»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
psch m: my $x; $x++ = 5 00:00
camelia rakudo-moar : OUTPUT«Cannot modify an immutable Int␤ in block <unit> at /tmp/QvTRPVk_PG:1␤␤»
psch so it would behave like that
AlexDaniel m: my $x; $x**25 = 5
camelia rakudo-moar : OUTPUT«Use of uninitialized value of type Any in numeric context in block <unit> at /tmp/3ncEDgJwHy:1␤Cannot modify an immutable Int␤ in block <unit> at /tmp/3ncEDgJwHy:1␤␤»
AlexDaniel same error though 00:01
00:01 Peter_R left 00:03 n0tjack left 00:04 nowan_ joined 00:06 nowan left 00:20 kjs_ left 00:30 Psyche^ joined 00:34 Psyche^_ left
psch well, almost vOv 00:39
Skarsnik night here
psch .tell hahainternet gist.github.com/peschwa/7202b9f622de517cd881 is what i hacked around a bit with re: superscript. the only things that's missing is proper integration, afaict. as in, the nqp internals complain that the lhs isn't there... 00:40
yoleaux psch: I'll pass your message to hahainternet.
00:40 Skarsnik left
psch .tell hahainternet line 27 is just a debugging aid... 00:41
yoleaux psch: I'll pass your message to hahainternet.
00:45 diakopter___ left
psch is pretty sure there's only a minor oversight hidden somewhere 00:48
unfortunately it's hidden in EXPR and that thing's scary :P
m: say "\x[E0100]".codes 00:49
camelia rakudo-moar : OUTPUT«1␤»
psch j: say "\x[E0100]".codes
camelia rakudo-jvm 0b6328: OUTPUT«2␤»
psch r: sub foo (@a) { say @a.WHAT }; my $a = Seq.new((1..2).iterator); foo($a) 00:51
camelia rakudo-jvm 0b6328: OUTPUT«Type check failed in binding @a; expected Positional but got Seq␤ in sub foo at /tmp/tmpfile:1␤ in block <unit> at /tmp/tmpfile:1␤␤»
..rakudo-moar: OUTPUT«(List)␤»
psch r: use nqp; sub foo (@a) { say @a.WHAT }; my $a = Seq.new((1..2).iterator); foo(nqp::decont($a))
camelia rakudo-moar, rakudo-jvm 0b6328: OUTPUT«(List)␤»
00:54 grondilu left 01:05 geraud joined 01:10 znpy left 01:20 zengargoyle left, tokuhiro_ joined 01:25 tokuhiro_ left, kid51 joined 01:32 aborazmeh joined, aborazmeh left, aborazmeh joined 01:36 n0tjack joined
timotimo huh, newlines on osx aren't \r? 01:36
konobi nope 01:37
good ol' unix 01:38
01:40 n0tjack left
arnsholt Yeah, \r newlines are MacOS Classic 01:49
That's possibly the best part about OS X: It's pretty much all Unix underneath
But BSD rather than GNU userland, which can trip you up every so often if you're not used to it 01:50
01:50 kid51_ joined 01:52 kid51 left, kid51_ is now known as kid51
Hotkeys if only windows had jumped on the unix train 01:52
arnsholt MS actually sold a Unix for a long time. Think it's dead now, though 01:53
psch solaris was the one with the surprising killall, wasn't it? 01:54
arnsholt But the Windows kernel is based on a completely different approach from Unix. Apparently the NT kernel is super-solid though
Most of the problems with Windows are due to crap third-party software or crap hardware, I think
pink_mist psch: what did the solaris killall do? 0_o 01:55
psch pink_mist: i'm not sure it was solaris, but i remember a story of some fresh sysadmin being used to gnu/linux killall and finding himself on a server where killall literally meant kill all 01:56
pink_mist ohlord
wonder who thought that was a good idea :P 01:57
psch tehtable.wordpress.com/2009/12/07/...ris-linux/ fits my memory :)
konobi arnsholt: it's an _actual_ unix... it was even certified at one point
psch pink_mist: aparently i was right in remembering it as solaris, cf. the link 01:58
konobi psch: iirc, it was solaris
01:58 skids joined
konobi but i think it was pre-10.0 01:58
01:59 rurban left
arnsholt konobi: OS X, you mean? Yeah, that's exactly it. Only atypical thing is pretty much the GUI stuff not being X11 01:59
konobi arnsholt: yup. 02:00
psch konobi: i've never actually spent time with solaris, i just remembered the story...
konobi psch: kicked me in the ass a couple of times =0) 02:01
TimToady pink_mist: the $ in $ = stuff is not a declaration, it's more like (state $) = stuff, so the = is just ordinary assignment
pink_mist TimToady: ahhh, I see 02:02
mst pink_mist: solaris killall is a famous trap
konobi now that I think about it... SDC would be a good platform to do infrastructure support for builds/testing/dev
mst: o/
mst konobi: \o 02:03
02:04 BenGoldberg joined
konobi mst: long time no chats 02:04
02:06 johan joined
mst yar, last time was the nix related discussion on #catalyst IIRC 02:06
mst is still iterating slowly towards having a working version of the thing in question 02:07
02:07 leont left
mst (which is eventually going to force me to work with Inline::Perl6, which I'm rather looking forward to ;) 02:07
02:08 zengargoyle joined 02:09 colomon left
konobi mst: oh? 02:11
02:12 colomon joined
mst konobi: my plan for handling config files will require grammars 02:14
konobi ah ha... as i suspected. 02:15
then you could use another grammar to use nix config files and just steal the project... *mwuahahaha*
mst yes. yes, I could. that's ... actually a rather interesting idea. 02:16
mst tries to forget it to avoid being distracted from the core ;)
02:17 lab_ joined 02:18 xenu_ joined, Juerd_ joined, zoosha_ joined, BinGOs_ joined, ab5tract joined, f3ew_ joined, bhm_ joined, btyler_ joined, leedo_ joined, metasin_ joined, maddingu1 joined, jojotus_ joined, Upasaka_ joined, yubimusubi joined, [Coke]_ joined, sftp left, tony-o joined 02:19 renormalist left, ab5tract_ left, bhm left, zoosha left, [Coke] left, Possum left, f3ew left, BinGOs left, Juerd left, Juerd_ is now known as Juerd, jnthn_ joined, sftp joined
arnsholt mst: Why Inline::Perl6 for grammars, instead of something like Regexp::Grammars, if I may ask? 02:19
02:20 MadcapJake left, MadcapJake joined
dalek c: 58e9172 | skids++ | doc/Type/X/AdHoc.pod:
Add some cultural guidance
 02:21
02:21 grondilu joined
konobi mst: indeed... core project first! 02:22
02:22 Zoffix joined
mst arnsholt: because, generally, when something is that important to one of my plans, I end up bending it -very- hard 02:22
konobi and the ability to just use it in perl6
02:23 Upasaka left, lab left
mst the only thing that seems like an advantage of Regexp::Grammars is "not needing to manage a rakudo install" 02:23
and the codebase in question is going to have to do that anyway
because one of the things I'm going to build is a carton-like thing designed to manage hybrid p5/p6 projects
psch skids: i'd rephrase that to "in some cases using the string returned might be preferable, but tests usually want to check against Exception if no subclass is known." or somesuch
mst at which point I may's well use the real thing 02:24
plus, every time I discuss my full plans, people give me confused looks, with the exception of jnthn and pmichaud who said "yeah, that's not going to be trivial, but I've done something like it before and I totally think it should be possible"
02:24 telex left
psch still, skids++ for culturally applicable documentation 02:25
02:26 telex joined
konobi i think i was pretty thumbs up about it 02:26
arnsholt mst: That sounds pretty cool, TBH. I think you're right that hybrid 5/6 applications are going to be a thing, and a tool to manage those is definitely going to be awesome 02:27
skids Well, you'd probably want to ensure you have an Exception before trying to call .message.
arnsholt And with a bit of luck it might help cross-pollinate the five and six worlds a bit
mst that's pretty much what I have in mind
plus the fact that people working on such apps are going to be a lot more willing to try something new
seems like a positive sum game to me
arnsholt Indeed
02:28 n0tjack joined
konobi mst: now you just gotta start a repo! 02:28
dalek href="https://modules.perl6.org:">modules.perl6.org: 4377db9 | (Zoffix Znet)++ | web/templates/layouts/default.html.ep:
Use title() helper in <title> element
psch skids: oh, right. i didn't check the surroundings of the patch, if it's clear that this specifically about checking for Exceptions and their message it's probably fine
mst konobi: core algorithm first. then the perl5 client. then the hybrid client. then grammar support. then C dependencies. then I take on puppet/chef/etc. 02:29
arnsholt And I agree that the Perl 6 grammars are pretty awesome. Even parsing Python turned out to be pretty straightforward (once I stopped being completely obtuse)
konobi arnsholt: think... a parser for ansible that isn't just a plain ol' YAML loader 02:30
mst augeas 02:31
dalek href="https://modules.perl6.org:">modules.perl6.org: ae688d0 | (Zoffix Znet)++ | web/templates/root/index.html.ep:
Add explicit dimensions to camelia's image

Google cries we load non-critical CSS/JS and based on screenshots it is actually due to camelia changing layout on load and not actual CSS/JS
mst great idea
the lens file format is horrific
grammars will be way way way way nicer 02:32
02:32 n0tjack left
konobi mst: also the ability to extend the grammars for "modules" or what not 02:33
mst however. burn that bridge when I come to it.
02:36 yqt left
konobi got the core algo into notes yet? 02:36
02:36 edehont joined 02:37 edehont left
mst I'm implementing as I go. 02:38
each prototype then gets beaten up until I manage to break it utterly, and then I start a new git repository and try again
mst is optimising for having a working implementation, I'll backfill the algo into the docs 02:40
but I can't motivate myself to do that until I know it actually works
konobi ah... probably no re-useable tests then 02:41
02:42 xenu_ is now known as xenu 02:43 BenGoldberg left, BenGoldberg joined
dalek href="https://modules.perl6.org:">modules.perl6.org: 0c3b817 | (Zoffix Znet)++ | web/ (3 files):
Show Perl 6-ish error pages instead of Mojolicious's default (Closes #30)
 02:43
02:46 edehont joined 02:47 modcure left
mst not any time soon, I'm afraid. I am aware that making it possible to re-implement it is a worthwhile goal, but right now everything is secondary to either proving the current algo works or breaking it and figuring out the next attempt 02:49
02:50 kid51_ joined
mst konobi: being right at the edge of my capabilities is both fun, and somewhat restricting :) 02:52
02:52 kid51 left, kid51_ is now known as kid51
dalek href="https://modules.perl6.org:">modules.perl6.org: 0fb8a2f | (Zoffix Znet)++ | / (3 files):
Document and improve handling of META.list.local
 03:00
03:02 johan left 03:06 kid51 left 03:07 edehont left
konobi mst: =0P 03:08
03:08 majid joined
konobi mst: yeah... just thinking that someone might know of some smart math that can be applied 03:08
dalek href="https://modules.perl6.org:">modules.perl6.org: 1d97568 | (Zoffix Znet)++ | web/lib/ModulesPerl6/Model/ (2 files):
Fix Unicode handling in db (Closes #34)
03:08 oceanpollen left
psch m: IO::Notification.watch-path(".") 03:11
camelia ( no output )
psch m: IO::Notification.watch-path(".").tap( -> $f { say $f }); sleep 10
camelia ( no output )
psch well, it compiles at least...
03:11 espadrine left
mst konobi: I want the initial implementation to be only as smart as strictly necessary, anyway, in the names of being able to explain it 03:11
psch i guess that means dying with an NYI is preferable to what r-j currently does... 03:12
j: IO::Notification.watch-path(".").tap( -> $f { say $f });
camelia ( no output )
03:12 aindilis joined
psch ...that doesn't match my local output o.o 03:12
03:12 majid left
zengargoyle
.oO( should user's module install dir be XGD_DATA_HOME vs ~/.perl6 )
 03:12
psch zengargoyle: XDG, no? 03:13
zengargoyle gah, XDG_DATA_HOME/perl6 that is :)
psch m: say %*ENV<XDG_DATA_HOME> 03:14
camelia rakudo-moar : OUTPUT«(Any)␤»
psch camelia doesn't know about any of that vOv
zengargoyle defaults to $HOME/.local/share
psch or well, camelia only knows Any about that, maybe...
zengargoyle i initially didn't like the XDG stuff, but have grown acustomed to .config .cache .local/share with non-hidden directories per app. 03:15
psch zengargoyle: so you're saying something like < %*ENV<XDG_DATA_HOME> // "$HOME/.perl6".IO > for module installation path? 03:16
zengargoyle: 'cause clearly XDG_DATA_HOME isn't set everywhere, camelia and windows as two examples
dalek href="https://modules.perl6.org:">modules.perl6.org: f41ff22 | (Zoffix Znet)++ | web/ (2 files):
Improve log messages for db generation/app restart (Closes #28)
zengargoyle more like < (%*ENV<XDG_DATA_HOME> // "$HOME/.local/share") ~ '/perl6' 03:17
no more new .foo files/dirs in $HOME 03:18
psch that leaves windows
zengargoyle standards.freedesktop.org/basedir-s...c-0.6.html
psch there we probably want %appdata& or however that is called
s/\&/%/
zengargoyle yeah, i totally have never used windows. 03:19
.oO( slight lie, never have used windows of my own accord )
 03:20
03:22 tokuhiro_ joined
dalek p: b5008ad | peschwa++ | src/vm/jvm/QAST/Compiler.nqp:
Correct types for op nqp::watchfile.

