vrurg .tell MasterDuke gist.github.com/vrurg/ef54cb5f45f8...be1ba3f8c5 00:22
tellable6 vrurg, I'll pass your message to MasterDuke
00:28 sena_kun left, sena_kun joined 00:36 frost-lab joined 00:42 raydiak joined 01:31 kvw_5 joined
ugexe s/@*ARGS/*@ARGS/ 01:32
01:34 kvw_5_ left
Geth rakudo: vrurg++ created pull request #4280:
Make subtest output a header
roast: vrurg++ created pull request #726:
Improve test for subtest output
02:52 frost-lab left 02:54 frost-lab joined 03:16 sxmx joined 03:36 Kaiepi left, Kaiepi joined 04:03 Kaeipi joined, Kaiepi left 06:38 domidumont joined 06:40 nebuchad` is now known as nebuchadnezzar 08:12 bobv joined
Geth roast/6.c-errata: d3e66a35c6 | (Elizabeth Mattijsen)++ | S06-signature/errors.t
unival(42.2) shouldn't fail if (42.2).unival doesn't

Built-in sub versions should just pass on to their method counterparts.
  (42.2).unival returns NaN, unival(42.2) should as well, therefore this
test is incorrect.
Same as c22b7e9b680f5bf0f68dd on master but not cherry-pickable.
08:43 Geth left 08:44 Geth joined
Geth rakudo: 9f73552aac | (Elizabeth Mattijsen)++ | 2 files
Move a lot of stuff from Str to unicodey

There should be no functional changes.
rakudo: 009d90b83b | (Elizabeth Mattijsen)++ | 3 files
Move unicodey bits from Int / Cool

No functinal changes should have occurred here.
rakudo: 069c0b205d | (Elizabeth Mattijsen)++ | src/core.c/unicodey.pm6
Move all sub uni* sub protos to one place

To make sure that all (conditional) backend code actually sees them.
No functional changes again, except maybe unbreaking the js backend because it referred to a multi sub *before* its proto was defined, which generally is a nono in the setting.
rakudo: Altai-man assigned to codesections Issue recent changes broke non-simple MAIN usages github.com/rakudo/rakudo/issues/4278
9d7279d521 | (Elizabeth Mattijsen)++ | 2 files

No functional changes expected either
10:40 Xliff joined
rakudo: 2251b70c32 | (Elizabeth Mattijsen)++ | 3 files
Move chr(s) subs to unicodey

There should be no functional changes. This most likely concludes the rearranging of unicode related functionality. *phew*
Xliff So I have a question about dispatch ordering. 10:54
m: sub a ($a where 2 | 3 | 4) { say 'Limit!
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in single quotes; couldn't find final "'" (corresponding starter was at line 1)
at <tmp>:1
------> 3sub a ($a where 2 | 3 | 4) { say 'Limit!7⏏5<EOL>
expecting …
Xliff m: sub a ($a where 2 | 3 | 4) { say 'Limit!' }; sub a (Int() $a) { say 'Forced Int!' }; a(3); a(42); a('a') 10:55
camelia 5===SORRY!5=== Error while compiling <tmp>
Redeclaration of routine 'a'. Did you mean to declare a multi-sub?
at <tmp>:1
------> 3; sub a (Int() $a) { say 'Forced Int!' }7⏏5; a(3); a(42); a('a')
expecting any of:
Xliff m: multi sub a ($a where 2 | 3 | 4) { say 'Limit!' }; multi sub a (Int() $a) { say 'Forced Int!' }; a(3); a(42); a('a')
camelia Forced Int!
Forced Int!
Forced Int!
Xliff I would almost say that this is not according to spec.
I'd consider the where multi to be more restrictive than the coercive, so tested first. 10:56
Is this worth a write-up in problem-solving?
10:57 lizmat_ joined 11:00 lizmat left 11:21 lizmat_ is now known as lizmat 11:38 Xliff left 12:24 linkable6 left, linkable6 joined
Geth rakudo: dc74f08173 | (Elizabeth Mattijsen)++ | 6 files
Move all ord/ords/chr/chrs methods to unicodey

