»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
Skarsnik AlexDaniel, fun, if I remove the wget and replace by slurping a local copy of this page memory consuption (110Mb) is consistent across rakudo version, otherwise I can get crazy stuff like 2017-06 and 2017 taking over 250-400Mb 00:00
*2017-07 00:01
well time to sleep 00:02
00:03 Skarsnik left
AlexDaniel MasterDuke: can you try dd %tree.kv; ? :) 00:06
MasterDuke: nvm 00:07
MasterDuke with what in %tree?
AlexDaniel MasterDuke: it's from the dugword's issue, but nvm I think I'm progressing with golfing 00:08
00:08 Ben_Goldberg joined 00:09 BenGoldberg left, Ben_Goldberg is now known as BenGoldberg
MasterDuke nice 00:09
00:18 itaipu joined
Zoffix "zengargoyl+│ 13:36 < Zoffix> Stop adding features just because you can. There's a word for it: featuritis. -- lol ²² :P" 00:41
zengargoyle: yes, and now we have to support it!
MasterDuke: "have you seen dugword's problem" I see a giant wall of text.. What's the problem or did someone answer already? 00:42
00:44 gdonald left, eyck joined
MasterDuke he has a List with an Any in it that gets turned into a Mu when the List is Slipped 00:46
dugword Zoffix: The flatten "|" operator returns different values than what appears with say .perl or dd 00:47
Zoffix zengargoyle: in fact. The powers were first implemented in a module, but then were rushed into core becausee in the module you couldn't chain more than one char and also custom ops at the time were unusably slow. And the superscript ops do have a problem with precedence that Texas versions don't have. Not to mention the endless grief over `²²` being valid code. If there ever were an example that you can't 00:48
anticipate all the issues, power ops are it. And given that unlike the power ops the `:?foo` thing affects much broader part of the language it'll likely have more severe issues, and given that unlike the power ops the `:?foo` is entirely useless, any issues it'll introduce will have zero justification behind them.
00:48 nadim left
Zoffix dugword: how do I repro it? 00:49
MasterDuke git clone [email@hidden.address] 00:52
dugword Then run perl6 t/00prim.t
Zoffix ok. Then?
SmokeMachine I prefer :?$foo than :foo($foo.Bool) 00:53
dugword The output near the end will have output with Mu
github.com/dugword/perl-Convert-AS....pm6#L1100
Zoffix SmokeMachine: that's not what we're discussing though. It's adding `:?foo` to do the same thing as `:foo`
SmokeMachine Or :foo(?$foo) 00:54
Zoffix dugword: ok. Gonna have a quick dinner first and then take a look
SmokeMachine Isn’t it the same usage? 00:55
dugword I'm trying to golf it down to a more simple example
00:56 gdonald joined
Zoffix SmokeMachine: no, `:foo` is equivalent to `:foo(True)` `:$foo` is equivalent to `:foo($foo)`. The usage you showed sticks an op between `:` and a variable 00:57
00:57 Cabanossi left 00:58 itaipu left
SmokeMachine m: my $foo =42; say :?$foo 01:00
camelia 5===SORRY!5===
Argument to "say" seems to be malformed
at <tmp>:1
------> 3my $foo =42; say7⏏5 :?$foo
Confused
at <tmp>:1
------> 3my $foo =42; say :7⏏5?$foo
expecting any of:
colon pair
Other potential diffi…
01:00 Cabanossi joined
SmokeMachine m: sub a(:$f) {say $f}; my $foo = 42; a :?$foo 01:01
camelia 5===SORRY!5=== Error while compiling <tmp>
Confused
at <tmp>:1
------> 3sub a(:$f) {say $f}; my $foo = 42; a :7⏏5?$foo
expecting any of:
colon pair
SmokeMachine I mean: if :?$foo was possible, :?foo should be possible to (imho)... and I think :?$foo should be possible
ugexe its bad enough the 'required' marker (!) does one thing on the right side of a signature param and something else on the other when give as a param 01:03
e.g. :!foo and :$foo! 01:04
01:05 itaipu joined
Zoffix It's bad enough we have fifty ways to write colon pairs. No need to invent a million more. 01:07
Consdering SmokeMachine was using Perl 6 longer than me and still didn't know :?$foo wasn't one of the valid syntaxes :) 01:09
dugword: so you want to have a Mu but somehow getting an Any? 01:12
SmokeMachine Were I?
MasterDuke Zoffix: other way around 01:13
Zoffix MasterDuke: so the code wants to convert a Mu to an Any? 01:14
MasterDuke: I see the test code with Compiled.new being created by being given one Mu and two Anys
MasterDuke i don't think he wants to convert anything, it's just happening when slipping
dugword No, I have an Any but somehow I'm getting a Mu 01:15
I have a work around by not using |, but I thought it was weird I got one set of values using it and a different set using anything else. Thought it might be a perl6 bug 01:16
Zoffix dugword: but why do you say "somehow"? You got one here github.com/dugword/perl-Convert-AS...rim.t#L121 and here github.com/dugword/perl-Convert-AS...rim.t#L141
dugword Yeah, that's the expected value that the test is testing against. I changed it to Mu so the test would pass. 01:18
Zoffix Ah
m: my @a = <a b c>; @a[1]:delete; dd @a 01:19
camelia Array @a = ["a", Any, "c"]
Zoffix m: my @a = <a b c>; @a[1]:delete; dd [|@a]
camelia ["a", Mu, "c"]
AlexDaniel don't we have a ticket for this ↑? 01:20
BenGoldberg m: my @a = <a b c>; @a[1] = Nil; dd @a; 01:23
camelia Array @a = ["a", Any, "c"]
BenGoldberg m: my @a = <a b c>; @a[1] = slip(); dd @a;
camelia Array @a = ["a", Empty, "c"]
Zoffix m: use nqp; my @a = <a b c>; nqp::bindpos(nqp::getattr(nqp::decont(@a), List, '$!reified'),1,nqp::null); dd @a
camelia Array @a = ["a", Any, "c"]
Zoffix m: use nqp; my @a = <a b c>; nqp::bindpos(nqp::getattr(nqp::decont(@a), List, '$!reified'),1,nqp::null); dd |@a
camelia Str @a = "a"
Mu
Str @a = "c"
BenGoldberg m: my @a = <a b c>; @a[1] = slip(); dd @a[1]; 01:24
camelia Slip @a = Empty
BenGoldberg m: my @a = <a b c>; @a[1]:delete; dd @a[1];
camelia Any @a = Any
BenGoldberg m: my @a = <a b c>; @a[|1]:delete; dd @a[1];
camelia Any @a = Any 01:25
BenGoldberg m: my @a = <a b c>; @a[|1]:delete; dd [@a[1]];
camelia [Any]
luxio BenGoldberg: camelia works in queries, you know
Zoffix .ask lizmat should "holes" in Arrays that get created from `:delete` use `is default` value maybe? IIRC we talked about it long time ago and at the time the `nqp::null` was decided to be good enough, but now some users are having issues that they get an unwanted Mu in a hole ( irclog.perlgeek.de/perl6/2017-07-23#i_14909593 ). I figured maybe it should be what you'd get as if an element was never there. 01:31
yoleaux Zoffix: I'll pass your message to lizmat.
luxio Is perl generally considered a lightweight language?