Previously this would complain about not being able to unbox a str to int. Now we at least get the NYI error.		 03:23
03:23 noganex_ joined
psch zengargoyle: fwiw, i'd prefer something that works with a simple substitution 03:23
zengargoyle: as in "if you're on something *nixy it's in $HOME/$path, on something doze-y it's in %appdata%\$path" 03:24
zengargoyle: i recognize that following XDG here might mean its adoption increases, but i'm not sure it's our responsibility to push it
03:25 noganex left
zengargoyle well it's a nice way to let people store it where they want without resorting to making symlinks or something. the ability to override that environment variable. 03:26
03:27 tokuhiro_ left
zengargoyle most more-popular linux desktops use the XDG stuff (probably falling back to the defaults for some). that's where all the Documents and Videos and Music type folders come from. 03:28
psch yeah, that's what came to mind, and i agree with it
if we want to allow overriding of where to store installed stuff going with the (even if not 100% adopted) standard is the best choice 03:29
and if we don't want to allow it we might as well bundle local::lib in CORE...
well, an equivalent anyway 03:30
zengargoyle like me, running bare X+openbox, i only have some that X or systemd sets up... XDG_RUNTIME_DIR=/run/user/1001 and SESSION_ID / SESSION_COOKIE 03:31
mst adding XDG support to File::HomeDir was a huge mistake that has made it a complete disaster to deal with
zengargoyle care to elaborate? 03:32
mst I would strongly recommend an old-school "use $HOME" approach by default
it's un-unixy and oriented around shiny desktops
so it completely fails POLS for anybody who expects unix
zengargoyle true, but dotfiles were an accident. 03:33
mst yes, and now they're a well understood standardised accident that people actually like
following XDG by default is largely a recipe for "where the fsck did you put my files?!" moments for your users
psch actually remembers that from trying to customize awesomewm
mst XDG support is absolutely a nice thing to have. but let the people who want it turn it on explicitly. 03:34
psch i did resort to < find . -iname 'rc.lua' >
mst *snrk*
psch: which is precisely what I don't want to do to perl6 users :)
zengargoyle but once you know, i find it easier/nicer to just ls .config and see stuff. 03:35
each app with it's own nice directory
mst sure. and anybody who already knows, will turn on the option
I know it's nicer for you. it's just the failure modes are horrible for people who aren't you.
psch FSVO "horrible" 03:36
it is inconvenient if you don't expect it, yes
mst confusing the heck out of a new user is more than inconvenient
zengargoyle maybe true, but still everything else is moving that way already.
so it's not like we'd be protecting them from anything. 03:37
mst and GUI distros are welcome to package it with the option turned on by default
but dotdirs are still fundamentally the standard for terminal programs designed to be used in a unixish environment 03:38
the only thing in my .config/ on my main development system is Qt related 03:39
so I think your 'everything else' assertion is optimistic :)
zengargoyle Perl 6 is for 7 year old girls! :P
think of the future.. :)
where i have 65 dirs in .config
mst in the future, the distros will probably have invented yet another standard and moved to that 03:40
zengargoyle and that's 65 things not cluttering my $HOME
true that.
psch for reference, i have 43 non-dot dirs in my $HOME on my linux partition
probably twice as many dot dirs, one of them .config which might have like 7 or so..? 03:41
zengargoyle 62 :)
psch not sure how to teach explorer.exe to show dot dirs :P
mst psch: 'ls -la' in your cygwin install :D 03:42
psch right, i *did* install cygwin somewhen...
mst zengargoyle: basically, I don't think defaulting to something that, as yet, most things don't default to, is a good idea 03:43
also I bet the things in .config are largely GUI apps
which I -would- expect to default to that
zengargoyle mine are sort of a mix, depends on where you draw the line between desktopy, guiy, graficy, texty. 03:48
i count it as something i got used to the same way i got used to ls -a mixing dot and regular and upper and lower cases all together. 03:53
geekosaur LC_COLLATE=C
zengargoyle yeah, just too much trouble to keep fighting the change. 03:54
03:54 xpen joined
zengargoyle like textfiles not being latin1 any more. 03:55
and \r\n being a single thing! the horrors! 03:56
i'm still all inner-ragey about newer input methods not bringin in good old XCompose functionality. 03:58
04:00 molaf joined 04:04 metasin_ is now known as metasin 04:19 AlexDaniel left 04:22 xpen left 04:25 herby_ joined
herby_ Good evening, everyone! 04:25
konobi like ldap/nis+ 04:28
herby_ I asked yesterday about parsing html and someone gave me a link to an example but I can't seem to find it 04:34
Any recommendations for parsing a web page after I've retrieved it using HTTP::Client?
geekosaur someone still uses nis+?! 04:35
konobi geekosaur: you'd be surprised 04:36
geekosaur my condolences
konobi haven't had to deal with it myself for a while
04:38 hankache joined
geekosaur doesn't see anything from you yesterday. or about HTTP::Client for that matter 04:39
04:41 lostinfog left
herby_ geekosaur, no? 04:41
geekosaur not in this channel
herby_ not sure how you're searching, are you able to see back to wednesday? 04:42
geekosaur previous time you said anything in my log was about a week ago
herby_ whaaat
geekosaur unles syou were using a different nick (I dropped the underscore for the search)
herby_ i was some version of herby :) 04:43
geekosaur 2015 Nov 12 03:47:58 <herby_>now it works!
herby_ i'm not loco, I swear
skarnsik, or skarnskik or something gave me the link
geekosaur last thing in my log from anyone whose nick contained "herby" in #perl6, before you joined a few minutes ago 04:44
04:44 aborazmeh left
herby_ in your log, do you see anything about rosettacod from yesterday? 04:45
because that was a response to my question
geekosaur ahhh. I didn't do case independent search
04:46 Ben_Goldberg joined, BenGoldberg left
geekosaur closest thing I see is 2015 Nov 20 02:12:40 <Skarsnik> Herby_, parse-html is from the Gumbo module if you are wondering, not in perl6 x) 04:47
herby_ yesss. skarsnik replied with a link to hit github
as a parsing example
psch herby_: the topic has a link to the public log for this channel, in case you want to search yourself
04:48 rigsby joined
herby_ perfect, i'll give that a shot. thanks! 04:48
psch not that i don't admire geekosaur++ for helping with that... :)
geekosaur 2015 Nov 20 02:00:44 <Skarsnik> Herby_, if you want an example gist.github.com/Skarsnik/cc6a997fe72b1caae73d
herby_ that's it! thanks geeko!
rigsby copied "words from file" code from perl6.org . getting the following error; No such method 'words' for invocant of type 'IO::Handle'
psch rigsby: does dict.txt exist? 04:50
Ben_Goldberg m: IO::Handle.^methods.say
camelia rakudo-moar : OUTPUT«(FALLBACK gist new)␤»
psch m: say $*IN.WHAT; say $*IN.can('words')
camelia rakudo-moar : OUTPUT«(Handle)␤(words)␤»
rigsby dict.txt is there
psch m: say $*IN ~~ IO::Handle
camelia rakudo-moar : OUTPUT«False␤»
Ben_Goldberg m: Handle.^methods.say
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/PyAh9SP7qk␤Undeclared name:␤ Handle used at line 1␤␤»
Ben_Goldberg m: say $*IN 04:51
camelia rakudo-moar : OUTPUT«IO::Handle<<STDIN>>(opened, at octet 0)␤»
Ben_Goldberg m: Handle.can('words').say
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/UqeqCbCMV2␤Undeclared name:␤ Handle used at line 1␤␤»
Ben_Goldberg m: IO::Handle.can('words').say
camelia rakudo-moar : OUTPUT«()␤»
psch rigsby: it works locally... what's your perl6 --version?
geekosaur m: 'dict.txt'.IO.^can('words').say 04:52
camelia rakudo-moar : OUTPUT«(words words)␤»
rigsby This is perl6 version 2013.12 built on parrot 5.9.0 revision 0
hmm
my bad
geekosaur oh. trather old
psch yes
rigsby did apt-get inst rakudo just now -- but maybe not using that
psch that's a bit old indeed 04:53
rigsby yea looked at IO:handle.^methods.say and there was no words
geekosaur rakudo has changed a lot just in the past couple months, let alone compared to 2013
04:53 skids left
psch rigsby: the recommended way of installing currently is rakudobrew, from github.com/tadzik/rakudobrew 04:53
geekosaur winding toward release
rigsby alrightey will try that psch
04:54 sprocket left
psch rigsby: feel free to ask any questions that might pop up, there's usually someone around who can at least point in a direction that might help 04:54
05:04 xpen joined 05:07 eliasr left 05:10 rigsby left 05:11 rigsby joined 05:12 xpen left, n0tjack joined
rigsby got rakudobrew added init to .zshrc ran build moar.. still no new perl6 yet will look at rakudobrew github page a bit more 05:13
guess i have to get rid of my /usr/bin/perl6 , or link to wherever rakudobrew puts things 05:14
geekosaur if you're in a shell that you didn't start after adding that, you'd need to run the same command in that shell 05:15
rigsby though my .rakudobrew/bin is prior to /usr/bin so if it's in there shouldn't it prefer the rakudo version?
geekosaur also: hash -r
shells remember where they saw things 05:16
rigsby ahh hash -r did it
niiice
thxx!!
geekosaur I'd still remove the package managfer one just to avoid confusion 05:17
05:17 n0tjack left
zengargoyle rigsby: which distro was that with the old perl6? 05:17
rigsby but the package manager one is the shiny new one that allows for method like IO::words 05:18
05:18 khw left
rigsby xubuntu 05:18
one sec and ill be more specific
Ubuntu 14.04.3 LTS 05:19
konobi ugh... i miss namespaces
zengargoyle ah, older LTS release
05:20 hankache left, herby_ left 05:21 Ben_Goldberg left
rigsby zengargoyle: were you suggesting getting rid of the .rakudobrew? 05:21
zengargoyle no, your ubuntu is a year and a half old, and one of the Long Term Stable releases, so isn't likely to get new updated packages. 05:22
05:23 raiph left
zengargoyle you should probably uninstall the old package you installed...: apt-get purge rakudo 05:23
most everybody here uses the rakudobrew'd version... you're good to go. 05:24
05:24 tokuhiro_ joined
zengargoyle if your system were Ubuntu 15.something non-LTS we could complain about the old rakudo package. :) 05:25
like Debian sid had the 2015.09 version now... probably newer Ubuntu also has a fresher version in the repositories. 05:26
05:26 kaare_ joined
geekosaur just wouldn't rely on distro packages for something that changes as quickly as rakudo has been 05:28
likewise I don't rely on calibre packages from a package manager since there's a new version ~ every Friday
05:29 tokuhiro_ left
awwaiid can someone explain .+ and .* to me? 05:51
m: say <5 6 7 8 23>.*roll 05:52
camelia rakudo-moar : OUTPUT«(7 6)␤»
awwaiid m: say 23.*is-prime 05:53
camelia rakudo-moar : OUTPUT«(True)␤»
awwaiid m: say <5 6 7 8 23>.roll
camelia rakudo-moar : OUTPUT«5␤»
05:55 llfourn joined
geekosaur foo.meth calls the most applicable method for the object, foo.?meth calls if it exists (without failing if it doesn't), foo.+meth calls all applicable methods instead of the most applicable one, foo.*meth like foo.?meth does that without failing if there aren't any 05:55
awwaiid ya. that's the doc that I want to update because I have no idea what the junk after .? means :) 05:56
geekosaur <5 6 7 8 23>.^can('roll')
er
m: <5 6 7 8 23>.^can('roll').say 05:57
camelia rakudo-moar : OUTPUT«(roll Method+{<anon|61865648>}.new)␤»
awwaiid maybe rakudo isn't doing what it is supposed to, but in the above you can see that .roll and .*roll are returning different things; also .* gives back a list instead of a single value
geekosaur there are two applicable methods returned by "can" there; I don't know what the second one is...
anyway .* or .+ will call all of them and return a list of the results
awwaiid ahhh... interesting 05:58
why in the world would anyone want to do that?
those other methods come from like inheritance or roles or some such thing, ya?
geekosaur tbh I don't know why you'd need that 05:59
design.perl6.org/S12.html#Multisub_Resolution might be enlightening 06:00
awwaiid oh, you think it is applicable 'multi' methods rather than overridden inherited ones? interesting 06:01
geekosaur I think inheritance is just a special case of that
awwaiid even better
m: 23.^can('floor').perl.say 06:03
camelia rakudo-moar : OUTPUT«(method floor (Int:D $: *%_) { #`(Method|43504952) ... }, method floor (Cool $: *%_) { #`(Method|43535808) ... })␤»
geekosaur so anyway .+ calls all applicable methods instead of just the "best" one; .* is to .+ as .? is to plain .
(not failing if no methods are considered applicable)
awwaiid gotchya
ok. Not clear on the value, but that helps me understand the mechanics
thanks! 06:04
06:10 geraud left 06:11 redhands joined 06:22 karim32 joined
karim32 I have code in Python which makes an http post request and works well. The same code returns 400 Bad Request in Perl 6. Could anyone help me? 06:23
It's very simple, a couple of lines of code.
pastebin.com/DapXdVMt
awwaiid can you post it in a gist or something?
ah, you beat me to it
karim32 I'm using this lib github.com/sergot/http-useragent It seems to be de-facto for Perl 6 at the moment 06:24
awwaiid interesting. do you own the server as well? It'd be interesting to see the different log entries
konobi awwaiid: oh... long time no see.
karim32 No, it's not mine server
awwaiid greetings konobi 06:25
karim32 Or if there's a better http library, please share its name
redhands why does you $http-client variable have a dash in it? 06:26
awwaiid redhands: that is the perl6ish way to name things
redhands ah I see
awwaiid lisp-style (kabob)
redhands neat 06:27
awwaiid interestingly, you're not allowed to put the dash at the beginning or end, and can't have two in a row
makes it very much a word-separator
redhands ah I suppose that makes sense 06:28
awwaiid karim32: I'm checking this lib out, see if anything jumps out at me 06:29
karim32 ok, thanks 06:30
I'm going offline, I don't get along with irc, how can I get back to you? Or will you ping me? 06:31
email?
06:31 polettix joined
karim32 I've opened an issue at the "issue" section at that library 06:31
at github
awwaiid karim32: maybe try HTTP::Request.new(POST => 'api.some_website.org/some_params', var1 => 1234, var2 => 'some text') 06:32
ah cool -- I'll watch the issue on that lib
and reply there if I learn anything useful
karim32 not working 06:34
ok
06:34 karim32 left
awwaiid my guess is still that there is some encoding issue going on that the server doesn't like 06:34
I should go to sleep anyway though :) 06:35
awwaiid &
redhands tempted to boot up a test webserver with one post endpoint to recreate/confirm the issue 06:36
gfldex m: my $b = 1; my Int $a := Proxy.new(FETCH => { $b * 2}, STORE => -> $, $new { $b = $new + 1 } ); say $a; $a = 4; say $a; $a.VAR.^name.say; $a.WHAT.say; $a.^name.say; 06:40
camelia rakudo-moar : OUTPUT«2␤10␤Proxy␤(Int)␤Int␤»
gfldex m: my Int $a := 1.2; 06:41
camelia rakudo-moar : OUTPUT«Type check failed in binding; expected Int but got Rat␤ in block <unit> at /tmp/TJ0TvR5RB3:1␤␤»
gfldex is Proxy magic when it comes to type checks when binding? 06:42
konobi probably no more than anything else 06:43
gfldex m: my \r = Proxy.new(FETCH => -> {Nil}, STORE => -> $,$ {}); my $a := r; $a = 1; 06:49
camelia ( no output )
gfldex m: my \r = Proxy.new(FETCH => -> {Nil}, STORE => -> $,$ {}); my $a := r; $a = 1; say $a;
camelia rakudo-moar : OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in block <unit> at /tmp/o_WdlmrS2q:1␤␤»
gfldex m: my \r = Proxy.new(FETCH => -> {2}, STORE => -> $,$ {2}); my $a := r; $a = 1; say $a;
camelia rakudo-moar : OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in block <unit> at /tmp/18_f7g9aEd:1␤␤»
konobi gfldex: multimethods, no?
gfldex there are no multies involved 06:50
m: my \r = Proxy.new(FETCH => -> {2}, STORE => method ($name) {2}); my $a := r; $a = 1; say $a; 06:51
camelia rakudo-moar : OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in block <unit> at /tmp/ylKkfOMNSq:1␤␤»
konobi gfldex: um... isn't that more mop worthy?
gfldex i'm not trying to implement anything
i'm trying to break rakudo, what is getting hard lately :) 06:52
konobi so a proxy that does nothing?
gfldex i doubt it would change if it wouldn't be a black hole 06:53
m: my $b = 1; my \r := Proxy.new(FETCH => { $b * 2}, STORE => -> $, $new { $b = $new + 1 } ); my $a := r; say $a; $a = 4; say $a; $a.VAR.^name.say; $a.WHAT.say; $a.^name.say; 06:55
camelia rakudo-moar : OUTPUT«2␤10␤Proxy␤(Int)␤Int␤»
gfldex m: my $b = 1; my \r := Proxy.new(FETCH => {}, STORE => -> $, $new { $b = $new + 1 } ); my $a := r; say $a; $a = 4; say $a; $a.VAR.^name.say; $a.WHAT.say; $a.^name.say; 06:56
camelia rakudo-moar : OUTPUT«Cannot find method 'CALL-ME'␤ in block <unit> at /tmp/n8ZHpgdfHh:1␤␤»
gfldex m: my $b = 1; my \r := Proxy.new(FETCH => sub {}, STORE => -> $, $new { $b = $new + 1 } ); my $a := r; say $a; $a = 4; say $a; $a.VAR.^name.say; $a.WHAT.say; $a.^name.say;
camelia rakudo-moar : OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in sub at /tmp/nwvDmZ8hTl:1␤ in block <unit> at /tmp/nwvDmZ8hTl:1␤␤»
gfldex technically LTA, but i guess if you are in Proxy territory you stand on quicksand anyways 06:57
m: gist.github.com/gfldex/a45c8c592ff963ba53bc 07:00
camelia rakudo-moar : OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in sub at /tmp/ILpByKKREm:3␤ in block <unit> at /tmp/ILpByKKREm:7␤␤»
gfldex the line number is helpful so it's all good
07:02 TEttinger joined 07:07 rurban joined 07:12 rigsby left
gfldex m: my \a = Proxy.new(FETCH => { state $b; print 'fetched '; $b++ }, STORE => method ($new) { self } ); say a,' ',a; 07:12
camelia rakudo-moar : OUTPUT«fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched 8 17␤»
gfldex what explains whay a Proxy that does nothing wont work. Rakudo really likes to call FETCH. 07:14
moritz well, it has to multi-dispatch, type check, coerce, ... 07:16
gfldex m: my \a = Proxy.new(FETCH => { state $b; print 'fetched '; $b++ }, STORE => method ($new) { self } ); say q{I ain't do nothing!}; 07:17
camelia rakudo-moar : OUTPUT«I ain't do nothing!␤»
Woodi m: my $a = Proxy.new(FETCH => { state $b; print 'fetched '; $b++ }, STORE => method ($new) { self } ); say $a,' ',$a; 07:22
camelia rakudo-moar : OUTPUT«fetched 0 0␤»
Woodi that's surprising to me... 07:23
gfldex m: my $a = Proxy.new(FETCH => { state $b; print 'fetched '; $b++ }, STORE => method ($new) { self } ); $a = 'changed'; say $a;
camelia rakudo-moar : OUTPUT«fetched changed␤»
gfldex m: my $a := Proxy.new(FETCH => { state $b; print 'fetched '; $b++ }, STORE => method ($new) { self } ); $a = 'changed'; say $a;
camelia rakudo-moar : OUTPUT«fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched 9␤»
gfldex \a forces binding
m: my Int $a := Proxy.new(FETCH => { state $b; print 'fetched '; $b++ }, STORE => method ($new) { self } ); $a = 'changed'; say $a; 07:25
camelia rakudo-moar : OUTPUT«fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched fetched 10␤»
07:26 tokuhiro_ joined
gfldex what gets me back to my original quesion, how Proxy binding and type checks work together 07:26
Woodi looks current Proxy is very volatile-like... 07:27
gfldex m: my $b=0; my Int $a := Proxy.new(FETCH => { print 'fetched '; $b++ }, STORE => method ($new) { $b = $new; } ); $a = 'changed'; say $a; 07:28
camelia rakudo-moar : OUTPUT«fetched fetched fetched fetched fetched fetched changeh␤»
07:28 soshi18 joined
soshi18 I need to uninstall/install or upgrade a Perl 6 package using Panda, I did this: 07:28
panda uninstall package1
but it ran the tests for some reason. How can I do that properly?
07:31 tokuhiro_ left
gfldex m: sub silently-returns-a-proxy () is rw { my $b = 10; Proxy.new(FETCH => { print '.'; $b++ }, STORE => method ($new) { $b = $new; } )}; my Int $a := silently-returns-a-proxy(); say $a; $a = q{that's bad}; say $a; 07:32
camelia rakudo-moar : OUTPUT«...........20␤.....that's bah␤»
gfldex m: sub silently-returns-a-proxy () is rw { my $b = 10; Proxy.new(FETCH => { print '.'; $b++ }, STORE => method ($new) { $b = $new; } )}; my Int $a = silently-returns-a-proxy(); say $a; $a = q{that's bad}; say $a; 07:33
camelia rakudo-moar : OUTPUT«.10␤Type check failed in assignment to $a; expected Int but got Str␤ in block <unit> at /tmp/svlSVcVwWS:1␤␤»
gfldex so if you suspect some sub in module might return a proxy, never ever mix type checks with bindings.
07:33 soshi18 left
gfldex m: sub silently-returns-a-proxy () is rw { my $b = 10; return-rw Proxy.new(FETCH => { print '.'; $b++ }, STORE => method ($new) { $b = $new; } )}; my Int $a = silently-returns-a-proxy(); say $a; $a = 'bad'; say $a; 07:39
camelia rakudo-moar : OUTPUT«.10␤Type check failed in assignment to $a; expected Int but got Str␤ in block <unit> at /tmp/hGdl_C13em:1␤␤»
gfldex m: sub silently-returns-a-proxy () is rw { my $b = 10; return-rw Proxy.new(FETCH => { print '.'; $b++ }, STORE => method ($new) { $b = $new; } )}; my $a = silently-returns-a-proxy(); say $a; $a = 'bad'; say $a;
camelia rakudo-moar : OUTPUT«.10␤bad␤»
gfldex m: sub silently-returns-a-proxy () is rw { my $b = 10; return-rw Proxy.new(FETCH => { print '.'; $b++ }, STORE => method ($new) { $b = $new; } )}; my $a = silently-returns-a-proxy(); say $a; $a = 'bad'; say $a, $a.VAR.^name; 07:40
camelia rakudo-moar : OUTPUT«.10␤badScalar␤»
gfldex as I understand it Proxies are ment to be used as lvalues with state. If I asign that lvalue to a variable (and don't use :=) the bahaviour changes. If I do use binding, the behaviour changes back. 07:44
Woodi m: my Int $a = Proxy.new(FETCH => { state $b; print 'fetched '; $b++ }, STORE => method ($new) { self } ); $a = 1; say $a;
camelia rakudo-moar : OUTPUT«fetched 1␤»
gfldex I don't think I like this.
japhb .tell timotimo Thanks for taking care of perl6-bench; much appreciated!
yoleaux japhb: I'll pass your message to timotimo.
gfldex On top of that, if I use binding, type checks vanish. 07:45
07:47 domidumont joined 07:51 rindolf joined 07:52 domidumont left, domidumont joined 07:54 CIAvash joined 07:56 polettix left, polettix_ joined
Woodi gfldex: if you assign Proxy somewhere then it is not used anymore... 07:56
08:04 dwarring joined 08:09 polettix_ left 08:10 polettix joined 08:16 darutoko joined 08:24 dwarring left, arnsholt_ joined
dalek osystem: 23139e0 | (David Warring)++ | META.list:
Add PDF-Tools

