🦋 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.
lizmat Files=1351, Tests=117115, 293 wallclock secs (35.02 usr 9.81 sys + 4087.29 cusr 333.99 csys = 4466.11 CPU) 10:15
dogbert17 m: multi foo (int) { say "OK!" }; foo 2 # should this work? 12:51
camelia Cannot resolve caller foo(Int:D); none of these signatures matches:
(int)
in block <unit> at <tmp> line 1
Voldenet m: multi foo (Int) { say "OK!" }; foo 2 # should this work? 12:56
camelia OK!
Voldenet I'm betting it should 12:57
dogbert17 I agree :) 12:58
m: sub foo (int) { say "OK!" }; foo 2 # there's also this
camelia OK!
Voldenet m: multi foo (int $i) { $i + $i }; say foo 2
camelia Cannot resolve caller foo(Int:D); none of these signatures matches:
(int $i)
in block <unit> at <tmp> line 1
Voldenet this should definitely work
dogbert17 yes it's odd 13:00
perhaps lizmat knows more 13:02
Voldenet m: say \(2) ~~ :(Int) 13:04
camelia True
Voldenet m: say \(2) ~~ :(int)
camelia Lexical with name '<anon>' does not exist in this frame
in block <unit> at <tmp> line 1
Voldenet hmmmmmm
dogbert17 your latest case seems to be failing around here I think: github.com/rakudo/rakudo/blob/mast...P.nqp#L424 13:12
lizmat that hasn't worked like forever :-( 13:34
multi and natives don't work, jnthn knows the specifics :-) 13:35
nine That specific case (int literal) could be made to work via the optimizer and it could be quite beneficial. IIRC it's just the Want node that for now is in the way. 13:45
MasterDuke nine: you mean collapse all Wants down to just the IVal if there's a compile-time-known value? 14:22
nine Yes 14:30
bartolin just for the record: the jvm backend prints '4' for the 'multi foo' example from above 14:31
nine MasterDuke: I think I did send you my work in progress for optimizations of multi dispatch to native candidates with compile time constants, didn't I?
MasterDuke this topic sounds somewhat familiar, but if you sent me something i've lost it
nine Well, GitHub doesn't forget :) gist.github.com/niner/05f73c3288e4...3288c8d0f6 14:33
MasterDuke oh, i think i've seen that before, but iirc i was doing something slightly different (don't remember what that was either) 14:35
thanks 14:36
nine: btw, don't remember asking before, but what's wip about it? 14:49
nine Actually I don't remember 14:54
MasterDuke m: multi foo (int $i) { $i + $i }; my $a = 2; say foo $a 14:57
camelia ===SORRY!=== Error while compiling <tmp>
Calling foo(Mu) will never work with any of these multi signatures:
(int $i)
at <tmp>:1
------> oo (int $i) { $i + $i }; my $a = 2; say ⏏foo $a
MasterDuke why is ^^^ saying `foo(Mu)`?
nine no idea 14:58
MasterDuke committable6: releases multi foo (int $i) { $i + $i }; my $a = 2; say foo $a
committable6 MasterDuke, gist.github.com/09cadebc97fb51e305...5cadac7821 14:59
MasterDuke bisectable6: old=2019.11 new=2020.01 multi foo (int $i) { $i + $i }; my $a = 2; say foo $a
bisectable6 MasterDuke, Bisecting by output (old=2019.11 new=2020.01) because on both starting points the exit code is 1
MasterDuke, bisect log: gist.github.com/8c6f40017b0f5613cb...2669d835c2 15:00
MasterDuke, (2019-10-14) github.com/rakudo/rakudo/commit/73...0668913bc8
jdv starting the release... 15:35
dogbert11 ++jdv 15:42
AlexDaniel for the record, tellable is probably still broken (will try to fix today), let me know if there are issues with anything else 15:54
does anybody know why colabti stopped working? colabti.org/irclogger/irclogger_logs/raku-dev
a, ok, that is known but feb can't be reached 15:55
I see…
bartolin committable6: fae105b070bb multi foo (int $i) { $i + $i }; my $a = 2; say foo $a 16:36
committable6 bartolin, ¦fae105b: «Cannot resolve caller foo(Int:D); none of these signatures match:␤ (int $i)␤ in block <unit> at /tmp/ayhfcbRR8w line 1␤␤ «exit code = 1»»
bartolin committable6: 6f086e65ee multi foo (int $i) { $i + $i }; my $a = 2; say foo $a ## just to be sure 16:37
committable6 bartolin, ¦6f086e6: «Cannot resolve caller foo(Int:D); none of these signatures match:␤ (int $i)␤ in block <unit> at /tmp/j0limhGhDF line 1␤␤ «exit code = 1»»
bartolin hmm, so the old one was a runtime error 16:38
MasterDuke ah, i didn't notice that. compile time is better... 16:39
bartolin I guess my former self thought that as well. But than, it probably didn't consider the details in the error message :) 16:40
probably there is no way to give something else than Mu at compile time, right? 16:41
MasterDuke well, it does get the type if you give it one
m: multi foo (int $i) { $i + $i }; my Int $a = 2; say foo $a 16:42
camelia ===SORRY!=== Error while compiling <tmp>
Calling foo(Int) will never work with any of these multi signatures:
(int $i)
at <tmp>:1
------> int $i) { $i + $i }; my Int $a = 2; say ⏏foo $a
bartolin yeah, that's good at least
I'll add that to the issue (github.com/rakudo/rakudo/issues/4823) 16:46
Xliff_ p6-Suite Compile Times: 9094.51s (Non-Parallel ) / 8841.291 (Parallel) 17:39
This represents a serious regression in parallel compilation time. 17:40
For reference, last week's time was almost 7 times faster for parallel.
In some cases, parallel times were longer than non-parallel. Might this be a file locking issue? 17:41
lizmat Xliff_: I'm actually looking at that at the moment, but am waiting for the release to finish 17:43
Xliff_ lizmat: Thank you. 17:53
Geth rakudo/release-2022.03: 1920a58258 | (Justin DeVuyst)++ | 3 files
Update changelog + announcement

