»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.
RabidGravy release the goats 00:02
RabidGravy toodles people! 00:07
average I was looking at New Zealand's TZ 00:26
it's GMT+12:45
average that kinda throws off some of my calculations 00:26
does it make sense if I consider .nz to be on GMT-11:15 ? 00:27
or is it unsound ? 00:28
btw, TZs never cease to amaze me en.wikipedia.org/wiki/List_of_UTC_time_offsets
leont That would be a day off
average turns out, they span from GMT-12 to GMT+14
that means they cover 26 hours, that's very weird 00:29
leont Kiribati has +14 because a country being split by the day separator is suboptimal
average I would expect the distance between the earliest and the latest timezone to be at most 24h
average that's how much time it takes for Earth to revolve around its axis 00:30
leont: is it suboptimal ?
leont: or are they just pampered and wanted to be different ?
leont Euphemisms, I like using them ;-)
average if someone asked me I would just tell them Kiribati is full of hipsters
"Oh we're on this timezone called GMT+14, you've never heard of it" 00:31
leont Try -2:00, that's the one no one ever heard of
average leont: because it's in the ocean ? 00:32
leont Population estimate of Fernando de Noronha and South Georgia and the South Sandwich Islands combined: 51 00:33
average leont: but they made sure they got their own timezone 00:34
those 51 people must be super-important
AlexDaniel m: say DateTime.new(‘2015-12-24T12:23:00+69:42’) 00:37
camelia rakudo-moar 3259ba: OUTPUT«2015-12-24T12:23:00+69:42␤»
AlexDaniel m: say DateTime.new(‘2015-12-24T12:23:00+69:42’).timezone 00:37
camelia rakudo-moar 3259ba: OUTPUT«250920␤»
AlexDaniel what's that 00:38
m: say DateTime.new(‘2015-12-24T12:23:00+02:00’).timezone
camelia rakudo-moar 3259ba: OUTPUT«7200␤»
AlexDaniel ah right
leont timezone is the wrong word for that, really :-/ 00:39
AlexDaniel average: what kind of calculations are you doing? :)
leont: yeah, it is clock offset
leont: why do we even have this crap
AlexDaniel the whole assumption that the sun should be shining with some specific brightness at some specific hours is just not relevant anymore. We don't use sundials, right? 00:43
AlexDaniel though these are separate issues 00:45
leont Humans and habits… 00:46
AlexDaniel what completely escapes me is how humans strive to improve all aspects of life except calendars and clocks 00:49
every time I look at monthly stats I wonder “was it adjusted according to the number of days in the month or not?” 00:50
come on people, we can do better
joydo Can somebody help me explain the dynamic variable "$*REPO" ? or some documents I can refer to!!! 00:53
AlexDaniel m: say $*REPO 00:54
camelia rakudo-moar 3259ba: OUTPUT«inst#/home/camelia/.perl6/2015.12-212-g3259ba3␤»
AlexDaniel m: say ‘🖕’.uniname 01:01
camelia rakudo-moar 3259ba: OUTPUT«REVERSED HAND WITH MIDDLE FINGER EXTENDED␤»
average does Perl6 have a monotonic clock ? 01:05
you need a monotonic clock guys..
leont Not sure that has to be part of the language, but they're sure darn useful 01:18
autarch and a pentatonic scale - need that too 01:20
TimToady wonders what "monotonic" means there 01:22
skids Well, the definition is solid in non-parallel applications, at least.
leont Steadily increasing 01:23
As opposed to wallclock which follows human time, and can move around at will
TimToady our instants are certainly monotonic in that sense
leont monotonic is what you'd use for a benchmark, wallclock is for end-users
That doesn't mean the clock that generates them is
skids Oh, well that wasn't the definition I thought then. 01:24
TimToady that's a problem with the computer, not the language
leont Time is difficult :-/
It's neither, or both, but probably neither. It's time.
TimToady Instants are supposed to be sync'd to TAI to the extent possible to determine on the computer
leont You're missing the point 01:25
Unless your computer is/has an atomic clock, it doesn't have a monotonic timesource that's relatable to human time (and even one could argue TAI is an average)
A typical monotonic clock is based on seconds since boottime, since that won't skip when the user decides their clock is wrong. 01:26
TimToady a user is not supposed to be able to change Instants at their whim 01:29
but I see that backtranslating from the system clock can be a problem 01:30
we could peg Instants to a monotonic clock at the beginning of a process, but then we could have concurrent processes that disagree on the current Instant by a significant degree 01:33
there's no easy answer that solves every problem
average linux.die.net/man/3/clock_gettime 01:35
see the part about CLOCK_MONOTONIC
skids Well, that resets on reboot. 01:37
average "continuity is more important than having actual time" (I think he means monotonicity is more important than actual time) www.youtube.com/watch?v=-5wpm-gesOY#t=9m13s 01:39
average he mentions Google has a clock that smears leap seconds all throughout the day 01:39
I suppose they are doing that in the attempt of having an actual monotonic clock that is not affected by leap seconds 01:40
"A leap second is a one-second adjustment that is occasionally applied to Coordinated Universal Time (UTC) in order to keep its time of day close to the mean solar time, or UT1."
average ^^ and that's a leap second 01:40
TimToady we'd rather have another second in our monotonic time than to have seconds vary in length 01:41
mspo that assumes monotonic means monotonic in your OS
yoleaux 15 Jan 2016 22:11Z <sortiz> mspo: I'm working in a p6-LMBD
mspo sortiz: nice
TEttinger they do though, if you're working at a different speed
approaching the speed of light, seconds do vary in length 01:42
TimToady we aren't worried about that for Perl 6 01:42
TEttinger haha
100 years not 100,000
TimToady TAI is good enough for earthbound earthlings :)
and if it's not, you're gonna have your own clock already :) 01:43
TEttinger ha yes
skids Eh, if reality is only partially ordered in the first place a Lamport timestamp should do just fine. 01:46
average skids: lamport is that distributed systems guy ? 01:49
skids: he made that clock for distributed systems right ?
skids Yeah. One of the distributed partially ordered causality schemes. 01:50
leont If you want monotonic time, you probably don't need to map it to human time. Instants don't need to have a meaningful value, you'd only use them to create Durations 01:56
TimToady that's pretty much how they're defined; epoch-less atomic seconds 01:57
people keep wanting to confuse them, however :) 01:58
leont They do! 01:58
Herby_ Good evening, everyone! 03:01
Herby_ m: comb(<[a..z]>, "h.e.l.l.o.").say 03:02
camelia rakudo-moar 3259ba: OUTPUT«()␤»
Herby_ m: comb(/./, "h.e.l.l.o.").say
camelia rakudo-moar 3259ba: OUTPUT«(h . e . l . l . o .)␤»
Herby_ how would I go about stripping punctuation out of a string?
