»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.
Geth modules.perl6.org: 86dddb247f | (Zoffix Znet)++ | update-modules.perl6.org
Temporarily always force full rebuild

For some reason it appears to never full rebuild even when commit message has `[REBUILD]` in it.
Temporarily force the value to 1, to start eliminating suspects.
Part of #76		 00:03
00:05 devmikey left, ChoHag left, ijneb left
Geth modules.perl6.org: 13c48046df | (Zoffix Znet)++ | 2 files
Add MODULES_PERL6_DO_NOT_REMOVE_OLD_DISTS env var

To be used a debugging aid to the live site		 00:08
Zoffix ohhhhh... oh ok 00:09
Yeah, it only deletes at the end by matching the build ID, so actually it's safe to kill it
Geth modules.perl6.org: 15f5404e1d | (Zoffix Znet)++ | 2 files
Revert "Add MODULES_PERL6_DO_NOT_REMOVE_OLD_DISTS env var"

This reverts commit 13c48046df719075aebdb2e846326cee4d481be7.
On second thought, not needed.
modules.perl6.org: a02751d7e3 | (Zoffix Znet)++ | update-modules.perl6.org
[REBUILD] Undo always-rebuild
 00:13
Zoffix :S 00:14
Now it's just screwing with me. Full rebuild works when starting by hand :\
Geth modules.perl6.org: 160e2ced47 | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/AppVeyor.pm
[REBUILD] Clean up AppVeyor postprocessor docs
 00:16
Zoffix sacrifices a small mammal and starts a rythmic chant
\o/ hey it worked \o/ 00:23
00:29 ijneb joined 00:31 mr-foobar left 00:32 mr-foobar joined 00:37 lookatme joined
lookatme morning 00:37
ijneb howdy lookatme!
again lol
lookatme hey ijneb
yeah, again :) 00:38
ijneb, so you are learning Perl 6 ? 00:42
Zoffix k. seems to be working now and appveyor update came through. No idea why it was broke tho. Ghost in the code, I guess. 00:43
00:47 entonian joined 00:48 jameslenz left 00:49 andreoss left 00:51 entonian left, jameslenz joined
ugexe buffer 3 00:53
Zoffix /win3 00:55
buggable Zoffix, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 23 hours, 4 minutes, and 19 seconds
Zoffix :)
01:00 Kyo91 left 01:05 mr-foobar left 01:06 mr-foobar joined 01:07 TEttinger joined 01:12 Cabanossi left 01:13 Cabanossi joined 01:32 wamba left, w_o_w joined
Geth_ perl6-examples: mattwhisenhunt++ created pull request #50:
Change URL to a real http address		 01:35
protium /win3 01:38
buggable protium, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 22 hours, 21 minutes, and 39 seconds
01:41 Cabanossi left 01:43 Cabanossi joined 01:46 ilbot3 left 01:52 ilbot3 joined, ChanServ sets mode: +v ilbot3
mspo wow package namespaces really start to sound like a good idea with all that npm stuff :) 01:57
02:02 w_o_w left
lookatme ?_? 02:05
02:08 http_GK1wmSU joined 02:11 http_GK1wmSU left 02:22 BenGoldberg joined 02:25 pelev left, pelev joined 02:28 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg 02:32 Ptolemarch left, Ptolemarch joined 02:41 Cabanossi left 02:43 Cabanossi joined
ryu0 /win3 02:45
buggable ryu0, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 21 hours, 14 minutes, and 30 seconds
lookatme What is `win3` for ? 02:47
/win3
buggable lookatme, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 21 hours, 12 minutes, and 24 seconds
lookatme /win5 02:48
buggable lookatme, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 21 hours, 11 minutes, and 58 seconds
lookatme /win
Geth ecosystem/html-escape: 95a3903c17 | moznion++ | META.list
Add HTML::Escape
 02:49
ecosystem: moznion++ created pull request #357:
Add HTML::Escape		 02:50
02:57 skids left 02:58 noganex_ joined 02:59 raschipi joined
raschipi lookatme: the win stuff is for a lottery that awards an imaginary wd-40 can, the number is the chosen ticket. 03:01
03:01 noganex left 03:02 skids joined
BenGoldberg Well, Since my imagination is full of bubble gum, I could definitely use an imaginary can of WD-40, so... 03:07
/win
/win3
buggable BenGoldberg, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 20 hours, 52 minutes, and 39 seconds
lookatme /win99
buggable lookatme, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 20 hours, 52 minutes, and 21 seconds
BenGoldberg Actually, I'd want a *different* ticket number, to avoid sharing my imaginary spray lube.... 03:08
/win42
buggable BenGoldberg, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 20 hours, 51 minutes, and 37 seconds
raschipi g'night 03:09
03:14 raschipi left
awwaiid Sorry I forgot to send a reminder -- tomorrow is the ICFP Contest, which is specifically designed for people to show off their favorite language / programming tech. icfpcontest.org . This is a 72 hour marathon challenge, starting at 12:00 UTC (in 8 hours 40ish minutes) when the problem will be announced. You can participate with any team in any location 03:19
03:25 Cabanossi left 03:26 Cabanossi joined 03:34 pharv joined 03:52 skids left 03:54 Cabanossi left 03:55 Cabanossi joined 03:58 pharv left 04:01 cpage_ left 04:05 BenGoldberg left 04:06 astj left
zengargoyle lazyirc: do we have flock and compress/gzip like IO? 04:10
04:12 vendethiel joined
zengargoyle /win666 04:23
buggable zengargoyle, Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 6 days, 19 hours, 36 minutes, and 48 seconds
04:43 lizmat left 04:46 cpage_ joined 04:56 RockyBalboa joined 04:57 lizmat joined 04:59 astj joined
RockyBalboa Hello, world! I have a quick question. If I do that: gist.github.com/anonymous/56ca994d...72fa55c069 ; will the child sub be defined each time I call the parent sub (and that may lead to poor performances)? Or only at compile time? Thx for your help. :) 05:01
zengargoyle yes i would expect. 05:03
new sub each time.
RockyBalboa Ok! Thank you very much, zengargoyle :) 05:04
zengargoyle i think...
method-y might have variables from method-x.... 05:06
sub method-x($a) { sub method-y { $a ~ "hello" }; return "say" ~ method-y } 05:07
RockyBalboa Yes, I think your are right then, there is obviously some context involved here, so the sub is redefined each time. (Ohhh and btw, these are not methods, sorry, I'm just starting Perl 6). 05:08
zengargoyle there are ways to make method-y not be created each time, but depends on your usage....
maybe perl6 is smart enough to know that it doesn't change and doesn't make new sub each time.... not sure... 05:09
usually... sub x { return sub y {} } means you want y to change each time... 05:11
RockyBalboa And if I do EVALFILE a few times on the same file that doesn't change, will the file be evaluted each time from scratch or is there some kind of caching? 05:13
05:15 xtreak joined
zengargoyle m: sub x($a) { state $x = 0; sub y { "world" ~ $x++ }; return "hello" ~ y }; say x(1), x(2); 05:16
camelia helloworld0helloworld1
RockyBalboa Ohhh that's nice this state variable! :) 05:18
It amazes me all these features in Perl6!
05:19 bwisti left, labster left
zengargoyle m: sub x($a) { state $x = 0; state &y = sub { "world" ~ $x++ }; return "hello" ~ y }; say x(1), x(2); 05:20
camelia helloworld0helloworld1
zengargoyle expected that to work.. :P 05:21
zengargoyle has no clue how to make method-y only once.... :/ 05:24
i think maybe it is only compiled once, but it has access to the lexical scope.... so it's the same sub { ... $x++ ... } as a compiled thing, but the $x++ is still the outside thing.... 05:26
RockyBalboa What is this amperand &y sigil here? A kind of block/lamda var? 05:27
zengargoyle m: sub x($a) { my $x = 0; state &y = sub { "world" ~ $x++ }; return "hello" ~ y }; say x(1), x(2);
camelia helloworld0helloworld1
zengargoyle ouch
m: sub x($a) { my $x = 0; &y = sub { "world" ~ $x++ }; return "hello" ~ y }; say x(1), x(2); 05:28
camelia 5===SORRY!5=== Error while compiling <tmp>
Unsupported use of y///; in Perl 6 please use tr///
at <tmp>:1
------> 3 { "world" ~ $x++ }; return "hello" ~ y 7⏏5}; say x(1), x(2);
zengargoyle m: sub x($a) { my $x = 0; my &y = sub { "world" ~ $x++ }; return "hello" ~ y }; say x(1), x(2);
camelia helloworld0helloworld0
zengargoyle m: sub x($a) { my $x = 0; first my &y = -> { "hello" ~ $x++} ; return "hello" ~ y }; say x(1), x(2); 05:29
camelia hellohello0hellohello0
RockyBalboa Thanks zengargoyle, I am going to study the doc in depth. I just started Perl 6 yersteday, so I need more work to understand the synthax (&, -> and so). :) 05:31
zengargoyle RockyBalboa: it's the wrong time of the day for experts.... you would probably get a better answer in a few hours... 05:32
RockyBalboa Ok lol. Are they all sleeping? :-P
zengargoyle it's 05:32 UTC and most experts are in UTC-ish timezone and asleep.... :P
RockyBalboa Where are U form?
zengargoyle west coast USA. 05:33
it's only 22:33
but east coast people are asleep too. :)
RockyBalboa Ohhh ok! :) Here 07:33 Switzerland time
zengargoyle ah, yes, you'll have better help in a few hours... :) 05:34
RockyBalboa Yeah sometimes I like to watch NHL playoff hockey, and I can only follow the western games, because of time zone :-P
zengargoyle i'm just here to play and fill up the logs.
RockyBalboa Ohhh I see lol 05:35
So have a good night, thx for you help. CU later maybe :)
zengargoyle and tell people it doesn't really get rolling for a few more ours.. best of luck.
05:36 lowbro joined, lowbro left, lowbro joined 05:37 andrzejku left
zengargoyle m: sub x($a) { my $x = 0; constant &y = -> { "hello" ~ $x++} ; return "hello" ~ y }; say x(1), x(2); 05:38
camelia hellohello0hellohello1
zengargoyle m: sub x($a) { my $x = 0; constant &y = -> { "world" ~ $x++} ; return "hello" ~ y }; say x(1), x(2); 05:39
camelia helloworld0helloworld1
zengargoyle m: sub x($a) { my $x = 0; once my &y = -> { "world" ~ $x++} ; return "hello" ~ y }; say x(1), x(2); 05:40
camelia Cannot invoke this object (REPR: Uninstantiable; Callable)
in sub x at <tmp> line 1
in block <unit> at <tmp> line 1
zengargoyle m: sub x($a) { my $x = 0; my &y = once -> { "world" ~ $x++} ; return "hello" ~ y }; say x(1), x(2);
camelia helloworld0helloworld1
zengargoyle grrr
m: sub x($a) { my $x = 0; my &y = -> { "world" ~ once $x++} ; return "hello" ~ y }; say x(1), x(2); 05:42
camelia helloworld0helloworld0
RockyBalboa You can eval code with camelia bot ?! OMG 05:43
m: say "Hello, World!";
camelia Hello, World!
RockyBalboa I guess we better not try some infinite loop :P 05:44
zengargoyle BuildTheRobots: really, i think it is 'compiled' once as far as work goes, but has access to the lexical scope...
so no extra work, it is not compiling method-y each time... only once. but in the compiled code, variables may be outside.... 05:45
it's not an optimization sort of thing you need worry about... 05:46
part of perl 6 is that the parser is one pass only. 05:47
except for thing like EVAL i guess.
m: say 1 ... Inf
camelia (...)
zengargoyle m: say flat 1 ... Inf 05:48
camelia (...)
zengargoyle m: say flat 1 .. Inf
camelia (...)
zengargoyle m: .say for 1 .. Inf
zengargoyle hopes it times out.....
camelia (timeout)1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51…
RockyBalboa Thx! That's nice! :) 05:49
zengargoyle they make the bot safe enough.... there are many things you can't do... :)
and if you break it, it's probably a bug they want to know about.... 05:50
RockyBalboa We can "use MyModule;" in an evaluated file, so it seems compiled too, not just interpreted at runtime. 05:51
zengargoyle yes, there is no interpreter...
RockyBalboa LOOL. Poor Camelia. :P
zengargoyle this is why the `perl6` REPL is sort of broken...
each line you type is a new { .... } scope that is compiled and then executed. 05:52
RockyBalboa Ok! Sorry, I am used to Ruby, I'm just starting to understand Perl. But I really like all the stuff in there so far! That's amazing all we can do. Ruby is much more narrowed.
zengargoyle yes, Perl 5 or 6 rocks! :) (we're so much better.... ) 05:53
zengargoyle is biased.
you will like it more and more. 05:54
and wish it was a bit faster and a bit more 'production' ready because it's so nice.
RockyBalboa Yeaahh I think so. :) It's very cool! 05:55
I am trying to port a micro homemade web framework from Ruby to Perl 6.
Perl 6 grammar is very cool! I mean it's so easy and fast to build a template engine with that! 05:56
05:57 wamba joined
zengargoyle yes, a bit slow compared to say perl 5 regex or PCRE, but nice and hopefully gets faster. 05:57
i did a strange AI visual box stacking thing using a Grammar because when you have Grammar everything is a Grammar. :) 05:58
RockyBalboa Ruby is kind of slow too, but it's definitly fast enough for the web. How does Perl 6 compare to Ruby right now with performances? 05:59
lookatme :) 06:00
ryu0 And Perl6 leaves it open to the possibility of being ported to other major VMs.
zengargoyle i don't do much web stuff, but i think Ballidor is the 'framework' sort of thing at the moment.
ryu0 I just hope the current major implementation is the one they settle on. 06:01
RockyBalboa Looool. Yeaahhh Grammar is this hammer thing: everyithing look like a nail :P
ryu0 They've gone through a lot over the years.
zengargoyle i am Perl 5 hacker, but i got the memo that Perl 6 would take long time (15 years -ish) but waited and am mostly happy. 06:03
RockyBalboa Yes what a long road :-/
zengargoyle i did not jump to Ruby because it was slow-ish and i saw Matz?? once complain about Rails and popularity and having to maintain a 'used' language. 06:04
ryu0 Me, i'm entirely new to Perl and not too experience with dynamic typing. I've largely used static typed languages to date.
zengargoyle he wrote Ruby for himself and was a bit *something* about other people using it because it means he can't do whatever he wants.
lookatme Perl 6 is quite stable. I have a script( a process manager) running with old rakudo version. It has running well about 2 x 31 x 24 hours.
ryu0 lookatme: hopefully it's more useful than a PHB. =p 06:05
RockyBalboa I never used Perl 5, started with C/Java/PHP, then Ruby, and here I am (I am a mostly a web developer) :P
zengargoyle lookatme: i ran into everything that changed.... 06:06
ryu0 RockyBalboa: how much did you use C? It is the main language I am adept with.
zengargoyle i've had to change all 3 of my modules because of changes to certain things.... :) 06:07
lookatme :) 06:08
ryu0 p6: say [+] "a", "b"
camelia Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5a' (indicated by ⏏)
in block <unit> at <tmp> line 1
zengargoyle but it gets better and better, so i'm mostly happy camper. 06:09
ryu0 Huh. Thought it would concatenate.
zengargoyle m: say [~] <a b>
camelia ab
lookatme yeah, it must be!
