Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2015-12-17

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:11 [ptc]_ joined #perl6
00:18 BenGoldberg joined #perl6
00:24 zjmarlow joined #perl6
00:27 b2gills m: [[-3,2,-5],[-1,0,-2],[3,-4,1]]
00:27 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/OLRf3yAM5a␤Unable to parse expression in bracketed infix; couldn't find final ']' ␤at /tmp/OLRf3yAM5a:1␤------> 3[[-7⏏3,2,-5],[-1,0,-2],[3,-4,1]]␤»
00:27 b2gills [[ -3,2,-5],[-1,0,-2],[3,-4,1]]
00:27 b2gills m: [[-3,2,-5],[-1,0,-2],[3,-4,1]]
00:27 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/yvs3rofZJK␤Unable to parse expression in bracketed infix; couldn't find final ']' ␤at /tmp/yvs3rofZJK:1␤------> 3[[-7⏏3,2,-5],[-1,0,-2],[3,-4,1]]␤»
00:27 b2gills m: [[ -3,2,-5],[-1,0,-2],[3,-4,1]]
00:27 camelia rakudo-moar 33ac49: ( no output )
00:31 ShimmerFairy .tell FROGGS I tried writing a slang to add a general metric postfix (needs to be more than just basic sub postfix:<>, due to ambiguity and speed issues), and I get "Cannot look up attributes in a type object" without, "Too many positionals passed; expected 1 argument but got 2" with any code after the 'use'.
00:31 yoleaux ShimmerFairy: I'll pass your message to FROGGS.
00:46 chatterman joined #perl6
00:53 zjmarlow joined #perl6
00:57 AndyDee joined #perl6
00:58 Sqirrel joined #perl6
01:03 cognominal m: my $i = 0;  my @a;  gather { loop { take $i++ }}; say @a.WHAT
01:03 camelia rakudo-moar 33ac49: OUTPUT«(timeout)»
01:03 cognominal .WHAT seems to try to reify @a
01:04 cognominal that seems wrong to me.
01:04 yeahnoob joined #perl6
01:06 gfldex m: my $i = 0;  my @a;  gather { loop { take $i++ }}; say 'here'; say @a.WHAT
01:06 camelia rakudo-moar 33ac49: OUTPUT«(timeout)»
01:07 gfldex m: my $i = 0;  my @a;  my $target := gather { loop { take $i++ }}; say 'here'; say @a.WHAT
01:07 camelia rakudo-moar 33ac49: OUTPUT«here␤(Array)␤»
01:07 gfldex cognominal: i dont think so ^^^
01:10 cognominal oops, I misttped
01:10 cognominal m: my $i = 0;  my @a :=  gather { loop { take $i++ }}; say @a.WHAT
01:10 camelia rakudo-moar 33ac49: OUTPUT«Type check failed in binding; expected Positional but got Seq␤  in block <unit> at /tmp/hmOyg2INU7:1␤␤»
01:11 cognominal m: my $i = 0;  my @a =  gather { loop { take $i++ }}; say @a.WHAT
01:11 cygx joined #perl6
01:11 camelia rakudo-moar 33ac49: OUTPUT«(timeout)»
01:11 cygx m: my $i = 0; my @a := (gather { loop { take $i++ } }).list; say @a.WHAT
01:11 camelia rakudo-moar 33ac49: OUTPUT«(List)␤»
01:12 cygx m: my $i = 0; my @a = (gather { loop { take $i++ } }).list; say "unreachable"
01:12 camelia rakudo-moar 33ac49: OUTPUT«(timeout)»
01:14 cognominal m: my @a = gather { loop { take $i++ }}; say @a # otherwise it works fine
01:14 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/uvGwgYZ4v2␤Variable '$i' is not declared␤at /tmp/uvGwgYZ4v2:1␤------> 3my @a = gather { loop { take 7⏏5$i++ }}; say @a # otherwise it works fin␤»
01:14 cognominal oops
01:14 cognominal m: my $i = 0; my @a = gather { loop { take $i++ }}; say @a # otherwise it works fine
01:15 camelia rakudo-moar 33ac49: OUTPUT«(timeout)»
01:15 cognominal hum
01:15 cognominal m: my $i = 0; say gather { loop { take $i++ }}; say @a
01:15 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/tROMmPPsJG␤Variable '@a' is not declared␤at /tmp/tROMmPPsJG:1␤------> 3; say gather { loop { take $i++ }}; say 7⏏5@a␤»
01:16 cognominal m: my $i = 0; say gather { loop { take $i++ }};
01:16 camelia rakudo-moar 33ac49: OUTPUT«(0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 8…»
01:17 cognominal m: my $i = 0; my @a = gather { loop { take $i++ }}; say @a[1000]
01:17 camelia rakudo-moar 33ac49: OUTPUT«(timeout)»
01:17 cognominal I probably miss something obvious
01:18 cygx cognominal: the assignment alredy tries to access an infinite number of elements
01:18 cygx m: [ gather { loop { take $i++ } } ]
01:18 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/0zy1e87MNd␤Variable '$i' is not declared␤at /tmp/0zy1e87MNd:1␤------> 3[ gather { loop { take 7⏏5$i++ } } ]␤»
01:19 cygx m: [ gather { loop { take $++ } } ]
01:19 cygx ^ same thing happens here
01:19 camelia rakudo-moar 33ac49: OUTPUT«(timeout)»
01:20 cognominal m: my $i = 0;  my @a :=  gather { loop { take $i++ }}; say @a[10]
01:20 camelia rakudo-moar 33ac49: OUTPUT«Type check failed in binding; expected Positional but got Seq␤  in block <unit> at /tmp/6NHzXUudng:1␤␤»
01:21 cognominal ok
01:21 cognominal m: my $i = 0;  my $a =  gather { loop { take $i++ }}; say $a[10]
01:21 camelia rakudo-moar 33ac49: OUTPUT«10␤»
01:21 cognominal got it.
01:22 cognominal too bad I can't use the @ sigil
01:22 [Coke] ZOMG.
01:23 cygx cognominal: http://irclog.perlgeek.de/perl6/2015-12-17#i_11730808
01:23 zjmarlow joined #perl6
01:24 Lucas_One joined #perl6
01:24 cognominal indeed, cygx++ but you need to cast to list
01:26 cognominal bad time, I can't think clearly anymore
01:26 Hotkeys I assume arrays can't be lazy
01:27 mrsolo joined #perl6
01:28 cognominal m: my @a := 1..*; say @a[100]
01:28 camelia rakudo-moar 33ac49: OUTPUT«101␤»
01:28 clkao joined #perl6
01:29 cognominal Hotkeys++, Arrays can't be lazy indeed, but @a holds any positional, not only an array.
01:29 Hotkeys ah
01:30 cognominal An array is the reification of something that gives a finite number of elements.
01:31 cognominal Hotkeys, but as you see, I am confused about the specifics
01:31 Hotkeys oh I see, Array is just the default type for @
01:32 cygx m: my @a := list(1, 2 ... *)
01:32 camelia rakudo-moar 33ac49: OUTPUT«Type check failed in binding; expected Positional but got Seq␤  in block <unit> at /tmp/d2mOOz9_wF:1␤␤»
01:32 cygx m: my @a := List(1, 2 ... *)
01:32 camelia rakudo-moar 33ac49: ( no output )
01:33 cygx m: my @a := List(1, 2 ... *); say @a[^2]
01:33 camelia rakudo-moar 33ac49: OUTPUT«((...) Nil)␤»
01:33 cygx ;)
01:33 Hotkeys m: my @a = <a b c>.List; say @a.WHAT
01:33 camelia rakudo-moar 33ac49: OUTPUT«(Array)␤»
01:34 cygx m: say Array ~~ List
01:34 camelia rakudo-moar 33ac49: OUTPUT«True␤»
01:35 cognominal m: say (my @a).WHAT
01:35 camelia rakudo-moar 33ac49: OUTPUT«(Array)␤»
01:36 cygx 'night o/
01:36 cognominal so why not List, that seems to me the more general type
01:40 ShimmerFairy cognominal: Lists aren't mutable, so they wouldn't make a good default considering what people will usually expect from @a variables.
01:54 zjmarlow joined #perl6
02:01 cognominal I agree. Being the least derived type is not good enough.
02:06 dalek roast: 21bf7a4 | hoelzro++ | S32-io/IO-Socket-Async-UDP.t:
02:06 dalek roast: Handle zero-length UDP datagrams
02:06 dalek roast:
02:06 dalek roast: libuv may signal us with zero-length datagrams on MoarVM using certain
02:06 dalek roast: OSes, so we need to handle this
02:06 dalek roast: review: https://github.com/perl6/roast/commit/21bf7a4996
02:09 Ben_Goldberg joined #perl6
02:12 kittenlips joined #perl6
02:14 molaf_ joined #perl6
02:16 adu joined #perl6
02:17 geekosaur why do people insist "on certain OSes"?!!
02:17 geekosaur a UDP stack that does not preserve packet boundaries even for 0-length packets is broken. Period.
02:17 geekosaur it's *every* OS that supports UDP.
02:21 kittenlips_ joined #perl6
02:23 zjmarlow joined #perl6
02:25 kittenlips_ left #perl6
02:25 FROGGS_ joined #perl6
02:26 kittenlips joined #perl6
02:27 xpen joined #perl6
02:31 skids joined #perl6
02:32 Zoffix joined #perl6
02:33 b2gills m: say List ~~ Positional # Actually @a accepts anything that does the Positional role
02:33 camelia rakudo-moar 33ac49: OUTPUT«True␤»
02:34 awwaiid geekosaur: maybe not osx -- https://github.com/nodejs/node-v0.x-archive/issues/8023
02:34 geekosaur I think I saw that bug reported in libuv itself, yes
02:34 awwaiid yeah
02:35 awwaiid *shrug*
02:35 geekosaur interesting someone tried to finger kqueue because OS X doesn't have kqueue; it uses Mach ports
02:36 awwaiid "Hooray! This one is fixed in El Capitan ;)" is the most recent note :)
02:36 geekosaur and fixed in 10.11 it seems
02:37 awwaiid El Capitan is, incidentally, also the name of my favorite pirate themed mexican food restaraunt
02:38 Ben_Goldberg m: { my $answer will leave { .say } = 42; say "What is 6 x 9" };
02:38 camelia rakudo-moar 33ac49: OUTPUT«What is 6 x 9␤42␤»
02:39 Zoffix wtf
02:39 Ben_Goldberg I wanted to create a socket, and did not want it to outlive the block in which it's created, would 'will leave { .close }' be the best way to do it?
02:40 Ben_Goldberg s/I wanted/If I wanted/
02:41 awwaiid this will leave is a new concept for me
02:41 kittenlips joined #perl6
02:42 timotimo "will" can attach a phaser to a variable or sub
02:45 awwaiid fascinating
02:45 kittenlips Hi everyone
02:45 TimToady howdy
02:45 timotimo good night everyone! i'm already way late for bed ...
02:45 Zoffix night
02:45 TimToady o/
02:45 BenGoldberg_ joined #perl6
02:45 Zoffix m: my $x will INIT; say [*] 1..4000; say now - $x now
02:45 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ewFIyTM8Vs␤Missing block␤at /tmp/ewFIyTM8Vs:1␤------> 3my $x will INIT7⏏5; say [*] 1..4000; say now - $x now␤    expecting any of:␤        block or pointy block␤»
02:45 Zoffix m: my $x will INIT { now }; say [*] 1..4000; say now - $x
02:45 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/O9PHgc2vBh␤Can't use unknown trait 'will INIT' in a variable declaration.␤at /tmp/O9PHgc2vBh:1␤------> 3my $x will INIT { now }7⏏5; say [*] 1..4000; say now - $x␤    expecting any of:␤        be…»
02:45 kittenlips Have a design question, why are :exists and :delete not method calls like e.g. %hash<key>.delete
02:45 Zoffix kittenlips, 'cause they would be methods on "nothing" should those keys not exist
02:45 Zoffix but as adverbs they work
02:45 TimToady so that they work on anything you can subscript, including multi-dimensional arrays and slices
02:45 kittenlips Ah ok, makes sense, thought maybe they would be a class method on the Hash class
02:45 TimToady otherwise we'd have to reinvent subscript syntax as arguments
02:46 BenGoldberg_ m: my $x will init { now }; say [*] 1..4000; say now - $x
02:46 camelia rakudo-moar 33ac49: OUTPUT«1828801951514065013314743175573919044217377710730439219706452695420895979797317736485037028687048410733644304156928557175467246186154355733394261561795699671674528483159731749881876093748280498041957651294872061055892812978809780062059342953770532674062445…»
02:46 Zoffix :/
02:46 BenGoldberg_ The trait version of a phaser is lowercase.  Dunno why.
02:46 Zoffix weird
02:46 TimToady I'm a little surprised it managed 4000!
02:47 BenGoldberg_ m: my $x will init { now }; my $y = [*] 1..4000; say now - $x
02:47 camelia rakudo-moar 33ac49: OUTPUT«Use of uninitialized value of type Any in numeric context  in block <unit> at /tmp/elMhzYuB1d:1␤Instant:1450320463.122963␤»
02:47 Zoffix Takes just 0.328533s according to earlier tests
02:47 BenGoldberg_ m: my $x will init { now }; say $x
02:47 camelia rakudo-moar 33ac49: OUTPUT«(Any)␤»
02:47 TimToady m: my $x will init { $_ = now }; say $x
02:47 camelia rakudo-moar 33ac49: OUTPUT«(Any)␤»
02:48 BenGoldberg_ m: my Instant $x will init { .new }; say $x
02:48 camelia rakudo-moar 33ac49: OUTPUT«(Instant)␤»
02:48 TimToady m: my $x will init { $x = now }; say $x
02:48 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/fo0wcZ8BUU␤Variable '$x' is not declared␤at /tmp/fo0wcZ8BUU:1␤------> 3my $x will init { 7⏏5$x = now }; say $x␤»
02:48 BenGoldberg_ I'm not sure 'will init' is really useful
02:48 TimToady m: my $x will init { .say }; say $x
02:48 camelia rakudo-moar 33ac49: OUTPUT«(Any)␤»
02:49 BenGoldberg_ m: { my $x will init { say 'foo' }; ++$x }
02:49 camelia rakudo-moar 33ac49: ( no output )
02:49 TimToady m: my $x will init { .=now }; say $x
02:49 camelia rakudo-moar 33ac49: OUTPUT«(Any)␤»
02:49 BenGoldberg_ It's not reached
02:49 Zoffix m: my $x will init { .say }; say "bazinga"\
02:49 camelia rakudo-moar 33ac49: OUTPUT«5===SORRY!5=== Error while compiling /tmp/51RarbrcYp␤Confused␤at /tmp/51RarbrcYp:1␤------> 3my $x will init { .say }; say "bazinga"\7⏏5<EOL>␤    expecting any of:␤        postfix␤»
02:49 Zoffix m: my $x will init { .say }; say "bazinga"
02:49 camelia rakudo-moar 33ac49: OUTPUT«bazinga␤»
02:49 TimToady m: my Instant $x will init { .=now }; say $x
02:49 camelia rakudo-moar 33ac49: OUTPUT«(Instant)␤»
02:49 TimToady hmm
02:49 kittenlips Could you give me example of adverb usage?
02:49 Zoffix kittenlips, %hash<key>:exists is one
02:49 BenGoldberg_ m: for 1..3 { my $x will leave { say 'foo' }; say 'bar'  }
02:49 camelia rakudo-moar 33ac49: OUTPUT«bar␤foo␤bar␤foo␤bar␤foo␤»
02:50 kittenlips Thank you
02:50 BenGoldberg_ m: for 1..3 { my $x will enter { say 'foo' }; say 'bar'  }
02:50 camelia rakudo-moar 33ac49: OUTPUT«foo␤bar␤foo␤bar␤foo␤bar␤»
02:50 TimToady dinner &
02:50 BenGoldberg_ m: my $x will enter { $_ = 42 }; say $x
02:50 camelia rakudo-moar 33ac49: OUTPUT«42␤»
02:51 kittenlips To all the Perl 6 team have to say you've made it amazingly quick and easy to get everything installed
02:52 kittenlips Easier than even Perl 5 with perlbrew and cpanm
02:54 zjmarlow joined #perl6
02:55 Zoffix :/
02:56 Zoffix perlbrew and cpanm is about the same TBH
02:56 PerlJam joined #perl6
03:03 kittenlips joined #perl6
03:03 awwaiid Zoffix: I keep wanting to seed docs from synopsis -- right now I am thinking of taking the Phasers docs from S04 into doc/Language/phasers.pod; what think you?
03:04 kittenlips Are there plans to incorporate support for numerical computing in Perl 6?
03:04 awwaiid kittenlips: Yes, I think that some of the "shaped" array syntax is to make it transparent even
03:05 kittenlips Does the shape array support provide the foundation for building a Perl 6 version of PDL or NumPy?
03:05 awwaiid I think
03:05 awwaiid mmm
03:05 awwaiid dunno
03:08 awwaiid Zoffix: also, is doc 'make html' supposed to work on nom?
03:09 BenGoldberg_ So, here's a silly question:  If you look at S04, and search for 'per-variable', you'll see an example of 'will enter', 'will undo' ... are there tests for these in roast?
03:09 awwaiid (getting a "Cannot unbox a type object" right after "Composing doc registry ..."
03:10 awwaiid )
03:11 kaare_ joined #perl6
03:12 stmuk https://github.com/perl6/doc/issues/241
03:13 cibs joined #perl6
03:14 stmuk maybe related to https://rt.perl.org/Ticket/Display.html?id=126927 (?)
03:15 stmuk republic++ # trout mask replica pilots </virtualsue>
03:16 awwaiid gfldex: In day 13 advent you link to http://192.168.0.2:3000/type/Capture.html
03:21 awwaiid ah, thanks stmuk
03:22 PotatoGim joined #perl6
03:24 zjmarlow joined #perl6
03:28 noganex_ joined #perl6
03:34 davercc joined #perl6
03:49 Actualeyes joined #perl6
03:54 zjmarlow joined #perl6
03:59 adu joined #perl6
04:20 Hotkeys I just discovered hyperwhatever
04:21 Hotkeys m: my $foo = ** x 3; say $foo("A".."F")
04:21 camelia rakudo-moar 33ac49: OUTPUT«(AAA BBB CCC DDD EEE FFF)␤»
04:24 zjmarlow joined #perl6
04:29 adu joined #perl6
04:32 dj_goku Hotkeys: haha that is interesting
04:43 dalek rakudo/nom: b6ba2b8 | TimToady++ | src/Perl6/Optimizer.nqp:
04:43 dalek rakudo/nom: remove debugging line
04:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b6ba2b84a0
04:43 dalek rakudo/nom: 2d98de9 | TimToady++ | src/Perl6/Optimizer.nqp:
04:43 dalek rakudo/nom: speed up if/while on native int
04:43 dalek rakudo/nom:
04:43 dalek rakudo/nom: Makes "loop (my int $i = 100000000; $i; $i--) {}" runs about 15x faster
04:43 dalek rakudo/nom: by turning $i into $i != 0.
04:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2d98de93ba
04:45 BenGoldberg_ joined #perl6
04:48 raiph joined #perl6
04:53 zjmarlow joined #perl6
04:55 llfourn do we have a module yet that does a bunch of introspection with .WHY and produces a README.md for a module?
05:05 ribasushi joined #perl6
05:12 dj_goku llfourn: after some googling some have suggested for built-ins it could go and fetch them and output that POD.
05:12 llfourn dj_goku: mmm but how?
05:13 dj_goku llfourn: build that something?
05:13 llfourn heh I'm trying :D
05:13 llfourn m: say CompUnit::RepositoryRegistry.candidates("Test")
05:13 camelia rakudo-moar 2d98de: OUTPUT«Earlier failure:␤ Dynamic variable @*INC not found␤  in block <unit> at /tmp/8gaZS_zXoN:1␤␤Final error:␤ Type check failed in binding $spec; expected Str but got Failure␤  in block <unit> at /tmp/8gaZS_zXoN:1␤␤»
05:14 llfourn the idea would be to get the compunit handle and introspect what it exports and then .WHY everything
05:14 llfourn (I think)
05:21 adu joined #perl6
05:24 zjmarlow joined #perl6
05:40 Sqirrel joined #perl6
05:46 sno joined #perl6
05:47 snarkyboojum joined #perl6
05:54 zjmarlow joined #perl6
06:02 protium joined #perl6
06:03 protium p6: say ‘Hello’
06:03 camelia rakudo-moar 2d98de: OUTPUT«Hello␤»
06:09 xfix joined #perl6
06:09 llfourn m: say "welcome to $*PERL, protium"
06:09 camelia rakudo-moar 2d98de: OUTPUT«welcome to Perl 6, protium␤»
06:09 protium heh, just checking it out
06:10 protium I have always liked perl 5
06:10 protium But haven’t really followed perl 6 until now
06:10 llfourn m: say Perl, Cool with any Version
06:10 camelia rakudo-moar 2d98de: OUTPUT«(Perl)(Cool)␤»
06:11 protium Agreed
06:11 protium But wow, it’s different
06:12 llfourn it sure is very different :)
06:12 protium Think they’ll make the Christmas release date?
06:12 llfourn but if you want to write p5 code in p6 it's not so different
06:12 llfourn I'm not sure.
06:14 Exodist joined #perl6
06:14 TimToady joined #perl6
06:14 ShimmerFairy joined #perl6
06:24 zjmarlow joined #perl6
06:43 diakopter llfourn: lol
06:46 llfourn m: say Perl, not .can("lol") with Version, v6.b
06:46 camelia rakudo-moar 2d98de: OUTPUT«(Perl)True␤»
06:54 zjmarlow joined #perl6
07:05 pierre-vigier joined #perl6
07:12 _nadim joined #perl6
07:18 CIAvash joined #perl6
07:19 _nadim Good morning.
07:22 _nadim any link to some documentation for 'handles' nothing in the doc page
07:24 zjmarlow joined #perl6
07:25 nwc10 joined #perl6
07:25 nwc10 If it's April 1st already, how come it's so dark outside? -- http://www.businesswire.com/news/home/20151216005376/en/Unicode-Launches-Adopt-a-Character-Campaign-Support-World%E2%80%99s-%E2%80%9CDigitally
07:25 nwc10 (that might be old news)
07:26 glaukommatos joined #perl6
07:28 CIAvash _nadim: Not sure it's useful to you, but 'handles' is mentioned here: https://perl6advent.wordpress.com/2009/12/11/day-11-classes-attributes-methods-and-more/
07:32 secwang joined #perl6
07:36 zjmarlow joined #perl6
07:38 _nadim it is :)
07:40 darutoko joined #perl6
07:40 CIAvash :)
07:41 firstdayonthejob joined #perl6
07:46 cognominal joined #perl6
07:50 dalek rakudo/nom: e800e38 | TimToady++ | src/Perl6/ (2 files):
07:50 dalek rakudo/nom: don't use optimization level for debug
07:50 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e800e38ed9
07:50 dalek rakudo/nom: 323a5c0 | TimToady++ | src/Perl6/Actions.nqp:
07:50 dalek rakudo/nom: ifnull was hiding &foo from 'useless use' sinking
07:50 dalek rakudo/nom:
07:50 dalek rakudo/nom: Fixes RT #121406
07:50 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/323a5c077e
07:50 dalek rakudo/nom: a779f7b | TimToady++ | src/Perl6/ (3 files):
07:50 dalek rakudo/nom: warn on useless adverbials
07:50 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a779f7b615
07:51 RabidGravy joined #perl6
07:54 FROGGS joined #perl6
07:54 rurban joined #perl6
07:56 andreoss joined #perl6
07:58 abraxxa joined #perl6
07:59 JimmyZ TimToady: no test for RT #121406?
08:04 TimToady feel free to add one
08:04 darutoko joined #perl6
08:05 kanishka joined #perl6
08:10 JimmyZ TimToady: I'd like to suggest tagging RT #121406 as testneed if there is no test yet but fixed :)
08:14 nine protium: yes the Christmas release date is still on
08:17 TimToady JimmyZ: okay, you can suggest that :)
08:17 dalek rakudo/nom: cf72c2d | TimToady++ | src/Perl6/ (2 files):
08:17 dalek rakudo/nom: also check non-adverbial pairs for useless use
08:17 dalek rakudo/nom:
08:17 dalek rakudo/nom: (They were hidden inside a Want that didn't propagate sink.)
08:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/cf72c2d38a
08:18 TimToady m: &say; :a(42)
08:18 camelia rakudo-moar a779f7: OUTPUT«WARNINGS for /tmp/YEzK6dfvA2:␤Useless use of ":a(42)" in sink context (line 1)␤Useless use of &say in sink context (line 1)␤»
08:19 TimToady m: "a" => 42
08:19 camelia rakudo-moar a779f7: ( no output )
08:19 TimToady that should warn in another 12 minutes or so
08:28 TimToady m: loop (my int $i = 100000000; $i; $i--) {}; say $i; say now - INIT now
08:28 camelia rakudo-moar a779f7: OUTPUT«0␤0.60463868␤»
08:28 ribasushi joined #perl6
08:29 nine TimToady: that's indeed a tiny bit faster than the 11 seconds on yesterday's version :)
08:29 sno joined #perl6
08:31 TimToady m: "a" => 42
08:31 camelia rakudo-moar cf72c2: OUTPUT«WARNINGS for /tmp/zHSjMwlXBc:␤Useless use of "=>" in expression "\"a\" => 42" in sink context (line 1)␤»
08:31 JimmyZ m: loop (my int $i = 10000000; $i; $i--) {}; say $i; say now - INIT now
08:31 camelia rakudo-moar cf72c2: OUTPUT«0␤0.0623971␤»
08:33 nine m: loop (my int $i = 1000000; $i; $i--) {}; say $i; say now - INIT now # even faster!!!
08:33 camelia rakudo-moar cf72c2: OUTPUT«0␤0.00849470␤»
08:34 FROGGS :P
08:34 yoleaux 00:31Z <ShimmerFairy> FROGGS: I tried writing a slang to add a general metric postfix (needs to be more than just basic sub postfix:<>, due to ambiguity and speed issues), and I get "Cannot look up attributes in a type object" without, "Too many positionals passed; expected 1 argument but got 2" with any code after the 'use'.
08:35 FROGGS .tell ShimmerFairy can I see the code?
08:35 yoleaux FROGGS: I'll pass your message to ShimmerFairy.
08:39 cpage joined #perl6
08:42 ShimmerFairy FROGGS: sure :)    https://gist.github.com/ShimmerFairy/30a3d44da2af9254eda1
08:42 yoleaux 08:35Z <FROGGS> ShimmerFairy: can I see the code?
08:42 ShimmerFairy FROGGS: seems to be centered on the bindkey lines at the end
08:44 ShimmerFairy FROGGS: also, that arg_flat_nok thing is just a leftover debug thing, if you couldn't tell
08:45 smls joined #perl6
08:46 espadrine joined #perl6
08:47 g4 joined #perl6
08:47 g4 joined #perl6
08:51 rindolf joined #perl6
08:54 smls m:  my @a = [2, 4], [6, 8];   my ([$a, $b], [$c, $d]) := @a;   say ($a, $b, $c, $d);
08:54 camelia rakudo-moar cf72c2: OUTPUT«((Mu) (Mu) (Mu) (Mu))␤»
08:54 smls Am I imagining that this used to work?
08:55 FROGGS ShimmerFairy: I'd suggest that you look at the last few lines of NativeCall.pm6
08:55 FROGGS ShimmerFairy: especially the bits about Mu $/ and also how to access matches
08:56 FROGGS ShimmerFairy: and you have to 'use nqp'
08:57 andreoss joined #perl6
08:57 ShimmerFairy I left out the use nqp snippet (just copy-pasted a commented-out section), but I do see now the bit where I have to remember to live in NQP land a bit :)
08:57 FROGGS yeah
08:58 FROGGS brb
09:01 FROGGS joined #perl6
09:05 dakkar joined #perl6
09:07 El_Che is using nqp::force_gc something that is OK to do in user code? I read huge files (secuentially through read and parse channels) and calling the gc as soon as possible seem handy to me
09:08 nine El_Che: I'd argue that it's better to not bother. Rakudo will garbage collect quite often anyway if you have lots of memory churn.
09:08 nwc10 left #perl6
09:08 El_Che I see, thx
09:10 PotatoGim joined #perl6
09:11 hankache joined #perl6
09:11 rindolf joined #perl6
09:12 zakharyas joined #perl6
09:12 Skarsnik joined #perl6
09:16 brrt joined #perl6
09:16 brrt \o
09:17 hankache hola
09:18 RabidGravy erp
09:19 brrt :-)
09:23 vytas are tthere plans to replace libuv ?
09:24 cxreg should there be?
09:25 Skarsnik Hello, someone should write something to get substring from json string ~~
09:25 DrForr Uh... Just so people here know...
09:25 cognominal vytas, jnthn once talked about it
09:25 DrForr Tutorial "Introduction to Perl 6" accepted for OSCON 2016.
09:25 RabidGravy :)
09:25 moritz vytas: I don't think so
09:25 moritz vytas: what should it be replaced with?
09:26 Skarsnik nice DrForr :)
09:26 DrForr Meep.
09:26 vytas moritz, well libev for *nix, but then windows...
09:27 cxreg libuv does use libev on unix, doesn't it?
09:27 vytas cxreg, it doesn't. though it used to ages ago
09:27 RabidGravy I think that's the point, libuv is designed to be cross platform
09:28 cxreg vytas: ok, guess my info is dated :)
09:28 moritz vytas: sounds like not much of a plan :/
09:29 vytas not really in position to suggest anything, I was just interested to see what's going on :)
09:29 RabidGravy there's always the option for someone port to a different VM of course
09:29 moritz vytas: afaict libuv is less worse than all the alternatives, so there are no plans or incentives to replace it
09:30 nine I think we should first simply try to upgrade libuv to a less ancient version. AFAIK that's on the table for some time after the release. It would be much too risky to do it now.
09:30 cxreg is libuv still primarily authored by the nodejs guys?
09:30 vytas moritz, wasn't there issue with sockets though? let me find bug report
09:30 koo8 joined #perl6
09:32 iH2O joined #perl6
09:32 vytas ignore me, can't find it now
09:33 * iH2O ignores vytas
09:34 moritz vytas: there being an issue with sockets doens't seem like a good reason to ditch a library :-)
09:36 PotatoGim joined #perl6
09:40 Ven joined #perl6
09:40 Ven \o
09:40 * Ven should really find the time to finish his blog post for tomorrow :-)
09:40 iH2O why is tomorrow that important?
09:41 Ven ..because that's my advent day :P
09:41 n0tjack joined #perl6
09:42 * Ven notices he didn't put his draft on the internet and frowns
09:42 Ven (people who fixed the install bug)>>++
09:51 iH2O left #perl6
09:54 au joined #perl6
09:56 Skarsnik *is annoyed by leak in the XML module* 873.01 Mb /root/.rakudobrew/moar-nom/install/bin/moar --execname=/root/.rakudobrew/bin/../moar-nom/install/bin/perl6 .... -I . Ferrel.p6
09:59 brrt are any of the LPW videos online yet?
10:00 lizmat good *, #perl5!
10:00 * lizmat is not awake yet, clearly
10:00 lizmat #perl6!
10:00 llfourn s/perl5/perl6/
10:00 lizmat brrt: not afaik
10:00 brrt ohahi lizmat
10:00 brrt ah, too bad
10:01 brrt what about slides?
10:01 Skarsnik hm, I managed to have moarvm stuck in a boucle just with use it seems x)
10:01 Hotkeys .u 五
10:01 yoleaux No characters found
10:01 Skarsnik *loop
10:05 dalek rakudo/nom: f609147 | lizmat++ | src/core/ (2 files):
10:05 dalek rakudo/nom: Make error on 123 +< (1 +< 64) little more awesome
10:05 dalek rakudo/nom:
10:05 dalek rakudo/nom: Also for +> .  Adds error class X::NYI::BigInt, perhaps also useful for other
10:05 dalek rakudo/nom: cases.  In response to RT #126941.
10:05 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f609147648
10:06 Hotkeys m: 六 - 1
10:06 camelia rakudo-moar cf72c2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/x0YHJnHghO␤Undeclared routine:␤    六 used at line 1␤␤»
10:06 Hotkeys Aw
10:06 dalek roast: 9dccebf | lizmat++ | S02-types/WHICH.t:
10:06 dalek roast: Add test for X::NYI::BigInt error class
10:06 dalek roast: review: https://github.com/perl6/roast/commit/9dccebf762
10:06 Hotkeys I was hoping it magically knew Chinese numerals like it does the strange Arabic ones
10:08 lizmat m: say uniprop("六")
10:08 camelia rakudo-moar cf72c2: OUTPUT«Lo␤»
10:08 DrForr This weekend I want to put together Roman numerals, but the thing is '0rXVII << 2' doesn't seem right, because ... Roman numerals don't have 0s.
10:08 brrt what strange arabic ones? /me thought the arabic ones were the ones we use
10:08 Ven lizmat: good #perl11 to you too :P
10:08 lizmat m: say uniprop("5")
10:08 camelia rakudo-moar cf72c2: OUTPUT«Nd␤»
10:09 Hotkeys Not like 1 2 3 etc
10:09 Hotkeys It's a different thing
10:09 lizmat Hotkeys: seems like 六 is not marked as being numeric ?
10:09 Hotkeys I forget what they're called
10:09 Hotkeys .u Arabic digit
10:09 yoleaux U+0660 ARABIC-INDIC DIGIT ZERO [Nd] (٠)
10:09 yoleaux U+0661 ARABIC-INDIC DIGIT ONE [Nd] (١)
10:09 yoleaux U+0662 ARABIC-INDIC DIGIT TWO [Nd] (٢)
10:09 Skarsnik when I have a case of class A { has B $} and class B { has B $}; I just need to stub one class in one pm file?
10:09 Hotkeys The Arabic indic digits
10:09 lizmat m: say uniprop("٢")
10:09 camelia rakudo-moar cf72c2: OUTPUT«Nd␤»
10:09 brrt aha
10:10 Hotkeys You can use them just like regular digits
10:10 moritz Skarsnik: just  class B { ... }; before class A
10:10 Hotkeys In perl 6
10:10 lizmat Hotkeys: because the have an N in the uniprop, not an L
10:10 Skarsnik ===SORRY!=== Error while compiling /root/piko/Discord/lib/Discord/Channel.pm6
10:10 Skarsnik The following packages were stubbed but not defined:
10:10 Skarsnik Discord::User
10:10 Skarsnik ~~
10:10 Hotkeys Ah
10:11 llfourn m: .say when 4 for '۰'..'۹'
10:11 camelia rakudo-moar cf72c2: OUTPUT«۴␤»
10:11 llfourn so cool :)
10:11 Hotkeys Ya
10:11 Skarsnik hexchat does not see your weird symbol :(
10:11 llfourn Skarsnik: :(
10:11 Hotkeys Weird
10:11 Hotkeys Must be your font
10:12 Skarsnik or windows x)
10:12 Hotkeys Nah
10:12 llfourn m: .say when 5 for '۰'..'۹'
10:12 camelia rakudo-moar f60914: OUTPUT«۵␤»
10:12 Hotkeys I was able to see them on windows hexchat
10:12 Hotkeys With consolas I think
10:13 nine m: say [+] '۰'..'۹'
10:13 camelia rakudo-moar f60914: OUTPUT«45␤»
10:14 llfourn m: say "WIN" if ۵۵ == 55
10:14 camelia rakudo-moar f60914: OUTPUT«WIN␤»
10:14 llfourn :D:D
10:14 llfourn m: say "WIN" if "۵۵" == 55
10:14 camelia rakudo-moar f60914: OUTPUT«WIN␤»
10:14 llfourn m: say "WIN" if "5۵" == 55
10:14 camelia rakudo-moar f60914: OUTPUT«WIN␤»
10:14 llfourn lol
10:14 Hotkeys Iol
10:14 nine m: say [+] ۰..۹ # because, why should we use strings there?
10:14 camelia rakudo-moar f60914: OUTPUT«45␤»
10:15 llfourn I find that sometimes my editor swaps the ۹ to the front... which is apparently the right way around according to wikipedia
10:15 llfourn but p6 still gets it
10:16 Hotkeys That's because it just displays it right to left on the front end
10:16 ShimmerFairy It's just a display thing; you could use LRO for example to force it to be left->right, but unfortunately P6 will die on those kinds of characters
10:16 Hotkeys But behind the scenes all is well
10:16 llfourn I see...
10:16 Hotkeys m: (۰..۹)».say
10:16 camelia rakudo-moar f60914: OUTPUT«0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤»
10:16 Hotkeys Oh
10:17 Hotkeys Oops
10:17 Hotkeys m: ("۰".."۹")».say
10:17 camelia rakudo-moar f60914: OUTPUT«۰␤۱␤۲␤۳␤۴␤۵␤۶␤۷␤۸␤۹␤»
10:17 lizmat Hotkeys: note that the order of the say could be random in the future
10:17 Hotkeys I know
10:17 Hotkeys It is just a bad habit
10:17 Hotkeys :p
10:18 Hotkeys Quick and dirty
10:18 lizmat m: say "۰".."۹"
10:18 camelia rakudo-moar f60914: OUTPUT«"۰".."۹"␤»
10:18 Hotkeys I avoid doing it in real code
10:18 lizmat m: say flat "۰".."۹"
10:18 camelia rakudo-moar f60914: OUTPUT«(۰ ۱ ۲ ۳ ۴ ۵ ۶ ۷ ۸ ۹)␤»
10:18 lizmat m: say |"۰".."۹"
10:18 camelia rakudo-moar f60914: OUTPUT«1.."۹"␤»
10:18 lizmat m: say |("۰".."۹")
10:18 camelia rakudo-moar f60914: OUTPUT«۰۱۲۳۴۵۶۷۸۹␤»
10:18 Hotkeys But I like hyper :p
10:18 llfourn | is flatter than flat..
10:18 El_Che if you people would add colours, this would remind me of mirc in the '90s
10:19 moritz m: say "۰"..."۹"
10:19 camelia rakudo-moar f60914: OUTPUT«(۰ ۱ ۲ ۳ ۴ ۵ ۶ ۷ ۸ ۹)␤»
10:19 moritz El_Che: somehow I doubt mirc in the '90s did Unicode so well :-)
10:20 El_Che moritz: it did it best to annoy users by weird coloured ascii combinations in the channel
10:20 * El_Che returns to pointing angrily at the clouds :)
10:20 vytas what's the difference between .. and ... ?
10:20 llfourn m: |("۰".."۹").WHAT.say; (flat "۰".."۹").WHAT.say
10:20 camelia rakudo-moar f60914: OUTPUT«(Range)␤(Seq)␤»
10:21 llfourn m: (|("۰".."۹")).WHAT.say; (flat "۰".."۹").WHAT.say
10:21 camelia rakudo-moar f60914: OUTPUT«(Slip)␤(Seq)␤»
10:21 llfourn so slips dispaly closer together :)
10:21 llfourn vytas: series operator
10:21 Skarsnik https://github.com/Skarsnik/perl6-discord this code make moar loop forever (and crash when no free memory) just on the use Discord in the example. (on my october rakudo/moar)
10:21 Hotkeys Also re: 0r prefix for Roman numbers, isn't/wasn't the 0 just to represent that it is a constant to the parser
10:22 moritz vytas: .. constructs a Range, which is just "from here to there"
10:22 Hotkeys hence thither
10:22 moritz vytas: ... is the series (lazy list) operator which supports all kinds of funky sequence generation
10:22 DrForr I could easily make it IrMMXV, but that (while more correct) would just look like a weirdly-named function or something.
10:22 Hotkeys hence .. thither
10:23 moritz m: say 1, 2, 4 ... 256
10:23 camelia rakudo-moar f60914: OUTPUT«(1 2 4 8 16 32 64 128 256)␤»
10:23 Hotkeys I think 0r would be best for consistency's sake
10:23 moritz it autodetected a geometric sequence here, for example
10:23 lichtkind joined #perl6
10:23 moritz m: say 1, 1, *+* ... *>1000
10:23 camelia rakudo-moar f60914: OUTPUT«(1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597)␤»
10:23 moritz fibonacci numbers!
10:24 DrForr Well, I'm also thinking of cuneiform with 0B :)
10:24 zjmarlow joined #perl6
10:24 Hotkeys Do it yo
10:24 Hotkeys B might be confusing, isn't 0b binary
10:25 DrForr Well, yes. I suppose 0c would be a little better...
10:26 Hotkeys Can we get unary
10:26 vytas moritz, thanks for explaining, I must admit my brains stopped for a second when I saw Slip, Range, Seq :)
10:26 Hotkeys eg 0u11111 = 5
10:26 jnthn vytas: from backlog: we'll almost certainly stop doing sync I/O in Moar through libuv; for all the async stuff, doesn't expect any change for some years. There's just no point to re-do all that hard work for different platforms async I/O differences.
10:26 DrForr I've got the weekend :)
10:26 espadrine joined #perl6
10:27 Woodi hi #perl6
10:27 vytas Thanks jnthn :)
10:27 DrForr At the moment I'm starting to recover from getting the tutorial accepted.
10:28 Woodi jnthn: I looked into uv_ lastly and there is no aio_ or io_ calls... so, is it async what uv_ is doing ?
10:29 dalek nqp: 74d2a80 | moritz++ | tools/build/MOAR_REVISION:
10:29 dalek nqp: Bump MOAR_REVISION to get UDP-on-linux fix by hoelzro++
10:29 dalek nqp: review: https://github.com/perl6/nqp/commit/74d2a80413
10:30 timotimo o/
10:30 dalek rakudo/nom: b422f44 | moritz++ | tools/build/NQP_REVISION:
10:30 dalek rakudo/nom: Bump NQP version to get UPD-on-Linux fix
10:30 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b422f44e67
10:31 vividsnow joined #perl6
10:34 jnthn moritz: "What keeps the supply alive" - supplies are push-based, so it's actually the event loop that keeps them alive. Going out of scope doesn't do a thing, since the chain is held from the top; you have to explicitly .close the tap to free resources (but the test just lets process exit do it).
10:34 _nadim nay unicode character one can use to show that a line being wrapped? ASCII "glyph" wuld be useful too
10:34 DrForr There should be an arrow for that.
10:35 nine m: say "hello"; say "world";
10:35 camelia rakudo-moar f60914: OUTPUT«hello␤world␤»
10:35 Begi joined #perl6
10:35 jnthn Woodi: Yes, decidedly async
10:35 nine _nadim: ^^^
10:35 Skarsnik I am stuck with my class definition :( etheir it make rakudo loop forever or it does not want my stubbed class ~~
10:35 nine err....is there a NL in there? My terminal doesn't show anything
10:35 DrForr Yes, there is.
10:35 moritz jnthn: ah, ok
10:36 DrForr hello>N/L<world
10:36 moritz .u ␤
10:36 yoleaux U+2424 SYMBOL FOR NEWLINE [So] (␤)
10:36 _nadim it shows, but >N/L< is a mouthfull in ascii. I need to provide both
10:36 timotimo that's not really for line-wrap, though? that's for explicit newlines?
10:36 moritz typically the browser is better at displaying Unicode than the terminal, so check http://irclog.perlgeek.de/perl6/today in case of doubt :-)
10:37 Begi2 joined #perl6
10:37 _nadim bleah, my browser does not show it, google chrome
10:40 |Tux| since panda fails big-time in rakudo, are there any alternatives to install modules under rakudobrew?
10:40 |Tux| or will panda be fixed shortly
10:40 Peter_R joined #perl6
10:41 moritz |Tux|: no idea if it works at the moment, but rakudobrew also supports building zef
10:41 CIAvash joined #perl6
10:41 |Tux| Initializing
10:41 |Tux| ===SORRY!=== Error while compiling /pro/3gl/CPAN/rakudobrew/moar-nom/zef/lib/Zef/Distribution/Local.pm6
10:41 |Tux| Undeclared name:
10:41 |Tux| PARSE-INCLUDE-SPEC used at line 123
10:42 |Tux| :( currently that also barfs
10:42 n0tjack joined #perl6
10:42 timotimo oh, it relies on old rakudo internals, eh?
10:42 lizmat PARSE-INCLUDE-SPEC  was an internal thing that I removed, if I remember correctly
10:44 rurban_ joined #perl6
10:44 |Tux| so rakudowbrew currently has neither, as both fail
10:44 timotimo :(
10:45 lizmat fwiw, at the current state of curli , I would no longer recommend using zef anyway
10:46 nine |Tux|: what's the issue with panda anyway?
10:46 TEttinger what is zef?
10:46 moritz TEttinger: an alternative module installer
10:46 DrForr . o ( Zero Extraction Force? )
10:46 nine lizmat: I thought you wrote PARSE-INCLUDE-SPEC? It's now a private sub in CompUnit::RepositoryRegistry as all the $*REPO setup code is in there
10:47 jnthn lizmat: I'd not be surprised if zef bounces back once the comp unit stuff settles down :)
10:49 |Tux| nine: https://gist.github.com/Tux/42ddea88b3f52f45ff43
10:52 nine |Tux|: I assume your rakudo is very up to date?
10:52 lizmat jnthn: as long as zef uses $*REPO.install   that should be ok, I think?   nine ?
10:53 |Tux| is 15 minutes up-to-date enough?
10:53 |Tux| This is Rakudo version 2015.11-601-gb422f44 built on MoarVM version 2015.11-60-g9fd64bd
10:53 |Tux| implementing Perl 6.b.
10:53 stmuk |Tux|: you could try ufo and make as a panda replacement
10:54 zjmarlow joined #perl6
10:54 |Tux| ?
10:54 nine lizmat: almost. It should call .install on a CompUnit::Installable repo that reports .can-install
10:54 |Tux| never heared of ufo? what should I do?
10:54 Skarsnik hm, I am confused. I have: use DC; use DU; use DG; DC need DU class so I stub it in DC (DU need DC); DG need DC. But it complain that DU is stubbed in DC
10:54 stmuk https://github.com/masak/ufo generates a makefile from the meta file
10:54 stmuk I've found it useful for module installs when panda is broken for a bit
10:55 nine |Tux|: yes, recent enough
10:55 moritz Skarsnik: wherever you stub a class, you also need to at least load its definition afterwards
10:55 moritz Skarsnik: and you can only have cyclic deps between classes in the same file
10:55 moritz (at least compile-time deps)
10:56 nine |Tux|: can your panda install other modules, like File::Temp?
10:57 stmuk when I tried yesterday panda was unable to find Test.pm
10:57 Skarsnik moritz, I don't get it. I use the class just after, it should work?
10:57 nine Then please let's fix the panda issue instead of suggesting alternative installers
10:58 nine stmuk: do you use rakudobrew, too?
10:58 |Tux| nine, https://gist.github.com/Tux/1b7f08d40791e78a4962
10:58 stmuk nine: yes
10:59 stmuk I'm trying totally from scratch now
10:59 nine |Tux|: maybe the issue is with Slang::Tuxic then
11:00 |Tux| without that I cannot test :(
11:00 |Tux| FWIW CSV::Parser also fails to install with panda
11:01 |Tux| Could not find Test in …
11:03 pierre-vigier joined #perl6
11:04 stmuk the can't find Test error is relatively new (maybe 2/3 days old) compared with the other panda/rakudo errors
11:05 nine nine@sunshine:~/install/rakudo (nom=)> perl6 -Ilib -e 'use lib "lib"; say $*REPO.repo-chain'
11:05 nine (file#/home/nine/install/rakudo/lib)
11:05 nine That's the underlying issue ^^^
11:05 nine It's just because I fail repeatedly at implementing a first grader's example, i.e. a dead simple linked list
11:06 |Tux| $ ls -l ~/.perl6
11:06 |Tux| lrwxrwxrwx 1 merijn users 24 Nov 30 09:30 /home/merijn/.perl6 -> /pro/3gl/CPAN/rakudobrew
11:07 stmuk nine++ # repeatedly succeeding too
11:07 nine In other words: if you passed second grade and operations on a linked list are no challenge for you, you're already better than the one who implemented curli. That's how simple most of this code is!
11:07 nine So please join :)
11:07 dalek rakudo/nom: 8d50dab | lizmat++ | src/core/IO/Path.pm:
11:07 dalek rakudo/nom: Fix for #126935
11:07 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8d50dabfa9
11:08 zjmarlow joined #perl6
11:10 woolfy joined #perl6
11:10 woolfy left #perl6
11:12 vividsnow joined #perl6
11:15 Zoffix awwaiid, I don't know anything about the doc stuff
11:17 * lizmat confirms that UDP tests are now passing on OS X
11:17 Skarsnik nic
11:19 cygx_ joined #perl6
11:19 cygx_ o/
11:19 timotimo yo cygx_
11:20 cygx_ lizmat: you implemented some Cool methods in IO::Path (split, comb, ...)
11:20 cygx_ do you think they should fail or hrow if .open fails?
11:20 cygx_ *throw
11:20 dalek roast: e41c661 | lizmat++ | S32-io/io-path.t:
11:20 dalek roast: Add test for #126935
11:20 dalek roast: review: https://github.com/perl6/roast/commit/e41c661785
11:20 lizmat cygx: good question
11:20 lizmat I have no real opinion about that atm
11:21 cygx_ right now, they will return handles failures, which is wrong
11:21 cygx_ I have a pull request that makes them fail, but I wonder if throwing would be more appropriate
11:21 cygx_ (thinking about binding their return values)
11:22 cygx_ *return handled failures
11:23 lizmat yes, that part is maybe not so good...
11:23 lizmat will think about it over brunch&
11:24 cygx_ bon appétit
11:25 andreoss joined #perl6
11:30 llfourn m: { berp => "lerp", ( with "herp" { $_ => "derp" } ) }
11:30 camelia rakudo-moar 8d50da: OUTPUT«When invoking cuid_1_1450351810.18985 '', provided outer frame 0x31d0630 (cuid_3_1450351810.18985 '<unit>') does not match expected static frame 0x31d0760 (cuid_2_1450351810.18985 '')␤  in block <unit> at /tmp/MDAsAMUnCb:1␤␤»
11:30 llfourn o.o I will RT
11:34 moritz it's surprising how many corner cases a language can have :-)
11:35 dalek nqp/sized-natives: 792fa61 | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
11:35 dalek nqp/sized-natives: Map objprimbits and objprimunsigned ops.