Seems I wasn't ready yet. Again, no functional changes, all unicodey stuff should now be in one place, so work on optimizing and making them more consistent should now be able to start.
lizmat afk for a few hours& 13:01
vrurg m: multi sub a (Int $a where 2 | 3 | 4) { say 'Limit!' }; multi sub a (Int() $a) { say 'Forced Int!' }; a(3); a(42); a('a'); # Xliff type goes first in this case 13:23
camelia Limit!
Forced Int!
Forced Int!
13:35 frost-lab left
Geth rakudo: a7c57a01e8 | (Vadim Belman)++ | lib/Test.rakumod
Make subtest output a header

Something I always wanted it to do. Now it behaves similar to Perl's Test::More module and prints `# Subtest: description` before actually starting doing its work.
This commit also introduces `:force-informative` argument for `_diag`.
rakudo: 8fc4cb56c6 | (Vadim Belman)++ (committed using GitHub Web editor) | lib/Test.rakumod
Merge pull request #4280 from vrurg/rakudo_4266

Make subtest output a header
roast: f5c971b1b7 | (Vadim Belman)++ | S24-testing/3-output.t
Improve test for subtest output

Take into account additional header from a subtest output which reports subtest message and looks like:
   # Subtest: message
... (6 more lines)
roast: 0069f6927a | (Vadim Belman)++ (committed using GitHub Web editor) | S24-testing/3-output.t
Merge pull request #726 from vrurg/rakudo_4266

Improve test for subtest output
16:41 domidumont left 17:06 linkable6 left 17:07 linkable6 joined 17:10 sourceable6 joined
Geth rakudo/new-disp: 6d6017b315 | (Jonathan Worthington)++ | src/vm/moar/dispatchers.nqp
Prepare multi dispatch for new-disp resumption

We'll use the resumption mechanism not only for `callsame` and the like, but also as a means to try candidates with `where` clauses and the like with only a single invoke, using a NYI mechanism to be added to MoarVM in order to turn bind failures into a dispatch resumption. This re-works the new-disp multi dispatch handling in preparation for both of these, while trying to keep the cost low for the common case of a multiple dispatch that can be decided on nominal types only and never resumes.
17:35 HarmtH_ left 18:06 HarmtH joined 18:08 linkable6 left 18:09 linkable6 joined 18:36 ggoebel joined
Geth rakudo: 2a972b2408 | (Elizabeth Mattijsen)++ | src/core.c/unicodey.pm6
Make the .ord method a proper multi
18:39 HarmtH left 18:42 HarmtH joined
Geth rakudo: 92dc60b768 | (Elizabeth Mattijsen)++ | src/core.c/unicodey.pm6
Make Str:U.ord die instead of silently returning Nil