RockyBalboa For example, a 30 lines template engine I use in my web framework written in Ruby: gist.github.com/anonymous/5ba0f6b9...0fab07ecd6 I will try to port my stuff to Perl 6 and publish it
zengargoyle ~ is string concat... + is number...
ryu0 Ah, so that's it. D uses the same syntax for appending strings.
lookatme RockyBalboa, Perl 6 now is strong type language 06:10
ryu0 Made more sense to me since concatenation is not really like addition.
RockyBalboa @zengargoyle: ohhh I just have the basics in C, not much an expert :P
ryu0 1, different types. and 2, it's not commutative.
p6: say ("a" ~ "b") == ("b" ~ "a") 06:11
camelia Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5ab' (indicated by ⏏)
in block <unit> at <tmp> line 1
ryu0 ... then what's equality?
zengargoyle eq
ryu0 Oh, it's a keyword.
How pascalish.
p6: say ("a" ~ "b") eq ("b" ~ "a") 06:12
camelia False
ryu0 say (1+2) eq (2+1)
evalable6 True
ryu0 ... Huh.
zengargoyle m: say ("a" ~ "b") cmp ("b" ~ "a")
camelia Less
RockyBalboa m: 'hello' eq 'hello'
camelia WARNINGS for <tmp>:
Useless use of "eq" in expression "'hello' eq 'hello'" in sink context (line 1)
zengargoyle m: say so ("a" ~ "b") cmp ("b" ~ "a")
camelia True
zengargoyle eh...
m: say so ("a" ~ "b") === ("b" ~ "a")
camelia False
zengargoyle m: say ("a" ~ "b") === ("b" ~ "a") 06:13
camelia False
ryu0 do any perl6 implementations support compiling standalone executables yet?
lookatme ryu0, no
zengargoyle ryu0: not to my knowledge... and i ask the same frequently. :)
ryu0 huh. Hope it gets it one day.
zengargoyle it's more that MoarVM is still changing so much and nobody has the time. 06:14
but it is on the roadmap and should be relatively easy.
ryu0 It's not a priority for what I intend to do so i don't really care but it would add more options.
RockyBalboa Just a few basics questions. Indentation: 4 spaces? Naming: kebab-case? File naming: .p6, .pm6 ? I think it would be cool to have some kind of conventions guideline
ryu0 zengargoyle: so what options exist for exposing C APIs to rakudo? 06:15
zengargoyle no indentation arguments. kebab-case since we can. Camel::Case for modules, UPPER for special things, i use pl6 and pm6 just to make vim do the right thing. 06:16
lookatme I prefer .p6 amd .pm6
and
zengargoyle ryu0: it's called NativeCall
ryu0 Ah. 06:17
thanks.
zengargoyle and it's pretty simple sub foo () is native ('libfoo.so', :version(1)
but not even that hard, but not even that simple.... :)
RockyBalboa Ok thanks lookatme! Yes, I found too that Vim didn't synthax highlight .pm files, only .pm6 06:18
lookatme Hmm, I use emacs, don't know that 06:19
zengargoyle ryu0: i have a module that builds C code into a .so and includes it in the install and calls function.....
lookatme It's quite easy use NativeCall make a binding for C library 06:20
RockyBalboa We have for example, DBIish ".allrows" method, why not ".all-rows" ? It sometimes confusing having kebacase or singlewordcase? And if you had some snake_case naming, it quickly confusing. :-/
zengargoyle Algorithm::Trie::libdatrie -- it not the best, but if i can manage... it's not too hard... :)
yeah, everybody has their own naming by habit, Perl 5 does not do kebab-case so some do snake_case out of habit. 06:22
or finger memory... :)
i cringe if a class is not Camel::Case
but also 'lowercase' package/module/namething are technically reserved for internal use. 06:23
for pragmas like 'use foo;' 06:24
06:24 nattefrost joined
RockyBalboa Yes, but overall, all the core libs are very consistent CamelCase for class and kebab-case for the rest. Pyhton API is much more a mess 06:24
zengargoyle and all UPPER are sorta reserved for special things like 'use MONKEY-SEE-NO-EVAL'
RockyBalboa Yes, like use lib::MyClass? It will look on my local lib dir? 06:25
zengargoyle eww, no.. 06:26
./lib is just for organizing.... then you do 'use MyClass;' and run with `perl6 -Ilib` to include the lib.... 06:27
RockyBalboa Ok! :)
zengargoyle lib is not part of the name.... if you install your program... it will be MyClass and not lib::MyClass....
it works... but not a good idea and doesn't work for long.... :)
RockyBalboa Ok lol thx! :) 06:28
zengargoyle that might be leftover Perl 5 module layout thing.... ./bin ./lib ./t are just where everybody and every tool expects things to go.
RockyBalboa So I am just starting this Perl 6 Fundamentals book :P
See U later, have a good night zen, thx again! 06:29
Ok /bin, /lib, /t !
zengargoyle RockyBalboa: check out App::Mi6
it is developer tool that makes directories, runs tests, makes thing 10x easier than doing it yoursefl. 06:30
RockyBalboa Oh and another cool stuff in Perl 6: multi sub! :)
Ok thx, I will add that to my stuff :P
06:30 nige joined
zengargoyle have a nice day. 06:31
06:31 nige is now known as n1ge
RockyBalboa U2 :) 06:31
Or night :P
zengargoyle lookatme: where in the world are you? 06:33
n1ge have been trying to get rakudobrew + zef + Bailador to work well together - I can't seem to get a combination that is working ... does anyone know of a combination of versions that work from a scratch install on linux? 06:34
I needed to use TWEAK - which was the motivation for upgrading rakudo / moarvm / zef - and I'm currently on day 3 of yak shaving 06:36
06:37 szabgab joined
zengargoyle i have never used Bailador so not sure... 06:37
szabgab hello there! 06:38
n1ge hi :-)
zengargoyle woot
szabgab There seem to be some memory leaks in Bailador I've started to track down 06:39
I wonedr if this makes any sense: www.mail-archive.com/perl6-users@pe...04118.html
zengargoyle oh, ask szabgab i think.
n1ge yes - I saw the memory leaks message
zengargoyle yeah, zef install Bailador fails for me....
n1ge yes - it's a bit of a bummer 06:40
I was running it with ubic (perl5) - which would usefully keep the Bailador app up and running
zengargoyle it's like the weather... wait a few days and it will change... :P 06:41
szabgab zengargoyle: have you opened an issue with the output of the failure?
n1ge so it was just about managing - to do the job
06:41 nadim_ joined
zengargoyle szabgab: i just now tried for the first time... much checking before complaining. :) 06:42
n1ge indeed - normally I'm happy to wait (key trait of a Perl 6 early adopter) 06:43
szabgab :) We seem to keep bumping into various installation issues, but I hope we are solving them. (Actually I am only talking others are doing the solving :)
zengargoyle i've never used it before and i'm probably on recent-enough rakudobrew built perl6 and new-ish zef but still ....
06:43 RockyBalboa left
zengargoyle szabgab: ah, i just thought i remembered your name coming up in regards to Bailador... :) 06:44
szabgab The thing is that I even see the Travis flucktuating between success and failure in a rather unpredictable way. A change to the README file and the report might change. 06:45
zengargoyle have you talked to ugexe? 06:46
lookatme zengargoyle, Hmm?
szabgab anyway, if there are issues with Bailador I can try to help here, but I was mostly interested in that memory usage code.
nadim_ .tell sena_kun can you please try the zef install Data::Dump::Tree again? Should work now. 06:48
zengargoyle i'll try install Bailador again with verbose and debug and maybe file bug if i can figure out where-ish.
n1ge thanks Gabor - generally Bailador was working OK - ubic did help with keeping it alive
thanks zengargoyle will give that a go 06:49
zengargoyle there has been a high variability in latest rakudobrew + zef + things....
n1ge (before I give go a go ;-) )
lookatme zengargoyle, I'm in China. <- you mean this ? 06:50
zengargoyle and i think that why some try to point people to lastest YYYY-MM release ... :)
lookatme: yeah, timezone... i see you more than others... just curious..
i tend to be very late nigth and catch opposite sides of the world. :) 06:51
06:52 darutoko joined
lookatme zengargoyle, Oh, it's 2pm here 06:52
n1ge Gabor - there are a couple of things on the bailador wishlist 06:54
06:54 Cabanossi left, nadim_ left
n1ge I've missed some things from Mojolicious - the flash and easily moving controllers into other files (did this with roles in the end) 06:54
zengargoyle Bailador.... t/20-cli.t Failed test: 6 Non-zero exit status: 1 06:55
szabgab that t/20-cli is known to be problematic
zengargoyle at t/20-cli.t line 127 06:56
n1ge - it does the trick for me mostly - as the backend perl5 service talks to the db etc
06:56 Cabanossi joined
szabgab try zef install --serial Bailador 06:56
zengargoyle ah, for me zef fails because that.
06:56 rindolf joined
zengargoyle one moment. 06:56
szabgab I have plenty of time. I am tying to install the Redis module now :) 06:57
zengargoyle so it's deps inside of the temp CuR? still waiting... 06:58
n1ge ok - trying zef install -serial Bailador now
szabgab doubldash ! 06:59
n1ge yep - got that
zengargoyle oh, try zef --deps-only Bailagor; zef install Bailador...
ah, too late. :P 07:00
szabgab that would probably also a good idea
n1ge will try that next ;-)
szabgab what zengargoyle wrote
but I guess those two do effectively the same
n1ge ok trying zen version now 07:01
zengargoyle yep, installed: 1 bin/ script [bailador] installed to
well, serial is each at once as one thing, --deps-only would still be all of them at once but just not the Bailador code.
depends on where the actual problem is i guess..... 07:02
this is probably a ugexe thing.
szabgab as far as we understood it is that in the t/20 we ru perl6 exeternally and that command does not see the not-yet installed dependencies
and this seems to be new in zef 07:03
and I think it has been fixed in the GitHub of Bailador.
zengargoyle szabgab: makes sense... 07:04
szabgab Get ugexe release it :)
zengargoyle gist.github.com/86ca199c8fc77757d4...690b4ee487
install log i hope.
i got :ver<0.0.9>:auth<>:api<> 07:05
szabgab looks good to me :) 07:06
zengargoyle i wonder if zef prefers cpan and git does HEAD and somebody didn't bump versions... (if it's fixed in git) 07:07
i could see zef pulling 0.0.9 from cpan instead of git when 0.0.9 in git is diferent because ..... 07:08
szabgab and I see I am mixing up nicknames again.. 07:11
apparently it is not fixed in git, but zef prefers cpan, and it would not see our fix anyway because we use a dev branch 07:12
zengargoyle gah, uninstall and install from git works, but dependencies already installed now. :P 07:13
szabgab but there is no fix in git yet and maybe it is not even the problem in Bailador
zengargoyle oh, nm. :P
szabgab Now I am confusing myself as well :)
The issue is clearly that t/20-cli does not see the dependencies 07:14
Unless they were really installed earlier
zengargoyle like using $^X in perl6 right?
07:15 andrzejku joined
szabgab The only question why and who should fix this 07:15
$^X is not $*EXECUTABLE but yes
s/not/now/
zengargoyle i think i get it.
but yeah, not my fixability. :) 07:16
and the answer for the moment is --serial
or mayby (--deps-only; Bailador) 07:17
n1ge ec2 instance chugging along in the background - doing the zef install Bailador set now
step
zengargoyle really needs to get my lxc instances back up and ready for perl 6 stuff... 07:19
i would like to see the --deps-only case vs --serial. my understanding is that --serial negates some mulitprocessing magic. 07:22
i.e. --deps-only would use all your cores while --serial is just one at a time.
07:24 xtreak left 07:25 xtreak joined 07:29 robertle_ left
n1ge hmm - no worky - pastebin.com/AGH94zd1 07:29
07:29 xtreak left
lookatme It's test failed ? 07:31
zengargoyle wow, All candidates are currently installed.
how is `zef install --deps-only Bailador` come back with that? 07:32
still fails though... what's the `perl6 -v` on that box? i only know travis which builds *everything* from scratch. 07:37
07:38 http_GK1wmSU joined
n1ge This is Rakudo version 2017.07-126-g86cb136 built on MoarVM version 2017.07-293-g3f33a84 07:38
implementing Perl 6.c.
zengargoyle hehe, i'm 2017.07-54-g640641876 07:39
07:40 http_GK1wmSU left
zengargoyle you *might* want to try 2017-07 tag and stick to the release rather than building the absolutely latest if you don't know you need the absolute latest. 07:41
07:41 nadim joined
zengargoyle unless you're testing an just willing to report broken... picking the rakudo of the day is a rough gamble. 07:42
n1ge yes - zen - to be honest I've tried lots of different versions 2017.01, 2017.05, 2017.07 etc 07:43
zengargoyle those errors look a lot like network/firewall problem. 07:44
expected a <h1> title, got nothing.
n1ge I probably should have kept a log of failures on different versions / combinations - made various bug reports along the way though
zengargoyle there is no compile error or start error or load failure but a failure to get a '<h1>Bailador App</h1>' response. 07:46
07:46 ChoHag joined
zengargoyle i would think that the test can't talk to the server from that. 07:46
unless that testing is not a server but more direct access through an instance... (no network involved) 07:47
n1ge ok - that's possible 07:48
it seems to install with the serial flag 07:49
zengargoyle that's just my guess... :)
or what i would be looking at first.
n1ge although it noted: WARNING: unhandled Failure detected in DESTROY. If you meant to ignore it, you can mark it as handled by calling .Bool, .so, .not, or .defined methods. The Failure was:
in any event starting the app causes this ...
===SORRY!=== 07:50
P6M Merging GLOBAL symbols failed: duplicate definition of symbol Response
07:50 szabgab left
zengargoyle that sounds like two modules both try to do someting like class Response is export without using namespaces. 07:52
n1ge yes - I need HTTP::UserAgent - I think there may be a clash there 07:53
the Bailador app uses that to talk to the backend REST service
07:55 baest left
zengargoyle yeah, my fuu would be download all the things and grep for Response and complain loudly at whomever dediced to export that name unqualified. :P 07:55
07:56 cschwenz joined
n1ge yep :-) 07:57
zengargoyle i feel your graar.
n1ge think I need to wait for a change in the ecosystem weather 07:58
zengargoyle if it's doable, you could 'need HTTP::UserAgent' and maybe use long names...
07:58 baest joined
zengargoyle avoid the export/import bits. i'm not sure this would work... 08:00
n1ge yes - that could work - thanks for your help zengargoyle - much appreciated
Geth ecosystem: 95a3903c17 | moznion++ | META.list
Add HTML::Escape
ecosystem: be1d0eae6c | (Jonathan Stowe)++ (committed using GitHub Web editor) | META.list
Merge pull request #357 from perl6/html-escape

