»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.
MadcapJake phew, finally back to a desktop after botching a distro ugprade 00:54
geekosaur can 00:55
geekosaur can't wait for his turn at that 00:56
(been putting off mint 17.3 upgrade on desktop)
(laptop will probably be worse, though, as I'd been using secure boot and ubuntu dropped it so it's gone in mint 17.3)
MadcapJake luckily i put my /home in a separate partition 00:57
geekosaur I do have backups, but.
MadcapJake not me, I like to live on the edge :P 00:58
AlexDaniel .tell TimToady qww<foo „hello world” bar> – no quote protection here. Was it supposed to be this way or is it a bug? 01:00
yoleaux AlexDaniel: I'll pass your message to TimToady.
pierrot Good night. 04:26
pierrot if I have an array of strings and want to build a regular expression for the first string of the array followed by some text followed by the second string of the array and so on.. how could I do that in p6? 04:30
in perl 5, I would do something like $pattern = join ".*", map{quotemeta} @keywords and then qr/$pattern/.. 04:32
CIAvash pierrot: I think this should work: $text ~~ /<{@keywords.join: '.*'}>/ 05:59
CIAvash or my $pattern = @keyworkds.join: '.*'; /<$pattern>/ 06:00
CIAvash That's without the quotemeta part though 06:35
RabidGravy badoom! 07:20
azawawi . 08:03
RabidGravy: ping 08:04
RabidGravy bloop!
azawawi RabidGravy: good morning :) 08:05
RabidGravy: strange output after adding provides on 2016.05 gist.github.com/azawawi/72729d6b5a...8bc0dbe5f5
RabidGravy: did you encounter such a problem?
RabidGravy nope 08:06
that's odd
something is doing "fail()" without a message 08:08
azawawi Build.pm without 'use LWP::Simple' works btw... 08:09
strange
tried adding "build-depends": [ "panda", "Shell::Command", "LWP::Simple" ]
same problem
RabidGravy looks at the Build.pm 08:11
azawawi actually to be exact even 'use NativeCall' causes that
with a bare class Build is Panda::Builder { ... }
azawawi nukes 2016.05 and rebuilds it 08:13
azawawi RabidGravy: `zef install .` exhibits the same problem btw 08:46
RabidGravy: im suspecting that the first `panda install GTK::Simple` which fails somehow corrupts installation metadata 08:47
RabidGravy: even zef uninstall GTK::Simple does not fix it
RabidGravy :_O 08:57
RabidGravy azawawi, I've just made a reset branch without that stuff at github.com/perl6/gtk-simple/tree/before-split 09:16
a get out of jail free card
azawawi RabidGravy: confirmed a fresh 2016.05 'panda install .' without doing a failed `panda install GTK::Simple` worked. It is a metadata corruption bug. 09:18
RabidGravy: sending a PR to fix the issue
RabidGravy if nothing else it will allow for a bisect to determine which part breaks it
cool 09:19
ab6tract o/ #perl6 09:19
i noticed that this section is still TBD in the docs on NativeCall: docs.perl6.org/language/nativecall#..._and_Blobs 09:20
ab6tract i was wondering if anyone knows of any ecosystem examples? 09:20
also, the zavolaj-for-examples section of this doc feels a bit out of date 09:23
tadzik wow, metadata corruption is something new indeed
RabidGravy Yeah, I think it needs some love 09:24
RabidGravy tadzik, there was a change at some point in the last few weeks that got rid of some of the locking I think 09:25
tadzik huh 09:26
ab6tract RabidGravy: do you have any experience with NC blobby/buffy interop?
i'll gladly write the section, once i've put it to some use
RabidGravy I'm not sure (too many modules) 09:27
I do know that it works however
dalek c: 501dd46 | ab5tract++ | doc/Language/nativecall.pod:
Remove tiny typo
ab6tract cool, i'll poke around your github :) 09:28
azawawi RabidGravy: travis-ci.org/azawawi/gtk-simple/b.../133062628 and ci.appveyor.com/project/azawawi/gt...uild/1.0.2 09:29
RabidGravy ab6tract, I can short cut that for you, I don't think that I have any modules that pass a Buf to a native sub directly (or any that return one) 09:33
azawawi, yeah, the tests won't run because no DISPLAY
(just grepped all my modules) 09:34
azawawi github.com/perl6/gtk-simple/pull/35 09:36
RabidGravy I'm just going to merge that without testing ;-) 09:40
dalek k-simple: ebaf0b1 | azawawi++ | META6.json:
Add missing provides. Bump version to 0.1.2
k-simple: e929c3c | azawawi++ | / (3 files):
Test on Linux/MacOSX (Travis CI, latest and 2016.05) and windows (AppVeyor, 2016.05). Add badges for AppVeyor
k-simple: b69c6aa | RabidGravy++ | / (4 files):
Merge pull request #35 from azawawi/master

Fix #34 and add mac os x and windows testing
ab6tract RabidGravy: actually, your Audio::PortMIDI does do, for an error message
RabidGravy yeah I thought it was familiar
timotimo so, can we get an x server, like Xvnc or something running on travis-ci to have actual tests? 09:41
also the %*ENV<DISPLAY> is only sensible for linux, obviously
ab6tract m: class A does Positional { has @.a handles Positional; method gist { "i'm all like aaaaaa {@!a}" } }; my $a = A.new; say $a.push: 42 09:43
camelia rakudo-moar a5c46b: OUTPUT«Cannot call push(A: Int); none of these signatures match:␤ (Any:U \SELF: |values is raw)␤ in block <unit> at /tmp/u_eMfQ4Mj6 line 1␤␤»
ab6tract i'm clearly holding it wrong, but i thought something like the above used to work 09:44
no results for 'handles' on docs.perl6