11:35 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/792fa61dbe
11:35 dalek nqp/sized-natives: 36c6701 | jnthn++ | src/ (3 files):
11:35 dalek nqp/sized-natives: Support sized native types in NQP MOP.
11:35 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/36c6701c2f
11:36 kid51 joined #perl6
11:36 ollej joined #perl6
11:38 lucasb joined #perl6
11:38 Zoffix m: say ''''''''foo''''''''
11:38 camelia rakudo-moar 8d50da: OUTPUT«5===SORRY!5=== Error while compiling /tmp/eqrfrk0px5␤Two terms in a row␤at /tmp/eqrfrk0px5:1␤------> 3say ''7⏏5''''''foo''''''''␤    expecting any of:␤        infix␤        infix stopper␤        postfix␤        statement end␤   …»
11:38 Zoffix Damn... I had a dream last night this was working in Perl 6 and I was bragging that no other language can do it :)
11:39 jnthn m: say Q''''''''foo''''''''
11:39 camelia rakudo-moar 8d50da: OUTPUT«5===SORRY!5=== Error while compiling /tmp/JbvkAtoe1W␤Two terms in a row␤at /tmp/JbvkAtoe1W:1␤------> 3say Q''7⏏5''''''foo''''''''␤    expecting any of:␤        infix␤        infix stopper␤        postfix␤        statement end␤  …»
11:39 jnthn aww :)
11:39 jnthn You're likely tinking of the matching open/close though
11:39 jnthn m: say q{ say { 'omg' } }
11:39 camelia rakudo-moar 8d50da: OUTPUT« say { 'omg' } ␤»
11:40 Zoffix I think this what gave my brain the idea:
11:40 Zoffix m: say 'fo/o' ~~ /'/'/
11:40 camelia rakudo-moar 8d50da: OUTPUT«「/」␤»
11:40 Zoffix the delimiter appears inside, but still doesn't break the regex
11:40 jnthn Sure, we parse regexes as a real language
11:40 jnthn Same for strings fwiw
11:41 PotatoGim joined #perl6
11:41 Zoffix m: say q{{{{{{{foo}}}}}}}
11:41 camelia rakudo-moar 8d50da: OUTPUT«foo␤»
11:41 Zoffix Well, that's interesting
11:42 jnthn m: my $meth = 'flip'; my $foo = 'desserts'; say "$foo backwards is $foo."$meth"()"
11:42 camelia rakudo-moar 8d50da: OUTPUT«desserts backwards is stressed␤»
11:42 nine Can I check if a named argument was passed in at all, i.e. even when it's passed an undefined value?
11:42 jnthn nine: |c then c.hash<foo>:exists
11:42 moritz nine: you can, if you don't declare it specifically, but with a slurp *%named
11:43 moritz *slurpy
11:43 * arnsholt goes to review NativeCall/JVM
11:43 n0tjack joined #perl6
11:44 arnsholt Prospective employer wants me to present something I've worked on, and that's probably the most interesting stuff I've worked on
11:44 dalek nqp/sized-natives: 017aff3 | jnthn++ | src/vm/moar/stage0/ (11 files):
11:44 dalek nqp/sized-natives: Update bootstrap.
11:44 dalek nqp/sized-natives:
11:44 dalek nqp/sized-natives: Otherwise we can't actually declare sized natives in the NQP setting.
11:44 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/017aff34f6
11:44 dalek nqp/sized-natives: 028062f | jnthn++ | src/core/NativeTypes.nqp:
11:44 dalek nqp/sized-natives: Add sized native types.
11:44 dalek nqp/sized-natives:
11:44 dalek nqp/sized-natives: Only on Moar for now, though that's fine since I only plan to use them
11:44 dalek nqp/sized-natives: in fixing Moar code-gen for the time being.
11:44 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/028062fc43
11:47 lucasb m: /o/
11:47 camelia rakudo-moar 8d50da: OUTPUT«getlexrelcaller needs a context␤  in block <unit> at /tmp/w4BrjD5iUB:1␤␤»
11:53 koo8 joined #perl6
11:53 virtualsue joined #perl6
11:59 virtualsue joined #perl6
11:59 PotatoGim joined #perl6
12:01 dalek nqp/sized-natives: d1a15a1 | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
12:01 dalek nqp/sized-natives: Make QAST -> MAST respect size differences.
12:01 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/d1a15a150f
12:13 grondilu m: sub infix:<choose>($n, $k) { ([*] ($n ... 0) Z/ 1 .. min($n - $k, $k)).Int };  say 1e9 choose (1e9 - 3);
12:13 camelia rakudo-moar 8d50da: OUTPUT«166666666166666665274638336␤»
12:19 timotimo http://thonny.cs.ut.ee/ - this looks cute
12:20 PotatoGim joined #perl6
12:24 moritz it also nearly spells "cute" in the domain name :-)
12:27 administrator joined #perl6
12:28 andreoss masak: consider my PR
12:33 Begi2 Hello ! What's the best doc to start concurrency with Perl 6 ?
12:33 Skarsnik the advent post? x)
12:34 llfourn Begi2: https://perl6advent.wordpress.com/2015/12/14/day-14-a-nice-supplies-syntactic-relief-for-working-with-asynchronous-data/
12:35 jnthn http://doc.perl6.org/language/concurrency is probably good to read before that post :)
12:35 Begi2 Thanks llfourn & jnthn !
12:35 smls Begi2: Also, some of the PDFs here: http://jnthn.net/articles.shtml  (the newer ones at least)
12:35 llfourn yes watch this: https://www.youtube.com/watch?v=JpqnNCx7wVY jnthn++
12:36 moritz jnthn: maybe link to that form the advent post? (unless you already do, of course :-)
12:36 gfldex jnthn: is the following a good way to change the default scheduler?
12:36 moritz *from
12:36 gfldex m: PROCESS::<$SCHEDULER> = ThreadPoolScheduler.new(initial_threads => 0, max_threads => %*ENV<THREADS>.?Int // 2);
12:36 camelia rakudo-moar 8d50da: ( no output )
12:36 moritz gfldex: I'd just do  my $*SCHEDULER = ...
12:36 jnthn gfldex: I prefer to declare a my $*SCHEDULER = ...
12:37 moritz gfldex: at the top of the program; that way everything is in the dynamic scope in which this variable is visble
12:37 rindolf joined #perl6
12:37 smls timotimo: Yes, that's neat.
12:38 dalek nqp/sized-natives: 6dc2a6f | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
12:38 dalek nqp/sized-natives: num64/num32 truncate/extend coercion code-gen.
12:38 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/6dc2a6f831
12:38 dalek nqp/sized-natives: 1fbfd04 | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
12:38 dalek nqp/sized-natives: Use correct sized lex/local ref ops.
12:38 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/1fbfd04e91
12:38 dalek tablets: 08ad96d | (Herbert Breunung)++ | docs/tablet-9-regexes.txt:
12:38 dalek tablets: -typo, -unecessary file endings
12:38 dalek tablets: review: https://github.com/perl6/tablets/commit/08ad96d285
12:39 jnthn Aside from the fact they break a bunch of other things, those patches fix RT #124084 :)
12:40 ollej left #perl6
12:40 gfldex m: $*SCHEDULER =  ThreadPoolScheduler.new(initial_threads => 0, max_threads => %*ENV<THREADS>.?Int // 2);
12:40 camelia rakudo-moar 8d50da: OUTPUT«Cannot modify an immutable ThreadPoolScheduler␤  in block <unit> at /tmp/9gmMbymSP9:1␤␤»
12:40 gfldex i think i like my version better :->
12:40 andreoss masak: i take it back, some tests fail
12:41 jnthn gfldex: You missed the "my"
12:43 _nadim is there a way to run the tests in parallel?
12:43 jnthn lunch &
12:44 n0tjack joined #perl6
12:44 vytas joined #perl6
12:50 gfldex jnthn: that seams not to work in a module. And i have to do it in the module because of perl6 --doc=BigPage
12:51 nine _nadim: TEST_JOBS=5
12:54 kaare_ joined #perl6
12:54 moritz ... which is even mentioned in the INSTALL.txt document in rakudo
12:55 _nadim nine++
13:00 advwp joined #perl6
13:01 sufrostico joined #perl6
13:06 dalek rakudo/nom: 300bddc | grondilu++ | README.md:
13:06 dalek rakudo/nom: mention rlwrap in README
13:06 dalek rakudo/nom:
13:06 dalek rakudo/nom: Linenoise is slow, we should at least mention an alternative.
13:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/300bddcf62
13:06 dalek rakudo/nom: cfaa921 | lizmat++ | README.md:
13:06 dalek rakudo/nom: Merge pull request #636 from grondilu/patch-10
13:06 dalek rakudo/nom:
13:06 dalek rakudo/nom: mention rlwrap in README
13:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/cfaa921370
13:10 dalek modules.perl6.org: 4096ab1 | (Zoffix Znet)++ | lib/ModulesPerl6/DbBuilder/Dist/Source/GitHub.pm:
13:10 dalek modules.perl6.org: Revert "Anchor GitHub source regex so it does not incorrectly match weird URLs"
13:10 dalek modules.perl6.org:
13:10 dalek modules.perl6.org: This reverts commit ae432076a9571fbb4c47bae37b5452c268ad90be as it breaks the entire build. I'll fix the original issue more correctly later today.
13:10 dalek modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/4096ab1e5b
13:12 ZoffixW joined #perl6
13:13 ZoffixW grondilu, would you mind submitting your version and simply closing my PR? RE https://github.com/rakudo/rakudo/pull/635
13:13 smls Is it possible that there has never been an advent post about Inline::Perl5 and Inline::Python?   https://www.google.com/search?q=inline++site:https://perl6advent.wordpress.com
13:13 smls nine ^^
13:14 ZoffixW Would be very nice to write one about Inline::Python
13:14 nine smls: seems like :/
13:14 smls A shame
13:14 nowan joined #perl6
13:14 virtualsue joined #perl6
13:16 ZoffixW m: say (1..6) Z+ (1..6).reverse
13:16 camelia rakudo-moar cfaa92: OUTPUT«(7 7 7 7 7 7)␤»
13:16 ZoffixW m: say (1..7) ZZZZZZ (1..7).reverse
13:16 camelia rakudo-moar cfaa92: OUTPUT«(((((((1 7)))))) ((((((2 6)))))) ((((((3 5)))))) ((((((4 4)))))) ((((((5 3)))))) ((((((6 2)))))) ((((((7 1)))))))␤»
13:17 ZoffixW hehe, neat
13:21 sufrosti1o joined #perl6
13:21 Skarsnik my rakudo is very slow to start something weird
13:21 lichtkind m: @a = 1..5; say @a[2]
13:21 camelia rakudo-moar cfaa92: OUTPUT«5===SORRY!5=== Error while compiling /tmp/PQ3CcJ3wNs␤Variable '@a' is not declared␤at /tmp/PQ3CcJ3wNs:1␤------> 3<BOL>7⏏5@a = 1..5; say @a[2]␤»
13:21 lichtkind m: my @a = 1..5; say @a[2]
13:21 camelia rakudo-moar cfaa92: OUTPUT«3␤»
13:21 lichtkind m: my @a = 1..5; say @a/   [2]
13:21 camelia rakudo-moar cfaa92: OUTPUT«5␤»
13:22 lichtkind syn 2 says it should be 3 too
13:22 lichtkind any idea what happened?
13:22 andreoss m: (1 Z 2).Str.say
13:22 camelia rakudo-moar cfaa92: OUTPUT«1 2␤»
13:22 andreoss m: (1 ZZ 2).Str.say
13:22 camelia rakudo-moar cfaa92: OUTPUT«This Seq has already been iterated, and its values consumed␤(you might solve this by addding .cache on usages of the Seq, or␤by assigning the Seq into an array)␤  in block <unit> at /tmp/CrYzcN6ppQ:1␤␤»
13:24 zjmarlow joined #perl6
13:24 andreoss m: (1 ZZ 2).say
13:24 camelia rakudo-moar cfaa92: OUTPUT«(((1 2)))␤»
13:25 andreoss m: (1 ZZ 2).cache.Str.say
13:25 camelia rakudo-moar cfaa92: OUTPUT«This Seq has already been iterated, and its values consumed␤(you might solve this by addding .cache on usages of the Seq, or␤by assigning the Seq into an array)␤  in block <unit> at /tmp/_PlUjRAsFD:1␤␤»
13:26 sufrostico joined #perl6
13:26 PotatoGim joined #perl6
13:26 grondilu ZoffixW: I don't have rights to do the merge
13:27 grondilu oh sorry you meant submitting an other PR
13:28 rurban joined #perl6
13:28 grondilu I suggest you just edit your PR
13:29 lichtkind nine is this a bug  or intended?
13:31 moritz lichtkind: aren't you confusing \ with / ?
13:31 ZoffixW m: my @a = 1..5; say @a\   [2]
13:31 camelia rakudo-moar cfaa92: OUTPUT«3␤»
13:31 moritz say @a/   [2] does a division
13:31 sufrostico joined #perl6
13:32 moritz (number elements in @a) / (number of elements in [2])
13:32 nine it's actually 5/1 which is 5. Nice obfuscation :)
13:32 ZoffixW m: say [3]/[2]
13:32 camelia rakudo-moar cfaa92: OUTPUT«1␤»
13:33 moritz the so-called lichtkind manouver
13:35 PotatoGim_ joined #perl6
13:35 lichtkind thats not i wanted be famous for
13:36 lichtkind but your right i looked it up and the misspelled :)
13:38 Skarsnik m: say Any ~~ Int;
13:38 camelia rakudo-moar cfaa92: OUTPUT«False␤»
13:45 n0tjack joined #perl6
13:46 sammers joined #perl6
13:48 andreoss where does method .get_value come from?
13:49 Skarsnik In what context? x)
13:50 andreoss it's used in 007
13:50 andreoss lib/_007/Q.pm
13:50 andreoss 47:        sub avalue($attr) { $attr.get_value(self).?quoted-Str }
13:50 andreoss
13:52 jnthn Attribute
13:54 zjmarlow joined #perl6
13:54 Skarsnik it the read accessor?
13:54 jnthn Well, directly reads the attribute, doesn't go via the accessor
13:58 Skarsnik hm, is there a way to say class B is rw(for class A) ?
13:58 Begi joined #perl6
14:03 moritz no
14:03 moritz there is the "trusts" trait
14:03 gfldex Skarsnik: http://design.perl6.org/S12.html#Trusts
14:03 moritz but it only allows you to call private methods
14:03 moritz so you need to manually write private setters that the trusted class can call for you... yuck.
14:04 adu joined #perl6
14:04 jnthn In short, take this as a hint from Perl 6 to reconsider your design.
14:04 Skarsnik a bit... not that useful so
14:05 Skarsnik I don't know. having some class in a module you want to modify freely inside your module make sense for me
14:09 xfix joined #perl6
14:18 dalek nqp/sized-natives: 2884c8a | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
14:18 dalek nqp/sized-natives: Fill out sized num box/unbox handling.
14:18 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/2884c8a651
14:19 thundergnat joined #perl6
14:19 dalek rakudo/nom: 54457ad | lizmat++ | src/core/Str.pm:
14:19 dalek rakudo/nom: Fix some issues wrt :global/:g and extras
14:19 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/54457ad57b
14:20 thundergnat Is it a known issue that Xop and Zop are broken for multi dim arrays? I couldn't find anything on RT about it.
14:21 thundergnat m: my @a = ^10; my @b = [10..20]; @a = @a X* 2; say @a; @b[0] = @b[0] X* 2; say @b;
14:21 camelia rakudo-moar cfaa92: OUTPUT«[0 2 4 6 8 10 12 14 16 18]␤Method 'push' not found for invocant of class 'Mu'␤  in any  at /home/camelia/rakudo-m-inst-1/share/perl6/runtime/CORE.setting.moarvm:1␤  in sub THROW at /home/camelia/rakudo-m-inst-1/share/perl6/runtime/CORE.setting.moarvm…»
14:21 lizmat thundergnat: they should be NYI, afaik
14:21 lizmat at least before 6.c
14:23 dalek rakudo/nom: 898da16 | (Stefan Seifert)++ | src/core/CompUnit/RepositoryRegistry.pm:
14:23 dalek rakudo/nom: Fix $*REPO chain broken by use lib
14:23 dalek rakudo/nom:
14:23 dalek rakudo/nom: If a "use lib" statement occurs before $*REPO is initialized and the
14:23 dalek rakudo/nom: repository would already be in the $*REPO chain, it would never get its
14:23 dalek rakudo/nom: .next-repo attribute set. This is caused by repository-for-spec caching
14:23 dalek rakudo/nom: the created repositories, so only the first access determined the
14:23 dalek rakudo/nom: next-repo. The "use lib" implementation doesn't pass a :next-repo, so it
14:23 dalek rakudo/nom: was always undefined. OTOH if we just blindly overwrite .next-repo, we
14:23 dalek rakudo/nom: may create loops in the chain.
14:23 dalek rakudo/nom:
14:23 dalek rakudo/nom: Fixed doing the full "cut out of the repo-chain and re-insert properly"
14:23 dalek rakudo/nom: dance that use-repository does, even when setting up the initial chain.
14:23 dalek rakudo/nom:
14:23 dalek rakudo/nom: Fixes lots of "Cannot find Test in..." errors.
14:23 dalek rakudo/nom: Thanks to Tux++ for the report!
14:23 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/898da16765
14:23 nine |Tux|: ^^^^
14:23 thundergnat lizmat Ah. thanks.
14:23 * |Tux| runs to try, hold on
14:23 nine stmuk: ^^^
14:24 zjmarlow joined #perl6
14:24 Skarsnik hm, is there an operator to append a Str if the value at the right of ~= is defined?
14:24 Skarsnik +only
14:25 lizmat m: use experimental :cached
14:25 camelia rakudo-moar cfaa92: OUTPUT«WARNINGS for /tmp/_OrGnrQ3le:␤Useless use of ":cached" in sink context (line 1)␤»
14:25 lizmat jnthn:  ^^^ is that to be expected ?
14:26 raiph joined #perl6
14:26 jnthn TimToady: sink is leaking ^^
14:26 moritz lizmat: I think you have to ask TimToady
14:26 lizmat ah, ok
14:26 jnthn (it's wrong, but I'm delegating :))
14:26 lizmat I guess that's also responsible for all of the linenumbers being off
14:27 lizmat well, maybe not all, but quite a few
14:27 diakopter how could it affect the line numbers?
14:28 moritz I'd guess the sink context analysis does some AST wrapping, which uses :node information
14:28 moritz s/uses/loses/
14:32 smls moritz: Is htmlify known to be busted on nom?
14:33 Skarsnik it's still busted? :(
14:33 moritz smls: yes (unless somebody fixed it since yesterday)
14:34 smls I get a huge amount of "Use of uninitialized value of type Str in string context" warnings, and then "Cannot unbox a type object" error.
14:34 smls moritz: What rakudfo version does it work with?
14:34 moritz smls: possibly none :(
14:34 smls ok
14:34 diakopter aw, modulecounts.com went negative
14:35 Skarsnik lol what
14:35 diakopter p6 ecosystem
14:35 dalek rakudo/nom: 7188229 | lizmat++ | src/core/ (2 files):
14:35 dalek rakudo/nom: Fix for RT #126936
14:35 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7188229344
14:39 dalek nqp/sized-natives: 28b6851 | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
14:39 dalek nqp/sized-natives: Various sized int box/unbox/coercion cases.
14:39 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/28b68515b1
14:40 smls Heh. Running htmlify.p6 through the debugger aborts with "Unhandled exception: ctxlexpad needs an MVMContext" ... :D
14:44 skids joined #perl6
14:47 n0tjack joined #perl6
14:47 inokenty joined #perl6
14:48 _nadim joined #perl6
14:50 lizmat nine: is it still so that Inline::Perl5 doesn't build on OS X ?
14:51 RabidGravy joined #perl6
14:51 nine lizmat: worse :( It's not even installable on Linux.
14:51 nine Will have a look on it in the evening
14:51 lizmat ok  :-(   I was just trying it again in a long time...
14:52 lizmat Text::CSV doesn't install either because of test failures
14:52 lizmat :-(
14:53 Skarsnik hm, I think DateTime does not follow ISO 8601 complety according to https://en.wikipedia.org/wiki/ISO_8601#Times 2015-12-11T20:41:10.562000+00:00 is valid but it get rejected
14:53 |Tux| lizmat, looking, but $work interrupted
14:53 Skarsnik Should I fill a RT?
14:53 lizmat in my case, $flu interrupted  :-(
14:54 zjmarlow joined #perl6
14:54 nine lizmat: note that having a local clone and doing panda install . actually works
14:54 diakopter flu--
14:57 chansen_ Skarsnik: It's a valid ISO 8601 representation and equivalent to Z (zulu). ISO only states that negative zero offsets (-00:00) is not permitted
14:57 shibly joined #perl6
14:57 shibly Is perl6 stable and released?
14:58 tadzik give it another week or two :P
14:58 tadzik it's coming out this christmas
14:58 timotimo the perl6 language specification is about to become a whole lot stabler; after that, a lot more care will be put into backwards compatibility and such
14:58 Skarsnik chansen_, yes but DateTime does not want the extra precision on second and reject this string as valid ISO 8601 string
14:59 masak ahoj, #perl6
14:59 smls Hi
14:59 shibly https://perl6.org/
15:00 moritz \o/ people use the SSL version of the website
15:00 * moritz is happy he put the effort into the ssl certs
15:00 * diakopter too # moritz++
15:00 |Tux| lizmat, nine, still running into Cannot find method 'run_alt'
15:01 lizmat even if you try it again immediately again?
15:01 lizmat afaik, the "run_alt" is a symptom of on the fly precomp, and then trying to run the precomped
15:01 lizmat the second time, there is already a precomp, so no on the fly precompe
15:02 shibly for(i=0;i<5;i++) ... , is it valid for loop syntax in perl 6?
15:02 diakopter if you replace for with loop
15:02 timotimo and put a space before the opening parenthesis
15:02 shibly http://perl6intro.com/#_for
15:03 shibly for @array -> $array-item { , is it the only syntax as for loop ?
15:03 dalek rakudo/nom: 1b846f1 | (Stefan Seifert)++ | t/harness:
15:03 dalek rakudo/nom: Fix Inline::Perl5 detection in the test harness
15:03 dalek rakudo/nom:
15:03 dalek rakudo/nom: require Inline::Perl5; returns an undefined value, so we cannot re-use
15:03 dalek rakudo/nom: it as value for exit()
15:03 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1b846f1929
15:03 ibnz joined #perl6
15:03 |Tux| rm -rf .precomp && p6 t/10_base.t ; p6 t/10_base.t
15:04 chansen_ Skarsnik: I would consider that as a bug
15:04 jnthn shibly: Yes, for loops are only for iterating over things in Perl 6. But note you could write your example as just for ^5 { ... }
15:04 molaf joined #perl6
15:05 nine lizmat: unfortuantly not. The run_alt failure can also appear when loading an already precompiled file. My guess is that it could have happened even before curli. We're just much more agressive with regards to precompiling
15:05 lizmat ah,  :-(
15:06 ellybelly joined #perl6
15:08 shibly left #perl6
15:09 cdg joined #perl6
15:10 dalek nqp/sized-natives: 063d3be | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
15:10 dalek nqp/sized-natives: Always receive parameters at full width.
15:10 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/063d3be6d7
15:13 gtodd joined #perl6
15:16 dalek rakudo/nom: 0e24208 | lizmat++ | src/core/Int.pm:
15:16 dalek rakudo/nom: Temporary fix for #126942
15:16 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0e24208abf
15:18 jnthn lizmat: The int candidate part of that should be reverted
15:18 lizmat why ?
15:18 lizmat the int version doesn't do it right either
15:18 jnthn Because TimToady and I seem to have settled on native behavior
15:18 jnthn For native ints
15:20 jnthn All native int ops should really be calls straight down to the VM, which will in turn be JITted to something equally low level
15:20 dalek rakudo/nom: 63d34fc | lizmat++ | src/core/Int.pm:
15:20 dalek rakudo/nom: http://irclog.perlgeek.de/perl6/2015-12-17#i_11733657
15:20 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/63d34fc21e
15:20 jnthn Thanks
15:20 lizmat but if it is currently incorrect ?
15:20 jnthn Why is it currently incorrect?
15:21 jnthn We're leaning towards defining it on int as "what the CPU does"
15:21 pyrimidine joined #perl6
15:22 lizmat m: m: my int $a = -123; my int $b = 128; say $a +> $b
15:22 camelia rakudo-moar 1b846f: OUTPUT«-123␤»
15:22 lizmat m: m: my int $a = -123; my int $b = 127; say $a +> $b
15:22 camelia rakudo-moar 1b846f: OUTPUT«-1␤»
15:22 lizmat m: m: my int $a = -123; my int $b = 129; say $a +> $b
15:22 camelia rakudo-moar 1b846f: OUTPUT«-62␤»
15:23 jnthn If you write a C program doing that, you'll get just the same answers.