Deliberately not logged:
  [0b11a143][83b2417a][ce07254a][b5f49ace][a2cf0ea5][b67d38c4][fafcca31]
  [fbe73152][2de6d63d][ad85660c][451ad0ea][0dead3c6][d53b3ee8][68a43f9d]
  [8fb0a66d][1d733cbf][1cd72f1b][1823cdad][5d1056a0][be86adba][66a67dee]
  [54f98fd7][0831a3b5][26e77d98][430d3dd4][f1eccb39][3bde138c][90724381]
  [0e9b2b7c][6e6ef03c][a14b92e2][cd02552c][29fc6507][599e33ef][6e383c20]
  [8d284d8c][46ec1ce7]
18:05
jdv should the t/spec in the release tarball be the 6.d-errata branch? 19:15
lizmat sena_kun ^^ 19:25
afk& 19:29
jdv can someone unblock me from updating the roast errata branches? 20:37
who is a admin or whatever on that repo?
vrurg jdv: as a member of Raku group you must have the full access. 20:47
write access, to be precise
Ah, there seem to be branch restrictions. 20:49
jdv: try now.
jdv same
maybe takes a moment to take? 20:50
vrurg Sorry, forgot to click "save". :) Now?
Geth roast/6.d-errata: 1cabed24d8 | (Daniel Sockwell)++ (committed by Justin DeVuyst) | S12-attributes/instance.t
Fix over-specific regression test (#798)

Two S12-attribute regression tests checked the exact value of an error's $.declaring property (expecting it to be `n attribute`). This changes the matcher to test that $.declaring contains `attribute`, which seems more robust.
jdv thanks
Geth roast/6.d-errata: d2a147d479 | (Vadim Belman)++ (committed by Justin DeVuyst) | 2 files
Accommodate tests to fixes of classification over junctions