RabidGravy the way handles works it doesn't over-ridge the one it gets from 'Any' 09:46
so I don't think it ever worked
m: class A does Positional { has @.a handles <push>; method gist { "i'm all like aaaaaa {@!a}" } }; my $a = A.new; say $a.push: 42
camelia rakudo-moar a5c46b: OUTPUT«[42]␤»
azawawi RabidGravy: thx 09:47
RabidGravy I do have a completely separate problem with GTK::Siimple as it is, it hangs "forever" precompiling 09:48
which is almost certainly something wrong with some repository 09:49
azawawi RabidGravy: it takes a white 09:51
RabidGravy: s/white/while
lizmat waves from NLUUG Conference where [Tux] will give a presentation about Text::CSV in Perl 6: www.nluug.nl/activiteiten/events/v.../ab06.html 09:52
RabidGravy no, this is "forever" (like twenty minutes before I kill it,)
:) 09:54
anyway off out shopping
lizmat hmmm... RT #126391 09:58
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=126391
ab6tract RabidGravy: i could have sworn that at some point i was able to do `has %.hash handles Associative` but maybe it was only a dream... 10:12
but it's easy enough to get the semantics via handles <AT-KEY EXISTS-KEY ASSIGN-KEY> 10:13
ab6tract of course, then you need to remember <keys kv values> as well 10:17
azawawi 172 wallclock seconds after GTK::Simple refactor (classes/roles out) on 2016.05 10:18
lizmat ab6tract: not sure about that 10:20
jnthn doesn't recall that ever being something that would work
timotimo i think keys kv values happens via Any, which your class likely already derives from
jnthn You'd still need to delegate 'em to the right place though 10:21
lizmat as long you have a .pairs, I think all will derive ok
timotimo oh, right, you need to have the source of "what keys exist at all" in the original handles 10:23
we don't have a single underlying thing for that 10:24
like, no LOUD-KEBAB
DrForr Or -LOUD-ARROW-> :) 10:25
ab6tract jnthn, lizmat: okay, conceded that the memory is not based on reality. it was a good dream, though... 10:26
m: class B { has %!f handles <AT-KEY EXISTS-KEY ASSIGN-KEY pairs> }; my $b = B.new; $b<foo> = 42; say $b.keys 10:28
camelia rakudo-moar a5c46b: OUTPUT«0..0␤»
dalek k-simple: ff396b5 | finanalyst++ | README.md:
Improve README
 11:06
k-simple: 585f611 | azawawi++ | README.md:
Merge pull request #36 from finanalyst/master

Improve README
azawawi dzone.com/articles/four-reasons-wh...mming-lang # Perl 6 is mentioned in comments 11:24
grondilu frankly I'm not sure Python is that easy to learn. Sure, for basic stuff it *looks* easy, but the devil is in the details and things gets messy quite quick. I remember learning about its OO model and all the __method__ syntax for stuff and it seemed pretty complicated to me. 11:44
perlawhirl funny conversation to walk into... i was just discussing perl5 v. python to a guy at work 11:45
cosimo what is the __method__ syntax?
perlawhirl i'm comfortable with python, but some things would be downright difficult for a biginner... eg. var = sys.argc[1]... if you don't supply an arg it gives you one of pythons trademark ugly stack traces... followed by "Index Error", which for a beginner is very unhelpful 11:46
DrForr And I just had someone on FB ask me yesterday because I'm on "the Perl group" if I had any advice on how to learn Python.
perlawhirl i meant... argv, but my point stands
also! try re.compile('[') and look at the very unhelpful wall of garbled stack trace. whereas perl will tell you "Unmatched '[' in regex HERE<---" 11:48
grondilu cosimo: I hope I remember well but IIRC it's a special syntax for meta-method names 11:49
like __new__, __init__ and so on.
perlawhirl __repr__ and __str__, which are kinda like Perl 6's Str and gist methods 11:50
grondilu I used to be excited about python. It lasted about a month. Now I hate it. It's dull and verbose. To me it's the new Fortran. 11:51
azawawi github.com/azawawi/gtk-simple/blob.../README.md # finanalyst++ 11:56
cosimo it's not that bad
brrt what kind of article is that, even 12:02
i mean 12:03
psch "agile methodology is scalable" made me stumble
brrt "python is highly efficient" made me 12:04
and i /like/ python...
hmm, question
brrt if we were to put out articles of *ahem* similar quality, but with similar ferocity, would it work? 12:05
does this work, for python, to compete for mindshare?
psch it probably would invite the same quality of comments, but more and more contrarian
i'm not really sure what "work" means here, though. will people talk about it? probably 12:06
but it probably won't motivate people to honestly examine what Perl 6 brings - or rather, only very few 12:07
but i guess that's true regardless, and putting out this kind of content definitely would increase exposure
brrt exposure is usually a good thing 12:08
DrForr WTH does "Python even uses real-time wordings" mean? It's written in that Indian English dialect that makes me want to facepalm. 12:09
psch i understood it as "words that programmers use conversationally" or somesuch 12:14
although i don't quite know how "def" fits in there :)
pmurias grondilu: dull is kind of an advantage compared to some other things 12:15
DrForr Every section of that article uses circular reasoning. 12:18
moritz ... and we all know that circular reasoning is the best, because it's circular! 12:54
El_Che_ moritz: don't be a square!