15:23 lizmat the shift factor is mod 64
15:23 lizmat ok
15:23 hankache joined #perl6
15:23 jnthn We have Int as the default in Perl 6, where we can put the safety valves etc.
15:24 lizmat BTW, is there a reason why int literals don't use the native int candidate ?
15:24 zjmarlow joined #perl6
15:24 Juerd What do int literals look like?
15:24 lizmat m: 123
15:24 camelia rakudo-moar 1b846f: OUTPUT«WARNINGS for /tmp/cUyV9pado4:␤Useless use of constant integer 123 in sink context (line 1)␤»
15:24 dalek doc: b6c0999 | smls++ | lib/Perl6/Documentable/Registry.pm:
15:24 dalek doc: Make Perl6::Documentable::Registry.lookup more robust
15:24 dalek doc:
15:24 dalek doc: With this, `./htmlify.p6 --sparse=1000` works again for me.
15:24 dalek doc: Full run is still busted.
15:24 dalek doc: review: https://github.com/perl6/doc/commit/b6c0999bb8
15:24 Juerd Those are Int, right?
15:25 grondilu Zoffix: I created PR as suggested.  https://github.com/rakudo/rakudo/pull/637/files
15:25 moritz Juerd: int literals are allomorphic; the can be Int or int
15:25 Juerd m: 123.WHAT.say
15:25 camelia rakudo-moar 1b846f: OUTPUT«(Int)␤»
15:25 jnthn lizmat: Yes, because you'd always get wrong results if the result doesn't fit on a native int
15:25 lizmat m: my int $a = 42; say $a.WHAT   # you can't tell
15:25 camelia rakudo-moar 1b846f: OUTPUT«(Int)␤»
15:25 Juerd Oh
15:25 Juerd Hmm
15:25 Juerd That's weird
15:25 andreoss masak: what attributes-order suppose to do in _007?
15:25 n0tjack joined #perl6
15:26 jnthn An int is allomorphic, but will only show its native side if there is at least one declared native involved in the expression also
15:26 jnthn $a-native-int + 1 # the 1 will be treated as native
15:26 dalek nqp/sized-natives: e9dcf6b | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
15:26 dalek nqp/sized-natives: Always pass arguments at full width.
15:26 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/e9dcf6b38f
15:27 jnthn uh, s/An int/An integer literal/ above
15:28 _nadim joined #perl6
15:29 masak andreoss: (a) impose a natural order in Q node stringification (the one I got from .^attributes wasn't always the one that felt most natural)
15:30 dalek roast: 83057c9 | lizmat++ | S03-operators/numeric-shift.t:
15:30 dalek roast: Add test for #126942
15:30 dalek roast: review: https://github.com/perl6/roast/commit/83057c9b8f
15:30 masak andreoss: (b) selectively hide attributes that are more "internal" and shouldn't be initialized from 007 userland. Q::Block's %.static-lexpad is an example.
15:31 lucs m: my $func = 'elems'; say <1 2 3>.$func  # Can something like this be made to work?
15:31 camelia rakudo-moar 63d34f: OUTPUT«Cannot find method 'CALL-ME'␤  in block <unit> at /tmp/JFRz8GTmLz:1␤␤»
15:31 jnthn quote it
15:31 masak andreoss: reason (b) also affects equivalence comparison in infix:<==> between Q nodes. i.e. %.static-lexpad does not enter into such a comparison.
15:32 lucs jnthn: Not sure what to quote :/
15:32 FROGGS m: my $func = 'elems'; say <1 2 3>."$func"()
15:32 camelia rakudo-moar 63d34f: OUTPUT«3␤»
15:32 FROGGS lucs: ^^
15:32 masak andreoss: you can see all three of these usages by doing a `git grep \\.attributes`
15:32 lucs Aha, thanks
15:32 andreoss masak: i added attribute if Q::Statement::If, but it's not available when .Str is caleed
15:33 masak andreoss: I don't understand; could you rephrase?
15:33 andreoss *called
15:33 moritz masak: or if you're not eager to type backslashes, 'git grep -F .attributes'
15:33 masak moritz: oh, nice.
15:33 masak moritz: I actually had it as `git grep '\.attributes'`
15:33 andreoss masak: i added attribute to Q::Statement::If, but it's not available when .Str is called . Is it hidden somehow?
15:33 moritz masak: (traditional grep has the same option)
15:36 regreg joined #perl6
15:37 andreoss .Str on Q, which produces AST dump. So now all tests which use if-statement failing, i believe because of the attribute $.else was not introduced properly somewhere
15:40 dalek nqp/sized-natives: 7b32c87 | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
15:40 dalek nqp/sized-natives: Return values at full width.
15:40 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/7b32c87e5d
15:43 khw joined #perl6
15:45 nine masak: I just realized: Perl 5 people have often described Perl 6 as a language research project. 007 is a research project for a macro system for the language research language Perl 6. That kina makes 007 a meta-research language in two ways :)
15:46 FROGGS such meta
15:47 jnthn very reserch
15:47 jnthn so spelling
15:47 FROGGS :P
15:48 gfldex could 'Internal error: invalid thread ID 29492288 in GC work pass' indicate a locking problem?
15:48 jnthn No, that's a memory corruption problem
15:50 jnthn Phew, I can pass make test again
15:51 gfldex who is General Failure and why is he corrupting my memory? o.0
15:52 diakopter Husband of Gamma Ray
15:53 Skarsnik .tell RabidGravy I have to fix some stuff on JSON-Unmarshal to handle null field ("key" : null) and number passed as string ("key" : "42") with the attr declared as Int. do you want me to make a PR?
15:53 yoleaux Skarsnik: I'll pass your message to RabidGravy.
15:53 Skarsnik .botsnack
15:53 yoleaux :D
15:53 stmuk w00t! lots of apparent fixage in the last couple of hours!
15:54 zjmarlow joined #perl6
15:54 jnthn m/sg camelia m: use Test; sub i16_o(int16 $i) { isnt $i, 32770, 'sub with int8 arg will see it truncated'; }; i16_o(32770);
15:54 jnthn oops
15:55 masak andreoss: yes, unless you also add it to .attribute-order, it's not visible in .Str, and not possible to pass as a property when instantiating from inside of 007 itself.
15:55 masak andreoss: aha, you're adding $.else -- cool!
15:56 masak andreoss: I think the trick is to make sure $.else is always initialized to either another Q::Statement::If, or to a Val::None instance.
15:56 masak andreoss: that's how we do it with Q::Statement::My and Q::Statement::Return, which see for inspiration.
15:56 timotimo stmuk: i know of no fixage :o
15:56 dalek nqp/sized-natives: bd55852 | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
15:56 dalek nqp/sized-natives: Attributte binding is full-width.
15:56 dalek nqp/sized-natives: review: https://github.com/perl6/nqp/commit/bd55852241
15:56 japhb .
15:57 stmuk the Test.pm related panda fail and some htmlify of doc progress I mean
15:57 masak nine: I would find the "meta-research project" thing more funny if it didn't mean buying into the slightly mean and IMO inaccurate view that Perl 6 is a research project (as opposed to a usable language that has a place next to Perl 5)
15:57 timotimo to me, perl 5 is now only a language used to provide modules for perl 6 to use via Inline::Perl5 :P
15:58 Skarsnik huhu
15:59 nine masak: no one claimed that a research project may not yield something immediately useable
16:00 * Skarsnik used LAST in a for loop for something
16:00 Skarsnik This thing is handy ^^
16:01 raiph p6: say "षि".encode # seems to disagree with http://demo.icu-project.org/icu-bin/nbrowser?t=&amp;s=0937+093f&amp;uv=0
16:01 camelia rakudo-moar 63d34f: OUTPUT«utf8:0x<e0 a4 b7 e0 a4 bf>␤»
16:01 eyck joined #perl6
16:01 psy_ joined #perl6
16:02 jnthn m: say "षि".encode.decode eq "षि"
16:02 camelia rakudo-moar 63d34f: OUTPUT«True␤»
16:02 eyck joined #perl6
16:03 jnthn Well, if we're doing something wrong we've a bug in both encoding and decoding...
16:03 raiph I believe that's supposed to be NFC. The ICU tool has all of NFC, NFKC, NFD, NFKD having the same values. None are like the Rakudo values.
16:03 jnthn huh?
16:04 jnthn But that page is showing you codepoint values
16:04 jnthn Not how it encodes to UTF-8
16:04 jnthn m: say "षि".NFC
16:04 camelia rakudo-moar 63d34f: OUTPUT«NFC:0x<0937 093f>␤»
16:04 jnthn So, wrong level :)
16:05 jnthn m: say "षि"."$_"() for <NFC NFD NFKC NFKD>
16:05 camelia rakudo-moar 63d34f: OUTPUT«NFC:0x<0937 093f>␤NFD:0x<0937 093f>␤NFKC:0x<0937 093f>␤NFKD:0x<0937 093f>␤»
16:05 jnthn Those match
16:05 diakopter slick
16:05 raiph So I'm *still* confused by Unicode. I'm so glad P6 is going to simplify this stuff!
16:06 raiph jnthn++ # brilliance and patience is a potent combo
16:06 jnthn raiph: I tend to think of it as 3 levels: graphemes, codepoints, bytes
16:06 jnthn where encodings are about bytes <-> codepoints, and NFK?C/NFK?D are about codepoints <-> codepoints
16:07 jnthn And graphemes are a level about that, consisting of one or more codepoints
16:09 dalek rakudo/nom: 467582d | lizmat++ | src/core/IO/Path.pm:
16:09 dalek rakudo/nom: Some more #126935 fixage
16:09 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/467582dbc9
16:09 raiph jnthn: Thanks. I thought I understood that. But my mind is a mysterious thing...
16:09 dalek nqp: 792fa61 | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
16:09 dalek nqp: Map objprimbits and objprimunsigned ops.
16:09 dalek nqp: review: https://github.com/perl6/nqp/commit/792fa61dbe
16:09 dalek nqp: 36c6701 | jnthn++ | src/ (3 files):
16:09 dalek nqp: Support sized native types in NQP MOP.
16:10 jnthn 15 commits, dalek. That was 15 commits. It's a merge! :P
16:10 jnthn How much do I have to do in a branch to his the heuristic :P
16:10 diakopter heheh
16:10 dalek joined #perl6
16:11 * jnthn retests Rakudo with that + latest
16:11 jnthn It fixes a bunch of misbehavior around sized native types as lexicals.
16:12 jnthn As in, actually makes them behave sized :)
16:14 JimmyZ jnthn++ # another hard work done
16:15 jnthn Hopefully
16:15 jnthn The nativecall tests and spectest looks good with this
16:15 jnthn Well, spectest still running on latest merges of stuff
16:15 diakopter yayay
16:16 jnthn ecosystem fallout is relatively unlikely, though a number of things have become a bunch stricter
16:16 Skarsnik "unlikely" = "Will break thing in a weird way"
16:16 jnthn In summary, Moar failed to properly validate lexical/register ops that invovled a typevar.
16:16 abraxxa how often is https://doc.perl6.org built and deployed?
16:16 jnthn *involved
16:17 Ven joined #perl6
16:17 Skarsnik it's not up to date since 5 days I think
16:17 abraxxa 7
16:17 Skarsnik there is an issue in the update script x)
16:17 jnthn Which hid a lot of code-gen issues
16:17 stmuk the doc process is broken by recent changes .. someone did commit one fix today
16:18 RabidGravy Skarsnik, JSON::Unmarshal is tadzik's module so yeah I'd go with a PR :)
16:18 yoleaux 15:53Z <Skarsnik> RabidGravy: I have to fix some stuff on JSON-Unmarshal to handle null field ("key" : null) and number passed as string ("key" : "42") with the attr declared as Int. do you want me to make a PR?
16:18 Skarsnik Oh right I mixed the two xD
16:19 andreoss masak: thanks, making $.else to be Val::None by default solves the problem with stringification
16:19 xpen joined #perl6
16:19 RabidGravy I thought, however that it actually dealt with those cases, but I may be mistaken
16:20 Skarsnik RabidGravy, you can't commit into it? *too lazy to create a fork for like 3 lines patch*
16:21 Skarsnik hm or can we create a PR from a .diff file?
16:23 masak andreoss: for the longest time it was enough for Q nodes to contain Any when things were empty.
16:23 stmuk if you use the github "hub" wrapper around git forking and doing PRs is about 3 one liners
16:23 masak andreoss: but since a while back, Q nodes need to work as normal 007 objects, and then they can only contain Val objects or other Q objects.
16:23 TimToady good native morning
16:24 zjmarlow joined #perl6
16:24 JimmyZ good morning TimToady
16:24 jnthn o/ TimToady
16:24 abraxxa Hi!
16:26 dalek rakudo/nom: ec18f24 | jnthn++ | tools/build/NQP_REVISION:
16:26 dalek rakudo/nom: Bump for sized natives improvements.
16:26 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ec18f24d27
16:26 dalek roast: 9fddf87 | jnthn++ | S02-types/native.t:
16:26 dalek roast: Unfudge test for RT #124084.
16:26 dalek roast: review: https://github.com/perl6/roast/commit/9fddf872fa
16:26 dalek roast: 53a062f | jnthn++ | S02-types/native.t:
16:26 dalek roast: Some extra sized natives tests.
16:26 dalek roast:
16:26 dalek roast: Many of which failed before recent Rakudo fixes.
16:26 dalek roast: review: https://github.com/perl6/roast/commit/53a062f3e0
16:27 jnthn Down to 8 xmas RTs
16:28 JimmyZ \o/
16:29 abraxxa jnthn: wow, it was 20 last time i've read that from you
16:29 TimToady jnthn: I also looked at #125769 and #126005 yesterday, but didn't figger them out
16:30 TimToady the one does not appear to be related to sink context, at least as far as the compiler is concerned, but could still be related down at a semantic level somehow
16:30 RabidGravy jnthn, does that fix the regression that required https://github.com/jonathanstowe/Audio-Encode-LameMP3/commit/44f0c4c911845938d49bd13b7812e6a8d928d034
16:31 jnthn TimToady: The problem with #125769 is that we sink an if statement's result rather than propagating need to sink into the blocks of the if statement
16:31 TimToady it might be the difference between lexical an dlocal
16:31 jnthn RabidGravy: Looks unrelated to me
16:32 RabidGravy :-\
16:32 jnthn TimToady: But if you --optimize=off you don't get lexical -> local stuff going on, so I'm not sure it's that...
16:32 TimToady good point
16:32 jnthn (And it's still wrong with that off)
16:32 domidumont joined #perl6
16:33 TimToady I can bang on #125769 some more, but my reduced test case doesn't even use 'if'
16:33 TimToady maybe this is two bugs
16:33 jnthn Maybe
16:33 jnthn TimToady: On an "easier" one, https://rt.perl.org/Ticket/Display.html?id=114946
16:34 TimToady m: { my $f := Failure.new("bar"); }
16:34 camelia rakudo-moar ec18f2: OUTPUT«bar␤  in block <unit> at /tmp/_vhkxGImKT:1␤␤Actually thrown at:␤  in block <unit> at /tmp/_vhkxGImKT:1␤␤»
16:34 TimToady see, no if
16:34 jnthn Yeah, that's the same issue
16:34 jnthn Or another manifestation of it
16:34 TimToady that is in sink context, and it does propagate
16:35 smls m: my %foo := :{ a => 1, b => 2 };   sub f (*%args) { dd %args };   f |%foo;
16:35 camelia rakudo-moar ec18f2: OUTPUT«Hash $var = ${"Str|a" => 1, "Str|b" => 2}␤»
16:35 smls ^^ is this a known bug?
16:35 TimToady m: { my $f = Failure.new("bar"); }
16:35 camelia rakudo-moar ec18f2: ( no output )
16:35 smls happens with typed hashed which are declared with  :{  }
16:35 smls this is what causes our htmlify.p6 troubles.
16:35 nine I think, after concurrency and naming things, the hardest problem in software engineering is picking the right music to code to..
16:35 smls moritz ^^
16:36 jnthn TimToady: My expectation on #114946 is we declare the curernt seamntics as what we want and maybe path the design docs if we well like doing so :)
16:36 lizmat https://www.youtube.com/watch?v=__-lpvXfI6E   # an oldy, but goody
16:36 lizmat nine: ^^^
16:37 jnthn TimToady: If it *does* propagate, then maybe we should not be emitting sinks on if ops that are at statement list level.
16:37 nine lizmat: yeah, then let's get going :)
16:38 timotimo nine: https://ottsonic.bandcamp.com/album/skylon - i quite like this, but right now i'm not sure whether it's good for coding to ...
16:40 dalek mu: 2a06169 | (Zoffix Znet)++ | misc/SHELTER/ (3 files):
16:40 dalek mu: Move SHELTER from ecosystem to mu
16:40 dalek mu:
16:40 dalek mu: See https://github.com/perl6/modules.perl6.org/issues/49#issuecomment-165381505
16:40 dalek mu: review: https://github.com/perl6/mu/commit/2a061690d2
16:41 dalek ecosystem: 7ec46ab | (Zoffix Znet)++ | / (3 files):
16:41 dalek ecosystem: Move SHELTER from ecosystem to mu
16:41 dalek ecosystem:
16:41 dalek ecosystem: See https://github.com/perl6/modules.perl6.org/issues/49#issuecomment-165381505
16:41 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/7ec46ab6ec
16:42 virtualsue joined #perl6
16:45 dalek mu: c2cc3b1 | (Zoffix Znet)++ | misc/SHELTER/README.md:
16:45 dalek mu: Use a better link for IRC chat
16:45 dalek mu: review: https://github.com/perl6/mu/commit/c2cc3b16cb
16:46 dalek rakudo/nom: 09983ae | grondilu++ | src/core/List.pm:
16:46 dalek rakudo/nom: add count-only method to combinations Iterator
16:46 dalek rakudo/nom:
16:46 dalek rakudo/nom: as suggested on IRC.  Replaces https://github.com/rakudo/rakudo/pull/635
16:46 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/09983ae1a9
16:46 dalek rakudo/nom: e0647c6 | lizmat++ | src/core/List.pm:
16:46 dalek rakudo/nom: Merge pull request #637 from grondilu/patch-11
16:46 dalek rakudo/nom:
16:46 dalek rakudo/nom: add count-only method to combinations Iterator
16:46 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e0647c67a0
16:47 dalek ecosystem: 808e996 | (Sylvain Colinet)++ | META.list:
16:47 dalek ecosystem: Added Discord - Client for the Discord API (hey, something perl5 does not have)
16:47 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/808e9967a8
16:47 jdv79 ugexe: how does it look to install from an non-p6c auth with zef?
16:47 yoleaux 16 Dec 2015 14:57Z <ZoffixW> jdv79: what is the status of MetaCPAN and what were the objections you had in mind for making it the "standard" Perl 6 module ecosystem instead of having two ecosystems? There were a discussion on the ecosystem stuff: http://irclog.perlgeek.de/perl6/2015-12-16#i_11727140
16:47 * flussence makes another attempt to figure out the right voodoo for installing-to-system-dirs
16:48 nine Perl 6 is sooo nice for typical shell one liners:
16:48 nine git log --format=%ae | perl6 -e 'lines().BagHash.pairs.sort(*.value)>>.say;'
16:48 dalek ecosystem: fb22993 | sergot++ | META.list:
16:48 dalek ecosystem: add bamboo, the dependency manager (bundler)
16:48 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/fb22993ac3
16:49 Skarsnik need 20 more modules for 500 x-mas? x)
16:49 lizmat nine: nice, but >>.say might throw the order away again
16:50 lizmat git log --format=%ae | perl6 -e '.say for lines().BagHash.pairs.sort(*.value);
16:50 nine ok ok :) git log --format=%ae | perl6 -e 'lines().BagHash.pairs.sort(*.value).map(*.say);'
16:51 jnthn Is the .pairs actually needed?
16:51 nine It's not!
16:51 jdv79 Zoffix: what objections?
16:52 jdv79 Zoffix: i'll try to read that backlog by tomorrow at hte latest.  got a little busy the last few days but i'll be back on this full force tomorrow.
16:53 dalek rakudo/nom: 12ba341 | lizmat++ | src/core/IO/Path.pm:
16:53 dalek rakudo/nom: Fix () vs <> thinko, zefram++
16:53 dalek rakudo/nom:
16:53 dalek rakudo/nom: There does not seem to be a spectest to catch this  :-(
16:53 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/12ba3410a1
16:53 stmuk m: say Str.can(Num)
16:53 camelia rakudo-moar e0647c: OUTPUT«Use of uninitialized value of type Num in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed.  in any can at gen/moar/m-Metamodel.nqp:1099␤Use of uninitialized value of type Num in string context␤Any of .^na…»
16:54 zjmarlow joined #perl6
16:54 autarch joined #perl6
16:56 skids m: my $f = * or 1; $f.WHAT.say # There was an RT for this which I commented on, but cannot find now.  Struck me as semantic enough to maybe make a call on pre-xmas (whether or/and/etc should curry, or in general review which ops should/not)
16:56 camelia rakudo-moar e0647c: OUTPUT«(Whatever)␤»
16:56 skids Maybe the call has already been made.
16:56 domidumont joined #perl6
16:56 prammer joined #perl6
16:56 JimmyZ nine: https://gist.github.com/zhuomingliang/f1b7a6bd502c47f5ab49
16:57 nine JimmyZ: is that with today's rakudo?
16:57 JimmyZ nine: I just type `make`
16:57 jnthn skids: We've generally followed the principle that short-circuits ones don't
16:57 JimmyZ nine: yes, newest
16:57 jdv79 lizmat: did you come up with the auth structure in S22?
16:57 lizmat jdv79: yes, 2+ years ago
16:57 dalek ecosystem: e77c5b0 | (Zoffix Znet)++ | README.md:
16:57 dalek ecosystem: Update common errors
16:57 dalek ecosystem:
16:57 dalek ecosystem: We now normalize source URLs, so that part is not needed. Panda now cares greatly about correct `provides` section, so mention that
16:57 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/e77c5b01d0
16:58 lizmat jdv79: atm, I'm not sure how that matches the current implementation
16:58 jdv79 its seems a bit off where the "storage loc" is equated to an uri scheme
16:58 lizmat well, it was very handwavy at that point
16:58 * nine is happy to answer any questions about the current implementation
16:58 jdv79 but what if i want to have a new one, jdvpan or something, now i have to getan installer to hardcode in my details?
16:58 lizmat jdv79: ^^^
16:59 jdv79 perhaps it should be more like the "host" in a uri where it can be a resolved name or an ip or sommething
17:00 vytas in perl6 how do u print to STDERR, or how do you warn without associated location information
17:00 dalek ecosystem: 0dd6111 | (Zoffix Znet)++ | README.md:
17:00 dalek ecosystem: Actually link to `provides` spec, in case reader doesn"t know about it
17:00 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/0dd6111048
17:00 * flussence has a weak opinion that the module names in META's provides/depends sections should all support the full syntax with auth/ver -- right now they definitely don't
17:02 nine vytas: note
17:02 vytas thanks nine++
17:02 nine JimmyZ: please delete your install/share/perl6. From then on the underlying issue should be fixed forever
17:02 ZoffixW joined #perl6
17:03 xpen joined #perl6
17:03 ZoffixW jdv79, well, not objections, but I recall I said when do you think P6 MetaCPAN will be on modules.perl6.org and you said that won't happen and we'll have two ecosystems
17:03 TimToady skids: as jnthn indicated, || and && and such are really control flow operators, turned rather rapidly into 'if' and 'unless'
17:04 TimToady those ops can be curried with .assuming, I assume
17:04 jdv79 ZoffixW: the thing is metacpan is built around the cpan model which is quite different from the way we are using github
17:04 TimToady m: say &infix:<or>.assuming(*,42)
17:04 camelia rakudo-moar e0647c: OUTPUT«sub __PRIMED_ANON (|_ is raw) { #`(Sub|69131864) ... }␤»
17:04 TimToady m: say &infix:<or>.assuming(*,42)(0)
17:05 camelia rakudo-moar e0647c: OUTPUT«42␤»
17:05 TimToady yeah, you gotta use .assuming
17:05 jdv79 that's not to say we couldn't adapt the gh usage to it.
17:05 moritz jdv79: but our model (like, not doing explicit releases) is insanity in the long run
17:05 ZoffixW jdv79, but I think that naturally brings us to a point that we shouldn't be using github. It's just a temporary thing. We need to move to a CPAN structure with proper versioned releases
17:05 nine Do we have an easy way to get a local copy of the whole ecosystem?
17:05 jdv79 i'm not sure if we want to have and/or support multiple "main" ecosystems
17:05 jdv79 i think i've never gotten a clean answer to that
17:06 moritz nine: the perl6-all-modules repo on github (iirc under my username) has a copy
17:06 jdv79 some people like gh and travis and shiny stuff and some peopple like old skool perl
17:06 TimToady we expect multiple distributions to happen
17:06 JimmyZ nine: thanks.
