00:27 titsuki left 00:45 sena_kun left 00:59 sena_kun joined 02:45 sena_kun left 02:59 sena_kun joined 04:46 sena_kun left 05:01 sena_kun joined
sjn is getting some build warnings when compiling 2020.01 05:06
src/core/interp.c: In function ‘MVM_interp_run’ 05:07
hm. moarvm warnings... 05:08
AlexDaniel sjn: yep I've seen these too 06:12
I typically don't complain because I'm on the latest gcc which tends to have more warnings for everything, but others won't see them 06:13
stoned75 hi. in rakudo?src/core.c/Bool.pm6 I read "These operators are normally handled as macros in the compiler" 06:25
where can I find these macros' definition ? 06:26
hum I guess I found them 06:40
06:45 sena_kun left 07:01 sena_kun joined 07:06 AlexDaniel left 07:11 AlexDaniel joined 07:12 AlexDaniel left, AlexDaniel joined 07:50 stoned75 left 08:23 domidumont joined 08:45 sena_kun left 09:01 sena_kun joined
sena_kun is it just me or 2020.01 has visibly longer build time? Parse stage is usually 50-52 seconds for me, but last night it was 67, which is not really cool. I wonder if it's my system issue or. 09:12
lizmat Files=1301, Tests=109816, 212 wallclock secs (28.20 usr 8.12 sys + 2948.93 cusr 271.94 csys = 3257.19 CPU) 09:42
.ask [Coke] there is hardly anything in Perl.pod6 anymore? 09:44
tellable6 lizmat, I'll pass your message to [Coke]
lizmat . 09:45
10:08 AlexDaniel left 10:09 squashable6 left 10:10 squashable6 joined
Geth nqp: 7e0ac3da16 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump NQP to get IPNS changes
10:11
rakudo/master: 4 commits pushed by (Ben Davies)++, (Elizabeth Mattijsen)++
lizmat .ask tyil please let me know how I can help to get a Rakudo Star 2020.01 out before FOSDEM :-) 10:13
tellable6 lizmat, I'll pass your message to tyil
tyil lizmat: I'm currently contemplating to make just a rakudo-star 2020.1 source-only release (so no .msi, no .dmg) if I get no response from the official maintainers before FOSDEM 10:14
I can start preparations on it, but I can't make a guarantee it'll be done before FOSDEM hits :( 10:15
lizmat there are no guarantees in life :-)
++tyil for starting preps
Geth rakudo: bae5fc7d49 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get Moar support for IP6NS changes
10:36
lizmat sena_kun Kaiepi ^^ 10:37
10:45 sena_kun left 11:00 sena_kun joined 11:20 Guest78373 joined 11:26 Guest78373 left 11:27 Guest48371 joined 11:34 Guest48371 left
sena_kun we have a looot of PRs 11:37
11:38 dogbert1 joined
Geth rakudo: bd68955b76 | (Ben Davies)++ | src/core.c/Parameter.pm6
Add Parameter.prefix and Parameter.suffix

The prefix method returns the *, **, or + markers a parameter was declared with, if any, and the suffix method returns the ? or ! marker a parameter was declared with, if any.
11:42
rakudo: 3f6a2a2225 | (Ben Davies)++ | src/core.c/Parameter.pm6
Refactor Parameter.raku and improve how it handles parameter names

Parameter.raku was pretty complex and prone to bugs, particularly with how it determined what name the parameter was declared with. Now that other methods can handle most of that work for this method, use them instead of reimplementing their behaviour.
rakudo: c9a6b02de2 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/core.c/Parameter.pm6
Merge pull request #3430 from Kaiepi/parameter-raku

Simplify Parameter.raku, add Parameter.prefix and Parameter.suffix
rakudo: b2c809e4f3 | vendethiel++ | 2 files
Make Parameter's BUILD use named args

Between 3d3578 and d31fb2, the by-name marker has been lost. This commit restores it, allowing a Parameter to be constructed manually.
11:44
rakudo: 749ab904f2 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | t/02-rakudo/99-misc.t
Merge pull request #3330 from vendethiel/manual-parameter-new

Make Parameter's BUILD use named args
12:02 titsuki joined
dogbert1 clickbaits lizmat with github.com/perl6/nqp/pull/596 12:05
tellable6 2020-01-26T18:09:58Z #raku-dev <lizmat> dogbert11 yes, a PR for 6d23891a3a7e github.com/perl6/nqp/issues/559 would be great!
linkable6 (2020-01-03) github.com/perl6/roast/commit/6d23891a3a Some s/perl/raku changes
Geth nqp: 9f3f0a530d | (Jan-Olof Hendig)++ | src/QRegex/Cursor.nqp
Fix for github.com/perl6/nqp/issues/559

During timtoady's cleanup of the Braid API a copy paste error made it through. This was discovered by cognomial++ some time ago. This commit fixes the error where the Braid grammar class attribute should have been initialized with a grammar instance.