Zoffix luxio: which one?
luxio perl6
Zoffix don't think so 01:32
"A lightweight programming language is one that is designed to have very small memory footprint, is easy to implement (important when porting a language), and/or has minimalist syntax and features."
Yeah, we're definitely not minimalist syntax
luxio I'm more concerned about memory footprint
geekosaur luxio, lightweight languages are generally things like lua and tcl that are extremely compact 01:33
AlexDaniel luxio: definitely not yet
Zoffix Ohhhh 01:35
.tell lizmat disregard. I see it already does use is default :)
yoleaux Zoffix: I'll pass your message to lizmat.
Zoffix dugword: well, I see the problem but I think the current behaviour might be correct. 01:42
dugword: deleted hole has `is default` value, which is an `Any` by default in Arrays, which is what you get when you dump it. But Slips ain't got an `is default`, so when you convert your array to a Slip, you get the hole in all of it's Mu glory 01:44
dugword Interesting
01:44 Cabanossi left 01:45 Cabanossi joined 01:46 ilbot3 left
dugword Thank you for looking into it. Either way is fine, I can extract the values using an Array method instead. The inconsistency is what threw me. Should [].flat also return (Any)? It currently does. 01:49
Zoffix dugword: filed it as a ticket. I think the behaviour does make sense, but it's not quite ideal, so maybe someone has a better idea: rt.perl.org/Ticket/Display.html?id=131783
dugword: ok, then at least something is definitely buggy. [].Seq does get `is default` value too. So either it should get a Mu or .Slip and .List should get `is default` 01:51
01:51 ilbot3 joined, ChanServ sets mode: +v ilbot3 01:52 cdg left
AlexDaniel actually, the problem in this case is not with :delete 01:56
but it is about holes, yes
m: my @a = 42; @a[5] = 49; say |@a
camelia 42(Mu)(Mu)(Mu)(Mu)49
Zoffix aye 01:57
AlexDaniel dugword: and here is the exact line where it appears in your case :) github.com/dugword/perl-Convert-AS...6.pm6#L873
dugword Great, thanks Alex 01:58
AlexDaniel there are many similar lines, so I guess they are problematic as well
timotimo luxio: rakudo perl6 on moarvm takes about 63 megabytes of ram for bare startup. there's room for improvement, though
AlexDaniel but you can do this before and after that line and see what it says: say |$vs[4];
02:06 noganex_ joined 02:09 noganex left 02:17 mr-fooba_ left 02:29 alexi5 left 02:36 mr-foobar joined 02:40 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg 02:45 skids joined 02:51 skids left
BenGoldberg m: my @a; ++@a[1]; dd @a.slip 02:52
camelia No such method 'slip' for invocant of type 'Array'. Did you mean any of these?
Slip
flip
sin
skip

in block <unit> at <tmp> line 1
BenGoldberg m: my @a; ++@a[1]; dd @a.Slip
camelia slip(Mu, 1)
02:58 itaipu left 02:59 Cabanossi left 03:00 Cabanossi joined 03:18 skids joined 03:20 dj_goku left 03:23 perigrin left 03:37 Cabanossi left 03:40 Cabanossi joined 03:55 Sgeo_ left 03:56 skids left 04:19 araraloren_ joined 04:21 araraloren left 04:22 skids joined 04:23 Cabanossi left 04:25 Cabanossi joined 04:41 raiph left 04:51 luxio left 04:56 khw left 05:11 Morfent left 05:12 skids left 05:16 Sgeo joined, CorePower joined 05:30 CorePower left 05:31 skids joined 05:32 dugword left 05:35 BenGoldberg left 05:45 mr-foobar left 06:16 skids left 06:34 setty1 joined 06:50 darutoko joined 06:58 vendethiel left 07:00 vendethiel joined 07:18 dwarring left 07:26 rindolf joined 07:28 mr-foobar joined 07:32 dugword joined
AlexDaniel fire in the hole!! 07:34
07:34 bloatable6 left, benchable6 left, evalable6 left, coverable6 left, quotable6 left, greppable6 left, bisectable6 left, unicodable6 left, committable6 left, nadim joined 07:37 dugword left, Cabanossi left 07:38 evalable6 joined, committable6 joined, ChanServ sets mode: +v evalable6, quotable6 joined, ChanServ sets mode: +v committable6, ChanServ sets mode: +v quotable6, unicodable6 joined, bisectable6 joined, bloatable6 joined, ChanServ sets mode: +v unicodable6, ChanServ sets mode: +v bisectable6, ChanServ sets mode: +v bloatable6 07:39 greppable6 joined, ChanServ sets mode: +v greppable6, benchable6 joined, ChanServ sets mode: +v benchable6, statisfiable6 joined, ChanServ sets mode: +v statisfiable6, Cabanossi joined 07:41 coverable6 joined, ChanServ sets mode: +v coverable6 07:42 parv joined 07:49 setty1 left 08:08 domidumont joined 08:12 domidumont left 08:13 domidumont joined 08:19 wamba joined 08:22 jeromelanteri joined, Cabanossi left
jeromelanteri imagebin.ca/v/3UKfuqqjDgZI 08:22
f.perlbot.pl/p/ncocds
the second prompt question does not appear but the prompt is running and waiting for an answer. After you hit y or n, the prompt question appera with the answer. 08:23
Just because i use Term::Prompt (if not use, all works fine).
i can not resolve this because i definitly not understand why this happen. 08:24
*try the code and see by yourself (i do this little code specificaly for show the problem and find someone should ba able to help me on that)
08:24 Cabanossi joined 08:31 rindolf left 08:36 rindolf joined 08:39 Dunearhp_ joined 08:40 Dunearhp left 08:43 parv left
nine jeromelanteri: that is Perl 5 code. This channel is about Perl 6 which is an entirely different (though related) programming language. Please try #perl instead 09:02
09:03 nebuchadnezzar left
jeromelanteri nine, oh ok, sorry, bye 09:04
nine jeromelanteri: no worry :) You can also stay here anyway. This is a very friendly place. 09:05
jeromelanteri i was banned from perl by Grinnz because of i ignored him on an other one serveur then... i tryed here (and i see there were no one on line speaking also)
nine, thank you, but i just start learning perl5
nine And FWIW I see nothing obviously wrong with your script 09:06
jeromelanteri nine, ho thank you to look at it (very sympatic). did you try to run it ?