Herby_ m: comb(/<[a..z]>/, "h.e.l.l.o.").say 03:03
camelia rakudo-moar 3259ba: OUTPUT«(h e l l o)␤»
Herby_ :)
autarch I could use some help debugging a confusing warning in github.com/autarch/perl6-Test-Stream
Herby_ m: comb(/<[a..z]>/, "h.e.l.l.o.").join.say 03:04
camelia rakudo-moar 3259ba: OUTPUT«hello␤»
Herby_ m: comb(/<[a..z]>/, "hello. this is a test string. \n and another line!").join.say
camelia rakudo-moar 3259ba: OUTPUT«hellothisisateststringandanotherline␤»
Herby_ wish I could help, autarch, but I'm still fumbling my way through the basics 03:05
autarch hmm, now maybe the warning went away ...
alright, that's odd 03:06
Herby_ m: $string = "Was it a car or a cat I saw?"; if $string.lc.comb(/<[a..z]>/) eq $string.lc.comb(/<[a..z]>/).reverse { say "Its a palindrome" }; 03:07
camelia rakudo-moar 3259ba: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Ce35O6S5QM␤Variable '$string' is not declared. Did you mean 'Stringy'?␤at /tmp/Ce35O6S5QM:1␤------> 3<BOL>7⏏5$string = "Was it a car or a cat I saw?"␤»
Herby_ m: my $string = "Was it a car or a cat I saw?"; if $string.lc.comb(/<[a..z]>/) eq $string.lc.comb(/<[a..z]>/).reverse { say "Its a palindrome" }; 03:08
camelia rakudo-moar 3259ba: OUTPUT«Its a palindrome␤»
autarch ok, here's the warning - check out the repo, then run ... perl6 -Ilib -e 'use Test::Predicates' 03:10
and the warning I get is - "Use of uninitialized value of type Mu in string context
Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in code at /home/autarch/projects/perl6/Test-Stream/lib/Test/Predicator.pm6 line 66"
but what's so confusing is that this happens at compilation time and this code is not being run
the warning used to be repeatable, but now it seems to only happen sometimes, if I run that command repeatedly it doesn't always show up 03:11
autarch I've got to stop waiting to ask questions til after the europeans have all gone to sleep 03:20
skids try --ll-exception 03:21
autarch skids: it's just a warning - will that still help? 03:23
skids maybe not.
autarch on another topic, how do I fork?
skids autarch: as it so happens, I braindumped what I do to send PRs last night at github.com/perl6/user-experience (the two "Rakudo" links in the README.md) 03:27
autarch skids: I really think travis should be doing this - the point is to make it easier for core devs to accept/reject PRs 03:28
I mean, it's good to get people to run the tests themselves too, but the CI system if the protection against lazy people like me who always forget
skids Yeah, I havent had the time to stat to learn travis yet, not being a actual "developer" really. 03:29
skids But It's good to tell people what travis would be testing anyway so they can dig into what is failing. 03:30
I can't seem to get that warning to happen here. 03:32
autarch skids: I just pushed a commit that may have made it go away 03:35
the last commit that I know produced the warning is e0083d5c31a3b28d070fbdb506eee5a30a5542ad
skids Yeah that made the difference. 03:36
autarch but what did I change? that's what's so confusing - I basically copied that code from Test.pm6 and that one never warns 03:38
I wonder if it's some very odd compiler bug
skids It seems to happen when some other file under lib/ has been recently touched and needs to be recompiled, other than Predicates.pm 03:39
skids Predicator.pm6 actually 03:40
autarch I just don't understand the warning 03:45
skids m: Mu.Str 03:48
camelia rakudo-moar 3259ba: OUTPUT«Use of uninitialized value of type Mu in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block <unit> at /tmp/WbOEF2_20y line 1␤»
dalek kudo-star-daily: e4d2e7d | coke++ | log/ (6 files):
today (automated commit)
 03:53
skids m: sub f ($op) { qq{&infix:«$op»}; } 04:00
camelia rakudo-moar 3259ba: OUTPUT«Use of uninitialized value of type Mu in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in code at /tmp/WdRUE5CITk line 1␤»
skids golfed so far.
only happens on precomp.
erm, when precomping
autarch yeah, that's just really odd, unless & is interpolated somehow? 04:01
ah, maybe that's the issue?
autarch yeah, I think it is 04:01
m: say qq{&infix<==>}
camelia rakudo-moar 3259ba: OUTPUT«5===SORRY!5=== Error while compiling /tmp/OMUWEf_M2p␤Undeclared routine:␤ infix used at line 1. Did you mean 'indir', 'index'?␤␤»
autarch m: say qq{\&infix<==>}
camelia rakudo-moar 3259ba: OUTPUT«&infix<==>␤»
skids missing colon 04:02
autarch m: say qq{&infix:<==>}
camelia rakudo-moar 3259ba: OUTPUT«&infix:<==>␤»
autarch hmm, no clue then
m: say qq{\&infix:<==>}
camelia rakudo-moar 3259ba: OUTPUT«&infix:<==>␤»
autarch m: sub foo { 42 }; say "&foo" 04:03
camelia rakudo-moar 3259ba: OUTPUT«&foo␤»
autarch m: sub foo { 42 }; say "&foo()"
camelia rakudo-moar 3259ba: OUTPUT«42␤»
autarch yeah, definitely interpolated
skids m: sub f ($op) { qq{&infix:($op)}; }; f(">").perl.say
camelia rakudo-moar 3259ba: OUTPUT«"\&infix:(>)"␤»
skids m: sub f ($op) { qq{&infix:«$op»}; }; f(">").perl.say
camelia rakudo-moar 3259ba: OUTPUT«Use of uninitialized value of type Mu in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in code at /tmp/TzD4Y9bjWs line 1␤"\&infix:«>»"␤»
skids m: sub f ($op) { qq{&infix:<$op>}; }; f(">").perl.say 04:04
camelia rakudo-moar 3259ba: OUTPUT«"\&infix:<>>"␤»
autarch m: say &infix:<>>.perl 04:08
camelia rakudo-moar 3259ba: OUTPUT«Use of Nil in string context in any canonicalize_pair at src/Perl6/World.nqp line 4039␤Use of Nil in string context in any canonicalize_pair at src/Perl6/World.nqp line 4045␤5===SORRY!5===␤Undeclared routine:␤ infix:<> used at line 1. Did …»
autarch you need to use «>»
TimToady infix references evaluate the name as a constant at compile time, so you can't parameterize it like that 04:21
skids Right but why is it interpolated? 04:22
(the actual code this was golfed from used EVAL on it after makin the string)
average hey kids, did you eat your vegetables today ? 04:23
Herby_ o/ 04:26
average \o 04:28
Herby_ \o/ 04:28
skids m: "&abs(1)".perl.say; # it's hitting the "looks like a &-qualified function call" 04:30
camelia rakudo-moar 3259ba: OUTPUT«"1"␤»
skids ...but it isn't
autarch skids: I think it's interpolated because you can do this ... 04:36
m: sub foo { 42 }; say "&foo()"
camelia rakudo-moar 3259ba: OUTPUT«42␤»
skids Right that's what I eant with the "&abs(1)" line. 04:37
I think (?) its a bug
autarch no idea
m: sub foo { 42 }; say "foo()"
camelia rakudo-moar 3259ba: OUTPUT«foo()␤»
autarch it seems like it might be intentional, since it's not that likely to be typed by accident
skids It's intentional that "&foo()" works but probably not that "&infix:«" without a following () tries to do the same thing 04:38
m: "&infix:<+>(1,2)".say 04:39
camelia rakudo-moar 3259ba: OUTPUT«3␤»
skids m: "&infix:«+»(1,2)".say
camelia rakudo-moar 3259ba: OUTPUT«3␤»
skids m: "&infix:«+»".say
camelia rakudo-moar 3259ba: OUTPUT«&infix:«+»␤»
autarch skids: design.perl6.org/S02.html#Sub_calls 04:40
yeah, that lack of parens means it shouldn't interpolate according to that synopsis
skids files a rakudobug
skids RT#127284 04:53
average m: say "Eat your veggies kids, very important" 04:54
camelia rakudo-moar 3259ba: OUTPUT«Eat your veggies kids, very important␤»
autarch skids++ 04:58
AlexDaniel I have to admit that this one is pretty clever: tinyurl.com/juynybx 06:22
though I'm not sure if it is still a butterfly 06:23
dalek osystem: eaedef1 | azawawi++ | META.list:
Add OpenCV to ecosystem
 06:53
st_iron good morning 07:22
azawawi hi 08:26
how does one force panda install XYZ ? --force does not seem to do it
nevermind, 'zef install XYZ' works perfectly and skips installed modules :) 08:34
Zero_Dogg Is there any way to have run wait until the command exits, like system() did in perl5? 09:10
nine Zero_Dogg: which command? 09:14
Zero_Dogg nine: whatever is being executed with run() 09:15
azawawi Zero_Dogg: design.perl6.org/S29.html#shell ?
