»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.
geekosaur (granted that this is a perl 6 specific issue, but there are equivalents with every other language's string handling.) 00:00
tyil opened an issue 00:06
for now I'll just run << curl >>
thanks for the assistance geekosaur
AlexDaniel- “Use of uninitialized value of type Any in string context.” 00:10
that's not a good sign when it comes from zef :(
can somebody try ｢zef install Image::PNG::Portable｣ ? 00:11
tyil AlexDaniel-: works on my desktop 00:13
This is Rakudo version 2017.09 built on MoarVM version 2017.09.1
- Identity: zef:ver('0.1.29'):auth('github:ugexe') 00:14
AlexDaniel- yep, nuking my install helped 00:23
tyil: thanks
AlexDaniel` meh, if I'm using two nicknames, what should I do? 00:30
` looks fine I guess
[Coke] . 00:39
tyil AlexDaniel`: setup a bouncer and go back to using one nickname :p 00:43
AlexDaniel` tyil: … tells me somebody who is using two nicknames at the same time 00:44
AlexDaniel tyil: o/
AlexDaniel` tyil[m]: o/ I mean 00:45
tyil oh, thats a whole 'nother chat service
which just happens to be bridged to irc
AlexDaniel` well… practically it's a bouncer :)
tyil (I dont use it anymore either)
it just, idles there
forever
and ever
wasting precious cpu cycles 00:46
AlexDaniel` well, I love Matrix and Riot, if only it was not built on top of chromium
tyil if it had a good client I might reconsider
AlexDaniel` tyil: what do you use for IRC? 00:47
tyil weechat
AlexDaniel` tyil: matrix.org/docs/projects/client/weechat.html ? 00:48
tyil I'd rather not touch that again
AlexDaniel` never tried it
tyil unstable client, and the encrypted chat doesnt work at all
Geth doc: d02509493c | (Will "Coke" Coleda)++ | doc/Type/Hash.pod6
trailing whitespace
 00:50
doc: cb5a24c8b7 | (Will "Coke" Coleda)++ | xt/code.pws
learn new code snippet
synopsebot Link: doc.perl6.org/type/Hash
doc: 06f0875a46 | (Will "Coke" Coleda)++ | xt/words.pws
learn new word
AlexDaniel` tyil: I see 00:51
tyil: fwiw, if you're not using tyil[m], you might as well want to bring it down
so that I wouldn't need to cycle thru two nicknames for autocompletion… :) 00:52
ye-ye, I know, 4 characters. But I'd much rather type two and press Tab
which also types “: ”
tyil I get highlighted on either 00:53
AlexDaniel` … but I have to disambiguate anyway… :S
geekosaur I thought matrix was discouraging unpaid idlers these days 00:55
AlexDaniel` geekosaur: “unpaid”? 00:58
geekosaur I was under the impression they provided limited unpaid service and a full paid/subscription service 00:59
maybe that was one of the other gateways instead
tyil afaik its all gratis
EtomicBomb hey 00:59
geekosaur (the eternal question being, if it's free, who's paying the bills? or are you, the user, also the product?)
EtomicBomb yes 01:00
tyil geekosaur: if it worked well I was going to put up my own server
tyil just like IRC, matrix runs on people donating their server power 01:00
travis-ci Doc build passed. Will "Coke" Coleda 'learn new word' 01:16
travis-ci.org/perl6/doc/builds/288365566 github.com/perl6/doc/compare/abefd...f0875a46a6
AlexDaniel` [Coke]++ 01:17
japhb .seen salortiz 01:33
yoleaux I haven't seen salortiz around.
japhb Anyone know who github:salortiz is here?
perigrin wm 19
AlexDaniel` .seen sortiz 01:46
yoleaux I saw sortiz 17 Jan 2017 23:36Z in #perl6: <sortiz> 'gnight lizmat, and thanks again
AlexDaniel` japhb: ↑ 01:47
geekosaur 10 months ago... 01:50
kjk hello, trying to use the 'rename' sub, but notinced that it returns a failure instead of throwing an exception... Currently I'm calling it like: sink rename ...; is this how it's expected to be used? 02:32
kjk basically I'd it to throw an exception instead of failing silently on me, which seems to be the case if I just call it without "sinking" it. 02:34
perigrin a) that's possibly a question better suited for #perl-help than #p5p, b) if you want it to throw an exception you probably want to use autodie 02:36
Juerd Sink context is what happens implicitly if you don't use the value
perigrin c)I'm in the wrong damn channel
stupid irssi
perigrin sigh
geekosaur er, this is neither #p5p nor #perl5 :)
whoops
Juerd perigrin: "sink" is Perl 6's void context.
perigrin geekosaur: yeah irssi just renumbered everything on me and I'm an idiot :) 02:37
perigrin moves some things around
geekosaur this is why I don't like irssi and stick to hexchat. channel pane behaves :) 02:38
Juerd kjk: For example, if you assign the return value of rename to a variable, the failure will be assigned. No exception is thrown until you use the value later on. But if you just have a bare function call, it should fail immediately.
kjk @Juerd right, I wonder why it was decided to return a Failure instead of throwing an Exception 02:38
Juerd kjk: That's the easiest way to enable ignoring the problem if your program doesn't mind 02:39
perigrin geekosaur: it usually isn't a problem unless my host restarts over night without my being aware of it
Juerd kjk: And it's the easiest way to allow simple boolean error handling: rename ... or die "didn't work :("
AlexDaniel` perigrin: this reminds me how one time I was explaining to somebody the difference between "" and '', only to realize that I'm on #python
kjk since I think most people just call it like this: rename from-name to-name; but then that would silently fail :( 02:40
AlexDaniel` perigrin: I ran away once I realized
Juerd kjk: If it would throw an exception, then you would need to actually CATCH the error to get your own error message, or to ignore it.
perigrin I'll own being an idiot. I do it often enough.
