»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.
zengargoyle right, but how does it know about things on CPAN? 00:00
Zoffix Same way METACpan knows of things on CPAN 00:02
zengargoyle just guesses you've got this sorted and i won't worry :)
Zoffix From index files
zengargoyle until i try. 00:03
:)
eater[sha2017] is there a way to open unix pipes?
or is it not there yet?
Zoffix which I guess aren't there (yet?). ugexe's ecogen just rsyncs stuff: github.com/ugexe/Perl6-App--ecogen...pm6#L6-L19 00:04
zengargoyle eater[sha2017]: how do you mean?
eater[sha2017] zengargoyle: create an unix named pipe, like `mknod(2)` does with S_FIFO or S_IFSOCK does 00:11
-does
zengargoyle i'd expect you can shell and pass a pipeline or run something and get stdin/stdout/stderr and you could open a named pipe, but don't know if there's a mknod.
yeah..... i'm not sure there's a builtin mknod. but expect that one created otherwise would work. 00:12
eater[sha2017] hmm
nadim I am surprised that 10e10 put in a num32 gives 99999997952!
zengargoyle p6 tries to avoid being *too* POSIXy
eater[sha2017] let's see how hard it is to add into Perl6
nadim jnthn: I am surprised that 10e10 put in a num32 gives 99999997952!
zengargoyle so you could do a NativeCall wrapper and make a mknod() if you desired.... 00:13
or run a `mknod` to do it.
eater[sha2017] to which lib tho, I'm pretty sure it's syscall
zengargoyle is it not in clib ? 00:14
eater[sha2017] hmm
let's see
Zoffix nadim: looks like a bug to me 00:15
timotimo nadim: 16777216 is the highest number that a 32bit num can exactly represent
Zoffix Is it?
huh
timotimo wikipedia claims that
let's see 00:16
Zoffix Ah, ok
timotimo 33554432 is the upper bound for represented to the next multiple of 2
zengargoyle my fuu is weak, but someing along: sub mknod(somethinghere) is native {*} .... calling clib things is pretty simple and the docs probably have at least a simple example of that.
timotimo you have to "is native(Str)" i think
nadim let me paste the code
eater[sha2017] zengargoyle: yeah 00:17
timotimo m: say 2 ** 25, 10 ** 10
camelia 3355443210000000000
timotimo m: say 2 ** 25; say 10 ** 10
camelia 33554432
10000000000
timotimo that's still only 8 decimal points
eater[sha2017] zengargoyle: already have one for reboot :') github.com/the-eater/shinit/blob/m...Kernel.pm6
timotimo if you're below that you'll get rounded to a multiple of 4
sorry, wrong 00:18
zengargoyle timotimo: yeah, i'd just have to actually look at docs and try to get it right.
timotimo twice that
nadim nopaste.linux-dev.org/?1160478
meh, I can write it here
zengargoyle eater[sha2017]: ah, you were the syscall question many moons ago?
nadim m: class Point is repr('CStruct') {has num32 $.x} my $p= Point.new: :x(10e10) ; dd $p 00:19
camelia 5===SORRY!5=== Error while compiling <tmp>
Strange text after block (missing semicolon or comma?)
at <tmp>:1
------> 3Point is repr('CStruct') {has num32 $.x}7⏏5 my $p= Point.new: :x(10e10) ; dd $p
expecting any of:
infix…
zengargoyle that question made me shutdown my desktop server because the example was for shutdown and not reboot! :P
nadim m: class Point is repr('CStruct') {has num32 $.x} ;my $p= Point.new: :x(10e10) ; dd $p
camelia Point $p = Point.new(x => 99999997952e0)
timotimo nadim: i just compiled short c program
99999997952.000000 00:20
that's the output i get for 'printf("%f", 10e10f);'
nadim interesting!
eater[sha2017] zengargoyle: don't think so, I wrote this 2 days ago
timotimo single precision floats have 8 bits exponent and 23 bits fraction 00:21
zengargoyle oh, nm, there was a syscall and reboot question a while back.... i proved it worked and made myself have to go and press a button to get my server back up. :) good times.
timotimo hah :D 00:22
eater[sha2017] zengargoyle: :')))
nadim timotimo: imgur.com/X6tcR04 some support for Native call in the dumper
timotimo: I don't know who is the doc master but in docs.perl6.org/language/faq#How_ca..._similar)? there is a reference to Data::Dumper 00:23
It's better to write nothing than that, it's useless and broken. 00:24
timotimo feel free to file an issue or pull request, i've gotta go sleep, my brain isn't working any more :)
nadim I work in a few hours, will do in the morning
nadim good night 00:24
eater[sha2017] o/ 00:29
Zoffix dam, no idea why live modules.perl6.org NEWDB build failed. On cron logs, I see it failed at running `perl bin/build-project-list.pl` (no idea why tho) but when I recreate that locally everything work 00:31
s
oohhh 00:32
zengargoyle cron has really different environment.... 00:33
Zoffix Damn, it's the whole "it's already running" thing biting me twice in a week ---
zengargoyle oh, nm.
[Coke] is there a standard in p5 for allowing multiple threads inside a single .t file? (TEST_JOBS lets you control # of simultaneous files, wondering if there's a related env var for concurrency)
eater[sha2017] zengargoyle: cannot locate symbol 'mknod' :( 00:34
Zoffix Basically, I need the same thing but for "new db" too: github.com/perl6/modules.perl6.org...c868f98316
timotimo [Coke]: ln the same file multiple times and decide based upon the script name which part to execute? :P 00:35
Geth doc: ebfdfebf63 | (Will "Coke" Coleda)++ | t/pod-valid.t
Add test to verify pod.

Closes #1398
doc: 1180e7223a | (Will "Coke" Coleda)++ | Makefile
Make 'make test' respect TEST_JOBS
doc: b2b56f61a7 | (Will "Coke" Coleda)++ | t/pod-valid.t
allow some concurrency
ugexe [Coke]: Test2 has things for that I believe 00:36
Zoffix (or just figure out a whole new system for marker files; considering it's spaghettified a bit ATM)
Prolly easiest is to have some sort of pid file for the current run and have bash script bail out if it's still running; so that it doesn't even pull anything new before a run finishes 00:37
Geth modules.perl6.org: 9b87209021 | (Zoffix Znet)++ | update-modules.perl6.org
Remove stray temp db file setting
 00:38
timotimo hm. if we installed github web hooks in a big chunk of the module repos we have ...
we'd have to engineer a bunch, but we could perhaps cut down on build time significantly
we're bound by the github api rate limit, aren't we?
[Coke] timotimo: see github.com/perl6/doc/commit/b2b56f61a7
Zoffix no hooks, please 00:39
timotimo sure
Zoffix ZofBot: it's not WEB SCALE!
geekosaur eater[sha2017], linux is "special" 00:40
Zoffix timotimo: what you mean by webhooks? Having modules site subscribe to 1000+ web hooks? 00:41
And not bother updating those repos via current system?
eater[sha2017] geekosaur: no, I am! 00:41
Zoffix See also github.com/perl6/modules.perl6.org/issues/79
Geth doc: bd51763948 | (Sam Morrison)++ | doc/Type/IO/Path.pod6
routine link creates hard links
 00:42
doc: 12f1cc0bee | (Will Coleda)++ (committed using GitHub Web editor) | doc/Type/IO/Path.pod6
Merge pull request #1434 from scmorrison/master