17:06 jdv79 (cpan,pause,testers,etc)
17:06 flussence I'd feel a little more comfortable if we could have multiple mirrors listed in source-url, at least
17:06 nine moritz: for easier 'git grep'ability # that's exactly what I need it for :)
17:06 ZoffixW jdv79, you can have all that and still make regular uploads
17:07 eyck joined #perl6
17:07 moritz nine: and that's exactly why I didn't use submodules :-)
17:07 TimToady I'd put it this way; you can have a simple push model, but a complex pull model
17:08 ZoffixW jdv79, with P5 tools like Dist::Zilla the upload process is entirely transparent. All I do is type "dzil release" and blaamo...
17:09 ZoffixW (and the actual code is stored on github, with travis and all the shiny things)
17:09 moritz ZoffixW: at the cost of having to get dzil to work first, which isn't always easy with its bazillion deps
17:09 moritz but yes, I want tooling that's as pleasant to use
17:10 Skarsnik I think a github based ecosystem could work a cpan at the side if you have an easy tool to tag realease and up them to cpan/pause
17:10 ZoffixW Well, there's also App::git::ship without the crazy deps :P
17:10 jdv79 but then you have 2 ecosystems Skarsnik
17:11 nine Looks like perl6-CSS-Specification is the only distribution depending on Build.pm running with the build dir already in $*REPO. And that should be easy to replace
17:11 ZoffixW I wouldn't want a GitHub based anything in the long run. That's putting too much into an entity we have no control over.
17:11 flussence +1
17:11 jdv79 that's also a good piont
17:11 jdv79 *point
17:11 moritz we need tooling
17:12 moritz once we have tooling, it's easy to get module authors to upload releases
17:12 ZoffixW k
17:12 jdv79 i think maybe we just get both up and running and see how it falls out
17:12 jdv79 the gh side needs versioning
17:12 vividsnow joined #perl6
17:13 jdv79 and cpan6 needs to spring into existence
17:14 jdv79 moritz: so is the concensus that we abandon gh as an eco?
17:14 * flussence got an early and terrifying christmas present today, in the form of gcc5 on gentoo
17:14 jdv79 or have 2 seperate?  or try to bridge the 2 at some level like search/isntall aka something like metacpan6
17:15 AndyDee joined #perl6
17:15 jdv79 i actually have to do laundry or i'll have no clothes tomorrow:(
17:15 jdv79 sorry.  later.
17:15 Skarsnik the two could work probably if we have nice tools to create a realase from git to cpan6
17:16 stmuk panda probably shouldn't install from github but cpan6
17:16 jdv79 stmuk: that's more the answer i was looking for
17:16 Juerd Wasn't the name "CPAN6" kind of taken by a non-Perl6-project by MARKOV?
17:16 jdv79 we'll see i guess:)
17:16 stmuk but I don't see github for dev going away anytime soon
17:16 jdv79 oh, was it?  no!
17:17 Juerd http://cpan6.org/
17:17 stmuk it might be nice to still allow panda to use github installs as well but just not by default
17:18 Juerd Just allow any URL supported by git, and let authors pick their favourite site?
17:18 Juerd I don't mind using github today, but I don't want to be forced to use it once I start hating it in the future.
17:18 timotimo the cpan6.org website steals from the ubuntu logo
17:18 jdv79 who takes cpan6 and pause6 like that?
17:18 jdv79 gah
17:18 lizmat Juerd: I think CPAN6 was a project for Perl 6, much like Parrot, but they lost touch  :-(
17:18 RabidGravy it should be able to install from wherever the MET6.json says it should
17:19 stmuk not using git by default could make the dependency of panda on git optional
17:19 stmuk so windows users don't need git to just install modules
17:19 Juerd lizmat: Originally, yes, but it deviated from that in an early stage, even before the nlnet grant iirc.
17:19 timotimo not having to have git to install stuff on windows would be worth alot
17:20 Juerd Oh, Windows :(
17:20 flussence not having to use git would make distro maintainers happier too...
17:20 Juerd Maybe we can have a pure P6 Git::Download
17:20 stmuk panda is effectively dependent on both perl5 and an external git
17:21 moritz stmuk: does it use p5 for anything except prove?
17:21 stmuk moritz: yes I meant prove
17:21 jdv79 RabidGravy: for deps or the current dist?
17:22 stmuk actually using the master.zip from github.com wouldn't need git just a http client
17:22 Juerd lizmat: Wow, did you know the grant was 30k €?
17:22 jdv79 once you are looking at a meta6.json you likely have that dist or know where it is (same place), no?
17:22 Juerd That's much more than I had assumed
17:22 masak andreoss++ # adding `else` to 007
17:23 stmuk maybe panda should fall back to the .zip download off GH if git isn't found in the path
17:23 jdv79 quick whose got better names for the p6 variants of pause,cpan,metacpan
17:23 lizmat Juerd: would be nice if we could salvage something/all of that project    :-)
17:23 timotimo rauser?
17:23 jdv79 i was gonna use metacpan6 and ...6 but didn't check if all were unused
17:23 timotimo Repository of All User Supplied Ecosystem Rubbish
17:24 zjmarlow joined #perl6
17:24 Juerd lizmat: Perhaps. I don't really know if it's any good. Haven't checked it out.
17:24 moritz python used "cheese shop", ruby has "gems"
17:24 jdv79 i see a lot of XML stuff on cpan6.org.  seems dated.
17:24 flussence "butterfly net"
17:25 Juerd lizmat: But for that amount of money it sure should be useable.
17:25 moritz we can do better than acronyms, no?
17:25 timotimo it also says "it could use anything, like json or yaml", jdv79
17:25 timotimo hmm... pupae?
17:25 moritz "bugnet" :-)
17:25 ZoffixW "CPAN6 is under heavy development"... That's like the 3rd project this week I see that's "under heavy development" and seems to have been dead for a number of years
17:25 cygx joined #perl6
17:25 ZoffixW "skynet"
17:25 moritz CPAN6 doesn't really try to solve the Perl 6 module distribution problem
17:26 jdv79 maybe i'll just email and see if we can take the names
17:26 ZoffixW jdv79++ that's a good plan
17:26 moritz it tries to solve a much more general problem that grew out of the module distribution problem
17:26 jdv79 laundromat &
17:26 nine Second system syndrom done wrong?
17:27 stmuk is it possible to use the existing CPAN intrastructure for random binaries (say zip files) I wonder?
17:27 nine stmuk: it is
17:27 Juerd stmuk: Yes, it is.
17:27 nine And has been abused because of that a couple of times...
17:27 Juerd You could in theory upload your encrypted backups to it.
17:27 Juerd Or host a website on a thousand mirrors
17:28 mnewtj joined #perl6
17:28 cygx note that we also have the option to switch to libgit2 to ease "git pains" on windows
17:28 stmuk so can't we just stick https://github.com/tadzik/panda/archive/master.zip type archives up? (maybe with version nos in filename?)
17:28 mnewtj Is it possible to work in ASCII in perl6?
17:28 masak mnewtj: yes
17:28 Juerd mnewtj: Yes
17:28 moritz stmuk: we really, really want to work with Perl 6 modules that know their own version
17:28 mnewtj please tell  me how to do that
17:29 ZoffixW mnewtj, all the unicode stuff has "texas variants"
17:29 ZoffixW mnewtj, what in particular are you asking about?
17:29 lizmat stmuk: yes, it would be possible, but not sure how the PAUSE powers that be would like that
17:29 Juerd mnewtj: Can you give an example of what you wish to do?
17:29 lizmat dinner&
17:29 mnewtj mmh, that not what I meant
17:29 moritz but don't even already have PAUSE indexer support for Perl 6 modules? ISTR FROGGS++ working on that...
17:29 mnewtj I meant that if you don't need any unicode, than disable encoding
17:30 ZoffixW mnewtj, but ASCII is an encoding
17:30 mnewtj well not really, not anything magic about it at least
17:30 ZoffixW mnewtj, on the grand scheme of things, it's same as Unicode, except much smaller.
17:30 stmuk moritz: sorry can you clarify what you mean by "known own version"?
17:31 ZoffixW mnewtj, what particular problem are you trying to solve?
17:31 dalek rakudo/nom: 1ac59b2 | lizmat++ | src/core/ (2 files):
17:31 dalek rakudo/nom: Move Str perlification logic to Rakudo::Internals
17:31 dalek rakudo/nom:
17:31 dalek rakudo/nom: And hopefully speed it up a little bit
17:31 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1ac59b2b5d
17:31 moritz stmuk: you should be able to ask a module what version it is
17:31 moritz stmuk: and if we just use github-provided archives, we'll never have that
17:31 stmuk before installation?
17:31 moritz stmuk: always :-)
17:31 nine moritz: but the META6.json contains the version?
17:32 stmuk isn't that just extracting META6.json from a zip?
17:32 mnewtj ZoffixW I am  trying to increase performance
17:32 mnewtj I really like the language
17:32 masak mnewtj: increase performance? by coding only in ASCII?
17:32 moritz nine, stmuk: but the META6.json doesn't generally have the right version, unless you have a mechanism that ensures that ever commit changes the version
17:32 Juerd mnewtj: Could you share your code? The unicode stuff in Perl 6 is unlikely to slow down your program if you're not using it.
17:33 timotimo mnewtj: if you're only handling ascii, you can use Blob and Buf types with int8 types. however, that will leave you without a lot of methods and functions you'd usually use on strings.
17:33 ZoffixW mnewtj, I really doubt using a different encoding would help you with that. Have you ran the profiler? What's the bottleneck?
17:33 vytas ruby's gem, python's egg, Perl's ... hmm :))
17:33 stmuk maybe just a SHA hash of the zip?
17:33 jnthn dinner &
17:33 nine moritz: why would you want the version to change with every commit?
17:33 mnewtj Maybe it is not the unicode part
17:33 mnewtj I will have to profile a bit
17:33 nine moritz: if you do that, then what about uncommitted changes?
17:33 mnewtj but I experience strange slowdowns with simple things and I was wandering
17:33 ZoffixW mnewtj, perl6 compiler has a --profile flag
17:34 nine I really wonder what the point is of panda copying lib into blib/lib during build
17:34 nine Does anyone know?
17:34 abraxxa i'd like to merge DBDish::Oracle into DBIish on github
17:34 moritz nine: because people are arguing for using gibhub-provided zip archives as a source for module installations
17:34 moritz nine: and if you do that, every push becomes a possible version
17:35 moritz nine: and then you should try really hard to avoid multiple zips with same version but different contents
17:35 stmuk maybe the CPAN6ish stuff associates the archive SHA hash with the first version number seen and just warns if the two are out of sync
17:36 mnewtj ZoffixW I am not sure the internal profiler will help: I get things like it start fast at 100% cpu then slow down to 25%cpu usage (but really it is at least 10x slower)
17:36 abraxxa can i get access to perl6/DBIish on github?
17:37 lichtkind joined #perl6
17:37 moritz abraxxa: sure. What's your gh user name?
17:37 abraxxa abraxxa
17:37 TimToady m: 1 / 2
17:37 camelia rakudo-moar 12ba34: ( no output )
17:37 TimToady huh, I wonder if there's any deep reason / isn't marked pure
17:37 stmuk moritz: ideally the CPAN6ish would package zips .. I'm not even sure if the GH zips are constant with no git pushes
17:37 moritz abraxxa: you should already have access
17:37 ZoffixW mnewtj, why wouldn't it help? It's much more helpful than looking at how fast your CPU is spinning...
17:37 timotimo TimToady: maybe someone thought "well, if you divide by zero, it could throw"?
17:37 abraxxa moritz: as my work has also extended the general tests I'd like to test testing on MySQL and Pg as I'm only testing sqlite here
17:38 abraxxa moritz: possible, just wanted to check if it's ok
17:38 ZoffixW m: 1/0; say "hm"
17:38 camelia rakudo-moar 12ba34: OUTPUT«hm␤»
17:38 abraxxa is someone the main DBIish owner?
17:38 vytas Perl6's cub - for a perl6 project release  :)
17:38 stmuk the GH zips could be a stepping stone to CPAN6ish archives
17:38 timotimo abraxxa: moritz is, i believe
17:38 stmuk there probably should be some clever crypto signing of the archives anyway
17:38 moritz abraxxa: can you please open a pull request, and I'll try to review it on the weekend?
17:39 nine moritz: Github supports downloading versions as zip: https://github.com/rakudo/rakudo/archive/2015.11.zip
17:39 abraxxa moritz: i'd change t/99-common.pl6 to be a class with some attrs for db-specific things and a run-tests method
17:39 nine moritz: so panda could just look for the latest version and download that instead of the latest commit
17:39 abraxxa moritz: to get rid of this global %*opts
17:39 moritz abraxxa: sounds good
17:40 abraxxa moritz: is there already a Test module that does that which I can use?
17:40 lichtkind i just wrote a memory bomb
17:40 nine tadzik: I really wonder what the point is of panda copying lib into blib/lib during build
17:40 lichtkind perl 5 creid out of memory
17:40 moritz abraxxa: not that I'm aware of
17:40 lichtkind got to kill it
17:40 abraxxa lichtkind: you invented malloc? :P
17:40 lichtkind but rakudo still computes happily along
17:40 lichtkind better
17:40 ZoffixW lichtkind, what is it?
17:40 moritz lichtkind: join 'x', ('abc') xx 99e99
17:40 abraxxa moritz: ok, then i'll have a go so the PR will not break the other DB tests knowingliy
17:41 lichtkind something of that kind
17:41 masak TimToady: can infix:</> throw an exception and still be considered pure?
17:41 abraxxa moritz: your pc just doesn't have enough memory :)
17:41 timotimo m: 1 / 0; say "well..."
17:41 camelia rakudo-moar 12ba34: OUTPUT«well...␤»
17:41 lichtkind @a = 1..10e8
17:41 Begi Is there a probleme with the syntax highlighting on this page ? http://doc.perl6.org/language/objects
17:41 ZoffixW lichtkind, are you sure it's not because P5 simply does that computation like a 1000 times faster? :)
17:41 moritz masak: for the sake of constant folding, yes
17:41 geekosaur abraxxa, an interesting thing about eceptions is that throwing them is pure but catching them is not
17:41 timotimo Begi: it seems like syntax highlighting and embedded Pod6 format codes don't work together especially well :(
17:41 ZoffixW Begi, that's a wont-fix ATM.. It's due to bolds used in the code samples
17:41 abraxxa geekosaur: hm?
17:42 flussence IIRC someone complained last time an unused divide-by-0 caused an exception
17:42 geekosaur whoops
17:42 ZoffixW Begi, and last we thought about it, it was too much work to make it work, so *shrug*
17:42 moritz geekosaur: I think you meant masak
17:42 geekosaur yes
17:42 geekosaur misread scrool
17:42 geekosaur masak, an interesting thing about exceptions is that throwing them is pure but catching them is not
17:43 stmuk nine: I wonder if archive/2015.11.zip for example is always exactly the same binary under GH or if its dynamically generated?
17:43 geekosaur (the alternative is that all code is impure because there are things that can generate exceptions at any random time)
17:43 Begi so, *shrug* :p
17:43 geekosaur stmuk, removed when unaccessed for som time and then regenerated again when needed
17:43 _nadim joined #perl6
17:43 lichtkind my point is not that p5 is faster but that p6 was much gentler, p5 like ate all cpu and my computer froze
17:43 ZoffixW ¯\_(ツ)_/¯
17:43 geekosaur ...not always with the same checksum :/
17:44 stmuk geekosaur: makes sense .. thanks
17:44 lichtkind but under moar all worked still
17:44 rindolf lichtkind: hi, sup?
17:44 moritz lichtkind: I think you've just discovered that p6 lists are potentially lazy :-)
17:44 lichtkind no sudo
17:44 lichtkind normal
17:44 ZoffixW lichtkind, but you said p6 was still chugging along... Which made me think p6 will gobble up all memory just like p5, it'll just take it longer to do it :)
17:44 TimToady masak: yes, I think it's still pure even if it can fail
17:45 lichtkind if its just about laziness it would do nothing
17:46 masak geekosaur: yes, that's an interesting point.
17:47 ZoffixW lichtkind, perl6 -e 'my @a = 1..*; for @a -> $n { say $n; sleep 1 }'
17:47 stmuk which pure perl6 file archive extractions exist?
17:48 lichtkind yes but i did nothing with the list
17:48 lichtkind i dont even understand why it even computet at all
17:48 abraxxa where can i find the Pod docs?
17:48 flussence stmuk: github archives seem stable enough that gentoo uses them directly, and that has a gazillion client-side checksums
17:48 flussence (gitlab's tarballs by comparison last about 5 minutes, and are worthless for that type of use)
17:49 lichtkind in my understanding perl 6 would say ok we have a list from one to 100 million , noone wants an element good i quite
17:49 timotimo stmuk: i think we have tar and gzip
17:50 abraxxa none of the intro docs even mentions it, I've only found http://design.perl6.org/S26.html
17:50 timotimo um ... apparently no tar.
17:53 abraxxa i don't think that using =head5 for a method pod is correct as DBIish currently does
17:54 zjmarlow joined #perl6
17:58 abraxxa seems nobody writes docs
17:58 abraxxa the cheetsheat also doesn't include any pod hints
17:59 flussence I just pretend it's HTML5 with weird syntax
18:00 flussence .oO( can't imagine why anyone would ever use <h5> in the real world either )
18:00 firstdayonthejob joined #perl6
18:00 abraxxa have to go, bye
18:01 Juerd http://juerd.nl/i/5a60276c583205e275eeed034c2cf9ae.jpg
18:01 Juerd ^ Camelia rubber stamp :)
18:02 timotimo hah, neat!
18:02 timotimo very good result!
18:02 Juerd Yeah, I'm quite satisfied with it
18:04 stmuk cool!
18:06 dalek rakudo/nom: 7673d10 | (Stefan Seifert)++ | src/core/CompUnit/RepositoryRegistry.pm:
18:06 dalek rakudo/nom: Have CompUnit::RepositoryRegistry.use-repository always return the $repo
18:06 dalek rakudo/nom:
18:06 dalek rakudo/nom: Fixes PERL6LIB=$(pwd)/lib perl6 -Ilib -e '' dieing with a "Type check failed
18:06 dalek rakudo/nom: in binding; expected CompUnit::Repository but got Nil"
18:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7673d10deb
18:07 aborazmeh joined #perl6
18:07 aborazmeh joined #perl6
18:09 dalek rakudo/nom: 882649f | TimToady++ | src/ (2 files):
18:09 dalek rakudo/nom: grade school students also allowed to use unicode
18:09 dalek rakudo/nom:
18:09 dalek rakudo/nom: This adds support for × and ÷ operators, which at minimum have the nice
18:09 dalek rakudo/nom: property that you can now write * × * without confusing people (so much).
18:09 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/882649fedc
18:09 dalek rakudo/nom: 629b67c | TimToady++ | src/core/Numeric.pm:
18:09 dalek rakudo/nom: mark division as pure
18:09 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/629b67cd75
18:09 dalek rakudo/nom: b58b837 | TimToady++ | src/Perl6/Optimizer.nqp:
18:09 dalek rakudo/nom: 1 / 0 is special cased to not be useless
18:10 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b58b837c6a
18:15 smls m: dd :{ a => 1, b => 2 }
18:15 camelia rakudo-moar 7673d1: OUTPUT«Hash[Mu,Any] $var = :{:a(1), :b(2)}␤»
18:15 smls TimToady: Is this^^ correct?
18:16 smls Or was it supposed to produce a Hash[Str, Int] instead?
18:17 timotimo it's not, no
18:17 timotimo :{ } is about "object hash", meaning the keys stay objects
18:17 timotimo rather than being stringified
18:17 smls ok
18:18 smls I wonder why htmlify.p6 uses :{...} hashes all over the place, when - as far as I can see - the keys are all strings.
18:20 timotimo no clue
18:21 lucasb joined #perl6
18:21 timotimo also, assigning them to a regular % variable can get rid of the objectness right away
18:21 smls it uses binding
18:21 timotimo OK
18:23 smls Aha! https://github.com/perl6/doc/commit/b45a10ec3796553d0c45250e52b5181600c4145d
18:24 zjmarlow joined #perl6
18:24 smls Apparently they were introduced to work around previous errors, and now cause new ones... :/
18:25 timotimo weird
18:25 lucasb smls: having %h := :{ ... } looks strange. it that really necessary?
18:25 timotimo 4 days ago?
18:25 smls lucasb: That's what I'm talking about, yeah.
18:26 Juerd Why even have anything around the list?
18:27 smls I'll try reverting and fixing the previous errors properly. ;)
18:30 lucasb m: my %h = arr=>[]; say \(|%h).perl
18:30 camelia rakudo-moar 7673d1: OUTPUT«\(:arr($[]))␤»
18:30 lucasb m: my %h; %h<arr> := []; say \(|%h).perl
18:30 camelia rakudo-moar 7673d1: OUTPUT«\(:arr([]))␤»
18:30 lucasb I think the trouble with that doc commit is because of itemized vs non-itemized arrays
18:31 lucasb RabidGravy++ showed me in the conversation starting here: http://irclog.perlgeek.de/perl6/2015-12-13#i_11707328
18:31 lucasb smls: ^^ If that information helps...
18:32 smls thanks
18:33 lucasb m: my %h := :{ arr=>[] }; say \(|%h).perl
18:33 camelia rakudo-moar b58b83: OUTPUT«\("Str|arr" => $[])␤»
18:33 krunen joined #perl6
18:34 lucasb ^^ I have no idea why these key "Str|arr" is generated
18:34 lucasb *this
18:34 timotimo it's .WHICH
18:34 timotimo m: say "Hello!".WHICH
18:34 camelia rakudo-moar b58b83: OUTPUT«Str|Hello!␤»
18:35 lucasb timotimo: thanks. but it looks wrong in this context, doesn't it? It's a capture...
18:35 timotimo yeah, somehow it's mistreating the key
18:36 smls I'm rt'ing that bug right now
18:39 nine Attention please! Are there any objections to me ripping out blib handling code from panda? That means it will no longer copy lib to blib/lib and the build runs directly in the checkout directory
18:40 diakopter ask the continuous integration integration tests... oh wait..
18:40 ugexe jdv79: I'm trying to model it loosely on CompUnit::Repository. Each phase tries to load whatever plugins are defined in the config file (and the config can be used to set arbitrary options like mirrors). The config can set a content storage as default (if a content storage isn't specified), but the plugin itself must also define its str-id so a command like `install cpan:FOO:Bar:1.0` does something like `my
18:40 ugexe @uris = @plugins.first({ "cpan:FOO:Bar:1.0" ~~ $_.storage-matcher }).query("cpan:FOO:Bar:1.0", :include-depends)`
18:40 smls_ joined #perl6
18:41 ugexe This work is not finished though (nor on github) because i'm aiming for a polished end result without the workarounds I would have to implement to get around current behaviors (like precomp issues) or redoing chunks of code that were unknowingly initially wrong when things like Supply change
18:41 jdv79
18:42 jdv79 ok
18:42 hankache joined #perl6
18:44 spider-mario joined #perl6
18:44 hankache hello #perl6
18:45 timotimo heyo
18:46 [Tux] I still run into run_alt
18:47 rurban joined #perl6
18:47 hankache hola timotimo [Tux]
18:47 dalek rakudo/nom: 887ca40 | TimToady++ | src/Perl6/Grammar.nqp:
18:47 dalek rakudo/nom: improve diagnostic on detached nonalpha postfixes
18:47 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/887ca40ec6
18:48 [Tux] o/ hankache
18:49 vendethiel TimToady: now that's a fun commit message.
18:49 TimToady m: $_ .++
18:49 camelia rakudo-moar b58b83: OUTPUT«5===SORRY!5=== Error while compiling /tmp/VOWNPX_bw1␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/VOWNPX_bw1:1␤------> 3$_ .+7⏏5+␤»
18:50 TimToady that's what it fixes
18:50 vendethiel oh yeah, that I get :). I just liked the commit message
18:50 dalek Inline-Perl5: 5e1d9b3 | (Stefan Seifert)++ | / (3 files):
18:50 dalek Inline-Perl5: Fix installation with panda
18:50 dalek Inline-Perl5:
18:50 dalek Inline-Perl5: Now that blib is gone, there is no difference anymore between a local build
18:50 dalek Inline-Perl5: during development and building with panda. Should simplify finding
18:50 dalek Inline-Perl5: p5helper.so.