Add HTML::Escape
08:00 sena_kun joined
zengargoyle good luck. :) 08:00
and then complain. :P 08:01
n1ge 17 years waiting - what's a few days - I'll be back ;-)
zengargoyle find where and complain or it won't get fixed. 08:02
squeaky wheel gets the grease.
n1ge yep - that too
08:04 ijneb left 08:07 n1ge left 08:10 Cabanossi left 08:11 Cabanossi joined 08:13 zakharyas joined 08:24 dustinm` left, nige joined 08:30 vendethiel- joined, nige left 08:32 vendethiel left, dustinm` joined 08:34 astj left 08:35 astj joined
nadim ugexe: your help would be appreciated for this one github.com/jonathanstowe/Test-META...-319090770 08:38
08:43 mcmillhj joined 08:46 nige joined 08:48 mcmillhj left, vendethiel joined 08:49 vendethiel- left 08:50 ufobat joined, nige left 08:53 nige joined 08:54 Cabanossi left 08:55 xtreak joined, RockyBalboa joined 08:56 Cabanossi joined 08:57 konsolebox left, edehont joined
RockyBalboa Hello! Two quick questions :) 09:00
1. Is there a sub to unaccent a string? Like converting from "âé" to "ae"?
2. How can I interpolate a string when there is no space around? my $var = "hello"; say "$helloworld"; # doesn't work
ryu0 hm. 09:02
ufobat didn't get what you mean in 2 09:03
ryu0 Thought so.
m: my $var = "hello"; say "{$var}world" 09:04
camelia helloworld
ufobat m: $hi = "foo"; say "$hi"
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$hi' is not declared
at <tmp>:1
------> 3<BOL>7⏏5$hi = "foo"; say "$hi"
ufobat m: my $hi = "foo"; say "$hi"
camelia foo
ryu0 ufobat: his question is related to a similar problem in shell scripts.
ufobat your reply makes sense :)
ryu0 Turns out perl6 uses a similar construct.
ufobat or code
ryu0 ${var} in shell; {$var} in perl6 09:05
09:05 sena_kun left
RockyBalboa Perfet! Thx a lot ryu0, I tried ${var} and it failed, but your 09:05
but your solution works :)
ufobat i've got a LibCurl question, what is the equivalent for 'curl -k'
ryu0 RockyBalboa: i got the hint from the REPL output: Unsupported use of ${var}; in Perl 6 please use {$var}
maybe it was valid in perl5. 09:06
09:06 konsolebox joined
RockyBalboa OMG I didn't even read the error message => lesson learned :) 09:06
ryu0 I also found out the think perl book is free if you don't mind using the digital version. 09:07
greenteapress.com/wp/think-perl-6/
i assume this is the book website.
RockyBalboa Thanks ryu0! It will help 09:08
ryu0 it's funny. sometimes amazon tries to sell you a kindle version of a book that the author has released for free.
moritz and why shouldn't they try? 09:09
ryu0 shrugs.
RockyBalboa I bought "Perl 6 at a Glance" and "Perl 6 Fundamentals": they are nice too. "Think Perl 6" seems more for people that are new to programming 09:10
09:10 pecastro joined
ryu0 maybe so. 09:10
i've always preferred books that take time to explain the language.
even though i'm no longer a beginner.
moritz RockyBalboa: thanks for buying my book! :-) 09:11
RockyBalboa: It would be very helpful if you could write a review on amazon
RockyBalboa Ohhh Moritz is here! :))) Thanks for you work! I'm just starting Perl 6 today (and working on some Sudoku too lol). I will write a review yes. :) 09:14
ryu0 moritz: really? which is yours?
zengargoyle m: say "âé".samemark('a')
camelia ae
RockyBalboa Perl 6 Fundamentals
moritz ryu0: what RockyBalboa said 09:15
zengargoyle m: say samemark(""âé", 'a')
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in argument list; couldn't find final ')'
at <tmp>:1
------> 3say samemark(""7⏏5âé", 'a')
expecting any of:
infix
infix stopper
zengargoyle m: say samemark("âé", 'a')
camelia ae
zengargoyle RockyBalboa: Q1 ^^ 09:16
09:16 xtreak left
moritz I'm also writing leanpub.com/perl6regex right now 09:16
at about 80 pages A4 now 09:17
Altreus Pretty sure I bought one of these books but I have no idea where it is :(
moritz (and negotiating with apress if they publish it)
ryu0 RockyBalboa: so you went to the applied books?
RockyBalboa: what's the benefit there? i usually prefer reading about my tools before trying to use them. 09:18
moritz not everybody wants to read 200 pages about how variables, loops, condtionals and subroutines work if they already know that from other languages
but somebody new to programming needs those 200 pages 09:19
so IMHO it makes sense to have different books for different audiences
ryu0 I see.
i always thought those were meant for people already familiar with the language. 09:20
RockyBalboa Thx a lot zengargoly! Not sleeping yet? :)
moritz ryu0: at least "Perl 6 Fundamentals" is targeted at programmers who are familiar with other languages
I've made perl6book.com/ to explain the differences :-) 09:21
zengargoyle RockyBalboa: my internal clock is weird.
ryu0 Maybe my mindset is from having to use C where language subleties can make or break your program in major ways.
RockyBalboa Yes, like Morirtz said, I didn't want an introduction to OOP and SO. "Just bring PERL 6" :)
moritz: Ok nice, I bookmarked your book about regexes :) 09:22
ryu0 There was a lot to learn that wasn't even covered in the ANSI C book that I remember.
Mostly related to machine architecture. 09:23
zengargoyle and since we don't have a translation army, a multitude of examples can be better than a paragraph of english for some....
RockyBalboa i just skimmed through "Think Perl 6": it looks cool too, more like a language reference yes 09:25
ryu0 i've had people complain about 400 page books being too long, when i've seen ones for C++, C#, or Java that can easily exceed 600.
zengargoyle has 1978 edition of K&R C Programming sitting on shelf. :P 09:26
RockyBalboa "Give me six hours to chop down a tree and I will spend the first four sharpening the axe."
ryu0 "And one of those remaining hours will be spend ensuring I learn how to not cut off my hand."
RockyBalboa :)))) 09:27
ryu0 C joke.
RockyBalboa: bored? 09:28
if you're interested i can entertain you with some C language trivia.
AlexDaniel for some reason I'm always using {$foo} instead of ${foo}
RockyBalboa A lot of C people in there! :) 09:29
AlexDaniel I wonder if it is significantly slower
ryu0 int x <:50:> = {};
Is this valid C? If so, what is it using?
RockyBalboa And you guys, do you like C++?
ryu0 Not really, but it's tolerable.
lookatme I'm C++ developer 09:30
zengargoyle i skipped it because when it was new it was just a bunch of pre-processor wrapper around plain C.
RockyBalboa I don't have enought knowledge, I come from web dev (PHP/Ruby/JS) and now... Perl 6
ryu0 C++ wasn't standardized until 1998.
zengargoyle hah, standard. 09:31
ryu0 For what that's worth.
zengargoyle yeah, C++ was a thing at least in early 90's
it just wasn't much more than macros/pre-processing.
ryu0 Anyone care to try to answer my C trivia question? ;) 09:32
zengargoyle an int x[50];
?
ryu0 yes, it's an array. 09:33
Can you name what feature I used?
RockyBalboa I had to learn a bit of FORTRAN for computation. That one is old too :)
zengargoyle probably NULL.
trigraph
ryu0 digraph actually, but close.
zengargoyle <: == [ :> == ]
ryu0 trigraph is when the sequence is 3.
digraph when the sequence is 2.
zengargoyle same diff. :P
ryu0 <: = [
09:33 TEttinger left
ryu0 :> = ] 09:33
RockyBalboa lol
ryu0 Also applies to C++ since it inherits from C. 09:34
zengargoyle it's C version of Texas.
ryu0 but in practice you shouldn't be using these. they just make code harder to understand.
zengargoyle wait until you get a load of p6 unicode. :) 09:35
ryu0 ... it makes p5 seem readable? =p 09:36
zengargoyle nah, p6 has much more spaces. 09:37
ryu0 i noticed some of the perl6 examples regarding REs look like vim/sed/ed style patterns.
lookatme What would happened when I call slurp-rest on Proc.out after Proc exit with -1?
zengargoyle and no short $X vars... makes one-liner's in p6 a pain. :P
RockyBalboa Hey what do you suggest as a good IRC client for terminal?
ryu0 RockyBalboa: what are you using now?
Ah? The mozilla IRC client extension? 09:38
zengargoyle i use irssi, but no reason.
RockyBalboa I am using the chat from per6.org... Last time I used IRC, it was like in 1999 with mIRC. I didn't know IRC was still in use!
ryu0 IRC isn't dead by a long shot.
RockyBalboa That a good surprise yes! 09:39
ryu0 hexchat is a decent cross-platform GUI irc client.
successor to xchat.
09:39 vendethiel left
RockyBalboa IRSSI! Oh that's exactly what I was looking for! 09:39
ryu0 but if you really want command line, i'd suggest weechat probably.
RockyBalboa Yeah something like Vim but for IRC
ryu0 it's a bit friendlier to people not used irssi's minimalism.
zengargoyle lookatme: i'd guess Nil or an Exception. 09:40
ryu0 weechat has more whistles out of the box like a nick list.
RockyBalboa Ok I will try them thx
ryu0 irssi doesn't have that setup by default.
09:40 nowan left, vendethiel joined
ryu0 also, weechat comes not only in an ncurses version but also in gui ones too. 09:41
09:41 melezhik left
ryu0 anyway. 09:41
zengargoyle run('foo', :out).out.slurp-rest.perl.say
RockyBalboa Ok thx! I check that. It's cool to see there are all these guys on IRC (and not on Facebook) 09:42
ryu0 Facebook... bleh.
zengargoyle m: run('foo', :out).out.slurp-rest.perl.say
camelia run is disallowed in restricted setting
in sub restricted at src/RESTRICTED.setting line 1
in sub run at src/RESTRICTED.setting line 14
in block <unit> at <tmp> line 1
RockyBalboa :)
zengargoyle doh! it's ""
ryu0 p6: "Beetlejuice!" x 3 09:43
camelia WARNINGS for <tmp>:
Useless use of "x" in expression "\"Beetlejuice!\" x 3" in sink context (line 1)
ryu0 aww.
p6: say "Beetlejuice!" x 3
camelia Beetlejuice!Beetlejuice!Beetlejuice!
ryu0 'lol
cheesy example.
09:44 nowan joined
RockyBalboa m: my $username = 'RockyBalboa'; say "You should get back to work, {$username}!"; 09:45
camelia You should get back to work, RockyBalboa!
RockyBalboa CU later guys, thx :)
AlexDaniel m: say "You should get back to work, %*ENV<ME>!";
camelia You should get back to work, AlexDaniel!
RockyBalboa Haha nice Alex!
ryu0 m: say "Help me, AlexDaniel. You're my only hope." 09:47
camelia Help me, AlexDaniel. You're my only hope.
ryu0 lol
AlexDaniel already has his superman suit on 09:48
zengargoyle lol, funny story since it's late. i knew somebody who called tighty-whiteys "dun-dun-dun's" because superman. 09:50
09:50 nige left, RockyBalboa left
lookatme zengargoyle, seems like it hang on 09:55
I'll do more test later
10:01 ijneb joined 10:02 lookatme left 10:09 Cabanossi left 10:11 Cabanossi joined
zengargoyle .tell lookatme also try just .slurp -- i think .slurp-rest and .slurp have merged into the same thing.... i know i used slurp-rest once upon a time and was told to use just slurp... 10:12
AlexDaniel where's yoleaux ? 10:13
zengargoyle .tell lookatme but Proc is something i haven't played with much so...
10:14 prabh joined
prabh Hi 10:17
How can I contribute in perl coding?
AlexDaniel prabh: start using it for your projects :) 10:18
10:18 vendethiel left
prabh ok 10:19