nine I also haven't seen a for (my $i = 0; $i < scalar(@selection); $i++) { style loop in a long while. In Perl the loop is better written as for my $item (@selection) { push @table, [ $list_of_items[$item], $link_with_items[$item] ]; } or just my @table = map { [ $list_of_items[$_], $link_with_items[$_] ] } @selection; 09:09
jeromelanteri nine, yes sure, i also cut this part for check if this have an influence on the result code... 09:10
f.perlbot.pl/p/l9bnkt
and in fact, same problem...
nine jeromelanteri: oh, before you do anything else at all, add "use strict; use warnings;" before line 2 of your script 09:12
jeromelanteri: oh, disregard that. I see it in line 3 09:13
jeromelanteri nine, yes, and i do an other change for clarify the situation (just change sub name and remove table part totaly)
f.perlbot.pl/p/uno44m
nine I jumpted to conclusions because I can't find -y1 as a valid option in IO::Prompter's docs
jeromelanteri yes, and this works fine. 09:14
if i just remove Term::Prompt, IO::Prompter works well (not if i use Term::Prompt BEFORE).
nine, ok then i know something more because of you: 1/ my code is not wrong, 2/ you havbe the same problem as me: your IO::Prompter prompt line appear AFTER it wait for an answer (that's not normal). 09:15
09:16 nebuchadnezzar joined
nine jeromelanteri: in fact, the script seems to work just fine for me 09:19
jeromelanteri: gist.github.com/niner/9563745f2db9...91297eef6a
jeromelanteri nine, you have the second prompt question appear BEFORE it wait for answer ?
nine The "Now could you read.."? yes 09:20
jeromelanteri ho... i have it after i answer...
09:20 Skarsnik joined
jeromelanteri what is your OS running ? 09:20
i can make a video for show that... 09:21
nine jeromelanteri: openSUSE Tumbleweed, perl 5.24.1 IO::Prompter 0.004014 Term::Prompt 1.04 using bash in konsole 17.04.2 09:22
jeromelanteri ho, you think it should be linked because i use zsh ?
Skarsnik Hello
nine jeromelanteri: terminals are arcane beasts full of surprises. Shells, too. That's software with 4 or 5 decades of history. 09:23
jeromelanteri nine, mmmh, yes.
i may have to look at the module Term::Prompt code, but i think my level should be not suffisant for find bug related to shell console used. 09:24
ok, thank you
nine jeromelanteri: to come back on topic though, may I interest you in writing that script in Perl 6 just for fun? Perl 6 does have a builtin "prompt" function and fewer ways to screw up. E.g. you wrote @list_of_items[$selection[$i]] which would be somewhat correct Perl 6 but not actually Perl 5 ;) 09:25
jeromelanteri nine, yes sure, why not ? if for me use perl6 for write little script (actually i plan to script for vim plugins implementation from git remote repo helper)... 09:26
is more easy
nine, i'm on archlinux and or antergos OS... any pacman packages for that ?
nine jeromelanteri: not sure. I've never used arch 09:27
And never heard about the other one :) 09:28
jeromelanteri ok, i will find it this day and try to copy then change my code relatively to perl6. Any link for learn "quickly" this perl6 langage ?
09:28 rindolf left, domidumont left
nine jeromelanteri: if not, installation is really just: git clone github.com/rakudo/rakudo.git && cd rakudo && perl Configure.pl --gen-moar --make-install 09:29
jeromelanteri nine, ok i will before to try this search for nice pacman packages of perl6, it should have some... 09:30
nine jeromelanteri: docs.perl6.org/ is your place to go. perl6.org/resources/ lists a lot more
jeromelanteri: package could be named rakudo (as that's the implementation) 09:31
jeromelanteri ok
09:33 TEttinger left, rindolf joined 09:49 nadim left
jeromelanteri f.perlbot.pl/p/umuq64 09:52
nine, i find rakudo, zef, alacryd, panda, and some "modules" ? 09:55
bamboo 09:56
and is ther any cpan related modules for perl6 ? (or an other than mspan ?) 09:57
s/mspan/mcpan/
10:01 jeromelanteri left, nadim joined 10:14 setty1 joined 10:15 jeromelanteri joined 10:16 espadrine left 10:22 nadim left
nine jeromelanteri: what do you mean by "cpan related"? 10:28
10:29 lizmat left 10:30 lizmat joined
jeromelanteri nine, i have cpan command for install module (from cpan site i think... or mcpan). I think i understand that it is zef who does this... install modules... 10:30
nine, i'm just reading doc of perl6 now... it seems to be more OO than perl5 10:31
nine jeromelanteri: yes, zef is the too to use
s/too/tool
much more so, yes
jeromelanteri syntax is nice
nine Though it's perfectly fine with being used as a functional or procedural language as well. Whatever fits better with the problem 10:32
jeromelanteri yes, i read this also... like haskell then ?
nine It got much inspiration from haskell. But also tries to be more easily accessible
jeromelanteri i have some little problem for install zef from archlinux (actually, packages are not in official maintained repos, but in AUR repos... then... fight a little for install things, it should run soon) 10:33
10:33 setty1 left
jeromelanteri nine, can it be also linked with a GUI ? 10:33
nine Installing zef from source is rather trivial if using the package doesn't work out
There's a Gtk::Simple module 10:34
jeromelanteri nine, package ask for perls-file-find and perl6-shell-command (that i allready install...)
nine, only tk GUI ?
oh ... gtk... 2 or 3? 10:35
nine GTK 3
jeromelanteri (gtk is fine for me)
good
nine modules.perl6.org/#q=gtk%3A%3Asimple
jeromelanteri yes... i think i would like to try that, it seems to be powerfull and funny for use, let's try that... 10:36
El_Che jeromelanteri: I thought arch got rakudo pkgs covered. If not I can have a look at creating a pkg 10:46
jeromelanteri El_Che, yes, i do install rakudo well, now i'm trying to install zef, but there is some problem (it search for alacryd... who is allready installed but whereis alacryd show nothing). 10:47
i also have to install pand who actually failed for same reason.
panda
(and i see my vim has allready perl6 syntax ability) 10:48
El_Che jeromelanteri: panda is obsoleted 10:50
jeromelanteri El_Che, ok, a problem fixed then... :)
El_Che I provide 2 scripts in my rakudo pkgs: 10:51
github.com/nxadm/rakudo-pkg/blob/m...ef_as_user
github.com/nxadm/rakudo-pkg/blob/m...ef_as_root
most users will use the first 10:52
jeromelanteri well.. let's see that...
El_Che the magic is just /opt/rakudo/bin/perl6 -Ilib bin/zef --install-to=perl install .
(I just returned from holidays todat, I'll try to push new rakudo pkgs today or soonish) 10:53
jeromelanteri El_Che, then i have just to change my rakudo real directory position...