18:50 dalek Inline-Perl5: review: https://github.com/niner/Inline-Perl5/commit/5e1d9b31d4
18:51 TimToady m: $_ .++
18:51 camelia rakudo-moar 887ca4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/MnSIAnyq56␤Malformed postfix call (only alphabetic methods may be detached)␤at /tmp/MnSIAnyq56:1␤------> 3$_ .7⏏5++␤»
18:51 TimToady m: $_ . "foo"
18:51 camelia rakudo-moar 887ca4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/CVtgjgY8qJ␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/CVtgjgY8qJ:1␤------> 3$_ .7⏏5 "foo"␤»
18:52 TimToady m: $_ X. "foo"
18:52 camelia rakudo-moar 887ca4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/W_ZiW2Aied␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/W_ZiW2Aied:1␤------> 3$_ X.7⏏5 "foo"␤»
18:52 TimToady m: $_ X.++
18:52 camelia rakudo-moar 887ca4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/fVjNAifnfs␤Malformed postfix call (only alphabetic methods may be detached)␤at /tmp/fVjNAifnfs:1␤------> 3$_ X.7⏏5++␤»
18:53 joeschmoe joined #perl6
18:53 TimToady I guess it's reasonable there
18:53 TimToady m: $_ X. foo
18:53 camelia rakudo-moar 887ca4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/LkaPXtM7Rq␤Cannot cross with . because dotty infix operators are too fiddly␤at /tmp/LkaPXtM7Rq:1␤------> 3$_ X.7⏏5 foo␤»
18:53 vendethiel- joined #perl6
18:53 TimToady well, it's really reserved if we decide to make it work
18:54 dalek modules.perl6.org: d548fca | (Zoffix Znet)++ | / (3 files):
18:54 dalek modules.perl6.org: Do not operate on weird URLs, even if they are on GitHub
18:54 dalek modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/d548fcac28
18:54 zjmarlow joined #perl6
18:54 TimToady m: $_ X. foo,bar  # could work someday
18:54 camelia rakudo-moar 887ca4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/_0dUHfHFsS␤Cannot cross with . because dotty infix operators are too fiddly␤at /tmp/_0dUHfHFsS:1␤------> 3$_ X.7⏏5 foo,bar  # could work someday␤»
18:55 joeschmoe m: say 'Fizz' x $_ %% 3 ~ 'Buzz' x $_ %% 5 || $_ for 1 .. 100;
18:55 camelia rakudo-moar 887ca4: OUTPUT«1␤2␤Fizz␤4␤Buzz␤Fizz␤7␤8␤Fizz␤Buzz␤11␤Fizz␤13␤14␤FizzBuzz␤16␤17␤Fizz␤19␤Buzz␤Fizz␤22␤23␤Fizz␤Buzz␤26␤Fizz␤28␤29␤FizzBuzz␤31␤32␤Fizz␤34␤Buzz␤Fizz␤37␤38␤Fizz␤Buzz␤41␤Fizz␤43…»
18:55 TimToady we'd have to distribute . the way we distribute thunkityness
18:56 TimToady and that sort of thing is why we detached the precedence of x from *
18:57 vendethiel- TimToady: probably is, you probably want to compose that?
18:58 vendethiel- well,you can always just add [+], I guess
18:58 dalek Inline-Perl5: f4aff79 | (Stefan Seifert)++ | Build.pm:
18:58 dalek Inline-Perl5: Only create the resources dir if it doesn't exist
18:58 dalek Inline-Perl5: review: https://github.com/niner/Inline-Perl5/commit/f4aff7904b
18:58 dalek Inline-Python: 6122cec | (Stefan Seifert)++ | / (3 files):
18:58 dalek Inline-Python: Fix installation with panda
18:58 dalek Inline-Python:
18:58 dalek Inline-Python: Now that blib is gone, there is no difference anymore between a local build
18:58 dalek Inline-Python: during development and building with panda. Should simplify finding
18:58 dalek Inline-Python: p5helper.so.
18:58 dalek Inline-Python: review: https://github.com/niner/Inline-Python/commit/6122cec6a4
19:00 nine moritz: Inline::Python should now be installable again with latest panda
19:00 vendethiel- nine++
19:00 hankache nine++
19:01 timotimo ninf \o/
19:01 tadzik yay :)
19:01 Amnez777 joined #perl6
19:02 Amnez777 joined #perl6
19:02 dalek modules.perl6.org: 2ac27f1 | (Zoffix Znet)++ | Build.PL:
19:02 dalek modules.perl6.org: Bump version of Mew prereq
19:02 dalek modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/2ac27f1873
19:03 nine All my commits to panda combined: 16 files changed, 113 insertions(+), 227 deletions(-)
19:03 TimToady m: 1 / 3
19:03 camelia rakudo-moar 887ca4: OUTPUT«WARNINGS for /tmp/_SiSxH2FhE:␤Useless use of "/" in expression "1 / 3" in sink context (line 1)␤»
19:03 TimToady m: 1 / 0e0
19:03 camelia rakudo-moar 887ca4: OUTPUT«Attempt to divide 1 by zero using /␤  in block <unit> at /tmp/g8fEtj3t5Q:1␤␤Actually thrown at:␤  in block <unit> at /tmp/g8fEtj3t5Q:1␤␤»
19:03 nine in 24 commits
19:03 TimToady m: 1 / 0
19:03 camelia rakudo-moar 887ca4: ( no output )
19:03 cygx_ joined #perl6
19:03 timotimo nine: pretty good :)
19:03 Juerd Ooh, we can /0 now? Fancy feature that very few languages have :)
19:03 diakopter TimToady: maybe it's already a Rat?
19:03 TimToady the last is deemed to have the side effect of telling you Rat division is lazy
19:04 TimToady so not completely useless :)
19:04 _nadim joined #perl6
19:04 hankache joined #perl6
19:06 * vendethiel- starts writing his advent blog post
19:06 timotimo nice :)
19:06 vendethiel- FWIW, I prepared two, so I'll post either: do people prefer one about sized/typed/shaped variables/arrays, or about introspection?
19:07 hankache vendethiel BOTH!!
19:07 vendethiel- well, I can publish the other one on my "own" blog (if I make one)
19:08 vendethiel- but not as part as the advent, as there are no slots left
19:08 hankache i see
19:09 timotimo i wasn't against the "on some days have two posts instead of just one" idea :P
19:09 joeschmoe so im waffling back and forth between perl6 and python3 ive taken beginer online cources in both and probly will learn both; however what advantages does perl6 have
19:09 vendethiel- timotimo: well, that's fine by me
19:09 _nadim joined #perl6
19:09 joeschmoe advantages over perl3
19:10 joeschmoe *python3 (wow im terible)
19:10 timotimo :D
19:10 timotimo "perl6's advantages over perl3" would be a fun advent blog post
19:10 joeschmoe LOL
19:10 vendethiel- :D
19:11 TimToady .oO(not much, perl3 already handled binary data)
19:11 hankache perl3!! I wasn't even born yet
19:12 hankache what year was it again?
19:13 TimToady oh, 'bout 1990 or so
19:13 joeschmoe i googled and your about right 1989
19:13 joeschmoe So now that I have made a fool of myself what do yo think does perl6 have any advantages over python3
19:13 TimToady perl4 was '91
19:13 hankache great I was 1 year old
19:14 TimToady joeschmoe: yes :)
19:14 joeschmoe Im guesing Roles is one
19:14 _nadim joined #perl6
19:14 hankache joeschmoe absolutely
19:14 pierre-vigier joined #perl6
19:14 TimToady error messages
19:15 TimToady decent lexical scoping
19:15 TimToady language mutability
19:15 TimToady not confusing strings with arrays :)
19:15 timotimo lexical scoping is a big one that very often gets overlooked
19:15 joeschmoe I noticed that python error messages are kinda cryptic
19:15 RabidGravy multiple dispatch
19:15 TimToady it's a very, very long list
19:16 flussence consistent string quoting syntax
19:16 TimToady sigils ;)
19:16 TimToady laziness
19:16 joeschmoe what about graphical aplication at this point it looks like perl6 does not have a lot of GUI hooks
19:17 hankache joeschmoe most of all, beyond the technical stuff listed above, in #perl6 land it's Christmas all year long
19:17 timotimo yes, one christmas every month, roughly
19:17 TimToady on GUI people just need to keep hooking up libraries
19:17 timotimo joeschmoe: we have a simple GTK binding, electron and we can use PyQt and PySide via Inline::Python
19:17 TimToady we ain't gonna bulid in any one particular flavor of the month on GUI
19:17 TimToady *build
19:18 timotimo yeah, no Tk for you!
19:18 joeschmoe PyQt's not bad
19:18 timotimo so ... is there a simple commandline tool that'd take two lines in a row and spit them out again as one line? i bet awk can do this really quick like
19:18 timotimo PySide is better than PyQt :)
19:18 flussence .oO( HTML is a better GUI anyway, it works in terminals... ;)
19:19 timotimo flussence: and it's automatically searchable! usually ...
19:19 Skarsnik Qt binding will be nice if we could have supply and signal/slot work together x)
19:19 FROGGS joined #perl6
19:19 joeschmoe Im sure you can mangle HTML5 pretty fast in and well in perl6
19:19 Skarsnik No
19:19 _nadim joined #perl6
19:19 FROGGS o/
19:20 Skarsnik html5 is not easy to parse
19:20 hankache a graphical interface!!! Blasphemy!! go back to your terminal spoiled kids
19:20 timotimo hum. do we have something that'd let me align a string with spaces given its size? i know python has something
19:21 joeschmoe Skarsnik: Well aparently im told HTML5 is "The Future".
19:21 flussence timotimo: fmt('%-20s')?
19:21 flussence (you can replace "20" with .chars)
19:21 timotimo oh, indeed
19:21 Skarsnik webbased Ui sucks comparing to native one x)
19:22 hankache joeschmoe i believe so.
19:22 timotimo also, my number was way too low, so i wouldn't have ended up with spaces anyway
19:22 timotimo i've tried "% 20s" and "%020s"
19:23 timotimo also, my "two lines into one" now reads perl6 -ne '$_.fmt("%-30s").print; "\n".print if $++ %%2'
19:24 joeschmoe Skarsnik: the trade off in html is portabilty over speed and reliability
19:24 zjmarlow joined #perl6
19:24 _nadim joined #perl6
19:24 Skarsnik html5 do nothing alone. it need JS x)
19:25 hankache skarsnik or Perl6 ;)
19:25 lucasb timotimo: sed 'N;s/\n/ /'  # two lines into one :)
19:26 joeschmoe JS is like starbucks its on evrry corner but the coffe is kinda over rosted and overpriced but if you need a cup quick what are you going to do
19:27 TimToady I'm thinking we should maybe do a security evaluation for EVAL before we release, and allow EVAL in user code only if MONKEY-SEE-NO-EVAL is used, because injection attacks
19:28 _nadim joined #perl6
19:29 hankache ++ what TimToady said
19:29 Juerd use experimental :eval;
19:29 TimToady not harsh enough :)
19:30 Skarsnik and find a solution to replace the eval in DBIish test x)
19:30 hankache MONKEY-SEE-NO-EVAL is Perl6ish
19:30 hankache i like it
19:30 * TimToady has been tempted to rename experimental to MONKEY-AROUND
19:30 Juerd The results of the careful security evaluation are in: it's very dangerous.
19:30 hankache hahaha please do so :)
19:31 TimToady as a sop, we could just turn on all the monkeys with 'use MONKEY;'
19:31 Juerd dial M;
19:32 hankache excellent
19:32 Juerd I think it should be called ONKEY because of perl6 -MONKEY
19:33 timotimo oh lord
19:33 cygx .oO( use UNSAFE :eval )
19:33 TimToady and a module that uses any MONKEYs would automatically have them marked on the website as dangerous
19:33 timotimo use DANGER
19:33 Juerd IMHO the allcaps of EVAL is a strong enough signal
19:33 TimToady no, it isn't
19:34 TimToady not the moment you do web programming
19:34 timotimo only allow use MONKEY at the top of a file
19:34 Juerd Then instead of enabling it with something else, just rename it?
19:34 RabidGravy or -DONKEY
19:34 TimToady well, grep is cheap
19:34 _nadim joined #perl6
19:34 hankache monkey monkey monkey
19:34 vendethiel joined #perl6
19:35 TimToady and lots of great puns on "get the monkey off our back" and such :)
19:35 vendethiel TimToady: "you think it's cute now"? :P
19:35 TimToady I think it's important to recognize dangerous constructs, seriously
19:35 timotimo m: my $name = "EV" ~ "AL"; ::($name)("say 'hi'")
19:35 camelia rakudo-moar 887ca4: OUTPUT«No such symbol 'EVAL'␤  in block <unit> at /tmp/HmXVTtLiRr:1␤␤Actually thrown at:␤  in block <unit> at /tmp/HmXVTtLiRr:1␤␤»
19:35 timotimo ^- cannot be grepped
19:36 TimToady sure, but we're looking for global footguns here
19:36 hankache and see/speak/do no evil
19:36 timotimo ah, not for "people are intent on hurting you"
19:36 vendethiel yes, recognize them, but I'm not sure tying them to some monkey is really good
19:37 cygx sub MURDER-DEATH-KILL-187-EVAL() {...}
19:37 pdl joined #perl6
19:37 lsm-desktop joined #perl6
19:38 * TimToady wonders how much of the ecosystem is already depending on EVAL for webcode...
19:38 RabidGravy I have a test that does
19:38 TimToady we have no tainting mechanism yet to prevent abuse
19:38 timotimo Mustache generates its template code via EVAL i think
19:38 Juerd sub EVAL ($code, :$I-accept where :so) { ... }
19:38 hankache TimToady what alternative do we have?
19:39 Juerd sub EVAL ($code, :suppress-important-warning) { ... }
19:39 TimToady parsers that aren't injectable
19:39 Skarsnik Mojo::Template+
19:40 Skarsnik all Template probably use eval
19:41 TimToady you realize that, currently, if you say / <{ $outsiderdata }> / it's GAME OVER, and you are pwned?
19:41 Juerd I did not realize that.
19:41 Juerd <EVAL { ... }>?
19:42 Skarsnik same for "{$somevar}" I guess?
19:42 dalek modules.perl6.org: 187c4f8 | (Zoffix Znet)++ | / (5 files):
19:42 dalek modules.perl6.org: CAN HAS KOALITY?
19:42 dalek modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/187c4f8aec
19:42 timotimo no, {$somevar} is harmless
19:42 Juerd Skarsnik: That's just a string injection. Could be harmful.
19:43 Juerd But that's not really easy to prevent at this point
19:43 TimToady only if you eval it downstream
19:43 timotimo hm. in the to-match part of the regex, as a literal?
19:43 Juerd TimToady: Which the SQL interpreter will do for you :P
19:43 TimToady why one should *always* use placeholders, never interpolation
19:43 timotimo maybe if you have a database (text file) with username:password and you search for / {$username}:{$password} /
19:43 Juerd Agreed
19:44 timotimo then if the supplied $password is empty (or a prefix of the actual password), it'd pass
19:44 timotimo because not anchored
19:44 TimToady sorry, went to a security seminar last week, and got scared :)
19:44 Juerd TimToady: This brings me to an unanswered question I asked a while ago: can users build their own custom quote operators? I can see how accepting only q:sql(...), where $foo is automatically placeholderified would be nice.
19:44 TimToady but I'd like to not go down in history as enabling the next round of security attacks
19:45 Juerd TimToady: Be more terified over allowing nullbytes in syscalls (e.g. filenames) than over deliberate uses of eval.
19:45 TimToady Juerd: once we have slangs, sure
19:45 Juerd The ungreppable things are always the scariest ones
19:45 yqt joined #perl6
19:45 * Juerd <- security guy
19:46 Juerd But, /me also has to run because the supermarket closes in 14 minutes
19:46 Juerd bbl
19:46 TimToady o/
19:46 timotimo i have an hour longer for supermarket grocery shopping :)
19:48 hankache i think security should be a focal point as Perl 6 opens up to the world
19:48 [Tux] https://gist.github.com/Tux/9daa5f64f42395fc2941 <= Inline::Perl5 functions again, CSV::Parser got slower
19:48 timotimo hankache: just a *little* bit late on that :)
19:48 Sqirrel joined #perl6
19:49 timotimo [Tux]: could i ask you to run csv-parser with MVM_JIT_LOG=some_file.txt and grep that file like this: | grep Constructing\\\|BAIL | grep -B1 BAIL | grep -v -- --
19:49 hankache timotimo: going forward maybe
19:50 timotimo sure
19:50 [Tux] https://gist.github.com/Tux/7621ffd6b91f55c2e4c8 <= csv-parser slowed down by a factor TWO!
19:52 b7j0c joined #perl6
19:53 b7j0c is there a meaningful difference between `constant $foo = 'foo'` and `$foo := 'foo'` ??
19:53 smls_ Alright, so the original htmlify.p6 problem that pdl tried to solve with commit b45a10e and for which I am trying to find a better solution now, is the result of an unfortunate interaction of two Perl 6 behaviors:
19:53 smls_ 1) The default BUILD method for objects uses assignment rather than binding semantics to map named arguments to attributes:
19:53 smls_ class A { has @.a }; say A.new(a => 42)       # @!a is now [42]
19:53 smls_ class A { has @.a }; say A.new(a => $[3, 4])  # @!a is now [$[3, 4],]
19:53 smls_ 2) Hashes containerize their values:
19:53 smls_ my %hash = a => [3, 4]   # %hash<a> is now $[3, 4]
19:53 smls_ 3) As a result of 1 and 2, naively building an argument hash first and then doing  A.new(|%hash),  blows up with array attributes.
19:54 zjmarlow joined #perl6
19:54 [Tux] timotimo, https://gist.github.com/Tux/af8ccc15da2e8138b11d
19:54 smls_ Behavior (1) came as somewhat of a surprise to me. Is that new?
19:55 * diakopter vendethiel please feel free to take my advent slot
19:55 diakopter vendethiel: ^
19:55 [Tux] I have to go. cu l8r
19:56 vendethiel diakopter: you sure?
19:56 timotimo [Tux]: i expect eof and get may be called often-ish from your code?
19:57 timotimo hmm. but just "we don't jit that" doesn't explain the 2x time unimprovement
19:57 [Coke] looking at the advent schedule, we go from 24 TimToady to 26 ?
19:57 lizmat [Tux]: any idea when this slowdown occurred ?
19:58 timotimo still, the jit should jit that. i'll look into that tonight. but first: grocery shopping
19:58 masak [Coke]: I only see the advent schedule go up to 25
19:58 hankache it's supposed to stop at 25
19:58 adhoc joined #perl6
19:59 hankache it's a countdown to 25
19:59 dalek mu: fd2ae70 | (Carl Masak)++ | misc/perl6advent-2015/schedule:
19:59 dalek mu: retcon the 17th topic
19:59 dalek mu:
19:59 dalek mu: *jedi mind trick*
19:59 dalek mu: review: https://github.com/perl6/mu/commit/fd2ae70d3f
20:02 TimToady looks like we use EVAL two places internally, in INTERPOLATE and in .assuming
20:03 TimToady INTERPOLATE is certainly injectable, and .assuming might or might not be
20:04 lizmat TimToady: Cursor.MAKEREGEX>
20:04 TimToady INTERPOLATE needs to allow only a non-injectable subset of regex syntax
20:04 lizmat ?
20:05 TimToady yes, that's what INTERPOLATE calls
20:05 lizmat ah, ok
20:05 TimToady .assuming probably needs to sanitize typenames for weird characters
20:05 lizmat I think this was mentioned in a P6Advent post as well: we don't have an MOP API for creating regexen
20:05 nine lizmat: Inline::Perl5 should now install with the latest panda
20:06 lizmat nine: will try
20:06 TimToady or just futz with the metadata
20:07 smls_ I wouldn't mind if regex interpolation went away for now (or marked experimental), so a safe version of it can come back later.
20:08 [Coke] wonder if I had a local edit, weird.
20:08 [Coke] so there's no post, per se, on teh 25th, then. I will probably post the release announcement or something on that day on the blog.
20:11 _nadim joined #perl6
20:12 lizmat FWIW, I'm seeing core parse times go above 50 seconds again  :-(
20:14 anon joined #perl6
20:16 dalek mu: ed97f15 | ven++ | misc/perl6advent-2015/schedule:
20:16 dalek mu: Take over @diakopter++'s slot for Fun And Profit(tm)
20:16 dalek mu: review: https://github.com/perl6/mu/commit/ed97f156f7
20:16 b7j0c left #perl6
20:22 dalek rakudo/nom: af74f04 | lizmat++ | src/core/Rakudo/Internals.pm:
20:22 dalek rakudo/nom: Add \0 as special case, better handle multiples
20:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/af74f048a0
20:22 dalek rakudo/nom: 2735116 | lizmat++ | src/core/IO/Path.pm:
20:22 dalek rakudo/nom: Show non-printables in IO::Path.perl
20:22 dalek rakudo/nom:
20:22 dalek rakudo/nom: Related to #126935
20:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2735116bfb
20:23 _nadim joined #perl6
20:24 zjmarlow joined #perl6
20:24 dalek modules.perl6.org: b00cb8e | (Zoffix Znet)++ | lib/ModulesPerl6/Model/SiteTips.pm:
20:24 dalek modules.perl6.org: Fix heading in docs
20:24 dalek modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/b00cb8e480
20:26 dalek roast: 94aeb45 | usev6++ | S04-declarations/my.t:
20:26 dalek roast: Fudge tests for sigilless variables for JVM
20:26 dalek roast: review: https://github.com/perl6/roast/commit/94aeb45f3a
20:26 dalek roast: d5eb43e | usev6++ | S11-compunit/rt126904.t:
20:26 dalek roast: Use $*EXECUTABLE instead of 'perl6', fudge for JVM
20:26 dalek roast: review: https://github.com/perl6/roast/commit/d5eb43ea0f
20:29 FROGGS m: my $a = '. }; for ^3 { say 42 '; 'a' ~~ /<$a>/
20:29 camelia rakudo-moar 887ca4: OUTPUT«42␤42␤42␤Cannot find method 'Nil'␤  in block <unit> at /tmp/u6JMJUG3WX:1␤␤»
20:31 lostinfog joined #perl6
20:32 dalek roast: 9ab7485 | lizmat++ | S32-io/io-path.t:
20:32 dalek roast: Add test for non-printables in path
20:32 dalek roast: review: https://github.com/perl6/roast/commit/9ab7485fde
20:33 masak FROGGS: O.O
20:34 vendethiel m: my int @a[8]; @a++;
20:34 camelia rakudo-moar 273511: OUTPUT«Cannot call postfix:<++>(array[int]); none of these signatures match:␤    (Mu:D $a is rw)␤    (Mu:U $a is rw)␤    (Int:D $a is rw)␤    (int $a is rw)␤    (Bool:U $a is rw)␤    (Bool:D $a is rw)␤    (Num:D $a is rw)␤    (Num:U $a is rw)␤  …»
20:34 Juerd lizmat: Is there any IO backend in which passing \0 is actually sane thing to do?
20:34 lizmat FROGGS: looks like we *really* need to fix INTERPOLATE
20:34 Juerd lizmat: Because \0 has historically been a source of security bugs, also in Perl 5.
20:35 lizmat no, but it is a known way to hide info
20:35 lizmat indeed
20:35 FreezerburnV joined #perl6
20:35 Juerd I'd opt for disallowing it if we can. I don't know if Str.IO is the right place though.
20:35 lizmat perhaps we need to indeed not allow \0, but any other non-printable could be ok, no ?
20:36 Juerd Yes, only 0 and / are special. Disallowing / is kind of problematic though ;)
20:36 brrt joined #perl6
20:36 Juerd s/0/\0/
20:36 masak Juerd: there's a network protocol in Git that uses \0 in the middle of a line to transmit "meta data". I read that in a post somewhere.
20:36 dalek rakudo/nom: 6ca94e8 | lizmat++ | src/core/IO/Path.pm:
20:36 dalek rakudo/nom: Fix IO::Path.perl thinko
20:36 dalek rakudo/nom:
20:36 dalek rakudo/nom: Since we properly escape all of the characters in the path, we can now just
20:36 dalek rakudo/nom: use double quotes, which also improves readability
20:36 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6ca94e8332
20:36 Juerd Specifically, maybe check that \n stays \n (without \r) there :)
20:36 masak Juerd: it feels presumptuous to disallow something and then leave no way to implement things like that protocol.
20:37 Juerd masak: In data streams it's okay, including in files. In *paths*, hostnames, etc, it's an issue.
20:37 ZoffixW joined #perl6
20:37 lizmat well, for now, it at least makes them visible
20:38 vendethiel FROGGS: lol, it's -pe all over again :P
20:38 ZoffixW I'd leave the decision of whether it's an issue for the programmer.
20:38 vendethiel m: my int @a[8]; @a[0]++;
20:38 camelia rakudo-moar 273511: ( no output )
20:38 vendethiel m: my int @a[8]; @a[0]++; say @a[0];
20:38 camelia rakudo-moar 273511: OUTPUT«1␤»
20:38 vendethiel m: my int @a[8]; @a[0] += 3; say @a[0];
20:38 camelia rakudo-moar 273511: OUTPUT«3␤»
20:39 vendethiel jnthn++ # fixing these up in time so I can include them in the advent blog post :-)
20:39 vendethiel m: my int @a[4][4]; @a[0][0]++; say @a; # or was it this one?
20:39 camelia rakudo-moar 273511: OUTPUT«5===SORRY!5=== Error while compiling /tmp/dzvzpcloTs␤Multiple shapes not yet understood␤at /tmp/dzvzpcloTs:1␤------> 3my int @a[4][4]7⏏5; @a[0][0]++; say @a; # or was it this o␤    expecting any of:␤        constraint␤»
20:39 vendethiel m: my int @a[4;4]; @a[0][0]++; say @a; # or was it this one?