12:12
nqp: 75c01d8c70 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/QRegex/Cursor.nqp
Merge pull request #596 from dogbert17/fix-braid-init-copy-pasto

Fix for github.com/perl6/nqp/issues/559
dogbert1 lizmat++ 12:16
Geth rakudo/master: 10 commits pushed by (Elizabeth Mattijsen)++
review: github.com/rakudo/rakudo/compare/7...abd58861ac
12:20
roast: e92bc17a24 | (Elizabeth Mattijsen)++ | S06-signature/introspection.t
Fix test for unnamed named parameter in signature
12:33
12:46 sena_kun left 12:49 AlexDaniel joined, AlexDaniel left, AlexDaniel joined 12:51 patrickb joined
patrickb o/ 12:52
tellable6 2020-01-28T10:09:55Z #raku <El_Che> patrickb I am not territorials about packages. The more the merrier and even better if we can work together!
patrickb El_Che: Great! Then I'll give that a shot. 12:53
tellable6 patrickb, I'll pass your message to El_Che 12:54
patrickb lizmat, tyil: I'll probably get 2020.01.1 precomp packages for mac/win/linux ready before fosdem starts (maybe I'll even manage to automate the process).
lizmat patrickb: for Star ? cool!
patrickb lizmat: No, not star. 12:55
lizmat ah, bummer :-)
patrickb Only rakudo + zef.
tyil patrickb: nice work :D 12:59
13:00 sena_kun joined
AlexDaniel weekly: please mention old-issue-tracker squashathon 13:02
notable6 AlexDaniel, Noted! (weekly)
AlexDaniel weekly: #perl6 irc channel now forwards to #raku
notable6 AlexDaniel, Noted! (weekly)
AlexDaniel by the way
I can't do it for #perl6-dev 13:03
so, whatever
lizmat AlexDaniel : *you* can't, or it can't be done ? 13:05
AlexDaniel I can't
I asked for permissions but didn't get them
anyway, whoever has the founder flag can do it relatively easily 13:06
you just set the flags with mlock and possibly also turn the guard on
that's what I did for #perl6 and it should work fine
lizmat do you know who the founder flag has ? 13:07
AlexDaniel jnthn, TimToady, masak, mst 13:11
I think jnthn tried to give it to me too but turns out the limit is 4 founders in total
I suggested we remove TimToady so that I can set the redirects, but that didn't happen :) 13:14
btw he is currently idling on #perl6-dev but not here
lizmat that's a very striking image :-( 13:16
AlexDaniel jnthn: want me to guide you through the commands you need to run? 13:17
jnthn: /msg ChanServ set #perl6-dev guard on 13:19
jnthn: /msg ChanServ set #perl6-dev mlock +Cinptzf #raku-dev
I think
the important ones are +f and +i 13:20
the rest are just ones that are already there 13:21
and guard should make a bot join the channel so that it won't get fully empty 13:22
+i is what is going to make the redirect work (and you'll be able to invite yourself to that channel if you ever want to join it to see what's going on there)
as I've said, I have no idea what I'm doing so there's possibly a better way, but this is going to work too 13:23
AlexDaniel sleep & 13:26
13:44 lucasb joined
|Tux| Rakudo version 2020.01-34-g14abd5886 - MoarVM version 2020.01.1-2-gcbb04d462
csv-ip5xs0.690 - 0.733
csv-ip5xs-205.673 - 5.814
csv-parser23.149 - 23.623
csv-test-xs-200.362 - 0.367
test7.087 - 7.317
test-t1.803 - 1.823
test-t --race0.780 - 0.795
test-t-2028.891 - 29.618
test-t-20 --race8.869 - 8.877
13:45
[Coke] (Perl6.pod) it's a class definition. With the latest changes made to xt/examples-compilation, I don't have an easy way to get the error on the compile, so I don't know what it's complaining about. 13:53
tellable6 2020-01-28T09:44:55Z #raku-dev <lizmat> [Coke] there is hardly anything in Perl.pod6 anymore?
[Coke] ^^ lizmat 13:54
lizmat will look in a mo
[Coke] adds github.com/Raku/doc/issues/3188 14:05
lizmat: oh. OK, figured out the problem. :) 14:08
lizmat *phew*
[Coke] 1) I have a script that only runs xt/ against files changed since I last updated the git checkout. so it saw the class Perl is Raku ... on my slightly older perl6, this complained (presumably because it didn't know what Raku was) - then I rebuilt perl6 with git latest. re-ran the test. it failed. couldn't see why, if I tried to test it separately, the code worked. Turns out that the nqp::getcomp that you 14:09
were asking about is in the examples compilation test file itself - so after rakudo update it was failing to find the compiler and failing because of that.
updating the testing file... 14:10
lizmat aaaahh good catch!
[Coke] what's the new preferred way to getcomp? 14:11
sena_kun the old one but with `Raku` instead of `perl6` 14:12
?