El_Che, an idea should be to use command find in the script for find rakudo directory also ? 10:54
nine El_Che: why --install-to=perl?
10:55 itaipu joined
jeromelanteri also, why (but maybe it is your distro who ask for that, i don't know) not just link on /usr/local/bin ? 10:55
El_Che nine: I need to check that. If I remeber correctly that will it will end in whatver .perl6 dir is 10:56
jeromelanteri: the idea is to have it locally contained without needing root to install modules
10:56 lizmat_ joined
jeromelanteri ho yes, why not... 10:56
nine El_Che: no, that would be home. perl is the name of the repo that's meant to contain only the CORE dist which is installed with rakudo itself 10:57
El_Che everything in the pkg is in /opt/rakudo. The user modules in .perl6 (expect if installed as_root, then in /opt/rakudo)
I need to check with ugexe that game me the info. However, this works as expected (despite the cargo cult)
nine: I had trouble getting it to install correctly at the beginning. 10:58
El_Che off
10:58 lizmat left
jeromelanteri i see there also have a solution for install from docker. 10:59
with docker
nine m: say $*REPO.repo-chain.grep({$_.can("name") and $_.can("prefix")}).map({"$_.name(): $_.prefix()"}).join(", ") 11:00
camelia home: /home/camelia/.perl6, site: /home/camelia/rakudo-m-inst-1/share/perl6/site, vendor: /home/camelia/rakudo-m-inst-1/share/perl6/vendor, perl: /home/camelia/rakudo-m-inst-1/share/perl6
nine El_Che: ^^^
jeromelanteri El_Che, from pacaur AUR repos installation, i have my perl6 (so rakudo) inside /usr/share/perl6 and binary linked to /usr/bin/perl6 11:04
nine jeromelanteri: installing zef is a matter of: git clone github.com/ugexe/zef.git && cd zef && perl6 -Ilib bin/zef install . 11:05
jeromelanteri nine, yes, i'm just doing that.
nine, then now, how to use zef ? 11:10
nine zef install Whatever::Module
zef --help
jeromelanteri nine, no
we just install zef somewhere distro can not find it... 11:11
nine jeromelanteri: zef should have told you where it installed itself, so you can add that path to your $PATH 11:12
11:12 mr-fooba_ joined
jeromelanteri i just see that... alacryd is inside /usr/share/perl6/vendor/bin/ and zef is inside: /usr/share/perl6/site/bin (w t f) 11:12
nine That actually sounds quite sane 11:13
11:13 mr-foobar left
jeromelanteri nine, i would prefer to just create a symbolic link and not do things dirty inside my PATH variable env 11:13
nine What is dirty about that?
That's exactly what PATH is there for
jeromelanteri nin, ok... good news then ?
nine /usr/share/perl6/vendor is for stuff installed from packages, /usr/share/perl6/site is for manually installed stuff 11:14
jeromelanteri nine, yes but if i loose this or for reason of backup things, it will become dirty and boring to remember all the (unorganized) place of every thing in the room.. (i like when things are sorted and clean)... 11:15
nine
well
i will use ln -s
nine, why is there: zef zef-j zef-m ?
nine Because rakudo can run on different backends (JVM, MoarVM) and sometimes you want to be explicit about which one you use 11:16
jeromelanteri (same for alacryd ?)
ok, fine
it will find them alone or i need to syboli linked them also ?
nine You can pretty much ignore that issue 11:17
jeromelanteri ok
11:17 lizmat_ is now known as lizmat
jeromelanteri fine... well, something more usual to install for use perl6 ? 11:19
El_Che lizmat_ dropping the _ is like a Clarck Ken/Superman experience 11:20
lizmat yeah, exposed by a flaky Ziggo ubee base station
jeromelanteri nine, a last question... is there any problem relative to use perl6 things from zsh shell (doesn't want bash anymore) ? 11:22
nine jeromelanteri: none that I'm aware of 11:25
jeromelanteri cool 11:26
11:27 mr-foobar joined
El_Che jeromelanteri: just backup your profile file containing your PATH config. It's a lot less messy than remembering syslinks 11:29
11:29 mr-fooba_ left
jeromelanteri El_Che, that is right. 11:32
well... tomorrow i will be back if i failed to convert my code or for ask questions. thank you nine and El_Che. 11:36
11:39 itaipu left, BenGoldberg joined
El_Che np 11:43
11:47 nadim joined 12:06 espadrine joined 12:08 Cabanossi left 12:09 Cabanossi joined 12:24 bjz joined 12:42 bjz left
nadim Hi, can we return a list with named elements? sub x { :name<...>. :other<...>, 123, 456 } my (:$other, @positional, :$name) = x() ; # I will return a Hash but I am curious 12:45
Zoffix: that sounds like a question for you ^
12:49 khw joined 13:00 mr-fooba_ joined, mr-foobar left 13:07 Cabanossi left
Skarsnik AlexDaniel, here? 13:08
13:09 Cabanossi joined 13:10 bioexpress joined
bioexpress Hello, is it save to splice more elements than the array contains? 13:13
Juerd Yes 13:16
bioexpress Thx! 13:18
MasterDuke Skarsnik: i tried running your leaktest script, but didn't notice any leaking
Skarsnik witch one? did it output stuff? x) 13:19
MasterDuke i modified gist.github.com/Skarsnik/03b970d2a4b827ba1e1d 13:20
jeromelanteri it is oo really... this concept is colser than C++... il love it.
MasterDuke i took out the Linux::Proc::Statm stuff since i didn't have that installed. ran it under heaptrack and memory consumed plateaued at ~120mb 13:21
jeromelanteri is there any prompt module for multi-selections menu ?
Skarsnik hm weird 13:22
did you ran it long?
MasterDuke i reduced the loop to 10
Skarsnik well it leak very little now 13:23
so it hard to notice
13:24 eliasr joined 13:29 zakharyas joined 13:30 bioexpress left
timotimo jeromelanteri: we do have "Term::Choose" 13:40
.eco Term::Choose
13:40 nadim left
timotimo buggable: eco Term::Choose 13:40
buggable timotimo, Term::Choose 'Choose items from a list interactively.': github.com/kuerbis/Term-Choose-p6 2 other matching results: modules.perl6.org/#q=Term%3A%3AChoose
jeromelanteri timotimo, thank you 13:44
timotimo this website there has pretty much all of our modules
14:01 zakharyas left
jeromelanteri timotimo, very good. now... how to use class/modules from different files (for perl5, create a .pm file, then import it by use package_name, use of @INC, etc...) but for perl6 ? i not find any tuto on that. any link for tutos ? 14:06
Skarsnik docs.perl6.org/language/modules.html 14:07
jeromelanteri Skarsnik, thank you 14:08
Skarsnik docs.perl6.org/language.html lot of stuff in there already
Zoffix .tell nadim yes you can. Just use binding instead of asignment to cause signature binding instead of list assignment. `my () :=` is just a signature so you can use whatever you can use in normal signatures (sans anything currently not implemented yet). sub x { :name<foos>, :other<meows>, 123, 456 }; my (:$other, *@positional, :$name) := x; dd [$other, $name, @positional] 14:09
yoleaux Zoffix: I'll pass your message to nadim.
