM issue 09:01
bartolin thanks, nine++ ## not only for the small fix for the JVM
Geth rakudo/revert-4629-new-disp-nativecall: 0c60d65ca8 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 6 files
Revert "New disp nativecall (#4629)"

This reverts commit ae8fec579aaea59a52fa664311513ab8002bb280.
09:15
rakudo: lizmat++ created pull request #4644:
Revert "New disp nativecall"
rakudo: 1cf360dbf7 | (Elizabeth Mattijsen)++ | 6 files
Revert "New disp nativecall (#4629)"

This reverts commit ae8fec579aaea59a52fa664311513ab8002bb280.
09:16
Geth rakudo/new-disp-nativecall: 13 commits pushed by (Stefan Seifert)++
review: github.com/rakudo/rakudo/compare/6...93c054d5ad
09:17
rakudo: lizmat++ created pull request #4645:
New disp nativecall
rakudo/master: 14 commits pushed by (Stefan Seifert)++, (Elizabeth Mattijsen)++
review: github.com/rakudo/rakudo/compare/1...4a9cb33f6a
09:18
lizmat nine: nativecall newdisp now properly merged 09:19
nine lizmat: I just did that?!
lizmat then I'm not sure what I did :-(
lizmat should have more tea in the morning -( 09:20
nine I actually do not understand that git history anymore 09:25
lizmat or maybe I'm just getting too old for this stuff :-(
nine But my fix for the JVM build issue definitely vanished
lizmat :-(
I was taking my cue from "[08:57:14] <nine>Ah yes, then we need to fix it there as well" 09:26
and missed the "[10:01:05] <nine>That's the revert + branch merge with an amended last commit fixing the JVM issue" :-(
nine I also said that I was going to fix that, pushed a branch and said that this was the revert + proper merge
And git must have told you that origin/master changed in the mean time?! 09:27
lizmat no, it didn't
well, Github didn't
nine Ah, the JVM build fix is still there. I checked the wrong repo 09:28
lizmat *phew*
frost nine by my observation, after these operations, in /tools/build/install-core-dist.raku of the master of rakudo, %provides<NativeCall::Dispatcher> is always set to "lib/NativeCall/Dispatcher.rakumod", no matter on any backend, just like before you fix the jvm issue 12:46
the fix seems to be reverted 12:48
jdv is there any reason releases could be not on a saturday? its the freakin weekend:) 13:04
nine frost: oh, indeed! I only looked for the conditional adding. But the original unconditional one appeared as well 13:29
I wonder if this is that one time where a force push to master might be an appropriate solution. I'd reset it to commit f67fb4ab2ea5f5714275644c0ca54c22cb6167bc
frost nine++
nine I'm not sure if that would make things worse though :) 13:30
lizmat couldn't be worse than the mess I made of it :-( 13:57
nine Well, when you start digging... 14:01
! [remote rejected] master -> master (protected branch hook declined) 14:03
Ok, disabled that protection and force pushed. Doesn't mean that it's over though. If someone who checked out master including the faulty commits pushes to master again, they will reappear 14:05
Everyone: please reset your master branches to the current origin/master. Sorry for the inconvenience! 14:06
nine lizmat: IIRC Ecosystem::Archive had issues? Are they still there? How can I reproduce? 14:07
lizmat the problem with Ecosystem::Archive are basically the same that prevent you from installing Cro::TLS 14:07
Cro::TLS testing hangs 14:08
MVM_JIT_DISABLE=1 makes it work
[Coke] please announce force push on #raku as well, we have a lot of people building from source. (if I read it right that it was rakudo) 14:12
lizmat checks out a fresh clone to be on the safe side 14:13
nine Oh, cro-tls also succeeds with MVM_JIT_EXPR_DISABLE=1. lizmat: is the same true for Ecosystem::Archive?
lizmat yes 14:13
nine Ok, so it's definitely not NativeCall as the expr jit can't cope with that 14:14
I wonder why Cro-TLS didn't appear in the Blin output? 14:15
lizmat well, maybe it's similar to the issue you found with that module that had a missing return signature ?
nine No, then it wouldn't depend on the expr jit 14:16
Geth rakudo: Xliff++ created pull request #4646:
Main default values
15:08
[Tux] Rakudo v2021.10-123-gf67fb4ab2 (v6.d) on MoarVM 2021.10-110-g81082e1c3
csv-ip5xs0.875 - 0.893
csv-ip5xs-205.629 - 5.656
csv-parser4.192 - 4.443
csv-test-xs-200.407 - 0.411
test7.260 - 7.266
test-t1.712 - 1.718
test-t --race0.961 - 0.974
test-t-2024.846 - 25.378
test-t-20 --race7.422 - 7.462
16:11
Geth rakudo: Kaiepi++ created pull request #4648:
Consider Positional/Associative type parameters in optional defaults
23:48