[Coke] Yes, but there was a conditional to allow it to work on older versions.
my $parser = nqp::getcomp('perl6') // nqp::getcomp('Raku'); ?
lizmat no, please use ||
Brian Duggan found that // fails on older rakudo's :-( 14:13
[Coke] that did it (nqp obj. is too low level for ||)
lizmat for // or || ?
sena_kun sent a bunch of PRs with `//`
lizmat yeah :-(
sena_kun woah
lizmat bisectable6: use nqp; nqp::getcomp("Raku") // nqp::getcomp("perl6") 14:14
bisectable6 lizmat, Bisecting by exit code (old=2015.12 new=14abd58). Old exit code: 1
lizmat, bisect log: gist.github.com/c8c94d2fe9d37831ff...9cd7f7b854
lizmat, (2020-01-24) github.com/rakudo/rakudo/commit/b7...54ef5afd58
lizmat bisectable6: use nqp; nqp::getcomp("Raku") || nqp::getcomp("perl6")
bisectable6 lizmat, On both starting points (old=2015.12 new=14abd58) the exit code is 0 and the output is identical as well
lizmat, Output on both points: «»
lizmat *sigh*
sorry sena_kun, messed up that one 14:15
[Coke] "//" failed for me in a very recent build
sena_kun well, ones which were accepted can be accepted again
those which are not yet can be fixed
[Coke] ... which you may not have noticed if you were looking for Raku first. I happened to have perl6 first in my initial guess. 14:17
sena_kun well, yeah...
so `nqp::getcomp('Raku') || nqp::getcomp('perl6');` is safe? 14:19
[Coke] I think so. if you do perl6 || Raku on a later version, it works.
lizmat yeah, that's what Brian Duggan found: github.com/bduggan/p6-jupyter-kern...-579242473 14:20
dogbert1 m: my $s = Str(123456789) for ^100_000; say now - INIT now
camelia 0.136214
dogbert1 m: my $s = ~123456789 for ^100_000; say now - INIT now
camelia 0.0120425
lizmat wow 14:21
sena_kun updated patches
dogbert1 it's from rindolfs example yesterday
lizmat sena_kun++ 14:22
[Coke] m: my $s = 123456789.Str for ^100_000; say now - INIT now 14:27
camelia 0.03648555
lizmat that last one is even more suspect 14:29
.Str should be the fastest?
as prefix:<~> calls that ?
TreyHarris A I misremembering, or did Pugs-era proto-Raku not have a "backwards" Range, e.g. a 3-element 5..3? 14:35
double negative, sorry... "didn't Pugs have such a Range"? 14:36
[Coke] you might be able to figure that out based on the last roast skips for pugs which were eventually removed. 14:38
TreyHarris [Coke]:good call
Geth rakudo: d8e859d000 | (Elizabeth Mattijsen)++ | 3 files
Make has Str $.a = 42 a compile time failure

  - adapts default BUILDPLAN logic to perform some type tests
  - adds X::TypeCheck::Attribute::Default class
  - adds X::TypeCheck::Attribute::Default stub for setting compilation
14:42
tyil lizmat: gitlab.com/tyil/rakudo-star/pipelines/112963099 perhaps it'll just work :>
lizmat m: .say for 10 ... 1 # TreyHarris
camelia 10
9
8
7
6
5
4
3
2
1
lizmat this is actually statically optimized 14:43
m: .say for ^10 .reverse # as is this, if I recall correctly
camelia 9
8
7
6
5
4
3
2
1
0
14:45 sena_kun left
Geth roast: 218eae0a44 | (Elizabeth Mattijsen)++ | APPENDICES/A02-some-day-maybe/misc.t
Typecheck on default value for attribute changed

with github.com/rakudo/rakudo/commit/d8e859d000
14:46
14:47 |Tux| left
dogbert1 lizmat: ~ seems to be the fastest 14:51
lizmat wonders why
dogbert1 and why is Str() so slow, does it do a lot of extra stuff
lizmat --profile ? 14:52
dogbert1 I did profile yesterday and noticed that when using Str() 10% of the code was interpreted while 0% was interpreted when using ~ 14:53
TreyHarris lizmat: yes, but ... is a Seq constructor, not a Range constructor
dogbert1 here's the code from yesterday: gist.github.com/dogbert17/9e039d6a...b4e0aba220 14:54
lizmat indeed... but that case is actually optimized
14:54 |Tux| joined
lizmat TreyHarris: there is no other syntax for reverse ranges, afaik 14:54
TreyHarris lizmat: understood. If you killfiled todd&margo you may not understand why I'm zooming in on this particular case of Range rather than the pragmatics 14:55
lizmat indeed I have and indeed I haven't 14:56
:-)
TreyHarris Yeah--well, I was trying to explain why "my Int $x; $x.Range" produces something and "my Int $x = 3; $x.Range" fails 14:57
lizmat yeah... people who want to know answers but are unwilling to teach themselves basics, will never understand that
TreyHarris I did notice something that may be a doc error? class Range says for .elems: "Returns Inf if either end point was specified as Inf or *." But it doesn't, it fails for being called on a lazy range. 14:58
And that wording seems wrong, it seems like if it does work on calculable lazy's it should return Inf if the right-hand is Inf, or zero if the left-hand is Inf (but the right-hand is not Inf) 15:00
lizmat that's probably a bug
I feel the doc is correct
bisectable6: dd ^Inf .elems
15:00 sena_kun joined
bisectable6 lizmat, Bisecting by output (old=2015.12 new=d8e859d) because on both starting points the exit code is 0 15:00
jnthn I don't think .elems ever returns Inf these days 15:01
bisectable6 lizmat, bisect log: gist.github.com/731edfbffd86ada516...d9d9faf40f
lizmat, (2017-06-07) github.com/rakudo/rakudo/commit/20...f715bf1503
TreyHarris lizmat:even to my suggestion that (Inf..3).elems should be 0, not Inf?