The previous version of these tests was a statement of buggy behavior where classify was autothreading over junctions internally and producing incorrect results. This version establishes less confusing and more useful standard of method's behavior.
... (8 more lines)
vrurg Good!
roast/6.c-errata: 561e795bb1 | (Daniel Sockwell)++ (committed by Justin DeVuyst) | S12-attributes/instance.t
Fix over-specific regression test (#798)

Two S12-attribute regression tests checked the exact value of an error's $.declaring property (expecting it to be `n attribute`). This changes the matcher to test that $.declaring contains `attribute`, which seems more robust.
roast/6.c-errata: c36d783518 | (Vadim Belman)++ (committed by Justin DeVuyst) | 2 files
Accommodate tests to fixes of classification over junctions

The previous version of these tests was a statement of buggy behavior where classify was autothreading over junctions internally and producing incorrect results. This version establishes less confusing and more useful standard of method's behavior.
... (8 more lines)
nqp: e777039f10 | (Justin DeVuyst)++ | tools/templates/MOAR_REVISION
[release] Bump MoarVM revision to 2022.03
21:38
nqp: dcb1b6baa1 | (Justin DeVuyst)++ | VERSION
[release] Bump VERSION to 2022.03
rakudo/release-2022.03: ef5c08b99f | (Justin DeVuyst)++ | tools/templates/NQP_REVISION
[release] Bump NQP revision to 2022.03
rakudo/release-2022.03: 2c3bce8bf4 | (Justin DeVuyst)++ | VERSION
[release] Bump VERSION to 2022.03
rakudo: jdv++ created pull request #4824:
Release 2022.03
21:48
rakudo/master: 4 commits pushed by (Justin DeVuyst)++ 21:50
lizmat whee!
jdv++
jdv almost done, maybe 21:51
lizmat builds and tests ok on MacOS jdv++ once again! 22:01
jdv can i get commit on rakudo.org? i thought that had happened already.
lizmat: cool 22:02
lizmat should be able now
invite sent 22:03
jdv I believe my part is done. woohoo. 22:05
patrickb: release
.tell El_Che release 22:06
i guess not then:(
lizmat: thanks! 22:07
lizmat AlexDaniel said he'd look at tellable I guess we need to start looking at hosting that elsewhere potentially 22:09
jdv oh right - i forgot already
i'm sure El_Che will notice eventually
dinner and stuff & 22:10
Geth rakudo: c081f32838 | (Elizabeth Mattijsen)++ | tools/build/install-core-dist.raku
Don't create a lib/.precomp, we don't need it

All we need is to load the source of CURS and EVAL it.
22:12
lizmat afk&
dogbert11 jdv++ 22:55
sena_kun jdv, "should the t/spec in the release tarball be the 6.d-errata branch?" <- I don't think there is a discussed convention right now, so whatever technically happens is "canon". The release should pass all three branches though (main and erratas), so it doesn't matter much in my understanding. 23:06
congrats with the release!
jdv i had just assumed it would be master but then realized its 6d. 23:23
but now that i think about it should it even be included in any form?... 23:24
ugexe What is the purpose of changing the CUR used for install dist? 23:34
Was there any consideration that perhaps it should use the same loading machinery everything else uses so design decisions like e.g. emulates are still relevant? 23:35
im still out of town for a few days, but I hope to see things are being contemplated before implementation
ugexe Anyway that commit is wrong other than removing the one use lib li e 23:39
line
Actually even that line would be wrong then 23:43
lizmat: please request reviews for these things
yes I saw your commit to master but I also felt compelled to comment on it during vacation because it was just pushed right in 23:44
jdv how is your vaca goin? 23:45
jdv needs a vaca
drakonis a vaca? 23:49
oh
vacation