Zoffix m: my @a is default(42); @a[2] = 70; { dd [$^a, $^b, $^c ] }(|@a) 14:14
camelia [Mu, Mu, 70]
Zoffix m: my @a is default(42); @a[2] = 70; dd @a.FLATTENABLE_LIST 14:15
camelia (Mu, Mu, 70)
Zoffix wc
14:21 Cabanossi left 14:24 Cabanossi joined 14:28 gdonald left 14:35 pilne joined 14:40 nadim joined, gdonald joined 14:41 zacts joined 14:58 zakharyas joined 15:02 BenGoldberg left 15:04 BenGoldberg joined
El_Che if someone asks, new rakudo 2017.07 linux packages: twitter.com/nxadm/status/889139909860130817 15:08
15:08 skids joined 15:10 Guest90987 joined, macsnowball joined 15:23 Cabanossi left 15:24 Cabanossi joined
timotimo cool, El_Che++ 15:30
BenGoldberg Ubuntu version numbering has me confused -- why is 16.10 at the end of it's life now, whereas 16.04 expires in April 2021? 15:38
15:40 itaylor57 left, nowan_ left 15:43 nadim left 15:44 nowan joined
skaji X.04 where X is even is LTS (long term support). 15:46
15:52 itaylor57 joined 16:01 nadim joined 16:03 zakharyas left 16:06 zakharyas joined, eyck left 16:11 araraloren_ left 16:22 Cabanossi left 16:24 Cabanossi joined
stmuk_ pl6anet.org/drop/rakudo-star-2017.07-RC1.tar.gz 16:32
that's likely to be the release tomorrow 16:33
16:33 dugword joined 16:37 dugword left
El_Che stmuk_++ 16:38
stmuk_: although my main OS is linux and I use my own packages, I have a win and osx around where precompiled pkgs are handy :) 16:39
BenGoldberg: non LTS version have a 9 month support window
stmuk_ El_Che: the Mac DMG RC will follow shortly! 16:40
El_Che BenGoldberg: so when 18.04 comes out, I'll end up having 16.04, 17.10 and 18.04 debs
stmuk_: appreciated!
16:44 mr-fooba_ left 16:49 mr-foobar joined
stmuk_ El_Che: see #perl6-dev (if you haven't) 16:56
16:58 rubio_terra joined 17:02 zakharyas left 17:03 zakharyas joined
SmokeMachine I was thinking: if :!$foo is valid, shouldn’t :?$foo be valid? (As I said earlier) but now I’m not sure what I think, because if that’s true, so I think :+$foo, :-$foo, :~$foo, :++$foo, etc should all be valid... I don’t know what I think about it... 17:03
17:07 Cabanossi left 17:09 Actualeyes joined, Cabanossi joined
zengargoyle hehe, i now want ://$foo, :||$foo, for like the p5 (foo => $foo)x!!$foo, or (:$foo if defined $foo) like things where you only want to pass the named param if it is true/defined but otherwise leave it out and accept the default... 17:13
maybe ':' should be a slang like 'Q' :P 17:14
zengargoyle jests
17:16 domidumont joined
zengargoyle i did tend to use the Enterprise operator ()x!! quite a bit, not sure of the most p6 way ATM. 17:18
Juerd I have a hash that I'm passing to a standard .new with Foo.new(|%hash); one of its values is an array, which gets into the "has @.bar;" as item 0 17:25
Instead I'd like to assign the elements from that array to @.bar flattened
The hash comes from JSON::Tiny
Is there a way to do this, without specifying the specific key? 17:26
As a work-around I've changed "has @.bar;" to "has $.bar;" :( 17:30
zengargoyle that was sorta going to be my suggestion, or make a BUILD/TWEAK to patch it up otherwise. 17:31
what would you want? some sort of trait to say 'flatten any has @. attributes'? 17:33
or 'has @.bar something-that-says-to-flatten' 17:34
17:40 itaipu joined, zakharyas left
AlexDaniel Skarsnik: yes 17:43
17:45 mr-foobar left 17:48 mr-foobar joined
Skarsnik AlexDaniel, I did a small version of the gumbo/xml module for test purpose, sadly no crash 17:51
Zoffix SmokeMachine: :!$foo is not valid
Skarsnik want it to play around? 17:52
Zoffix Seriously, all the people who want to pile in on feature proposals should at least learn the current ones >_<
17:53 parsnip joined
parsnip after doing `brew install rakuda-star`, i have issues like sprunge.us/ 17:53
oops, sprunge.us/fMNA 17:54
SmokeMachine Zoffix: you are right! 17:55
AlexDaniel Skarsnik: hm, yeah? I can try running it for a little longer… :)
geekosaur parsnip, sounds like something replaced your perl 5 in the process without rebuilding a perl-enabled vim
which would be a brew packaging/maintenance issue
parsnip thanks 17:56
geekosaur (in fact the kind of issue which is why I don't use homebrew)
too easy to get into that kind of situation
parsnip but also, when i open a terminal, it seems .bashrc isn't loaded anymore
Skarsnik www.nyo.fr/~skarsnik/tmp/minigumbo/
AlexDaniel Skarsnik: wow that's really small 17:57
Skarsnik You will need the Binding.pm6 from Gumbo 17:58
17:59 setty1 joined
AlexDaniel Skarsnik: well, yes, I don't see any issue with this 18:02
18:02 itaipu left
AlexDaniel Skarsnik: would be nice to have a minimal example that *does* crash :) 18:03
zengargoyle Zoffix: i think we're just thinking out loud and not terribly serious....
Skarsnik My best guess at this point is stuff inside the xml module 18:04
geekosaur parsnip, there's only a billion things that can lead to that :/ 18:05
zengargoyle m: my $x; sub foo(:$x = 42){ say $x }; foo( |%(:$x if $x) ); 18:06
camelia 42
zengargoyle is that a decent thing?
Zoffix zengargoyle: no need foe % 18:07
*for
|(:) aka the ninja operator 18:08
zengargoyle m: my $x; sub foo(:$x = 42){ say $x }; foo( |($x andthen :$x) ); 18:10
camelia 42
zengargoyle depending on which direction you think in. :)
Zoffix oh hah 18:12
I looked in the logs and 42 has a question marl after i
had me confused
moritz: looks like log encoding isn't 100% back to normal unicode n/l char turns up as a question 18:13
18:13 eyck joined
Zoffix Unless it's my font :/ though I'm pretty sure it used to work 18:13
parsnip ah, my .bashrc wasn't being loaded because installing rakudo-star also added a file .bash_profile with the line `eval "$(perl -I$HOME/perl5/lib/perl5 -Mlocal::lib)"`
zengargoyle parsnip: ouch! 18:14
Zoffix :o 18:15
MasterDuke anybody have a machine with more than 16gb ram to test something on?
parsnip could rent a VPS 18:17
zengargoyle is OSX default shell bash?