woolfy Tux++ now does a presentation at NLUUG conference about Text::CSV in Perl 6. :-) www.nluug.nl/activiteiten/events/v.../ab06.html 13:03
DrForr I've got 4 proposals in for OSCON London. If I'd known last year's was in Amsterdam... 13:12
RabidGravy Oooh that's a bit special, "RAKUDO_MODULE_DEBUG=1 perl6 -Ilib -MGTK::Simple -e1' seems to hang forever on attempt to write to STDOUT 13:13
DrForr (and that I thought I'd stand a chance of getting in the door...) 13:14
gregf_ i guess the point azawawi was trying to make was ....'Perl6 is getting mentioned somewhere.. by some silly blogger'.. rather than ..'hey.. heres 4 reasons out of a million to make you feel good about python' ;) 13:26
[Coke] alexdaniel: please don't "bombard" any of our backend support staff. 13:28
[Coke] thanks everyone getting Perl 6 out on the conf. circuit. 13:49
DrForr bows. And likewise to the others. 13:50
RabidGravy I can't remember the last time I even went to a conference 14:05
_4d47 p6: my @tens = 0, 10, 20, 30; @tens[2]:delete; dd @tens; 14:07
camelia rakudo-moar 498d0a: OUTPUT«Array @tens = [0, 10, Any, 30]␤»
azawawi RabidGravy: what's a conference? :) 14:08
RabidGravy :)
DrForr A way to get your company to pay for your beer? 14:09
gregf_ we have conferences every morning....er... standups i mean ;)
azawawi RabidGravy: seems that use-ing ::NativeLib is causing the slowness in the refactor in GTK::Simple. im testing it to improve performance atm 14:10
tony-o DrForr: if you find a way for that outside of travel expenses please let me know 14:11
azawawi RabidGravy: travis-ci.org/azawawi/gtk-simple/b.../133113522 14:12
RabidGravy nicie! I'm wondering if the nqp and NativeCall can go from the main GTK::Simple if they are in all the other places they are needed 14:15
azawawi RabidGravy: im doing it :)
RabidGravy cool
azawawi but .precomp is certainly a problem in 2016.05 14:16
azawawi everytime i move subs from one class/module with atom open and run a prove... .precomp corruption 14:16
RabidGravy that's not good 14:17
llfourn azawawi: does it look like RT #128156 ?
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128156
azawawi yup same error message 14:19
llfourn yeah nine++ is looking into it 14:19
azawawi nine++ # cool 14:24
_4d47 p6: my @tens = 0, 10, 20, 30, 40; @tens[3]:delete; dd @tens 14:41
camelia rakudo-moar 498d0a: OUTPUT«Array @tens = [0, 10, 20, Any, 40]␤»
_4d47 p6: my @tens = 0, 10, 20, 30, 40; @tens[4]:delete; dd @tens
camelia rakudo-moar 498d0a: OUTPUT«Array @tens = [0, 10, 20, 30]␤»
llfourn if you wanna remove it completely use splice I think 14:42
grondilu m: say 0, 10 ... 40
camelia rakudo-moar 498d0a: OUTPUT«(0 10 20 30 40)␤»
_4d47 looks like it change .elems but only for the last elems
llfourn mmm I think that's intended 14:43
gfldex m: my @tens = 0, 10, 20, 30, 40; @tens[3]:delete; .say for @tens; 14:44
camelia rakudo-moar 498d0a: OUTPUT«0␤10␤20␤(Any)␤40␤»
llfourn m: my @tens = 0,10 ... 40; say @tens.splice(2,1); say @tens;
camelia rakudo-moar 498d0a: OUTPUT«[20]␤[0 10 30 40]␤»
ugexe m: my @a = 1,2,3,4,5; @a[1]:delete; .say for @a[0..*]; # >:) 14:46
camelia rakudo-moar 498d0a: OUTPUT«1␤»
psch ugexe++
that's pretty similar to what i was trying to hack together :)
ugexe m: my @a = 1,2,3,4,5; @a[1]:delete; .say for @a[0..4] 14:47
camelia rakudo-moar 498d0a: OUTPUT«1␤(Any)␤3␤4␤5␤»
ugexe .splice(0) follows example 1, .splice follows example 2 14:48
RT 127573
pierrot Good morning. I had asked something yesterday and CIAvash gave me an answer but without considering quoting. Which is the p6 equivalent to p5's quotemeta?