zengargoyle ask questions and write docs.
10:19 jaffa4 joined
Zoffix blog about it 10:21
jaffa4 hi all.. what is the state of Perl6? 10:22
zengargoyle needs a blog.
Zoffix fix issues: github.com/issues?utf8=%E2%9C%93&a...r%3Aperl6+ 10:23
zengargoyle needs to just get a VPS.
Zoffix jaffa4: we generally try to be stateless pirates
^_^
jaffa4: "stable and actively developed". Is that what you mean? 10:24
jaffa4 Is it used widely now?
10:25 rindolf left
Zoffix jaffa4: it's less than 2 years old. What would "widely" mean? 10:26
jaffa4 world domaination... number 1
lizmat Rakudo Perl 6 is considered a big language :-)
Zoffix jaffa4: it's not the 90s anymore. I don't think any language can be an absolute number 1 anymore. We have just too many different domains, each with its own leaders 10:28
zengargoyle i would say like early JVM. remember when it was nice and worked but was horribly slow and had rough edges. now it's good. perl 6 is on same sort of path. it's nice, maybe a bit slow and rough, but always getting better and better.
jaffa4 is it as fast as Perl 5 at least? 10:29
Zoffix I don't feel the roughness anymore. At least compared to like a year ago, when any time I tried to write a program bugs happened.
jaffa4: you're very good at asking very vague questions :)
jaffa4 does the same program run in Perl 6 as fast as Perl 5? 10:30
zengargoyle yeah, more tooling. and waiting for certain expected bits to be realized (like maybe native shaped arrays) or such.
Zoffix jaffa4: what task exactly? A few might be faster. Many might be slower. A lot has to do with the languages not being equivalent. Perl doesn't deal in graphemes; we do. You can bench line reading, but you're not measuring the same things
10:30 xtreak joined
Zoffix jaffa4: "the same program" is tough to come by. An empty Rakudo program in Perl 5 would need to load dozens of modules. 10:31
jaffa4: but in general, I'd say the answer is "not yet".
jaffa4 yes, I am asking in general
Zoffix jaffa4: there's still work to do to optimize things. And Perl folks aren't standing still either. A new grant to work on Rakudo's stability in performance was just approved, so there's at least 200 hours of work yet to be done :) news.perlfoundation.org/2017/07/gra...perl-.html 10:32
s/stability in/stability and/; 10:33
prabh What are the cool features of perl6 ?
10:33 Slaash joined
Zoffix m: sub infix:<¯\(°_o)/¯> { @_.pick }; say 'Coke' ¯\(°_o)/¯ 'Pe 'Pepsi'; 10:34
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3 { @_.pick }; say 'Coke' ¯\(°_o)/¯ 'Pe '7⏏5Pepsi';
expecting any of:
infix
infix stopper
postfix
statement …
Zoffix bah. My broken terminal ruins everything :)
m: sub infix:<¯\(°_o)/¯> { @_.pick }; say 'grammars' ¯\(°_o)/¯ 'ease of concurrency';
camelia grammars
zengargoyle make it good, make it work, make it work fast. it's mostly good in idea. it mostly works in practice. it gets faster every time i look.
10:35 xtreak left
Zoffix prabh: my favourite "cool" feature is how consice the language is without going into the unreadable code golf territory. Compared to Perl, some of my programs I converted ended up half the size 10:35
concise
zengargoyle there is a lot of do the simplest thing that works correctly and then go back and do more complicated thing that works correctly but is faster. 10:36
Zoffix There's even a poster on the subject: raw.githubusercontent.com/perl6/ma...300ppi.png
prabh: another is I somehow no longer have Unicode problems I tend to have in other languages 10:37
'cause Rakudo was designed with Unicode in mind from scratch, whereas in many popular langs of today it's mostly a bolted-on hack 10:38
zengargoyle heh, yeah, what's a Unicode problem?
El_Che zengargoyle: it's like a Unicorn but less cute
Zoffix m: await ^4 .map: { start sleep 1 }; say now - INIT now 10:39
zengargoyle biggest unicode problem i have is terminal fonts. :P
camelia 1.00311816
zengargoyle and typing the **** things. 10:40
Zoffix prabh: and that's my other favourite feature ^. 4 sleeps of 1 second that took just 1 second to run. Ease of parallelism and concurrency in Rakudo is handy. Something also that was in mind from the start.
prabh: actually, we have a marketing brochure that probably answers your question better than my rambling :D github.com/perl6/marketing/blob/ma...ochure.pdf 10:41
jaffa4 is there perl 6 conference now? 10:42
zengargoyle modules are Name:ver<1.0.0>:auth<github:zengargoye>:api<2> and many modules of same name an different versions and different authors can be used at the same time in different bits of your code....
Zoffix jaffa4: yeah, it's part of The Perl Conference. Latest one in Amsterdam is days from happening (or maybe even started already): act.perlconference.org/tpc-2017-amsterdam/ 10:43
ufobat where are ubuntu rakudo packages? 10:44
zengargoyle use Foo:ver<1.0> in legacy bits of code and Foo:ver<2.0> in new bits of code. use Bar:from<Perl5> and use Perl5 modules, same goes for Python.
Zoffix huggable: deb 10:45
huggable Zoffix, CentOS, Debian, Fedora and Ubuntu Rakudo packages: github.com/nxadm/rakudo-pkg/releases
Zoffix ufobat: ^ there
ufobat rakudo.org is pretty missleading :/
regarding finding the ubuntu package
10:45 Slaash left
Zoffix ufobat: the messenging bot is ded, but I left you a message with a hint for how to work around the resource thing. Apparently there *is* a way to access the meta: irclog.perlgeek.de/perl6/2017-08-02#i_14960004 10:46
ufobat: that site is in need of updates ( rt.perl.org/Ticket/Display.html?id...et-history )
ufobat my coworker just complains he coudnt find a ppa for ubuntu
thanks :)
jaffa4 do the original Perl6 developers work on Perl6 stilll? 10:48
Zoffix jaffa4: some, yeah
Including Larry Wall
zengargoyle ufobat: yeah, i dont' think there's a Debian package whatever either, but the .deb are there and making a full per-distro support their special package manager logic just to get a .deb is a bit much.
Zoffix Tons of new ones too :)
w00t \o/ managed to repro the disappearing appveyor issue on modules.perl6.org locally \o/ 10:49
ryu0 ufobat: there isn't a PPA really, but there are packages.
zengargoyle making a Debian repo just to hold one .deb is a PITA specific work just for one thing.
10:49 Slaash joined
ryu0 I could probably throw together a PPA. 10:50
but not right this moment. 10:51
zengargoyle ppa may be easier... :) i gave up trying to 'repo' the two .debs i make.... not worth the hassle!
ryu0 zengargoyle: yea. but i have a background doing linux packaging. shouldn't be too hard.
zengargoyle cool, better you than me. :P 10:52
ryu0 i had to fiddle with the packages for rakudo awhile to figure out how it fits together. it's not as easy as perl5 is.
maybe that'll be fixed as it matures but you can't really package it directly from rakudo star as it is.
because it insists on writing directly to the root directories during the build. 10:53
and debian packages don't work like that.
Zoffix ewww
I think star isn't what you're meant to package, tho? Package rakudo compiler itself + zef + whatver modules
zengargoyle yeah, there's a lot of automatic building if it's not already there.
doesn't fakeroot handle that?
ryu0 it fakes root, but it still won't let you write to /usr/bin directly. 10:54
and the existing stuff uses hard-coded pathes so best workaround is to split into separate packages.
10:54 enheh joined
zengargoyle eh... it writes to ./fakeroot/usr/bin 10:54
ryu0 ... you don't know what fakeroot does, do you?
it's used to pretend you're root so files in an archive show up a root user or so. 10:55
useful when doing make install and such.
zengargoyle the one time i did try, fakeroot manage to wrangle /etc scripts /usr/bin scripts.
ryu0 i tried though.
zengargoyle ah, maybe rakudo has no DESTDIR like......
i can see it now. 10:56
ryu0 Yes, and rakudo star also tries to build moarvm and pqp
and then it expects them to be installed in the root directories.
See the problem now?
zengargoyle yes, but only if you ask.
10:56 andreoss joined
ryu0 which is the case if they don't already exist as packages. 10:56
regardless i'll give it a spin later. should be doable regardless. 10:57
zengargoyle i only did manual build moar, build nqp, build rakudo as separate steps.... but it worked.
10:58 noganex_ left
ryu0 manual always works, but in an automated environment you usually have to work within a greater set of restrictions. 10:58
zengargoyle there should be a 'once' in there... i did it once. but i don't see a real problem with 3 packages.
ryu0 afaik, rakudo star really isn't designed with this in mind right now. 10:59
i assume it'll get addressed eventually.
10:59 astj left
zengargoyle ah, 'rakudo star'. i was thinking just rakudo iself. 10:59
ryu0 yes.
zengargoyle nm.
ryu0 that's what i tried the first time. 11:00
11:00 noganex joined
ryu0 think i should try to package all the addons rakudo star likes to include? 11:00
i was thinking only zef seems essential.
zengargoyle i sorta think rakudo star is gone. there was once a Task::Star that installed all of the rakudo star modules, but it was deemed a bad idea and removed. 11:01
i don't even know what rakudo star is now.
ryu0 still exists. it's a bundle of convenience. 11:02
rakudo.org/downloads/star/?asset=star
zengargoyle yeah, i just think they used to have a Task::Star bundle such that moar+nqp+rakudo+zef+install Task::Star == rakudo star. 11:03
ryu0 ah.
11:03 markmont left
ryu0 i'm tempted to package them individually and then add a meta package to depend on the components of the whole set. 11:04
zengargoyle and you could do same with moar+nqp+rakudo+zef+ recommended modules ...
ryu0 so people can choose what parts they want.
why not? ubuntu provides free hosting for PPAs at launchpad.net 11:06
zengargoyle in the end... there are a few special modules you might want, debuggers and the like, and a few other JSON or WWW type things, and all the rest are acutally dependencies that would get installed automatically anyways... :)
ryu0 yea, i suppose. i can take requests for extras but i'd just package the foundation first. 11:07
hm. reminds me.
i'll need to split the dev files too. :)
this'll be fun when i get some time. 11:08
zengargoyle i think the choice of useful modules varies widely from person to person, but maybe like Term::ANSIcolor is just something you're eventually going to want, or some other package will want, so might as well install it now.
and Readline for the REPL
ryu0 yea. Linenoise seems silly to install when you got readline available.
zengargoyle i think there are GPL vs BSD license things there but yeah, 11:10
ryu0 true enough, but i doubt it'll matter for just usage.
it's not even part of rakudo's normal distribution. 11:11
kinda like a GPLed plugin if you will.
Geth_ perl6-examples: 0d47b45137 | mattwhisenhunt++ (committed using GitHub Web editor) | categories/cookbook/20web-automation/20-01-fetching-uri.pl
Change URL to a real http address