lizmat ah, I see that it was changed on purpose, so I guess it's a doc issue
TreyHarris Okay 15:02
jnthn Yeah, my recollection is that .elems should always return an Int or a Failure
Inf..3 should almost certianly say 0 elems though
m: say (Inf..3).elems
camelia 0
TreyHarris Arguably the lhs > rhs => 0 check should come first
yeah
jnthn And it does :)
TreyHarris huh
lizmat m: dd Inf ~~ Int
camelia Bool::False
TreyHarris jnthn: ah-hah, i typoe'd. Good. 15:03
jnthn Phew, a bug we don't have. :) 15:04
TreyHarris In that case it's purely a doc issue
tyil on R*: gitlab.com/tyil/rakudo-star/-/jobs/417548433 seems like it builds fine 15:06
(2020.01-rc1)
the spectest takes almost 1 hour, if it doesn't fail too much, I'll be able to publish 2020.01-rc1 on dist.tyil.nl before I sleep 15:07
TreyHarris m: say (*..3).elems
camelia Cannot .elems a lazy list
in block <unit> at <tmp> line 1
Geth roast: 8db6dba6af | (Elizabeth Mattijsen)++ | S12-attributes/smiley.t
Typecheck on default value for attribute changed

with github.com/rakudo/rakudo/commit/d8e859d000
lizmat tyil++ 15:08
TreyHarris m: say(∞..3).elems 15:09
camelia Inf..3
TreyHarris I'm fixing the docs for elems now
m: say(Inf..3).elems
camelia Inf..3
TreyHarris m: say((Inf..3).elems)
camelia 0
TreyHarris m: say((∞..3).elems)
camelia 0
lizmat afk for a few hours& 15:13
TreyHarris Huh, to make my editor happy, I had to blow away and recreate my doc repo from scratch under the Raku org rather than via the perl6 redirect. Now I'm getting permission denied on push via SSH key. my remote push url is origin `ssh://git@github.com/Raku/doc.git (push)` and I also tried it without the `.git`. 15:22
Anyone know what might be going on?
lizmat new public keys ?
TreyHarris No, before I got that error, it made me go through the "re-approve your public key" rigamarole, and I did that
lizmat then I have no idea :-( 15:23
really afk&
[Coke] is there extra debug output you can see from the command line?
TreyHarris Yes: `Pushing to ssh://git@github.com/Raku/doc.git␤ERROR: Permission to Raku/doc.git denied to treyharris.␤fatal: Could not read from remote repository.␤␤Please make sure you have the correct access rights␤and the repository exists.␤` 15:27
hmm, 'read'. let me do a fetch
Nope, fetch worked, push still fails
tyil tfw local installation seems to work for rakudo-star-2020.01-rc1 15:29
spectest at gitlab so far only failed for t/spec/S02-magicals/KERNEL.rakudo.moar, which it did last time as well 15:30
TreyHarris An ssh -T [email@hidden.address] works fine, so it's not my keys
[Coke] TreyHarris: what's your github id? 15:33
TreyHarris [Coke]: treyharris
I thought I'd pushed since the switch to the Raku org, but it's possible I haven't yet 15:34
If access wasn't copied over wholesale, maybe mine wasn't?
[Coke] I don't see you in the explicit list. perhaps you're under the "Raku" team, but I don't have rights to see that.
TreyHarris [Coke]: do you have rights to add me? 15:35
I have rights to close issues in any Raku-org repo, so I figured I was set
[Coke] ah, I can see it. you're not on the team.
TreyHarris weird
[Coke] er. I think? 15:36
TreyHarris Could you add me? Since I was going to move the emacs mode over soon, I'll need full org access like I had on perl6
s/had/have/
[Coke] I'm not on IRC much these days: can someone say "yes that's cool"? 15:38
tyil yes that's cool 15:40
(I'm not sure what it's about tho)
TreyHarris [Coke]: if you look, you'll see I've been a member of the org admins of perl6 forever, so I'm safe ;-) 15:43
github.com/orgs/perl6/teams?query=%40treyharris
[Coke] ok, you seem trustworthy. :)
TreyHarris :-) 15:44
lmk when I can try pushing again
[Coke] ok, you should have an invite.