MasterDuke parsnip: seems like overkill to run one line of code though... 18:19
zengargoyle that would bonk a Debian as well, uses .profile instead of .bash_profile
parsnip not sure what did that, i don't see it at github.com/Homebrew/homebrew-core/...do-star.rb 18:20
geekosaur bash uses .bash_profile if it exists, otherwise .profile 18:22
18:22 Cabanossi left
parsnip hmm, but file creation time was 13 hours ago, so i must have done something else... 18:22
geekosaur so yes this situation is a nasty corner case by someone who didn't think through how shells work. (also doesn't allow for people using zsh, fish, etc.) 18:23
parsnip i don't think this is perl6 related
i don't know where that file came from
stmuk_ Outer Space?
or local::lib
geekosaur the latter from what I am seeing 18:24
homebrew got "clever:
well, the hb maintainer for local::lib
18:24 Cabanossi joined
geekosaur there's some hb bug reports for past instances of that maintainer being "clever" too 18:25
in any case it'd be related to the same issue of something upgrading your hb-installed perl with one that breaks stuff
(perl5)
parsnip the same line appears at search.cpan.org/~haarg/local-lib-2....cal/lib.pm 18:28
zengargoyle that's a pretty standard thing for installing local::lib, you put that in your .profile/.bash_profile. hb should have checked if .profile existed and put it there instead of to .bash_profile. 18:30
18:31 macsnowball left
zengargoyle i don't think it's even bash specific and would be fine for plain old sh as well. 18:32
is your bash also homebrew?
parsnip no, i think my bash is from macos 18:33
so like, gplv2
does this edit of .profile take into account how macos does things differently? i'm trying to recall, but i think macos loads .profile on every terminal. 18:34
*terminal emulator instance 18:35
zengargoyle *maybe* homebrew bash is setup to use a .bash_profile vs falling back to .profile. so *maybe* hb rakudo-star is built with hb bash in mind.... but it's still a bit wonky.
parsnip i'm not even sure rakudo-star did it, i can't reproduce 18:36
zengargoyle in general, bash does .bash_profile (if it exists) or .profile otherwise (but only on 'login' shells) and does .bashrc on non-login shells. IIRC
geekosaur I don;t think it did it directly, most likely it just depended on a new perl5 package
and that package is .. unfortunate
18:37 pompomcrab joined
geekosaur anyway "brew doctor" might help with the collateral damage from the new perl5 package 18:38
zengargoyle i don't ack local::lib or .bash_profile in rakudo source.
i'd guess it's in one of those bottle things... 18:39
parsnip timestamp is 13 hours ago, i installed rakudo-star only an hour ago, so i either started trying perl last night, or something else provoked it. 18:40
18:41 pompomcrab left
zengargoyle some systems use a .bash_profile that sources .profile (if it exists) and then source .bashrc (if it exists). 18:45
18:45 mr-foobar left
zengargoyle it's actually a bit of a horrorshow on like older Solaris when you have a /bin/sh to get everything *just right* so you have the environment you want both when you login and when you just `ssh host comman`. 18:49
geekosaur ob github.com/geekosaur/dotty 18:51
(I haven't completely removed the horrorshow to support ultrix, much less solaris (which I actually have to still care about because of customers)
zengargoyle geekosaur++ :) i went mad with like 4 versions of Solaris, RedHat, Gentoo, and Debian and finally just mostly gave up. 18:54
i keep around ~/profile.XXX scripts and just source them as needed. (i don't have to worry about such things ATM) 18:56
18:57 ufobat joined
geekosaur well, that came from a university environment where I didn't get to say what platforms I had to deal with 18:59
zengargoyle same. 30 years of cruft basically... 19:05
Juerd zengargoyle: Maybe, or maybe something that can un-itemize the array in a |, or ... I don't know. The current things just don't combine well. 19:06
19:06 gfldex left 19:07 gfldex joined
zengargoyle Juerd: i totally see where it could be a handy thing. i could imagine a Role or something that does the BUILD thing and introspects for @/% attributes and does stuff. but it's way beyond me. 19:11
19:11 domidumont left
ugexe thats just how binding works 19:14
otherwise method BUILDALL(|) { my $self = callsame; @.foo = |@.foo; $self; } or some such works
19:25 DBeepBeep joined
DBeepBeep How do we run a script with multiple files? 19:26
19:26 |oLa| joined, TEttinger joined
zengargoyle DBeepBeep: what do you mean? 19:30
DBeepBeep zengargoyle, I've got a grammar in one file that I'd like to test, but I can't seem to get the test file to find it. 19:32
I'm running with a -I to the directory. 19:33
19:34 darutoko left
zengargoyle sounds like you need to make a module. 19:37
DBeepBeep: is your grammar file a module that can be use'd? 19:44
DBeepBeep Yeah 19:45
It's just a grammar surrounded by module. The file has a pm6 extension
zengargoyle try putting it in a ./lib/NAMEGOESHERE.pm6 19:46
i'm not sure if -I. looks in the current directory or actually expects loadable things to be in ./lib
but modules go under ./lib is the norm.... 19:47
AlexDaniel u: combining 19:49
unicodable6 AlexDaniel, U+0300 COMBINING GRAVE ACCENT [Mn] ( ̀)
AlexDaniel, U+0301 COMBINING ACUTE ACCENT [Mn] ( ́)
AlexDaniel, 440 characters in total: gist.github.com/6c5a7d6aaf79516e65...bb39364c96
zengargoyle i think most people now are using App::Mi6 to generate a skeleton 'distribution' that creates some files and tests and has some commands to make module building and testing a bit easier. if you havn't tried it yet, it's worth a look.