routine link creates hard links
geekosaur "mknod" is a preprocessor macro. the real call is apparently __xmknod ad the first parameter specifies the ABI version used by the other parameters 00:43
timotimo Zoffix: yeah, those that promise to keep us up to date with webhooks we wouldn't pull manually
except for full rebuilds, or twice a day, or whatever
Zoffix timotimo: looking at how much trouble we have keeping Geth subbed to half a dozen hooks, I'd imagine it'd be pretty tough to handle a thousand of them. 00:44
Wonder what the node.js folks are doing 00:45
timotimo mhm
maybe they have buy-in from github themselves 00:46
like, the permission to be pulling loads of times more than usually acceptable
Zoffix I guess by the time we reach GitHub's limits (and after we implement smarter rate limited) AND staying within the limits gives too-long builds, we'd be large enough to scrub up the $7 for premium GitHub with more API limits
zengargoyle C include grepping.... the `stat' family of functions and `mknod' are in fact inline
Zoffix (I'm just guessing at pricing there)
geekosaur zengargoyle, I went beyond grep
mknod is an inline wrapper for __xmknod 00:47
timotimo i didn't know we were not bound by request throttling
zengargoyle oh, me was busy looking and not reading. :)
eater[sha2017] geekosaur: pfff
thanks
geekosaur and you have to dig out _MKNOD_VER (which is _MKNOD_VER_LINUX and has *two* incompatible definitions in the same header; I didn't look to see which was which, but likely depends on which compatibility ABI you specify) 00:48
Zoffix It's currently mostly a guestimate "If we run continuous runs while sleeping for 1s between dists; then we won't hit the limit" kind of thing and smarter rate limiter would just figure out the needed delay between dists (or even if it's needed at all) on each run, based on how many requests we have left 00:49
zengargoyle geekosaur++ 00:50
at this point i'd just make a simple C wrapper and include it with the module and call that and let build figure out the details. 00:51
i.e. a libmymknod.so with a single function.
eater[sha2017] yeah 00:52
tbh
so much effort for a joke project
:'_
zengargoyle i'm sure there are better examples, but that's how Algorithm::Trie::libdatrie works... builds a .so from source that is inclueded in the $?RESOURCES and mostly magically works. 00:53
lookatme morningh 00:53
morning
eater[sha2017] zengargoyle: it's the best option tho 00:54
zengargoyle there's decent tooling for building a library and bundling it up with the module so it's not a dependency or something.
the .so is in the CUR with the rest of the code...
zengargoyle eater[sha2017]: and yeah, considering... i suspect you will *eventually* have some C stubs or something out of necessity. 00:57
given the problem domain.
eater[sha2017] haha 00:58
will be fun
never really written C
zengargoyle hey, don't we have Inline::C ??? 01:00
MasterDuke eater[sha2017]: heh. i've written more C in the past 15 months working on Rakudo/NQP/MoarVM than i have in the past 15 years on anything else
zengargoyle buggable: eco Inline::C 01:00
buggable zengargoyle, Nothing found
zengargoyle buggable: eco Inline
buggable zengargoyle, Inline 'use C-code from within Perl': github.com/FROGGS/p6-Inline-C 8 other matching results: modules.perl6.org/s/Inline
eater[sha2017] :'))
Zoffix bah "'message' => 'API rate limit exceeded for 209.*.*.*. (But here\'s the good news: Authenticated requests get a higher rate limit. Check out the documentation for more details.)'" 01:02
The README fetcher uses up a request too
lookatme Haven some one faced this problem: A Proc will be hang on except you call .exitcode on it
Zoffix lookatme: are you closing yer pipes? 01:03
lookatme yeah, I capture in out err 01:04
and close in
github.com/araraloren/perl6-app-sn...t.pm6#L117
The code
Geth modules.perl6.org: 83235aaad0 | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/p05DataFetcher.pm
Propagate API errors when fetching READMEs
 01:05
eater[sha2017] well
lookatme If I add `say $proc.exitcode;` after 120, it works fine
eater[sha2017] guess I'm tpying C tomorrow
ugexe you would still need to close :$out and :$err if you open them
eater[sha2017] thanks for the help zengargoyle++ o/
Zoffix lookatme: you're not closing your pipes: github.com/araraloren/perl6-app-sn...#L333-L334 01:06
lookatme I close it in fetchMessage 01:06
Zoffix lookatme: ah right 01:07
lookatme: ok, what do you mean by "Proc will be hang on"? 01:08
lookatme Zoffix, right, If there raised an exception what would happen?
Zoffix I don't understand the question. 01:09
Zoffix goes back to fixing modules.perl6.org before getting too tired to finish it today..
lookatme Do you have time try some code ?
:/
Geth doc: 5e47aad3e4 | (Will "Coke" Coleda)++ | 3 files
Provide :method to explicitly tag code as such

Allows us to have nicely formatted method signatures without having to make the examples detection smarter
Closes #1368		 01:11
lookatme gist.github.com/araraloren/1b857a1...26d9a129a8 01:14
jdv79 .tell Skarsnik I used your Gumbo module but now its failing like in the one bug against the dist 01:18
lookatme Somebody have interest on my question ? :)
AlexDaniel where's the bot again? 01:19
jdv79 odd
MasterDuke something happened with a bunch of the bots earlier 01:20
Geth modules.perl6.org: e771371a9c | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/PostProcessor/p05DataFetcher.pm
Fix failure to fetch README due to exhausted API limit

Pass authorization token so GitHub knows we're using an account with higher limit than the pub interface		 01:22
AlexDaniel I msg-ed dpk about yoleaux just in case 01:24
but somehow I feel like he knows already :)
lookatme :) Oh, have no idea about this problem 01:49
gist.github.com/araraloren/1b857a1...26d9a129a8
can someone help me ?
Zoffix lookatme: what problem are you trying to solve? 01:50
lookatme: .close returns the Proc which for procs without 0 exit code will throw when sunk
change to $ = $proc.in.close
lookatme Zoffix, Not working. The problem is the Proc hang on When I read **err** and **out** message in sub &fetchMessage 01:54
Zoffix What does "hang on" mean? 01:55
jdv79 did you try using Proc::Async?
just a random idea
lookatme "hang on" mean nothing happen and rakudo not exit 01:56
It should print error message and raise exception.
Zoffix lookatme: ok, first remove this .resume crap: github.com/araraloren/perl6-app-sn...t.pm6#L124
Most exceptions aren't resumable. 01:57
lookatme I already try it replace `...` with `.resume`, not working 01:58
Zoffix Alright. modules.perl6.org rebuilt and repulled and is working.... Nobody use `NEWDB` trigger :) 01:59
Zoffix &
lookatme And as I said before, if I add code `say $proc.exitcode` before &fetchMessage, it works fine
jdv79, I haven't try it 02:00
jdv79 idk. solved my isssues. 02:07
lookatme :/ 02:10
Geth doc: f7a524d6c2 | (Will "Coke" Coleda)++ | xt/space-after-comma.t
also test .md files
 02:11
doc: bd0790571f | (Will "Coke" Coleda)++ | t/pod-valid.t
Add TEST_THREADS env var support

Default to 2 threads.
Zoffix Well, you have a `loop` in your code. Start sticking print statements in places.
doc: 69a1ff0a87 | (Will "Coke" Coleda)++ | xt/space-after-comma.t
Add TEST_THREADS env var support
Zoffix Then start deleting stuff until the problem disappears. 02:12
Rinse and repeat until you golf to the part that fails and understand why :)
Zoffix and get rid of all the CATCHES 02:14
You're deliberatelly silencing errors and errors usually tell you why your stuff hangs 02:15
Zoffix and all the try's 02:19
lookatme Zoffix, I know where the problem is . 02:21
I remove all slurp-rest in fetchMessage
it works fine 02:22
Geth doc: b9a60c9280 | (Will "Coke" Coleda)++ | doc/Language/about.pod6
change from skip to show reason to skip
 02:26
Zoffix did you remove all the try's? .slurp-rest sounds like it might be dying due to encoding errors 02:31
ijneb Can I do a multiple gather thingy 02:32
like this:
my ($a, $b) = gather { take 1; take 2; } 02:33
so then $a would have 1, and $b would have 2
lookatme yeah, I remove try around that code
Zoffix, encoding error ?
Zoffix m: buf8.new(200).decode 02:34
camelia Malformed termination of UTF-8 string
in block <unit> at <tmp> line 1
lookatme Zoffix, But there no encoding error. It is still not working even I get one line from the err or out IO::Pipe of Proc 02:37
Zoffix no idea
lookatme What the app doing is use clang compiler compile that c++ code. In same condition, I change compiler to gcc, it working again. :/ 02:39
Zoffix ijneb: yeah you can do it. 02:40
lookatme Thought I have idea get rid of this trouble. But I don't want let this problem be there. :/ 02:41
Geth doc: e978918428 | (Will "Coke" Coleda)++ | 4 files
Don't skip tests; setup correct preambles/tags instead
 02:45
modules.perl6.org: aad538dbc6 | (Zoffix Znet)++ | 2 files
Improve home search form display on mobile/tablet media