Juerd kjk: Actually, that shouldn't fail silently. Can you share the code that does?
geekosaur kjk, if they call it like that, the Failure should be sunk and get thrown
kjk really? hold on I'm sure it didn't get sunk, let me check 02:41
geekosaur if they assign the result, that is a Failure object; testing it as a Boolena lets you see if it succeeded or not as normal flow control instead of as an exception
but using it in any other way should throw the deferred exception
also, for when you really want stuff to just explode immediately, there is 'use fatal' 02:42
AlexDaniel` this reminds me of github.com/perlpilot/p6-sake/issues/12
(lol, everything reminds me)
Juerd wonders if there's implicit returning going on, obscuring the non-sink context through propagation, as in sub mv { rename $^a, $^b }; my $x = mv $foo, $bar; 02:43
AlexDaniel` that's exactly what I'm thinking
kjk yeah, just read about 'use fatal'. nice. anyway, coming up with a short example to reproduce it 02:46
ok can't reproduce it with new code, I'm just going to share my existing code that has the problem in a gist 02:49
japhb Are there any references on writing (and/or using) a grammar to make it produce decent error messages on failure?
kjk @Juerd, sorry to dump this on you, but here's the script I'm writing that's having the issue with rename. gist.github.com/kjkuan/1f575bb323b...06193548c2 02:53
Juerd kjk: I don't mind but I do feel free to ignore it because I'm almost going to bed :) 02:54
I'll have a quick look though
kjk oh no worries then, thanks for helping 02:54
Juerd It seems that for doesn't propagate sink context? 02:55
I wonder if that's a bug or an obscure feature.
geekosaur I was just thinking that. probably returns the last thing in the block or somethin
then MAIN's magic becomes an edge case
Juerd geekosaur: But the loop is the last thing in MAIN and I'd expect *that* to be sunk...
The same thing without the loop would have thrown the failure 02:56
geekosaur Im not sure I'd expect that. depends on whether MAIN was supposed to inspect the final value to decide perl 6's exit code
in which case that might be an NYI or something with an unfortunate side effect
Juerd I don't think it has to do with MAIN
The same thing happens with a regular named sub
geekosaur ...
Juerd perl6 -e'sub mv ($a, $b) { for 1 { rename $a, $b } }; my $x = mv "does-not-exist", "meh";' 02:57
Err 02:58
Shouldn't have that assignment obviously :)
Same result without the "my $x ="
geekosaur m: sink [1,2,Failure.new('boo!')]
camelia ( no output )
geekosaur thinkI just golfed it
for returns a list of the final block values. one of those values is a Failure. sinking the list doesn't fire the Failure 02:59
Juerd kjk: A work-around is to add a statement after the rename, like simply "True;" as in the git issue AlexDaniel` linked.
Good night #perl6
kjk k, thanks
geekosaur m: sink [Failure.new('boo!')]
camelia ( no output )
AlexDaniel` “use fatal;” is probably better
m: use fatal; say sink [1,2,Failure.new(‘boo’)]
camelia boo
in block <unit> at <tmp> line 1
 02:59
kjk so is this a bug or just a trap?
AlexDaniel` it's a trap, but maybe we don't have it documented yet 03:00
geekosaur mm. I think for now I say trap, because having sink context force evaluation of a possibly infinite list to trip embedded Failures seems kinda bad
AlexDaniel` kjk: maybe you can open an issue on github.com/perl6/doc/issues ? 03:00
kjk ok @AlexDaniel, I'll just say document this trap in the issue 03:02
AlexDaniel` kjk: yes
wander how to use X<> in pod6 file 03:11
see X«Blocks and Lambdas|syntax,->» but don't know what it means
it said the form is X<if|control>, while I can't figure it out 03:12
geekosaur before | is the text rendered at th epoint of use, after is the index entry to link to 03:15
design.perl6.org/S26.html#Indexing_terms (I think for pod6 currently that's the best doc we have :/ )
wander ty 03:16
AlexDaniel` kjk: thanks! 03:22
kjk cool, good night 03:24
wander m: multi sub f() { say "Hi" }; multi sub f(:$g) { say $g; }; f(); 06:44
camelia (Any)
wander why it calls `multi sub f(:$g)` 06:45
m: multi sub f() { say "Hi" }; multi sub f(:$g) { say "Hello"; }; f(); 06:47
camelia Hello
wander ok, I find the sentence `Named parameters participate in the dispatch even if they are not provided in the call. Therefore a multi candidate with named parameters will be given precedence.' 07:13
wander m: class A { multi method f(:$g) { say $g }; multi method f(:$g, :$h) { say $g ~ $h } }; A.new.f(:g('g'), :h('h')); 07:23
camelia g
wander so this one? I provide both $g and $f, but it calls the method that only use $g 07:24
whlie 07:26
m: multi sub f(:$g) { say $g }; multi sub f(:$g, :$h) { say $g ~ $h }; f(:g('g'), :h('h'));
camelia gh
atroxaper wander: docs.perl6.org/language/functions#...i-dispatch says that you have to use required named parameters. But in doesn't work too. I can't find any tests in roast for that. Probably it is a bug. 07:59
wander m: class A { multi method f(:$g is required) { say $g }; multi method f(:$g is required, :$h is required) { say $g ~ $h } }; A.new.f(:g('g'), :h('h')); 08:03
camelia g
wander maybe