nine Unless you use any options run() does wait for the command to exit 09:16
Zero_Dogg ah, ok, the docs (doc.perl6.org/type/IO#sub_run) was not clear on that, and Proc was described as «running process», so I assumed it referred to the process still running 09:17
azawawi Zero_Dogg: Proc::Async is the one that does not run if i remember correctly 09:22
RabidGravy Proc::Async runs things and returns before they complete, straight Proc (i.e run() ) doesn't
azawawi Zero_Dogg: i.e. async
Zero_Dogg thanks
RabidGravy :)
azawawi RabidGravy: :)
azawawi retupmoca: ping 09:52
[Tux] csv-ip5xs 50000 17.653 12.150 10:01
test 50000 22.528 21.830
test-t 50000 12.284 11.400
csv-parser 50000 50.566 -0.131
timotimo o/ 10:18
azawawi timotimo: hi there 10:23
modules.perl6.org/#q=Opencv # is finally working :) 10:24
github.com/azawawi/perl6-opencv/bl...o-world.p6 # hello world
timotimo neat :) 10:25
azawawi now it is time to do OO wrappers :) 10:27
and more API wrapped
timotimo bubble wrap!
azawawi github.com/abidrahmank/OpenCV2-Pyt...ic_ops.rst
timotimo: any progress on your ipython p6 kernel? 10:28
RabidGravy cool, azawawi++
timotimo no :( 10:29
i haven't even started looking at arnsholts work in that repo much :(
azawawi timotimo: metacpan.org/release/Devel-IPerl # Perl 5 ipython kernel 10:33
lucs In Synopsis 26 mentions ｢=head86 A "Missed it by I<that> much!" heading 10:37
｣
And it makes me smile :)
DrForr s/head/agent/ 10:38
lucs Yeah, I grew up with that :) 10:39
timotimo i don't know what that means :<
lucs Old running gag in an old TV show: youtu.be/oPwrodxghrw 10:40
DrForr /Get Smart!/ ref.
timotimo ah 10:41
i saw the movie they put out recently-ish
lucs He was Agent 86.
timotimo it was amazing 10:42
masak good antenoon, #perl6 10:43
azawawi do we have something like python's "import something as st" ?
masak: hi
RabidGravy marning masak 10:44
azawawi, you could do it with an EXPORT sub 10:45
azawawi RabidGravy: do you have an example?
RabidGravy not in my code, but basically the EXPORT sub returns a Map with symbol => thing pairs 10:47
timotimo um, i don't think that's what he's asking for? 10:48
isn't what he wants to have something used and installed under a different name than it itself wants?
RabidGravy well, it can do that
it doesn't have to be just subroutine, it can be types too 10:49
timotimo we're not going to go around telling people "you have to write an EXPORT subroutine so that others can use your module nicely" 10:49
nine S11 speculates about more possibilities, but I don't think we're even close to that. 10:51
RabidGravy well no, but it was a short-cut for "no there is no built in mechanism as yet, but you could emulate it with an EXPORT" :) I'm sure azawawi's clever enough to read between the lines ;-) 10:53
anyway, off to the seasidde before the impending snowpocalypse 10:55
timotimo i didn't realize we don't do that at all yet :|
azawawi im clever enough? :) 11:09
azawawi eats tea + cake :)
i was asking for a nice way to namespace (prefix) a module exports 11:10
DrForr BUT THE CAKE... never mind.
azawawi github.com/azawawi/perl6-opencv/bl...o-world.p6 # a bit of OpenCV::Mat Perl 6 OO 11:38
zhmylove masak: Hi! my $c = 1; for @*ARGS {qqx"curl -o $c++.fmt('%02d').htm $_.subst("'","\\'",:g)"} # tested on the recently built rakudo 2015.12, so actually I _can_ do c++ without { } 11:42
zhmylove azawawi: great! What is $img.clone ? 11:46
azawawi zhmylove: get a new copy of the image matrix so you can manipulate it
zhmylove: e.g. you want to make it grayscale and then put them side by side 11:47
zhmylove azawawi: actually I can't get the difference with `my OpenCV::Mat $denoised_img = $img;`... Could you please describe or just give a link where I can read about it? 11:48
azawawi zhmylove: they point to the same C++ object reference. It is an OpenCV library thing. 11:50
let me simplify the example
azawawi zhmylove: docs.opencv.org/2.4.11/modules/core...=clone#Mat Mat::clone() const 11:51
zhmylove azawawi: got it! Thank you a lot! Actually `method clone() returns OpenCV::Mat ` helped too. 11:53
azawawi zhmylove: np, added an example of how to load in grayscale if you're interested 12:04
azawawi starts working on wrapping drawing api
lucs How do I do the equivalent of ｢$k ~~ %h｣ to check for a value instead of a key? 12:52
geekosaur match against %h.values? 13:16
lucs Fails, %h.values is a Seq, not sure how to handle that. 13:17
timotimo it'll have to be something like "defined %h.values.index($needle)" or so 13:35
alternatively, $needle ~~ any(%h.values)
kid51 At rakudo.org/how-to-get-rakudo/, I read: "NOTE: the Rakudo Star with 6.c Christmas Perl 6 release is not yet available. Check back soon or try one of the older versions."
When will "soon" happen?
lucs timotimo: Thanks. 13:36
kid51 I ask because I'm about to start a Perl6 study group in NYC and, for the people likely to participate in this group, I've been advised that Rakudo Star will be a better bet.
DrForr A study group? Nice.
brrt .seen konobi 13:37
yoleaux I saw konobi 4 Jan 2016 17:55Z in #perl6: <konobi> pmurias: just how things on the JS backend are going... has beena while since i had a chance to look
brrt darn, that's some time ago
timotimo kid51: sorry about that. FROGGS_ has been putting in time to bring the latest rakudo star release to fruition
brrt we could say we're a bit understaffed in the polishing-and-releasing department 13:39
timotimo mhm 13:40
lucasb sorry, I'm one week without reading the logs. just wanted to know if a release will happen today, or if you decided to delay it? 13:52
timotimo hm. looks like [Coke] signed up for this month's release 13:54
sena_kun Hi, #perl6. 13:58
timotimo hello sena_kun 13:59
sena_kun I was thinking this morning about bindings to C libs. In some way we just need to read bunch of signatures(and structures/variables) and convert it into perl6 expressions. Maybe it can be done my perl6 program? Something like c2hs library in haskell, which can generate bindings automatically. I searched a bit and found nothing like this. Is this a bad idea or just nobody has enough time to try to do it? 14:05
Skarsnik huggable not here :( 14:06
Did you have a look at github.com/Skarsnik/gptrixie ?
DrForr sena_kun: You need a full C parser and c preprocessor to do it that way.
Not saying it's a bad idea, I've thought about it myself. 14:07
timotimo DrForr: did you have a look at gptrixie? 14:08
DrForr Didn't know of its existence.
brrt senia_kun; if you have a compiler ready, you can do something like gcc -E to do macro evaluation
sena_kun DrForr, also, to just do bindings we don't need a full C parser. We need to parse only .h files, which is not so complex as .c, aren't we? 14:39
mspo hat if you used nm and objdump instead? 14:41
DrForr Try 'cpp -E /usr/include/stdio.h' and have a look at the output.
sena_kun DrForr, ah, sorry, you're right. 14:42
DrForr Not trying to dissuade, it's just quite a bit of work.