  - Move logo to top
  - Fix up margins when search/lucky buttons are atop each other		 02:51
gilbert according to docs.perl6.org/language/traps#The_^_twigil 03:01
for 1..4 { say "$^one $^two $^three $^four"; } # produces OUTPUT: << 2 4 3 1 >>
I don't get it !
raschipi The parameters are in alphabetical order. 03:05
m: for 1..4 { say "$^one $^two $^three $^four"; } 03:06
camelia 2 4 3 1
geekosaur " but assigns values based on the variable's Unicode ordering."
raschipi m: <one two three four>.sort 03:07
camelia ( no output )
raschipi m: <one two three four>.sort.say
camelia (four one three two)
raschipi m: (%(<one two three four>.sort Z=> (1..4))){'one', 'two', 'three', 'four'}.say #gilbert 03:11
camelia (2 4 3 1)
Geth doc: 5f5a4a67b7 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/traps.pod6
Improve $^foo var naming example
 03:18
gilbert but perl6 -e 'no strict; @a = <one two three four>; say @a.sort'
raschipi m: my @a = <one two three four>; say @a.sort 03:19
camelia (four one three two)
raschipi m: my @a = <one two three four>.sort Z=> 1..4; say @a 03:20
camelia [four => 1 one => 2 three => 3 two => 4]
geekosaur yes, so perl6 took you to mean for 1...4 -> $^four $^one $^three $^two { ... }
gilbert yields 'four one three two' not "two four three one" !! 03:21
Zoffix gilbert: alphabetically `four` comes first and that variable gets value 1; `one` is second, alphabetically, so it gets value 2; `three` is third, so it gets 3; and `two` is fourth, so it gets value 4. 03:22
Zoffix gilbert: if sorted came out as "two four three one", you'd have to have them in the source as "$^two $^four $^three $^one" to get 1 2 3 4 03:23
gilbert thanks - back in 30 minutes 03:24
Zoffix gilbert: you're confusing meaning of the text encoded in variable name with the value that variable contains. Exactly what the trap is telling you you should avoid
gilbert That's why come hereto ask questions ;\) 03:27
Zoffix :) 03:29
zengargoyle feels sorry for $^x, $^y, $^z, $^w users. :) 03:40
geekosaur "if you care about the names, use actual names not placeholders" 03:41
zengargoyle m: say <x y z W>.sort
camelia (W x y z)
geekosaur m: say <X Y Z w>.sort 03:42
camelia (X Y Z w)
zengargoyle my head maths and ijk and xyzw are sorta where not naming would sorta jus fit the common things. 03:43
lookatme m: for 1..4 { say *, *, *, * }; 03:44
camelia ****
****
****
****
zengargoyle but yeah, we have unicode so ....
could put hats on them or such. 03:45
geekosaur someone's probably going to write a slang which puts them in order used, and then discover even more lovely edge cases...
zengargoyle lol 03:46
zengargoyle just found out that ibus compose doesn't do jhat or khat but has î 03:47
geekosaur ĵ 03:49
ibus but with the magic to have it fall back to X11 compose. <compose> ^ j 03:50
zengargoyle oh, *please* tell me how.
i've tried to do that for ages and failed.
geekosaur uh, I got it from the p6 docs page? 03:51
zengargoyle hrm, maybe it's new info.... 03:52
geekosaur but you have to export some stuff into the environment before ibus starts, debian will read your dotfiles before starting stuff but other distros/OSes will need special configuration
debian/ubuntu/mint
zengargoyle goes woot and is off to look. 03:53
:)
geekosaur the real question might be whether you're getting ibus or gnome's pathetic simulation thereof
zengargoyle i may have some gnome stuff floating about, but mostly use i3 03:54
geekosaur I probably should have said gtk
zengargoyle i think my only gnome is gdm login bit.
geekosaur it's not the window manager, it's the toolkit used by the terminal or irc client
(in my case, hexchat uses gtk but allows overriding the XIM so I set it to use ibus directly) 03:55
zengargoyle ah... :< i use gnome-termnal + irssi for better unicode font things.
so i'm like halfway between using xterm for stuff where i don't care as much and it's good enough, and gnome-terminal because it does the whole totally weird unicode font fallback thing better. 03:56
zengargoyle wants a terminal that does full font magic, uses XCompose, supports ibus for other input methods. i have not succeded as of yet. 04:00
Geth modules.perl6.org: a5714b21d9 | (Zoffix Znet)++ | templates/layouts/default.html.ep
Fix rendering of "home" nav on mobile
 04:01
zengargoyle maybe try going to xim or scim (if it still works at all) 04:02
zengargoyle ah, yep, new info in docs and yep xim... will give it a shot. 04:08
Geth modules.perl6.org: 9ce47a932c | (Zoffix Znet)++ | assets/sass/main.scss
Add a bit of shadows
 04:12
Geth modules.perl6.org: 165d9ef1f4 | (Zoffix Znet)++ | lib/ModulesPerl6/Controller/Root.pm
Add `from:` search feature

  - Filter by dist source:
   from:github or from:gitlab or (future) from:cpan		 04:33
jdv79 .tell Skarsnik I used your Gumbo module but now its failing like in the one bug against the dist 04:46
yoleaux jdv79: I'll pass your message to Skarsnik.
ryu0 m: say "I like to eat 3.14 PIs a day." 04:47
camelia I like to eat 3.14 PIs a day.
Zoffix m: say "I like to eat τ PIs a day" 04:49
camelia I like to eat τ PIs a day
Zoffix m: say "I like to eat {τ} PIs a day"
camelia I like to eat 6.28318530717959 PIs a day
BenGoldberg Wait, I thought perl was going to converge on 2pi, not in this tau thing ;) 04:51
raschipi BenGoldberg: Everyone measures circles by their radius, stop being weird. 04:56
Geth modules.perl6.org: 4266508225 | (Zoffix Znet)++ | 3 files
Add help page