  perl6.org is getting redirected to perl6.org.		 11:12
perl6-examples: 82eeb3018c | (Zoffix Znet)++ (committed using GitHub Web editor) | categories/cookbook/20web-automation/20-01-fetching-uri.pl
Merge pull request #50 from mattwhisenhunt/patch-1

Change URL to a real http address
zengargoyle yeah, but could be included in full in some dist with rakudo/nqp/moar/etc other Artistic/etc license without worry and all self inclusive.
but LineNoise is bad with unicode so i don't think of it much. :P 11:13
ryu0 i'm surprised editline wasn't mentioned though. 11:14
it's a BSD licensed readline alternative.
lizmat ryu0: probably because nobody in the Perl 6 bubble knew about it ? 11:16
zengargoyle don't think i've ever heard of it.
heh
11:16 zakharyas left
ryu0 really? i guess i know a lot about linux libraries from reading a lot... 11:16
thrysoee.dk/editline/ 11:17
zengargoyle ugg, tell you something, last $WORK had a readline.a that wasn't GNU readline and was from the '80s 11:18
confused me to no end for a while.
program uses readline but doesn't do readline arrrrgh. 11:19
ryu0 let me guess. it was a getline/fgets clone?
zengargoyle i didn't dive into the details once i realized it was just somebodys code from ancient times. 11:21
ryu0 let me guess. it uses K&R style functions? 11:22
11:22 astj joined
zengargoyle much like the command processor that did TAB completion on a command in a way that is different from any other thing i've ever seen. 11:22
ryu0 stuff like how arguments used to be declared and defined between the ) and {. lol 11:23
i've only seen that convention in a few rare places.
BASH uses it extensively in some places.
zengargoyle i $WORKed at .edu that had Sun network over thickwire in '87 and still has some same code running today.
ryu0 I've also seen it in BSD code.
ANSI C changed the rules. 11:24
but K&R style is still legal C.
zengargoyle and is on the earliest maps of the internet when there was only like 5 lines...
it's not COBOL or anything, but there was tons of code that has been used since mid '80s unchanged. 11:25
11:26 xtreak joined 11:27 nhywyll joined 11:28 markmont joined
zengargoyle is old and doen't know why K&R C is such a bad thing. :) 11:30
ryu0 same reason pre-standardized C++ is a bad thing. =p 11:31
zengargoyle by the time it was ANSI C 98 i was doing mostly Perl. :)
11:32 TimToady left
ryu0 zengargoyle: uh, ANSI C arrived in 89/90. It's also called C89/C90. 11:33
Do you mean C99?
Or C++98?
11:33 http_GK1wmSU joined
zengargoyle i probably should have picked up some C++ 11:34
11:34 mcsnolte left
zengargoyle oh, yeah, C99 11:34
11:34 TimToady joined
ryu0 have you given D a look? 11:34
zengargoyle i learned from the '78 K&R 11:35
11:35 http_GK1wmSU left
zengargoyle no D, for the most part i often look and say meh and really i don't have the use case for that. 11:37
ryu0 fair enough.
i've been looking for an all-in-one high level language, and i thought perl6 might be what i'm looking for.
including a flavoring of functional but doesn't force me into that model all the time. 11:39
11:40 lichtkind joined
zengargoyle i think it will get there. if moar can JIT along the same lines of old JVM was dog slow, now it's fast. 11:41
11:42 araraloren joined
zengargoyle and in my mind, the parts of p6 that are written in p6 were originally CS101 code, correct, works, not necessarily the fastest thing. 11:42
bits and pieces keep getting improved, VM keeps getting better. not sure where JVM backend stands, but that's a thing, JS backend is a thing in progress. 11:43
ryu0 was compiling directly to native code ever considered? 11:49
zengargoyle don't know. i think you could say that perl6 is written in perl6 is written in nqp and nqp is an abstract syntax tree. whether you could turn that into raw native code, i don't know. 11:52
11:52 AlexDani` joined
ryu0 wouldn't surprise me if it didn't. seems like that's more common with static typed languages. 11:53
zengargoyle i think the vm does do some native code replacement. if i read jnthn's blog posts correctly.
11:53 AlexDaniel left
Geth doc: scmorrison++ created pull request #1434:
routine link creates hard links		 11:54
zengargoyle there is JIT and inlining of code and some native bits. but that's all way above my head... :)
jnthn Yeah, we do JIT compilation
So hot bits of code are typically turned into machine code
ryu0 maybe eventually they'll also target CIL too.
if JVM works out. 11:55
araraloren The compiler will compile some code to native code according some indicator
11:56 meemoo joined
lichtkind moritz, ping 11:56
jnthn Not only that, it may compile multiple versions of it specialized for certain input types. :)
lichtkind cheers jnthn
jnthn On MoarVM, at least.
o/ lichtkind
araraloren I know a little about JIT
they have optimization base on Path or Method 11:57
jnthn fwiw, I've been doing a bunch of work on that area recently, so will write some blog pots about it soon
*posts :)
araraloren jnthn++, sound greater
sounds great
11:59 meemoo left
ufobat zengargoyle, ryu0 yeah i know, its fine, i just wanted to mention that he (my coworker) is complaining in other words some users would expect that to be there. BTW there are official ubuntu packages that have a 2015.xx 12:01
12:02 w4and0er96 left
ryu0 ufobat: i'm aware. i'm planning to make a PPA for people to use a more recent version. 12:02
ufobat kk 12:03
:) awesome!
araraloren ryu0, there are binary package for Ubuntu
ryu0 araraloren: i saw.
araraloren yeah 12:04
ryu0 doesn't have the full advantages of a PPA though.
12:07 AlexDani` left 12:08 Slaash left, slaash joined 12:11 slaash left
zengargoyle ufobat: i understand, it's put a line in a textbox or file and forget about it. i think there are acually Debian people looking into the whole packaging thing so that recent rakudo+etc and precompiled modules are in the standard places. 12:11
ufobat: i.e. people much better than i are working on it on the Debian side of things... :) 12:13
and maybe ubuntu borrow debian when done. :) 12:16
12:17 konsolebox left
zengargoyle m: sub x($a) { state $x = 0; state &y = sub { "world" ~ $x++ }; return "hello" ~ y }; say x(1), x(2); 12:18
camelia helloworld0helloworld1
zengargoyle parting question for the others... can &y be done only once yielding 'helloworld0helloworld0"? 12:20
zengargoyle goes to bed... 12:21
12:21 AlexDani` joined
araraloren good night! 12:21
ufobat zengargoyle, i coudn't do it as well.. 12:22
12:23 konsolebox joined
ufobat i just wrote a useful little tool for my $daylightjob and my coworker wanted to use it as well. and he is already not perl6 loving and failing to install it quickly made it even worse for him 12:23
at least he likes my tool ;)
12:25 Slaash joined 12:27 AlexDani` is now known as AlexDaniel, sherwool joined 12:28 xtreak left 12:29 xtreak joined, nhywyll left 12:30 prabh left 12:31 xtreak left 12:33 ChoHag left 12:34 astj left
lizmat ufobat: installing rakudo, or the specific modules ? 12:35
Zoffix .tell zengargoyle m: sub x($a) { state $x = 0; state &y = sub { "world" ~ once $x++ }; return "hello" ~ y }; say x(1), x(2); 12:36
12:36 grumble left
nebuchadnezzar zengargoyle: yes, we need to find a proper way to handle packages, it looks like we will do like Emacs or Python packages, byte-compile at install time 12:36
Zoffix misses yoleaux
zengargoyle Zoffix: i did that one, but desired someting that operated on sub level like; my &y = once sub { ... } 12:38
zengargoyle really gong to bed now. :)
Zoffix zengargoyle: what are you trying to do exactly?
huggable: xy problem 12:39
huggable Zoffix, You want to do X, but don't know how. You think you can solve it using Y, but don't know how to do that, either. So, you ask about Y in order to solve X, which doesn't make sense. You should ask about X.
ufobat rakudo
12:41 Aaronepower joined 12:45 andreoss left
Geth modules.perl6.org: b6d153d023 | (Zoffix Znet)++ (committed using GitHub Web editor) | lib/ModulesPerl6/DbBuilder/Dist/Source/GitHub.pm
[REBUILD] Do not loose has_appveyor status...

  ... went going through half-build due to no new commits
Fixes github.com/perl6/modules.perl6.org...-320154980		 12:45
Zoffix s/loose/lose/;
s/went/when/; sheesh.. I need typing lessons 12:48
12:52 setty1 joined 13:03 wamba left
lichtkind lizmat: i now change Perl6::Doc as agreed 13:05
13:06 zakharyas joined
lizmat lichtkind: is that a question ? 13:07
lichtkind lizmat, judt a message :) . see foreward to see you the xxx :) 13:08
13:08 Cabanossi left 13:09 astj joined 13:11 Cabanossi joined 13:13 astj left 13:14 wamba joined 13:22 andrzejku left, enheh left, enheh joined 13:26 cdg joined
lizmat as do I :-) 13:27
Geth modules.perl6.org: 30e49f7e52 | (Zoffix Znet)++ (committed using GitHub Web editor) | lib/ModulesPerl6/DbBuilder/Dist/Source/GitHub.pm
OCD
 13:30
modules.perl6.org: 1e1d67f1ec | (Zoffix Znet)++ (committed using GitHub Web editor) | bin/build-project-list.pl
Decrease dist fetch interval

The original limit was to account for us having more dists in the future, but there's actually an API end point we can poll to figure out the limits. So reduce the interval for now and will implement smarter rate limiter that uses that endpoint later.		 13:32
Zoffix m: say 868*2/60
camelia 28.933333
Zoffix half hour shaved
Can probably shave even more once I add smarter rate limiter :) 13:33
moritz do we use the public API, or through an account?
ISTR that rate limits were higher when you use API tokens 13:34
13:35 edehont left
Zoffix Account 13:36
13:38 rindolf joined
moritz ok 13:40
13:42 g0d355__ joined
moritz yay, Perl 6 Fundamentals now officically "In Stock" on Amazon: www.amazon.com/dp/1484228987 13:46
Zoffix: now you should be able to write a review, should you still be willing :-)
Zoffix \o/
13:48 grumble joined 13:52 mcsnolte joined 13:53 skids joined 13:54 itaipu joined 13:55 cdg left 14:00 Kyo91 joined, jferrero joined 14:01 setty1 left 14:02 pecastro left, pecastro joined 14:03 lowbro left 14:06 astj joined
mspo is sprintf the best built-in "template" system? 14:09
14:10 astj left
Zoffix mspo: yeah. There's fmt method though that's an alternative way to use sprintf 14:10
m: :foo.fmt("%s ♥ %s").say
camelia foo ♥ True
mspo I'd just like named parameters instead 14:11
Zoffix Ah.
skids Close around a qq?
Zoffix m: {"$:key ♥ $:value"}(:foo.Pair).say 14:12
camelia Too many positionals passed; expected 0 arguments but got 1
in block <unit> at <tmp> line 1
Zoffix m: {"$:key ♥ $:value"}(:key<meow>, :value<foos>).say
camelia meow ♥ foos
mspo m: say sprintf("%s, and %s", "foo", "bar"); 14:13
camelia foo, and bar
Zoffix Doesn't give you formatting options tho.
14:13 zakharyas left
Zoffix m: {sprintf "%.4f ♥ %.4f", $:key, $:value}(:42key, :70value).say 14:14
camelia 42.0000 ♥ 70.0000
Zoffix heh
skids m: {"$:key ♥ {$:value.fmt("%2.2d")}"}(:key<meow>, :value<45.2>).say 14:15
camelia 5===SORRY!5=== Error while compiling <tmp>
Placeholder variable $^value may not be used here because the surrounding block takes no signature
at <tmp>:1
------> 3{"$:key ♥ {$:value.fmt("%2.2d")}7⏏5"}(:key<meow>, :value<45.2>).say
…
14:15 zakharyas joined
skids hrmf 14:15
Zoffix That error sucks 14:16
mspo m: say sprintf("%2\$s, and %1\s", "foo", "bar");
camelia 5===SORRY!5=== Error while compiling <tmp>
Unrecognized backslash sequence: '\s'
at <tmp>:1
------> 3say sprintf("%2\$s, and %1\7⏏5s", "foo", "bar");
expecting any of:
postfix
Zoffix m: "{$:x}"
camelia 5===SORRY!5=== Error while compiling <tmp>
Placeholder variable $^x may not be used here because the surrounding block takes no signature
at <tmp>:1
------> 3"{$:x}7⏏5"
expecting any of:
horizontal whitespace
sta…
mspo m: say sprintf("%2\$s, and %1\$s", "foo", "bar"); 14:17
camelia bar, and foo
mspo okay, explicit numerics aren't bad
Zoffix Filed as rt.perl.org/Ticket/Display.html?id=131838 14:18
AlexDaniel ∞ 14:21
«∞»² 14:22
14:23 Cabanossi left
ufobat moritz, i can buy it "used for 3dollars cheeper? :D 14:24
Zoffix hah, what a scam, 14:25
I get "5 Used from $20.64"
14:26 Cabanossi joined
Zoffix I guess not really a scam, since you're getting a new book instead of used. 14:27
Still a lie and fake numbers.
14:30 brabo left
Zoffix grrrrr 14:30
modules site refused to do full rebuild again. The hell is its problem :/ 14:31
14:31 brabo joined
Zoffix wonder if there's any problem with the bash script github.com/perl6/modules.perl6.org...l6.org#L18 14:33
14:34 devmikey joined
Geth modules.perl6.org: 6be4574b90 | (Zoffix Znet)++ (committed using GitHub Web editor) | TODO-IDEAS.md
[REBUILD] Remove Koality from TODO

It was already implemented and subsequently killed		 14:36
14:44 Exodist left, Exodist joined 14:51 enheh left 14:53 nhywyll joined, Cabanossi left 14:56 Cabanossi joined 15:03 TEttinger joined 15:07 itaipu left 15:08 itaipu joined 15:18 wamba left
Zoffix :| 15:21
Above commit didn't trigger full build 15:22
Ohhhhh 15:24
*phew* finally figured it it out: updater script pulls, starts the builder script with the full rebuild flag set, BUT, if the previous run is currently running, the builder script just exits and the next time it's started it no longer sees the commit for rebuild 15:25
moritz the danger of having rebuilds take longer than the interval they're scheduled for 15:27
Zoffix :) 15:28
gonna fix it shortly
AlexDaniel www.cs.cmu.edu/%7Etom7/papers/sigb...twords.pdf :) 15:33
15:35 nadim left 15:43 jaffa4 left 15:45 araraloren left
Geth modules.perl6.org: 2863c4b697 | (Zoffix Znet)++ | bin/build-project-list.pl
[REBUILD] Fix failures in full-rebuild-from-commits