RabidGravy well for the most part, but there are libraries that are entirely composed of .h files (hey i2c_dev I'm looking at you) so it is entirely possibly that you need to parse C completely
Skarsnik's gptrixie makes a decent fist of making the binding code from a .h fie though 14:43
RabidGravy you have seen the github.com/Skarsnik/gptrixie ? 14:44
sena_kun RabidGravy, oh, I was sure it exist. Thank you. Skarsnik posted it already, but I missed the message. 14:45
RabidGravy I stuck it at the PortMidi.h and it made a reasonable go at it for a weird interface, would need some tweeking but a good start 14:47
Skarsnik Feel free to fill issue on what is wrong x) 14:59
sena_kun gist.github.com/Altai-man/ea83a55ffa51ec88089a - here is a gist, which can produce "Unhandled kind 1" error. If you use native perl6 Bool code will be correct. It will be good to replace this message with something more readable. 15:00
dalek c: f614b01 | (David Brunton)++ | doc/Language/operators.pod:
Cross meta operator

Some of the meta operator documentation is repeated in the infix operator sections. This makes the cross section match the other meta operators, though.		 15:09
c: ef9708a | sylvarant++ | doc/Language/operators.pod:
Merge pull request #356 from dbrunton/patch-3

Cross meta operator
Skarsnik sena_kun, well you are not supposed to use bool outside a native x) 15:14
but I think it's like an Int for perl 6 15:15
m: use NativeCall; say bool.WHAT; say bool.var.WHAT
camelia rakudo-moar 3259ba: OUTPUT«(bool)␤Method 'var' not found for invocant of class 'NativeCall::Types::bool'␤ in block <unit> at /tmp/oJ74kfCAEf line 1␤␤»
Skarsnik m: use NativeCall; say bool.WHAT; say bool.VAR.WHAT
camelia rakudo-moar 3259ba: OUTPUT«(bool)␤(bool)␤»
Skarsnik m: use NativeCall; say bool.repr
camelia rakudo-moar 3259ba: OUTPUT«Method 'repr' not found for invocant of class 'NativeCall::Types::bool'␤ in block <unit> at /tmp/adYqFESKhp line 1␤␤»
Skarsnik m: use NativeCall; say bool.REPR
camelia rakudo-moar 3259ba: OUTPUT«P6int␤»
sena_kun Skarsnik, I know, but 1)emacs perl6-mode didn't highlight Bool properly(fixed); 2)someone can just do a typo and we need pretty error messages anyway. If it's NativeCall "error", I can try to fix it and send a PR. But greping zavolaj with "unhandled" or "kind" gives me nothing. 15:19
Skarsnik zavolaj is "dead" 15:20
everything is in rakudo
masak is that clearly signaled in the zavolaj github repo?
Skarsnik and it's probably not possible to catch the missuse of those type :(
I know I should not have called it bool xD 15:21
but like c_bool or something
ely-se If I want to export a multi sub, do I have to add is export to each definition or only to the proto? 15:23
ely-se apparently only on the proto 15:24
RabidGravy I would expect that the proto would be anough
ely-se m: [+] (1,) 15:36
camelia ( no output )
ely-se m: say [+] ("1",)
camelia rakudo-moar 3259ba: OUTPUT«1␤»
ely-se m: ([+] ("1",)).perl.say 15:37
camelia rakudo-moar 3259ba: OUTPUT«1␤»
ely-se nice, it converts
b2gills m: say 0.6931471805599453 == log 2; say log 2; say log(2) == ~log 2 15:54
camelia rakudo-moar 3259ba: OUTPUT«True␤0.693147180559945␤False␤»
lucasb one digit of difference? 15:55
b2gills The same bug showed up in Perl 5: rt.perl.org/Public/Bug/Display.html?id=127182 15:56
masak interesting.
I mean, not surprises me with floating-point rounding errors, but I don't think I'd've been able to predict that one. 15:57
not much*
b2gills I was reading the P5P mailing list summary blogs.perl.org/users/sawyer_x/2016/...-10th.html 15:58
masak nice.
sawyer_x++ b2gills++
ely-se Is there a function to get the full path to a program given its name, looking up in the PATH environment variable?
lucasb ely-se: I'm not aware of any builtin that does this. so you'll probably have to traverse the dirs in PATH yourself 16:00
ely-se ok! 16:00
RabidGravy I think File::Which does that 16:01
b2gills It was already tested with Rakudo if you read the summary. (I don't know if a bug was reported yet) The fix doesn't seem to work though
m: printf "%.18g\n", log 2
camelia rakudo-moar 3259ba: OUTPUT«0.693147180559945␤»
ely-se brb
azawawi RabidGravy: im back :) 16:10
RabidGravy RARR!
Skarsnik azawawi, go finish pg array in dbiish support! 16:11
azawawi RabidGravy: tested the ethernet shield one one of the 2 Arduino unos i got yesterday :)
RabidGravy: the pins were twisted so i had to fix them lol
RabidGravy oopsie
Skarsnik I use my arduino to capture my snes controller input x) 16:13
azawawi RabidGravy: they didnt have the camera shield/addon though... i wanted to use it to transmit images via a cheap 433 MHz transmitter/receiver back to my PC for Serial to OpenCV/Perl6 program :)
azawawi RabidGravy: the local electronics store did not have also an RPi2 box too... 16:15
RabidGravy: my luck :(
RabidGravy :-\ you can get an ov7670 based camera thing from nearly anywhere for a few quid 16:16
azawawi anywhere but here... lol 16:17
b2gills m: say $_ = log(2).base(10,17); say $_ == log 2 16:19
camelia rakudo-moar 3259ba: OUTPUT«0.69314718055994531␤True␤»
azawawi github.com/azawawi/perl6-opencv/bl...o-world.p6 # Loading image, grayscale, denoise and writing image and moving image preview windows :) 16:25
ely-se Why does dir pass Strs to test and not IO::Paths? 16:33
Skarsnik I think it's because you can Str.IO 16:35
dalek p: 481d81a | (Pawel Murias)++ | src/vm/js/nqp-runtime/ (3 files):
[js] Use the same exception class for both types of exceptions thrown.
 16:41
p: e71a1bd | (Pawel Murias)++ | src/vm/js/nqp-runtime/ (2 files):
[js] nqp::{push,pop,unshift,shift,splice} throw apropriate exceptions on fixed arrays.
ely-se lucasb: I think this will do: gist.github.com/rightfold/0869cee15c3735b43417 16:51
lucasb ely-se: pretty cool :) just repeating that RabidGravy also mentioned File::Which. 16:54
ely-se oh hmm didn't notice that
gfldex m: class C {}; class W {}; sub outer(::T $c){ sub (T){ &?ROUTINE.signature.say } }; my $inner = outer(C); $inner(C.new); $inner(W.new); 17:12
camelia rakudo-moar 3259ba: OUTPUT«(T)␤Type check failed in binding <anon>; expected C but got W␤ in sub at /tmp/RkOzlgeovb line 1␤ in block <unit> at /tmp/RkOzlgeovb line 1␤␤»
gfldex the type check is correct but stringy on the type capture in the inner function does not resolve the type
gfldex m: class C {}; class W {}; sub outer(::T $c){ sub (T){ &?ROUTINE.signature.params[0].say } }; my $inner = outer(C); $inner(C.new); 17:14
camelia rakudo-moar 3259ba: OUTPUT«T $␤»
gfldex m: class C {}; class W {}; sub outer(::T $c){ sub (T){ dd &?ROUTINE.signature.params[0], T === C } }; my $inner = outer(C); $inner(C.new); 17:15
camelia rakudo-moar 3259ba: OUTPUT«T $␤Bool::True␤»
gfldex m: class C {}; class W {}; sub outer(::T $c){ sub (T){ dd &?ROUTINE.signature.params[0], T === C, T.new } }; my $inner = outer(C); $inner(C.new); 17:16
camelia rakudo-moar 3259ba: OUTPUT«T $␤Bool::True␤C.new␤»
gfldex grabs his rakudobug form
dalek p: d2f4787 | (Pawel Murias)++ | src/vm/js/nqp-runtime/ (2 files):
[js] Type checking on fixed arrays.