To document new search features, the TODO issues feature, as well as how users can include Travis/AppVeyor stuff to their modules.		 05:03
Geth modules.perl6.org: 152e838ad7 | (Zoffix Znet)++ | templates/todo/index.html.ep
Improve Module TODO page header
 05:06
Zoffix \o/ yey help page modules.perl6.org/help 05:07
Geth modules.perl6.org: ca215799ad | (Zoffix Znet)++ | site-tips.txt
Add a couple more site tips
 05:17
Geth doc: nkh++ created pull request #1435:
Changed link in the data dump section of the FAQ		 08:14
jnthn nadim: Welcome to floating point. ;) 08:57
Skarsnik Hello 08:57
yoleaux 04:46Z <jdv79> Skarsnik: I used your Gumbo module but now its failing like in the one bug against the dist
nadim jnthn: I had my problems with them, on multiple architectures, 25 years ago. Silly me, I thought we were doing floating points internally. ala Big::Int. 09:04
jnthn num32 is explicitly asking for floating point 09:05
As is a literal wiht an e in it
zengargoyle jnthn: is there any magic anywhere in the floating point Num chaing of things. 09:06
or is using e guaranteed to actuallly get you an IEEE floating point thing like you would expect in other languages. 09:07
nadim jnthn: I added "some" Native support to DDT, but I'd like to do a bit more. timotimo showed me some of the ropes yesterday with explanations, Skarsnik also helped a bit. Now I am wondering if there is a way, via nqp if necessary, to look at an Int and say you are an int32. let me post a link to an image. 09:09
jnthn zengargoyle: We use the presence of the e in there to know to treat it as floating point and produce a Num literal 09:10
And Num is just a boxed num
Which you can assume to be the same as num64 in practice
nadim jnthn: imgur.com/X6tcR04 here list[4], which displays as '7', is received as an Int, but I know it was declared as an int32, I'd like to dig that out. timotimo exaplined that it was a property of the variable but by the time I display it, the original variable is gone. 09:11
jnthn Yup
You'd have to track it from the container
Yourself
zengargoyle hrm, i thought i noticed somewhere when a trivial case (like .1e0 or such) didn't behave like i would expect from floaing point. 09:12
nadim jnthn: then it will stay like that, it is impossible to track it from the container as the user, not DDT, may have put it in another container before the dump 09:13
zengargoyle like 1e0 is just 1 .... but it was a while back and i forget the details. i'll assume i was wrong in this case if 'e' makes floating....
jnthn nadim: If they do that, it will be an Int by then :)
Unless the container is typed
nadim ok
jnthn Natives aren't objects. They don't know what they are.
That's the point :)
m: say 1e0.WHAT 09:14
camelia (Num)
nadim I had to be imaginative to show the type of variable in CStructs already, I may be able to do that for typed containers, but it may be ugly
lookatme jnthn, do you have time have a look at my module problem, it's about capture stdout or stderr of Proc 09:16
gist.github.com/araraloren/1b857a1...26d9a129a8
jnthn lookatme: And you get a hang on the line you linked? 09:18
lookatme No, it's in sub fetchMessage
jnthn But general advice: if you're *ever* dealing with more than one handle when spawning a process, use Proc::Async, not Proc.
lookatme line 333 09:19
jnthn, yeah I replace with Proc::Async, it was working again
but I want know if there are problem with my code
jnthn Yes
You can't just read from $out and then $err
Because there's buffers involved 09:20
And if the program you're running fills its error buffer then it will block
zengargoyle yeah, i tried that myself.... there was some discussion a long while ago where some bit didn't make sense...
zengargoyle maybe Z fixed it.... it sorta seemed like e0 or e-1 things were being turned into Rat. 09:21
like DWIM overrided the 'e' in 'i want a float' 09:22
jnthn This is kinda why I wanted to deprecate Proc. People say it's simpler than Proc::Async. But for use cases like this, it only *looks* simpler, but hides complexity that you have to deal with anyway. :S 09:23
zengargoyle or optimized into 3e0 is really a 3.
lookatme jnthn, sorry, I can't understand full of you. Do you mean I can not capture both out and err in same time ? 09:24
jnthn, I think Proc::Async is better
zengargoyle and the expected floating point error (that p6 goes to great lengths to avoid) wasn't there.
jnthn lookatme: You'd have to start two threads to do it, and then do blocking reads on them 09:25
nadim jnthn: I have had error message like "cannot unbox to a native integer: P6opaque, Num" ie: when assigning a num with 10 or an int32 with a Num. That all makes sense but couldn't P6opaque carry some information, it still be opaque in its working. Obviously the compiler knows it's wrong so it could say why it is wrong, no? 09:26
jnthn lookatme: Yes, Proc::Async is always better if you're trying to deal with stdout *and* stderr of a process.
nadim jnthn: something like "cannot unbox to a native integer: P6opaque(int32), Num" 09:27
jnthn nadim: What wsa the exact code that gave the error? 09:28
*was
lookatme jnthn, Hmm, thanks, I understand a little.
nadim jnthn: my $point = Point.new: :x(2e56), :y(10) ; where $.y is a num32 09:29
m: class Point is repr('CStruct') { has num32 $.y } ; my $point = Point.new: y(10) ; 09:31
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
y used at line 1 (in Perl 6 please use tr instead)
andreoss could parametric types be possible with rakudo? what should i use instead? 09:32
nadim m: class Point is repr('CStruct') { has num32 $.y } ; my $point = Point.new: :y(10) ;
camelia This type cannot unbox to a native number: P6opaque, Int
in block <unit> at <tmp> line 1
moritz andreoss: you can parameterize roles
nadim jnthn: ^^
andreoss cool 09:34
but why not classes?
zengargoyle marvels at the morning pounce.
jnthn nadim: It'd probably be less confusing it if didn't mention P6opaque there 09:35
nadim yes
jnthn Should just say Int 09:35
nadim and what native number type if possible 09:36
jnthn I don't think we have that to hand 09:36
We'll unbox it full width and then coerce it to 32 bit
nadim andreoss: www.jnthn.net/papers/2009-yapc-eu-r...slides.pdf there is a section about parametrized roles in it, it's an old presentation, things may have changed
jnthn And it's the first step that fails, and it doesn't (and can't easily, I don't think) know about the second 09:37
Righty, I'm going to and hide away and try to concentrate and getting control flow graphs more precise around exception handlers so dominance calculations come out differrently and PHI nodes get placed better and maybe this arcane stuff will make something faster... :) 09:39
*concentrate on
nadim andreoss: docs.perl6.org/language/typesystem...s_as_Types
zengargoyle chuckles 09:49
nebuchadnezzar hello 09:51
zengargoyle hello back.
Geth ecosystem: bfa016b75f | (Ramiro Encinas)++ (committed using GitHub Web editor) | META.list
Add System::DiskAndUpdatesAlerts
 10:10
nadim Skarsnik: since it was you idea to ask about NativeCall support, maybe you can give me some feedback on the result so far? ;) 10:11
Skarsnik where is your module? x) 10:12
nadim hehe, you get a screenshot to start with, and if you hate it I fix it, otherwise I think I can stamp that as experimental and throw it in in the next release, I'd like to add more test case (your help appreciated with that). let me give you the code that generates the screenshot and the screenshot. 10:14
nadim Skarsnik: nopaste.linux-dev.org/ imgur.com/X6tcR04 10:16
meh! nopaste.linux-dev.org/?1160483 10:17
Skarsnik: I think the <VMArray> is overkill since the type already contains [int64], I think removing it or replacing it with something like < 10:20
<native>
Skarsnik I will look after lunch ^^ 10:24
domm is there something like CPAN::Mini for Perl6? I'm spending the whole tomorrow on a train / offline but might still want to hack some Perl6? 10:58
Skarsnik *Star? 10:59
nadim domm: Hi, not that I know of but it feels like downloading the list of modules (ecosystem) and using git to clone localy, zef may have the option to point somewhere else. if you run zef without any option it list a config file at the bottom 11:02
Skarsnik *need to buy more ram* 11:03
moritz you might want to clone github.com/moritz/perl6-all-modules to have all the modules in one place
Skarsnik nadim, you can still commit stuff on your repo and add the module later in the ecosystem :) 11:06
nadim I can even make you a branch to try if you want to 11:07
the module is in the ecosystem already 11:08
Skarsnik Hoo
nadim takes 5 minuted if you want an unstable branch 11:08
Skarsnik make me wonder if we have something to pin a version for the ecosystem
nadim that, IMO would make little sense 11:09
Skarsnik dunno, you could pin a release to have a stable version and still work on the master branch 11:10
nadim I have a master and a release branch, the release branch is in the ecosystem
but even my development branch is not enough, sometimes multiple development branches makes sense 11:11
but the branches for, IE, NAtive support, are not on github. they can be added when necessary
Skarsnik: github.com/nkh/P6-Data-Dump-Tree/tree/native 11:19
Skarsnik: test you can play with would be in example/int32.pl 11:20
ryu0 how interesting. for my PPA, the precomp files are only good for one architecture's rakudo. 11:23
Skarsnik *rebuild rakudo* 11:26
ryu0 haha. yea.
i'm figuring out how to make rakudo play nice with distro packaging.
made a lot of progress.
Skarsnik hm, I though the work on that was done? 11:27
ryu0 it has some sharp edges still.
but it's doing fairly well.
ryu0 just noticed i need to make the perl modules architecture specific to work around this issue. 11:27
otherwise the perl6 recompiles them in the user's .perl6 directory. 11:28
Skarsnik did you ask stuff in #perl6-toolchain ?
ryu0 No? I was digging around in how ARCH's AUR handles it.
ilmari ISTR a thread about this on debian-devel recently
ryu0 anyway i seem to have solved my problem. 11:29
ryu0 and there we go. that should solve the issue. 11:31
ryu0 i noticed that i couldn't take the precomp files from amd64 and get them to work easily with i386. X) 11:31
Skarsnik well, yes lol 11:33
ryu0 hence why i'll just build them separately.
hoping my ppa will make it easier to get rakudo running on ubuntu at least. 11:34
ryu0 i've been using zef as a test package until i figure out how i'm going to handle the rest of the modules. 11:35
Skarsnik hm, debian package does not work already? 11:36
Skarsnik rakudo dependacy are rather low 11:36
domm moritz: thanks 11:48
ryu0 Skarsnik: yes, but they're very old. 11:49
Geth doc/molecules-patch-1: ae8d174803 | (Christopher Bottoms)++ (committed using GitHub Web editor) | doc/Language/syntax.pod6
Added mention of if/elsif/else special case

Does this look okay? Inspired by brian d foy's question (stackoverflow.com/questions/455204...paration).		 11:57
Geth doc/molecules-patch-1-1: 1608091e88 | (Christopher Bottoms)++ (committed using GitHub Web editor) | doc/Language/syntax.pod6
Added mention of if/elsif/else special case

Does this look okay? Inspired by brian d foy's question (stackoverflow.com/questions/455204...paration).		 11:59
doc: molecules++ created pull request #1437:
if/elsif/else series of blocks as a special case for implied separator		 12:00
_ramix_ Hi guys. I just posted a new module in the ecosystem a couple of hours ago but it is still not available. The new module (System::DiskAndUpdatesAlerts) appears at ecosystem-api.p6c.org/projects.json but no error appears at ecosystem-api.p6c.org/errors.json. Any ideas? 12:08
Skarsnik hm weord 12:11
araraloren _ramix_, try to do test with Test::META github.com/jonathanstowe/Test-META/ 12:17
travis-ci Doc build passed. Christopher Bottoms 'Added mention of if/elsif/else special case 12:20
travis-ci.org/perl6/doc/builds/261797654 github.com/perl6/doc/commit/ae8d17480354
travis-ci Doc build passed. Christopher Bottoms 'Added mention of if/elsif/else special case 12:22
travis-ci.org/perl6/doc/builds/261798574 github.com/perl6/doc/commit/1608091e88a0
tbrowder hi #perl6 12:25
tbrowder is there any way to use NativeCall on binary objects that are not library functions such as "app" created from one or more libraries? 12:27
_ramix_ araraloren, According with Test::Meta... all ok and Project META file is good 12:28
Skarsnik Nooo my vm is dead 12:34
tbrowder, just call your binary with shell or run
araraloren _ramix_, maybe it not update 12:35
see the last update log
modules.perl6.org/update.log
El_Che Skarsnik: restore a snapshot? 12:35
_ramix_ araraloren, thanks. What is the frequency of the update? 12:37
Skarsnik I did not have a snapshot ^^
tbrowder yes i know that works, but everyone always warns about security issues using them
Skarsnik the disk file is dead I think
araraloren _ramix_, According here mentioned: docs.perl6.org/language/modules#Di...g_Modules, it is one hour 12:40
_ramix_ araraloren, Then I suspect the update is not working well. The module insertion in META.list was 3 hours ago and the last activity of update.log was Sun Aug 6 23:04:07 2017 12:44
araraloren _ramix_, maybe :) 12:46
samcv . 12:48
_ramix_ araraloren, I'm going to open a issue at github.com/perl6/ecosystem/issues with this situation. Thanks! 12:53
araraloren :) _ramix_ you welcome 12:54
Skarsnik tbrowder, I dunno what security issue, it's not more secure to call an external lib or a binary. you can already avoid lot of issue using run and not shell
[Coke] win 16 12:56
buggable [Coke], Thank you for entering Accidental /win Lottery! The next draw will happen in 3 weeks, 3 days, 11 hours, 3 minutes, and 54 seconds
ryu0 How fiendishly cool. 12:57
[<] 1, 2, 3
m: [<] 1, 2, 3
camelia Potential difficulties:
Useless use of [<] in sink context
at <tmp>:1
------> 3<BOL>7⏏5[<] 1, 2, 3
ryu0 m: say [<] 1, 2, 3
camelia True
ryu0 test if a list is sorted :)
m: say [<] "a", "b", "c" 12:58
camelia Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5a' (indicated by ⏏)
in block <unit> at <tmp> line 1
ryu0 Huh.
m: say ("a" < "b")
camelia Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5a' (indicated by ⏏)
in block <unit> at <tmp> line 1
ryu0 O_o
Skarsnik this error is not that good
araraloren m: say 1 < 2 < 3; 12:59
camelia True
ryu0 I was expected it to perform a lexicographic comparison.
Skarsnik probably another operator x) 13:00
ryu0 m: say ("a" lt "b") 13:01
camelia True
ryu0 m: say [lt] "a", "b", "c"
camelia True
ryu0 Huh. somewhat like using shell test utility.
i guess it works. I was just expecting it to be overloaded for all comparable types.
or rather, ones that can have an ordering. 13:02
Skarsnik m: say (9..2).sort;
camelia ()
ryu0 O_o
jnthn ryu0: < consistently means numeric comparison, lt means string comparison. It's so even if you don't know the exact types you have, you'll get consistent semantics. 13:03
ryu0 m: say (2..9)
camelia 2..9
ryu0 jnthn: i see.
jnthn Same reason ~ is used for concat, rather than overloead +
Skarsnik m: say 9..2.sort; 13:03
camelia Seq objects are not valid endpoints for Ranges
in block <unit> at <tmp> line 1
ryu0 jnthn: so i have to choose operators based on types
ok. 13:04
jnthn There's also before which *is* overloaded by type
When you know that's what you want
ryu0 overloading has its own perils i know. 13:04
jnthn This saves the problems in many dynamic languages where you can't look at a piece of code and know whether < will do numeric comparison or not :)
ryu0 it can create a lot of "humorous" problems in C++. :/ 13:04
jnthn In Perl it always will.
ryu0 lol 13:05
jnthn And if one side (or either side) are strings, it'll coerce them to numbers first
ryu0 And if that fails?
jnthn m: say "a" < "b" 13:05
camelia Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5a' (indicated by ⏏)
in block <unit> at <tmp> line 1
ryu0 m: say "a" + "1" 13:06
camelia Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5a' (indicated by ⏏)
in block <unit> at <tmp> line 1
ryu0 Oh.
jnthn It complains :)
ryu0 so it keeps the JS like behavior but actually throws an error instead of silently giving garbage. 13:06
jnthn Well, except in JS, + might do a concat. 13:07
Depending on what it gets.
ryu0 yea...
i'm aware.
i meant in general
not to mention it has no integral types, only floating point.
araraloren jnthn, :) A quick question. Why we should use CArray[uint8] not Str if C function requires the lifetime of a string to exceed the function call ? 13:08
jnthn araraloren: Because with Str a buffer with the enoded data is made for you, but you never get access to it, so you can't store it somewhere to make sure it won't be deallocated 13:09
fiwiw, you can pass a Blob or Buf instead of CArray[uint8]
Which may be more convenient
araraloren Blob or Buf suit for NativeCall ? 13:10
jnthn Yeah
jnthn Just make sure that you keep a reference to it in your program so long as the C code also will 13:10
araraloren Thanks, I will try it.
ryu0 launchpad.net/~ryu0/+archive/ubuntu/perl6 <-- my work so far.
jnthn And don't mutate it :) 13:11
araraloren jnthn, I will
ryu0 i'll probably do some more later.
araraloren ok
ryu0 i plan to try to package everything that is shipped in rakudo star. 13:12
but for now i need to go.
araraloren jnthn, thanks. Cool!!! working perfectly! 13:13
jnthn ryu0++ 13:15
nadim araraloren: Test::META itself does not work for me because META.version xxxx is not found but Zef says it is up to date. May be a problem related to the one ramix has. let me dig out a ticket 13:19
araraloren: github.com/jonathanstowe/Test-META...-319090770 13:20
araraloren oh 13:21
:)
samcv grant status update 3 is up cry.nu/perl6/grant-status-update-3/ 13:26
[Coke] samcv++ # also reminding me of many small grant things i need to do 13:37
pmurias ryu0: small floating point numbers sort of act like integers in js 13:44
ryu0: rakudo.js/nqp.js sort of gets them to work that way with adding |0 everywhere 13:45
nadim samcv++
pmurias jnthn: will you be at TPC::EU? 13:48
jnthn pmurias: No, 'fraid not 14:01
pmurias jnthn: :( 14:04
eater[sha2017] what would be the best way to implement unix:// sockets? make a new class for it or put it in IO::Socket::Async ? 14:08
jnthn eater[sha2017]: IO::Socket::Async; since we pass the port number as a 64-bit integer but it's only 16 bits, we can use some of the upper bits to indicate socket type, so it won't changes to the op interface 14:14
jnthn *change 14:14
andreoss how should i prevent .new from being called on an 'abstract' role? 14:15
eater[sha2017] jnthn: and use the host bit for the full path? 14:16
jnthn eater[sha2017]: Yeah 14:17
eater[sha2017] haha
let's go
ilmari andreoss: perl6 will refuse to auto-pun a role with stub methods 14:20
m: role Foo { method bar {…} } Foo.new
camelia 5===SORRY!5=== Error while compiling <tmp>
Strange text after block (missing semicolon or comma?)
at <tmp>:1
------> 3role Foo { method bar {…} }7⏏5 Foo.new
expecting any of:
infix
infix stopper
state…
Altreus I've lost my perl6 book :(
ilmari m: role Foo { method bar {…} }; Foo.new
camelia Method 'bar' must be implemented by Foo because it is required by roles: Foo.
in block <unit> at <tmp> line 1
merced strange tet!
xxxxxxt
Altreus Is there a perl6 book I can just print? like MP for Perl5? 14:21
timotimo m: role NoNotInTheFace { die "have to apply me to a role" unless ::?CLASS eq "NoNotInTheFace" }; class omg does NoNotInTheFace { }; say "built an { omg.new }"; NoNotInTheFace.new()
camelia 5===SORRY!5=== Error while compiling <tmp>
concatenate requires a concrete string, but got null
at <tmp>:1
Altreus use Modern::Perl 6;
timotimo whoopsie
andreoss ilmari: thanks, i've tried it with `proto` but it didn't work 14:22
ilmari andreoss: proto is for multidispatch, not requiring methods 14:22
docs.perl6.org/language/objects#Stubs
Zoffix Altreus: there's Think Perl 6 digital edition of which is free: greenteapress.com/wp/think-perl-6/
timotimo m: role Test { say ::?CLASS }; Test.new 14:23
camelia concatenate requires a concrete string, but got null
in any protect at gen/moar/stage2/NQPCORE.setting line 1033
in block <unit> at <tmp> line 1
Zoffix All the rest of the books are listed on perl6book.com/
timotimo :\
is ::?CLASS even what i'm looking for?
m: role Test { say ::?WHATEVER }; Test.new
camelia 5===SORRY!5=== Error while compiling <tmp>
No such symbol '::?WHATEVER'
at <tmp>:1
------> 3role Test { say ::?WHATEVER7⏏5 }; Test.new
timotimo it at least exists
Zoffix m: role Test { method x { say ::?CLASS } }; Test.new.x
camelia (Test)
Altreus I think I'm ahead of my time 14:24
Zoffix m: role Test { say ::?ROLE }; Test.new
camelia (Test)
timotimo i was hoping to get the thing the role was being applied to 14:25
m: role Test { method x { say ::?CLASS } }; Test.new.x; class OtherThing does Test { }; OtherThing.new.x
camelia (Test)
(OtherThing)
andreoss is there a way to make AbstractRole.new impossible on compile-time?
Zoffix timotimo: it's ::?CLASS, but you're running it inside the body, so it ain't got value yet 14:26
timotimo the body runs at composition time i thought so it ought to be okay i thought
Zoffix Ah 14:27
I thought it ran at compilation time
[Coke] ponders a pod marker for sample code like :dies-with
andreoss m: role Option[::A] { method get {...}}; role Some[::A] does Option[::A] { has A $!value; method get return A { $!value }}; my Option[Int] $x = Some[Int].new(value => 1); say $x.get; 14:32
camelia 5===SORRY!5=== Error while compiling <tmp>
Invalid trait modifier (did you mean 'returns'?)
at <tmp>:1
------> 3[::A] { has A $!value; method get return7⏏5 A { $!value }}; my Option[Int] $x = Som
andreoss m: role Option[::A] { method get {...}}; role Some[::A] does Option[::A] { has A $!value; method get returns A { $!value }}; my Option[Int] $x = Some[Int].new(value => 1); say $x.get; 14:32
camelia Died with X::TypeCheck::Return
in method get at <tmp> line 1
in block <unit> at <tmp> line 1
pmurias should EVAL q:to/END/; be allowed with MONKEY-SEE-NO-EVAL turned off? 14:34
andreoss m: role Option[::A] { method get {...}}; role Some[::A] does Option[::A] { has $.value; method get { $.value }}; my Option[Int] $x = Some[Int].new(value => 1); say $x.get;
camelia 1
andreoss why the first one doesn't "type check"? 14:35
timotimo you want "does Option[A]" instead 14:36
i find it surprising it compiles at all :)
hm. or maybe it's fine
andreoss m: role Option[::A] { method get {...}}; role Some[::A] does Option[A] { has A $!value; method get returns A { $!value }}; my Option[Int] $x = Some[Int].new(value => 1); say $x.get; 14:40
camelia Died with X::TypeCheck::Return
in method get at <tmp> line 1
in block <unit> at <tmp> line 1
andreoss it works without `returns A` in either case 14:41
Zoffix damn, so yesterday I had 3 fuck ups on modules.perl6.org that affected users. 14:54
Maybe talking on IRC while working on the site at the same time ain't a good idea. 14:55
(1) Nuked a db with a broken updater script and I forgot to back it up before doing so; (2) Killed the site due to another bug in updater script that I thought was running fine and someone had to tell me there was a 500 error; (3) While debugging the previous issue, I disabled cron job and forgot to turn it back on, so a user opened an issue about missing module 14:56
What am I doing wrong? :} 14:57
[Coke] m: say 3.skip-test 14:58
camelia No such method 'skip-test' for invocant of type 'Int'
in block <unit> at <tmp> line 1
[Coke] m: say Any.skip-test
camelia No such method 'skip-test' for invocant of type 'Any'
in block <unit> at <tmp> line 1
[Coke] ... oh, duh.
m: say 3.sum 14:59
camelia 3
[Coke] hurls github.com/perl6/doc/issues/1438 - a skipped code sample hiding a bad example. 15:11
Geth doc: 492896feb7 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/subscripts.pod6
Fix bad example