AlexDaniel Skarsnik: I let it run 10000 times and it did not crash 19:50
Skarsnik: so I'd say there's no bug in this one for sure
u: combining letter
unicodable6 AlexDaniel, U+0363 COMBINING LATIN SMALL LETTER A [Mn] ( ͣ)
AlexDaniel, U+0364 COMBINING LATIN SMALL LETTER E [Mn] ( ͤ)
AlexDaniel, 145 characters in total: gist.github.com/8acbef9bf1f5d04372...64daf5d8d8
zengargoyle or i have a module that uses a grammar with actions and tests that you could look at. or there are some other modules that use grammars that you could look at how they are laid out... 19:51
SmokeMachine zengargoyle: I’m using 6pm... :P 19:52
AlexDaniel u: -2
unicodable6 AlexDaniel, U+106A MYANMAR SIGN WESTERN PWO KAREN TONE-2 [Mc] (ၪ)
AlexDaniel, U+1087 MYANMAR SIGN SHAN TONE-2 [Mc] (ႇ)
zengargoyle SmokeMachine: haven't even heard of it... :)
unicodable6 AlexDaniel, 854 characters in total: gist.github.com/ab97f1a0c8e64d2fbe...fd1e1c2909
SmokeMachine zengargoyle: github.com/FCO/6pm 19:53
AlexDaniel test (◌ၪ)
moritz: well… irclog currently does something weird
moritz: compare this irclog.perlgeek.de/perl6/2017-05-25#i_14635499 and this irclog.perlgeek.de/perl6/2017-07-23#i_14911623 19:54
19:54 DBeepBeep left
AlexDaniel (oops, I meant the line below my message) 19:55
zengargoyle SmokeMachine: looks neat. like Carton-ish.
ugexe if you point -I to a directory the name of the module must be split into subdirectories by name part unless the -I directory contains a meta6.json 19:56
SmokeMachine zengargoyle: that’s the idea :) 19:57
Skarsnik AlexDaniel, could be because the MiniXML object are really small
zengargoyle ugexe: so more likely some naming problem? ./MyGrammar.pm6 and -I. should be able to 'use MyGrammar;' 19:59
Skarsnik or because parse-html in gumbo create a whole class that is not here 20:00
ugexe well you can still do -Isome/dir and `use NameOfFile;` but NameOfFile wouldnt be the name of your grammar/classes
zengargoyle guess i haven't really tried things apart from mi6 and a probably decent starting point for a module..
20:01 eyck left
ugexe but -I. usually refers to something containing a META6.json file, which won't just guess the filename if it exists 20:02
zengargoyle ah, think i probably got it but woud have to play around to grok the details. :) 20:03
20:10 |oLa| left 20:11 rindolf left 20:12 rindolf joined
zengargoyle does rakudobrew build take advantage of multiple cores? 20:14
20:14 perigrin joined
zengargoyle heh, 6pm seems bork'd 20:17
20:21 mr-foobar joined, rindolf left 20:23 rindolf joined
SmokeMachine zengargoyle: really? Could you send me the error? 20:27
zengargoyle SmokeMachine: gist.github.com/4f614ac3d10f884382...e942310625 20:33
with the zef debug magic: ZEF_PLUGIN_DEBUG=1 zef --debug install 6pm 20:34
it's a fresh (like just nuked and rebuilt via rakudobrew) moar... so no surprise something may have changed.. :) 20:36
20:37 |oLa| joined
zengargoyle but i nuked and built because it failed the same way on like 2017-07 from a few days back... 20:38
20:38 eyck joined 20:45 itaipu joined
ugexe maybe try `zef install 6pm --deps-only && zef look 6pm && perl6 --ll-exception -Ilib t/01-six-pm.t` to show the stack trace 20:52
20:55 Skarsnik left 20:58 konsolebox left
zengargoyle ok, weird. Installing: JSON::Fast:ver('0.9.2') ... # Could not find JSON::Fast:ver<0.4..*> at line 62 in: 21:15
did zef look 6pm and zef install --deps-only and JSON::Class fails from JSON::Marshal which fail from JSON::Fast. 21:16
21:19 test joined
test hello? 21:20
21:20 test is now known as Guest31196
Guest31196 hello to everyone... 21:20
anyone there?
p6: say (9+9); 21:21
camelia 18
geekosaur sort of/not very...
Guest31196 anyone? 21:22
there are lots of people here apparently
zengargoyle SmokeMachine: ugexe: i'm getting crazy stuff, might be some local issue (though i nuked and rm-rf ~/.perl6 just in case). 21:23
gist.github.com/2abf87cd6740d64d1d...96945b8bfc
21:23 Guest31196 left
ugexe you should zef update, those modules were updated recently fixing that 21:23
zengargoyle deps-only croaks in some JSON::Class -> JSON::Marshal -> JSON::Fast. 21:24
i literally nuked and rebuilt via rakudobrew and build zef just hours ago.
ugexe rakudobrew doesnt nuke anything related to zef 21:25
21:25 Manifest0 left, Manifest0 joined
ugexe its cache/store anyway 21:25
zengargoyle does it not nuke site? i install under opt/rakudobrew (via permissions) and also nuked ~/.perl6 just in case. 21:26
i'll try a hard rm-rf of moar-nom 21:27
ugexe zef caching in ~/.zef, like the project list
zengargoyle oh, that update. i thought that always happened anyway for convienence....
zengargoyle thinks i complained about it once... :)
no help. zef update; zef install --deps-only . -> same JSON error. 21:29
maybe because still in look directory. 21:30
no love. think i'll go on actual rm-rf purge. 21:32
21:32 itaipu left
timotimo you'll go on a rmpage? 21:32
21:33 espadrine left
zengargoyle lol 21:33
21:33 itaipu joined 21:37 Cabanossi left 21:38 JAP{NECRO}H joined
JAP{NECRO}H p6: say 3; 21:38
camelia 3
JAP{NECRO}H hello 21:39
zengargoyle huh .... i had a lib/.precomp directory in /opt/rakudobrew probably from an ancient PERL6LIB=lib in .profile long since removed... that caused 'lib' to show up under 'rakudobrew versions'.
JAP{NECRO}H everyone
21:39 Cabanossi joined
Zoffix JAP{NECRO}H: hi 21:39
JAP{NECRO}H have any books about p6? 21:41
timotimo perl6book.com/ - here's a page about perl 6 books 21:42
JAP{NECRO}H ok. 10q 21:43
zengargoyle that lib/.precomp must have been in a CUR or something.... success after removing stray lib/.precomp in rakudobrew.
that PERL6LIB thing in evil and infectious.
timotimo being worked on 21:44
zengargoyle i'm not sure how rakudobrew picked it back up... it's just a directory named lib with a .precomp in it. how did 'build moar; build zef' end up borked? 21:46
timotimo at least i believe the "unexpected lib folders get made" thing will be fixed
zengargoyle stumbler-upon-the-wtf 21:47
21:47 setty1 left
zengargoyle now i wish i had done perl6 -V sometime in the middle.... 21:49
21:49 itaipu left 21:50 dogbert17 left
ugexe easy enough to replicate 21:51
zengargoyle sadly /opt isn't under my backup plan, wish i knew what was in that .precomp directory.