 17:18
cognominal perl6 verbose for one liner. Can I do better ? 17:21
$ a h
alias h='history'
$ a 6e
alias 6e='perl6 -MMONKEY-SEE-NO-EVAL -e'
$ h | 6e 'my $rx = EVAL "rx/@*ARGS[0]/"; $*IN.lines.grep($rx).tail(5).eager.join("\n").say' git
nine cognominal: just lines.grep is enough. The lines() sub uses $*IN by default 17:27
Also the .eager seems superfluous as the join needs to read all lines anyway 17:28
cognominal nope, lines uses $*IN if no arguments 17:32
thx anyway
I mean if @*ARGS == 0 17:34
gfldex do i have access to named arguments inside EXPORT ? 17:36
azawawi m: sub c( &a (Int) ) {}; sub b($) {}; c( &b ) 17:48
camelia rakudo-moar 3259ba: OUTPUT«Too few positionals passed; expected 1 argument but got 0 in sub-signature of parameter &a␤ in sub c at /tmp/1w2aE5v0OB line 1␤ in block <unit> at /tmp/1w2aE5v0OB line 1␤␤»
azawawi what's wrong with the above code? 17:49
timotimo i think it has to be &a:(Int) 17:50
otherwise it tries to unpack &a into one positional argument that has to be an Int
azawawi so what's the correct syntax for doing it? 17:52
im trying to use callbacks via nativecall for value sliders
gfldex m: sub c( &a:(Int) ) {}; sub b($) {}; c( &b )
camelia rakudo-moar 3259ba: OUTPUT«Constraint type check failed for parameter '&a'␤ in sub c at /tmp/RN1HIQHyKL line 1␤ in block <unit> at /tmp/RN1HIQHyKL line 1␤␤»
gfldex m: sub c( &a:(Int) ) {}; sub b(Int $) {}; c( &b ) 17:53
camelia ( no output )
azawawi gfldex: thanks 17:55
so basically the documentation here is wrong, right? 17:57
doc.perl6.org/language/nativecall#F..._arguments
cognominal If I don't want to go pure Perl6 I can do :
$ a t5
alias t5='tail -5'
$ h | 6ne '.say if /git/' | t5
with 17:58
$ a 6ne
alias 6ne='perl6 -MMONKEY-SEE-NO-EVAL -ne'
azawawi github.com/azawawi/perl6-opencv/bl...o-world.p6 # slider callbacks are now working. Enjoy :) 18:16
timotimo azawawi: i wrote the correct syntax right before i told you what's wrong :S 18:17
then i went AFK
azawawi timotimo: fixed :) 18:18
ely-se can a promise return a promise?
timotimo sure
ely-se nice 18:19
After the ECMAScript 6 Promise fiasco you can never be sure 18:22
azawawi pasteboard.co/QggG9WB.png # OpenCV fun with Perl 6 :)
timotimo i don't know anything about that, ely-se
ely-se timotimo: in ECMAScript 6, the then method automatically flattens. Thus you can't write any generic code that uses promises 18:23
basically then checks if the callback returns a Promise, and if it does, recurs
timotimo oh 18:24
RabidGravy m: (await Promise.in(1).then({ Promise.in(1))).then({say "baa"}); 18:30
camelia rakudo-moar 3259ba: OUTPUT«5===SORRY!5=== Error while compiling /tmp/5Z2kpkRBdF␤Missing block␤at /tmp/5Z2kpkRBdF:1␤------> 3await Promise.in(1).then({ Promise.in(1)7⏏5)).then({say "baa"});␤ expecting any of:␤ statement end␤ statement modifier…»
RabidGravy m: (await Promise.in(1).then({ Promise.in(1)})).then({say "baa"});
camelia ( no output )
pnu should panda read the extra_statefiles from _all_ Installable repos in the repo-chain? -> github.com/tadzik/panda/pull/289
RabidGravy m: await (await Promise.in(1).then({ Promise.in(1)})).then({say "baa"});
camelia rakudo-moar 3259ba: OUTPUT«baa␤»
MadcapJake Translated a presentation on first-class commands into Perl 6 yesterday: github.com/MadcapJake/Command6 18:47
grondilu ~/perl6/rakudo $ git pull 18:51
Already up-to-date.
^what's going on??
MadcapJake grondilu: git status; git branch 18:54
TimToady or just don't pull so often :) 18:55
grondilu :P
TimToady nobody's pushed to nom in the last 24 hours 18:57
must be some kind of a record :)
RabidGravy must be finished then 19:00
MadcapJake is there a module for parsing Perl 6 19:16
Skarsnik Perl 6 ?
MadcapJake lol
DrForr It's written in Perl 6 grammars. 19:17
MadcapJake So no module then? I need something that I could plug into a program.
DrForr Look at the Rakudo:: modules on modules.perl.org. 19:18
MadcapJake sweet
MadcapJake I'm gonna try and write a command line tool for generating autocompletes (also will have a linter but I think I'll write my own parser for that) 19:20
MadcapJake I realized that writing a linter directly in JavaScript (for my linter-perl6 package) was not very modular, so I'm gonna write it in perl6 and then create easily parsed output for various linters/editors 19:22
RabidGravy MadcapJake, the perl6 parser that it uses to parse the programs is available at runtime 19:23
this may however be overkill
MadcapJake well the problem is, that parser is setup to run the program. I think I'll need to create a custom parser for both features of this tool because I need to make sure it doesn't stop when there are errors, it just skips over whatever it doesn't understand 19:24
RabidGravy well the actions are setup to run the program, you can just supply your own actions 19:26
MadcapJake i could try that for sure but really i don't need to know everything. The autocomplete parser could just look for my/has/our/class/sub/[sub]method declarations and add the name to an array. 19:28
The linter parser I think I would actually have a pluggable grammar that you select which rules you want to check for in the META.info file. Something along the lines of eslint eslint.org/docs/user-guide/configuring.html but perhaps with some saner defaults
pmurias MadcapJake: why skip errors in a linter? 19:34
MadcapJake pmurias: what do you mean?
pmurias MadcapJake: "I'll need to create a custom parser..." 19:35
MadcapJake because `perl6 -c` stops at the first error, I want a linter to capture everything it can find in a file, not just the first error in the first stage that it hits.
gfldex can i overload EXPORTALL? 19:36
pmurias MadcapJake: keep in mind that it's literally imposible to parse Perl 6 with anything that's not Perl 6 19:37
MadcapJake pmurias: well I'm not trying to parse an executable program, I'm just looking for errors. So each rule will be setup to find a given kind of error. 19:38
but yes this program will be written in perl6 :)
pmurias MadcapJake: what sort of error do you want to find? 19:39
* errors
cognominal pmurias, MadcapJake is the great author of atom p6fe syntax hiliter and knows very well the limit of adhoc perl6 parsing
MadcapJake, what is needed is an alternative dumper for --target=parse 19:40
also the current dumper does not descend into EXPRs 19:41
MadcapJake that'd be one way to do it, but i still feel like you're limited by the fact that the parser is designed for creating an executable program. whereas an error parser could just have each rule be specifically designed for a given error.
cognominal apparently need a rewrite of the EXPR parser to do that
MadcapJake++ for P6FE 19:42
MadcapJake pmurias: Good question! Anything that could be discovered via parsing the text of a file.