The bash script pulls and based on commit message titles tells the build script to do a full rebuild. However, if another build run is currently running, the build script will just exit. The next build run now won't see those commits and no full rebuild is done.
Fix by leaving a marker file lying around when we're bailing out due to another run being active and a full rebuild was requested.		 15:49
15:51 sufrostico left 15:53 ufobat left 15:54 sivoais left
moritz Zoffix++ 15:54
Zoffix well, it's not yet working for some reason. I don't see it create any marker files :D 15:55
15:57 setty1 joined, zakharyas left 16:00 Slaash left, araraloren joined 16:01 robertle left, sivoais joined
Geth modules.perl6.org: f270659f11 | (Zoffix Znet)++ | TODO-IDEAS.md
[REBUILD] Fix typo
 16:01
Zoffix ok, that did work. OK. I think all good now 16:02
16:05 Kyo91_ joined 16:08 Kyo91 left
Zoffix AlexDaniel: did you mail or email your CLA? You can *email* it and that generally has fast turn around time, so you could do that so you have the keys to the repo in time for release 16:09
huggable: CLA
huggable Zoffix, www.perlfoundation.org/contributor_..._agreement print it out, sign it, scan it in. Email to trademark -at- perlfoundation.org
AlexDaniel Zoffix: I snail mailed it. Yep, you are right, I should probably send another one via email… 16:12
16:14 TimToady left 16:15 astj joined 16:16 TimToady joined 16:19 astj left
[Coke] moritz: I can already buy your book used. 16:23
16:25 Cabanossi left
Zoffix :) 16:25
16:25 meemoo joined 16:26 Cabanossi joined
moritz [Coke]: I know. I think there are vendors who sell every book used, and speculate that they can obtain a copy in time, or get a bulk discount on a new book or so 16:26
or make their margin through inflated shipping costs or whatever 16:29
16:29 meemoo left
Zoffix or lure in buyers who may have thought price for new is too high and would've left; smaller margin, but better than zero 16:29
moritz or that, yes 16:30
Zoffix or make buyers of new books feel like they're getting a good deal buy paying just $3 more bucks for a new instead of used copy
s/deal buy/deal by/;
ryu0 \o/ moarvm packaged for my PPA. Now to let to build. 16:31
lichtkind greeting moritz 16:32
Zoffix ryu0: BTW, saw some earlier convo about letting the user pick which parts they want to install... keep in mind a particular rakudo release typically requires a particular nqp release (which often requires a particular release of moarvm). The only mix-and-matching to be done is MoarVM backend vs. JVM or JavaScript backenedd
moritz \o lichtkind
lichtkind moritz i currently rewrite Perl6::Doc wher shoud in point beside .doc.perl6.org?
16:32 meemoo joined
lichtkind any deep links to recommend 16:33
ryu0 Zoffix: OK. I was planning to only update these when there's a new rakudo star bundle. 16:34
Zoffix Ah.
ryu0 Since that seems to be what most people would be using.
Zoffix FWIW, rakudo star releases are quarterly and rakudo+nqp+moarvm releases are monthly
ryu0 I noticed. 16:35
But monthly seems a bit excessive to me.
moritz lichtkind: well, perl6book.com/ has an overview of Perl 6 books, so that might be worth a link
Zoffix Certainly what they should be using :) Though I see a lot of people sporting random dev commits that happened to be HEAD when they ran rakudobrew :)
16:36 ChoHag joined
moritz lichtkind: perl6intro.com/ and learnxinyminutes.com/docs/perl6/ are pretty popular too 16:36
lichtkind yes intro i know too do you have access to the author since he doesnt respond to me 16:37
Zoffix .seen hankache 16:38
Who runs the yoleaux bot, again?
timotimo yoleaux still goe
Zoffix huggable: yoleaux
huggable Zoffix, nothing found
Zoffix bah
moritz lichtkind: not more than his email address and IRC nick; I guess you have those too
timotimo github.com/dpk/yoleaux - maybe?
16:38 http_GK1wmSU joined
Zoffix lichtkind: this is their twitter: twitter.com/hankache 16:38
16:38 http_GK1wmSU left
Zoffix oh, right, user dpk on IRC 16:39
Zoffix sent a /msg
16:40 meemoo left 16:41 edehont joined 16:42 meemoo joined, meemoo left 16:44 kaare_ joined, haha joined, kaare__ left, haha is now known as Guest85096, cschwenz left 16:47 Guest85096 is now known as hawtf
lichtkind thanks moritz 16:47
16:50 brrt joined 16:51 hawtf left, haha_ joined 16:53 brabo left
Zoffix So AppVeyor doesn't use the same user/repo as GitHub? 16:57
e.g this project github.com/ugexe/Perl6-Net--HTTP appears to have just a single hyphen before HTTP on appveyor ci.appveyor.com/project/ugexe/Perl...nch/master
And github.com/sergot/openssl has a badge linking to ci.appveyor.com/project/skinkade/o...nch/master (different username). 16:58
16:58 edehont left
Zoffix How does it know which project is what? Is there a way to figure it out if all you got is the gihub repo and its contents? 16:58
k, there's a way to get the badge by tacking on `github/` into the URL: ci.appveyor.com/api/projects/statu...-Net--HTTP 17:01
Zoffix looks for a way to do that same with history
17:02 ramillies joined
Zoffix Ohohoh. Actaully, we can cheat with the badge, by getting it in SVG format and figuring out status like that :P 17:02
17:05 haha_ left 17:06 haha_ joined 17:13 andreoss joined, pilne joined
Zoffix damn. badge can say "running"? 17:15
I guess it's OK actually. We can make the site use that and refetch later 17:16
ah. Makes my job easier www.appveyor.com/blog/2014/11/26/c...us-badges/ 17:17
moritz passingText=failed :-) 17:21
17:22 espadrine joined
ramillies Hello everybody. Could you please explain what this error message means: "Serialization Error: missing static code ref for closure 'run' (gen/moar/stage2/NQPHLL.nqp:1866)"? 17:24
(I'm sorry I can't provide more context, it seems to happen only in one piece of code at one place if the code is supposed to do what I want... :--( ) 17:25
(and that piece of code works OK when isolated from the rest of the module.) 17:26
Zoffix ramillies: that sounds like an internal error that regular users are never meant to see :/ 17:27
ramillies: what's your rakudo version? perl6 -v
ramillies The newest one, 2017.07.
Exactly "This is Rakudo version 2017.07-126-g86cb1363d built on MoarVM version 2017.07-293-g3f33a841" 17:28
Zoffix err
ramillies: well, you're using a random dev commit and are hitting some bug.
m: say $*PERL.compiler.version
camelia v2017.07.129.g.5.d.200.f.1.e.5
Zoffix ramillies: it would be helpful if you could provide some code that reproduces that error
ramillies I see. That was what rakudobrew gave me. // The bad thing is that I can provide only a whole module. 17:29
(Terminal::Print to be exact)
[Coke] and also verify that it still happens against a stock 2017.07 install.
ramillies Whenever I try to extract a MWE it behaves exactly as intended.
Zoffix ramillies: but as far as resolving your problem, you could try a tested, stable release. (rakudobrew build moar 2017.07; rakudobrew switch moar-2017.07)
ramillies: extract it into a module that you still precompile 17:30
ramillies OK, thank you. Currently trying.
ugexe always rakudobrew nuke moar $whatever first
Zoffix ramillies: the "serialization error" is something that happens with precompilkation
ugexe moar-$whatever rather
Zoffix ugexe: how come?
rakudobrew is designed for running multiple versions, I mean :/ 17:31
ugexe because otherwise it will rebuild the moar-$whatever, which seems to cause weird problems
so it doesn't hurt to run it first, and it helps if they happen to already have it installed
Zoffix huh, never had them
ramillies OK, nuked and going anew. 17:33
17:34 devmikey left
ugexe there is also the hail mary `rm -rf ~/.perl` 17:34
er ~/.perl6
Zoffix update-perl6 is aliased to `rm -fr ~/.perl6; rm -fr ~/.zef; rm -fr ~/.rakudobrew/; git clone github.com/tadzik/rakudobrew ~/.rakudobrew; rakudobrew build moar; rakudobrew build zef; zef install TAP::Harness; zef install WWW' 17:35
that's my alias. Though normal users need to specify a proper release
ramillies The stable compiler is now installed, still rebuilding zef and pulling prerequisites. 17:39
And that one sadly gives something quite similar: "Code ref '' does not exist in serialization context." 17:40
17:42 hoelzro left, hoelzro joined
Zoffix ramillies: do you have any BEGIN blocks or constants? You said you can share the module? What's the code 17:43
Zoffix wags the "you didn't run tests" at the previous hackers of modules.perl6.org :) 17:45
*figner
awwaiid Remind me -- I used to have MAIN(IO() $file where :e) { ... }, is there a new way to do htat (take a paramter filename that exists and turns into an IO)
Zoffix tho I guess the tests suck; too specific 'cause I was going for 100% coverage
ramillies Yes, there is one. 17:46
Zoffix awwaiid: that's still the way
ramillies I'll put the code on github shortly.
awwaiid Zoffix: oh. you're right htat works. I was trying to do `subset FileName of IO() where :e;` which didn't work out so well 17:48
ramillies OK, I managed to extract quite a manageable chunk (30 lines). 17:50
Zoffix yeah, that won't work
The coercer would go to the signature but even then, it'd have to have method FileName on Str :/
Geth ecosystem: 514f75a791 | loren++ (committed using GitHub Web editor) | META.list
Change URL

change to master branch		 17:53
ramillies gist.github.com/ramillies/edb326b0...ed8c176035 — Here it is. 17:54
Geth modules.perl6.org: 2ee3f47d16 | (Zoffix Znet)++ | 3 files
Fix failure to fetch AppVeyor status / Add 'pending' status

Turns out we can't just use github-username/repo-name as names for AppVeyor project as they can differ. Haven't found a way to get to data via REST API, but the badge URL allows "github/" specifier that does work. So I fixed that problem by parsing the SVG badges.
... (5 more lines)		 17:55
Zoffix hopefully that works; tests are bitrotten so I can't even tell which are my failures and which aren't.
ramillies: why the BEGIN block? 17:56
ramillies It's a little piece of Terminal::Print. 17:57
And there it is to query tput capabilities only once, when the module is compiled.
I guess at least.
This is just an excerpt that manages to still reproduce the bug. 17:58
(so the code has no actual sense to it)
(Btw. note that this MUST be loaded as a module — if you make an ordinary script out of it, the error won't manifest) 18:00
Zoffix Yeah, I reproed it. Trying to golf it down more.
18:00 brrt left
haha_ why do you like rakudo? because the syntax is fun? 18:02
Zoffix hehe "Missing serialize REPR function for REPR ConditionVariable (Lock::ConditionVariable)" 18:03
haha_: because it's concise yet still readable.
18:03 lostinfog joined
Zoffix As the poster says: raw.githubusercontent.com/perl6/ma...300ppi.png 18:03
ugexe wouldnt that be serializing all the failures?
araraloren I start an appveyor build, it build failed with error: Could not find Text::Tabs at line 3 in: 18:04
haha_ Zoffix: that sounds good enough. 18:05
18:05 nadim joined
Zoffix buggable: eco Text::Tabs 18:05
buggable Zoffix, Text::Tabs 'Perl 6 port of Perl's Text::Tabs': github.com/Altai-man/perl6-Text-Tabs
Zoffix araraloren: did you add it to `depends` meta section?
or whatever it was
araraloren yeah, I added, it was ok on travis
haha_ i actually like how someone explained why OOP is at the heart of the rakudo. - Because Larry likes it :D
Zoffix OK, golfed one of the issues: unit module Foo; BEGIN shell 'dir', :out; That gives `Missing serialize REPR function for REPR ConditionVariable (Lock::ConditionVariable)` 18:06
18:06 astj joined
araraloren I copy the appveyor config from ugexe repo 18:06
Zoffix no idea 18:07
araraloren github.com/araraloren/perl6-termin...pveyor.yml
Zoffix Filed as rt.perl.org/Ticket/Display.html?id=131839 18:08
araraloren gist.github.com/araraloren/f793d1e...0908a07cd2 18:09
build log
Zoffix yey, gofled the other one too. ugexe++ indeed that was a failure
18:10 astj left 18:12 devmikey joined
Zoffix ramillies: well, I filed two tickets to fix two core bugs. rt.perl.org/Ticket/Display.html?id=131840 and rt.perl.org/Ticket/Display.html?id=131839 18:14
ugexe araraloren: you have dependencies. you need to install them
Zoffix ramillies: if you change that `fail` to `die` that should avoid your bug (and probably what you want there anyway)
araraloren ugexe, how to install ? use zef ? 18:15
ugexe yes
ramillies OK, thank you. As to "die", that's exactly what I didn't want. 18:16
The idea was that the exception should be thrown only when those capabilities are actually about to be used.
Geth modules.perl6.org: 24c2161504 | (Zoffix Znet)++ | TODO-IDEAS.md
[REBUILD] Trigger full rebuild

To rebuild bad appveyor statuses
ugexe return a 0 or 1, unless you want to have the exceptions serialized. throw your exception outside of BEGIN based on that 18:17
ramillies Thanks, I'll try that. 18:18
ugexe araraloren: github.com/ugexe/Perl6-CompUnit--R...or.yml#L19
araraloren ugexe, thanks very much
Geth modules.perl6.org: c758b7949d | (Zoffix Znet)++ | 2 files
Remove a bunch of duplicated CSS
 18:22
Zoffix broke something 18:24
ah; I fixed my test code but forgot to copy to real code -_- 18:25
Geth modules.perl6.org: c4756595ed | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/AppVeyor.pm
[REBUILD] Fix typo
ugexe ramillies: i'd suggest checking at runtime - the availability of tput depends on the env 18:26
you could use `once { }`... here is one pattern I experimented with gist.github.com/ugexe/90b41dbf217c...xx-pl6-L35 18:28
basically use the proto to handle the logic for what to do if it knows its multis won't work, and once { } makes sures it only needs to makes the system call to check once 18:31
18:33 nadim left
ramillies ugexe: Thanks, it's not my code so I didn't want to change it more than necessary. I'll still have to read up on the "proto" in order to understand the example. 18:36
(and I don't get the '{*}' construct either...) 18:37
ugexe the proto is not neccesary if you are not repeating this pattern much
you could do `die "xxx" unless once { do-check() }` as the first statement of your routine 18:38
Zoffix ramillies: the `{*}` is where the body of the matching candidate "goes". So that code reads run the candidate's code if this condition is true 18:39
I got work yoleaux will end its robotstrike soon
ramillies Ah so, thanks. 18:40
Zoffix m: proto foo (|) { say "and.. "; {*}; say " and.. "; {*}; say "weeee" }; multi foo (Int $x) { say $x }; multi foo(Str $s) { say "[STR $s]" }; foo 42; foo "meows"
camelia and..
42
and..
42
weeee
and..
[STR meows]
and..
[STR meows]
weeee
ramillies A nice example, thank you. 18:42
18:42 nadim joined 18:50 sufrostico joined
tbrowder hi, #perl6 18:51
Zoffix \o\\
tbrowder hi, Z! i want to try nativevcall to use an openssl function. is there 18:52
18:52 CacoS joined
tbrowder any special handling to grab stdout and stderr with that interface like there is with proc run? 18:53
18:55 andreoss left, Cabanossi left 18:56 Cabanossi joined
haha_ what do you think of the rakudo mascot? i remember how i forgot perl6 when i opened the website and saw that butterfly. it took like 2 years until i found it again and considered it as a normal language. now i kinda like it. it looks crazy. 18:57
[Coke] tbrowder: have you seen the eco SSL modules? 18:58
tbrowder yes, and i'm looking at maybe a PR if I can grok what's happening 18:59
i can use proc run but a native call iface is probably safer, no? 19:00
19:00 robertle joined
ugexe capture the stdout of the process like you would capture output from any other non-proc source I would think 19:01
the 'ol temp $*OUT = class :: { method print(|c) { $some-storage-var ~= c.join } } 19:02
Zoffix haha_: looks fine to me; the color saturation is a bit too much on the eyes, but I like the benefits an awful logo brings. It gives people who are just there to complain about anything something to complain about, so you can just tune them out; without a logo they'd start complaining about feature X, drawing people into a heated discussion :) 19:03
There's also a write up somewhere about the logo being a cultural hack to weed out the (what I call) "elitist types"; they can't bring themself to fly the banner of butterfly :D 19:04
skids SSH stdout/stdin handling is somehwat special. There's some safeguards in there specifically to fudge up people trying to programmatically use password-based auth methods. 19:05
(when not using a vt)
jnthn is now confused if this is about OpenSSL or OpenSSH... :)
skids No, it is I who is confused 19:06
jnthn hah
19:06 espadrine left
Zoffix haha_: plus, I can't say I spend much time looking at that butterfly :) Here's a book cover I designed for "The White Book". It's a butterfly, but a completely different one: i.imgur.com/HsivBCT.jpg 19:06
jnthn If it's SSH I wrote a module to do that, and you can use it rather a lot like Proc::Async :)
tbrowder i'm looking at x509 functions, nothing with ssh 19:07
skids Oh just write your own ASN.1 parser (snicker) 19:10
haha_ zoffix: that's an interesting thought 19:12
jnthn tbrowder: Native calling them is likely the more robust way in so far as all it needs is the library. I doubt that the C API thinks in terms of stdout/stderr and so forth, though. OTOH, this is OpenSSL, which has to be one of the less fun C libraries I've done Perl 6 bindings with :P
haha_ there should be more people like zoffix 19:13
[Coke] agreed, zoffix is on my clone-if-possible list. 19:15
Zoffix Ugh. Spotted another appveyor update bug: the links to results on the modules.perl6.org page link to url + github username/repo; and the same trick of adding `github/` to the URL that works for badges doesn't work for project page :/ 19:16
so some of those links are 404s 19:17
damn: "Do you mean default AppVeyor page and not badge? If yes, we do not provide this URL route, because we don't know what to do when more than one AppVeyor project uses single repository. This is very common" 19:20
19:20 wamba joined 19:22 brabo joined, ufobat joined 19:23 brabo left 19:24 brabo joined 19:31 darutoko left 19:33 haha_ left 19:41 cpage_ left 19:43 haha joined, haha is now known as Guest99050
Zoffix OK. I don't know a solution to that, so I re-opened the appveyor Issue: github.com/perl6/modules.perl6.org...-320336909 19:43
19:45 brrt joined 19:46 Guest99050 left 19:49 xinming joined 19:50 andrzejku joined
andrzejku hi 19:50
people
the Perl6 module file shouldn't be runnable right?
19:51 devmikey left
timotimo do you mean +x posix permissions? 19:53
[Coke] modules don't need to be executable. script could be. 19:54
19:57 astj joined 19:58 ijneb left
andrzejku I mean implementation fact 19:58
when I name something as pm6
and p6
so convention is all p6 should be executable
skids Personally, if I'm going to install a script, I'm not going to use a suffix at all, because it adds nothing to the CLI UX. 20:01
20:01 astj left
andrzejku I know I know 20:01
I just ask about convention 20:02
20:03 cpage_ joined
skids Convention? It's hard enough to get people not to just use "pm" :-) But yes, .pm6 files are not expected do anything when loaded alone/directly. They can, but don't have to. 20:03
andrzejku :< 20:05
skids I know thats possible to say lets fuck all conventions and write how we want and what we want 20:06
but I suppose the homogenous convention can help people to understand your code structure better
20:08 cpage_ left
geekosaur more than that; some platforms force such conventions on you (e.g. windows does things by file extension) 20:09
20:10 cpage_ joined
ramillies Zoffix: I don't want to sound scary, but please try this one: echo 'unit module Foo; die; CATCH { }' > Foo.pm6; perl -I. -MFoo -e '' 20:16
I get "Cannot invoke this object (REPR: Null; VMNull)" with this.
20:16 andreoss joined
ugexe i think you mean perl6 -I 20:17
ramillies (s/perl/perl6/ in the first)
exactly 20:18
Zoffix ramillies: I'm on the bus ATM. Would you report it as a bug?
huggable: rakudobug
huggable Zoffix, Report bugs by emailing to [email@hidden.address]
Zoffix &
ramillies Sure. I just wanted to ask first. 20:19
ugexe echo 'module Foo { die; CATCH { default { say $_ } }; }' > Foo.pm6; works though 20:21
ramillies Looks like the bug manifests only if there is a CATCH but the die is not caught. 20:23
20:26 AlexDaniel left
ugexe scoping matters too 20:26
unit module Foo; would not work in my example
had to be module Foo { }
20:27 andrzejku left
ramillies Hm, I just tried and you're right. 20:28
Actually it seems that if it's 'module Foo { }', it works OK, if 'unit module Foo;', the bug manifests. 20:29
(and I thought these two are equivalent...)
'module Foo { die; CATCH { } }' will work too. 20:30
20:33 CacoS left 20:34 wamba left, lostinfog left
andreoss m: class Foo::A {}; my $a = "A"; say Foo::{$a}.new; 20:37
camelia Foo::A.new
andreoss m: class Foo::A {}; my $a = "C"; say Foo::{$a}.new; 20:38
camelia Any.new
andreoss is it a bug?
m: class Foo::A {}; my $a = "C"; say "Foo::{$a}.new".EVAL;
camelia Could not find symbol '&C'
in block <unit> at EVAL_0 line 1
in block <unit> at <tmp> line 1
20:39 AlexDaniel joined
nadim Take the poll /if you have nothing better to do on a Friday evening), www.facebook.com/groups/1595443877...848371898/ 20:39
El_Che I don't have a facebook account to vote (do I exist?) so I'll vote here: => 20:41
:)
nadim haha! 20:42
nobody said "duh, where are the unicode arrows?" yet 20:43
skids Also no facebook. I'm kind torn between => and infix :. A case could be made for adverbial, but not without the corresponding parens. Definitely not -. 20:44
nadim skids: I agred with you I also like the idea of :name(123) 20:45
although it is a bit noisy 20:46
skids definitely '=>' if your keys aren't idents.
andreoss m: class Foo::A {}; my $a = "Foo::C"; say {$a}.new;
camelia Cannot make a Block object using .new
in block <unit> at <tmp> line 1
nadim unicode → looks nice in a terminal but dreadful in my irc client 20:47
ufobat m: sub foo (|a, :%e) { a.perl.say }; foo(1, :e( a => 1)) 20:49
camelia Too many positionals passed to 'foo'; expected 1 argument but got 1
in sub foo at <tmp> line 1
in block <unit> at <tmp> line 1
ufobat 1 > 1 is True
skids andreoss: docs.perl6.org/language/control#blocks
andreoss m: class Foo::A {}; my $a = "C"; say Foo::{$a}.new; 20:50
camelia Any.new
andreoss should this be Any?
not a compile-time error
or runtime rather 20:52
skids Indirect name lookups just tell you what is in the stash. If the stash is empty at that name, it's Any, which is the default value for Stash values. 20:53
20:54 Cabanossi left 20:56 Cabanossi joined
nadim and here is a stash dump imgur.com/fthEDkK 20:57
lizmat nadim++ 20:58
andreoss so this kind of lookups is not okay for a factory 21:00
21:00 alimon left 21:01 skids left
ugexe your factory wouldn't know that getting an Any means it failed? 21:04
21:08 markmont left 21:09 ufobat left 21:11 Kyo91_ left 21:12 brrt left 21:13 Kyo91_ joined
geekosaur andreoss, if you are accessing the symbol table ("stash") directly, of course all bets are off 21:15
m: class Foo::A {}; my $a = "C"; say Foo::{$a}:exists 21:16
camelia False
geekosaur m: class Foo::A {}; my $a = "A"; say Foo::{$a}:exists
camelia True
21:17 nhywyll left 21:19 nhywyll joined 21:20 nhywyll left, nhywyll joined 21:25 cpage_ left 21:30 Kyo91_ left 21:37 wamba joined 21:38 cpage_ joined 21:42 wamba left, wamba joined 21:44 markmont joined 21:45 astj joined 21:49 astj left 21:50 ramillies left
andreoss m: class Foo::A {}; my $a = "A"; say Foo::{$a}:exists ?? Foo::{$a}.new !! die; 21:53
camelia Foo::A.new
andreoss too wordy
21:54 Cabanossi left 21:56 Cabanossi joined
nadim m: class Foo::A {}; my $a = "A"; say Foo::{$a}.?new ; 21:58
camelia Foo::A.new
nadim shorter
m: class Foo::A {}; my $a = "B"; say Foo::{$a}.?new ; 21:59
camelia Any.new
andreoss it won't die
nadim m: class Foo::A {}; my $a = "B"; Foo::{$a}.?new or die;
camelia ( no output )
andreoss m: class Foo::A {}; my $a = "A"; say (my Foo::A $ = Foo::{$a}.new)
camelia Foo::A.new
andreoss m: role Foo {}; class Foo::A is Foo {}; my $a = "C"; say (my Foo $ = Foo::{$a}.new) 22:00
camelia Type check failed in assignment; expected Foo but got Any (Any.new)
in block <unit> at <tmp> line 1
nadim m: class Foo::A {}; my $a = "B"; die ;
camelia Died
in block <unit> at <tmp> line 1
nadim m: class Foo::A {}; my $a = "B"; say Foo::{$a}.?new ;
camelia Any.new 22:01
nadim m: class Foo::A {}; my $a = "B"; say Foo::{$a}.?new ~~ Any or die;
camelia True
nadim m: class Foo::A {}; my $a = "B"; say Foo::{$a}.?new ~~ Any and die;
camelia True
Died
in block <unit> at <tmp> line 1
andreoss is there a way to make a type assertion without `my Type $ = ...`? 22:02
and still return the value
nadim what do you want the type assertion for?
andreoss to avoid returning Any in this case 22:03
22:04 rindolf left 22:05 yoleaux joined, ChanServ sets mode: +v yoleaux, dpk joined
nadim m: class Foo::A {}; my $a = "B"; sub factory { my $x = Foo::{$a}.?new ; $x !~~ Any and $x} ; my $x = factory() ; dd $x ; 22:05
camelia Bool $x = Bool::False
yoleaux 31 Jul 2017 21:06Z <Zoffix> nadim: nicq20++ fixed the Hash.ACCEPTS: Hash issue you were having. It now does `eqv` of two hashes. 22:06
dpk fixed yoleaux
sorry it took so long, i was on a wild goose chase all evening
nadim Zoffix: nice work with the Hash, I can take away the work around :) 22:07
m: class Foo::A {}; my $a = "B"; sub factory { my $x = Foo::{$a}.?new ; $x ~~ Any ?? () !! $x} ; my $x = factory() ; dd $x ; 22:08
camelia List $x = $( )
nadim m: class Foo::A {}; my $a = "B"; sub factory { my $x = Foo::{$a}.?new ; $x ~~ Any ?? |() !! $x} ; my $x = factory() ; dd $x ;
camelia Slip $x = $(slip( ))
dpk (out putting up placards for my $political-party ahead of upcoming election; almost finished, when a politician from another party sees us and tells us we can't do that yet, by law it's only allowed from 7 weeks before the election and we're 7 weeks and two days away; argument ensues; we have to go round and take all of them down or he'll get the public order office to come and fine us all)
(so i was not at my computer because i was wasting time putting up and then taking down posters all evening)
nadim m: class Foo::A {}; my $a = "B"; sub factory { my $x = Foo::{$a}.?new ; $x ~~ Any ?? return !! $x} ; my $x = factory() ; dd $x ; 22:09
camelia ===SORRY!===
Cannot find method 'ann' on object of type NQPMu
nadim m: class Foo::A {}; my $a = "B"; sub factory { my $x = Foo::{$a}.?new ; $x ~~ Any ?? return Any !! $x} ; my $x = factory() ; dd $x ;
camelia Any $x = Any
nadim m: class Foo::A {}; my $a = "A"; sub factory { my $x = Foo::{$a}.?new ; $x ~~ Any ?? return Any !! $x} ; my $x = factory() ; dd $x ;
camelia Any $x = Any
nadim m: class Foo::A {}; my $a = "A"; sub factory { my $x = Foo::{$a}.?new ; $x} ; my $x = factory() ; dd $x ; 22:10
camelia Foo::A $x = Foo::A.new
nadim m: class Foo::A {}; my $a = "B"; sub factory { my $x = Foo::{$a}.?new ; $x} ; my $x = factory() ; dd $x ;
camelia Any $x = Any.new
nadim m: class Foo::A {}; my $a = "B"; sub factory { my $x = Foo::{$a} ; $x} ; my $x = factory() ; dd $x ; 22:11
camelia Any $x = Any
nadim m: class Foo::A {}; my $a = "A"; sub factory { my $x = Foo::{$a} ; $x} ; my $x = factory() ; dd $x ;
camelia Foo::A $x = Foo::A
nadim m: class Foo::A {}; my $a = "A"; sub factory { my $x = Foo::{$a} ; $x ~~ Any ?? $x !! $x.new} ; my $x = factory() ; dd $x ; 22:13
camelia Foo::A $x = Foo::A
nadim m: class Foo::A {}; my $a = "A"; sub factory { Foo::{$a} ~~ Any ?? Any !! Foo::{$a}.new} ; my $x = factory() ; dd $x ; 22:14
camelia Any $x = Any
22:15 cpage_ left
nadim m: class Foo::A {}; my $a = "A"; sub factory { dd Foo::{$a} } ; my $x = factory() ; dd $x ; 22:16
camelia Foo::A
Any $x = Any
22:16 nhywyll left
nadim m: class Foo::A {}; my $a = "A"; sub factory { dd Foo::{$a} === Any } ; my $x = factory() ; dd $x ; 22:17
camelia Bool::False
Any $x = Any
nadim m: class Foo::A {}; my $a = "A"; sub factory { Foo::{$a} === Any ?? Foo::{$a} !! Any} ; my $x = factory() ; dd $x ; 22:18
camelia Any $x = Any
nadim m: class Foo::A {}; my $a = "A"; sub factory { Foo::{$a} === Any ?? Foo::{$a}.new !! Any} ; my $x = factory() ; dd $x ;
camelia Any $x = Any
nadim m: class Foo::A {}; my $a = "A"; sub factory { Foo::{$a} === Any ?? Any !! Foo::{$a}.new} ; my $x = factory() ; dd $x ; 22:19
camelia Foo::A $x = Foo::A.new
nadim andreoss: is the above good enough? 22:20
m: class Foo::A {}; sub factory($a) { Foo::{$a} === Any ?? Any !! Foo::{$a}.new} ; my $x = factory('A') ; dd $x ; 22:23
camelia Foo::A $x = Foo::A.new
nadim m: class Foo::A {}; sub factory($a) { Foo::{$a} === Any ?? Any !! Foo::{$a}.new} ; my $x = factory('B') ; dd $x ;
camelia Any $x = Any
nadim andreoss: and you get to return what you want on error, including generating an exception if you prefer 22:25
Zoffix nadim: FWIW, camelia takes a /msg, if you didn't want to flood the channel.
It was nicq20++ that fixed the hash thing 22:26
dpk++ thanks for fixing the bot :D
nadim yes I have been spamming like a pig :)
good to know, thanks Zoffix
nicq20++
andreoss paste.debian.net/979898/ 22:29
could .factory be a meta-method of some kind? 22:30
22:31 cpage_ joined
nadim Zoffix: I worked around the Hash eqv by using ===. Now I wonder if I shouldn't short cut eqv for Arrays and other containers. it doesn't make sense to flatten Array or Hashes that are equivalent, and if that is needed a block can do it 22:32
andreoss also is something like `method factory(...) return self.WHAT` possible?
22:37 nattefrost left
nadim andreoss: the code you posted will return a Failure, is it what you want it to do? 22:37
andreoss yes, i want it to complain instead of returning Any.new, because the latter daesn't make sense at all to me 22:38
but the error message is cryptic though 22:39
nadim It doe not have to return Any.new, see the example I posted above, you can get an Any instead, and that is not defined
andreoss yeah, i saw it 22:41
but i would prefer the compiler cheching it for me instead (i.e checking the type of the value being returned) 22:42
nadim and, IMHO, the example you nopasted makes no sense, having a role does not make sense. 22:43
andreoss why?
Zoffix .tell ufobat please report the bugs you find, so they could be fixed. I filed your 1 > 1 True as rt.perl.org/Ticket/Display.html?id=131842 22:44
yoleaux Zoffix: I'll pass your message to ufobat.
nadim you're using the role as a factory name container. you add the role to A and then you call the role directly thus making the use of the role on A completely moot 22:45
even if you called the factory method through A but demanded a B it would not make sense, why would A give you a B object just because they have the same role? 22:46
andreoss this way it wrong yes
it's 22:47
nadim andreoss: also, you write "I prefer the compiler checking the type of the value being returned" but you're writing a factory that returns different types! Did you mean the type of the value being assigned? in that case you can type the receipient of the factory call. 22:50
andreoss it's supertype in this case 22:51
nadim someone need to confirm but I do not think Foo::A is a sub type of Foo. let's check. 22:52
Altreus Can someone point me in the right direction to understand the difference between Str:D and Str:D: in a signature?
nadim Altreus: did you type the same thing twice? 22:53
Altreus no, one has a colon after the Definite smiley
I'm noticing that the :D version is a sub and the :D: version is a method 22:54
does it simply mean this is the invocant object?
So Str:D: means it's invoked as a method on a Str:D
nadim m: class Foo::A {}; my $x = Foo::A.new ; dd $x ~~ Foo ;
camelia Bool::False
nadim andreoss: you can see above that Foo is not the suprclass of Foo::A 22:55
andreoss there's no Foo
m: class Foo { } ; class Foo::A is Foo {}; my $x = Foo::A.new ; dd $x ~~ Foo ; 22:56
camelia Bool::True
nadim m: ćlass Foo {} ; class Foo::A {}; my $x = Foo::A.new ; dd $x ~~ Foo ;
camelia 5===SORRY!5=== Error while compiling <tmp>
Illegally post-declared type:
Foo used at line 1
Undeclared routine:
ćlass used at line 1. Did you mean 'last'?
andreoss Altreus: yes design.perl6.org/S06.html#Invocant_parameters 23:00
23:00 robertle left
zengargoyle .tell moritz clicky clicky purchased dead tree. 23:00
yoleaux zengargoyle: I'll pass your message to moritz.
Altreus OK, great, thanks :) 23:02
nadim andreoss: nopaste.linux-dev.org/?1160341 23:04
andreoss m: class Foo { multi method bar(Foo:U:) { say "hi" }; multi method bar(Foo:D:) { say "bye" }}; Foo.new.bar; Foo.bar;
camelia bye
hi
Zoffix Altreus: it's basically same as in Perl's object, the invocant is just the first arg to the routine, except in Rakudo you don't have to specify it (and can refer to it via `self` in the method). But you can use the "invocant marker" (the extra colon you noticed) to specify what sort of the invocant the method is for. This, for example, can be used in multi method dispatch. The `:D` on the type means instances 23:05
only and `:U` means only the type objects, so you can write:
m: class Foo { multi method what-is-it(Foo:D:) {say "tis an instance" }; multi method what-is-it(Foo:U:) { say "tis a type object" } }; Foo.what-is-it; Foo.new.what-is-it
camelia tis a type object
tis an instance
nadim but a factory call that exits the program that wild is also non sense
Zoffix m: class Foo { method what-is-it ($arg) {say "arg is $arg" } }; Foo.what-is-it: 70; Foo.new.what-is-it: 42 23:06
camelia arg is 70
arg is 42
Zoffix m: class Foo { method what-is-it ($self: $arg) {say "arg is $arg; invocant is $self.gist()" } }; Foo.what-is-it: 70; Foo.new.what-is-it: 42 23:07
camelia arg is 70; invocant is (Foo)
arg is 42; invocant is Foo.new
Altreus Zoffix: and sub what-is-it (Foo:D) would be outside of the class entirely?
Zoffix Altreus: yeah, it'd just be a lexical sub
Altreus I'm seeing it in the Str documentation, you see, so that's where there's a sub and a method
This is why substr(Str:d) and Str:D.substr both do things 23:08
Str:D: just doesn't look like it knows whether or not it's happy about it
Zoffix :D
s: '', 'substr', \(1) 23:09
SourceBaby Zoffix, Sauce is at github.com/rakudo/rakudo/blob/abf1...ool.pm#L74
Zoffix s: &substr, \('', 1)
SourceBaby Zoffix, Sauce is at github.com/rakudo/rakudo/blob/abf1...r.pm#L3035
Zoffix ^ there are both a method defined in a class (Str is a subclass of Cool) as well as a separate sub defined.
The result is you can use both a method call or a sub; whatever floats your boat 23:10
(not all methods have subs; just the common ones)
23:11 skids joined
Altreus there's a lot to learn 🤔 23:11
23:12 BenGoldberg joined
Altreus the default for an argument is :_, I assume? 23:12
Zoffix :) Yeah, there's lots of stuff in core that in Perl you'd normally load up from some module (e.g Moose)
Altreus So (Str:) is (Str:_:) 23:13
that one looks *really* sad :(
Zoffix Altreus: currently yes, but there's a proposal for 6.d language version to make the default on `%` and `@` sigiled params to default to `:D` github.com/perl6/6.d-prep/blob/mas...ls-imply-d
Altreus because it would basically never make sense to provide Array when an @ is requested? 23:14
Zoffix right
Altreus okay I think I'm getting it at least somewhat
Zoffix \o/
Altreus I've already used my $var = do if $long-construct -> $n {...} else {...{ 23:15
uh, } 23:16
I dunno if I like it though but that's step 2
apparently it's bedtime :P
ttfn
23:17 Exodist left
andreoss m: constant Foo: = Metamodel::ClassHOW.new_type(name => 'Foo'); Foo.^compose; class Foo::A is Foo {}; 23:17
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing initializer on constant declaration
at <tmp>:1
------> 3constant Foo:7⏏5 = Metamodel::ClassHOW.new_type(name =>
andreoss m: constant Foo := Metamodel::ClassHOW.new_type(name => 'Foo'); Foo.^compose; class Foo::A is Foo {};
camelia 5===SORRY!5=== Error while compiling <tmp>
'Foo::A' cannot inherit from 'Foo' because 'Foo' isn't composed yet (maybe it is stubbed)
at <tmp>:1
andreoss why it's not composed? 23:18
23:20 Exodist joined
zengargoyle Zoffix: the sub thing was somebody else's question an i was just trying to tease out better answers. guess you had to be there to figure out just what the asker wanted to know. 23:24
probably like split code into sub for clarity but not run sub each time or something like that. 23:25
maybe my sub foo and state $foo = foo(); or similar i guess...
i didn't really get the what they wanted either.... :) 23:27
Zoffix hehe 23:28
zengargoyle the joys of being around when everybody else is asleep. :P 23:29
geekosaur andreoss, if you're messing with a *HOW directly, I imagine there is more involved than just .^compose 23:30
andreoss m: constant Foo := Metamodel::ClassHOW.new_type(name => 'Foo'); Foo.^compose; say Foo.new; 23:31
camelia Foo.new
zengargoyle i'm torn on rakudobrew vs package, but think rakudobrew just needs instructions to 'build moar 2017.07' release instead of default moar-nom. 23:34
or to have a 'latest-release' tag sort of thing to default to. let the wise select 'moar-nom' 23:35
because rakudobrew, perlbrew, rbenv, etc.... the install and switch is very common and the truth is often "never trust OS packages". 23:36
zengargoyle apologies to packaging people. :)
23:38 ChoHag left
zengargoyle and makes me think of old question of Task::Foo type modules and having zef do best-effort install. i.e. install this list of modules, but don't barf i one fails, skip and just say at the end that these installed and these failed. 23:40
because if you 'nuke moar' then you've lost all the modules you had.... and need to attempt to get back to what you had before nuke. 23:41
and that shouldn't fail if Grammar::Debug or whatnot fails with new build because all of zef install is one transaction or serial. 23:42
Zoffix: so, if you brand as 'Rakudo Perl 6' what happens when there is 'Not-Rakudo Perl 6'. and all of your searches now go to a specific implementation? 23:48
how does one search for 'Perl 6' when everything is branded 'Rakudo Perl 6' but their needs are 'web-asm-magic Perl 6'. 23:50
that is why i'd go more for "Camelia Perl 6" where it's trademarked or whatnot and there can be no actual 'camelia' 'perl6'. then Rakudo and web-asm-magic can both state that they are Camelia (Perl 6) implementations.... 23:52
there may be rakudo/moar and rakudo/js but in the future there may be bobs-your-uncle/squeakvm. 23:54
the branding modifier whould not be an instance of implementation. 23:57
AlexDaniel zengargoyle: well, at first I didn't understand it too… I still cringe, but at least I get the idea now 23:58
zengargoyle ends musings.
AlexDaniel zengargoyle: so the idea is not to pick a *better* name, but to *change* it, no matter to what 23:59
zengargoyle yeah, sorta get the idea.
AlexDaniel and sure, we can call it “tweak” if it sounds better…

Powered by Cro and the Raku® Programming Language.

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

Thank you!