llfourn RT #127573
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=127573
psch pierrot: we don't have quotemeta. interpolation of variables inside a regex depends on how you put the variable into the regex 14:49
geekosaur pierrot, in general quotemeta isn't needed (for its intended purpose at least) because the default for interpolation is to interpolate as literal string match instead of regex. (there's an alternative interpolation syntax if you want it to be treated as a regex)
gfldex i somehow feel that representing a hole in an array with Any is wrong. Nil would fit better. Nothing is more hole-like and any possible value. 14:50
pierrot geekosaur and psch : yes, I learnt that yesterday. interpolation is enabled with <$var> 14:50
gfldex s/and/then/ 14:50
psch m: my $x = '\w'; say '\w' ~~ /$x/; say 'a' ~~ /<$x>/
camelia rakudo-moar 498d0a: OUTPUT«｢\w｣␤｢a｣␤»
psch ah, yeah
m: my @a is default(10) = ^10; @a[1] = Nil; @a[2]:delete; .say for @a 14:52
camelia rakudo-moar 498d0a: OUTPUT«0␤10␤(Any)␤3␤4␤5␤6␤7␤8␤9␤»
psch that surprised me. i had thought :delete respects the default
ugexe Nil seems like it would muck up `is default`
this time i was beaten
lizmat psch: this feels like a bug to me 14:54
but it all depends on the semantics of :delete
m: my %h is default(10); say %h<a> 14:55
camelia rakudo-moar 498d0a: OUTPUT«10␤»
lizmat hmmm 14:55
ugexe i would not expect :delete to set a default value
_4d47 find weird .elems respond differently when :delete on last item
moritz if you want that, there's an undefine() function for that
gfldex if you revert to the default, the Array loses his ability to have a hole. Unless you define Arrays with defaults to be hole-less. 14:56
moritz _4d47: if you want it to keep the element alive, you can always just set it to Nil
lizmat moritz: undefine is just another way to say = Nil
moritz right
lizmat multi sub undefine(Mu \x) is raw { x = Nil }
moritz and if you want to remove an element from the middle of the list, making the higher indexes change, you can use splice
so, whatever behavior you want, there are ways to get it :-) 14:57
there simply isn't a way to make every way behave in ways that don't surprise anybody
lizmat m: my @a is default(10) = ^10; @a[1] = Nil; @a[2]:delete; say @a[2] 14:59
camelia rakudo-moar 498d0a: OUTPUT«10␤»
_4d47 moritz: ok thanks, will use splice
gfldex m: my @a = 1,2,3; @a[1]:delete; say @a[1]; say @a[100]
camelia rakudo-moar 498d0a: OUTPUT«(Any)␤(Any)␤»
lizmat ok, so the Iterator is wrong
ugexe if your array ends with Anys they get cleaned up 15:00
psch with i, personally, find completely reasonable. otherwise .elems should always report Inf (at least on default(Any) arrays)
s/with/which/
ugexe its just combined with rt 127573 can produce the type of .elems results that seem to match _4d47's confusion 15:02
pierrot geekosaur: I'd like to do something similar to: my $pattern=join ".*", map{quotemeta} @keywords;if ($text =~ m/$pattern/) { } In p6, CIAvash suggested: my $pattern = @keyworkds.join: '.*'; if $text ~~ /<$pattern>/ { } but would fail with quoting issues. 15:04
I'd like each individual element of @keywords to be treated as literal, but the whole string (with ".*" separating each element) to be treated as a regexp and I'm not sure how to do it. 15:05
psch m: my @keywords = '\wfoo', "bar"; my $pat = "'" ~ @keywords.join("' .* '") ~ "'"; say '\wfoo and eventually bar' ~~ /<$pat>/ 15:06
camelia rakudo-moar 498d0a: OUTPUT«｢\wfoo and eventually bar｣␤»
psch pierrot: does that ^^^ help? 15:07
gfldex doc.perl6.org/language/subscripts#%3Adelete doesn't explain that holes are filled with Any. We seam not to explain (anywhere) what happens when Nil is assigned.
pierrot psch: Yes, it does. Thanks :) 15:08
gfldex m: my @a = 1,2,3; @a[1]:delete; say @a[1]:exists;
camelia rakudo-moar 498d0a: OUTPUT«False␤»
gfldex m: my @a = 1,2,3,Any; @a[1]:delete; say @a[1]:exists; say @a[3]:exists; 15:10
camelia rakudo-moar 498d0a: OUTPUT«False␤True␤»
gfldex m: my @a = 1,2,3,Any; @a[1]:delete; say @a[1]:exists; say @a[3]:exists; say .defined for @a; 15:11
camelia rakudo-moar 1ab1fb: OUTPUT«False␤True␤True␤False␤True␤False␤»
gfldex Interation turns holes into undefineness.
psch m: my @a = 1,2,3,Any; @a[1]:delete; say @a[1]:exists; say @a[3]:exists; say ""; say .defined for @a; say @a[1]:exists; say ""; say @a[3]:exists 15:21
camelia rakudo-moar 1ab1fb: OUTPUT«False␤True␤␤True␤False␤True␤False␤False␤␤True␤»
psch eh, one of the extra says for clarity is actually at the wrong spot :| 15:22
anyway, i don't think iteration changes anything
Any is just undefined
lizmat not for .defined 15:23
but it does for just the for
psch right, there was the other WAT previously
pochi m: $?USAGE
camelia rakudo-moar 1ab1fb: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YzfhVNKYiS␤Variable '$?USAGE' is not declared␤at /tmp/YzfhVNKYiS:1␤------> 3<BOL>7⏏5$?USAGE␤»
lizmat m: my @a is default(42) = ^5; @a[2]:delete; .say for @a; for @a.keys -> { say @a[$_] } 15:24
camelia rakudo-moar 1ab1fb: OUTPUT«0␤1␤(Any)␤3␤4␤Too many positionals passed; expected 0 arguments but got 1␤ in block <unit> at /tmp/Ma0GoNIFCk line 1␤␤»
lizmat m: my @a is default(42) = ^5; @a[2]:delete; .say for @a; for @a.keys { say @a[$_] } 15:24
camelia rakudo-moar 1ab1fb: OUTPUT«0␤1␤(Any)␤3␤4␤0␤1␤42␤3␤4␤»
psch yeah, that is definitely weird
psch m: my @a is default(42) = ^5; @a[2]:delete; @a[$_].say for @a.keys; 15:25
camelia rakudo-moar 1ab1fb: OUTPUT«0␤1␤42␤3␤4␤»
psch m: my @a is default(Mu) = ^5; @a[2]:delete; @a[$_].say for @a.keys; 15:26
camelia rakudo-moar 1ab1fb: OUTPUT«0␤1␤(Mu)␤3␤4␤»
dalek c: 5049338 | (Wenzel P. P. Peppmeyer)++ | doc/Language/subscripts.pod:
clarify that :delete can create holes that are bit skipped by iteration
 15:27
gfldex looks like i need more tea
lizmat m: my Mu @a = ^5; @a[2]:delete; .say for @a; say @a[2] # seems it's always Any 15:30
camelia rakudo-moar 1ab1fb: OUTPUT«0␤1␤(Any)␤3␤4␤(Mu)␤»
gfldex m: my @a = 1,Nil,3; say @a[1]:exists; .say for @a; 15:32
camelia rakudo-moar 1ab1fb: OUTPUT«True␤1␤(Any)␤3␤»
gfldex So the value that indicates the absense of a value is present?