geekosaur I think there are some known issues with that? seem to remember a few instances of things depending on order in ways they're not really supposed to, possibly as a hack to get things working initially 08:04
m: class A { multi method f(:$g!, :$h!) { say $g ~ $h }; multi method f(:$g!) { say $g }; }; A.new.f(:g('g'), :h('h')); # order dependent :( 08:06
camelia gh
wander well, like protoregexes use first match alternation, not longest match alternation 08:09
timotimo no, they use longest token matching 08:10
wander timotimo, gist.github.com/W4anD0eR96/db0d6d5...7237130385 08:11
timotimo interesting! let me look closer at that 08:12
oh, the sigspace is not declarative, right? 08:13
the default ws you get, i mean
timotimo so the declarative prefixes of sub and num are the same length, and so a tiebreaker has to decide which gets precedence 08:13
timotimo hah, now i get it to fail every time, regardless of which i put first 08:16
wander don't understand declarative prefixes well, but by convert `rule' into `token', it uses longest match 08:18
timotimo doesn't work properly for me 08:20
if i put \s+ in between things and use token, it does work
wander right, i use "2-1" instead 08:21
jnthn rule means to insert <.ws> in place of (non-leading) whitespace. <.ws> is not declarative, unless you override the definition
timotimo yeah, i declared ws to be \s* or \s+ and that didn't help with "rule" 08:21
jnthn As for mutli-dispatch, named parameters work as constraints, not declaratively 08:22
timotimo good morning jnthn :)
jnthn o/ timotimo :) 08:23
Geth ecosystem: scovit++ created pull request #376:
Add Scheduler::DRMAA		 08:27
Geth ecosystem: d133303171 | (Vittore F. Scolari)++ | META.list
Add Scheduler::DRMAA
 08:28
ecosystem: 8a81ce1f3d | timo++ (committed using GitHub Web editor) | META.list
Merge pull request #376 from scovit/master

Add Scheduler::DRMAA
see github.com/scovit/Scheduler-DRMAA
moritz we should really update the modules guide to recommend uploading to CPAN 08:32
wander jnthn: multi-method seems odd, irclog.perlgeek.de/perl6/2017-10-16#i_15308827 08:33
tadzik scheduler drama :o 08:34
wander can't i distinguish multi-method by different numbers of named para? 08:35
moritz to first approximation, multi dispatch ignores named params 08:36
timotimo it'd also be nice if something told people to put a link to the github repo page in their commit messages
i find that a gigantic help
TEttinger eh? 08:37
that would be a mess browsing on github
DRY etc.
timotimo you think so?
i find it messy to have to either type the whole url myself or copy-paste and edit parts to get to it
DrForr Could you edit the boilerplate commit message to add a big ol' blinking notice? 08:38
timotimo no, that's the pr message, which is "just" a comment that github has 08:39
i was hoping for the message on the actual commit to have it, which github won't let us suggest anything for i don't think
so if they see it in the pr template, they'll have to go back and amend the commit to include the text
wander moritz, while multi-sub can distinguish named para? 08:40
DrForr Oh, I thought you meant the 'git commit' editing boilerplate, not.. right, ignore me :)
TEttinger "Fixed typo. github.com/MyXOF/Deep-Learning-Pap...ng-Roadmap "
timotimo what? 08:41
TEttinger putting a link in with the message 08:41
timotimo github.com/perl6/ecosystem/commits/master - i'm talking about these commits
DrForr Ah, gotcha. 08:42
timotimo it's already nicer to have "Add Foo::Bar to ecosystem" compared to "Update META.list"
TEttinger I don't know anything about the perl6 ecosystem. if people need to do this by hand, they won't
timotimo we can do it for them, but that'd mean rebasing and force-pushing, or doing the equivalent of github pull-requests some other way 08:43
anyway, the ecosystem is on the way out already
but yeah, like moritz said, we should update the module guide to discuss cpan prominently. also the ecosystem repo readme should mention it 08:44
DrForr I assume existing modules haven't already been pushed to CPAN? I.E. I need to do this for Readline &c? 08:50
timotimo yeah 08:51
it's something every distro author has to do themselves
rakudo.party/post/CPAN6-Is-Here 08:52
HoboWithAShotgun .tell DrForr On the site for Perl6Tidy (github.com/drforr/perl6-Perl6-Tidy) you claim it could be installed by zef, but it isn't in the eco system (modules.perl6.org/search/?q=tidy). 09:39
yoleaux HoboWithAShotgun: I'll pass your message to DrForr.
timotimo HoboWithAShotgun: you can git clone and zef install . 10:02
wander .tell wander test how yoleaux works 10:09
yoleaux wander: Talking to yourself is the first sign of madness.
wander lol 10:10
HoboWithAShotgun m: sub in { my $p = Supplier::Preserving.new; start for 1 .. 10 { .say; $p.emit($_); }; }; react whenever in() { ">$_".say; } 11:16
camelia 1
Use of Nil in string context
2
3
4
5
6
7
8
9
10
>
in block at <tmp> line 1
Use of Nil in string context
in block at <tmp> line 1
Use of Nil in string context
in block at <tmp> line 1
Use of Nil in …
HoboWithAShotgun what the heck am i missing here? 11:17
lizmat HoboWithAShotgun: add $p at the end of sub in() . You're not returning the supply :-) 11:23
m: sub in { my $p = Supplier::Preserving.new; start for 1 .. 10 { .say; $p.emit($_); }; $p }; react whenever in() { ">$_".say; };
moritz whenever in() ?
camelia (timeout)1
moritz shouldn't that be whenever in($seconds) ? 11:24
gfldex m: my $p = Supplier::Preserving.new; start for ^10 { $p.emit($_); }; react { whenever $p { $_.say; }; whenever Promise.in(2) { done } }; 11:25
camelia 0
1
2
3
4
5
6
7
8
9
timotimo it'd be a good idea to "done" the supplier when all values have been emitted 11:30
also, best to return $p.Supply rather than $p itself
moritz: no connection between sub in and Promise.in 11:31
timotimo m: sub in { my $p = Supplier::Preserving.new; start { for 1 .. 10 { .say; $p.emit($_); }; $p.done }; $p.Supply }; react whenever in() { ">$_".say; } 11:32
camelia 1
2
3
4
5
6
7
8
9
10
>1
>2
>3
>4
>5
>6
>7
>8
>9
>10
timotimo HoboWithAShotgun: ^
this simple case is best done using a supply block, though, IMO.
raschipi \/win 26 12:08
Geth doc: gfldex++ created pull request #1609:
doc Supplier::Preserving		 13:38
Geth doc: 524851f5a0 | (Wenzel P. P. Peppmeyer)++ (committed by Zoffix Znet) | doc/Type/Supplier/Preserving.pod6
doc Supplier::Preserving (#1609)
 13:47
synopsebot Link: doc.perl6.org/type/Supplier/Preserving
timotimo if there's ever a cro-redux integration module it'll surely be called Crodux 13:52
raschipi timotimo: better squat the namespace now, haha 13:56
timotimo crod.ux/ 13:56
buggable New CPAN upload: Grammar-ErrorReporting-0.2.tar.gz by MORITZ cpan.metacpan.org/authors/id/M/MO/...0.2.tar.gz 14:31
[Coke] .tell zoffix Thanks for Proc::Q, it's basically exactly what I needed to build. 14:32
yoleaux [Coke]: I'll pass your message to zoffix.
[Coke] also jnthn++ for all the plumbing and for reminding me to check the ecosystem. :) 14:35
Zoffix m: class Foo {…}; class Bar { method y($o) { $o!Foo::z } }; class Foo { trusts Bar; method !z {}}; Bar.new.y: Foo.new 16:00
camelia 5===SORRY!5=== Error while compiling <tmp>
Cannot call private method 'z' on package Foo because it does not trust Bar
at <tmp>:1
------> 3…}; class Bar { method y($o) { $o!Foo::z7⏏5 } }; class Foo { trusts Bar; method !z
expe…
Zoffix Do we have a way of making this work?
Seems `trusts` ain't working when you got a stub first 16:01
jnthn m: class Bar { ... }; class Foo { trusts Bar; method !z {}}; class Bar { method y($o) { $o!Foo::z } }; Bar.new.y: Foo.new 16:02
camelia ( no output )
jnthn I think it's 'cus of one-pass parsing 16:03
Well, and it doing the checking at the point it sees the call
Zoffix Ah
Zoffix m: class Foo {…}; class Bar {…}; BEGIN Foo.^add_trustee: Bar; class Bar { method y($o) { $o!Foo::z } }; class Foo { method !z { say "in it" }}; Bar.new.y: Foo.new 16:08
camelia in it
Zoffix :P
ZofBot: ♫ these chains ain't gonna hold me ♫ 16:09
Zoffix So looks like `trusts` trait is just missing from grammar: github.com/rakudo/rakudo/blob/nom/...3304-L3322 16:10
And if that's fixed, then `class Bar {…}; class Foo trusts Bar {…}; ....` will prolly work 16:11
|5d4h check out trusts stuff github.com/rakudo/rakudo/blob/d6a9...3304-L3322
ZofBot Zoffix, Will remind you on 2017-10-21T16:11:52.499374-04:00 about check out trusts stuff github.com/rakudo/rakudo/blob/d6a9...3304-L3322
Geth doc: ab366f63b3 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/concurrency.pod6
Remove confusing sentence part

  "Or alternatively..." alternatively to what? Previous para talks
about creating live supplies and para before that already differentiates between two kinds of supplies.		 16:22
synopsebot Link: doc.perl6.org/language/concurrency
Geth doc: 8810a1a942 | (Zoffix Znet)++ (committed using GitHub Web editor) | doc/Language/concurrency.pod6
Add missing verb
 16:24
Halldor Hi. The code examples on the perl6 website (eg examples.perl6.org/categories/best-...game.html) are buggy. All instances where a bare word is in angle bracket (<exp> in a regexp) does not show up in the webpage, since the html in browser eats it up. 16:40
Zoffix oops :) 16:41
Halldor Looking at the html source we see the <exp> words are there. Thanks for all your work. 16:42
Over and out.
timotimo ouchies! :) 16:44
Geth perl6-examples: 00cac09228 | (Zoffix Znet)++ | 2 files
Escape HTML when we're not highlighting code
 16:55
Zoffix sacrifices a goat to coding gods 16:58
K. That should do it.
If it doesn't... use more goats.
[Coke] mmm, goat. 17:00
Zoffix that thing builds a whole 'nuther rakudo on each change to repo? :o 17:01
Geth perl6-examples: 534ff9a24b | (Zoffix Znet)++ (committed using GitHub Web editor) | README.md
Remove mention of panda
 17:10
Geth perl6-examples: 8c0d0e1a6f | (Zoffix Znet)++ (committed using GitHub Web editor) | util/update-and-sync
Don't build a new rakudo each time

Use the one we use for everything instead		 17:13
perl6-examples: e5f39bbde5 | (Zoffix Znet)++ (committed using GitHub Web editor) | README.md
Link to chat
 17:19
Geth perl6-examples: d570b0f2b1 | (Zoffix Znet)++ (committed using GitHub Web editor) | META6.json
Fix broken JSON
 17:25
Zoffix k, I think it worked... 17:38
Zoffix is only 97% sure auto-build is working, tho
Site wasn't updating and build log said it was done, then I went inside and tried to find html files manually but there was nothing and then I ran `make html` manually and *while it was running* site got updated 17:39
So I'm not following wth happened there, but I think it should be working OK
And markup is working now \\o/ 17:40
Zoffix flies away into the horizon
[Coke] I am tempted to rewrite all of xt/ with Proc::Q. :) 18:59
(for perl6-docs)
lizmat starts on the P6W 19:00
let me know if you think I'm about to miss something
AlexDaniel` . 19:06
AlexDaniel .
AlexDaniel` lizmat: there were some notes in .tell, not sure if you noticed
lizmat yeah, I noticed :-)
fwiw . 19:08
yoleaux 14 Oct 2017 03:44Z <AlexDaniel-> fwiw: RT #132292. IMO it's blocker-worthy, but not marking it as such just yet. What do you think?
synopsebot RT#132292 [open]: rt.perl.org/Ticket/Display.html?id=132292 Recursively .emit-ing from the tap of the same supply bails out
El_Che lizmat: I almost had news for the P6W: I almost had alpine pkgs for a small rakudo footprint. Got around the FPM Ruby prob when building the packages, but I hit an an other bug (I reported upstream) 19:10
you can't win :) 19:11
AlexDaniel` you can, just not this week :) 19:12
AlexDaniel` (and a half-win is a worthy mention anyway I think) 19:13
El_Che nah, half of nothing is still nothing
we'll get there
Zoffix Is there a way to somehow signal to the channel or supply that when a value is received it should just close it? Basically, I got a setup like this currently: github.com/zoffixznet/perl6-Reminders#synopsis A react block just recieves values and my object has a marker that signals when all the values were *sent*. But that has a race and the marker triggers end of react if all the stuff is sent before 19:21
previous stuff got received: github.com/zoffixznet/perl6-Remind...m6#L45-L48
One thing I can think of is having a `.close` method on my object that will send some dummy value down the Supply or Channel and the receiver of it can be done based on that, but that's kinda iffy, since in some usecases the user may not want to look for dummy value 19:22
Basically, I need a way to Channel.close or Supply.done but for that to trigger only when all the current values chave been consumed. 19:23
hmm
Zoffix hackety hacks a bit
lizmat El_Che: looking forward to being able to report success 19:24
Zoffix: this feels similar to the issue with iterators signalling that they're done 19:25
create a Mu.new, send that raw and then do an =!= to check if it's done?
Zoffix The user would need to check that tho, no? 19:26
I don't want the user to do that
It'd normally work this way, but I got Promises future-sending values. I think I got it though: add marker to future-done in those cases 19:27
Zoffix compiles and runs
El_Che "Zoffix complies and runs" sounds cooler 19:28
:)
raschipi Run, Zoffix, run! 19:29
Zoffix Yeah, this seems to do that I want it to: github.com/zoffixznet/perl6-Remind...3e4ff7c0d2 19:30
Zoffix ... after the race is fixed :) 19:31
El_Che puts "Small Town Boy" by Bronski Beat on the turn table
lizmat Zoffix: how can I check new CPAN uploads of the past week ? 19:32
Zoffix ummm.. there's no easy way yet 19:33
Zoffix c: use lib <data/all-modules/zoffixznet/perl6-WWW>; use WWW; for (^7+9)».fmt: "%02d" -> { "irclog.perlgeek.de/perl6/2017-10-$....contains: "New CPAN upload")».say } 19:35
committable6 Zoffix, ¦use: «Cannot find this revision (did you mean “all”?)»
Zoffix c: HEAD use lib <data/all-modules/zoffixznet/perl6-WWW>; use WWW; for (^7+9)».fmt: "%02d" -> { "irclog.perlgeek.de/perl6/2017-10-$....contains: "New CPAN upload")».say } 19:35
committable6 Zoffix, gist.github.com/93ea03eff4d0334e98...aaae3c9ebe
AlexDaniel` ehhh it changed
AlexDaniel` data/github/all-modules I think 19:35
c: HEAD use lib <data/all-modules/github/zoffixznet/perl6-WWW>; use WWW; for (^7+9)».fmt: "%02d" -> { "irclog.perlgeek.de/perl6/2017-10-$....contains: "New CPAN upload")».say } 19:36
committable6 AlexDaniel`, gist.github.com/61b9682ed15319c4e3...80c268ccba
AlexDaniel` or that
c: HEAD use lib <data/all-modules/github/zoffixznet/perl6-WWW>; use lib <data/all-modules/cpan/TIMOTIMO/JSON-Fast/lib>; use WWW; for (^7+9)».fmt: "%02d" -> { "irclog.perlgeek.de/perl6/2017-10-$....contains: "New CPAN upload")».say } 19:37
committable6 AlexDaniel`, gist.github.com/1d077332721b10878e...46696d3f05
AlexDaniel` oh goddamit
c: HEAD use lib <data/all-modules/github/zoffixznet/perl6-WWW>; use lib <data/all-modules/cpan/TIMOTIMO/JSON-Fast/lib>; use lib <data/all-modules/github/sergot/http-useragent/lib>; use WWW; for (^7+9)».fmt: "%02d" -> { "irclog.perlgeek.de/perl6/2017-10-$....contains: "New CPAN upload")».say }
committable6 AlexDaniel`, gist.github.com/5d67ed586ec32bba44...df74c5c320
AlexDaniel` c: HEAD use lib <data/all-modules/github/zoffixznet/perl6-WWW>; use lib <data/all-modules/cpan/TIMOTIMO/JSON-Fast/lib>; use lib <data/all-modules/github/sergot/http-useragent/lib>; use lib <data/all-modules/github/sergot/perl6-encode/lib>; use WWW; for (^7+9)».fmt: "%02d" -> { "irclog.perlgeek.de/perl6/2017-10-$....contains: "New CPAN upload")».say } 19:38
committable6 AlexDaniel`, gist.github.com/d0581c276c7bcdb788...183e8c5392
AlexDaniel` … are you kidding me?
Zoffix lizmat: here's the list for past 7 days, if it's helpful: gist.github.com/zoffixznet/3d64648...f009a4740f 19:39
Eventually, modules.perl6.org/recent would show this stuff
AlexDaniel` c: HEAD use lib <data/all-modules/github/zoffixznet/perl6-WWW>; use lib <data/all-modules/cpan/TIMOTIMO/JSON-Fast/lib>; use lib <data/all-modules/github/sergot/http-useragent/lib>; use lib <data/all-modules/github/sergot/perl6-encode/lib>; use lib <data/all-modules/github/supernovus/perl6-http-status/lib>; use WWW; for (^7+9)».fmt: "%02d" -> { "irclog.perlgeek.de/perl6/2017-10-$....contains: "New CPAN
upload")».say }
committable6 AlexDaniel`, gist.github.com/e4512a26d854c1f9ec...95aeeb347c
AlexDaniel` and now it no longer fits in one message :D
Zoffix lizmat: I guess you need links to ecosystem. 1 sec
AlexDaniel` Zoffix: fwiw, it's probably much easier to just run ‘curl’ 19:41
Zoffix uuhh 19:50
Gotta relocate
This is what I had. Just need to make it compile and make urls poiting to modules.perl6.org/dist/$name:cpan gist.github.com/zoffixznet/9b097d4...277e730d12 19:51
where $name is the name of the dist without `-$version` part (and possibly with replacing dashes by colons)
Zoffix &
It's probably faster to implement /recent route on modules.perl6.org :P 19:52
HoboWithAShotgun m: sub x { return my $foo if $foo = 1; }; say x 20:03
camelia 1
HoboWithAShotgun interesting, that doesn't work at all in perl 5 20:06
but why do i have to put the my on the return? i mean, logically the if comes first
perigrin my $foo if ... is I believe officially undefined in perl5 20:07
since like 5.10 or something 20:08
HoboWithAShotgun i have 5.22 and i "get requires explicit package name" errors either way 20:09
Zoffix HoboWithAShotgun: only once executed, while compiling, the first one is first. The if is a statement modifier
Zoffix constant x = say 42 if 0 20:10
m: constant x = say 42 if 0 20:11
camelia 42
perigrin HoboWithAShotgun: see the note in perldoc.perl.org/perlsyn.html#State...-Modifiers 20:12
"The behaviour of a my, state, or our modified with a statement modifier conditional or loop construct (for example, my $x if ... ) is undefined."
Zoffix Actually, instead of parsing logs, we can just use JSON from modules site 20:13
it doesn't tell which dists are new tho, I don't think; onpy last updated 20:14
Zoffix gives up on this 20:21
lizmat Zoffif++ # looking into it 20:25
I'll figure out something :-) 20:26
patrickz Hi everyone! 20:27
HoboWithAShotgun perigrin: better to avoid such a construct then ;)# 20:28
perigrin quite :)
patrickz Is it possible with NativeCall to pass a perl6 function as a callback to a C function? (So the C code can call back into my perl6 code.) 20:29
HoboWithAShotgun yes, i remember that having been explicitly said in a talk i watched online 20:30
HoboWithAShotgun but don't ask ME how ^^ 20:31
ryn1x_ Is there a limit to how large of an array can be dynamically allocated, besided system limits? I am getting: > my @img = slurp "img2.tif", :bin Unable to allocate an array of 16781312 elements 20:34
lizmat ryn1x_: could you gist a stacktrace with --ll-exception ? 20:36
AlexDaniel` and how much memory do you have?
ryn1x_ AlexDaniel: 128GB on this box. File is only 16MB 20:37
AlexDaniel` very nice box you have there :)
lizmat 128G RAM ?
ryn1x_ lizmat: yes, 128GB. Let my try to get a stacktrace. I haven't done that with perl6 before. 20:38
lizmat add "--ll-exception' as a parameter :-)
AlexDaniel` xkcd.com/859/ ' " 20:39
lizmat " 20:40
hope that helps :-)
HoboWithAShotgun maybe that's the problem. too much ram, something overflows and perl thinks wrongly about how much memory it has 20:41
but then, they wouldn't repeat microsofts mistakes, would theyß 20:42
ryn1x_ Yeah this is a work computer and unfortunately it is a windows box. Our typical setup is 128-256GB of RAM. 20:44
ryn1x_ :lizmat waiting for a stacktrace now... the repl freezes for a while trying to read in the file... 20:45
lizmat ryn1x_: weird, slurping a 16MB should be less than a second
ryn1x_ :lizmat just un-froze. I am running the repl with the --ll-exception flag and did not get a stacktrace. Just this message again: Unable to allocate an array of 16781312 elements 20:47
ryn1x_ Guess my home laptop went to sleep ha. 20:50
lizmat ryn1x_: can you try running it outside of the REPL ? 20:53
ryn1x_ :lizmat ok 1 sec. I am also installing Rakudo Star on another box to try that out too. 20:54
ryn1x_ :lizmat looks like it might be doing the same thing... terminal is waiting... 20:57
lizmat feels to me something weird is going on there. Is the code you run just the slurp() or is there more? 20:58
ryn1x_ just slurp. I put it in a sub MAIN but I can pull it out to the top level... nothing else in the file 20:59
lizmat to get some metrics: slutping a 16MB file into a scalar takes 1.7 seconds for me, memory usage goes up to about 350 MB 21:04
lizmat *slurping 21:04
jnthn just did slurp with :bin of a 450 MB file
(And it worked out) 21:05
So not sure hwat's going on...16MB is pretty tiny
AlexDaniel` used to slurp and spurt half-G files a lot
ryn1x_ lizmat: it is working now. I put it in a file all alone and it ran instantly. Then added a say $img and it froze forezer. Then changed to say[0] and it printed the first value. Maybe the repl errors because it tries to print the large array after it is slurped?
jnthn Ah, it's probably that
Trying to nicely format it
m: say Buf.new(1,2,3)
camelia Buf:0x<01 02 03>
jnthn Like that
:) 21:06
jnthn afk for a bit
AlexDaniel` if it's the .gist that does that, then it should be smarter
lizmat hmnmm... but a gist should limit to 100 elements 21:07
ryn1x_: are you slurping with :bin ? 21:08
ryn1x_ lizmat: yes.
lizmat hmmm 21:09
nothing comes to mind atm :-(
but then I'm also writing the P6W atm
AlexDaniel` ryn1x_: maybe submit a ticket so that it doesn't fall through the cracks? 21:11
gfldex Geth: are you alright mate? 21:18
ryn1x_ :AlexDaniel' will do 21:22
Zoffix ryn1x_: what version of rakudo are you using? 21:42
Buf.gist was already fixed like a week ago
gfldex: what repo was it meant to report?
Zoffix something's wrong with it 21:43
gfldex github.com/perl6/doc/pull/1611
Zoffix oh, s/it/hack/;
can't ssh
Zoffix looks for other keys
Zoffix uhh, it's asking me for password 21:46
TEttinger "password" 21:47
Zoffix (connected, just trying to see why it's unhappy) 21:54
Zoffix no idea 21:57
Geth doc: gfldex++ created pull request #1611:
link to Supplier::Preserving, fixes #1610		 21:59
Geth doc: 91cd70de96 | (Wenzel P. P. Peppmeyer)++ (committed by Zoffix Znet) | 3 files
link to Supplier::Preserving, fixes #1610 (#1611)
 22:01
lizmat And another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2017/10/16/...seriously/ 22:02
Zoffix wooo
lizmat++ # good weekly 22:07
"Thanks to Zoffix Znet and Aleks-Daniel Jakimenko-Aleksejev we will soon have a page".... I guess I just walked right into that :P 22:08
|5d make recent modules pages on modules.perl6.org, bruh
ZofBot Zoffix, Will remind you on 2017-10-21T18:08:21.578495-04:00 about make recent modules pages on modules.perl6.org, bruh
ryn1x_ Zoffix: I am using 2017.07 22:09
Zoffix ryn1x_: yeah, that doesn't yet have Buf.gist fix. 22:10
(if that's what the problem was about_ 22:11
)
ryn1x_ I tried to reproduce the result on a unix laptop I have here with the same file and it has been paused at the repl for ~45mins now. Even small files, under 1MB take quite a long time to slurp when in the repl.
Zoffix Yeah, it was dog slow 22:12
c 2017.07,HEAD Buf.new(1..1000).gist; say now - INIT now
slurp itself is likely fast, just the gist is slow 22:13
c: 2017.07,HEAD Buf.new(1..1000).gist; say now - INIT now
committable6 Zoffix, ¦2017.07: «0.1359061» ¦HEAD(d6a9eda): «0.0196757»
ryn1x_ I am going home now, but I will check the laptop in the morning and see it actually printed the array after slurping or if it gives me the same "unable to allocate array of x elements" error as the windows box.
Zoffix: I think you are right.. because it slurps instantly in a source file. 22:14
Zoffix ryn1x_: there was a ticket for that; it's possible that got fixed after 2017.07 release
for allocate array on windows
lizmat goes to get some shuteye
ryn1x_ oh interesting.
Zoffix Can't find it. 22:18
Geth modules.perl6.org: 034202620c | (Zoffix Znet)++ (committed using GitHub Web editor) | use-me-for-commit-triggers
[REBUILD] due to cpan caching bug
 22:26
perllearner89 m: say ""; sub postfix:<!>(Int $n) { [*] 2..$n; }; say "6! is { 6! }."; 22:49
camelia 5===SORRY!5=== Error while compiling <tmp>
Negation metaoperator not followed by valid infix
at <tmp>:1
------> 3(Int $n) { [*] 2..$n; }; say "6! is { 6!7⏏5 }.";
expecting any of:
infix
infix stopper
perllearner89 m: sub postfix:<!>(Int $n) { [*] 2..$n; }; say ""; say "6! is { 6! }.";
camelia
6! is 720.
raschipi .seen zoffix 22:51
yoleaux I saw Zoffix 22:21Z in #perl6-dev: <Zoffix> ZofBot: must be a human thing...
raschipi .tell zoffix when will we get another toasting round? 22:52
yoleaux raschipi: I'll pass your message to zoffix.
Zoffix raschipi: I don't do it no more. Someone else can do it 22:54
yoleaux 22:52Z <raschipi> Zoffix: when will we get another toasting round?
Zoffix raschipi: github.com/zoffixznet/perl6-Toaster
Zoffix perllearner89: that's not fixed yet. Was gonna do it next weekend 22:55
perllearner89: ticket: rt.perl.org/Ticket/Display.html?id...et-history
Zoffix realizes the weekend quickly is getting packed... and it's just Monday :}
AlexDaniel` raschipi: last time it was done by Skarsnik 22:56
perllearner89 Zoffix: Cool. I just figured I would do a little test. Thanks for your help.
AlexDaniel` raschipi: if you're willing to toast yourself, let me know and I'll give you the DB file (so that we can compare previous results with your run)
raschipi: one way or another I want toaster results this week, so if you're willing to help that'd be great 22:57
raschipi I'm willing to help, but I don't have any resources. I wonder if some company interested in Perl6 can donate them. 23:01
Zoffix (related: Google Compute Engine gives out $300 free credit to try out) 23:02
AlexDaniel` ↑ this is exactly how Skarsnik did it the last time 23:03
raschipi Well, then each week someone different will have to run it... I don't want to put effort into a temporary solution. 23:04
AlexDaniel` Zoffix: how much does it take to use these $300 ? 23:05
AlexDaniel` Zoffix: do you know if we have a ticket for ｢say 1.111111111111111111111｣ ? 23:06
MasterDuke i've profiled the build of rakudo ~20 times in a vm with 4 cores and 26gb ram and used ~$2.47 23:08
callyalater m: .say for 1..10; 23:09
camelia 1
2
3
4
5
6
7
8
9
10
AlexDaniel` yet another drive-by-eval :) 23:12
Zoffix AlexDaniel`: depends on VM you run. The one I use is 70c/h or 1.5/h or something along those lines 23:18
They got a calculator somewhere. "pre_emptive" VMs are a lot cheaper 23:19
m: say 1.111111111111111111111
camelia 1.11111111111111111604544
Zoffix m: say 111111111111111111111.log: 2
camelia 66.5905649911923
Zoffix AlexDaniel`: it's in my private bug stash
And in a "definitely gonna fix it" section :) 23:21
AlexDaniel` well, you can now add a link to RT #132313 :) 23:23
synopsebot RT#132313 [new]: rt.perl.org/Ticket/Display.html?id=132313 .gist of some Rats does unnecessary Numification (say 1.111111111111111111111)
timotimo MasterDuke: all of them with the html profiler? or maybe with sql?
oh, bedtime for me 23:24
Zoffix AlexDaniel`: that's not a problem at all though. 1.111111111111111111111 is not a Rat
AlexDaniel` m: say WHAT 1.111111111111111111111
camelia (Rat)
AlexDaniel` isn't it?
Zoffix It's not supposed to be one. Your denominator is over 64 bits
AlexDaniel` ahhh… that's what happens
Zoffix m: say <1.111111111111111111111> 23:25
camelia 1.111111111111111111111
Zoffix AlexDaniel`: actually there's already is a ticket, sorta. One about smartmatches with ranges. Should be still open
AlexDaniel` m: say 1.111111111111111111111.FatRat 23:26
camelia 1.111111111111111111111
AlexDaniel` now I'm not even sure if it's a bug or not 23:27
AlexDaniel` what's the right behavior? 23:27
Zoffix AlexDaniel`: deciding that is part of the fix :) TimToady suggested generating a RatStr in those cases (it knows how to behave like a FatRat) and I said I'll try that first. Which is how the bug landed into my private stash :) 23:28
s/suggested/said "maybe it should..."/ 23:29
AlexDaniel` ok, I see. Thanks!
HoboWithAShotgun oh hey zoff. i am going to release this tomorrow. can you check for red flags? i am subclassing your Color.pm 23:36
hastebin.com/uwokivufux.go
raschipi HoboWithAShotgun: Are those HTML color names? 23:42
HoboWithAShotgun yeah 23:44
Zoffix HoboWithAShotgun: no idea. I don't remember writing Color 23:47
raschipi Should indicate that in the name, I think. Color naming is a huge topic and the HTML version is the least useful of them.
Zoffix HoboWithAShotgun: you can bind to attributes: `submethod BUILD(:$!name, :$!pretty-name) {}` 23:48
MasterDuke timotimo: sql. but maybe it wasn't as many as i thought. i did it at each release tag from 2017.03 to 2017.09 23:57
HoboWithAShotgun raschipi: which palette do you consider more "useful"? 23:59
raschipi Well, other usually try to at least represent actual colors. HTML color names ensure they will look different everywhere, instead of looking the same.

Powered by Cro and the Raku® Programming Language.

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

Thank you!