This is an intermediate modules for the reading and writing of PDF data.
It builds on PDF::Grammar adding serialization functions such as compression filters, encryption and indirect object fetching via index and writing of PDF files. It is capable of reading, editing and creation or incremental update of PDF files. Hopefully this will provide a useful set of classes and utilities for implementing higher level PDF modules.		 08:24
08:26 johan joined
dalek osystem: 32d73f5 | (David Warring)++ | META.list:
Update META.list

Correct just added PDF::Tools		 08:29
08:38 johan left, redhands left 08:52 rindolf left
dalek ast: 4a5edf7 | usev6++ | S02-types/array-shapes.t:
Unfudge and adjust tests for array shapes RT #124502
 09:11
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=124502
09:21 domidumont left 09:27 tokuhiro_ joined 09:30 domidumont joined 09:31 tokuhiro_ left
dalek ast: 0dd18f4 | usev6++ | S (3 files):
Unfudge tests for JVM RT #126679, psch++
 09:37
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126679
09:42 diana_olhovik_ joined
dalek ast: e2174a8 | usev6++ | S0 (3 files):
Unfudge tests for JVM RT #126673, psch++
 09:43
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126673
09:45 Peter_R joined 09:46 diana_olhovik_ left 09:53 nmp joined 09:54 diana_olhovik joined 09:55 xfix joined 09:56 nmp left 09:59 xpen joined 10:04 sprocket joined 10:06 Begi11111 joined 10:10 kjs_ joined 10:13 BinGOs_ is now known as BinGOs 10:14 BinGOs left, BinGOs joined
dalek kudo/curli: 1ca7b51 | (Stefan Seifert)++ | src/core/CompUnit/Repository/Installation.pm:
Do not load a module twice from Repository::Installation

Repository::FileSystem already had the check and the infrastructure was actually in place, so just use it.		 10:18
10:19 zhmylove left 10:28 znpy joined 10:31 diana_olhovik left 10:35 vendethiel joined
dalek ast: dd9eda9 | usev6++ | S32-str/index.t:
Fudge tests for JVM RT #126700
 10:39
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126700
10:40 noganex_ left 10:41 noganex joined
bartolin r: say index("xxy", "y", -1) 10:42
camelia rakudo-moar: OUTPUT«Nil␤»
..rakudo-jvm 0b6328: OUTPUT«2␤»
bartolin neither docs nor design documents seem to say how a negative start position should be handled 10:43
perl -E 'print index("xxy", "y", -1), "\n"'
2
10:44 arnsholt_ is now known as arnsholt 10:48 xpen left 10:50 hankache joined
moritz ß 10:51
0''''''
o
nine .tell jnthn I have an experimental patch for panda that makes it install into a CompUnit::Repository::Installation and successfully installed 87 distributions :) Startup time of perl6 -e 'use Test' is 0.239s 10:52
yoleaux nine: I'll pass your message to jnthn.
dalek ast: 2c9f42e | usev6++ | / (2 files):
Fudge tests for FIRST phase on JVM, RT #126701
 10:57
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126701
bartolin nine++
11:02 Actualeyes left 11:04 hunterzeroxxx joined 11:05 hunterzeroxxx left
dalek ast: a80e467 | usev6++ | S32-str/tc.t:
Fudge tests for .tc on JVM
 11:06
11:06 xfix_ joined 11:07 edehont joined, diana_olhovik_ joined, zhmylove joined 11:08 xfix left 11:10 Woodi left 11:15 rindolf joined 11:17 tokuhiro_ joined, domidumont left 11:19 g4 joined 11:20 Actualeyes joined, xpen joined 11:21 tokuhiro_ left 11:24 loren joined 11:25 domidumont joined 11:26 Woodi joined 11:30 rgrau joined 11:31 domidumont left 11:32 xpen left 11:37 jnthn_ is now known as jnthn 11:40 pmurias joined, zhmylove left 11:44 zhmylove joined
dalek kudo/curli: 3c814de | (Stefan Seifert)++ | src/core/CompUnit/Repository/Installation.pm:
Use the precomp repository of the head of the chain on installation

The need method is passed the precomp repository by its predecessor in the chain. The install method is called directly but it still needs to use the head's precomp repository as the full chain is used to resolve dependencies during precompilation and we want to install into the store that is actually used for loading.		 11:50
ast: 498ede8 | usev6++ | S (2 files):
Fudge tests for IN.nl-in for JVM
 11:54
kudo/curli: 4291711 | (Stefan Seifert)++ | / (4 files):
Clean up default INC paths and install CORE distribution

We install the modules that come with rakudo itself as a CORE distribution into a CompUnit::Repository::Installation. No more need to manually precompile modules in the Makefile		 11:55
11:56 Skarsnik joined
jnthn nine++ \o/ 11:59
yoleaux 10:52Z <nine> jnthn: I have an experimental patch for panda that makes it install into a CompUnit::Repository::Installation and successfully installed 87 distributions :) Startup time of perl6 -e 'use Test' is 0.239s
jnthn nine: What's perl6 -e '' for you, ooc?
12:01 xpen joined, kid51 joined 12:02 xpen left, loren left 12:03 FROGGS joined 12:06 black_ant joined
dalek ast: b4b4dfa | usev6++ | S32-basics/xxKEY.t:
Fudge tests for typename in X::Assignment::RO on JVM
 12:07
12:11 xpen joined 12:12 xpen left 12:13 black_ant left 12:18 rgrau left
dalek kudo/bool_enum: 40cba58 | (Stefan Seifert)++ | src/ (3 files):
Turn Bool into a proper enum
 12:22
kudo/bool_enum: 444b8cb | (Stefan Seifert)++ | src/core/operators.pm:
Fix SEQUENCE not expecting Bool ~~ Real to be true
kudo/bool_enum: e4af648 | (Stefan Seifert)++ | src/Perl6/Metamodel/EnumHOW.nqp:
Fix type check for binding parameters

EnumHOW just needed an is_composed method like many other HOWs. Fixes regression in t/spec/S06-signature/optional.t
kudo/bool_enum: 6d0ca40 | (Stefan Seifert)++ | src/ (2 files):
Bool improvements
jnthn I just rebased that on top of latest, fixing a conflict or two.
12:22 TimToady left 12:23 tokuhiro_ joined
jnthn spectests it to figure out the bustage that's blocking its merge :) 12:23
12:24 black_ant joined, TimToady joined 12:26 Zoffix left, black_ant left, xpen joined 12:28 tokuhiro_ left, black_ant joined 12:30 g4 left 12:31 xpen left 12:33 xfix_ left
nine jnthn: about 140mss 12:34
12:35 dbohdan_ is now known as dbohdan
jnthn nine: OK :) 12:35
nine: I'm looking into that weird BEGIN/add_method thing in bool_enum
nine jnthn: bool is blocking on a Bool.pick(*) segfault
as far as I remember 12:36
12:36 jsimonet1 is now known as jsimonet
jnthn Well, the SEGV in Moar got patched, so it's now an error 12:36
nine Visiting relatives right now
jnthn Enjoy :)
nine Thanks :)
12:40 black_ant left 12:43 hankache left 12:45 eliasr joined 12:46 black_ant joined
dalek href="https://modules.perl6.org:">modules.perl6.org: ad03cd7 | (Zoffix Znet)++ | / (3 files):
Toss module we no longer use form prereqs
 12:47
12:49 xfix joined
dalek href="https://modules.perl6.org:">modules.perl6.org: 4639c48 | (Zoffix Znet)++ | web/t/01-models/03-site-tips (2 files):
Added tests for site tips
 12:55
12:57 telex left 12:58 telex joined 12:59 black_ant left 13:03 CIAvash left 13:06 Begi11112 joined 13:07 Begi11111 left, sprocket left 13:14 johan joined
stmuk hmm the perl6 bigint performance seems a lot faster than perl5 13:26
13:28 leont joined
rindolf stmuk: even when using the Math::GMP module? 13:29
dalek kudo/bool_enum: 94385dd | jnthn++ | src/Perl6/ (2 files):
Fix BEGIN/CHECK block scoping precomp bug.

This in turn unbusts .pick and friends in Bool after the changes to make it a real enum.
pmurias nine's work will allow automatic precompilation?
jnthn pmurias: s/will allow/allows/ :)
pmurias that's really great, nine++ 13:30
jnthn Indeed
stmuk rindolf: I imagine if I use a C lib then perl5 would be faster 13:31
jnthn stmuk: Perl 6's bigint operations are implemented in C, so... :)
stmuk an external C library :)
jnthn nine: So, 94385dd hopefully unblocks thing 13:32
*thing
*thing*s*!
stmuk I can use the same script under perl5 and perl6 :D
well with -Mbigint in the perl5 case
timotimo urgh, it's so frustrating to watch people struggle with some rakudo/perl6 concept in the backlog and not being able to help them
yoleaux 07:44Z <japhb> timotimo: Thanks for taking care of perl6-bench; much appreciated!
arnsholt jnthn: Here's a question: Let's say I pass a Perl 6 sub as a callback off into C land, and a *different* thread invokes it. I'm gonna have a bad time, aren't I?
jnthn arnsholt: Uh... 13:33
arnsholt: Probably.
13:33 TEttinger left
jnthn arnsholt: We're not doing anything special to cope with that case 13:33
arnsholt Yeah, that's what I thought
13:33 hankache joined
jnthn arnsholt: I believe it's a solvable problem, but it's not an easy one. :) 13:33
I think we do already handle it on JVM 13:34
arnsholt Oh, neat
dalek href="https://modules.perl6.org:">modules.perl6.org: 70c9ef5 | (Zoffix Znet)++ | / (3 files):
ModulesPerl6::Model::SiteTips is complete
jnthn On Moar it'll need some effort
Though I can imagine not crazily much
arnsholt Reason I'm asking is because I suspect that's what happens in ZMQ when it's done with data passed into the library
jnthn ah 13:35
arnsholt There's an intermittent bug in the IPerl6 stuff where data gets corrupted, which I'm reasonably sure is because the GC tampers with data before ZMQ is done sending it
timotimo jnthn: something i saw in the backlog is that Proxy has no clue about typechecks; should it learn about types? 13:36
jnthn: because currently you can my Int $foo := Proxy.new(...) and then the proxy is allowed to willy-nilly return a Str instead
jnthn timotimo: Dunno. I'm not sure how to make that happen.
timotimo me neither
jnthn Well, maybe we can refuse to bind an untyped container or so 13:37
timotimo i'm also not sure if we can detect a Proxy being :='d into ... yeah
something that occupied my mind before i could fall asleep is: how do hyperops handle shaped arrays when shapes mismatch and the pointies are pointed at one of the shaped arrays?
will an element just get the value at the position modulo the size of the small array? 13:38
i'd like to put tests into roast for this if there's already a consensus about this
Skarsnik arnsholt, even with setting explicitly-manage on the data?
jnthn timotimo: I'm...not sure, and I don't know that we get them right on shaped arrays at all yet. 13:39
timotimo: That'd certainly be a good thing to make happen.
timotimo i'm sure we're not getting that right yet at all
jnthn Aye
timotimo after the tests exist, someone can go ahead and build hyperop candidates for shaped arrays
and when that exists, more thought can be put into how moar would be able to vectorize that 13:40
jnthn Feel free to work on it. ;) I'd say that for now we can just handle shaped
uh
Just handle identail shapes
*identical
pmurias jnthn: re my closures question from yesterday, it turns out that the block without captured lexicals is used when reseting the regex instead of the closure
jnthn pmurias: "oops" :) 13:41
pmurias: In Rakudo too, or just in NQP? 13:42
dalek c: 14fbd17 | (Steve Mynott)++ | doc/Language/faq.pod:
remove claims about slower and faster
jnthn timotimo: An slightly looser way to go would be to just handle upgrade in the (shaped, scalar) and (scalar, shaped) case 13:43
timotimo ah, sure
that's a very useful thing in many cases
jnthn And if we see two different shapes or a shaped and unshaped then we die
pmurias jnthn: not sure about Rakudo, will check and try to exploit it to create a failing test after family visit & lunch
dalek ast: 9aa4c45 | usev6++ | S14-roles/mixin.t:
Fudge test for '0 but True' on JVM, RT #100782
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=100782
jnthn Then we reserve it for when we've got a better idea of what, if anything, we should do 13:44
(I suspect it's the kind of area PDL-using folks might have some input on)
timotimo jnthn: i've also considered we could maybe annotate an assign metaop like += or -= so that the hyper sub could recognize it, unpack the inner operator and just work on the LHS instead of cloning
as in, the assign metaop would mix in a role that tells us what's up
13:45 znpy left
jnthn Hm, could be an interesting optimization also 13:45
timotimo or perhaps a special case in the grammar that only triggers when an assign metaop is inside a hyperop already
jnthn OK, lunch time here... :)
timotimo it's an optimization i suspect most people will expect
arnsholt Skarsnik: Yeah, explicitly-manage works around it, mostly. But only by extending the lifetime of the variable, making it almost inevitable that the data is transmitted before the GC gets around to it
But I wouldn't bet on that holding in a lossy enviroment, for example
13:46 Zoffix joined
dalek p: 0d93c7c | (Pawel Murias)++ | src/vm/js/nqp-runtime/code-ref.js:
[js] Add a hack to allow calling a codeRef whose closure hasn't been captured yet.

This replaces the hack which twisted the closure model. This should be done properly but is a step in the right direction.		 13:51
p: d44de00 | (Pawel Murias)++ | src/vm/js/QAST/Compiler.nqp:
[js] Avoid emitting js for closures in void context. This reduces the size of code a lot.
ast: 28cafa5 | usev6++ | S14-roles/basic.t:
Fudge test for role mixin on JVM
 13:54
ast: 5ab83b7 | usev6++ | S14-roles/basic.t:
Unfudge test for RT #120919 on JVM
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=120919
c: 59616f6 | (Steve Mynott)++ | doc/Language/faq.pod:
an example program using bigints which runs under 5 and 6
timotimo updating some rakudo 13:56
gfldex: are you awake right now?
13:58 Zoffix left, lab_ left 13:59 vendethiel left
dalek ast: a8e5644 | usev6++ | S12-subset/subtypes.t:
Fudge another test for role mixin on JVM
 14:05
stmuk is there a name for perl which runs under both perl 5 and perl 6 (apart from "very basic")!
universal perl maybe
timotimo Michael McDoesn'texist? 14:06
moritz poloyglot
pink_mist perlyglot? :P
14:08 lab_ joined
zengargoyle good * #perl6 14:08
dalek href="https://modules.perl6.org:">modules.perl6.org: e2f4cbc | (Zoffix Znet)++ | / (7 files):
Implement Site Tips (Closes #32)
 14:09
stmuk perlygot++ 14:12
timotimo zoffix is working on the site but not on irc? :(
dalek ast: 6fb9b12 | usev6++ | S06-multi/subsignature.t:
Fudge test for JVM, RT #126702
 14:13
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126702
14:15 znpy joined
dalek href="https://modules.perl6.org:">modules.perl6.org: ebfcfd7 | (Zoffix Znet)++ | web/t/ (2 files):
Fix failing sprite tests
 14:15
14:17 xpen joined
dalek ast: 3583fd3 | usev6++ | S06-multi/proto.t:
Fudge tests for JVM, wrong multi candidate called
 14:18
14:18 Zoffix joined
Zoffix timotimo, I'm more productive that way. Why did you want anything? 14:19
timotimo will there be an author:* search thingie?
Zoffix eventually yeah 14:20
timotimo neat :)
stmuk ewww did .cn DOS github by moving things around
Zoffix stmuk, no, it's the "new and more streamlined" look they went with 14:21
stmuk or worse still a "Designer" change it
pink_mist stmuk: they've made updates to the interface, yes
Zoffix I find I have to move my mouse farther to get to Issues/PRs now, so I'm not sure what's "streamlined" about that
It also looks much uglier
timotimo i prefer the tabs on top rather than the little icons to the right, tbh 14:22
Zoffix Traitor! :)
Skarsnik what is GitLike configuration in github.com/perl6/perl6-most-wanted...modules.md ?
14:22 sprocket joined
Zoffix goes back into the shadows 14:23
14:23 Zoffix left
timotimo Skarsnik: how the .git/config and ~/.gitconfig files look 14:23
14:24 tokuhiro_ joined
Skarsnik oh it's kinda like ini x) 14:24
14:25 sprocket left
jnthn timotimo: (tabs on top) me too, fwiw :) 14:26
timotimo jnthn: we can already @a >>+>> 1, but the result isn't shaped
14:29 hankache left, tokuhiro_ left, hankache joined 14:36 sprocket joined
jnthn timotimo: Yeah, .iterator on a shaped just gives you the values. 14:37
timotimo m: my Int @a[10;10] = (0 xx 10) xx 10; say @a # isn't this how you're supposed to initialize a shaped array like that? 14:38
camelia rakudo-moar : OUTPUT«Type check failed in assignment to ; expected Int but got List␤ in block <unit> at /tmp/pMaPButpx3:1␤␤»
14:38 diana_olhovik_ left 14:39 sprocket left
jnthn Not sure if Larry'll agree, but I think hypermap is probably where we want to put shape-retaining mapping, and .map is just over the values 14:39
timotimo how about mapping and modifying values inside the map? 14:40
jnthn m: say ((0 xx 10) xx 10).perl
camelia rakudo-moar : OUTPUT«((0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0, 0, 0, 0, 0, 0, 0, 0, 0), (0, 0…»
jnthn m: say ((0 xx 10) xx 10).WHAT
camelia rakudo-moar : OUTPUT«(List)␤»
jnthn Curious... 14:41
14:41 xpen left
timotimo how do you initialize the lists in your multidim tests? :) 14:41
jnthn Curious...I'd kinda have expected that to be a Seq...
m: my @a[3;3] = <a b c>, <d e f>, <g h i>; say @a 14:42
camelia rakudo-moar : OUTPUT«[[a b c] [d e f] [g h i]]␤»
timotimo mhh
jnthn m: my @a[3;3] = <a b c> xx 3; say @a
camelia rakudo-moar : OUTPUT«[[a b c] [a b c] [a b c]]␤»
timotimo m: my Str @a[3;3] = <a b c>, <d e f>, <g h i>; say @a
camelia rakudo-moar : OUTPUT«Type check failed in assignment to ; expected Str but got List␤ in block <unit> at /tmp/XObUYCKoZZ:1␤␤»
jnthn m: my @a[3;3] = ('a' xx 3) xx 3; say @a
camelia rakudo-moar : OUTPUT«[[a a a] [a a a] [a a a]]␤»
jnthn Oh...
timotimo it's really the type annotation; the type check seems to happen in a/the wrong place
jnthn Yes, looks like it.
Please RT
timotimo sure 14:43
jnthn Hm, I think I totally undertested typed shaped object arrays. :/
Natively typed shaped ones got some hundred plus tests.
timotimo my first instinct was to use those %)
psch jnthn: ooc, what was the hint to fix bool_enum by changing the QAST that has to be compiled instead of the compilation process itself?
jnthn psch: I first reduced it to a pre-comp bug, which showed it was about BEGIN and blocks inside them. Then was like "wait, but role bodies are BEGIN-time and things work there." Then went and looked at roles and...ooh. :) 14:45
psch ah, so the issue was that BEGIN blocks were missing the same fixup roles already got? 14:46
jnthn Yeah
I fixed CHECK ones while I was at it too
14:46 lab joined
jnthn Should probably add precomp regression tests, but I was getting lunch, and getting food/beer delivered. :) 14:47
lizmat jnthn; re xx 10 not being a Seq: it's not because of performance
14:47 lab_ left
jnthn lizmat: ? 14:47
lizmat jnthn: if someone asks for a finite repeat, they clearly know how much they want
xx Whatever is a Seq
psch jnthn: thanks for the explanation :)
14:48 stmuk_ joined
jnthn lizmat: for foo() xx 100000 { } # this will retain all the results of the foo() calls 14:48
lizmat yes, it would
but how often would you do that?
jnthn That's not good...we'll lose a bunch on GC
What benchmark showed a faster result? 14:49
I can quite easily imagine people using xx in benchmarking...
lizmat most uses I've seen of xx N have way lower values for N
jnthn It's also uncomfortable that xx has an inconsistent return type
14:49 edehont left
jnthn my @a = foo() xx 10; # I'd still expect this to work out faster with the Seq than List 14:50
14:50 stmuk left
jnthn 'cus it'd avoid the intermediate list creation 14:50
So even for lower values I'm still not quite seeing it.
lizmat jnthn: pretty sure it didn't, but let me double check 14:51
timotimo RT #126703
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126703
moritz m: for 1 xx * { .say; last }
camelia rakudo-moar : OUTPUT«1␤»
jnthn Of course, I'm assuming that push-exactly was implemented :)
m: for 1 xx 1000000 { last }
camelia ( no output )
jnthn oh, duh
wtf, brain :)
Oh, but no, that was what I wanted to try 14:52
m: dd (1 xx 1000000)
camelia rakudo-moar : OUTPUT«List $var = $(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…»
jnthn ah :)
m: say nqp::elems nqp::getattr((1 xx 1000000), List, '$!reified')
camelia rakudo-moar : OUTPUT«1000000␤»
jnthn hah, so it is just fast at doing that :)
timotimo m: my @a[4;4] = (1 xx *) xx *; 14:53
camelia rakudo-moar : OUTPUT«Index 4 for dimension 2 out of range (must be 0..3)␤ in block <unit> at /tmp/0sPnTSvfbc:1␤␤»
timotimo it would be cool if that could work, jnthn
agree/disagree?
14:53 vendethiel joined, Begi11112 left
lizmat jnthn: if you think the xx N case should be turned into a Seq again, then I'll do that 14:54
jnthn Disagree, the * isn't something the shaped array store can know about.
timotimo OK
14:54 loren joined
timotimo so i'll always have to copy-paste the dimensions from the LHS to the RHS 14:54
lizmat jnthn: but later today, now in the middle of packing
loren Hi, evening , #perl6
timotimo heyo loren 14:55
14:55 oceanpollen joined
jnthn lizmat: Yes, please. I can appreciate it may have been faster with List, but from a design point of view the List return type there worries me. 14:55
lizmat: Whenever it's comfortable :)
timotimo: Well, what we're trying to do is tell people about things like:
loren em, timotimo
jnthn m: my @a[2] = 0, 1, 2;
camelia rakudo-moar : OUTPUT«Index 2 for dimension 1 out of range (must be 0..1)␤ in block <unit> at /tmp/rt3Uon2tl2:1␤␤»
jnthn Which is a clear thinko.
loren I got some problem with new version rakudo 14:56
timotimo ah. right.
loren Is method ins deprecated in IO::Handle ?
timotimo will we have a way to assign a flat list of $x * $y values into a shaped array at some point?
jnthn loren: Yes
loren What should i use instead of it .
jnthn We could perhaps look at .is-lazy and decide not to complain on lazy things 14:57
timotimo loren: you can use .lines.kv for example
jnthn loren: If you're reading line by line, then for $handle.lives.kv -> $line-number, $line { }
loren Oh, the 'key' is line number ..
great !
14:58 lab left
loren thks help .timotimo jnthn 14:58
timotimo that'll start you at 0, btw
14:58 lab joined
loren Em, i see. 14:58
jnthn timotimo: The .is-lazy would make your example work out with xx *; I wonder what the WAT of that DWIM is :) 14:59
timotimo i wonder, too
jnthn Can't think of one right away, so maybe it would be OK, and helpful.
15:00 xfix left
jnthn lizmat: BTW, about the .lines de-optimization: losing 60% is rather sad, but it occurred to me we can still cheat perfectly fine if self.WHAT === IO::Handle - that is, there're no mixins. So we can have our cake and eat it there. 15:01
Again, no hurry. :)
jnthn gotta go for a bit
bbl
lizmat jnthn: on that note, I was also thinking of removing the $!chomp test in .get 15:04
and mixin a different .get if $!chomp is not set
would that make sense ? 15:05
15:12 oceanpollen left
bartolin lizmat: earlier today I found this differing behaviour for moar and jvm: 15:14
r: say index("xxy", "y", -1)
camelia rakudo-moar: OUTPUT«Nil␤»
..rakudo-jvm 0b6328: OUTPUT«2␤»
bartolin do you think moar is right here?
15:14 lab left
lizmat looks into it 15:15
bartolin lizmat++ 15:16
lizmat r: use nqp; say nqp::index("xxy","y",-1)
camelia rakudo-jvm 0b6328: OUTPUT«2␤» 15:17
..rakudo-moar: OUTPUT«-1␤»
15:17 lab joined
lizmat looks like a difference in implementation of nqp::index 15:17
hankache hello everyone
timotimo hey there
15:17 raiph joined
lizmat r: use nqp; say nqp::index("xxy","y",0) 15:18
camelia rakudo-moar, rakudo-jvm 0b6328: OUTPUT«2␤»
lizmat bartolin: seems worthy of an NQP ticket
or perhaps just a rakudobug :-)
bartolin lizmat: and what do you think is the "better" answer? Perl 5 allows negative start positions, if I'm not mistaken. 15:19
lizmat yes, but P5 also allows to use -1 to indicate from end
hankache In the process of learning Perl 6 (still learning) I managed to write an introductory document. The primary motive being to have an entry point for newcomers.
github.com/hankache/perl6intro
lizmat so perhaps it should just fail on a negative position
hankache perl6intro.com 15:20
bartolin I see. I'll open a RT for this (probably a bit later).
thanks for looking!
hankache i would be more than happy if anyone/someone/everyone take a look at it and give me their feedback
PS: This is the first time I write anything 15:21
and the first time I use github :)
15:22 lab left 15:23 kjs_ left
dalek kudo/nom: 72110bd | (Steve Mynott)++ | tools/CREDITS.pl6:
import of CREDITS.p6 script output
 15:23
kudo/nom: 0128437 | (Steve Mynott)++ | CREDITS:
start all records with N: for ease of parsing
kudo/nom: 66d46d4 | (Steve Mynott)++ | CREDITS:
import of result of CREDITS.p6 run
kudo/nom: 12b6564 | (Steve Mynott)++ | CREDITS:
some manual deduplication
kudo/nom: b235717 | (Steve Mynott)++ | CREDITS:
more dedup vendethiel++
kudo/nom: a8812c0 | lizmat++ | / (2 files):
Merge pull request #595 from stmuk/nom

Import a new CREDITS based on a new script CREDITS.p6		 15:24
15:24 lab joined, Zoffix joined
Skarsnik hankache, damn nice work 15:25
hankache thank you Skarsnik. this is draft 1 :) 15:26
15:27 Begi11112 joined
hankache please do not hesitate to send me your feedback if any (corrections/additions) 15:27
Zoffix hankache, looks very nice :) I've bookmarked and will read it in full soon 15:28
hankache Thank you Zoffix. This is my personal notes evolved.
15:28 xfix joined
hankache Zoffix my main motive is to facilitate the lives of newcomers to Perl 6. Something like an entry point 15:29
timotimo did you read learnperl6inyminutes yet? 15:30
15:30 skids joined
hankache timotimo yes 15:30
15:30 lab left
hankache maybe if my doc is good enough we can put a link for it in perl6.org/documentation/ 15:31
timotimo :)
surely can
hankache I am counting on your feedback though
This is the first time i write
zengargoyle hankache: looks nice. 15:33
hankache thank you zengargoyle
timotimo i'm juggling like three different things right now %) 15:34
zengargoyle how picky are we going to be re: use vs need distinction 15:35
[Tux] test 50000 21.301 21.188
test-t 50000 17.995 17.882
csv-parser 50000 28.538 28.426
FWIW I was at the universaty Utrecht today, and motivated the head teacher to have a look at perl6 (after he was very negative about perl5) \o/
diakopter where can I find a reference for what those numbers mean 15:36
[Tux] all the reasons he mentioned for not liking perl5 do not count for perl6 and native concurrency and laziness were quite a motivation for him to bite
tux.nl/Talks/CSV6/speed4.html
diakopter, look at those graphs 15:37
diakopter ok, they don't explain those numbers
there are 6 numbers there and 4 graphs
all four of those graphs say test-t
timotimo Juerd: your mqtt foo.pl script doesn't work on my machine; the connection never goes through ... do i have to do something to make it actually do something?
[Tux] just look at the middle number column. 50000 is a check (the number of fields parsed) 15:38
diakopter okay, but all those graphs say test-t
[Tux] test is my reference script (no features whatsoever), test-t is full-features Text::CSV, CSV::Parse is tony-o's module
yep
it is the performance of perl6 on Text::CSV over time 15:39
15:39 lab joined
timotimo Juerd: wireshark just shows a SYN and multiple retransmissions ... ?!? 15:39
[Tux] if that 17.995 I just posted was say 35, there would have been several people very unhappy
diakopter okay, but what differentiates the 4 graphs 15:40
timotimo i think just scale
both horizontally and vertically
15:44 Zoffix left
diakopter [Tux]: I don't understand what's different about those 4 graphs? 15:46
[Tux] they zoom in on the vertical axis
horizontal axis is the same for all
diakopter ok, thanks
timotimo is there some environment variable or something that'd prevent things from trying to ipv6? 15:53
i'll probably just hard-code the ipv4
now i receive messages 15:54
dalek kudo/nom: fa3ce11 | lizmat++ | src/core/Cool.pm:
Disallow positions < 0 on index/rindex, bartolin++
 15:55
zengargoyle timotimo: depends on system, but basically: no. on linux you can tweak some files and set some preferences, all in all it's still a bit PITA. 15:56
15:56 gtodd joined
lizmat commute to MIA& 15:56
15:56 lizmat left
zengargoyle IPv6 that is. 15:56
many programs have taken to adding a -4/-6 option to force things. 15:57
gtodd sometimes I think it would be fun to be able to load/use a module that lets me substitute math and symbolic logic characters for bits perl6 builtins :-P (unicode ∀ -> for etc.) 15:59
is there such a thing?
hankache see you later everyone 16:01
DrForr There can be, just write a bunch of 'multi prefix' operators.
zengargoyle gtodd: i think somebody is working on one. things like $x² for $x**2 etc. 16:02
skids will likely be working on github.com/skids/perl6-Proc-Screen over turkey days, anyone with an interactive terminal module they need to test may want to star it.
timotimo where the F does this script spend its time?!
flussence gtodd: I tried writing one for setops on arrays years ago, but then we got real Set objects with their own infix ops :) 16:03
timotimo mqtt's foo.pl only spends 5% cpu time :| 16:04
16:04 hankache left
flussence would be nice to have unambiguous ops for some existing things though... 16:04
.u 2212
yoleaux U+2212 MINUS SIGN [Sm] (−)
flussence .u ×÷
yoleaux U+00D7 MULTIPLICATION SIGN [Sm] (×)
U+00F7 DIVISION SIGN [Sm] (÷)
gtodd hee cool 16:05
flussence (can I combine those on one line?)
.u 2212 ×÷
yoleaux U+0020 SPACE [Zs] ( )
zengargoyle m: say 5 − 3;
yoleaux U+0031 DIGIT ONE [Nd] (1)
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/S0vfVloved␤Confused␤at /tmp/S0vfVloved:1␤------> 3say 57⏏5 − 3;␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement modifier␤ …»
yoleaux U+0032 DIGIT TWO [Nd] (2)
flussence guess not.
16:05 molaf left
zengargoyle i think there are some unicode ops, probably just ones TT thought of at the time. 16:06
m: my &safelog = &log ∘　&abs; say safelog -e; 16:07
camelia rakudo-moar : OUTPUT«1␤»
timotimo flussence: why GNU screen and not tmux? 16:08
zengargoyle that's U+2218 \x[RING OPERATOR]
timotimo er 16:09
flussence timotimo: ww?
timotimo skids: why screen and not tmux
flussence oh, wrong tab :)
timotimo sorry :)
flussence np
16:11 kjs_ joined, colomon left
zengargoyle the usual argument for screen is that it's more likely to already be installed. 16:12
16:12 skids left, CIAvash joined
timotimo pff. 16:12
so just stop working on perl6 then? :) 16:13
zengargoyle it's been around and in use fo much longer than tmux.
Skarsnik what tmux add to screen?
zengargoyle lol
Skarsnik *have over
timotimo working vertical splits
flussence I use tmux, but screen's useful for connecting to serial consoles
zengargoyle arguing yesterday to use XDG paths for module installation...
gtodd better name :-)
pink_mist never uses splits anyway 16:14
timotimo oooooh
flussence zengargoyle: I'd like that a lot, but I'm keeping my mouth shut until module installation *works* again...
timotimo i didn't know about synchronize-panes
zengargoyle think screen does xyz modem also. useful for flashing devices.
gtodd does screen have "zooming" ? C-b z 16:15
zengargoyle i use screen for a few things... generally see no problem with multiple xterms.... why split/tab a termnal?
gtodd :-D 16:16
when you really trick out tmux + zsh + vim + emacs it starts to look like ... errm a bunch of Xterms :-D 16:17
Skarsnik I only use screen on my dedicated servers x) 16:18
16:18 colomon joined
zengargoyle i use it for mutt 'cause my mailbox is so big it takes a *long* time to start. and i can grab the same screen from home/work 16:19
zengargoyle still needs to read the man page and figure out how to fix ESC key handling so vim/nvim isn't painful to use. 16:21
16:21 zed_ joined
gtodd screen/tmux ++ .... you can run important applications inside separate console windows so "when root logs out they keep running" :-D 16:23
haha
16:25 flaviusb left, zacts joined 16:26 tokuhiro_ joined
zengargoyle wonders if proc-screen could be replaced with an Expect module... 16:27
which will eventually be really handy to have...
16:27 zed_ left 16:28 z8_ joined
Skarsnik github.com/perl6/perl6-most-wanted...indings.md Can I add at the start something "A tool to generate part of the binding code"? 16:29
zengargoyle though the whole TTY thing is another *NIX thing who's time should have passed by now.
Juerd timotimo: I have no idea why it wouldn't work... It should work on ipv4 and ipv6, and I've tried it on several systems. 16:30
16:31 tokuhiro_ left
zengargoyle usually the problem is DNS lookup over IPv4 can return AAAA IPv6 records which the client will then try to use even if IPv6 isn't working on the client side. 16:31
16:34 colomon left
timotimo ipv6 was broken somewhere on the way 16:34
zengargoyle problematic if IPv6 works on the local network via auto configuration but is not routed.
you can tweak /etc/gai.conf on linux to prefer v4 addresses. 16:35
pink_mist 0_o I don't think I have an /etc/gai.conf 16:37
on any of my linux systems
16:37 colomon joined
zengargoyle try man getaddrinfo and see if it's mentioned under FILES. your distro might have moved it. 16:38
gtodd camelia: my %thing.push: <q w e r t y>; say "it us there" if "q" ~~ %thing ; 16:39
oops 16:40
timotimo gtodd: don't forget "push" is for pushing a single thing; we have .append, too now, which appends a list to the end of another list 16:43
zengargoyle: the thing is, my ISP gives me v6 16:44
autarch if I have a method with a signature like this - sub foo (Str $thing, Str %attrs) { ... } - how do I pass the second param like I would in Perl 5 with this - foo( 'thing', { hash => 'ref' } ) 16:47
basically, how do I construct an anonymous hash?
timotimo i don't really know how to debug my ipv6 stuff
16:47 Zoffix joined
timotimo you'll need Hash[Str].new(...) for that i believe 16:48
autarch m: sub foo (Str $thing, Str %attrs) { dd $thing; dd %attrs; } foo( 'thing', { 'I have' => 'no clue' } )
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/rwGzttfBqR␤Strange text after block (missing semicolon or comma?)␤at /tmp/rwGzttfBqR:1␤------> 3g, Str %attrs) { dd $thing; dd %attrs; }7⏏5 foo( 'thing', { 'I have' => 'no clue' }␤ expecting any of:…»
zengargoyle timotimo: do you know the v6 address of the other side and have ping6 / traceroute6 ?
timotimo yes and yes
zengargoyle m: sub foo (Str $thing, Str %attrs) { dd $thing; dd %attrs; } foo( 'thing', %('I have' => 'no clue') )
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/k6MydD9BVb␤Strange text after block (missing semicolon or comma?)␤at /tmp/k6MydD9BVb:1␤------> 3g, Str %attrs) { dd $thing; dd %attrs; }7⏏5 foo( 'thing', %('I have' => 'no clue') ␤ expecting any of:…»
autarch Type check failed in binding %attr; expected Associative[Str] but got Hash
m: sub foo (Str $thing, Str %attrs) { dd $thing; dd %attrs; } foo( 'thing', { 'I have' => 'no clue' } );
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/fozvtJdDHs␤Strange text after block (missing semicolon or comma?)␤at /tmp/fozvtJdDHs:1␤------> 3g, Str %attrs) { dd $thing; dd %attrs; }7⏏5 foo( 'thing', { 'I have' => 'no clue' }␤ expecting any of:…»
autarch m: sub foo (Str $thing, Str %attrs) { dd $thing; dd %attrs; }; foo( 'thing', { 'I have' => 'no clue' } ); 16:49
camelia rakudo-moar : OUTPUT«Type check failed in binding %attrs; expected Associative[Str] but got Hash␤ in sub foo at /tmp/qg3QjLgO8m:1␤ in block <unit> at /tmp/qg3QjLgO8m:1␤␤»
timotimo autarch: yeah, you're asking for a Hash[Str] to be passed, but you pass just a Hash
autarch: the pain point here is that you defined "Str %attrs"
autarch I see
timotimo zengargoyle: 6 * 100ge7-2.core1.fra1.he.net (2001:7f8::1b1b:0:1) 23.750 ms 23.733 ms
that's the last one i see
autarch m: sub foo (Str $thing, %attrs) { dd $thing; dd %attrs; }; foo( 'thing', { 'I have' => 'no clue' } );
camelia rakudo-moar : OUTPUT«Str $var = "thing"␤Hash $var = {"I have" => "no clue"}␤»
jnthn .tell lizmat (mixin if no chomp) maybe, but otoh most folks use lines, I think...so maybe not worth the complexity.
yoleaux jnthn: I'll pass your message to lizmat.
timotimo jnthn: i didn't get many tests written as i got massively sidetracked %) 16:50
zengargoyle timotimo: he, wonder if Hurricane Electric.... what's the address of the place you're trying to get to?
zacts yo, lot's of chats this AM 16:51
timotimo test.mosquitto.org (2001:ba8:1f1:f271::2)
and yeah, that's hurricane electric
16:52 Zoffix left, rangerprice joined
zengargoyle timotimo: can you curl -6 test.mosquitto.org and get their web page? 16:56
timotimo doesn't look like it
zengargoyle my tracerouts stop at urquell.bitfolk.com and i can get to the web page. 16:57
16:57 CIAvash left
zengargoyle via jump.net.uk on one path, and probably jump on the other (they don't name their interfaces well...) 16:58
flussence I can't get there from my desktop either, but I can ssh -6 to my remote server and get it from there just fine. I guess someone in the middle screwed up.
zengargoyle i go through he on one path and att + globalcrossing the other. 16:59
timotimo beh.
a friend of mine could reach my machine via v6
geekosaur can ping6 it from irc host (a linode in .tx.us) 17:00
timotimo so ... basically a netsplit 17:02
17:03 z8_ left
zengargoyle around 2001:7f8: land. 2001:7f8:4::22ef:1 is hop before bitfolk.com and 2001:7f8:: is hop where you stop. 17:05
timotimo hop where you stop 17:06
tee hee
17:06 johan left 17:08 zakharyas joined
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: d87f592 | (Zoffix Znet)++ | web/ (4 files):
Added DbBuilder::Log and its docs/tests
 17:15
zengargoyle between fra1.he.net (france?) and London Internet Exchange (LINX) if you wan't to call them. :P
timotimo jnthn: would X::HyperOp::NonDWIM be the right exception type for arrays with different shapes being >>+<<'d? if so, the type will ikely have to learn to speak of shapes properly. 17:16
17:18 rangerprice left
timotimo alternatively, a NonDWIM::Shaped could be built 17:19
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: d9f62a2 | (Zoffix Znet)++ | web/t/02-app/0 (2 files):
Git is too stupid to figure out renamed files are not new on its own
17:20 Begi11113 joined, khw joined 17:21 Begi11114 joined 17:22 Begi11112 left
nine I wonder where panda will install its bin files to in the new CompUnit::Repository world 17:23
17:23 lab left, lab joined 17:24 Begi11113 left 17:26 SunilJoshi joined, thowe joined 17:27 SunilJoshi left 17:28 lucasb joined 17:31 zakharyas left
hoelzro o/ #perl6 17:32
DrForr \o @oelzro 17:34
DrForr glares at keyboard. 17:35
17:36 zacts left
hahainternet 'oelzro' is how hoelzro's name is said in the north 17:38
yoleaux 00:40Z <psch> hahainternet: gist.github.com/peschwa/7202b9f622de517cd881 is what i hacked around a bit with re: superscript. the only things that's missing is proper integration, afaict. as in, the nqp internals complain that the lhs isn't there... 17:39
00:41Z <psch> hahainternet: line 27 is just a debugging aid...
hahainternet thanks yoleaux :)
hoelzro o/ DrForr, hahainternet
17:44 eliasr left
jnthn timotimo: Could work well as a subclass of the existing X::HyperOp::NonDWIM 17:44
17:45 geekosaur left
zengargoyle timotimo: funny, can ping 2001:ba8:1f1:f271::2 from he.net routers in france. :) 17:45
zengargoyle sorta bored.
17:45 ugexe joined
lucasb psch: ^^ about that gist, potentiation is another name for exponentiation? :) 17:46
hahainternet zengargoyle: what's so shocking about that? 17:47
17:47 eliasr joined
zengargoyle ah, frankfurt... 17:47
core1.fra1.he.net> ping ipv6 2001:ba8:1f1:f271::2 numeric count 5: 100% 5/5 22ms 36ms 44ms
hahainternet frankfurt? more like london :) 17:48
17:49 colomon left
zengargoyle going by last point on his traceroute: core1.fra1.he.net. that device can reach test.mosquitto.org 17:50
DrForr would mutter something about tetration but even the smallest powers quickly run out of room in the observable universe.
hahainternet oh i guess i missed scrollback :) 17:52
17:52 geekosaur joined
zengargoyle that narrows it down to London not being able to determine the route back to his IPv6 address. 17:54
Begi11114 I've a list of different values. Is there a module to trace a graph ?
17:57 tudorconstantin joined
DrForr waves at tudorconstantin. 17:57
ugexe rosettacode.org/wiki/Topological_sort#Perl_6 17:58
diakopter nine: well, if you make the lib/include system intelligent enough, they can be installed any old place, so that they can be symlinked to some bin dir on the path if necessary, and invoked from anywhere (and they can find their real home themselves)
tudorconstantin waves DrForr 17:59
17:59 tudorconstantin left 18:00 tudorconstantin joined
psch lucasb: uh, actually probably not... i think it was just a case of "i don't remember the english word, let's anglofy the german one"... :) 18:00
timotimo jnthn: if i throws-like X::HyperOp::NonDWIM, that'll refuse a subtype, won't it? 18:01
psch m: use Test; throws-like '0 + "f"', Exception 18:02
camelia rakudo-moar : OUTPUT« 1..2␤ ok 1 - '0 + "f"' died␤ ok 2 - right exception type (Exception)␤ok 1 - did we throws-like Exception?␤»
timotimo ah, neat
so i can write that test right away, eh?
18:02 colomon joined
dj_goku who would I contact about testers.perl6.org issues? 18:04
timotimo depending on whether it's a code issue or a server issue: FROGGS or any of the admins listed on p6c.org 18:06
dj_goku timotimo: ok, I am not sure what it is. re-building moar right now. No new reports since 2015-11-16.
18:07 raiph left
ugexe well, before that it was just marking *everything* as pass 18:07
bartolin psch: I tried to look at one of those UnwindExceptions we have on JVM. Does, perchance, the following ring a bell with you?
j: my $a := gather do for 1 { take 4; last; }; $a.cache.elems; say $a.gist
camelia rakudo-jvm fa3ce1: OUTPUT«(4)␤»
bartolin j: my $a := gather do for 1 { take 4; last; }; $a.cache; say $a.gist
camelia ( no output )
bartolin ^^ that was one of those UnwindExceptions 18:08
j: my $a := gather do for 1 { take 4; last; }; say $a.perl
ugexe i have bartolin
camelia rakudo-jvm fa3ce1: OUTPUT«(4,).Seq␤»
ugexe and camelia has always mocked me the same way when trying to show it here heh
bartolin \o/
dj_goku timotimo: I forgot it might be an issue with Panda. Going to jump in and try to diagnose it.
ugexe it seems if you can get the first run of some types of code it will actually show the unwind exception. after that, it will show nothing
bartolin j: my $a := gather do for 1 { take 4; last; }; say $a.gist # also UnwindException 18:09
camelia ( no output )
psch bartolin: gather/last as a combination is definitely broken, yes
bartolin: i think it's related to CX in general, we have a few hiccups there on r-j
j: O: for ^2 { for ^2 { last O } } # like this one 18:10
camelia rakudo-jvm fa3ce1: OUTPUT«===SORRY!===␤control exception without handler␤»
psch bartolin: i have a hunch that the unwinder has a similar problem as that bit, which is about not finding the right handler
18:10 colomon left
bartolin yeah, we have this pre-GLR ticket about gather/last: RT #124279 18:10
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=124279
bartolin I wondered, how adding that .elem avoided the UnwindException. but probably it does not help to pin the problem :-/ 18:11
dj_goku any idea how I can debug with perl6-debug-m or something else because I get this: Unhandled exception: ctxlexpad needs an MVMContext
nine dj_goku: you know about --ll-exception? 18:12
dj_goku nine: I do not.
nine dj_goku: perl6 --help 18:13
psch also, camelia actually should show UnwindExceptions since nqp 19a7c2e0d...
bartolin m: my $a := gather { take 4; last; }; say $a.gist # something I stumbled above -- looks like an LTA error to me
camelia rakudo-moar : OUTPUT«===SORRY!===␤Can only resume an exception object␤»
nine dj_goku: --ll-exception may point at where exactly in the core things go wrong
psch j: my $a := gather { take 4; last; }; say $a.gist
camelia ( no output )
psch heh, also unwinds... 18:14
bartolin nods
dj_goku nine: ok cool, it is just a wall of text to me: gist.github.com/djgoku/4b4d45062ba39043b6dc
18:15 colomon joined
psch bartolin: fwiw, the "control exception without handler" above happens because r-j doesn't seem to automatically register a handler for a labeled block, while moar somehow does 18:15
i *think* that's probably related to gather/take+last, because take is a CX and last is a CX 18:16
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 1b637f1 | (Zoffix Znet)++ | web/ (5 files):
Add Build ID to ::Dists model

This will allow us to rebuild modules database right into the live database and keep in RAM info for just one dist at a time instead of (potentially) thousands... MILLIONS!... of dists.		 18:17
dj_goku I had to resort to adding print statements to debug a module last week in Perl6. :(
18:18 colomon left
dj_goku oh wait, panda is perl5?! 18:19
ugexe no
dj_goku uh
timotimo rakudobrew is perl5 18:20
psch m: O: for ^2 { for ^2 { last O }; CONTROL { default { .perl.say } } } # this is also a bit interesting...
camelia rakudo-moar : OUTPUT«chars requires a concrete string, but got null␤ in block at /tmp/pcxUUi7pj9:1␤ in block at /tmp/pcxUUi7pj9:1␤ in block <unit> at /tmp/pcxUUi7pj9:1␤␤»
dj_goku well this is confusing. which panda shows ~/.rakudobrew/bin/panda is a p5 script 18:21
ugexe thats just a shim
dj_goku oh
lucasb yes, I think rakudobrew links every executable to itself, a perl5 script
ugexe it invokes panda depending on whatever directory rakudobrew is pointed at
dj_goku so am I doing this right: PANDA_SUBMIT_TESTREPORTS=1 perl6-debug-m --ll-exception ~/.rakudobrew/bin/panda install IDNA::Punycode 18:22
gfldex timotimo: i'm awake now and do not struggle with binding anymore
ugexe dj_goku: adding a `use trace;` might help point you to the correct line as well 18:23
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 93db555 | (Zoffix Znet)++ | web/ (2 files):
Update instead of croaking when adding existing dist
18:24 rindolf left
timotimo gfldex: ah, so you're no longer surprised by what proxies do and why things seemed so strange? 18:24
gfldex timotimo: indeed. However, I do believe now that the docu sucks :) 18:25
bartolin psch: that stuff is way over my head, unfortunately. I hoped to get some insights by golfing the bugs, but it didn't help much ...
leont Right. I still need to finish my PR to make panda not use prove, it's still indirectly using p5.
timotimo ah. well, docs are improvable :) 18:26
dj_goku ugexe: so is that a p5 or p6 use?
ugexe: searched cpanm didn't find trace.
ugexe dj_goku: `use trace;` is p6. your error mentions MVM context so i was guessing your problem was in p6 land
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 06592a8 | (Zoffix Znet)++ | web/lib/ModulesPerl6/Model/Dists/Schema/Result/BuildId.pm:
Fix incorrect ModulesPerl6::Model::Dists::Schema::Result::BuildId docs
 18:27
psch bartolin: well, stepping through org.perl6.nqp.runtime.ExceptionHandling.handlerDynamic showed me (a few weeks back) that we do have a handler for the "last OUTER", but it's always the one we're already in
dj_goku ugexe: ahh well the script is panda that I am trying to debug why reports are getting through
psch bartolin: which is what makes me think there should be another one...
dj_goku going through*
psch bartolin: but yes, in general this stuff seems pretty complicated and i don't quite get it either... :)
18:27 tokuhiro_ joined
ugexe dj_goku: probably related to newlines. but your error is certainly in the perl6 panda script that the perl5 panda shim points to 18:28
dj_goku lol
dj_goku wonders why it isn't all just perl6 18:29
ugexe because the perl5 is not panda, its generated by rakudobrew. rakudobrew is perl5 because you would otherwise need perl6 installed to install rakudobrew...
dj_goku ahh. ok that makes sense. 18:30
18:30 diana_olhovik_ joined
ugexe it used to just generate 1 line bash shims 18:31
18:32 rindolf joined, tokuhiro_ left 18:33 raiph joined
dj_goku ugexe: ok so added `use trace;` to the p6 panda, but no luck any getting more output. 18:34
ugexe the `use trace;` has to be in the correct scope, so either added to all modules or where you suspect it might be happening. So i would put it in Panda::Reporter 18:35
s/modules/submodules/
hoelzro .seen cygx
dj_goku ahh I was just wondering that too.
yoleaux I saw cygx 20 Nov 2015 18:53Z in #perl6: <cygx> bye again o/
dj_goku ugexe: I put it in moar-nom/panda/bin/panda 18:36
ugexe while you are inside Panda::Reporter, you might as well print debug output to show the raw http request and response to determine if its panda or testers.perl6.org 18:37
18:37 just_thanks joined 18:39 colomon joined
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: ffb04ac | (Zoffix Znet)++ | / (4 files):
Initial prototype of DbBuilder class and its supporting script
 18:40
ast: 6936a23 | usev6++ | S04-statements/for.t:
Fudge test not throwing X::Multi::NoMatch on JVM
 18:41
18:41 diana_olhovik_ left
just_thanks ilogger2: 18:42
ilogger2 just_thanks: I am a logger bot. Lines starting with [off] won't be logged. Extra help available at colabti.org/irclogger/irclogger_logs.
just_thanks dalek: 18:43
yoleaux:
hankache - great work on this your perl6Intro, I'd like how succinct it is at least for me( non-native English speaker) 18:45
18:45 just_thanks left
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 6c8bd64 | (Zoffix Znet)++ | web/lib/DbBuilder.pm:
Improve log messages when fetching META.list
 18:46
18:51 yqt joined
dj_goku ugexe: yay: 500 error is being thrown. 18:52
dalek ast: 71cbee5 | usev6++ | S04-declarations/will.t:
Fudge tests for 'will first' on JVM
 18:53
18:55 zengargoyle left
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 4ec675f | (Zoffix Znet)++ | web/lib/DbBuilder.pm:
Do not use line separator var
 18:56
18:56 zengargoyle joined
dj_goku FROGGS: I am not sure the dancer app is accepting reports. Still trying to figure out more. 18:57
18:59 hankache joined 19:00 kjs_ left
dalek ast: 3dbdc11 | usev6++ | S05-modifier/counted.t:
Fudge test for match with *-1 index on JVM
 19:00
kudo/curli: 9104866 | (Stefan Seifert)++ | src/core/CompUnit/Repository/Installation.pm:
Have Repository::Installation's id change when installing a module

The repository's identity needs to depend on its content so installation of a module into this repo can automatically invalidate precomp files of repos nearer the head of the repo chain. Those precomp ids contain the ids of all repos in the chain.		 19:03
dj_goku ugexe: thanks btw!
dalek ast: 3e28ff4 | usev6++ | S16-io/supply.t:
Fudge test for .Supply(:size(1)) on JVM
 19:05
ugexe dj_goku: sure. fwiw the test server has choked like this a few other times and FROGGS fixed it 19:11
FROGGS ahh 19:13
the P5 app on www does not like reboots on hack
it looses connection to the database
jnthn MadcapJake: Nice job on the Atom highlighter! src/Perl6/Grammar.nqp is quite a stress test for the grammars rendering. For some reason token quibble ends up totally unhighlighted, and then sometime before token longname it gets confused and thinks it's still in a string literal; seems it misses that \} is an escaped } a few lines before. 19:17
hoelzro Linenoise now works on Windows! \o/ 19:18
jnthn MadcapJake: Only other thing I immediately note is that the yellow for twigils feels a little "loud" for me. 19:19
MadcapJake: Oh, and I've love it if the supply/react were highlighted, like the gather on gather blocks are. And emit took. 19:21
*too
FROGGS dj_goku / ugexe: should work now again and should probably not fail anymore in future 19:22
jnthn MadcapJake: Examples in S17-supply/syntax.t.
vendethiel o/
jnthn: on the other hand, you might want the twigil to be loud? I've always found e.g. "!$!a" to be hard on the eyes 19:23
19:23 raiph left
FROGGS hoelzro: \o/ 19:24
hoelzro: what was it?
ilmari always misparses $!foo as !$foo if he's not careful
s/always/often/
jnthn vendethiel: I think it's fine to pick it out, but it's a bit too bright, imho
hoelzro FROGGS: amazingly, I didn't run into any of the linker issues we saw at SPW
FROGGS O.o
hmmm
hoelzro it was mostly a matter of not using fcntl on Windows, and using the right libc DLL for the environment 19:25
FROGGS maybe another build patch solved it
hoelzro yeah, I don't know
hopefully I got the DLL logic right; it works for Mingw32 19:26
FROGGS I can test it on MSVC later 19:27
hoelzro that would be much appreciated =)
dj_goku FROGGS: hmm still not working: gist.github.com/djgoku/85ead2d6f8e5c2107745
hoelzro if no one is opposed, I would like Linenoise to be included in Star: github.com/rakudo/star/pull/54 19:28
19:29 tokuhiro_ joined
ilmari Linenoise has no utf-8 support, though :( 19:30
hoelzro should .dll files for modules be installed to a bin directory instead of a lib one? or is it up to our Windows setup to make sure that the lib directory is in PATH?
ilmari: unfortunately not; I should probably work on Readline integration
but that brings certain questions about licensing up
jnthn hoelzro: Please check that we're clean with licensing stuff in the MSIs if you ship the readline DLL
dalek ast: 9817537 | usev6++ | S16-io/comb.t:
Fudge test for .comb with large buffer on JVM
jnthn hoelzro: Other than that, no objections
hoelzro jnthn: for the readline integration, I would only add "if Readline is available, use it" to Perl6::Compiler 19:31
as far as linenoise is concerned, it's MIT/X, I think, so it should be safe
DrForr Which means that I need to take some time to fill out the test sute...
jdv79 jnthn: you're from yorkshire? 19:32
jnthn jdv79: No, but grew up there 19:33
19:33 tokuhiro_ left
jdv79 that sounds like from to me 19:34
what do you think of ilkley brewing?
19:36 ifim joined, kjs_ joined 19:38 hankache left 19:42 kjs_ left
FROGGS dj_goku: I think it works now 19:43
silly typo -.-
jnthn jdv79: Ah, that's west Yorkshire; I was in North Yorkshire :)
jdv79: I've certainly had a couple of those, though...don't have any strong memories 19:44
19:44 vmbrasseur joined
FROGGS which might be a good sign? :o) 19:44
19:44 domidumont joined
jnthn ;-) 19:44
vmbrasseur Hey, gang. I just heard from the OSCON community manager that there are almost no Perl talks in the 2016 OSCON CFP right now and NO Perl6. 19:45
19:45 vendethiel- joined
vmbrasseur The CFP closes on Tuesday at midnight Pacific time. 19:45
jnthn My local pub did Black Sheep. :)
dj_goku FROGGS: HTTP/1.1 200 OK +++++++
FROGGS \o/
vmbrasseur Could we, maybe, get some talks in? Please?
FROGGS vmbrasseur: did you talk with woolfy? she organizes the CFP 19:46
jnthn FROGGS: That's...FOSDEM, no? :) 19:47
19:47 vendethiel left
vmbrasseur FROGGS Yeah, that's not OSCON. 19:47
FROGGS ohh, ups
vmbrasseur :)
FROGGS all those SHOUTS :o)
vmbrasseur We have a Perl6 release coming up. It's a big deal. We need to talk about it.
jdv79 i had no idea yorkshire was so big. nice
vmbrasseur I'm currently not qualified to talk about Perl6.
But y'all are. :)
DrForr If I could get reimbursed I'd do it. But work won't pay for YAPC::NA, so I doubt OSCON is even on the books.
grondilu m: sub foo (Str $thing, Str %attrs) { dd $thing; dd %attrs }; foo ( 'stuff', my % = why => "not this?"); 19:48
camelia rakudo-moar : OUTPUT«Too few positionals passed; expected 2 arguments but got 1␤ in sub foo at /tmp/ZyIyKQ2hbA:1␤ in block <unit> at /tmp/ZyIyKQ2hbA:1␤␤»
vmbrasseur DrForr OSCON provides speaker assistance for international speakers.
DrForr Oo. Hrm.
dj_goku FROGGS: thanks again. I didn't know if it was a problem I was causing. But I saw the issue on 11/17 just never got back to trying to figure it out until today
grondilu oops I did not scroll
DrForr When does the CFP close?
vmbrasseur DrForr Tuesday at midnight Pacific time.
DrForr I've got an intro to perl6 regex that I just did at Cluj. 19:49
FROGGS vmbrasseur: OSCON is out of reach for me :S
vmbrasseur DrForr That would be amazing.
FROGGS OSCON will give you financial assistance as a speaker.
19:49 domidumont left
vmbrasseur This is a big event with a lot of attendance. It would be a shame to miss the opportunity to get Perl6 in front of so many people so soon after its release. 19:50
DrForr Ya talked me into it :)
vmbrasseur If we could get at least 2 Perl6 talks accepted for OSCON it would be AMAZING
FROGGS vmbrasseur: there are better speakers, and my wife is not so fond of me travelling around :/
vmbrasseur DrForr <3!
FROGGS I can't help with the travel thing, but ANY speaker is better than the NO speaker we have right now.
DrForr Besides, having "Speaker at OSCON" on the ol' CV doesn't hurt.
vmbrasseur DrForr True dat ;)
If nothing else, FROGGS, could you and others help by talking to others to get THEM to CFP to OSCON? 19:51
FROGGS vmbrasseur: will do
vmbrasseur FROGGS <3 19:52
jdv79 DrForr: how is the hip?
DrForr Right now, I'm convincing myself that spending 880 lei.... fck it, buying.
FROGGS .tell lizmat are you going to OSCON? CFP is closing on Tuesday midnight Pacific time
yoleaux FROGGS: I'll pass your message to lizmat.
DrForr Let me buy this flght first...
I'll be the first to admit that I'm not the world's most dynamic speaker though, I think I saw a field where they ask you to submit URLs to previous YouTubed talks and I sort of shied away... 19:54
dj_goku ahh OSCON is in austin. 19:56
19:56 diana_olhovik joined
jdv79 oscon is too expensive and ive heard its all suits now 19:56
19:57 kjs_ joined
DrForr All the more reason :) 19:57
20:00 s_kilk joined, s_kilk left 20:01 diana_olhovik left, s_kilk joined 20:05 s_kilk left
TimToady .tell hankache re - and ' in names, alpha is only required to the right; to the left can be numeric; think of it as a sequence of identifiers separated by - or ' 20:05
yoleaux TimToady: I'll pass your message to hankache.
20:06 kjs_ left
vmbrasseur jdv79 If you're speaking, you get in for free & O'Reilly can give you financial assistance. 20:06
20:07 kjs_ joined, s_kilk joined
vmbrasseur jdv79 As well, it's not all suits. And if that WERE the case it wouldn't get any better by communities avoiding the conference. ;) 20:07
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 107e84a | (Zoffix Znet)++ | / (5 files):
Initial work on Dist::Source base class and GitHub dist source
 20:10
FROGGS $ perl6 -I. -e 'say (require Bar).new'
Bar.new(a => 42)
\o/
FROGGS spectests 20:11
jnthn
.oO( everyone requirs the bar now and then... )
20:12 bbkr joined 20:14 darutoko left 20:15 loren left
jdv79 vmbrasseur: speaking gets you in for free at a bunch of confs i thought 20:15
vmbrasseur jdv79 Most, but some (Pycon, for instance) have everyone (even organizers) pay to get in. 20:16
jdv79 vmbrasseur: the ticket priceis a pretty string detractor to normal folf^Hk
20:17 st_iron joined
Skarsnik FROGGS, when you have time can you check the patch in RT #126645 work on win32? 20:17
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126645
st_iron good afternoon my friends
psch jnthn: what would be needed to get signature compilation ported to r-j? as in, which bits of (i assume) BOOTSTRAP.nqp do i need to understand? :S
Skarsnik fun bot x)
jdv79 i will personally never pay 1k+ for a conf
psch ('cause there's at least one ticket that happens due to one of those magic deconts...)
jnthn psch: It's actually in Actions.nqp that it happens 20:18
psch: Search for lower_signature 20:19
psch ah, okay
jnthn I think it needs some QAST compiler updates
For the thing that falls back to the slow-path binder on error.
jdv79 its fairly clear that oreilly is trying to milk its conf attendees
FROGGS Skarsnik: will do 20:20
Skarsnik basicly it should do nothing x)
but thx! 20:21
20:22 tokuhiro_ joined
jnthn Updated version of the course slides from the Swiss Perl Workshop, fixing a couple of bugs: jnthn.net/papers/2015-spw-perl6-course.pdf 20:25
dalek ast: 9e99d8c | usev6++ | S04-statements/with.t:
Fudge tests for JVM: Failures not handled as expected

Had to duplicate code in order to fudge for JVM. Please revert if there is a better way to do the fudging.
jnthn (Same URL as before)
Think it got included into R*
So that copy may want updating
.tell lizmat Updated SPW slides, since I think you were going to use 'em: jnthn.net/papers/2015-spw-perl6-course.pdf 20:26
yoleaux jnthn: I'll pass your message to lizmat.
20:27 tokuhiro_ left, lizmat joined
lizmat waves from MIA 20:27
yoleaux 16:49Z <jnthn> lizmat: (mixin if no chomp) maybe, but otoh most folks use lines, I think...so maybe not worth the complexity.
19:52Z <FROGGS> lizmat: are you going to OSCON? CFP is closing on Tuesday midnight Pacific time
20:26Z <jnthn> lizmat: Updated SPW slides, since I think you were going to use 'em: jnthn.net/papers/2015-spw-perl6-course.pdf
jnthn Oh noes, lizmat is Missing In Action! 20:28
FROGGS /o\
lizmat hehe
from a business lounge that reminds me of the worst times at JFK during the late eighties 20:29
aka, packed with people, noisy , smelly, no room to sit
grrr.... :-)
FROGGS eww
lucasb I did a 'diff src/QAST/Stmt{,s}.nqp' in nqp and saw that the difference between QAST::Stmt and QAST::Stmts is just some int type annotation on the 'my int $i' counter. If those int types were added to QAST::Stmts, would it improve something?
dj_goku say (2 ** 99999).chars # 30103 20:31
lol
lizmat ah, woolfy found a place for us to sit!
diakopter MIA is the woooooooooorst 20:34
dj_goku FROGGS: uh, /report isn't accept reports again. :( 20:35
20:38 cdg joined
FROGGS O.o 20:38
20:38 s_kilk left, diana_olhovik_ joined
dj_goku FROGGS: I promise I didn't do it on purpose. :D 20:39
FROGGS dj_goku: works for me: ==> Test report submitted as: testers.perl6.org/reports/85170.html
20:40 rindolf left
dj_goku FROGGS: I got a 404 opening that link 20:40
FROGGS dj_goku: that's okay, it takes up to five minutes to show up 20:41
dj_goku ahh
FROGGS dj_goku: it is there now
dj_goku: but I see 500's for some reports...
dj_goku FROGGS: weird, so JSON::Fast worked, but Readline failed, and is sending a failure report. 20:42
FROGGS there was something wonky in the P5 script which I fixed now again 20:44
20:46 st_iron left 20:47 lab left
dj_goku FROGGS: so it should work now? I just tried to install readline and got a 500 still. 20:47
avar /w 6
FROGGS dj_goku: any the valid you send is valid? 20:49
dj_goku FROGGS: this was worked: ==> Test report submitted as: testers.perl6.org/reports/85171.html (JSON::Fast)
20:49 quietfanatic joined 20:50 cdg left
FROGGS I'm trying readline as well now 20:50
dj_goku FROGGS: that was before you said you fixed it again (14:44)
ok
quietfanatic Hey, forgive me if this has been asked before, because it seems kind of obvious, but 20:53
is there a way to get a grammar to tell me where it failed parsing?
I'm getting kinda tired of seeing "Method 'made' not found for invocant of class 'Any'" 20:54
dalek kudo/nom: 95558bb | FROGGS++ | src/Perl6/Actions.nqp:
make "require PAKAGE/STRING" return said PACKAGE/STRING

Previously it returned True on success which is not that helpful. Now we can use the return value as specified by S11: my $foo = (require Foo).new;
Fixes RT #126100
synbot6 Link: design.perl6.org/S11.html#my
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126100
ast: ff452eb | FROGGS++ | S11-modules/require.t:
add tests for RT #126100
 20:55
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126100
FROGGS quietfanatic: there exists a hack with setting $*HIGH_WATER or some such, also, sometimes it helps to call .subparse for debugging 20:57
quietfanatic Hmm, it just seems like there ought to be a standard way of doing this. 20:59
FROGGS aye
quietfanatic besides attaching "| <fail: blah>" on every failable token
FROGGS though I dunno how to expose that information
lizmat gate& 21:00
21:00 lizmat left
FROGGS maybe TimToady has a good idea on how to expose the highwater? 21:00
21:01 lucasb left
quietfanatic Perhaps. 21:01
FROGGS dj_goku: something in the test output is invalidating the json it seems 21:03
quietfanatic I'll be honest, I'm pretty annoyed. If I have to laboriously intuit every cause of parse failure, it'd be faster to write my own parser in standard C++.
dj_goku FROGGS: oh sweet. /me starts searching
FROGGS dj_goku: either that or it is about the size of the post data 21:04
ugexe FROGGS: it saves the last test report somewhere right? he could run that through a json linter probably
21:04 raiph joined
dj_goku I just testing the JSON in two linters and they both say valid json 21:06
FROGGS ugexe: it is just recording the result
but yeah, I also did that and it says it is valid
21:07 Begi11115 joined 21:10 Begi11114 left
dj_goku FROGGS: is the dancer code on github? 21:11
ugexe theres a fork me ribbon with a link on testers.perl6.org upper right corner anyway 21:12
FROGGS dj_goku: here it is: gist.github.com/FROGGS/cd2f1b94b3bd77609971 21:14
invalid character encountered while parsing JSON string, at character offset 13758 (before "\x{1b} my ( $of...") at...
dj_goku FROGGS: is there a reason this isn't being used: github.com/perl6/cpandatesters.per...pp.pl#L233 21:15
FROGGS so, it does not like colors
dj_goku: that's the very old site
dj_goku oh
FROGGS dj_goku: which delivered dynamic pages and was too lsow
slow*
dj_goku FROGGS: ahh. 21:16
FROGGS: so parts of that is used to generate some pages. 21:17
FROGGS aye
the jobs/
hmmm, and when I disable colors, it chokes on the eject sign
dj_goku when was the last time it was run dynamically with p6?
FROGGS malformed UTF-8 character in JSON string, at character offset 13795 (before "\x{fffd}src-offset )...") at... 21:18
almost exactly a year ago I think
21:18 kjs_ left
dj_goku FROGGS: ahh, a lot has changed since then! :D 21:19
FROGGS I know :o)
jferrero m: my $IP = "10.113.125.8"; say ($IP.split(".").cache.[0^..*].map: { 999 - $_ }).join("").flip 21:21
camelia rakudo-moar : OUTPUT«199478688␤»
dj_goku FROGGS: is there anything I can help with? 21:26
FROGGS dj_goku: no, I'm working at a patch atm 21:27
21:28 lab joined
dj_goku FROGGS: so is the issue unicode? because I see you have a "raw" column that is of type JSON in postgres. 21:30
dalek kudo/curli: ef03668 | (Stefan Seifert)++ | src/core/CompUnit/Repository/Installation.pm:
Fixup for previous id change