lizmat m: my Int @a = 1,Nil,3; say @a[1]:exists; .say for @a;
camelia rakudo-moar 1ab1fb: OUTPUT«True␤1␤(Int)␤3␤»
psch gfldex: assigning Nil always means "reset to default" 15:33
lizmat yes, if it exists, there's a container for it which has the default value
m: my Int @a is default(42) = 1,Nil,3; say @a[1]:exists; .say for @a;
camelia rakudo-moar 1ab1fb: OUTPUT«True␤1␤42␤3␤»
gfldex That means that :exists doens't test for the existence of a value but for a container that could contain the value or it's default. 15:35
psch fwiw, :exists and :delete (especially on arrays) are something i am staying clear of and don't want to argue for or against :) 15:37
lizmat well, in a case of a hash, it's clear what exists means
in the case of an array, less so
gfldex Nothing is hard. :-|
gregf_ coding is :| 15:38
ugexe only 1 thing is universally accepted as not easy 15:39
sjn ugexe: cache validation and off-by-one errors? ;) 15:47
dalek c: 892e14f | titsuki++ | doc/Language/glossary.pod:
Fix a typo "gitub" to "github"
 15:48
c: fd8f366 | RabidGravy++ | doc/Language/glossary.pod:
Merge pull request #549 from titsuki/fix-typo

Fix a typo "gitub" to "github"
MadcapJake oh man, the "Python is Highly Scalable" paragraph in that article is completely bonkers
"When a client requires a large amount of development for a software application that does not require for the project to be complete, the application can be moved to live at any moment of the development process by completing a certain amount of modules of the application" O_O 15:50
DrForr The article just made up reasons.
MadcapJake seriously that one sentence is just complete gobbledygook 15:51
tadzik wat
DrForr You could s/Python/ = 'Ruby' # and noone would notice.
RabidGravy read it four times to see if there was any sense to be extracted, failed
tadzik "when a client requires for a large building to be completed, you can just ship a standalone bathroom and call it a day"
huf i think it means that if the client wants something complicated, release some half-baked shit and they wont notice 15:52
most likely
MadcapJake every read makes me go from "WHAT?" to shear laughter, it's so completely nonsense
stmuk maybe its a Markov chain
huf it's a sortof depressive, nihilist twist on "release early, release often" :D
MadcapJake stmuk: I wouldn't doubt it, it's at that level of incomprehension
huf: lol that is hilarious
huf yeah but it's kinda real 15:53
client wants a bunch of undefined crap, you complete what you were able to figure out, the client sends incomprehensible feedback, you react by doing *something* 15:54
MadcapJake I like dark comedy so the more real, the more I'll laugh :)
RabidGravy yeah, it's aiming at rapid application development and continuous integration but somehow missing explaining either of those things
huf eventually the client just quiets down and goes away
psch "no bug reports means no bugs, right?"
MadcapJake RabidGravy: I think there's an element of modularity that the author is trying to describe too
huf psch: correction. no comprehensible bug report = no bug report
psch huf: that sounds more like corollary :) 15:55
MadcapJake I could seriously start a tumblr of some of the crazy nonsense people throw around in these "X reasons that X language is the BEST!" articles 15:56
psch .oO( Perl 6 reasons that Perl 6 language is the BEST )
MadcapJake lol 15:57
"Almost all of the syntax is very similar to "real-life" words." Oh. My. God. This article is a gold mine. 15:58
huf :D
when i ()()()() this morning i ()() ==
MadcapJake Ok, I seriously can't handle it. I'm starting a tumblr. Someone else has got to get as much joy out of this as I do. 15:59
RabidGravy :) 16:01
skids glfdex: Maybe link to docs.perl6.org/type/Nil when explaining :delete? 16:06
lizmat +1 from me 16:08
gfldex it's one sentence to tell the difference between assigning Nil and :delete 16:09
lizmat gfldex: but the only difference *should* be that :delete will make :exists for that element False 16:11
lizmat it's only the iterator on Array that seems wrong (and I'm looking into fixing that) 16:12
azawawi RabidGravy: re GTK using only ::App, ::VBox and ::Button in 01-hello-world.pl6 => ~36 second (down from ~170 sec) parse time, ~3 second runtime (second time)
tadzik that's still a crazy parsetime 16:13
azawawi so moving classes/roles out of ::Simple 16:14
shows a big performance bug in rakudo
dalek c: 3d247b7 | (Wenzel P. P. Peppmeyer)++ | doc/Language/subscripts.pod:
clarify interaction between assignment of Nil and subscripts
 16:15
azawawi i will try switching to moar-nom to verify whether it is fixed or not 16:16
RabidGravy azawawi, all good stuff though 16:17
gfldex lizmat: i'm not sure anymore of iteration of holes is wrong right now. If autoviv for elemets addressed with a subscript returns the default value, a hole in an Array should do so too. They are Arrays not linked lists.
RabidGravy azawawi, it could probably be shrunk a little by removing all of the unused native subs too 16:19
azawawi RabidGravy: im working on it github.com/azawawi/gtk-simple/commits/master :) 16:20
RabidGravy looks good 16:22
ZoffixW RabidGravy, which RPi did you install Perl 6 on? I'm wondering whether it'll work at all with Pi Zero 17:22
azawawi is there any performance advantage to using need or require over use?
ZoffixW wants to make his A/C Internet Connected :P
azawawi, highly doubt it, since you're loading the module at runtime 17:23
RabidGravy ZoffixW, I've got it installed on both a Rev. B and Pi 2, the former is way slow
azawawi ZoffixW: so we cant use 'require' to lazily load a module anymore, right? 17:24
ZoffixW azawawi, oh, I think that'll work just fine. And yeah, in that case it'll probably be a performance advantage, since you don't have to load the module when you don't need it 17:25
psch conditional loading isn't quite lazy loading, is it? 17:26
RabidGravy "but" any symbols defined in the module won't be known at compile time, so you may have to jump through hoops
ZoffixW Then I don't know what lazy loading is :)
psch might just be overly picky there
azawawi so github.com/azawawi/gtk-simple/blob...Simple.pm6 is a bad practice 17:27
psch i admit that "lazy loading" as "runtime loading" is probably usually used for what i called "conditional loading" just now
e.g. "when we need it, we load it"
ugexe lazy loading (like what was attempted at one time) would load the symbols and everything still
psch ...that doesn't really clarify i think
more like "when we know we are going to use it, we load it explicitly"
whereas lazy loading to me reads more as "we know we will use it eventually, but until we do we don't load it" 17:28
which isn't what <use> currently does
mostly because of RabidGravy++'s point above
Juerd I find 'lazy loading' quite vague as a term. "On-demand loading" is much clearer but I never see that used.
Maybe it's just me
psch Juerd: no, i do agree. i think that's a good alternative
the distinction of "the language user does it manually" vs "the runtime does it automagically" is what i'd but between on-demand loading and conditional loading 17:29
...that's chiastic btw :)
azawawi compares performance with github.com/perl6/gtk-simple/tree/before-split 17:30
ugexe i just hide all the hoop jumping inside a (caching) service locator
pmurias what would be a good way in JavaScript to make a method call on a Scalar be passed on to it's value? 17:36
a Proxy would work although I read somewhere it has performance issues on v8 17:38
jnthn pmurias: A method call on a Scalar is a call on the Scalar itself; it's that a bunch of ops imply a decont on a certain argument that makes it happen on the underlying value 17:40
pmurias: So they'd compile into a decont then a method call on the value
Or at least, that's how it happens elsewhere :)
azawawi RabidGravy: Before split => 24s startup, 3s runtime. After split => 2m52s startup, 11s runtime, Optimized => 1m57s startup, 11s runtime for 01-hello-world 17:43
RabidGravy if nothing else it has increased the amount of i/o involved 17:45
azawawi SSD here
github.com/perl6/gtk-simple/pull/37 17:48
azawawi movie night & 17:49
pmurias jnthn: thanks, got confused and what I planned would have made it impossible to call methods on the container itself 17:59
dalek k-simple: 6e69610 | azawawi++ | .travis.yml:
Install gtk3 on MacOS X Travis CI
 18:00
gtk-simple: e48a477 | azawawi++ | README.md:
gtk-simple: More documentation in README
sena_kun hi, #perl6. Is it achievable to pass a hash with arguments to a class constructor to create object? Class has many fields and pass them all to bless sub will be ugly, really ugly. 18:03
perlpilot sena_kun: Thingy.new(|%stuff)
sena_kun perlpilot, thanks! 18:04
smls_ sena_kun: Be careful with that though. If the Hash contains Array value, they will not initialize an `@.array` attribute the way you'd expect. 18:11
The `@.array` attribute will end up with a single element, which is in turn the Array value from the Hash. 18:12
smls_ And unfortunate consequence of the facts that Hashes containerize their arguments, and `bless` uses assignment rather than binding to attributes. 18:13
*An
psch m: class A { has @.a }; say A.new(:a[1,2,3]).perl 18:23
camelia rakudo-moar 1ab1fb: OUTPUT«A.new(a => [1, 2, 3])␤»
psch m: class A { has @.a }; say A.new(:a[1,2,3]).a.perl 18:24
camelia rakudo-moar 1ab1fb: OUTPUT«[1, 2, 3]␤»
psch m: class A { has @.a }; say A.new(:a[1,2,3]).a[1].perl
camelia rakudo-moar 1ab1fb: OUTPUT«2␤»
psch m: class A { has @.a }; say A.new(my % = a => [1,2,3]).a[1].perl
camelia rakudo-moar 1ab1fb: OUTPUT«Default constructor for 'A' only takes named arguments␤ in block <unit> at /tmp/Os6iMZ_WlK line 1␤␤»