The converse Int:U.chr also dies, so it only seems appropriate.
This does not make any spectest fail, but may have some ecosystem fallout. In which case, this commit should be reverted and we should consider making Int:U.chr also return Nil for consistency.
19:18 b2gills left
MasterDuke lizmat: did you catch my question last night? 19:18
tellable6 2021-03-31T00:22:52Z #raku-dev <vrurg> MasterDuke gist.github.com/vrurg/ef54cb5f45f8...be1ba3f8c5
lizmat MasterDuke : guess I didn't, on which channel ? 19:19
MasterDuke here i think
lizmat MasterDuke: I'm pretty sure it propped up in the spectests 19:21
so, if it doesn't anymore, then we're set I guess :-) 19:22
vrurg .ask sena_kun Trying to prepare for regular blin runs. Can't build docker, complains of 'E: The repository 'dl.bintray.com/nxadm/rakudo-pkg-debs buster Release' does not have a Release file.' 19:32
tellable6 vrurg, I'll pass your message to sena_kun
sena_kun vrurg, oh noes, bintray is now gone. :S
I have some replacement, but I don't immediately know where to apply it.
vrurg sena_kun: Ok. I have plenty of tasks anyway. Will run it manually when needed for a while. 19:33
sena_kun vrurg, I found it. :)
So now there is github.com/Raku/Blin/blob/master/d...le#L18-L20 19:34
But we need to use deb dl.cloudsmith.io/public/nxadm-pkgs...deb/ubuntu xenial (or not xenial) and key_url: dl.cloudsmith.io/public/nxadm-pkgs...C6CC6B.key
I'll try to patch it now
vrurg Ok, I'll wait for the patch. 19:35
19:52 bobv left
MasterDuke committable6: 2020.10,2020.11,2020.12,2021.02,2021.03 (^10_000_000).grep(*.is-prime).elems.say; say now - INIT now 19:59
committable6 MasterDuke, ¦2020.10: «664579␤2.38988048␤» ¦2020.11: «664579␤9.1360257␤» ¦2020.12: «664579␤4.35999185␤» ¦2021.02: «664579␤4.4333875␤» ¦2021.03: «664579␤4.368144␤» 20:00
MasterDuke uh, wtf?
lizmat yeah, that's weirtd 20:01
MasterDuke benchable6: 2020.10,2021.03 my $a := (^1_000_000).grep(*.is-prime).elems
benchable6 MasterDuke, starting to benchmark the 2 given commits
MasterDuke, ¦2020.10: «0.3841» ¦2021.03: «0.5610»
MasterDuke, benchmarked the given commits and found a performance difference > 10%, now trying to bisect
MasterDuke, benchmarked the given commits and found a performance difference > 10%, now trying to bisect 20:02
AlexDaniel` hmmm
benchable6 MasterDuke, benchmarked the given commits and found a performance difference > 10%, now trying to bisect
AlexDaniel` bisectable is probably more stable in that sense, just give it a condition like < 3.0
benchable6 MasterDuke, benchmarked the given commits and found a performance difference > 10%, now trying to bisect
MasterDuke yeah, benchable6 needs a little attention 20:03
benchable6 MasterDuke, benchmarked the given commits and found a performance difference > 10%, now trying to bisect
MasterDuke, gist.github.com/24279834ec1c82a6b3...5aee37af08
AlexDaniel` but it's not wrong
I think 20:04
MasterDuke 27fb3ec
linkable6: 27fb3ec
well, there appear to be some commits related to the new coercion stuff in the mix
lizmat ah, the coercion protocol
AlexDaniel` committable6: 27fb3ec^,27fb3ec (^10_000_000).grep(*.is-prime).elems.say; say now - INIT now
committable6 AlexDaniel`, ¦27fb3ec^,27fb3ec: ««timed out after 10 seconds» «exit signal = SIGHUP (1)»» 20:05
AlexDaniel` * committable6: 27fb3ec^,27fb3ec (^1_000_000).grep(*.is-prime).elems.say; say now - INIT now
committable6: 27fb3ec^,27fb3ec (^1_000_000).grep(*.is-prime).elems.say; say now - INIT now
committable6 AlexDaniel`, ¦27fb3ec^: «78498␤2.213962␤» ¦27fb3ec: «78498␤2.2355013␤»
20:05 Geth left
AlexDaniel` MasterDuke: or maybe not? 20:05
MasterDuke it sometimes seems to be off by one or two, but it's usually pretty close 20:06
committable6: 27fb3ec~5,27fb3ec (^1_000_000).grep(*.is-prime).elems.say; say now - INIT now
committable6 MasterDuke, ¦27fb3ec~5: «78498␤0.2546397␤» ¦27fb3ec: «78498␤2.2115329␤»
AlexDaniel` committable6: 27fb3ec~5...27fb3ec (^1_000_000).grep(*.is-prime).elems.say; say now - INIT now 20:07
committable6 AlexDaniel`, Bad start, cannot find a commit for “27fb3ec~5.”
AlexDaniel` rly
MasterDuke committable6: 27fb3ec~5..27fb3ec (^1_000_000).grep(*.is-prime).elems.say; say now - INIT now
committable6 MasterDuke, ¦fa2013d: «78498␤0.2494804␤» ¦b129f31: «78498␤0.2432479␤» ¦2a5dd9e: «78498␤0.2364339␤» ¦f2d7328: «78498␤2.169383␤» ¦2675813: «78498␤2.2807079␤» ¦81f3334: «78498␤2.13249087␤» ¦27fb3ec: «78498␤2.172161␤»
AlexDaniel` nods 20:08
MasterDuke heh, though i've thought for a while that `..` support should be removed, that might be the first time it's ever been used