MadcapJake cognominal: thanks! Glad you like it! 19:42
cognominal MadcapJake, also a good higlighter should be dynamic so that when cliking a operator it shows the relative precedences by higlighting the operator (say in backgournd red) and its operands (in background blue). I don't know any that does that. 19:45
MadcapJake cognominal: yeah that would be really awesome! Actually I think that could be done in atom. A plugin could get the text of the file, the cursor location, and then just create some rules for what the surrounding text is, then apply some CSS! 19:46
cognominal Really need a really iterate editing environment. The state of the art stopped at smalltalk which did not evolve :(
Anyway, in the meantime P6FE is good stopgap 19:47
jnthn P6FE++
pmurias for a lot of perlers the state of the art stopped at vim :/
MadcapJake I hope to add other cool packages to Atom, but I want to do the brunt of the computation in perl6 and then just have the atom packages be wrappers around shell commands 19:48
cognominal MadcapJake, when learning a new language, very often someone is stalled because he can't visually parse the code. Dynamic highliting would fix that.
P6FE++ indeed
cognominal I am happy to see that jhntn is moving to atom 19:49
jnthn Yeah, I need to get working (or patch) a multi-line editing module though... 19:50
jnthn The one I tried didn't seem to work, and I kinda can't live without it. 19:51
MadcapJake multi-line editing module?
MadcapJake I use ctrl-click all the time to create multiple cursors. it even works with ctrl-double-click 19:52
cognominal MadcapJake, is there any way in atom to compile nqp/rakudo and click on the error message to open the relevant file?. Especially difficult in nqp because the compilation units are concatenated files. 19:53
jnthn MadcapJake: ooh 19:54
ely-se > ‘rule’ causes whitespace in the regex to match literal whitespace in the matched text
Is this still true?
jnthn MadcapJake: So is there a keyboard shortcut for that?
ely-se I am so confused.
jnthn ely-se: Yes, inserts calls to <.ws> where there is whitespace between atoms 19:55
ely-se oh ok
does the default <.ws> match optional whitespace or required whitespace?
jnthn rule foo { bar baz } = token foo { bar <.ws> baz <.ws> } iirc
ely-se: required between alphanumeric, optional otherwise 19:56
ely-se ok, that explains
jnthn But you can override ws to do whatever you want
MadcapJake you could compile via atombuild.github.io - it supports GNU Make. I've never used it with make so I don't know if it matches error paths to files or not, but I know that atom-build *does* support that
cognominal: ^
jnthn MadcapJake: The one that's wired into my fingers is Alt + Shift + Up/Down arrow keys, which is what I'd really love to have work 19:57
MadcapJake jnthn: those should work in atom
jnthn: it works on my end 19:58
cognominal MadCapJake, multiple cursors are great in atom. I have yet to understand all the subtleties, though
Skarsnik multi cursors? x)
jnthn MadcapJake: Odd...it doesn't here...
MadcapJake: Version 1.2.3 fwiw
cognominal also, I use an emacs compatibility module so I can use my finger memory in atom. 19:59
thx MadcapJake, i will check it
MadcapJake jnthn: try `ctrl-shift-p` and type `add selection` and you should see two options for add above below, and there key shortcut next to it
jnthn MadcapJake: Aha, Ctrl + Alt + Up/Down 20:00
Thanks a lot! :)
MadcapJake jnthn: no prob! curious why your shortcut is different. 20:01
ely-se are there any tools for grammar debugging?
jnthn ely-se: Grammar::Tracer is one of the most useful...has a bug I should really patch, but tends to help quite a bit anyways
MadcapJake: Dunno. I'm on Windows...though Notepad++ has Alt + Shift + Up/Down... 20:02
MadcapJake jnthn: ahh, yeah i'm on linux, mine is alt-shift-up/down
jnthn guesses he can find where to tweak this... :)
ely-se jnthn: thanks 20:03
azawawi github.com/azawawi/perl6-opencv/bl...drawing.p6 # random rectangle fun with OpenCV/Perl6 :)
MadcapJake jnthn: i think i know why it's different for me: ctrl-alt-up/down is for linux desktop switching 20:05
cognominal jnthn++ for the not guts serie. jnthn got the guts to not gut.
jnthn MadcapJake: Ah, OK. :) 20:06
MadcapJake: Just found out how to make it what I want it to be in my custom keymap :)
azawawi pasteboard.co/Qn7mN5C.png # Camelia randomness :) 20:07
jnthn So, don't even have to re-muscle \o/
MadcapJake jnthn: nice! 20:08
ely-se jnthn: can you print the remaining text with the debugger? 20:16
jnthn ely-se: um...I can't remember...if you type help or so it'll tell you all the things it knows how to do. 20:21
ely-se printing text isn't in there, hence :p 20:22
ely-se something like {say $/.postmatch} in the regex maybe 20:24
ah, adding <ws> to the begin of TOP already helps 20:25
jnthn ely-se: Feel free to make a github issue requesting what you missed 20:26
jnthn I'd like to fix the LTM bug that folks sometimes hit in the next couple of weeks, so can probably put it what you're after also. :) 20:26
ely-se yay my parser finally works \o/ 20:28
jnthn \o/ 20:30
ely-se I forgot to implement some actions so it croaked. 20:31
azawawi Good night #perl6 20:47
rindolf azawawi: night. 20:50
autarch is there any way to make "CALLER::MY::<$foo> = ..." work if $foo hasn't been declared at all yet? 21:02
m: sub foo { CALLER::MY::<$foo> = 42 }; foo(); dd MY::' 21:03
camelia rakudo-moar 3259ba: OUTPUT«5===SORRY!5=== Error while compiling /tmp/uF1K5NyCrq␤Two terms in a row␤at /tmp/uF1K5NyCrq:1␤------> 3ALLER::MY::<$foo> = 42 }; foo(); dd MY::7⏏5'␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ s…»
autarch m: sub foo { CALLER::MY::<$foo> = 42 }; foo(); dd MY::
camelia rakudo-moar 3259ba: OUTPUT«Cannot modify an immutable Nil␤ in sub foo at /tmp/Z2gC93ZO2M line 1␤ in block <unit> at /tmp/Z2gC93ZO2M line 1␤␤»
autarch m: sub foo { CALLER::MY::<$foo> = 42 }; my $foo is dynamic; foo(); dd MY::
camelia rakudo-moar 3259ba: OUTPUT«PseudoStash.new(("!UNIT_MARKER" => !UNIT_MARKER,"\$!" => Nil,"\$/" => Nil,"\$=finish" => Mu,"\$=pod" => [],"\$?PACKAGE" => GLOBAL,"\$_" => Any,"\$foo" => 42,"\$¢" => Nil,"\&foo" => sub foo () { #`(Sub|54728688) ... },"::?PACKAGE" => GLOBAL,:EXPORT(EXPORT)…»
masak m: sub foo { CALLER::MY::<$foo> = 42 }; my $foo; foo(); say "alive" 21:05
camelia rakudo-moar 3259ba: OUTPUT«alive␤»
masak autarch: not sure what more beyond that you want.
autarch masak: I want to not have to declare $foo at all
jnthn Lexpads are immutable at runtime. 21:06
autarch basically, I want a way of putting some sort of marker in the caller's scope so I can later find it by looking at the packages in a trace
jnthn (A large amount of Perl 6's optimizability and error checking derives from that.)
So it'd need to be something that happens at compile time, like an import or use statement.
Ulti Skarsnik have you seen github.com/andydude/p6-c-parser 21:07
masak or a macro.
autarch jnthn: that makes sense
jnthn masak: I was going with things that work today 21:08
autarch: Do you need to make the thing statically or dynamically?
autarch jnthn: dynamically
autarch I'm trying to track the caller of a sub, but from elsewhere in the code 21:08
but I'm probably making this harder than it needs to be and I can just capture the caller immediately and save it somewhere 21:09
jnthn Wait, maybe that wasn't clear. Do you need to mark the block, or do you need to mark a particular invocation record?
callframe(1).code does role { method marker() { True } } # may work if you can cope with marking the block itself
autarch yeah, that wouldn't work, the markers needs to go away once the called sub exitrs 21:11
ely-se is it bad to override clone?
jnthn ely-se: No, it's fine
ely-se ok :)
jnthn autarch: Yeah, then we meant the same thing by dynamic... 21:12
autarch yep
Skarsnik Ulti, yes, but it's quite useless in its actual form x) And it's probably unreasonnable to use something else that a gcc/clang output x) 21:13
anyway, good night here 21:14
Ulti g'night 21:18
dalek p: 7c7ea34 | (Pawel Murias)++ | src/vm/js/nqp-runtime/reprs.js:
[js] Serialize MultiDimArray.
 21:22
ely-se can you assign a private property of another object of the same class?
pmurias autarch: what are you trying to accomplish? 21:25
autarch pmurias: I want to be able to find the caller of a particular sub (or subs) several layers further down the callstack
and I don't know what sub is being called in advance, so the subs themselves need a way to say "my caller is the one you're looking for"
pmurias why do you need to that? 21:26
cognominal jnthn, "lexpads are immutable at run time", assuming no EVAL in inner lexical scopes?
masak cognominal: that's not an exception to the rule.
ely-se hmm, it seems temp borks when used with a hash with non-string keys
masak cognominal: the variable declarations stay within the inner scopes. EVAL has its own compile time.
autarch pmurias: in order to provide useful error messages, basically
cognominal anyway it needs MONKEY-SEE-NO-EVAL 21:27
autarch in other words, if I call foo() which then calls 5 other things and one of those things generates an error, ideally that error says "called at line X" which is the line where foo() was called
ely-se m: my %h{Pair}; %h{a => 1} = 42; temp %h = %h.clone;
camelia rakudo-moar 3259ba: OUTPUT«Type check failed in binding key; expected Pair but got Str␤ in block <unit> at /tmp/7YVkdzDeCu line 1␤␤»
ely-se m: my %h{Pair}; %h{a => 1} = 42; temp %h{Pair} = %h.clone; 21:28
camelia ( no output )
ely-se interesting
cognominal ok, EVAL means a new lexical scope so it can't mess with outer scopes
jnthn ely-se: Feel free to file that one
ely-se oh right, that assigns the Pair property
jnthn: ok
jnthn cognominal: Correct. 21:29
ely-se where should I report it? Rakudo?
pmurias autarch: something like Carp?
cognominal so the REPL implies more nesting at each line?
timotimo yo
yeah, that's how it's implemented right now :(
autarch pmurias: yes, but to an arbitrary depth that isn't known in advance ;)
cognominal so the REPL implies one more lexical nesting at each newline line? # reformulated 21:30
timotimo cognominal: yes
jnthn ely-se: Mail [email@hidden.address]
ely-se jnthn: done 21:33
robfin how to call method with method-name contained in scaler string, like calling $obj.{$ChosenMethod}
jnthn $obj."$ChosenMethod"()
robfin perfect thanks! 21:34
masak robfin: note you need the parens at the end -- otherwise Perl 6 will assume you're confused and maybe trying to do Perl 5 style string concatenation 21:35
m: my $obj; my $ChosenMethod; $obj."$ChosenMethod"
camelia rakudo-moar 3259ba: OUTPUT«5===SORRY!5=== Error while compiling /tmp/eTPxYKs946␤Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.␤at /tmp/eTPxYKs946:1␤------> 3; my $ChosenMethod; $obj."$ChosenMethod"7⏏5<EOL>␤»
ely-se I'll use LEAVE for now then 21:36
robfin got it. thanks to you all 21:38
uruwi Hi 21:42
jnthn sleep & 21:43
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { }; class A { has $!a; has $!b is transient }; say trait_mod:<is>(A.^attributes[1], :transient) 21:44
camelia rakudo-moar 3259ba: OUTPUT«Nil␤»
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { True }; class A { has $!a; has $!b is transient }; say trait_mod:<is>(A.^attributes[1], :transient)
camelia rakudo-moar 3259ba: OUTPUT«True␤»
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { True }; class A { has $!a; has $!b is transient }; say trait_mod:<is>(A.^attributes[1], :transient); my $a = A.new(a => 1, b => 2) 21:45
camelia rakudo-moar 3259ba: OUTPUT«True␤»
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { True }; class A { has $!a; has $!b is transient }; say trait_mod:<is>(A.^attributes[1], :transient); my $a = A.new(a => 1, b => 2); say $a;
camelia rakudo-moar 3259ba: OUTPUT«True␤A.new␤»
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { True }; class A { has $.a; has $.b is transient }; say trait_mod:<is>(A.^attributes[1], :transient); my $a = A.new(a => 1, b => 2); say $a;
camelia rakudo-moar 3259ba: OUTPUT«True␤A.new(a => 1, b => 2)␤»
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { True }; class A { has $.a; has $.b is transient }; say trait_mod:<is>(A.^attributes[$_], :transient) for ^2; my $a = A.new(a => 1, b => 2); say $a; 21:46
camelia rakudo-moar 3259ba: OUTPUT«True␤True␤A.new(a => 1, b => 2)␤»
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { False }; class A { has $.a; has $.b is transient }; say trait_mod:<is>(A.^attributes[$_], :transient) for ^2; my $a = A.new(a => 1, b => 2); say $a;
camelia rakudo-moar 3259ba: OUTPUT«False␤False␤A.new(a => 1, b => 2)␤»
uruwi m: multi sub trait_mod:<is> (Attribute:D $attr, :$transient!) { True }; class A { has $.a; has $.b is transient }; say trait_mod:<is>(A.^attributes[$_], :transient) for ^2; my $a = A.new(a => 1, b => 2); say $a;
camelia rakudo-moar 3259ba: OUTPUT«True␤True␤A.new(a => 1, b => 2)␤»
masak uruwi: there's also /msg camelia ...
uruwi I was going to ask for help.
masak uruwi: if you feel like you're spamming #perl6 or something
uruwi: I'm not sure what you want to do above. mostly I'm just hypnotized by a dozen similar-looking one-liners 21:47
uruwi Trying to add an `is transient` trait so I can introspect objects for serialization. 21:48
masak why are you trying to add it both with an explicit `is transient` and then also with a run-time call?
why don't you actually do something useful in the `trait_mod:<is>` sub? 21:49
have you considered that `True` isn't enough? :)
uruwi I'm adding the `is transient` in the class definition, then trying to see which fields have it during runtime.
masak I think you're overestimating the magic that goes on with trait mods
you need to actually do something in the sub 21:50
typically, use `does` and mix in a role of some kind
which you can then introspect on the attribute using ~~
uruwi Ah, thanks! 21:51
ely-se I have a real hard time understanding the behaviour of clone 21:52
masak ely-se: .clone is similar to .new 21:53
ely-se: except it inherits attributes from the old object, which can then be overridden
ely-se: so `my $obj = C.new(:foo(1), :bar("mwhahaha")); my $obj2 = $obj.clone(:bar(2));` would create a C object with :foo(1) and :bar(2) 21:54
ely-se: but I bet you know this, and are confused by something else ;)
ely-se masak: gist.github.com/rightfold/bd6447b6f03f6b539a80 21:55
the last line I expected to print only "(a => 1) => 2" 21:56
but it seems that $c.clone.n affects %!h of the original C
geekosaur isn't clone shallow? 21:59
masak ely-se: I'm confused by line 19. which seems to override the %!h in the *original* object.
ely-se: recall that .clone is called on the original object, not on the clone
ely-se masak: it swaps them, effectively
masak ok, I suspected that's what you were doing 22:00
ely-se I'd use temp but temp is broken for hashes with non-string keys 22:00
masak maybe that doesn't work, for whatever reason?
ely-se: you can use .WHICH to see if you actually have two distinct hashes in the two objects
ely-se callsame uses the original %!h, then the original %!h is overwritten with a clone
masak: let's see
masak ely-se: it's a cute pattern, cloning the original object's hash. don't believe I've seen that before. ely-se++ 22:01
ely-se they are the same hash
ely-se WHICH returns the same value for both 22:01
Hash' clone method works though. It returns a new hash 22:03
m: my %h{Pair}; my %i{Pair} = %h.clone; %h.WHICH.say; %i.WHICH.say; 22:04
camelia rakudo-moar 3259ba: OUTPUT«Hash[Any,Pair]|78748640␤Hash[Any,Pair]|78748752␤»
masak m: class C { has %.h; method clone { C.new(:h( %!h.clone )) } }; my $c = C.new; say $c.h.WHICH; say $c.clone.h.WHICH 22:07
camelia rakudo-moar 3259ba: OUTPUT«Hash|66908304␤Hash|66908352␤»
masak that works too
ely-se I don't want it public though
masak I just wanted to be able to test it easily here
ely-se I have mutators. 22:08
maybe I'll just call new and then loop over the original hash and call the mutators
though I do want to know why this doesn't work
masak m: class C { has %.h; method clone { my $result = callsame(); %!h .= clone; $result } }; my $c = C.new; say $c.h.WHICH; say $c.clone.h.WHICH
camelia rakudo-moar 3259ba: OUTPUT«Hash|52028064␤Hash|52028064␤»
masak that... doesn't work.
ely-se very strange
it should shouldn't it 22:09
masak and we can see that the way it doesn't work is that it doesn't replace the %!h in the original object
m: class C { has %.h; method clone { my $result = callsame(); $result } }; my $c = C.new; say $c.h.WHICH; say $c.clone.h.WHICH 22:10
camelia rakudo-moar 3259ba: OUTPUT«Hash|48279952␤Hash|48279952␤»
masak i.e. the `%!h .= clone;` statement has no measurable effect
ely-se maybe clone can't mutate the original
masak well, it's not throwing any error 22:11
ely-se what about non-hash attributes?
masak and I don't see why it shouldn't be able to. it's just a normal method
ely-se m: class C { has $.x; method clone { my $result = callsame; $!x = 1 => 2; $result } }; my $c = C.new; say $c.x.WHICH; $c.x = 5 => 6; say $c.clone.x.WHICH 22:12
camelia rakudo-moar 3259ba: OUTPUT«Any␤Cannot modify an immutable Any␤ in block <unit> at /tmp/fe3D2wrhse line 1␤␤»
ely-se m: class C { has $.x is rw; method clone { my $result = callsame; $!x = 1 => 2; $result } }; my $c = C.new; $c.x = 5 => 6; say $c.x.WHICH; say $c.clone.x.WHICH 22:13
camelia rakudo-moar 3259ba: OUTPUT«Pair|Int|5|Int|6␤Pair|Int|5|Int|6␤»
ely-se oh lol
pairs are immutable
ely-se Ahh wait, don't hashes assign in list context? 22:13
it stays the original hash
it slurps instead of reassigning reference
masak ely-se: aaaaaah 22:15
ely-se that's nasty
masak ely-se: maybe binding would work, then
ely-se nope, lol
Type check failed in binding; expected Associative[Any,Pair] but got Hash[Any,Pair]
masak m: class C { has %.h; method clone { my $result = callsame(); %!h := %!h.clone; $result } }; my $c = C.new; say $c.h.WHICH; say $c.clone.h.WHICH
camelia rakudo-moar 3259ba: OUTPUT«Hash|73343744␤Hash|73343744␤»
masak um, uh?
ely-se which is stupid considering doc.perl6.org/images/type-graph-Hash.png
masak agreed 22:16
ely-se has no idea what is going on
masak ely-se: I think you have enough to submit a rakudobug, if you haven't already ;) 22:17
ely-se yummy, shrimps
m: my %h; %h := %h.clone; 22:18
Ulti ~/win 5
camelia ( no output )
ely-se m: class C { my %!h; method m { %!h := %!h.clone } }; C.new.m 22:19
camelia rakudo-moar 3259ba: OUTPUT«Use of uninitialized value of type Any in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in any throw at src/Perl6/World.nqp line 3740␤5===SORRY!5=== Error while compiling /tmp/15XJNu7GCs␤Cannot us…»
ely-se wtf
masak seems you got an error during the throwing of an error :/ 22:20
ely-se we can conclude that hashes are broken :P
time to submit my second bug today 22:21
masak aye
ely-se HUH 22:22
now it works
ah, it works only if the key type isn't set 22:23
ely-se hmm, it seems IO::Pipe.new isn't useful 22:47
oh, need to call IO::Pipe.pair
ely-se which is not implemented xD 22:48
ely-se can I create an IO::Handle from an Int file descriptor? 23:06
geekosaur @index fdToHandle
hm, did i break it? 23:07
hackage.haskell.org/package/unix/do...fdToHandle
also System.Posix.IO 23:08
ely-se geekosaur: Perl 6, not Haskell
geekosaur whoops, I am doing poorly today
xenu :D
geekosaur I onlu realized thechannel error when I tried to figure out the bot
ely-se :P
geekosaur I think... fd to handle has issues? because libuv wants to hide it 23:09
geekosaur got less sleep than intended last night and is still having coughing fits so naps are out :/
ely-se guess I'll open from /proc/seld/fd/* 23:14
if that works
geekosaur there's some internal thing iirc that can get the fd, for some things. (sockets were a problem iirc) 23:16
ely-se I don't want to get the fd. I have the fd and want to get an IO::Handle object
geekosaur yeh, but if you can't go one way easily you usually can't go theother way easily either 23:17
like, if you have an fd for a socket, you are already nonportable and libuv really wants to be portable. (windows sockets aren't fds) 23:18
ely-se eww gist.github.com/rightfold/eab266b8...m6-L12-L13 23:21
but at least it works :p
zostay m: for cross(0..2, 0..2) -> ($x, $y) { say ++$ } 23:35
camelia rakudo-moar 3259ba: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤»
yoleaux 7 Jan 2016 13:53Z <hoelzro> zostay: could you ping me when you have a chance? I have ideas about your abort() woes
zostay m: my $a = 0..2; my $b = 0..2; for cross($a, $b) -> ($x, $y) { say ++$ }
camelia rakudo-moar 3259ba: OUTPUT«1␤»
zostay why are those two lines different?
or better question: how do i make the second behave like the first? 23:36
abaugher I've used rakudobrew to build moar 2015.12 and panda. Now when I try to install modules with panda it says "Please remove leading 'v' from perl version in Text::CSV's meta info. Text::CSV requires Perl version 6.0.0. Cannot continue." 23:37
Is there a way I can get panda to ignore that, or otherwise make it work? --force and --notests don't get past it.
skids I had to remove 6.0.0 and just put 6 in all my modules. 23:39
zostay m: my $a = 0..2; my $b = 0..2; for cross($a.list, $b.list) -> ($x, $y) { say ++$ } # the best i can come up with 23:42
camelia rakudo-moar 3259ba: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤»
skids m: my @a = 0..2; my @b = 0..2; for cross(@a, @b) -> ($x, $y) { say ++$ }
camelia rakudo-moar 3259ba: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤»
AlexDaniel zostay: why do you expect $a to work like array? 23:43
m: my $a = 0..2; my $b = 0..2; for cross(@$a, @$b) -> ($x, $y) { say ++$ } 23:44
camelia rakudo-moar 3259ba: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤»
jvcakg Greetings 23:57
zostay AlexDaniel: because i've been messing with perl 6 for months and what i expect from things is still catching up from all the changes in the last few months 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!