psch m: class A { has @.a }; say A.new(|(my % = a => [1,2,3])).a[1].perl
camelia rakudo-moar 1ab1fb: OUTPUT«Any␤»
psch smls_++
m: class A { has @.a }; say A.new(|(my % = a => |[1,2,3])).a[1].perl 18:32
camelia rakudo-moar 1ab1fb: OUTPUT«Any␤»
psch yeah, there's probably no trickery possible there
psch well, aside from a custom BUILD probably 18:32
or new, or whichever vOv
ugexe i flatten inside a BUILDALL :/ 18:45
azawawi RabidGravy: ping 19:09
RabidGravy erp
azawawi RabidGravy: best optimization so far is 'use GTK::NeededModule' instead of 'GTK::Simple' 19:10
RabidGravy: 2s when using the modules needed
RabidGravy yeah, I figure that would 19:11
azawawi builder pattern for the rescue? :) 19:16
AlexDaniel [Coke]: perhaps you are right and I shouldn't've said that. However, there is a problem that is not going anywhere, what's your solution? 19:23
dogbert17 o/ #perl6 19:32
is anyone interested in taking a look at an, admittedly feeble, attempt document the 'default' method in the Hash class? 19:33
dogbert17 it's here gist.github.com/dogbert17/20650772...33d5df0f91 19:34
nicqbot It seems ok, for the 'Defined' part I would put "... returns Any or initalized value/object". Something like that. 19:39
dogbert17 nicqbot: thx 19:41
AlexDaniel m: my %h; %h.default = 42; say %h<abc> 20:13
camelia rakudo-moar 1ab1fb: OUTPUT«Cannot modify an immutable Any␤ in block <unit> at /tmp/m4nXmIulAJ line 1␤␤»
Emeric1 Are the Data::Dump tests broken ? I can't install it with panda 20:17
Emeric1 Is there a way to avoid erros durings the tests with panda ? 20:18
[Coke] AlexDaniel: what's the problem? 20:28
Emeric1: there's a force option, IIRC.
grondilu just received an email from Jeff Bezos 20:34
oops sorry wrong chan 20:35
[Coke] backlogs. 20:37
tbrowder bye 20:38
[Coke] .tell AlexDaniel so, if this happens again, have that person send an email, but be sure to reference that it's happened before (with ticket ids or account names if possible) and point out that it might be a trend.
yoleaux [Coke]: I'll pass your message to AlexDaniel.
pochi m: sub foo() returns Array of Int { return 1,2,3 }; foo() 20:38
camelia rakudo-moar 1ab1fb: OUTPUT«Type check failed for return value; expected Array[Int] but got List ($(1, 2, 3))␤ in sub foo at /tmp/61nLnD08VA line 1␤ in block <unit> at /tmp/61nLnD08VA line 1␤␤»
pochi is it possible to add a return signature for listy things? 20:39
something that catches both return 1,2,3 and return @x
ugexe m: sub foo(--> Array) { @ = 1,2,3 }; foo() # might work for you 20:42
camelia ( no output )
pochi I'd like to constrain it to a type though 20:43
ugexe m: my List of Int $list = 1,2,3 20:46
camelia rakudo-moar 1ab1fb: OUTPUT«5===SORRY!5=== Error while compiling /tmp/TgWDraLGuB␤List cannot be parameterized␤at /tmp/TgWDraLGuB:1␤------> 3my List of Int7⏏5 $list = 1,2,3␤»
ugexe otherwise wrap/multi-dispatch to check the return value and do any neccesary transformation into a single return type 20:47
Xliff roles can't inherit from a base class, correct? 20:55
dalek c: fda109d | (Jan-Olof Hendig)++ | doc/Type/Hash.pod:
Added docs for method default in class Hash
 20:56
ugexe m: class Foo { method x { 1 }; }; role Bar { method xx { $.x; }; }; class Foo2 does Bar { }; say Foo2.new.xx 20:57
camelia rakudo-moar 1ab1fb: OUTPUT«Method 'x' not found for invocant of class 'Foo2'␤ in method xx at /tmp/5sEwBHTCqJ line 1␤ in block <unit> at /tmp/5sEwBHTCqJ line 1␤␤»
ugexe m: class Foo { method x { 1 }; }; role Bar { method xx { $.x; }; }; class Foo2 is Foo does Bar { }; say Foo2.new.xx 20:58
camelia rakudo-moar 1ab1fb: OUTPUT«1␤»
Xliff ugexe: Hrm. I might need to go into the long explanation, here. However, this is still half baked stuff. 20:59
Maybe later. I need to give this more thought. 21:00
ugexe for the most part a role body is simply pasted into the class body 21:01
Xliff Right. OK. I have a class A::Base, A::A and A::B. A::A and A::B are functionally unrelated, but both inherit from A::Base 21:04
However A::B is more complex than A::A and has several "sections" I want to split into roles.
However the roles still need methods from A::Base.
What's the best way to handle that? Role parameterization?
ugexe because A::B is A::Base you should just be able to compose the roles into A::B as expected. you may ask yourself why you want to compose the parts into A::B at all if they aren't otherwise reusable or if A::B will always use the same parts 21:08
AlexDaniel [Coke]: I think that it's not just a trend, but any new account has this problem
yoleaux 20:38Z <[Coke]> AlexDaniel: so, if this happens again, have that person send an email, but be sure to reference that it's happened before (with ticket ids or account names if possible) and point out that it might be a trend.
AlexDaniel [Coke]: I'll create a new account to prove that…
Xliff ugexe, A::B will always use the same parts, but I'm hoping I don't have to maintain a large, monolithic file for A::B. 21:08
Xliff And I get the reusable part. However if I have to put all of the A::B methods into a single file, it will get really large. 21:09
Xliff I was hoping to split it into related, manegable chunks. 21:10
ugexe im not sure splitting a single functional class into multiple files helps with maintenance. it certainly makes grepping the code path more dificult
Xliff You have a point, there.
ugexe but if you want to it should work the way i think you want it to
Xliff OK. Another way, then.
If I split A::B into different classes. A::B::A … A::B::<n>, can I still have a class A::B that wraps the A::B::x classes in an a-la-carte manner? Kinda like the perl5 "use Module qw(<list>)" 21:12
ugexe the managable chunks are typically subroutines/methods themselves
AlexDaniel [Coke]: created a new account, problem not reproduced :/ 21:13
ugexe im not sure i follow. you can `class Foo is Foo::A is Foo::B { }`
AlexDaniel [Coke]: dunno then… alright, sorry for your time
ugexe you can also `does Foo::A does Foo::B` 21:14
Xliff Right.
However I'm trying to see if I can do A::B without having it pull in all of the classes in the A::B::* space. Just the ones I need to use for a particular purpose. 21:15
I'm probably mixing metaphors.
ugexe github.com/ugexe/Perl6-Grammar--HT...TP.pm6#L14 # this is split up into parts with roles, but i can also attest its hard to follow the code path around
you are talking about dependency injection/inversion type stuff 21:17
Xliff OK, thanks. I think I at least have something to start from.
Xliff ugexe, I guess. Yeah. 21:17
ugexe you'd need to use a service locator pattern type deal to call .^add_role(whatever)
Xliff Is there .^add_class()? 21:18
ugexe ^add_parent maybe
Xliff That works. Thanks! 21:20
Zoffix Worth mentioning in the Weekly (cc lizmat). Damian Conway will be doing two days of Perl 6 talks at Toronto Perl Mongers on June 27 and June 28: www.meetup.com/Toronto-Perl-Monger...231418224/ www.meetup.com/Toronto-Perl-Monger...231418278/ 21:31
Xliff next question: How do you pass parameters to superclass constructor? 21:34
YAY! perl6advent.wordpress.com/2010/12/...s-cousins/ 21:36
lizmat .tell Zoffix noted 22:01
yoleaux lizmat: I'll pass your message to Zoffix.
Xliff ARGH! 22:18
No https support for HTTP::Client, yet.
AlexDaniel Xliff: do you want to just download some page? 22:43
Xliff: perhaps try HTTP::UserAgent 22:46
Xliff AlexDaniel++ 22:47
Thanks!
AlexDaniel Xliff: I've also mentioned LWP::Simple here: github.com/supernovus/perl6-http-c.../issues/11 22:48
Xliff: but I don't think that I ever used it. I think HTTP::UserAgent always worked for me
Xliff Thanks. I was using HTTP::Client, before. 22:58
dalek Iish: 2bacdf9 | (Salvador Ortiz)++ | t/36-pg-native.t:
Keep a column-info example
 23:24
Iish: 2d273a2 | (Salvador Ortiz)++ | / (3 files):
mysql: Fix old typo in $dbh.quote-identifier

  kuerbis++ for spot that.
Anyway the Connection role implementation is not robust enough, so add a native supported .quote(:$as-id) following the Pg approach. Closes #66
Iish: cf198ad | (Salvador Ortiz)++ | / (2 files):
For panda's/zef's users, now v0.5.8
kalkin- hi 23:37
is there some clever way to detect if a string consists of only unique characters?
m: "CasdfghjkC234" ~~ /.*(.).*$0.*/ 23:41
camelia ( no output )
geekosaur you probably want to prefix that with `say so`
kalkin- m:say so "CasdfghjkC234" ~~ /.*(.).*$0.*/ 23:42
geekosaur m: say so "CasdfghjkC234" ~~ /.*(.).*$0.*/
AlexDaniel space
camelia rakudo-moar 1ab1fb: OUTPUT«True␤»
kalkin- Ahh thanks
so regex is probably the shortest solution for that 23:43
AlexDaniel I don't think so
sortiz kalkin-, that fails, 'C' isn't unique.
kalkin- m: say so "CasdfghjkC234" !~~ /.*(.).*$0.*/ 23:44
camelia rakudo-moar 1ab1fb: OUTPUT«False␤»
kalkin- m: say so "Casdfghjkc234" !~~ /.*(.).*$0.*/ 23:45
camelia rakudo-moar 1ab1fb: OUTPUT«True␤»
raydiak the leading and trailing .* is unnecessary
AlexDaniel m: say so ‘abcde’.comb.Bag.values.all == 1 23:46
camelia rakudo-moar 1ab1fb: OUTPUT«True␤»
AlexDaniel m: say so ‘abcdce’.comb.Bag.values.all == 1
camelia rakudo-moar 1ab1fb: OUTPUT«False␤»
AlexDaniel that's it
kalkin- m: say so "Casdfghjkc234" !~~ /(.).*$0/
camelia rakudo-moar 1ab1fb: OUTPUT«True␤»
kalkin- raydiak: Thanks 23:47
raydiak you're welcome :)
kalkin- is going to read the docs to understand what happens in AlexDaniel example
Of course thanks to all :)
AlexDaniel it also depends on how efficient you want it to be 23:49
and whether you are dealing with long strings or not
kalkin- AlexDaniel: the normal use case should be < 100 chars 23:50
AlexDaniel … and I have a feeling that regex solution is much slower than any other

Powered by Cro and the Raku® Programming Language.

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

Thank you!