Part of github.com/perl6/doc/issues/1438		 15:23
[Coke] Zoffix: that fixes part of it. the handles bit is still broken. 15:24
ah, you said part of. Danke!
andreoss m: role Foo[::A] { has A $.v; method get returns A { $.v }}; Foo[Int].new(v=>1).get
camelia Died with X::TypeCheck::Return
in method get at <tmp> line 1
in block <unit> at <tmp> line 1
andreoss what is wrong here? 15:25
Zoffix [Coke]: well, the handles is a rakudobug IMO 15:26
I commented on the Issue
andreoss: possible a Not Implemented Yet for this use of return value specifier. 15:27
Geth modules.perl6.org: 8cfdd25e3e | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder.pm
Reverse list of p6c eco metas for faster new module loading

We reverse the list, since users tend to add their modules to the bottom of the list, by reversing it, we can load new modules to the site at the start of the run, instead of at the end.		 15:31
Eddward p6: (1,2,3,4).print 15:36
camelia 1 2 3 4
Eddward p6: (1,2,3,4).map( * ~ "\n" ).print 15:37
camelia 1
2
3
4
Eddward For some reason (1,2,3,4).print inserts newlines for me locally in the repl.
oops\
nope wrong problem. 15:38
Eddward But is there a way to do it without the spaces. 15:38
Eddward so (1,2,3,4).?.print => «1234» 15:39
araraloren m: say (1, 2, 3, 4).gist;
camelia (1 2 3 4)
Zoffix Eddward: .join.print
araraloren m: say (1, 2, 3, 4).Str; 15:40
camelia 1 2 3 4
Eddward How is that with 50kb output strings?
araraloren There some default behavior when you use print or say. 15:41
Eddward I'm ultimately trying to track down a behavior change in the latest star release.
Zoffix Eddward: try and see? Another alternative is .print for (1,2,3,4)
Zoffix Eddward: what change? 15:41
araraloren docs.perl6.org/routine/put 15:42
[Coke] m: <1 2 3 4>.join('').print;
camelia 1234
Eddward I'm reading a sqlite db, processing and print rows into mpu files.
[Coke] m: <1 2 3 4>.join('').say
camelia 1234
Eddward Now every line but the first is getting a space inserted in the front.
araraloren How do you print the rows 15:43
Zoffix Eddward: what's the exact code you're using? `.print` the method call?
Eddward I was trying to find a recreate, but I can't find a toy example that doesn't insert space in 2017.01 or 2017.04
$fh.print( $sth.allrows.flat.map(...).map(...).map(...).map(...).map( * ~ "\n" ) ); 15:44
with various values of ...
Zoffix Eddward: and when was the "working" version working? 2017.04?
Eddward 2017.04 does not insert the spaces. 15:45
Zoffix Eddward: k, the old behaviour is a bug that was fixed. 15:45
Eddward ok
araraloren you can print them use a for loop
Eddward Think the flat is still needed. I'm not sure why I have that. I suspect it was a work around for something.
araraloren $fh.print($_) for $sth.allrows.flat.map(...).map(...).map(...).map(...).map( * ~ "\n" ) 15:46
Zoffix It might be faster to .join, as originally suggested
Eddward ok
Zoffix Though we have that grapheme limit that has potential to hit. I forget if we increased it recently.
araraloren you better use recent release, rakudo is still on developing 15:47
Zoffix It's like in 1GB> strings or so, tho
Eddward I'm guessing a playlist my song library won't get that big.
Zoffix :)
timotimo m: role Option[::A] { method get {...}}; role Some[::A] does Option[A] { has A $!value; method get(--> A) { $!value }}; my Option[Int] $x = Some[Int].new(value => 1); say $x.get; 15:49
camelia Died with X::TypeCheck::Return
in method get at <tmp> line 1
in block <unit> at <tmp> line 1
timotimo interesting
Eddward I'm trying the join. Thanks 15:49
timotimo i wonder why it doesn't get any interesting info in the error message
Zoffix buggable: eco author:zoffix 15:59
buggable Zoffix, Found 35 results: Acme::Anguish, IRC::Client, WWW, GlotIO, CoreHackers::Sourcery. See modules.perl6.org/s/author%3Azoffix
Zoffix buggable: eco from:github
buggable Zoffix, Found 861 results: panda, p6doc, Bailador, 007, zef. See modules.perl6.org/s/from%3Agithub 16:00
Zoffix :/
:D
buggable: eco from:gitlab
buggable Zoffix, Found 3 results: Task::Galaxy, WebService::FootballData, App::Football. See modules.perl6.org/s/from%3Agitlab
Zoffix \o/ really cool it Just Works™ in buggable :)
Zoffix Though looks like it couldn't hurt to do some optimizations :) 16:01
ChristopherBotto Hello everyone \o/ 16:01
p6: sub MAIN { fail 'oops!'; CATCH { put "Caught: $_" } }; MAIN(); 16:02
camelia oops!
in sub MAIN at <tmp> line 1
in block <unit> at <tmp> line 1
Zoffix Some of these author fields in META files have amusing values: modules.perl6.org/search?q=author%3A2015
ChristopherBotto: was there a question? 16:03
ChristopherBotto: `fail` returns a Failure. That CATCH is basically pointless.
you can use `use fatal` in that scope to make failures fatal
p6: sub MAIN { use fatal; fail 'oops!'; CATCH { put "Caught: $_" } }; MAIN(); 16:04
camelia Caught: oops!
oops!
in sub MAIN at <tmp> line 1
in block <unit> at <tmp> line 1
ChristopherBotto Zoffix: Thanks! 16:05
Zoffix (`use fatal` is also automatically enabled in `try` blocks) 16:07
Geth modules.perl6.org: 6a7573663b | (Zoffix Znet)++ | templates/root/search.html.ep
Display actual travis/appveyor status in title=""s
 16:10
Geth doc/molecules-patch-1-1: 05fde520e0 | (Christopher Bottoms)++ (committed using GitHub Web editor) | doc/Language/syntax.pod6
s/put/ say
 16:15
Zoffix ugh 16:17
GitHub automatically shows commit comments on PR comments, so I now sound like a moron: i.imgur.com/eEY3VqD.png 16:18
raschipi Zoffix: You can also edit your own comments. 16:20
Zoffix You can also edit someone elses comments.
Actually, I like to think I was the moving force for GitHub addign the "edited by" marker. After an altercation with a hostile author who edited my comments, I pointed out to GitHub that they could've edited my comments to say racist/sexist content without my knowledge and everyone would think I actually said them :) 16:22
ChristopherBotto Zoffix: That's GitHub feature confused me too just now. 16:28
[Coke] awe have a ticketing system at a previous employer that was setup to allow certain users to edit anyone's comments, and they did not see the issue with editing other people's sends. grumble. 16:29
s/awe /we /
ChristopherBotto s/That's/That/
pmurias sends?
[Coke] comments reads cleaner. 16:31
tbrowder Skarsnik: thank! 16:35
.tell Skarsnik thanks! 16:36
yoleaux tbrowder: I'll pass your message to Skarsnik.
Skarsnik for what, xD 16:37
yoleaux 16:36Z <tbrowder> Skarsnik: thanks!
raschipi I think this else exception has the same explanation why its 'for EXPR BLOCK' instead of 'for EXPR, BLOCK', the first being two terms in a row (normally not permitted in Perl6), TimToady thinks it's looks better. 16:38
it looks better* 16:39
Geth doc: 1608091e88 | (Christopher Bottoms)++ (committed using GitHub Web editor) | doc/Language/syntax.pod6
Added mention of if/elsif/else special case

Does this look okay? Inspired by brian d foy's question (stackoverflow.com/questions/455204...paration).		 16:40
doc: 05fde520e0 | (Christopher Bottoms)++ (committed using GitHub Web editor) | doc/Language/syntax.pod6
s/put/ say
doc: 80ae3f6662 | (Aleks-Daniel Jakimenko-Aleksejev)++ (committed using GitHub Web editor) | doc/Language/syntax.pod6
Merge pull request #1437 from perl6/molecules-patch-1-1

if/elsif/else series of blocks as a special case for implied separator
Geth doc: 6284299405 | (Aleks-Daniel Jakimenko-Aleksejev)++ | doc/Language/syntax.pod6
Whitespace
 16:42
[Coke] Not sure how that's a special case compared to the normal implied statement separator. 16:49
ilmari the special case is the lack of one between } and elsif/else? 16:51
AlexDaniel ilmari: yea, that was the idea
by the way, what's the normal implied statement separator?
raschipi Newline after } implies statement separator. 17:03
But if it's an else, it's always implied, even without the new line.
[Coke] m: if True { } else { } say "what?" 17:04
camelia 5===SORRY!5=== Error while compiling <tmp>
Strange text after block (missing semicolon or comma?)
at <tmp>:1
------> 3if True { } else { }7⏏5 say "what?"
AlexDaniel raschipi: no-no, not that
raschipi Between the if and the else 17:05
M: if True {say "oi"} else { put "hey";}
m: if True {say "oi"} else { put "hey";}
camelia oi
AlexDaniel “never” implied then
[Coke] that's not an implied statement separator since it's not a new statement.
raschipi m: if True {say "oi"}␤ else { say "hey"} 17:06
camelia oi
[Coke] I get what you're showing. it's not what you said, though.
raschipi Oh, I get it now. If it's followed by an else, the implied separator ISN'T inserted. 17:08
Perl6 inserts has implied statement separators if a new line follows a closing curlie, except if there's an else in the next line. 17:09
Defenestrated the grammar on this last one. 17:10
[Coke] not quite: if it's in the middle of parsing an if...
raschipi Let me try again: Perl6 has implied statement separators if a new line follows a closing curlie, except if there's an else in the next line after parsing the block for an if. 17:12
[Coke] squints. 17:13
[Coke] . o O (yah, that reads fine. :) 17:13
AlexDaniel raschipi: well, maybe there's a way to clarify it in the docs :) 17:15
[Coke] another doc-related pod question: adding a marker to say "include the next example with this one." (for large examples split up by descriptive text) 17:16
[Coke] s/question/pondering/ 17:17
anopperl6 Machine Learning in Perl 6 ? 17:22
help ? 17:23
Zoffix heh
raschipi anopperl6: Can you go to Amsterdam? 17:24
Zoffix anopperl6: that question is too vague
ChristopherBotto raschipi: Very recently updated docs.perl6.org/language/syntax#Sep...tatements. 17:26
raschipi ChristopherBotto: Yes, it's open here already, I'm thinking about how to explain it better... 17:27
andreoss m: role Foo[::A] { has A $.v; method get returns A { $.v }} my $x = Foo[Int].new(v=>1); say $x.get; 17:29
camelia 5===SORRY!5=== Error while compiling <tmp>
Strange text after block (missing semicolon or comma?)
at <tmp>:1
------> 3has A $.v; method get returns A { $.v }}7⏏5 my $x = Foo[Int].new(v=>1); say $x.get;
expecting any of:
in…
andreoss m: role Foo[::A] { has A $.v; method get returns A { $.v }}; my $x = Foo[Int].new(v=>1); say $x.get;
camelia Died with X::TypeCheck::Return
in method get at <tmp> line 1
in block <unit> at <tmp> line 1
andreoss is it a bug?
Zoffix andreoss: 1127 Zoffix │ andreoss: possible a Not Implemented Yet for this use of return value specifier. 17:31
ChristopherBotto raschipi: The if/elsif/else exception does muddy an otherwise straight forward explanation. A more clear explanation would be welcome. 17:34
andreoss thanks 17:38
jnthn ChristopherBotto: Maybe another way of looking at it is that, from the parser's point of view, an if/elsif/else sequence actually *is* a single statement :) 17:46
Geth doc: rafaelschipiura++ created pull request #1439:
Regarding separators in if/elsif/else blocks.
raschipi Did I made it easier to understand? 17:48
And is it correct? 17:49
geekosaur suddenly reminded of rc's else handling 17:51
Geth modules.perl6.org: d167f3461f | (Zoffix Znet)++ | 6 files
Improve TODO issues page

  - Reword descriptions of problems
  - Use proper icons for non-GitHub dists
  - Use BS styling for problems
  - Include author field
  - Sort problems by name as well as by severity to avoid sort floppiness
  - Sort dists by sum total of severities of problems
  - Use same method to handle both with/without author field routes
modules.perl6.org: 44c6d8f524 | (Zoffix Znet)++ | use-me-for-commit-triggers
[REBUILD] for TODO updates
andreoss is there a way to define a subset of a function type? 17:53
i.e `subset Fun of (Int --> Int)?`
Zoffix (Int --> Int) isn't a type... 17:54
timotimo return value is part of the Callable role's parameters isn't it?
andreoss Zoffix: what is it than? 17:55
Zoffix m: subset Fun of Callable where .signature :(Int --> Int); say sub {} ~~ Fun
camelia ===SORRY!===
This type (QAST::WVal) does not support positional operations
Zoffix booo
andreoss: signature
m: subset Fun of Callable where .signature ~~ :(Int --> Int); say sub (Int --> Int) {} ~~ Fun
camelia True
Zoffix m: subset Fun of Callable where .signature ~~ :(Int --> Int); say sub () {} ~~ Fun
camelia False
Zoffix While that error as rt.perl.org/Ticket/Display.html?id=131853 17:58
s/While/Filed/; # wat?
[Coke] is finding several sliiiiightly off examples when trying to get the code to compile. (mult vs. multi; compile time vs. runtime errors) 17:59
Geth modules.perl6.org: 9ba3d72e05 | (Zoffix Znet)++ | templates/todo/index.html.ep
Toss leftover unused code
Zoffix I also removed color coding of problems 'cause I couldn't manage to make them look good otherwise. 18:02
glyphicons now code for severity: fire, warning icon, and thumbs down 18:03
Geth modules.perl6.org: 096d8d4b9a | (Zoffix Znet)++ | templates/todo/index.html.ep
Context-colour TODO issues icons
 18:07
Geth modules.perl6.org: 3a48f9ff99 | (Zoffix Znet)++ | assets/sass/main.scss
Improve contrast on nav bar
 18:16
andreoss m: my &f = *.Str; say &f.signature.returns; 18:24
camelia (Mu)
andreoss why it's not Str? 18:25
timotimo because you can't know what a Str method does
perigrin one would assume it integrates teh cream with the tea
timotimo it's just a convention
you mean it stirs? 18:26
perigrin exactly.
jnthn *groan*
perigrin grins getting an actual _groan_ out of jnthn
andreoss is it possible to specify varience of a role? like Java's `? extends Foo`, '? super Bar`? 18:37
ryu0 Zoffix: did you include a steaming pile for code quality problems? =p 18:39
Zoffix heh
Geth modules.perl6.org: ecabc9d364 | (Zoffix Znet)++ | 3 files
[REBUILD] Allow lower-cased README files
 18:53
Geth doc/molecules-patch-2: c4fed83229 | (Christopher Bottoms)++ (committed using GitHub Web editor) | doc/Language/syntax.pod6
Is this better?
 18:53
Geth doc: molecules++ created pull request #1440:
if/elsif/else construct being one statement simplifies explanation		 18:55
Zoffix Wonder if we can find 3 people to click the "Star" button on the doc repo to make it the luckiest repo: github.com/perl6/doc 18:58
luckiest = dist you get when you click "I'm feeling lucky" on modules.perl6.org without entering any search terms; i.e. modules.perl6.org/search?lucky=1 18:59
moritz stars 19:00
yoleaux 14:36Z <Zoffix> moritz: FWIW, modules.perl6.org rejects any logotypes that aren't 32x32 px in size. I see a couple of your dists get reported with that problem
Zoffix ^ that's already fixed. It was just the two dists I sent (now merged) PRs to 19:01
moritz Zoffix: yes, thanks for those 19:02
Geth modules.perl6.org: 2bb3d9ceb0 | (Zoffix Znet)++ | 2 files
[REBUILD] Link author names in search results…

  …to search results of all the author's modules		 19:14
anopperl6 hii raschipi i am not going to Amsterdam . i have money to go there .... i know there will be talk on Machine Learning in Perl 6 19:22
<raschipi> hii i am not going to Amsterdam . i have money to go there .... i know there will be talk on Machine Learning in Perl 6 19:23
Zoffix wow damian is doing a ton of presentations on TPC::EU 19:25
mspo ways to use vim, how I use vim, vim rocks, and using vim with perl 19:26
Zoffix DrForr_: if you're around; anopperl6 ^ above was looking for stuff to read re Machine Learning in Perl 6
Geth modules.perl6.org: a261a0b612 | (Zoffix Znet)++ | assets/sass/main.scss
Toss dead SASS
 20:02
Geth modules.perl6.org: 0bc13c0bff | (Zoffix Znet)++ | assets/sass/main.scss
Re-style author links back to normal text
 20:09
nadim Just when one want to go to Amsterdam but work is oozing from everywhere and it's not even worth asking the boss for a break 20:16
moritz I just realized that I've been involved with Perl 6 for over 10 years 20:17
quite a scary thought :-)
nadim scares mainly those of us that have only been on the train for a year or two
maybe someone could help me understand some of the Native call example. I am working on rendering CStruct in Cstructs like the example nopaste.linux-dev.org/?1160500
mspo moritz: the mug was thrown at you? 20:18
[Coke] no, that was 17 years ago
nadim I don't understand how the struct of struct gets initialized, the code, lines commented out generate an "Cannot modify an immutable Point ((Point))" error
timotimo that looks kind of like type object vs instance problem? 20:21
or are you trying to assign a struct to a HASed struct?
nadim I am just trying to make sense of the example here docs.perl6.org/language/nativecall...nd_CUnions 20:23
it shows a struct embeded in another struct, and a struct referenced by another struct. I just want to know how to instianciate them 20:24
I am taken all the examples in that page and dumping them, thought it wuold be a good test
and yes, I'd like to know how to set a HASed struct too. 20:25
mspo [Coke]: oh right 20:29
nadim timotimo: any code tht shows how to initialize the C data structures? 20:38
nadim Yo anyone that can show me how to instanciate the examples in docs.perl6.org/language/nativecall...d_CUnions. I can't dump what I can't instantiate. 21:03
timotimo nadim: just use the accessors or .new 21:05
hm, actually
if it HAS a cstruct maybe you can't pass it via new
nadim none of them worked that way
code is here nopaste.linux-dev.org/?1160500 21:06
[Coke] someone want to fix the second mdash there to be an actual mdash?
(on nadim's link) 21:07
raschipi bye bots 21:10
mspo ha 21:10
Geth doc: b7370de194 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/nativecall.pod6
Use proper m-dash
 21:16
Zoffix hopes the spell checkers knows that you don't put spaces around mdash and won't think those are unknown words... 21:17
[Coke] promises to fix it if so. 21:26
[Coke] was pointing at github.com/perl6/doc/blob/master/d....pod6#L356 earlier. 21:27
[Coke] will grab that when he gets home if no one beats him to it. 21:28
Geth doc: 68057bd807 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/nativecall.pod6
Use MOAR m-dashes
 21:29
nadim I'll do it but give me the code to do the initialization god dam it! ;)
nadim goes to remove a dash
nadim meh! someone was on it already 21:42
nadim Zoffix: you're a robot! ;) 21:45
travis-ci Doc build errored. Zoffix Znet 'Use MOAR m-dashes' 21:50
travis-ci.org/perl6/doc/builds/262006738 github.com/perl6/doc/compare/b7370...057bd807ba
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 21:50
lucs Wow, that's a lot of bots. 21:56
lucs doesn't know what half of 'em do.
Zoffix nadim: I thought that was well-known :) 21:57
lucs Another one!
Zoffix samcv: ^ that failure might be related to recent rakudo changes? IO::String is failing tests; \r\n things are showing up 21:58
# expected: $("hello,", "world!")
# got: $("hello,\r\nworld!\r\n",)
Zoffix m: "foo\r\nmeow".lines.perl.say 21:58
camelia ("foo", "meow").Seq
Zoffix hmm 21:58
nadim Zoffix: not by me but I am now ;) 21:59
samcv how recent are these test failures?
i haven't touched that for a week+
Zoffix samcv: lastest commits to rakudo are yours and the tests build latest 22:00
samcv: and it wasn't failing on earlier commits
samcv ah
sorry i would have expected stuff i did a while ago to do \r\n stuff but not the more recent things
nadim Zoffix: be a good sport and tell me how to initialized the damned structures. Dumping of NativeCall data start to look nice, very I think, I'd like to test with those before I update the docs and release 22:02
Zoffix nadim: I don't know
nadim hahaha!
thank you in any case :)
I laugh because I am not use to read I don't know from you ;)
samcv i didn't bump MoarVM except for a week and a few days ago
nadim who is behind most of NativeCall? 22:03
samcv looks like bumped on Aug4th
Zoffix bisect: m: dd "hello,\r\nworld!\r\n".index: "\n" 22:04
stupid bot
m: dd "hello,\r\nworld!\r\n".index: "\n"
camelia Nil
Zoffix star: dd "hello,\r\nworld!\r\n".index: "\n"
camelia Nil
Zoffix hm
samcv checking the pod tests locally though 22:04
err doc tests 22:05
Zoffix samcv: the failure's in IO::String installion tho
nadim What's the nick of Tobias?
Zoffix FROGGS 22:06
samcv yeah. i'm going to try installing doc locally 22:06
Zoffix samcv: well, I'm just guessing it was one of the latest Rakudo commits that's problematic. I've not been able to golf anything weird 22:07
lizmat_ and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2017/08/07/...eekly-101/ 22:09
nadim lizmat_++
Zoffix lizmat++ good weekly 22:24
Zoffix m: sub foo (*@a) { @a.join: '|' }; say foo <h H>.any, 'amadryas' 22:27
camelia any("h", "H")|amadryas
nadim NativeCall support in ddt, I think there is an error in struct containing struct but I need to have an instancieted one to know :), I will write a little blog entry with explanations, imgur.com/YGDW0MX
nadim I also wonder if I should display the size of structs in the dump 22:29
ijneb Can you declare 32-bit floating point Nums? 22:47
Or just 64-bit
ijneb Is it just num32? 22:48
nadim there's a num32 and a num64 23:01
not all the types are in the screenshot
there's around 20 types 23:02
ijneb nadim: Are floating-point calculations faster with float32s? 23:14
geekosaur faster but less accurate 23:15
ijneb Makes sense, thanks
geekosaur they're *really* low resolution 23:16
ijneb Why are all types capitalized except for num32 and the variants like that 23:16
why not make it Num32?
geekosaur there is an uppercase version
lowercase is native machine types, uppercase is boxed perl 6 types
ijneb ohh 23:17
So Num32 works as well?
m: my Num32 $n;
camelia 5===SORRY!5===
Type 'Num32' is not declared. Did you mean 'num32'?
at <tmp>:1
------> 3my Num327⏏5 $n;
Malformed my
at <tmp>:1
------> 3my7⏏5 Num32 $n;
geekosaur well, there's supposed to be. iirc Num32 went missing 23:17
but there is Int vs. int, etc.
ijneb ah 23:18
geekosaur hm. guess all the sized ones are only available as machine, I thought there was also e.g. Int64
geekosaur but I suppose that makes sense insofar as the native ones care more about size, the Perl 6 ones should resize themselves as necessary 23:18
ijneb I see
ijneb How would I make a class convert given ints/other types to nums 23:20
class Point { has num32 ($.x, $.y) }
How would I make that usable even when I give ints `Point.new(x => 1, y => 1)` 23:21
I understand changing the .new method somehow
raschipi ijneb: You probably want to use TWEAK instead of writing a new new 23:22
raschipi Well, there's method BUILD that's called before the object has been initialized and TWEAK that's called after 23:24
BUILD will receive the same parameters as new 23:25
ijneb ah, I see
I'll use TWEAK to convert them
oh 23:27
nevermind, BUILD is better suited, thanks lol
BenGoldberg m: class Foo { has Num $.x; has Num $.y; method new(Num(Cool) $x, Num(Cool) $y) { self.bless: :$x, :$y } }; dd Foo.new( 3, pi ); 23:40
camelia Foo.new(x => 3e0, y => 3.14159265358979e0)
BenGoldberg ijneb, ^ 23:41
BenGoldberg Sadly, type coercions only work in subroutine or method signatures; you can't put them on variables or class fields. 23:42
Geth doc: e0c760dd65 | (Will "Coke" Coleda)++ | 13 files
Reduce number of skipped tests

Where needed, add a preamble Where we have to skip, minimize the scope or declare the reason.
Fix a few samples that were obv. broken.		 23:43
doc: c8dd129d35 | (Will "Coke" Coleda)++ | doc/Language/traps.pod6
no trailing whitespace
 23:44
BenGoldberg just thought of something cool! This might be a silly idea, but what if, for the ones which remained (partly) :skip-test, you had something like :expect-exception 23:49
[Coke] I had pondered that recently; bikeshedding on what it would look like welcome. 23:50
we'd need to specify what the failure mode was, at least.
BenGoldberg Maybe :throws-like<X::SyntaxError> 23:51
[Coke] I want to split out part of doc's CONTRIBUTING.md - I'd normally create a 'docs/' folder and put stuff in there; but with 'doc/' being the main folder, I suspect that would be confusing. Any other suggested names? 23:54
BenGoldberg subdock? 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!