ugexe perl6 -Ibaz -e 'use Test;'
cd baz 21:52
21:54 parsnip left 22:00 pmurias left
geekosaur really needs to do the "human fuzzer" shirt thing someday :p 22:01
22:03 BenGoldberg left, itaipu joined 22:04 JAP{NECRO}H left, JAP{NECRO}H joined, unop joined 22:09 JAP{NECRO}H left 22:10 ufobat left
zengargoyle lol, the stories i have about stumbing-upon-wtf. i thought complaining like an a** would get me out of vendor meetings... it just made vendors track me down and conferences to send me more sh*t to complain about. :P 22:11
unop Is there a way to .reduce a list where the operator is held in a variable (user-defined)? e.g. something like: my $op = <+ - * />.pick; (5, 10).reduce(&$op) 22:13
AlexDaniel why not? :) 22:14
jnthn &("infix:<$op>") or so
Or just my $op = (&[+], &[-], &[*], &[/]).pick to avoid the named lookup
22:15 mr-foobar left
jnthn D'oh, shoulda been 22:15
&::("infix:<$op>") or so 22:16
zengargoyle m: my &op = (&("infix:<+>"), &("infix:<*>")).pick; say &op.perl;
camelia Type check failed in assignment to &op; expected Callable but got Str ("infix:<+>")
in block <unit> at <tmp> line 1
AlexDaniel m: my @ops = &[+], &[-], &[*], &[/]; say (5, 10).reduce(@ops.pick)
camelia -5
AlexDaniel m: my @ops = &[+], &[-], &[*], &[/]; say (5, 10).reduce(@ops.pick)
camelia 15
zengargoyle m: my &op = (&::("infix:<+>"), &::("infix:<*>")).pick; say &op.perl;
camelia sub infix:<*> (Mu $?, Mu $?) { #`(Sub+{<anon|52354672>}+{Precedence}|42153536) ... }
zengargoyle nm 22:17
unop jnthn: Sweet!! That does it, Thanks! :)
AlexDaniel m: my &op = (&[+], &[-], &[*], &[/]).pick; say (5, 10).reduce(&op) 22:19
camelia 0.5
zengargoyle m:
m: my &op = (&::("infix:<+>"), &::("infix:<*>")).pick; say [&op] 3,2;
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3<+>"), &::("infix:<*>")).pick; say [&op]7⏏5 3,2;
expecting any of:
infix
infix stopper
postfix
statement end…
AlexDaniel I guess it's not a good idea to use & sigil in this case?
zengargoyle m: my &op = (&::("infix:<+>"), &::("infix:<*>")).pick; say [[&op]] 3,2; 22:20
camelia 6 22:21
22:21 Ptolemarch left
zengargoyle m: my &op = (&::("infix:<+>"), &::("infix:<*>")).pick; say [[&op]] 3,2; 22:21
camelia 5
22:21 Ptolemarch joined 22:22 Cabanossi left
jnthn AlexDaniel: Using the & sigil would handily catch the mistake I made in my first answer :) 22:23
zengargoyle jnthn: it would...
m: my &op = (&("infix:<+>"), &("infix:<*>")).pick; say &op.perl; 22:24
camelia Type check failed in assignment to &op; expected Callable but got Str ("infix:<+>")
in block <unit> at <tmp> line 1
22:24 Cabanossi joined
zengargoyle m: my &op = (&::("infix:<+>"), &::("infix:<*>")).pick; say [[&op]] 3,2; 22:24
camelia 5
AlexDaniel m: sub op() { say ‘hello’ }; my &op = (&[+], &[*]).pick; op 5, 10 22:27
camelia Potential difficulties:
Redeclaration of symbol '&op'
at <tmp>:1
------> 3sub op() { say ‘hello’ }; my &op7⏏5 = (&[+], &[*]).pick; op 5, 10
AlexDaniel nice, this warns
22:30 rindolf left, itaipu left
SmokeMachine zengargoyle: so, is 6pm working? 22:32
zengargoyle SmokeMachine: yes, was a rakudobrew/PERL6LIB/cruft problem. it installed. haven't played with it yet.. :) 22:33
i probably installed something that used JSON modules while i was in /opt/rakudobrew directory and had PERL6LIB=lib set and somehow got a /opt/rakudobrew/lib/.precomp that somehow was relevant. 22:35
ugexe PERL6LIB="lib" perl6 -e 'use Test;'
that creates a directory "lib"
so if you ever have PERL6LIB set... heh
to a non-absolute path anyway 22:36
zengargoyle ugexe: it's more that `rakudobrew nuke moar-nom; rakudobrew build moar; rakudobrew build zef` still picked up that stray lib/.precomp
SmokeMachine zengargoyle: thanks! 22:37
zengargoyle the .profile setting of PERL6LIB was *ages* ago. it interfered with rsyncing /opt/rakudobrew betewwn machines.
ugexe it wouldnt pick it up if it wouldnt get created in the first place
zengargoyle it was already there from the past.
ugexe right
now imagine instead that it was never created in the past 22:38
22:39 vendethiel left
zengargoyle i think me having PERL6LIB set way back and installing things when i was in /opt/rakudobrew somehow stuck because maybe rakudobrew with it's magic has a -Ilib somewhere and cd's during execution. 22:39
i noticed when `rakudobrew versions` showed up a mysterious 'lib' version... 22:40
Praise va a nascondere il suo cuore, che quest'arya pare minaciosa
zengargoyle it thought lib/.precomp was an actual previous install and added it to a CUR or something like that. 22:41
there are several lines in my previous gists where zef told me JSON::XXXX was already installed. 22:42
22:45 Ptolemarch left
zengargoyle ugexe: sorry, missed "now imagine instead..." 22:46
22:50 konsolebox joined 22:54 TEttinger left, JAP{NECRO}H joined
ugexe instead it should not be created if the CURFS target does not exist. presumably there was never a legit lib/ folder in your rakudobrew folder 22:56
22:59 unop left
zengargoyle sigh, Praise thinks we're fighting. :( 23:02
i'm just more curious as to how that lib directory broke things. because i rm-rf'd the moar-nom and ~/.zef and ~/.perl6 and re-built and still failed until i deleted the lib directory. 23:05
23:05 Guest90987 left
zengargoyle i know how it got there and that it soon would not happen again. 23:06
23:08 cognominal joined
zengargoyle the only option i see is that somehow rakudobrew noticed it and integrated it into the CUR. even --force-install of JSON modules after they were 'already installed' still picked up old broken versions of the module. 23:09
23:10 rindolf joined
zengargoyle *really* wish i had done a perl6 -V or more zef install location searching to figure out what was happening. 23:10
how does recomp work now? in the past when doing 'build moar' you had to re-install all modules, now it seems to notice and reinstall zef and modules get carried over and re-precompd if needed. 23:16
when do you actually need to nuke everything and start from scratch? 23:21
23:22 rindolf left
zengargoyle answers when it doesn't work. :) 23:22
23:22 Guest90987 joined 23:23 Cabanossi left, JAP{NECRO}H left 23:24 Cabanossi joined 23:25 wamba left 23:37 gdonald left 23:39 Guest90987 left, Guest90987 joined 23:50 gdonald joined 23:56 Guest90987 left

Powered by Cro and the Raku® Programming Language.

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

Thank you!