TreyHarris Okay
[Coke] probably have to click a button before you can push
TreyHarris [Coke]: oh, d'oh... Alex~~Daniel (not mentioning him to trigger spurious notifications) added me 27 days ago--but the button didn't appear until I hit reload just now (following your re-invitation, I assume). Must've been one of those "if you miss it the first time, you have to go a specific place to see it again" 15:47
[Coke] weird. I would have expected to find your name with a "invite sent" next to it. thought I saw that on others.
welcome aboard.
TreyHarris .tell AlexDaniel I'll need full access to the Raku org on GitHub to move the emacs mode repo over and get the package's canonicity validated so Emacs package managers will accept it as a valid redirect. I was a member of Org Admins for perl6 forever so I'm relatively trustworthy ;-) Send me a privmsg if you need to discuss further 15:51
tellable6 TreyHarris, I'll pass your message to AlexDaniel
TreyHarris oops, mentioned him there, didn't I? Hope I don't wake him
You can't privmsg tellable6, looks like. Maybe I'll drop an issue in whateverable for rout-tuit time 15:53
Appreciate a quick double-check of github.com/Raku/doc/commit/2c083530bc for correctness from those of you paying attention to this Range thing 15:58
tyil .tell lizmat gitlab.com/tyil/rakudo-star/-/jobs...8439#L1370 not sure how bad this is 16:03
tellable6 tyil, I'll pass your message to lizmat
TreyHarris tyil: I keep trying to parse 'tyil' as a 'itym'~'til'~'ty' mashup :-) 16:11
tyil just type t, press tab and hope for the best 16:12
:D
TreyHarris tellable6: okay i will
tellable6 TreyHarris, I haven't seen okay around, did you mean oka_?
TreyHarris ;-)
tyil haha
TreyHarris turns out tellable already strips out some things, so it should be possible to: 16:15
.tell ty_il I think you can do this
tellable6 TreyHarris, I'll pass your message to tyil
tyil wah
why does that work
albeit not completelyu
TreyHarris Oh, the code I thought would serve this purpose is only called in the output end, not the noticing-a-nick-is-active end. So no. Underscores are allowed in nicks, though, so I'm not sure why that part of the regex is there: github.com/Raku/whateverable/blob/...ble.p6#L41 16:17
hrm, actually I don't see why it wouldn't have told you, tyil, unless it doesn't tell until you've gone inactive first? 16:18
tyil what does it count as inactive? 16:19
because I get the tell updates usually without having to disconnect or anything
TreyHarris 10 minutes, it looks like: github.com/Raku/whateverable/blob/...ble.p6#L31 16:20
tyil then I guess I'll get them at some point tonight
cycling home takes about 10 minutes, at least
hmm 16:22
===> Install [FAIL] for DateTime::Format: ===SORRY!===
finish_code_object
on rakudo-star-2020.01-rc1, after installing on desktop
AlexDaniel . 16:26
tellable6 hey AlexDaniel, you have a message: gist.github.com/d7f68dad4fdd8f11d1...4246bde046
TreyHarris dalman: Gah, I've now reached the point of cube-solving practice where there are a couple YouTube videos I watch with unsolved cubes in the background of the set, and on one I get twitchy that I can't solve it :-) The other is a prob with an impossible cube state, so that one bothers me in a totally different way... 16:27
oops, fat-fingered paste there
wrong channel
Odd, why did tellable6 turn my message to AlexDaniel into a gist? (And AlexDaniel, if you want to bed < 3 hours ago, I hope I wasn't the one who woke you!) 16:29
AlexDaniel TreyHarris: see the graph on github.com/AlexDaniel/orgsleep
I had to receive groceries :) but it dooesn't really matter anyway 16:30
TreyHarris Ah-hah... do you do the "6 days a week" thing? 16:31
polyphasic sleeping, I meant--which is not 6 days a week, apparently. Been awhile since I looked at it
AlexDaniel TreyHarris: it's unintentional :( 16:32
TreyHarris: en.wikipedia.org/wiki/Non-24-hour_...e_disorder 16:33
TreyHarris understood. that sucks--i had insomnia for months after 9/11 (I lived nearby and worked in the WTC until the week before) and it's the worst
AlexDaniel TreyHarris: tellable6 is a relatively new bot so it's not perfect yet. The message was a bit too long so it gisted it :)
it adds extra text for the timestamp and stuff, so it's unavoidable. But the fixable thing is figuring out when exactly the message is too long 16:34
and honestly I don't know how to calculate that 16:35
last time I looked I didn't really understand
it depends on the… server you're on? Or something like that. It's weird
TreyHarris You could look at github.com/weechat/weechat/blob/d2...age.c#L954 which is how WeeChat decides whether and how to split messages 16:36
oops, I meant github.com/weechat/weechat/blob/d2...age.c#L983 -- I tried to scroll up to enclude that comment and insted scrolled up to the prior func 16:37
and I can't reed or spel very wele aparently... 16:38
[Coke] ooh, cuber. I got a teraminx for christmas. 16:40
AlexDaniel TreyHarris: github.com/Raku/whateverable/issue...2986379714 16:41
16:46 sena_kun left
AlexDaniel TreyHarris: you should have all permissions in the world now :) 16:46
TreyHarris: I'm still trying to figure out the exact list of limitations on github for those who are not Owners 16:48
and yeah, github in that regard is very limited :(
they recently added triage and maintain levels so that's nice, but it's completely different
TreyHarris yeah--GitHub absolutely loves to make world-breaking changesets that, if you opt into any, you opt into all 16:49
AlexDaniel TreyHarris: don't create a new issue, I'll explain :) 16:51
TreyHarris AlexDaniel:oops, too quick 16:53
Sorry
I contributed to the first general-purpose IRC bot back in the early 90's when I was 15 or so (eggdrop), so my mind is crammed with entirely too much IRC minutiae :-) 16:57
17:00 sena_kun joined 17:10 domidumont left 17:30 titsuki left
AlexDaniel .seen [trey-haris]` 17:32
tellable6 AlexDaniel, I saw [trey-haris]` 2020-01-28T16:57:02Z in #raku-dev: <TreyHarris> I contributed to the first general-purpose IRC bot back in the early 90's when I was 15 or so (eggdrop), so my mind is crammed with entirely too much IRC minutiae :-)
AlexDaniel :)
TreyHarris AlexDaniel: yep, I get the permmisiveness, but what will the bot do with people who have underscores in their nicks? 17:39
AlexDaniel TreyHarris: it will just work
TreyHarris it can't distinguish between nickname and nick_name though 17:40
AlexDaniel TreyHarris: that's correct
TreyHarris got it. that's probably an okay balance
especially given that several IRC clients do nick-avoidance by pre/appending underscores
AlexDaniel it uses that sub for lookups, so both nicknamme` and nick_name will turn into “nickname” 17:41
TreyHarris So sticking an extra underscore in already satisfies my other ticket
AlexDaniel the only issue is if you have two people with similar nicknames on the same channel
e.g. [Coke] is using [] because there's another user Coke 17:42
now if they both join this channel… :D
TreyHarris and if someone has a diacritic or other non-ASCII letter character--they'll be addressed as Andr rather than as André. And someone whose entire nick is non-ASCII Unicode letters characters will be just rejected entirely 17:43
AlexDaniel TreyHarris: no, because of this line :) github.com/Raku/whateverable/blob/...ble.p6#L55 17:44
the sub *looks* really stupid, but the behavior is actually pretty neat :)
TreyHarris: also, people are addressed by the nick of their last message 17:46
TreyHarris ah, yeah. if your nick is Ник12 you'll just become 12, though 17:47
(That first letter is not an H)
AlexDaniel “Ник12 is an erroneous nickname”
TreyHarris ник12 17:48
Is it on freenode? Then it doesn't matter.
that unicode nick extensions page says it should be valid if the server allows it
AlexDaniel if it was allowed I'd run that through confusables first :D 17:49
TreyHarris I'd test it but on freenode I will very quickly get my IP banned if I try more than a few
AlexDaniel maybe it's a good idea to pass it through confusables first for things like O0 and l1i 17:51
but I have never seen anybody do that mistake
.seen TreyHarr1s 17:52
tellable6 AlexDaniel, I haven't seen TreyHarr1s around, did you mean TreyHarris?
AlexDaniel and even then you get a helpful response :)
TreyHarris WeeChat actually chooses TreyHarr1s as my first nick avoidance
For reasons I can't fathom without examining the source
AlexDaniel TreyHarris: Because TreyHarris_ is too long, I think
that's why erc turns my nickname into AlexDani` 17:53
TreyHarris Ahh
WeeChat would probably do AlexDan1el then
AlexDaniel wonderful… yeah maybe I should add confusables mapping too 17:54
TreyHarris: btw here's that code working in real life: colabti.org/irclogger/irclogger_lo...01-28#l278
TreyHarris Ugh, looked at my logs and I've become TreyHarr1s but also TreyHarr1--off-by-one error that was corrected at some point I'd imagine 17:58
Yeah, I stopped becoming TreyHarr1s in 2016 and have only become TreyHarr1 since then. 17:59
m: say Str.Range 18:02
camelia No such method 'Range' for invocant of type 'Str'
in block <unit> at <tmp> line 1
TreyHarris Should that work analogously to Int.Range and Rat.Range?
[Coke] Nope. 18:03
AlexDaniel m: say Str.min
camelia Inf
AlexDaniel m: say Str.max
camelia -Inf
AlexDaniel … maybe?
I mean, what are these? :)
sourceable6: Str.min
[Coke] That is inconsistent, aye. I'd rather see max and min removed than range added.
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/d8e8....pm6#L1276
AlexDaniel sourceable6: Str.min() 18:04
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/d8e8....pm6#L1277
AlexDaniel m: say Any.min
camelia Inf
AlexDaniel that's weird, why would that be in Any 18:05
TreyHarris What's Raku for "is $x a single grapheme?" `.chars == 1`?
That's the only use for it I can imagine. Subset existence check
AlexDaniel TreyHarris: yep, .chars == 1 18:06
TreyHarris The only use I can see for Int.Range would seem to be analogous to the use for ["\0"..{$highest_known_codepoint}], and that doesn't seem very useful 18:07
er, strike those square brackets
it's pseudocode anyway 18:08
m: say ("\0"..'z').elems
camelia 123
18:09 patrickb left
AlexDaniel I've had a proposal drafted for changing how string ranges work but I got stuck 18:09
TreyHarris hmm, what's "sparse unicode padding"? i'm running into this term in a few places but I can't find a definition 18:10
AlexDaniel the final issue was having a convenient way for simple things like ‘a’..‘z’ 18:11
TreyHarris m: say so (3 ∈ (0..10)); say so (3 ∈ (0..Inf)); 18:14
camelia True
False
TreyHarris hm?
m: say so (3 ∈ Int.Range); 18:15
camelia False
TreyHarris guess these ranges aren't even useful for the purpose I thought they were useful for
m: say so (3 ∈ (0..10)); say so (3 ∈ (0..*)); 18:16
camelia True
False
TreyHarris m: say so (3 ∈ (0..10)); say so (3 ∈ (*..*));
camelia True
False
TreyHarris The -Inf..Inf (which is what Num.Range returns) would have to be special-cased, but 0..Inf seems like it should just work like any lazy range 18:17
m: say (0..Inf)[12] 18:18
camelia 12
Geth rakudo: a08f043953 | (Christian Bartolomäus)++ | src/core.c/VM.pm6
[JVM] Unbreak the build

The superfluous closing paren was accidentially added with aafa8ca228.
18:19
TreyHarris Range (elem)/∈ isn't documented 18:20
opened an issue about it github.com/Raku/doc/issues/3189 18:25
18:26 Kaiepi joined
TreyHarris Can anyone think of a core routine name that's used for two entirely different things depending on type? (semantically entirely different, not just multis) like, the old example Dog.bark() and Tree.bark() 18:41
there are operators I bet 18:42
18:44 sena_kun left
TreyHarris ah, `e` 18:44
18:59 sena_kun joined
lizmat TreyHarris: if they exist in core, we probably need to fix that 19:01
tellable6 2020-01-28T16:03:39Z #raku-dev <tyil> lizmat gitlab.com/tyil/rakudo-star/-/jobs...8439#L1370 not sure how bad this is
lizmat .tell tyil feels like a wrong version of roast ? 19:02
tellable6 lizmat, I'll pass your message to tyil
TreyHarris lizmat: I don't think so... `e` and `from` being two cases that seem obviously acceptable
lizmat methods ? subs? 19:03
bartolin++ # catching JVM issue with a08f04395345ff9f51b93f 19:05
linkable6 (2020-01-28) github.com/rakudo/rakudo/commit/a08f043953 [JVM] Unbreak the build
TreyHarris lizmat: docs.raku.org/routine/from or docs.raku.org/routine/e 19:07
lizmat I'm not sure the term "e" is a confusable with the .e method 19:08
and I always have thought List.from is a terrible, terrible hack 19:09
TreyHarris That's why I said we don't need to fix it
lizmat from 2014: 1eb7fada84 19:10
hmmm... why doesn't linkable6 pick this up?
linkable6: 1eb7fada84
1eb7fada843b75468e15d7e29b90008d68327ea4 19:11
commit 1eb7fada843b75468e15d7e29b90008d68327ea4
guess I don't understand linkable6
github.com/rakudo/rakudo/commit/1eb7fada84 19:12
Kaiepi hm, when i add tests for unix socket support to S32-io/IO-Socket-INET.t, moarvm throws EMFILE when trying to create the sockets 19:16
which doesn't make a whole lot of sense
maybe the threading in it's buggy
AlexDaniel 1eb7fada843b7 19:21
hmmmmm
lizmat: that's a good question!
I don't know!
foo d8e859d00 19:22
linkable6 (2020-01-28) github.com/rakudo/rakudo/commit/d8e859d000 Make has Str $.a = 42 a compile time failure
MasterDuke lizmat: did you happen to see my questions/comments last night ? 19:23
AlexDaniel lizmat: can you file a bug report please?
lizmat AlexDaniel: will do
MasterDuke: please remind me :-)
AlexDaniel it's a new feature that we didn't have before, but honestly I have absolutely no idea why it doesn't find that 19:24
it should work
MasterDuke lizmat: colabti.org/irclogger/irclogger_lo...01-27#l386 and then some discussion with AlexDaniel following 19:25
AlexDaniel oooh I think I know why… 19:26
lizmat: it, uh… found this result too: github.com/tony-o/deb-rakudodaily/...8d68327ea4 19:28
Kaiepi ah, i just need to do the UNIX socket tests before the IP ones
AlexDaniel and refused to print anything because it doesn't know which one you wanted
tony-o_: what is that repo, btw? 19:29
lizmat: that's easily fixable, I just need to figure out the best way to pick the right result 19:30
amount of github stars maybe?
AlexDaniel sleep & 19:31
lizmat rakudo/rakudo first ?
Geth roast: 7ae5e2be11 | (Ben Davies)++ | S32-io/IO-Socket-INET.t
Add tests for IO::Socket::INET UNIX socket support
19:34
19:35 MasterDuke left
tyil lizmat: idk what that implies, except that something needs fixing 19:35
tellable6 2020-01-28T19:02:24Z #raku-dev <lizmat> tyil feels like a wrong version of roast ?
Geth ¦ problem-solving: treyharris assigned to JJ Issue Routine doc pages need front matter github.com/Raku/problem-solving/issues/155 19:40
TreyHarris ↑ I wrote the above issue to consider the issues I've brought up before about needing a place for overall routine documentation aside from type-specific docs 19:43
Geth rakudo: 05b5f81b67 | (Elizabeth Mattijsen)++ | src/Perl6/Metamodel/BUILDPLAN.nqp
Move typecheck thrower to separate method

so we can use it from multiple places
19:57
19:59 squashable6 left 20:00 squashable6 joined
TreyHarris heh, when I got a reply to my answer to T&M, "That was one of the most illuminating things I have ever read. Thank you for taking the time to write that.", I actually thought it came from T&M initially and was gobsmacked. But, no, the earth is still in its orbit :-) 20:02
lizmat yeah :-) 20:03
TreyHarris++ # perseverance 20:21
20:30 MasterDuke joined
MasterDuke sena_kun: yeah, i'm also seeing a little slower parse times for rakudo recently 20:40
20:44 sena_kun left
TreyHarris The most-recent commit is no longer shown at the bottom of docs.raku.org or docs.p6c.org... is it hiding somewhere? I'm just trying to figure out what version I'm seeing in the index 20:53
[Coke] I believe that was disabled because of churn. 20:56
You could probably follow the build log.
jj would probably know for sure
lizmat bisectable6: subset A of Int; class B { has A:D $.a is default('') = 42 } 21:00
bisectable6 lizmat, Bisecting by exit code (old=2015.12 new=05b5f81). Old exit code: 0
lizmat, bisect log: gist.github.com/8eaabcbc417346f7ef...5a9bdc6104
lizmat, (2017-07-31) github.com/rakudo/rakudo/commit/14...042a171d82
21:00 sena_kun joined
TreyHarris [Coke]: "churn" meaning pages that didn't change changed since the commit was always changing? Could that be rectified by using the commit of the last change that changed a page's content, so if commit a1b2cd3 changed Foo::bar, only /type/Foo and /routine/bar would change? 21:03
(Notionally "could"; I'm aware it's not easy coding.)
lizmat bisectable6: subset A of Str where /^ foo/; class B { has A $.a = "foobar" } 21:04
bisectable6 lizmat, Bisecting by exit code (old=2015.12 new=05b5f81). Old exit code: 0
lizmat, bisect log: gist.github.com/dc0176feeac57de322...33b852e40b 21:05
lizmat, (2020-01-28) github.com/rakudo/rakudo/commit/d8...661dec7b0e
lizmat vrurg: that is what's killing URI atm
[Coke] TreyHarris: yes, I imagine that scheme would be acceptable. 21:06
lizmat ++TreyHarris
TreyHarris doesn't tellable6 have a trick to get it to reference the current moment in archives so you don't have to repeat info in-channel? 21:07
I thought I remembered such a thing
.tell jmerelo please see colabti.org/irclogger/irclogger_lo...01-28#l580 in re commit sha's and doc churn 21:11
tellable6 TreyHarris, I'll pass your message to jmerelo
TreyHarris [Coke]: do you happen to know where is the doc build log? I've been grepping around the repo hoping I'd stumble across a link to it, but no dice 21:17
Geth rakudo: d18c69b2bc | (Elizabeth Mattijsen)++ | src/Perl6/bootstrap.c/BOOTSTRAP.nqp
Simplify exception throwing in the bootstrap

  - use binding result of if to simplify code
  - don't use existskey, but atkey directly with binding
22:24
22:45 squashable6 left, sena_kun left 22:47 squashable6 joined 22:59 MasterDuke left 23:00 sena_kun joined
TreyHarris Oops, I just remembered you don't give contrafactual analogies to T&M. Expect a request for a way to know the colors of the bits in a native value to be incoming... :-/ 23:40