The previous commit went to far. While the precomp ids may not depend on the ever changing id of this repo, they still have to depend on the changing ids of the repos down the chain.
FROGGS dj_goku: it is failing when decoding the JSON before we hit the database 21:31
21:31 vendethiel- left
dj_goku FROGGS: ahh fun. :( 21:34
21:34 vendethiel joined
jferrero m: my $IP = "10.113.125.8"; say $IP.split(".").cache.[1..3].map({999 - $_}).join("").flip 21:35
camelia rakudo-moar : OUTPUT«199478688␤»
jferrero m: my @IP = <10.113.125.8 10.96.3.8>; say $IP.split(".").[1..3].map({999-$_}).join("").flip for @IP 21:37
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/2_kdzvuACE␤Variable '$IP' is not declared. Did you mean '@IP'?␤at /tmp/2_kdzvuACE:1␤------> 3my @IP = <10.113.125.8 10.96.3.8>; say 7⏏5$IP.split(".").[1..3].map({999-$_}).join␤»
jferrero m: my @IP = <10.113.125.8 10.96.3.8>; say $_.split(".").[1..3].map({999-$_}).join("").flip for @IP
camelia rakudo-moar : OUTPUT«199478688␤199699309␤»
21:39 diana_olhovik_ left 21:41 TEttinger joined 21:42 black_ant joined 21:50 kaare_ left 21:53 cibs left, cibs joined 21:55 black_ant left 21:57 eliasr left 22:02 diana_olhovik joined, Ven joined 22:04 rindolf joined 22:09 diana_olhovik left
FROGGS leont: is it possible that prove outputs a BEL character? 22:12
22:14 xargs joined 22:16 xfix left
Ven .tell zoffix seems like everything is "fixed" here :) ty 22:18
yoleaux Ven: I'll pass your message to zoffix.
22:18 rindolf left
FROGGS leont: that character seems to be at the beginning of this line: gist.github.com/FROGGS/4e2ce3d1f3f...ing-pl-L12 22:18
22:20 abaugher left 22:21 abaugher joined
nine quietfanatic: I guess you do know about Grammar::Debugger? 22:21
quietfanatic nine: seems like overkill for such a simple feature 22:22
nine quietfanatic: well I'd say the feature is just not that simple :) However Grammar::Debugger actually is 22:23
22:24 tokuhiro_ joined
quietfanatic How is getting the line and character number of where a parse failed not simple? 22:24
22:25 raiph left
nine quietfanatic: Backtracking. A failure for some subrule to match doesn't mean that the whole grammar fails. If there are 3 different changes to the input that would all individually make the parse succeed, which one should it flag? 22:26
quietfanatic Okay, it wouldn't be simple in the general case, I guess
22:26 diana_olhovik_ joined 22:28 tokuhiro_ left
quietfanatic Perhaps Perl 6 grammars are not as suited to parsing as I thought. 22:28
nine FWIW I'd love to see that feature in Perl 6. I guess the people who would be able to come up with a somewhat useful solution are just busy with other things right now. Like getting our 1.0 out of the door :) 22:29
quietfanatic yeah
22:31 Begi11115 left, Ven_ joined 22:32 BenGoldberg joined
Ven_ nine: how are the messages provided using :dba() (and maybe commit?) 22:32
22:33 FROGGS left, Skarsnik left 22:34 Ben_Goldberg joined 22:35 xargs left
Ben_Goldberg m: class Test is Iterator { sub pull-one { return 42 } }; say 'ok'; 22:35
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/UhflGUfaE4␤Method 'pull-one' must be implemented by Iterator because it is required by a role␤at /tmp/UhflGUfaE4:1␤»
22:35 Ven left, stux|RC-only left
Ben_Goldberg What am I doing wrong here? 22:35
Ooops, I see ;)
m: class Test is Iterator { method pull-one { return 42 } }; say 'ok';
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/aU8m4I6LLL␤Method 'pull-one' must be implemented by Iterator because it is required by a role␤at /tmp/aU8m4I6LLL:1␤»
Ben_Goldberg :(
22:36 stux|RC-only joined, diana_olhovik_ left 22:37 frew left, frew joined
nine Ven_: ??? 22:37
22:37 BenGoldberg left
masak nine: ¿¿¿ 22:38
Ven_ nine: whoops, it seems I had a *very* stale backlog
22:38 Zoffix joined
gfldex m: class Test does Iterator { method pull-one { return 42 } }; say 'ok'; 22:39
camelia rakudo-moar : OUTPUT«ok␤»
gfldex Ben_Goldberg: ^^^
Ben_Goldberg Thanks! :) 22:40
Zoffix Fun fact: 382 out of 462 (83%) are missing the required `perl` key from their META file :) gist.github.com/zoffixznet/2c56fa8071ecd6c47085 22:42
yoleaux 22:18Z <Ven> Zoffix: seems like everything is "fixed" here :) ty
Zoffix Ven_, awesome :D
Anyone's up for making 382 PRs? :P
stmuk_ what should the value be? 22:43
Zoffix
.oO( we can probably make a robot that does this stuff )
stmuk_, design.perl6.org/S22.html#perl
"The minimal perl version for which this distribution can be installed. Specified as a version string."
stmuk_ $*PERL.version doesn't resemble that string 22:46
22:47 bpmedley left
Zoffix 1 is missing `description` and 16 are missing `provides`, but I believe all of them already have PRs 22:47
m: say $*PERL.version 22:48
camelia rakudo-moar : OUTPUT«v6.b␤»
Zoffix stmuk_, ?
m: say v6.b before v6.0.1
camelia rakudo-moar : OUTPUT«True␤»
Zoffix m: say v6 before v6.0.1 22:49
camelia rakudo-moar : OUTPUT«True␤»
stmuk_ I thought that would be v6.c not v6.0.0 as shown in the example
Zoffix m: say v6.0.0 before v6.0.1 22:50
camelia rakudo-moar : OUTPUT«True␤»
Zoffix stmuk_, the example says specifying `perl` as v6.0.1 won't allow installation on earlier Perl versions, of which v6.0.0 is one.
22:51 sprocket joined
Zoffix It's a spec. It doesn't care that currently we're planning to release a particular implementation of Perl 6 with version of v6.c 22:51
22:52 Ben_Goldberg left, Ben_Goldberg joined
stmuk_ it's not very clear to anyone reading the spec what to put for a version 22:52
Zoffix A version string doc.perl6.org/type/Version 22:53
22:54 Ben_Goldberg left
Zoffix m: say $*PERL.version.WHAT 22:55
camelia rakudo-moar : OUTPUT«(Version)␤»
22:55 lab left
stmuk_ so what do we put now and on Dec 25? 22:56
22:56 znpy left, BenGoldberg_ joined
Zoffix stmuk_, you can put { ... perl: "v6.b" ... } 22:57
On Dec 25, if your code no longer works with v6.b, after fixing it up for v6.c, you'd put it as perl: "v6.c" 22:58
stmuk_ if I had set v6.b for code that worked with niceville it could have broken in quite a few ways by bent 22:59
23:00 cdg joined 23:02 lab joined 23:03 Ven_ left
dalek href="https://modules.perl6.org/new-db-builder:">modules.perl6.org/new-db-builder: 5f76c8b | (Zoffix Znet)++ | web/ (6 files):
First working Dist Source plugin pipeline with a dummy GitHub plugin
 23:03
stmuk_ maybe I will just put v6.* and hope for the best :) 23:05
Zoffix methinks that's be "v6"
m: say v6.* before v6.c
camelia rakudo-moar : OUTPUT«True␤»
Zoffix Oh.. It works
Perl6++
m: say v6.* after v6.c
camelia rakudo-moar : OUTPUT«False␤»
Zoffix m: say v6 before v6.c
camelia rakudo-moar : OUTPUT«False␤»
BenGoldberg_ m: my @a; push @a[2]: 'test'; say @a; if @a.shift -> $head { say "head is $head" } else { say "headless" } 23:06
camelia rakudo-moar : OUTPUT«[(Any) (Any) [test]]␤headless␤»
BenGoldberg_ m: my @a; push @a[2]: 'test'; say @a; if @a.shift -> $head { say "head is $head" } else { say "headless" }; say @a;
camelia rakudo-moar : OUTPUT«[(Any) (Any) [test]]␤headless␤[(Any) (Any) [test]]␤»
Zoffix Oh
weird
m: say v6 before v6.0.0
camelia rakudo-moar : OUTPUT«False␤»
Zoffix m: say v6 after v6.0.0
camelia rakudo-moar : OUTPUT«False␤»
stmuk_ m: say v6.a after v6.*
camelia rakudo-moar : OUTPUT«True␤»
stmuk_ m: say v6.a ~~ v6.*
camelia rakudo-moar : OUTPUT«True␤»
BenGoldberg_ Could someone tell me what I'm doing wrong? Why does my @a not get turned into [(Any) [test]] ? 23:07
Zoffix m: my @a; push @a[2]: 'test'; say @a; @a.shift; 23:09
camelia rakudo-moar : OUTPUT«[(Any) (Any) [test]]␤Cannot shift from an empty Array␤ in block <unit> at /tmp/fC34N2sl4U:1␤␤Actually thrown at:␤ in block <unit> at /tmp/fC34N2sl4U:1␤␤»
masak um.
...but it has three elements... 23:10
BenGoldberg_ Replacing the push with an assignment produces the same result:
m: my @a; @a[2] ='test'; say @a; shift @a; say @a;
camelia rakudo-moar : OUTPUT«[(Any) (Any) test]␤Cannot shift from an empty Array␤ in block <unit> at /tmp/cKxBG_Mzb0:1␤␤Actually thrown at:␤ in block <unit> at /tmp/cKxBG_Mzb0:1␤␤»
Zoffix m: my @a = 1; push @a[2]: 'test'; say @a.shift 23:11
camelia rakudo-moar : OUTPUT«1␤»
Zoffix m: my @a = Any; say @a.shift
camelia rakudo-moar : OUTPUT«(Any)␤»
Zoffix weird
BenGoldberg_ m: my @a = Any; say @a.shift; say @a.shift;
camelia rakudo-moar : OUTPUT«(Any)␤Cannot shift from an empty Array␤ in block <unit> at /tmp/b204AD5eF3:1␤␤Actually thrown at:␤ in block <unit> at /tmp/b204AD5eF3:1␤␤»
masak gentlemen, I think you have found a bug. 23:12
arnsholt m: my @a = Any, Any, 1; say @a; say @.shift; # What about this?
camelia rakudo-moar : OUTPUT«5===SORRY!5=== Error while compiling /tmp/01OZ1Kry8D␤Variable @.shift used where no 'self' is available␤at /tmp/01OZ1Kry8D:1␤------> 3my @a = Any, Any, 1; say @a; say @.shift7⏏5; # What about this?␤ expecting any of:␤ argument list…»
BenGoldberg_ m: my @a; @a[0] = Any; @a[1] = 42; say @a; shift @a; say @a;
camelia rakudo-moar : OUTPUT«[(Any) 42]␤[42]␤»
arnsholt m: my @a = Any, Any, 1; say @a; say @a.shift; # What about this?
camelia rakudo-moar : OUTPUT«[(Any) (Any) 1]␤(Any)␤»
arnsholt Weeeird.
masak arnsholt: ok, so the Anys produced by making holes are special somehow?
arnsholt Might look like it...
masak m: my @a; @a[1] = 42; @a.shift 23:14
camelia rakudo-moar : OUTPUT«Cannot shift from an empty Array␤ in block <unit> at /tmp/N7HgBgUm0E:1␤␤Actually thrown at:␤ in block <unit> at /tmp/N7HgBgUm0E:1␤␤»
masak m: my @a = Any, 42; @a.shift
camelia ( no output )
BenGoldberg_ m: my @a; @a[1] = 42; splice @a, 0, 1; say @a;
camelia rakudo-moar : OUTPUT«[42]␤»
masak looks like it to me.
Zoffix m: my @a; @a[1] = 42; say @a[0].WHAT, Any.WHAT; say @a; @a.shift 23:15
camelia rakudo-moar : OUTPUT«(Any)(Any)␤[(Any) 42]␤Cannot shift from an empty Array␤ in block <unit> at /tmp/jiN73UKf6P:1␤␤Actually thrown at:␤ in block <unit> at /tmp/jiN73UKf6P:1␤␤»
Zoffix Why is Mu class called that way? 23:17
TimToady m: my @a = 1,2; @a[0]:delete; @a.shift 23:18
camelia rakudo-moar : OUTPUT«Cannot shift from an empty Array␤ in block <unit> at /tmp/0mfbGccsyT:1␤␤Actually thrown at:␤ in block <unit> at /tmp/0mfbGccsyT:1␤␤»
Zoffix It's weird that it doesn't die on this: 23:19
m: my @a = 1,2; @a[0]:delete; say @a.shift ?? @a !! @a
camelia rakudo-moar : OUTPUT«[(Any) 2]␤»
23:20 sprocket left
BenGoldberg_ m: my @a; @a[1] = 42; so shift @a; say @a; 23:20
camelia rakudo-moar : OUTPUT«WARNINGS:␤Useless use of "so " in expression "so shift @a" in sink context (line 1)␤[(Any) 42]␤»
BenGoldberg_ m: my @a; @a[1] = 42; say so shift @a; say @a;
camelia rakudo-moar : OUTPUT«False␤[(Any) 42]␤»
23:21 cdg left
TimToady obviously return a failure that throws lazily 23:21
but the message is a bit off
actually, the shift should work, but return Nil
23:28 telex left 23:30 telex joined
BenGoldberg_ m: gist.github.com/BenGoldberg1/467df...d16fb4af23 23:31
camelia rakudo-moar : OUTPUT«(signal SEGV)ok 1␤»
BenGoldberg_ Ouch. 23:32
Locally, it works without the segv.
TimToady testing a patch fro shift 23:34
BenGoldberg_ j: say 'jakudo?'
camelia rakudo-jvm 95558b: OUTPUT«jakudo?␤»
BenGoldberg_ j: gist.github.com/BenGoldberg1/467df...d16fb4af23
camelia rakudo-jvm 95558b: OUTPUT«ok 1␤(2 3 5 7 11 13 17 19 23 29 31 37 41 43 47 59 71 77 83 89 101 107 113 119 143 149 161 167 191 203)␤ok 2␤»
Zoffix neat 23:35
timotimo moar does what jakudon't
BenGoldberg_ And vice versa?
23:36 BenGoldberg_ is now known as BenGoldberg
metasin primes? 23:36
BenGoldberg Yup.
Zoffix Next step: factor ones that are 200 digits long... ??? PROFIT! 23:37
flussence I'm trying to figure out how exactly ".return" behaves, but there's almost nothing in the obvious places (specs, doc, roast)
BenGoldberg It's not a prime factorizer, it's juse an unbounded list of primes. 23:38
23:39 rurban left
BenGoldberg OTOH, it's a unusual algorithm for the task, very cool :) 23:39
23:40 vendethiel left
timotimo multi-dim array access is significantly slower here than array-of-arrays used to be :S 23:40
23:40 Zoffix left
TimToady yes, that's been true from the start 23:40
hasn't been optimized yet
timotimo mhm 23:41
oh damn 23:42
--profile crashed!
diakopter it does when there's a signal handler
timotimo KnowHOW methods must be called on object instance with REPR KnowHOWREPR
diakopter: that's just because that introduces a second thread, i think 23:43
can't get --ll-exception from that :o
TimToady profiling forest, are we? 23:44
BenGoldberg Profiling my gist, perhaps. 23:45
TimToady doesn't have a single handler that I can see
BenGoldberg Well, no, but it produced a signal!
leont .tell FROGGS I don't think so, though the interactive output uses \r a lot 23:49
yoleaux leont: I'll pass your message to FROGGS.
dalek kudo/nom: 888ef1b | TimToady++ | src/core/Array.pm:
sparse array should not act empty on shift
TimToady BenGoldberg++ for the shift bug 23:54
BenGoldberg You're welcome :)
Hmm, looking at the fix though... just returning Nil isn't going to move the other elements closer to the front, I don't think. 23:56
23:58 sprocket joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!