AlexDaniel` yeah, I did use it sometimes too 20:09
for much wider ranges, like 50 or 100
MasterDuke f2d7328 is "The first draft of new coercion semantics"
AlexDaniel` I think there was one month when there were just a few commits, and I did a range between two releases once
MasterDuke it's not that much code removed to get rid of it, i guess there's no reason to given it is/has actually been used 20:10
sena_kun vrurg, I updated the dockerfile. 20:11
vrurg, github.com/Raku/Blin/pull/30
vrurg sena_kun: I saw the notification. :) Perhaps it's also time to merge ugexe branch? 20:12
sena_kun vrurg, probably, I'll take a look at it soon.
And yes, probably a good idea.
20:13 ggoebel left 20:16 ggoebel joined
MasterDuke AlexDaniel`: to be honest, i'm not quite sure why benchable6 gets off by a couple commits. i thought i'd worked out the logic (more than once!) and it was sound. istr we've even discussed this before and you agreed it was correct 20:17
AlexDaniel` I don't remember, need to look through it again 20:18
MasterDuke of course it might just be minor fluctuations in the timing throwing it off 20:21
oh, Geth is down again? 20:24
lizmat paging the tyil :-)
japhb Did tyil ever figure out what was causing it to fall off? ISTR there was something about it not responding to Freenode pings? 20:25
sena_kun vrurg, I applied the patch suggested for Blin and technically the branch is ready. Can you please do a blin run with the branch once again? If it will look OK, then we can merge it. 20:31
lizmat japhb: yeah, Geth is not seeing freenode's PINGs, so it doesn't PONG, so it's kicked off 20:35
japhb That's kinda odd 20:37
vrurg sena_kun: curious, but to pull ugexe-patch-1 branch I had to do 'pull --rebase' or it resulted in a conflict.
sena_kun vrurg, yes, I did a rebase to include dockerfile fix, sorry. 20:38
vrurg sena_kun: failed with 'no build for end point'
sena_kun vrurg, oops, someone has commited something and the build is not ready yet, I think. 20:39
vrurg, it'll fix it self in a couple of minutes I believe.
vrurg Ah, likely so. I'll try to remember to try it again.
sena_kun: meanwhile, docker build: 'Warning: apt-key output should not be parsed (stdout is not a terminal)' and 'no valid OpenPGP data found' 20:41
sena_kun vrurg, is it just a warning or an error? 20:42
Because I was able to build an image.
vrurg sena_kun: The failing command is: 'The command '/bin/sh -c apt-get update && apt-get install -y $pkgs_first && curl -1sLf 'dl.cloudsmith.io/public/nxadm-pkgs...6CC6B.key' | apt-key add - && curl -1sLf 'dl.cloudsmith.io/public/nxadm-pkgs...me=buster' > /etc/apt/sources.list.d/nxadm-pkgs-rakudo-pkg.list' returned a non-zero code: 2'
sena_kun argh 20:43
I wonder what's different.
vrurg, seems like we have to write it in some other way, but right now I have no time, I suspect you too. :S 20:44
vrurg sena_kun: you're right, I can run commands but no more than that. 20:45
sena_kun vrurg, sorry, cannot do a lot if I can't reproduce. Then some next time.
vrurg sena_kun: but there is no rush. I still have to decide how to post the results. So far adding a separate blog on vrurg.github.io looks like the best option. 20:46
lizmat blog posts are generally a good idea :-)
vrurg lizmat: this one would consist of blin reports only. 20:47
lizmat still.... :-)
vrurg sena_kun: blin runs ok so far. 200+ tested. Will see if it gets all the way down the end. 20:53
lizmat wonders whether it's too early to consider renaming Raku to "Perl 10" 20:55
vrurg lizmat: do really value your life? ;) 20:58
lizmat :-) I do
but I guess it's the time of the year...
sena_kun vrurg, wait, how is it running if you were enable to build an image?
vrurg sena_kun: on the second attempt docker failed much later with '/bin/sh: 1: zef: not found' 20:59
lizmat maybe I should wait a few hours
sena_kun vrurg, are you running it without docker iirc?
vrurg lizmat: I'm not worried, but some people will, for sure...
sena_kun: I do manual runs in a disposable VM. Docker I try on the host machine. 21:00
lizmat: I try to imagine Will's reaction at the joke. And then decide that watching a horror move would be less harmful... 21:01
sena_kun vrurg, oh, it clears the case up then.
lizmat hehe :-)
But Will was one step ahead of my consideration :-) 21:02
21:28 dogbert11 joined 21:31 dogbert17 left 21:32 dogbert17 joined 21:35 dogbert11 left
lizmat just pushed github.com/rakudo/rakudo/commit/18...f2cb4e30ea Make List.sum about 30% faster 22:18
sleep& 22:19
moon-child lizmat++ 22:32
[Coke] . 23:58
tellable6 2021-03-30T16:25:45Z #raku-dev <jmerelo> [Coke] thanks :-)