20:39 camelia rakudo-moar 273511: OUTPUT«Partially dimensioned views of arrays not yet implemented. Sorry. ␤  in block <unit> at /tmp/5ZCBW9o2_z:1␤␤»
20:39 FROGGS lizmat: IIRC both pmichaud and TimToady told me that one could implement INTERPOLATE without EVAL... but I was unable to find a solution, and there was no lack of trying
20:40 dalek modules.perl6.org: 2dd8327 | (Zoffix Znet)++ | / (2 files):
20:40 dalek modules.perl6.org: Increase test/doc coverage to 100% for ModulesPerl6::Metrics::Koalatee
20:40 dalek modules.perl6.org: review: https://github.com/perl6/modules.perl6.org/commit/2dd832791d
20:40 dwarring joined #perl6
20:42 Juerd A brainfart: subset CStr of Str where * !~~ /\0/, and make hostnames, io paths, etc, CStr instead of Str
20:42 lizmat FROGGS: as tadzik was saying "but since we have no other way of programmatically constructing the individual regexes (that I know of)"
20:42 lizmat in https://perl6advent.wordpress.com/2015/12/08/day-8-grammars-generating-grammars/
20:43 Juerd ZoffixW: It's been shown that most programmers are not aware of \0 being special in underlying system calls.
20:43 masak Juerd: ok -- no arguments against disallowing it in paths/hostnames, etc.
20:43 vendethiel m: my int @a[4]; my int @b[4]; say @a Z+ @b;
20:43 camelia rakudo-moar 273511: OUTPUT«(0 0 0 0)␤»
20:44 Juerd In fact, I think that many things could benefit from a \0-less Str
20:44 jnthn lizmat: I suspect QTrees will make that nicer
20:44 Juerd It's a great way to document the incompatibility too
20:44 ZoffixW m: subset CStr of Str where * !~~ /\0/; my $host = "\0"
20:44 camelia rakudo-moar 273511: ( no output )
20:44 lizmat jnthn: that would be post-christmas, no ?
20:44 ZoffixW m: subset CStr of Str where * !~~ /\0/; my CStr $host = "\0"
20:44 camelia rakudo-moar 273511: OUTPUT«Type check failed in assignment to $host; expected CStr but got Str␤  in block <unit> at /tmp/XveP3tsNs_:1␤␤»
20:45 jnthn lizmat: Sure
20:45 Juerd I don't know if it should be CStr or Cstr, or something else completely
20:45 masak lizmat: yeah, post christmas :)
20:45 jnthn lizmat: For now EVAL
20:45 masak lizmat: there's a foretaste of Qtrees in 007 already, if you're curious :>
20:45 lizmat m: subset CStr of Str where !*.contains("\0"); my CStr $host = "\0"
20:45 vendethiel https://perl6advent.wordpress.com/?p=3504&amp;preview=true&amp;iframe=true is the blog post scheduled for tomorrow, if people want to suggest some improvements, i'm all ears, thanks :-)
20:45 camelia rakudo-moar 273511: OUTPUT«Type check failed in assignment to $host; expected CStr but got Str␤  in block <unit> at /tmp/rENE3XiB_5:1␤␤»
20:45 masak lizmat: but 007 doesn't have regexes (yet)
20:46 masak lizmat: a couple months back I was close to creating a Qtree implementation (in module space) for Perl 6 regexes only. I might still do that, in fact.
20:46 timotimo vendethiel: ugh, didn't i fix my int @a[8]; @a[4]++ ?!?
20:46 lizmat well, FROGGS has just shown that we need to fix INTERPOLATE
20:46 vendethiel timotimo: I think you did
20:46 vendethiel did I leave something about that? whoops
20:46 timotimo it works locally
20:46 lizmat before Christmas
20:46 timotimo vendethiel: i only just saw it in the backlog
20:46 jnthn lizmat: Preferably before, yes
20:46 * flussence gently points out that docs/announce/2015.12.md still contains the paragraph about 2015.11's release name
20:46 Sqirrel_ joined #perl6
20:46 vendethiel timotimo++ <3
20:46 vendethiel got confused, sorry :)
20:47 timotimo but ...
20:47 timotimo m: my int @a[4]; @a[2]++
20:47 camelia rakudo-moar 273511: ( no output )
20:47 timotimo ooooh
20:47 [Coke] flussence: fixing.
20:47 vendethiel yep
20:47 timotimo you didn't [ ]
20:47 timotimo m)
20:47 timotimo man, you gave me a little shock right there
20:47 vendethiel oh, that. Yes, I went ahead and starting /w-ing camelia instead of spamming this channel...
20:48 ZoffixW m: my @dwarves[7] = ^7; my $x; for ^1000 { $x = @dwarves.elems }; say now - INIT now
20:48 camelia rakudo-moar 273511: OUTPUT«0.01995285␤»
20:48 ZoffixW m: my @dwarves = ^7; my $x; for ^1000 { $x = @dwarves.elems }; say now - INIT now
20:48 camelia rakudo-moar 273511: OUTPUT«0.00438673␤»
20:48 ZoffixW :\
20:48 dalek rakudo/nom: 6b5e9f5 | coke++ | docs/announce/2015.12.md:
20:48 dalek rakudo/nom: Post Christmas we get to pick a new naming scheme.
20:48 dalek rakudo/nom:
20:48 dalek rakudo/nom: Also, bend was last month. (flussence++)
20:48 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6b5e9f54dd
20:48 ZoffixW You'd think those times would be reversed :)
20:49 jnthn Nope, most recent addition = least optimized :)
20:49 vendethiel m: my int @dwarves[7] = ^7; my $x; for ^1000 { $x = @dwarves.elems }; say now - INIT now
20:49 camelia rakudo-moar 273511: OUTPUT«0.01929793␤»
20:49 ZoffixW Fair enough :D
20:49 vendethiel a *bit* more optimized :P
20:49 timotimo how come it's slower? because it has to consider if there's a shape?
20:50 lizmat possibly, no idea
20:50 lizmat most of the multidim stuff happened while I was partying in the Carribean
20:51 dalek doc: 6eeef84 | smls++ | htmlify.p6:
20:51 dalek doc: Revert "Fix type check failing (gh #227)"
20:51 dalek doc:
20:51 dalek doc: It didn't actually fix the problem, just hid it and replaced it with
20:51 dalek doc: a new problem by causing us to run into Rakudo bug RT 126951. Not to
20:51 dalek doc: mention it made the hash definitions kinda ugly.
20:51 dalek doc:
20:51 dalek doc: This reverts commit b45a10ec3796553d0c45250e52b5181600c4145d.
20:51 dalek doc: review: https://github.com/perl6/doc/commit/6eeef84e74
20:51 dalek doc: bd1480c | smls++ | lib/Perl6/Documentable.pm:
20:51 dalek doc: Another attempt to fix initialization of @!categories (gh #227)
20:51 dalek doc:
20:51 dalek doc: The problem is that when pre-building the attribute hash and then
20:51 dalek doc: creating the Perl6::Documentable object with .new(|%args), array
20:51 dalek doc: values are itemized by he hash, and thus the wrong thing happens
20:51 dalek doc: when the default constructor assigns them to the respective
20:51 dalek doc: array attributes of the class via single-argument rule.
20:51 dalek doc:
20:51 dalek doc: This commit overrides the default constructor to remove the
20:51 dalek doc: itemization, which fixes the problem.
20:51 dalek doc:
20:51 dalek doc: A better long-term solution might be to refactor htmlify.p6 to not
20:51 dalek doc: pre-build the attribute initialization lists as hashes.
20:51 dalek doc: review: https://github.com/perl6/doc/commit/bd1480c1e6
20:51 smls_ moritz: With this^^ htmlify works again for me.
20:53 timotimo smls++
20:54 zjmarlow joined #perl6
20:54 pierre-vigier joined #perl6
20:55 timotimo it seems like the int array should just use nqp::elems on its storage
20:55 timotimo it doesn't make sense to me that it'd be slower, unless there's some inlining/candidate choosing that's not going on?
20:55 jnthn timotimo: Not sure why elems would be any slower
20:56 timotimo time to dig in!
21:00 FreezerburnV Is there any way to define a function at compile time? (e.g.: make a function that logs debug output if the environment has a flag in it, or make the same function empty if it doesn’t)
21:01 Juerd FreezerburnV: BEGIN
21:01 grondilu my &f;  BEGIN { &f := $*VAR ?? sub {...} !! sub {...} } # should work
21:02 timotimo ugh
21:02 timotimo KnowHOW methods must be called on object instance with REPR KnowHOWREPR
21:02 timotimo when it tries to dump the profiler output :(
21:03 timotimo then it's gotta be jit log and spesh log, i suppose :\
21:03 brrt hey, i just noticed, the flags to the gcc invocation for the rakudo extension modules are passed dobule
21:04 brrt double
21:04 Juerd m: my &debug; &debug = BEGIN %*ENV<DEBUG_ME> ?? { note $^a } !! sub ($) { }; debug "hi";
21:04 camelia rakudo-moar 6ca94e: ( no output )
21:04 Juerd Oh. Heh. I guess I can't set environment varibales on camelia :P
21:04 brrt e.g.: gcc  -L/home/bart/usr/lib -shared -fPIC -O3 -DNDEBUG -Wl,-rpath,/home/bart/usr/lib -Wl,-rpath,/home/bart/usr/share/perl6/site/lib  -O3 -DNDEBUG -lm -lpthread -lrt -ldl -o dynext/libperl6_ops_moar.so dynext/perl6_ops.o dynext/container.o -lmoar
21:04 Juerd m: my &debug; &debug = BEGIN "cheating lalala" ?? { note $^a } !! sub ($) { }; debug "hi";
21:04 camelia rakudo-moar 6ca94e: OUTPUT«hi␤»
21:05 Juerd FreezerburnV: Something like that? :)
21:05 FreezerburnV Juerd: Awesome, thanks!
21:05 Juerd m: my &debug = BEGIN "cheating lalala" ?? { note $^a } !! sub ($) { }; debug "hi";
21:05 camelia rakudo-moar 6ca94e: OUTPUT«hi␤»
21:05 Juerd Separate assignment was silly
21:05 dalek rakudo/nom: 6b1b3ef | lizmat++ | src/ (2 files):
21:05 dalek rakudo/nom: Remove nqp deprecation
21:05 dalek rakudo/nom:
21:05 dalek rakudo/nom: MUST now specify "use nqp" in scope to be allowed to use nqp:: functions
21:05 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6b1b3ef068
21:06 lizmat wonder how much ecosystem fallout this will cause ^^^
21:06 timotimo time perl6 -e 'my @dwarves = ^7; my int $i; my int $elems; until $i++ == 10_000_000 { $elems = $elems + @dwarves.elems }; say $elems'
21:06 timotimo with this, the shaped version actually wins out by 3x
21:06 FreezerburnV Juerd: Does that allow for taking any number of arguments via a splat or something?
21:07 Juerd FreezerburnV: Sure. It's a sub like any other. Just make sure that the empty one takes the same thing :)
21:07 TimToady timotimo: that'll be a bit faster if you use ++$i there instead
21:07 * brrt has got a failing testcase on master/master/nom t/01-sanity/55-use-trace.t
21:07 masak pre-increment optimization is the root of all evil! :P
21:07 TimToady since postinc in non-sink has to subtract the 1 back out
21:07 timotimo fair enough, but the [7] is the only part that i'm changing ...
21:07 timotimo is the benchmark flawed in some other way?
21:08 TimToady well, it's just a little more overhead
21:09 Juerd TimToady: How do you feel about introducing a CStr type (subset CStr of Str where * !~~ /\0/) and limiting everything that will eventually go through a system call that doesn't support \0 to CStr instead of Str?
21:09 brrt this is because DISLOCATED MESSAGE] Useless use of &infix:<-> in sink context appears in the output
21:09 brrt bad test? just a random warning?
21:10 FreezerburnV Juerd: That works perfectly, thanks again for the help!
21:10 Juerd FreezerburnV: You're welcome. Have fun :)
21:10 Juerd FreezerburnV: Also thank grondilu btw :P
21:11 timotimo seems to maybe be an inlining thing
21:11 FreezerburnV Juerd: Having tons of fun already :) Perl6 is an interesting language that feels very alien from most languages I’m comfortable with, so it’s a nice getaway from those and fun for stretching the brain muscles
21:11 timotimo at least in the one log the "<unit>" shows up before "elems", in the other log it's the other way around
21:12 FreezerburnV grondilu: Thanks for the help as well! Sorry I didn’t send you a thanks earlier. Between stuff going on IRL, getting things working in my code, and reading IRC stuff is just crazy for me right now
21:12 TimToady brrt: does trace insert instructions?
21:12 brrt i have... no idea whatever
21:12 Juerd FreezerburnV: Let us know once you've gotten comfortable with Perl 6. We can get you the professional help you'll need ;)
21:12 brrt possibly, yes
21:12 TimToady that could certainly influence whether something thought it was in sink context
21:13 timotimo TimToady: switching $i++ and ++$i around doesn't make a noticable difference
21:13 Skarsnik how do you call configure.pl in a fork?
21:13 TimToady timotimo: not too surprised, it's small compared to what you're measuring
21:13 Juerd Skarsnik: Forked repository or forked process?
21:13 FreezerburnV Juerd: Heh. Might be a bit before I’m comfortable with it :P Going is pretty slow for what I’m working on, but still fun
21:13 Skarsnik repo
21:14 Juerd Skarsnik: Just like you would in the original?
21:14 dalek roast: 84a00b5 | lizmat++ | S02-types/WHICH.t:
21:14 dalek roast: Add basic tests for X::NQP::NotFound
21:14 dalek roast: review: https://github.com/perl6/roast/commit/84a00b5a9e
21:14 timotimo when i drop the $elems = $elems + part and my int $elems, the numbers change significantly:
21:15 timotimo with shape goes from 5.37 to 2.36
21:15 timotimo without shape goes from 15.30 to 11.37
21:16 timotimo so ... it's not about how the individual parts are optimized in the core setting code ... it's about when spesh decides to kick in
21:16 nine Funny: without a use lib a require '/full/path/to/file/name.pm'; won't work. But it doesn't matter which lib we use, because all we need is any FileSystem repository.
21:16 timotimo can't do much about that :(
21:17 TimToady m: use trace; $_ -= 1
21:17 camelia rakudo-moar 6b1b3e: OUTPUT«[DISLOCATED MESSAGE] Useless use of &infix:<-> in sink context␤2 (/tmp/zawbZm8sYr:1)␤$_ -= 1␤»
21:17 TimToady pretty easy to reproduce, anyway :)
21:17 brrt aye
21:18 TimToady DISLOCATED basically means some ast node didn't set :node($/) somewhere
21:19 cdg joined #perl6
21:21 smls_ Who came up with the name for this class http://doc.perl6.org/type/X::Syntax::AddCategorial::MissingSeparator
21:21 smls_ As a heading it barely fits on the page... ;)
21:22 timotimo *sigh* the + operator doesn't get inlined in the unshaped case. probably because of getlexref :|
21:22 masak smls_: must've been a Java programmer!
21:22 TimToady and it's misspelled
21:22 timotimo should be "categorical"?
21:23 TimToady yup
21:23 diakopter Categorial is a thing
21:23 * TimToady denies it categorially
21:23 timotimo we'll write that on your categorial when you've been laid to rest
21:23 diakopter https://en.wikipedia.org/wiki/Categorial_grammar
21:24 * masak .oO( to lady to rest )
21:24 zjmarlow joined #perl6
21:24 timotimo how is cattegory formed
21:24 masak I was wondering whether someone would pick up on that reference :>
21:24 jnthn a separator in arr
21:25 dalek nqp: 66e2e9d | jnthn++ | src/vm/moar/QAST/QAST (3 files):
21:25 dalek nqp: Prepare for better handling of unsinged ints.
21:25 dalek nqp: review: https://github.com/perl6/nqp/commit/66e2e9d76a
21:25 timotimo who had missed her three syntax
21:25 masak they need to do way instain "categorial"
21:27 diakopter unsinged, love it
21:27 dalek rakudo/nom: 90d7334 | lizmat++ | src/core/Temporal.pm:
21:27 dalek rakudo/nom: Make Date.new illegal
21:27 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/90d73348c1
21:27 geekosaur unhinged ints?
21:27 jnthn Apparently, my keyboard is toast :P
21:28 * bartolin_ looks up commit b249312b18 and blushes
21:28 lizmat .oO( Game Board Game Board Game Board )
21:29 lizmat .oO( which repo would bartolin_ mean )
21:29 bartolin_ the doc repo you just spoke about :-)
21:30 El_Che Today at work I got a visit from the Elastic Search guy that wanted help on his perl 5 ldap code (I didn't even know he wrote Perl), and from a python guy curious about perl6. Perl day \o/ :)
21:30 diakopter where did the python guy hear about it
21:30 masak El_Che: you should have them meet up, and they might turn into an Elastic Python!
21:31 masak diakopter: during the two minute hate, probably
21:32 El_Che diakopter: a long time a got I old him that perl6 was turning out very nicely. The guy got curious it seems and recently started looking at the docs. He loved the concurrency
21:32 masak ooh
21:33 El_Che a lot will depends on the library ecosystem I guess
21:35 * masak subscribes to nine++'s oddly controversial line "Perl 6 has all of CPAN!"
21:35 El_Che Inline?
21:35 masak aye
21:36 lizmat well, all of is still not true
21:36 El_Che I have a confession to make
21:36 lizmat I understand that Class::DBIx still has issues
21:36 diakopter well, someone could write a harness to run all of CPAN's test suites from within p6 code
21:36 El_Che I haven't tried it yet :)
21:36 El_Che concentrated on the new toys
21:36 masak lizmat: that's a good point.
21:36 masak El_Che: I think the Perl 6 community as a whole could make that confession.
21:37 masak El_Che: but it still seems to me that Inline::Perl5 is the natural conclusion of pragmatically wanting to use Perl 6 in production, today
21:38 Juerd The primary reason I'm not using Perl 6 in production, is that for the one thing I wrote in Perl 6, the memory usage is enormous
21:38 timotimo Juerd: are you using biiiig strings or biiiiig ints?
21:38 Juerd That is, using 500+ MB where a Perl 5 equivalent stays within 8 MB
21:38 timotimo or maybe biiiiiig bufs/blobs?
21:38 raiph joined #perl6
21:38 Juerd timotimo: Nope, lots of tiny ones that keep getting thrown away. So it might be leaky.
21:39 timotimo how tiny?
21:39 Juerd timotimo: Typically, under 100 bytes.
21:39 timotimo oh, mhm, ok
21:39 Juerd timotimo: It's the MQTT thing
21:39 timotimo ah
21:39 Juerd That I'm contemplating to rename from Net::MQTT to MQTT::Client. Any opinions on that?
21:39 timotimo the thing is we allocate those with malloc but we don't apply pressure to the GC, so it won't collect earlier
21:39 timotimo that sounds nice
21:40 Skarsnik *build rakudo with 100Mb of free space*
21:40 RabidGravy Juerd, sounds perfect
21:41 Skarsnik It seems unliky to work x)
21:41 Juerd Thanks for your feedback. Will rename now.
21:41 dalek rakudo/nom: bca46c1 | lizmat++ | src/core/IO/Socket/INET.pm:
21:41 dalek rakudo/nom: Remove IO::Socket::Inet.input-line-separator
21:41 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bca46c1f26
21:41 Juerd And after that I'll submit it to the ecosystem. I've only been putting that off because I was unsure about the name.
21:44 nadim joined #perl6
21:45 Skarsnik closer to 500 for x-mas then x)
21:45 lichtkind <prior> is gone?
21:45 Juerd Oh my, something got slow. The startup overhead for my one test script has gotten noticeable.
21:45 dalek rakudo/nom: 46c7659 | lizmat++ | src/core/IO/Socket/Async.pm:
21:45 dalek rakudo/nom: Remove IO::Socket::Async.(bytes|chars)-supply
21:45 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/46c76596e4
21:45 rindolf joined #perl6
21:46 nine lizmat: what issues?
21:46 lizmat I was discussing this with ribasushi at the LPW
21:46 masak lichtkind: yes, since bbebb37, December 2013. two years.
21:46 El_Che saw ribasushi declaration video recently :(
21:46 frew left #perl6
21:46 lizmat apparently some ways that Class::DBIx is generating packages, can't be mimicked on the Perl 6 side ?
21:46 lichtkind masak thanks you know i barely pay attention :)
21:47 masak lichtkind: that's not true :)
21:47 masak lichtkind: but <prior> is such a small thing, too
21:47 masak lichtkind: "a low-gain feature", TimToady says in the commit comment as he removes it
21:47 lizmat .oO( on PerlMonks, prior is a lot of work )
21:48 masak lizmat: it would've been in Perl 6 too. both for the implementor and the implementation :P
21:48 masak "match whatever the prior successful regex matched"
21:48 masak sounds like something I wouldn't mind settin gup myself if it came to that
21:48 dalek tablets: 17a65f2 | (Herbert Breunung)++ | docs/ (2 files):
21:48 dalek tablets: -prior
21:48 dalek tablets: review: https://github.com/perl6/tablets/commit/17a65f2d23
21:49 nine lizmat: well in August I had it fully functional: http://niner.name/talks/Building Bridges/examples/dbic_demo_p6.html http://niner.name/talks/Building Bridges/examples/dbic_demo_cds.html
21:49 masak actually, it's a bit under-specified, too. does it match the last successful thing as a *constant string*, or as a *regex*?
21:50 lizmat nine: I can only hope that ribasushi will chime in
21:50 masak that is, if the last successful regex was /foo.../, and it had matched successfully against "fools!", would <prior> then match successfully against "footie"?
21:52 * ribasushi reads up....
21:52 ribasushi nine: last you told me you couldn't do "p6 class visible to p5"
21:52 ribasushi nine: only "p5 class visible to p6 land"
21:52 ribasushi i.e. the mapping was somewhat one-way, once you are in p6 - you can't "leave"
21:53 ribasushi perhaps I misunderstood?
21:54 zjmarlow joined #perl6
21:55 dalek doc: 57cd2de | usev6++ | doc/Type/X/Syntax/AddCategorial/MissingSeparator.pod:
21:55 dalek doc: Remove misspelled exception
21:55 dalek doc:
21:55 dalek doc: it is named 'X::Syntax::AddCategorical::TooFewParts' now
21:55 dalek doc: smls++ for bringing this up
21:55 dalek doc: review: https://github.com/perl6/doc/commit/57cd2dec30
21:55 ribasushi basically DBICs model is built upon "plug subclasses at the proper override points"
21:55 ribasushi hence having p6 subclasses that are driven by the p5-side "core"
21:55 ribasushi is something quite useful
21:56 jnthn 'night, #perl6
21:56 ribasushi but the basic "p5-side core inflates to p6 objects" is something that nine did show working yes
21:58 ribasushi o/
21:59 pierre-vigier joined #perl6
22:01 dalek roast: 2a7596f | TimToady++ | S02-literals/radix.t:
22:01 dalek roast: correct exception type
22:01 dalek roast: review: https://github.com/perl6/roast/commit/2a7596fb62
22:04 nine ribasushi: at least all of DBIC's features that I know work :) Which is probably not everything. But if anything is missing, a determined user can probably still make them work.
22:04 Zoffix joined #perl6
22:04 dalek rakudo/nom: 77fc44d | lizmat++ | src/core/IO/Handle.pm:
22:04 dalek rakudo/nom: Remove IO::Handle.(ins|nl)
22:04 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/77fc44d9c0
22:04 dalek rakudo/nom: 6fccc52 | lizmat++ | src/core/IO/Handle.pm:
22:04 dalek rakudo/nom: Remove IO::Handle.seek(pos,Int:D whence)
22:04 dalek rakudo/nom:
22:04 dalek rakudo/nom: Must now specify a SeekType enum:
22:04 dalek rakudo/nom:  - SeekFromBeginning
22:04 dalek rakudo/nom:  - SeekFromCurrent
22:04 dalek rakudo/nom:  - SeekFromEnd
22:04 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6fccc52532
22:04 nine ribasushi: most of my work lately has been "make it pretty", not "make it work at all".
22:04 nine (lots of "work" in there obviously)
22:05 lizmat [Coke]: all DEPRECATED code is now removed from the setting
22:05 ribasushi nine: then I think I'm not up to date, as last conversation left me with the impression that a p6 class is not really visible/instantiable in p5 land
22:05 * lizmat stops wielding the chainsaw and goes to bed&
22:05 * ribasushi will check things some time early next year
22:05 ribasushi lizmat: ^^ note last 2 lines
22:06 Juerd Hm, is there any reason why SeekFromBeginning isn't the default whence?
22:06 dalek rakudo/nom: 54f6edf | TimToady++ | src/Perl6/Actions.nqp:
22:06 dalek rakudo/nom: METAOP_ASSIGN needs to want its op in sink context
22:06 dalek rakudo/nom:
22:06 dalek rakudo/nom: This fixes the 'use trace; $a -= 1;' problem.
22:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/54f6edffec
22:06 dalek rakudo/nom: 694a4b4 | TimToady++ | src/Perl6/Grammar.nqp:
22:06 dalek rakudo/nom: don't mention 'detached' if it wasn't
22:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/694a4b4807
22:06 lizmat Juerd: good point!
22:07 Juerd Especially with those constants being long, not having to type them in the common case would be nice :)
22:07 zengargoyle joined #perl6
22:07 Zoffix lizmat++ # fixing things
22:07 Juerd And having to type them in the uncommon case is even better because I keep forgetting what those magic values to seek are. I can only remember 0. :)
22:08 Zoffix Hmmm.... my note from work sent to self says "Fix PR" :/
22:08 Zoffix No idea what that was supposed to be about :(
22:08 TimToady fix our public relations?
22:08 dalek rakudo/nom: ae3a443 | lizmat++ | src/core/IO/Handle.pm:
22:08 dalek rakudo/nom: Oops, forgot to default to SeekFromBeginning
22:08 dalek rakudo/nom:
22:08 dalek rakudo/nom: Juerd++ for spotting
22:08 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ae3a4436ef
22:08 Zoffix heh
22:08 Juerd \o/
22:09 lizmat sleep&
22:09 nine ribasushi: with v6::invoke('Any::Perl6::Class', 'new') you can get at an instance of any Perl6 class from Perl5. Of course, creating wrapper Perl5 packages will make the integration more seamless. There's always ways to improve :)
22:09 Juerd Welterusten, lizmat :)
22:09 Zoffix Ahhh, it was about grondilu++ telling me a better way to do the combinations stuff :D And they fixed it for me :)
22:09 nine Looking forward to having more time for that after the big release :)
22:09 nine Good night all!
22:09 Zoffix night
22:13 timotimo gnite nine
22:14 TimToady m: use trace; $_ -= 1
22:14 camelia rakudo-moar 694a4b: OUTPUT«2 (/tmp/2W2WRG63gF:1)␤$_ -= 1␤»
22:15 TimToady m: 0b1.1e10
22:15 camelia rakudo-moar 694a4b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/xAHZ5_imHE␤Malformed postfix call␤at /tmp/xAHZ5_imHE:1␤------> 0b1.7⏏1e10␤»
22:15 TimToady m: 0b1 .1e10
22:15 camelia rakudo-moar 694a4b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/xP9qDXdln7␤Malformed postfix call (only alphabetic methods may be detached)␤at /tmp/xP9qDXdln7:1␤------> 0b1 .7⏏1e10␤»
22:15 TimToady m: 0b1 . 1e10
22:15 camelia rakudo-moar 694a4b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/q5Oe_Bv4nt␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/q5Oe_Bv4nt:1␤------> 0b1 .7⏏5 1e10␤»
22:17 travis-ci joined #perl6
22:17 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Remove IO::Socket::Inet.input-line-separator'
22:17 travis-ci https://travis-ci.org/rakudo/rakudo/builds/97549312 https://github.com/rakudo/rakudo/compare/90d73348c158...bca46c1f2681
22:17 travis-ci left #perl6
22:19 Zoffix "src/6model/reprconv.c:553:6: error: conflicting types for ‘MVM_repr_set_int’    src/6model/reprconv.h:93:17: note: previous declaration of ‘MVM_repr_set_int’ was here    make: *** [src/6model/reprconv.o] Error 1"
22:19 Zoffix from the above fail
22:19 timotimo yeah, already fixed
22:19 Zoffix Sweet
22:20 timotimo we just had the "luck" of using a moar version with the bug in one of the builds
22:22 Skarsnik dumb question: How I check if a float (Rat?) has a decimal?
22:22 Juerd Skarsnik: Most of them have an infinite number of "0" decimals.
22:22 RabidGravy Think there needs to be some documentation fixed
22:22 Juerd Well, not most of them, an infinite number of them do, just as an infinite number don't, so make that s/most/many/.
22:23 Skarsnik yeah but if I have 1 and 1.5 how I see that 1.5 is not a integer?
22:23 Juerd Skarsnik: But, you probably want to know if it could pass as an integer without losing any of its value, right? :)
22:23 Kyle joined #perl6
22:23 * vendethiel is happy with how christmas is shaping up
22:23 Juerd Skarsnik: Try to round-trip it: $num.Int == $num
22:23 vendethiel (my boss still want no code generation scripts in the codebase, but little does he know... :P)
22:24 zjmarlow joined #perl6
22:24 TimToady m: say 1.0.base-repeating; say 1.5.base-repeating; say (1+⅐).base-repeating
22:24 camelia rakudo-moar 694a4b: OUTPUT«(1 )␤(1.5 )␤(1. 142857)␤»
22:24 geraud joined #perl6
22:25 Skarsnik m: say 1.5.denominateur > 0
22:25 camelia rakudo-moar 694a4b: OUTPUT«Method 'denominateur' not found for invocant of class 'Rat'␤  in block <unit> at /tmp/CwCKvywjvi:1␤␤»
22:25 TimToady m: say 1.5.floor == 1
22:25 camelia rakudo-moar 694a4b: OUTPUT«True␤»
22:25 Skarsnik m: say 1.5.denominator > 0
22:25 camelia rakudo-moar 694a4b: OUTPUT«True␤»
22:25 [Tux] lizmat, as per that gist-paste between 14 Dec and today: in between I was not able to use panda
22:25 TimToady m: say 1.5.floor == 1.5
22:25 camelia rakudo-moar 694a4b: OUTPUT«False␤»
22:25 Skarsnik m: say 1.0.denominator > 0
22:25 camelia rakudo-moar 694a4b: OUTPUT«True␤»
22:25 * [Tux] => bed
22:25 Skarsnik m: say 1.0.numerator > 0
22:25 camelia rakudo-moar 694a4b: OUTPUT«True␤»
22:26 Skarsnik m: say 1.0.perl
22:26 camelia rakudo-moar 694a4b: OUTPUT«1.0␤»
22:26 Skarsnik m: say 1.0.gist
22:26 camelia rakudo-moar 694a4b: OUTPUT«1␤»
22:26 Juerd m: say 1.0.1.WHAT
22:26 camelia rakudo-moar 694a4b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ZIq6I5NDYB␤Malformed postfix call␤at /tmp/ZIq6I5NDYB:1␤------> 3say 1.0.7⏏1.WHAT␤»
22:26 Juerd :)
22:26 Juerd I'll miss writing literal IP addresses :)
22:27 Zoffix m: say -1.5.floor == -1
22:27 camelia rakudo-moar 694a4b: OUTPUT«True␤»
22:27 * TimToady has a soft spot in his heart (or maybe in his head) for v-strings
22:28 Juerd m: say (v1.0.1).WHAT
22:28 camelia rakudo-moar 694a4b: OUTPUT«(Version)␤»
22:28 Juerd m: say (v1.0.1).Str
22:28 camelia rakudo-moar 694a4b: OUTPUT«1.0.1␤»
22:28 Skarsnik m: say 1.1.WHAT
22:28 camelia rakudo-moar 694a4b: OUTPUT«(Rat)␤»
22:28 Juerd m: say (v1.0.1).Blob
22:28 camelia rakudo-moar 694a4b: OUTPUT«Method 'Blob' not found for invocant of class 'Version'␤  in block <unit> at /tmp/AKAO2ptOBU:1␤␤»
22:28 Juerd m: say (v1.0.1).Buf  # then?
22:28 camelia rakudo-moar 694a4b: OUTPUT«Method 'Buf' not found for invocant of class 'Version'␤  in block <unit> at /tmp/xFVGDI1MzB:1␤␤»
22:28 Juerd Aw
22:28 Juerd No v127.0.0.1 then :)
22:28 TimToady nope :D
22:28 Skarsnik m: say 1.2.denominator; say 1.2.numerator;
22:28 camelia rakudo-moar 694a4b: OUTPUT«5␤6␤»
22:29 Zoffix m: say 1.0.denominator; say 1.0.numerator
22:29 camelia rakudo-moar 694a4b: OUTPUT«1␤1␤»
22:29 Zoffix m: say 6.0.denominator; say 6.0.numerator
22:29 camelia rakudo-moar 694a4b: OUTPUT«1␤6␤»
22:29 Zoffix m: say 6.0.denominator == True+0
22:29 camelia rakudo-moar 694a4b: OUTPUT«True␤»
22:29 Zoffix :D
22:30 Zoffix m: say 6.0.denominator == True
22:30 camelia rakudo-moar 694a4b: OUTPUT«True␤»
22:30 Zoffix m: say 6.0.numerator == True
22:30 camelia rakudo-moar 694a4b: OUTPUT«False␤»
22:31 Juerd You have a fancy way of saying 1
22:32 Zoffix Yeah :)
22:34 TimToady m: say -(e ** (-i * pi))
22:34 camelia rakudo-moar ae3a44: OUTPUT«1+1.22464679914735e-16i␤»
22:34 TimToady m: say -(e ** (-i * pi)) == 1
22:34 camelia rakudo-moar ae3a44: OUTPUT«False␤»
22:34 TimToady m: say -(e ** (-i * pi)) =~= 1
22:34 camelia rakudo-moar ae3a44: OUTPUT«True␤»
22:34 TimToady m: say -(e ** (-i * pi)) <=> 1
22:34 camelia rakudo-moar ae3a44: OUTPUT«Same␤»
22:35 Zoffix *sigh* I don't get why the perl6.org recent post updater doesn't run unless you push some commits :S https://github.com/perl6/perl6.org/blob/master/update.sh#L11
22:35 TimToady m: say -(e ** (-i * pi)).Num == 1
22:35 camelia rakudo-moar ae3a44: OUTPUT«True␤»
22:35 lichtkind good night
22:35 Zoffix night
22:36 Juerd m: say e ** (i * τ) ≅ 1
22:36 camelia rakudo-moar ae3a44: OUTPUT«True␤»
22:36 Juerd Silly minuses ;)
22:37 dalek perl6.org: 65d2982 | (Zoffix Znet)++ | README.md:
22:37 dalek perl6.org: Linkify #perl6 in the README
22:37 dalek perl6.org: review: https://github.com/perl6/perl6.org/commit/65d298253d
22:37 TimToady m: say e ** τ\i ≅ 1
22:37 camelia rakudo-moar ae3a44: OUTPUT«True␤»
22:37 Zoffix :o
22:37 Juerd Wh...what's an, er... "\"?
22:38 Zoffix m: say e ** τ*i ≅ 1
22:38 camelia rakudo-moar ae3a44: OUTPUT«False␤»
22:38 Zoffix m: say e ** τ/i ≅ 1
22:38 camelia rakudo-moar ae3a44: OUTPUT«False␤»
22:38 masak Juerd: it's just an unspace
22:38 TimToady it's a degenerate unspace
22:38 Zoffix m: say e ** (τ*i) ≅ 1
22:38 camelia rakudo-moar ae3a44: OUTPUT«True␤»
22:38 TimToady m: say e ** (τ)i ≅ 1
22:38 camelia rakudo-moar ae3a44: OUTPUT«True␤»
22:38 Zoffix whoa you can do that!?
22:38 masak m: say "foo".comb\      [1]
22:38 camelia rakudo-moar ae3a44: OUTPUT«o␤»
22:38 masak m: say "foo".comb\[1]
22:38 camelia rakudo-moar ae3a44: OUTPUT«o␤»
22:38 Zoffix man, Perl 6 is pretty amazing on DWIMming
22:39 TimToady it's just a postfix, except it's alphabetic
22:39 Juerd So two terms in a row are bad, but with unspace in between it's multiplication? Am I understanding this correctly?
22:39 TimToady 2nd is a postfix, not a term
22:39 Zoffix Yes
22:39 Zoffix Oh?
22:39 TimToady it's only that postfix:<i> implies multiplication
22:40 Zoffix m: say τ\e
22:40 camelia rakudo-moar ae3a44: OUTPUT«5===SORRY!5=== Error while compiling /tmp/btAkrTgHmV␤Confused␤at /tmp/btAkrTgHmV:1␤------> 3say τ\7⏏5e␤    expecting any of:␤        postfix␤»
22:40 Juerd Oh, i is a postfix thing :)
22:40 Zoffix ahhh :(
22:40 * Juerd wipes sweat from forehead
22:40 Zoffix m: say (τ)e
22:40 camelia rakudo-moar ae3a44: OUTPUT«5===SORRY!5=== Error while compiling /tmp/lhPH3A1Rud␤Two terms in a row␤at /tmp/lhPH3A1Rud:1␤------> 3say (τ)7⏏5e␤    expecting any of:␤        infix␤        infix stopper␤        postfix␤        statement end␤        statement…»
22:40 * Zoffix is less amazed now
22:40 TimToady m: say -(e ** (-i × pi)).Num == 1
22:40 camelia rakudo-moar ae3a44: OUTPUT«True␤»
22:41 Zoffix .u minus
22:41 yoleaux U+002D HYPHEN-MINUS [Pd] (-)
22:41 yoleaux U+00B1 PLUS-MINUS SIGN [Sm] (±)
22:41 yoleaux U+02D7 MODIFIER LETTER MINUS SIGN [Sk] (˗)
22:41 TimToady we don't do that one yet
22:41 Zoffix .u minus sign
22:41 yoleaux U+00B1 PLUS-MINUS SIGN [Sm] (±)
22:41 yoleaux U+02D7 MODIFIER LETTER MINUS SIGN [Sk] (˗)
22:41 yoleaux U+0320 COMBINING MINUS SIGN BELOW [Mn] (◌̠)
22:41 Juerd Zoffix: I'm more comfortable with this, though, because having every term work as a postfix would be really weird and confusing.
22:41 Skarsnik I am confused how rakudo/moar saw that I give a named parameter in DateTime.Str under a condition when the spec tests try to set his own formatter https://gist.github.com/Skarsnik/5d33848627fe68400758
22:41 El_Che good night. Keep up the good work!
22:41 TimToady .u −
22:41 yoleaux U+2212 MINUS SIGN [Sm] (−)
22:41 Zoffix That's the one yeah :D
22:41 Juerd Oooh, does ± create an any() yet? :P
22:41 rurban joined #perl6
22:42 TimToady it would probably just create a range, or add error bars, or something
22:42 Juerd Hehe
22:42 TimToady it's not a single concept
22:42 TimToady so we'll leave that one up to the programmer
22:43 Juerd I don't know if it's common internationally because outside of Dutch I usually see ~ used for it, but at least in Dutch ± is used as 'approximately' (circa)
22:45 DrForr Thats also tolerance in engineering. (+/- as a single glyph.)
22:45 DrForr *That's
22:45 castelo joined #perl6
22:45 TimToady .tell lucasb I used || False because I was adding new lines all the time, and kept getting syntax errors, but they can probably come out now
22:45 yoleaux TimToady: I'll pass your message to lucasb.
22:48 TimToady and you can tell I'm still about 24 hours behind on the backlog...
22:48 Zoffix Ah. It runs, but it's mowyw that actually copies the file to the web dir. RE: [17:35:02] <Zoffix> *sigh* I don't get why the perl6.org recent post updater doesn't run
22:48 * Zoffix will fix it when soberer
22:50 travis-ci joined #perl6
22:50 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Remove IO::Handle.seek(pos,Int:D whence)
22:50 travis-ci https://travis-ci.org/rakudo/rakudo/builds/97553759 https://github.com/rakudo/rakudo/compare/46c76596e4b8...6fccc52532d2
22:50 travis-ci left #perl6
22:50 masak travis-ci: do you have to be so formal and name someone by their whole name when the build breaks? :)
22:50 Zoffix :)\
22:50 masak oh, ze already left.
22:51 TimToady if you're really in trouble, it puts in your middle name too
22:51 Zoffix heh
22:51 Juerd masak: It's like getting reprimanded by parents. Suddenly even your middle name counts.
22:51 Juerd TimToady: Yes, like that :)
22:51 masak yeah, but we try to keep it informal in here
22:51 masak like, "ey! lizmat! the build broke!"
22:51 castelo joined #perl6
22:51 masak see the difference? :)
22:52 Juerd I don't think travis-ci has that mode.
22:52 castelo_ joined #perl6
22:52 masak maybe something to think about for travis-ci 2.0...
22:52 Juerd 6, perhaps :)
22:52 Skarsnik I am trying to change Temporal.pm to support subsecond in ISO something timestamp (update of 2005 according to wikipedia), the default formatter has a named parameter to display them. (subseconds) but you can change the formatter. Should I test if the current formatter is the default?
22:52 castelo joined #perl6
22:53 Skarsnik In the Str method
22:53 flussence if you change the code's behaviour and nothing breaks, it needed more tests in the first place :)
22:53 Skarsnik tests break actually :)
22:54 RabidGravy boo
22:54 zjmarlow joined #perl6
22:54 Skarsnik Because they give a formatter that does not support this named parameter
22:55 castelo left #perl6
22:56 Zoffix Inline::Perl5 still fails on my box
22:56 Zoffix t/use.t (Wstat: 11 Tests: 4 Failed: 0)  Non-zero wait status: 11
22:56 Zoffix With Rakudo freshly squeezed with rakudobrew (maybe that's not recent enough tho)
22:57 Zoffix This is Rakudo version 2015.11-634-gae3a443 built on MoarVM version 2015.11-67-g278b18a
22:57 Skarsnik 15.11 ?
22:58 Zoffix Skarsnik, there hasn't been a .12 release yet, IIUC. It's the Christmas release
22:59 Skarsnik Oh I thought it get tagged with the commit date
23:00 * Zoffix runs TEST_JOBS=100 make spectest
23:01 Zoffix :) Would be nice to have a 100-core box to play with
23:01 Skarsnik xD
23:02 Skarsnik m: sub foo() {}; my &m = &foo; say &m eq &foo
23:02 camelia rakudo-moar ae3a44: OUTPUT«Sub object coerced to string (please use .gist or .perl to do that)  in block <unit> at /tmp/rr3xZH8ezq:1␤Sub object coerced to string (please use .gist or .perl to do that)  in block <unit> at /tmp/rr3xZH8ezq:1␤True␤»
23:02 Skarsnik How I do that without having a warning? x)
23:02 Zoffix tm: sub foo() {}; my &m = &foo; say &m === &foo
23:02 Zoffix m: sub foo() {}; my &m = &foo; say &m === &foo
23:02 camelia rakudo-moar ae3a44: OUTPUT«True␤»
23:03 Skarsnik Oh thx
23:03 Skarsnik There is too much operator!
23:04 Juerd Wow
23:04 travis-ci joined #perl6
23:04 travis-ci Rakudo build failed. TimToady 'don't mention 'detached' if it wasn't'
23:04 travis-ci https://travis-ci.org/rakudo/rakudo/builds/97554181 https://github.com/rakudo/rakudo/compare/6fccc52532d2...694a4b48074e
23:04 travis-ci left #perl6
23:04 TimToady don't blame me!
23:04 Zoffix :D
23:04 TimToady you've got CPU, do your own bisect!
23:06 Zoffix 6m33s run and I've seen RAM usage go up to 8.3GB. Here are the failures FWTW: https://gist.github.com/zoffixznet/4c15222fddf062958e9d
23:06 * Zoffix does TEST_JOBS=200 just for fun
23:07 Juerd How does one run the test suite?
23:07 Juerd So far I've only used individual test files manually
23:07 Zoffix 10.9GB ram :)
23:08 Zoffix Juerd, git clone https://github.com/rakudo/rakudo/ .; perl Configure.pl --gen-moar --gen-nqp --backends=moar; make; make install; make spectest
23:08 Skarsnik yay all test passed. Do I need something special to add something to roast or can I just commit like that? Or wait until my change in rakudo is merged?
23:08 Juerd Ah, make spectest! Thanks :)
23:09 Skarsnik I can't are and is today x)
23:09 pierre-vigier joined #perl6
23:09 lucasb joined #perl6
23:10 Juerd Skarsnik: If it's not yet merged, I'd send a pull request and refer to the PR in rakudo, and vice versa.
23:11 Juerd Although the back reference is added automatically
23:11 Skarsnik I love forking everything!
23:11 Juerd We all forking love it.
23:11 Skarsnik Yes if you mention in one, the other get noticed
23:12 * zengargoyle was hoping .IO.Handle.seek got slicker names than SeekFromBeginning, etc. :(
23:13 Zoffix TEST_JOBS=200 completed in 6m10s and RAM usage peaked at 13.5GB
23:13 * Zoffix needs moar cores
23:13 Zoffix zengargoyle, me too :(
23:14 Zoffix SeekFromBeginning makes me think "Java"
23:14 Juerd No, that would be seekFromBeginning
23:15 Zoffix heh
23:15 lucasb b2gills showed a bug in the backlog about [[-1]]. anyone knows why is that? how is it getting parsed?
23:15 yoleaux 22:45Z <TimToady> lucasb: I used || False because I was adding new lines all the time, and kept getting syntax errors, but they can probably come out now
23:16 lucasb ah, ok. thanks TimToady++
23:16 lucasb m: [[ -1]]
23:16 camelia rakudo-moar ae3a44: ( no output )
23:16 lucasb m: [[-1]]
23:16 camelia rakudo-moar ae3a44: OUTPUT«5===SORRY!5=== Error while compiling /tmp/OEAYTjR7HQ␤Unable to parse expression in bracketed infix; couldn't find final ']' ␤at /tmp/OEAYTjR7HQ:1␤------> 3[[-7⏏1]]␤»
23:17 stmuk_ joined #perl6
23:17 travis-ci joined #perl6
23:17 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Oops, forgot to default to SeekFromBeginning
23:17 travis-ci https://travis-ci.org/rakudo/rakudo/builds/97554562 https://github.com/rakudo/rakudo/compare/694a4b48074e...ae3a4436ef24
23:17 travis-ci left #perl6
23:19 Zoffix m: ([-1])
23:19 camelia rakudo-moar ae3a44: OUTPUT«WARNINGS for /tmp/20UMFgibnx:␤Useless use of "-" in expression "-1" in sink context (line 1)␤»
23:19 Zoffix m: [[\-1]]
23:19 camelia rakudo-moar ae3a44: ( no output )
23:20 Skarsnik Good night #perl6
23:20 Zoffix night
23:20 lucasb m: say (\-1).perl  # that one is interesting too Zoffix++. Maybe it is already known
23:20 camelia rakudo-moar ae3a44: OUTPUT«\(1)␤»
23:21 Zoffix wtf is \(...) ?
23:21 sftp joined #perl6
23:22 lucasb capture syntax
23:22 Zoffix ah
23:22 lucasb same as Capture.new(...)
23:22 sno joined #perl6
23:24 zjmarlow joined #perl6
23:24 Zoffix My guess it's trying to parse it as [-] operator and freaks out when it sees something other than ] after the minus
23:25 lucasb yes, it makes sense
23:25 Zoffix Though I may be stating the obvious :D
23:35 TimToady testing a fix
23:35 dalek ecosystem: 8fe42bb | (Juerd Waalboer)++ | META.list:
23:35 dalek ecosystem: Add MQTT::Client to ecosystem
23:35 dalek ecosystem: review: https://github.com/perl6/ecosystem/commit/8fe42bb052
23:36 kid51 joined #perl6
23:44 cognominal I don't get this one in Stringy.pm : multi sub infix:<eq>($x?)          { Bool::True }
23:44 cognominal are not infix supposed to have exactly two paremeters ?
23:45 rurban_ joined #perl6
23:45 timotimo m: say [eq] "Hi"
23:45 camelia rakudo-moar ae3a44: OUTPUT«True␤»
23:45 timotimo m: say [eq]
23:45 camelia rakudo-moar ae3a44: OUTPUT«True␤»
23:45 virtualsue joined #perl6
23:45 cognominal timotimo++
23:48 cognominal good to know, and doing so you answer another not yet asked question about neutral elements :)
23:54 zjmarlow joined #perl6
23:57 flussence m: say [**] # also this
23:57 camelia rakudo-moar ae3a44: OUTPUT«1␤»
23:57 flussence (and other Nummy things)
23:58 stmuk joined #perl6
23:59 edehont joined #perl6
23:59 dalek roast: 775039b | TimToady++ | S02-literals/char-by-number.t:
23:59 dalek roast: overspecific error type
23:59 dalek roast: review: https://github.com/perl6/roast/commit/775039bcd1
23:59 dalek roast: 363b7f1 | TimToady++ | S03-metaops/reduce.t:
23:59 dalek roast: test that [-1] and such parse right
23:59 dalek roast: review: https://github.com/perl6/roast/commit/363b7f1845
23:59 dalek rakudo/nom: f262b27 | TimToady++ | src/Perl6/Actions.nqp:
23:59 dalek rakudo/nom: remove temporary False values
23:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f262b2704d
23:59 dalek rakudo/nom: f0d3485 | TimToady++ | src/Perl6/Grammar.nqp:
23:59 dalek rakudo/nom: better redirect on /\1/ and such
23:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f0d3485538
23:59 dalek rakudo/nom: 03b52f9 | TimToady++ | src/Perl6/Grammar.nqp:
23:59 dalek rakudo/nom: don't treat things like [-1] as reductions
23:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/03b52f9f83

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo