»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend! | tinyurl.com/p6contest
Set by moritz_ on 28 December 2010.
masak I had some $life happen to me, ran out of time with my sekkrit project, so the graphics in the post are brought to you by ASCII. 00:00
oh well.
masak 'night, #perl6. 00:01
[Coke] if I want to test the exception message thrown for a compile time error, should: eval "thing that throws the error"; ok ~$! ~~ / 'portion of literal message' /, "reason"; // work? 00:34
sorear diakopter: well I'm here. 01:07
[Coke]: yes but I don't like it
[Coke] sorear: ok, well it /doesn't/ work. trying to figure out why - and why don't you like it? 01:22
aside from the usual "don't test the actual error message" (but in this case, that was the point of the ticket that was opened, that the error message didn't match STD) 01:24
rakudo: / <[a-z]> / 01:25
p6eval rakudo 549d2a: OUTPUT«===SORRY!===␤Unsupported use of - as character range;in Perl 6 please use .. at line 22, near "z]> /"␤»
[Coke] std: / <[a-z]> /
p6eval std 625303c: OUTPUT«===SORRY!===␤Unsupported use of - as character range; in Perl 6 please use .. at /tmp/f1feA04ho8 line 1:␤------> / <[a-z⏏]> /␤Parse failed␤FAILED 00:01 120m␤»
[Coke] GAH. That's why it hasn't been matching all this time. (note the extra space in STD. ;)
... why is that space missing? looking at src/Perl6/Grammar.pm, it looks like the space should be there. 01:32
[Coke] ... because it's coming from nqp-rx. 01:35
[Coke] cries.
thundergnat std: my @g; @g[^2]>>.[^6] >>[=]>> 0;
p6eval std 625303c: OUTPUT«===SORRY!===␤Can't hyper with = because list assignment operators are too fiddly at /tmp/gRZaVWEaGE line 1:␤------> my @g; @g[^2]>>.[^6] >>[=]>>⏏ 0;␤Check failed␤FAILED 00:01 122m␤»
thundergnat bummer
std: my @g; @g[^2]>>.[^6] >>&[=]>> 0; 01:39
p6eval std 625303c: OUTPUT«===SORRY!===␤Missing << or >> at /tmp/RRBCYt62uJ line 1:␤------> my @g; @g[^2]>>.[^6] >>&⏏[=]>> 0;␤ expecting infix_circumfix_meta_operator␤Parse failed␤FAILED 00:01 121m␤»
thundergnat std: my @g; @g[^2]>>.[^6] >>&infix:<=>>> 0; 01:40
p6eval std 625303c: OUTPUT«===SORRY!===␤Missing << or >> at /tmp/ECoeVF8mPS line 1:␤------> my @g; @g[^2]>>.[^6] >>&⏏infix:<=>>> 0;␤ expecting infix_circumfix_meta_operator␤Parse failed␤FAILED 00:01 121m␤»
thundergnat std: my (@a, $b); @a >>[>]>> $b; 01:42
p6eval std 625303c: OUTPUT«ok 00:01 123m␤»
dalek p-rx: ab1ffc0 | Coke++ | src/Regex/P6Regex/Grammar.pm:
Fix error message to match STD.
 02:08
p-rx: 96e4b67 | Coke++ | src/stage0/ (3 files):
update bootstrap for " in Perl 6" error update.
diakopter sorear: yes but this has occured for a few years now :) 02:19
diakopter wonders idly whether TimToady made it to India yet
sosey hello 02:41
i'm a newbie with perl, but i have a lot of experiencie working with python
do you recomend me to use perl? 02:42
dukeleto sosey: we recommend that you write code that you enjoy writing 02:45
sosey: also, hello
sosey dukeleto: i'm enjoy writting c code, so i think i can enjoy writting perl code 02:46
dukeleto sosey: perhaps :)
snarkyboojum sosey: writing C and Perl 6 is very different, but most people in here would at least say "Give it a try" :D
dukeleto sosey: what kind of code do you want to hack on? what interests you? It is hard for anybody to help you if we don't know what you are trying to do. 02:47
snarkyboojum interesting reaction :) 02:49
perigrin "I like writing device drivers, can I write those in Perl?"
sorear www.datsi.fi.upm.es/~jmpena/korbit.html 02:50
flussence_ that pretty much describes my dayjob
perigrin sorear: you scare me. 02:51
flussence_ "here's some undocumented hardware with a serial port. do everything kthx"
sorear incidentally, there was a GSoC2010 project to port Parrot to the RTEMS kernel environment 02:59
skangas Is there something like //= but for regular expressions? 03:18
I would like to do "$foo =~ /bar/" without having to prefix it with "defined $foo &&"
Otherwise if $foo is undefined, I get "Use of uninitialized value in pattern match" 03:19
Oh crap, wrong window, I thought this was #perl. Sorry :-( 03:20
colomon skangas: I was trying to figure out how to answer the question in Perl 6. :)
skangas colomon: That would be interesting, too, albeit not directly relevant to what I am doing right now. ;-) 03:21
colomon mind you, I don't know that I have a better answer than ($foo // "") ~~ /bar/
which still seems kind of awkward, IMO.
flussence_ couldn't it just do "class Regex { multi ACCEPTS(Any) { Bool::False } }", or would that break other things? 03:27
pmichaud draft of "What Rakudo needs from Parrot in 2011" available at 03:29
22:42 <pmichaud> but overall I don't want them to think "we keep doing things for Rakudo and they never recognize/appreciate it" :-)
22:42 <jnthn> *nod*
argggggh
wrong paste
(that was weird)
trying again
draft of "What Rakudo needs from Parrot in 2011" available at pmichaud.com/sandbox/rakpar.txt 03:30
flussence_ I think we've all mis-pasted something in here at some point :)
pmichaud comments and suggestions welcomed. I'll post that message to parrot-dev in a few hours or tomorrow morning
colomon seems a good message IMO. 03:34
"We know that any improvements serialization" probably wants a "to"?
pmichaud yes, added. 03:35
thanks
when I'm editing phrasing I tend to drop/add prepositions and articles
[Coke] wonders why this test file has a non-zero exit status. 04:48
fudge adds exit(1)? how do other tests that are fudged not fail the harness? 04:52
... arg. because I pre-fudged the file manually, I bet. 04:54
dalek ast: d9f5c2e | Coke++ | S05-metasyntax/charset.t:
Add a test for RT #69682
 04:57
[Coke] ok, that took longer than it should have. ;) 05:01
dalek kudo: fcc46ea | coke++ | build/PARROT_REVISION:
get latest nqp-rx fix for error message.
 05:02
[Coke] seen jnthn ? 05:14
aloha Sorry, I haven't seen jnthn .
[Coke] seen jnthn?
aloha jnthn was last seen in #perl6 6 hours 45 mins ago saying "They're just DSLs in Perl 6, right? ;_)".
[Coke] aloha: msg bacek - "seen foo ?" should work the same as "seen foo?"
sorear colomon: $foo andthen m/bar/ or $foo ~~ *.defined S& /bar/ 05:15
sorear [Coke]: THere's a --keep-exit-code or similar option that Rakudo uses 05:16
pmichaud: Doesn't PIRATE already implement #4? 05:18
pmichaud sorear: yes, apparently it does, although I'd like to see that declared "official" somehow for Parrot (either official tool or official approach for achieving that result) 05:19
[Coke] sorear: yup, danke. 05:20
hey, can we close 63584 with "well, don't DO that, then?" 05:21
sorear [Coke]: I'd rather we didn't. 05:22
One of my favorite things about the Perl5 toolchain is that it's very good about handling >1 perl in the $PATH
or even outside $PATH 05:23
[Coke] it's not that there's more than one perl in the path, it's that he re-arranged his path.
sorear I can install software using /opt/perl-5.11-debug/bin/perl $(which cpanm) ... and it all just works, even if the modules need to build XS and ...
[Coke] hurm. nevermind.
sorear oh, between steps?
[Coke] sorear: that was my impression, yes. 05:25
sorear yeah. don't do that then. 05:25
[Coke] rakudo: say q<< <<woot>> >>
p6eval rakudo 549d2a: OUTPUT«===SORRY!===␤Confused at line 22, near "say q<< <<"␤»
sorear if you want to switch perls in $PATH, use perlbrew
[Coke] std: say q<< <<woot>> >> 05:29
p6eval std 625303c: OUTPUT«ok 00:01 119m␤»
dalek ast: afcdfa2 | Coke++ | S02-literals/quoting.t:
track RT#
 05:31
[Coke] sorear: I misread. he used /foo/to/bar/perl, but first-in-path was being picked for other things. 05:32
I'll put in a poisoned perl and see if anything breaks.
*facepalm* 05:36
if you build parrot with "perl", that's what's passed along to Rakudo's build system as the perl to use. not /derived/path/to/perl, just "perl" 05:37
so if you build parrot, change your path, you're going to get a different perl. (or, if you run rakudo's configure.pl and change your path, the build will use a different perl.) 05:38
... not sure if this is worth fixing.
simon__ hey. can someone explain to me how chained comparison operators work? 05:55
simon__ at first i thought they were just syntactic sugar but then I tried to define another function that uses them internally and it could still be chained 05:56
rakudo: sub infix:<lesser> ($a, $b) { $a < $b }; 1 lesser 2 lesser 3; 05:59
p6eval rakudo 549d2a: ( no output )
simon__ rakudo: say 1 lesser 2 lesser 3 06:00
p6eval rakudo 549d2a: OUTPUT«===SORRY!===␤Confused at line 22, near "say 1 less"␤»
[Coke] (need to combine them.)
simon__ on my local machine it worked ;/
[Coke] each line here is a separate compiled program. 06:01
simon__ oh
[Coke] rakudo: sub infix:<lesser> ($a, $b) { $a < $b }; say 1 lesser 2 lesser 3;
p6eval rakudo 549d2a: OUTPUT«Bool::True␤»
[Coke] rakudo: sub infix:<lesser> ($a, $b) { $a < $b }; say 1 lesser 2 lesser -1;
p6eval rakudo 549d2a: OUTPUT«Bool::False␤»
simon__ how can this work? 06:02
[Coke] I'll leave that to someone more familiar with the grammar. 06:03
qiyong is perl6 out? 06:18
simon__ it has been for quite some time now 06:18
qiyong will it replace perl5? 06:19
[Coke] nope. New language. 06:20
simon__ probably not, at least not right away
[Coke] We hope people will like it and use it, but perl5 will be around for some time.
cxreg i keep having to correct people about that too. "no, you're wrong. perl6 _is_ out" 06:34
sadly, it's sometimes the same people over and over
[Coke] sorear: can you close out RT #75030? 06:38
msg sorear: can you close out RT #75030? 06:40
sorear [Coke]: that's "phenny: tell sorear ..." 06:57
colomon: and th answer is no.
[Coke]: and the answer is no. 06:58
simon__: Boolean values become 0 or 1 when passed to numeric operators like < 06:59
rakudo: sub infix:<lesser> ($a, $b) { $a < $b }; say 4 lesser 5 lesser 2 07:00
p6eval rakudo fcc46e: OUTPUT«Bool::True␤»
simon__ that can't be
sorear rakudo: say (4 < 5) < 2
p6eval rakudo fcc46e: OUTPUT«Bool::True␤»
sorear rakudo: say True < 2
p6eval rakudo fcc46e: OUTPUT«Bool::True␤»
sorear what can't be?
simon__ it tried it it still works when it would fail if they were interpreted as 1 07:01
at least i think so. ill try it again
you were right, sorry, my mistake 07:04
snarkyboojum lue: last post looks like some serious yak shaving :) 07:04
[Coke] sorear: ... no because you don't have privs or because you disagree? 07:08
[Coke] needs to stop poking RT with a stick. 07:29
moritz_ good morning 07:58
moritz_ phenny: tell pmichaud re nqp.txt and the parrot tools using nqp-rx, there's another option: they just can continue to use nqp-rx, and we can hand out commit bits to nqp-rx freely to all parrot folks (or even move the nqp-rx repo to the parrot organization on github) if that's what the parrot folks want 08:06
phenny moritz_: I'll pass that on when pmichaud is around.
masak [Coke]++ # RT queue traversal 10:52
cosimo aah 11:52
hallo
tadzik hello 12:08
masak ehlo 12:15
[Coke] masak: no worries. what else am I going to do when I'm up until 4am for no reason? 12:54
masak [Coke]: we need more insomniacs like you :) 12:57
[Coke] too bad I had to get up by 7:30 or I'd do it more often. ;) 12:58
masak maybe we need a slogan, too: "Can't sleep? Why not try to reproduce a couple of RT tickets?" 12:59
[Coke] perl6: /<[a-z]>/ 13:01
p6eval rakudo fcc46e: OUTPUT«===SORRY!===␤Unsupported use of - as character range; in Perl 6 please use .. at line 22, near "z]>/"␤» 13:02
..niecza v1-207-g786a885: OUTPUT«===SORRY!===␤␤Unsupported use of - as character range; in Perl 6 please use .. at /tmp/FD6qiWbDPI line 1:␤------> /<[a-z⏏]>/␤␤Parse failed␤␤»
..pugs: OUTPUT«Error eval perl5: "if (!$INC{'Pugs/Runtime/Match/HsBridge.pm'}) {␤ unshift @INC, '/home/p6eval/.cabal/share/Pugs-6.2.13.16/blib6/pugs/perl5/lib';␤ eval q[require 'Pugs/Runtime/Match/HsBridge.pm'] or die $@;␤}␤'Pugs::Runtime::Match::HsBridge'␤"␤*** '<HANDLE>' trapped by operat…
[Coke] hurm. why is std not in "perl6?"
std: /<[a-z]>/
p6eval std 625303c: OUTPUT«===SORRY!===␤Unsupported use of - as character range; in Perl 6 please use .. at /tmp/MjkysOSYRL line 1:␤------> /<[a-z⏏]>/␤Parse failed␤FAILED 00:01 120m␤»
takadonet morning all
masak takadonet: \i
er, \o :)
[Coke] œ∑´®†\¨ˆøø
ø/
moritz_ mojibake-smiley! 13:03
[Coke] oh, hey, look, if you hit enter, it sends it out to everyone!
masak we're happy for all contributions, regardless of quality.
moritz_ [Coke]: std is not in 'perl6:' because it doesn't produce run time output, and the perl6: target is supposed to return the same for all implementations for deterministic programs 13:04
perl6: say rand # of course not in such a program :-) 13:05
p6eval niecza v1-207-g786a885: OUTPUT«===SORRY!===␤␤Action method term:rand not yet implemented at /tmp/NUZQGAGRxp line 1:␤------> say rand⏏ # of course not in such a program :-)␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/SAFE.setting line
..377 (SAFE die @ …
..rakudo fcc46e: OUTPUT«0.337835111978592␤»
..pugs: OUTPUT«0.6156015820139091␤»
colomon and pugs kind of breaks that badly these days...
[Coke] listens to a story about a teleporting buddhist tyrannosaurus rex.
[Coke] wonders if something like:
std,rakudo: say "what?" ## would be helpful.
moritz_ might be. And not too hard to implement 13:06
[Coke]++ # new evalbot ideas
[Coke] (also "perl6,std:" to make checking diagnostics easier.) 13:08
moritz_ niecza: say 1, 2 Z 3, 4 13:09
p6eval niecza v1-207-g786a885: OUTPUT«Unhandled exception: Zip operators NYI␤ at /home/p6eval/niecza/lib/SAFE.setting line 377 (SAFE die @ 2)␤ at /home/p6eval/niecza/lib/SAFE.setting line 1022 (SAFE infix:<Z> @ 2)␤ at /tmp/KR83bnMFOC line 1 (MAIN mainline @ 3)␤ at /home/p6eval/niecza/lib/SAFE.setting
..line 1037 (SAF…
[Coke] hurls isitchristmas.com/ ... which is very funny and relevant to perl6, but has a very sad story behind it. 13:24
masak [Coke]: what's the story behind it? 13:30
heh, the JavaScript source for telling whether it's Christmas is very readable... and it even does a bit of i18n, it seems. but it doesn't take into account that the *date* for Christmas varies. :) 13:32
ask any Swede, and they'll say Dec 24 without hesitation.
colomon our old neighbors never put up their Christmas decorations until after Dec. 25th. I think they were perhaps Russian Orthodox? 13:33
moritz_ well, in .de we think Christmas is from Dec24 to Dec26 13:34
[Coke] masak - en.wikipedia.org/wiki/Bill_Zeller 13:35
moritz_ [Coke]: I don't quite see the connection 13:37
[Coke] sorry, he was the author. 13:38
moritz_ ah. 13:38
I looked into the whois record, and he wasn't the domain holder 13:39
that was I all I check though
[Coke] linked to from a news article about the suicide.
masak :( 13:41
I've used Graph Your Inbox, one of his creations. it's a nice program.
moritz_ medidates over perl6.org/page-stats#Domain%20referrals 13:44
1.5k visitors from planet.haskell.org. not bad.
takadonet wait a min... why can I see that link? I have no administration power on perl6.org? 13:46
moritz_ takadonet: what error do you get? 13:48
I don't think there are any restrictions on that page
takadonet moritz_: i get no error. I can see the page 13:48
moritz_: just thought it would have been restricted 13:49
moritz_ ah, I thought there was a negation somewhere :-)
why should it be restricted?
it doesn't reveal any private information
moritz_ and whom would we show the page, and whom not? 13:50
flussence_ The referrer links could use a rel=nofollow... 14:01
flussence_ or at least exclude that page in robots.txt 14:02
moritz_ flussence_: robots.txt is a good idea... care to patch it? 14:03
flussence_ er, ok 14:05
pmichaud good morning, #perl6 14:13
phenny pmichaud: 08:06Z <moritz_> tell pmichaud re nqp.txt and the parrot tools using nqp-rx, there's another option: they just can continue to use nqp-rx, and we can hand out commit bits to nqp-rx freely to all parrot folks (or even move the nqp-rx repo to the parrot organization on github) if that's what the parrot folks want
pmichaud moritz_: yes, that's definitely another option. 14:14
I'll try to add it to the text as well.
takadonet pmichaud: morning 14:15
moritz_ I think it's important we don't the parrot dev's tools away as we go along :-)
[Coke] chuckles at the latest perl5 bug: rt.perl.org/rt3/Ticket/Display.html?id=83086
pmichaud well, I'm rather hoping that Parrot will still decide to adopt the new NQP along with its object model. It's not that big a stretch.
masak morning, pm.
moritz_ pmichaud: sure. But some people like choices, just for the sake of it :-) 14:16
masak moritz_: important that we don't accidentally the parrot dev tools? :)
pmichaud I took it as "don't chase"
[Coke] pmichaud: is there anything left to do on RT #50118, especially in light of jnthn's recent work? 14:17
flussence_ moritz_: sent a pull request for the robots.txt on github
pmichaud [Coke]: no.
I think we can close that ticket
moritz_ flussence_: please commit directly. KTHXBY :-) 14:18
flussence_ oh :)
moritz_ flussence_: you now have commit access to most repos in the perl6/ organization, including perl6.org, mu, roast, specs and ecosystem
we have a 'perl6' group into which we stuff people, because handing out bits for each repo separately is just too cumbersome 14:19
pmichaud any other comments on nqp.txt or rakpar.txt before I post them? 14:20
dalek href="https://perl6.org:">perl6.org: 400c7a1 | (Anthony Parsons)++ | robots.txt:
Block /page-stats in robots.txt

The page contains hyperlinks to the top referrers; this should stop referral spammers from taking advantage of it.		 14:21
moritz_ pmichaud: if '+1' count as a comment... that would be mine :-) 14:22
pmichaud: I guess you've seen my other correction in #parrot (s/#2/#3/)
pmichaud no, didn't see that yet 14:23
looking
moritz_ at the bottom, where you refer back the serialization
pmichaud ah, yes, I reordered but forgot to fix the reference
Added the line:
"Or, the existing
nqp-rx could continue to live on as its own project targeting Parrot
libraries, perhaps even being adopted into the Parrot repository.
Whatever the Parrot team chooses to do ..."
s/libraries/subsystems/ 14:24
[Coke] pmichaud: done. 14:29
colomon rakudo: say ~reducewith(&infix:<*>, 1...10) 14:31
p6eval rakudo fcc46e: OUTPUT«No applicable candidates found to dispatch to for 'reducewith'. Available candidates are:␤:(&op, *@args, Any :chaining($chaining), Any :right-assoc($right-assoc), Any :triangle($triangle), Any :xor($xor))␤␤ in main program body at line 22:/tmp/3gymi_u02B␤»
colomon oh, right 14:32
moritz_ huh, what's missing?
all those anmed params seem optional
colomon not optional, because it's parsing as the sequence &infix:<*>, 1 ... 10 14:33
or something like that.
moritz_ oh
rakudo: say ~reducewith(&infix:<*>, (1...10)) 14:34
p6eval rakudo fcc46e: OUTPUT«3628800␤»
colomon just a side effect, I'm trying to sort out why metaops (etc) are failing so badly with Math::BigInt. 14:35
moritz_ I have a guess
if you write [*] 1, 2 14:36
flussence_ I'd like that error message above to say something like "Got signature: [...]; Available candidates are: [...]"...
moritz_ rakudo generates a prefix:<[*]>
[Coke] perl6: say [+] [1,2,3] # what's the right answer here?
p6eval niecza v1-207-g786a885: OUTPUT«Unhandled exception: System.Exception: Unable to find lexical &reduceop in mainline␤␤Server stack trace: ␤ at Niecza.CLRBackend.NamProcessor.ResolveLex (System.String name, System.Int32& uplevel, Boolean core) [0x00000] in <filename unknown>:0 ␤ at
..Niecza.CLRBackend.NamProcessor.…
..pugs: OUTPUT«1 2 3␤»
..rakudo fcc46e: OUTPUT«3␤»
moritz_ but I guess that doesn't properly multi-dispatch
pmichaud [Coke]: Rakudo is correct
moritz_ [Coke]: 3 is correct I think
colomon oh wow, here's a bad one: 14:37
1L ... 50L turns into 50 50 50 50 50 ... 50
oh, I think I know what's up with that.
moritz_ so [*] is "frozen" to a particular multi, instead of doing a new dispatch in the caller's scope
s/multi/candidate/
colomon the thing that's weird is [*] worked fine when I had it our multi infix:<*>(Math::BigInt etc) at the top level AND I called [*] in the same file. 14:38
moritz_ that's because the generated prefix:<[*]> is global (and not per scope or so) 14:39
colomon but when I made it our multi infix:<L*>(Math::BigInt etc) is export but defined it in the Math::BigInt class definition and then called it in a separate file, it fails miserably.
colomon called [L*], that is. 14:40
chaitanya9186 Hello 14:58
masak chaitanya9186: hi! 14:59
chaitanya9186 Can i knw 15:00
u
?
masak yes...?
[Coke] Can anyone really know anyone?
gfldex there are moments where i know myself 15:00
masak [Coke]: what, in the biblical sense? :P
mathw I learn about myself in aikido classes, there are these vivid flashes of insight about halfway between the throw and the floor... 15:02
mathw in mid-air, upside-down and rapidly rotating :) 15:03
I start to understand why people do acrobatics
masak from an outside perspective, that sounds like the totally wrong moment to have an insight :P
mathw no at that point it's fine 15:04
because if you've not set up properly for the fall by then so that you can just relax and let it happen, you're probably going to land wrong anyway
masak oh, ok.
frettled And that's when mathw solves Perl 6 problems.
moritz_ masak: I guess it's like seeing your live passing by while falling from the cliff
"I should have made peace with Aunty Foo, really" 15:05
frettled :D
mathw proper throws send you down too fast to be able to fall consciously 15:05
[particle] got a large rope burn on his ankle this weekend after a mistaken cliff fall
mathw ouch
[particle] my fault for having my foot behind the rope :( 15:06
mathw I hit the mat hard enough to numb three of the fingers on the hand that hit first
need to work on which bit of my body hits the ground first, spread it out a little more
just a tad too much force to dissipate into three fingers
[particle] obviously, you need more fingers. 15:07
masak imagines having an Auntie named "Foo"
mathw [particle]: what would be cool is hands which can expand into giant shock-absorbing pads in a fraction of a second just before landing :) 15:10
[particle] i'd certainly like skin that immediately stiffens when exposed to extreme friction 15:11
mathw That would also be useful 15:13
And maybe some sort of robot arm which comes out and whacks people who walk into me in the street because they forgot that other people use the pavement too... or am I going overboard now? 15:14
pmichaud wouldn't "going overboard" be something that would happen on a ship instead of a sidewalk? 15:14
frettled or boardwalk? ;) 15:15
mathw on a boardwalk, it's 'falling in' :)
or maybe 'falling off'
all purely accidental of course
mathw whistles innocently
pmichaud: Good email. 6model is leading to excitement
takadonet boardwalk empire : en.wikipedia.org/wiki/Boardwalk_Empire 15:16
?
moritz_ must remember to keep distance to mathw :-) 15:16
I do bump into people who just stop walking in front of me :-)
[particle] stop looking at your smartphone when you walk :P 15:19
masak I hear there are apps that show you the street while you type :P 15:20
moritz_ doesn't even have a smartphone 15:23
mathw that's the problem then 15:26
it can't show you the street while you're typing your SMSes
moritz_ has written about 2 SMS last year 15:27
Juerd I'm going to archive some user accounts on feather again 15:34
Juerd All homedirs with mtime < 2009 will be tarred, the corresponding user accounts removed 15:35
If you wish to keep your account, please shout :)
BinGOs tarred and feathered!
Juerd anatoly atnnn azuroth broquaint bsmith clkao cmarcelo domm er evalbot ferreira gaal geoffb henk jedai jesse joepurl johnpc karhu kolibrie lanny larry limbic liz lt marcel markstos mdiep morrow nothingmuch pdcawley pelagic scook0 shlomif simon sjp spinclad stevan szbalint tewk thoughtpolice zoo 15:36
masak if that's not two terms in a row, I don't know what is. :P 15:38
arnsholt "We could try each possible move in each possible heap, but that's inelegant." And what I did in my Kayles AI ^_^ 15:46
masak arnsholt: I haven't convinced myself that that particular feature of Nim carries over to .137 15:47
I'll know in a few days :P
arnsholt Oh, goody. =)
masak it's less trivial because of the mapping that goes on between .137 an Nim. 15:48
so trying each possible move might actually be necessary.
sorear good * #perl6 16:06
pmurias sorear: hi 16:08
sorear: i'm updating cl-backend to doc/nam.pod
sorear pmichaud: I was under the impression that the P in PCT, PAST, POST always stood for "pmichaud" 16:15
masak hi, sorear.
masak sorear: see also PmWiki, PMCs, and PGE :) 16:16
pmurias didn't they stand for "Parrot"? 16:16
sorear [Coke]: no because I don't have an RT account
masak pmurias: come to think of it, what does the 'pm' in pmurias stand for...?
diakopter :)
[particle] just call me pmichaudarticle 16:17
pmichaud when I was teaching at the university, some of my students would look at the many ".pm" files in /usr/lib/perl5 and marvel at how much code I was able to write.
masak *lol*
pmichaud "Dr. Michaud *really* likes Perl!"
pmurias heh
[Coke] heh.
pmichaud I did want to register "pm.org" for myself, but alas, it was taken by some subversive programming cult. 16:18
afaict, they're just squatting on the domain anyway
PerlJam pmichaud: I'd thought you'd lobby for a .pm TLD by now 16:19
pmichaud hadn't quite thought of that yet.
yay, new blog online: pmthium.com/ 16:20
pmichaud now to get the feeds into Planet Perl Six, Planet Parrot, and Iron Man Perl 16:20
masak pmthium! \o/ right there in the periodic table, between wonderflonium and unobtainium. 16:21
pmichaud actually, I think it's between Neodymium and Samarium. :-) 16:23
pmichaud But there are some new isotopes to be explored. 16:23
masak :)
PerlJam If Adam Ant can get his own element, anyone can.
masak groans 16:24
pmichaud ummmm, how do I update the location of my ironman feed, ooc?
PerlJam contact mst? 16:25
masak there's an ironman address on the Ironman page. 16:26
pmichaud maybe mailto:ironman@....
yeah
I'll try that
moritz_ iirc [Coke] can change planetsix URLs 16:32
[Coke] aye. 16:40
jnthn evening o/ 16:42
jnthn is pleased to see pmichaud has good taste in WordPress themes. ;) 16:43
pmichaud heh
well, I was working on an abstract header graphic, but it wasn't turning out well and I finally decided to just use a landscape 16:44
jnthn :)
jnthn It's a pretty one. :) 16:44
Looks a tad warmer than the one on 6guts. :)
masak hm, maybe I should have a landscape too, that's not too bad... 16:48
anyone have a strangely consistent langscape sitting around? :P
jnthn masak: Trying to think of somewhere that has old and new architecture side by side and makes it work... 16:50
jnthn Though that'd be a cityscape rather than a landscape. :) 16:51
pmichaud ("tad warmer") actually, that photo was taken in late november, and it was around 5 degC when I took the photo :) 16:53
so, not all that warm
jnthn Heh
Mine was taken in mid-May. :)
And was no warmer than that, but I doubt it was awfully colder. :) 16:54
pmichaud so it's entirely possible it was warmer when you..... right :)
[particle] jnthn: i'm thinking the guggenheim in bilbao or the mit campus
jnthn Yes, I think my photo taken in November woulda been pretty crappy.
Given it's in the Arctic Circle and there's no light in November. :)
masak [particle]: url? 16:55
[particle] www.kegz.net/archives/images/gehry_mit.jpg 16:56
pmichaud someone needed a better square and level when they built that.
moritz_ sees how it matches "strange", but doesn't see the consistency :-) 16:57
PerlJam wonders if hobgoblins factor into this somehow 16:58
jnthn Mmmm....Hobgoblin :D 16:59
[particle] that's just one angle, which shows old+new integrated 17:01
it's a fascinating building
dalek ecza: a7f0672 | pmurias++ | cl-backend/backend.lisp:
[cl-backend] port over unit definition from doc/nam.pod
 17:02
ecza: a7cd075 | pmurias++ | cl-backend/backend.lisp:
[cl-backend] refactor the way subs are compiled
jnthn [particle]: Wow, that's...quite a building.
moritz_ masak: maybe some buildings by Hunterwasser? 17:03
pmichaud fwiw, I like the strangelyconsistent header graphic 17:05
not saying "don't change it", but I like it fine as it exists now 17:06
daxim Hundertwasser!
masak moritz_: url?
masak [particle]: I look at that, and go "well, um, it gets the 'strangely' part down pat..." :) 17:07
daxim next time you come to vienna I could show you around. 17:09
masak yay 17:10
masak has been to Vienna twice so far
lovely city.
daxim www.kunsthauswien.com/en/museum # enjoyable geekery 17:11
masak pmichaud: *really*? it's a placeholder specifically designed to give me enough bad conscience to make a real one... :P
allbery_b glances up and imediately thinks "stata?" 17:18
moritz_ masak: images.google.com/images?q=hundertw...+buildings 17:22
masak whoa 17:23
masak this is what most gives me vibes of "strange consistency": www.roxanneardary.com/wp-content/up...-hajar.jpg 17:26
not too suitable as a header image, sadly... 17:27
PerlJam I keep thinking of things that are consistently strange (like Dali paintings :) 17:29
moritz_ PerlJam: I've thought of Dalí too
colomon is testing his internet connection 17:32
moritz_ test, 1, 2, 3 17:33
masak .oO( The Tango Maureen ) 17:34
dalek ast: b814b4d | moritz++ | S03-operators/eqv.t:
[eqv.t] tests for RT #75322
 18:02
moritz_ rakudo: class A is Array { }; my @a := A.new; @a.push: 'foo'; say @a 18:06
p6eval rakudo fcc46e: OUTPUT«foo␤»
dalek ast: bfe67a6 | moritz++ | S12-class/inheritance.t:
[inheritance.t] tests for RT #74844
 18:19
[Coke] moritz_++ # writing tests. 18:20
moritz_ [Coke]++ # identifying what to write tests for 18:22
colomon moritz_++ [Coke]++ # getting things done 18:25
muixirt rakudo: say 2**2**2**2**2 18:31
p6eval rakudo fcc46e: OUTPUT«Inf␤»
PerlJam Is that an associativity problem? 18:32
(perhaps in my head?)
moritz_ ** is right-assoc 18:33
muixirt inspired by twitter.com/#!/climagic/status/32140176075530240
moritz_ both in spec and rakudo
PerlJam then, yes, it's in my head :) 18:34
colomon whoa, that's a big number
colomon thinks he just created the biggest number ever calculated in Rakudo. :) 18:35
gist.github.com/804533
moritz_ colomon: strictly speaking it wasn't calculated *in* rakudo 18:36
colomon moritz_: by that logic, no math has ever been done *in* rakudo.
PerlJam but thankfully both your number and python's number agree :)
colomon \o/ 18:37
that's 19,729 digits long. 18:38
flussence_ rakudo: multi sub infix:<↑↑>(Int $a, Int $b) { return [**] $a xx $b }; say 2 ↑↑ 4
p6eval rakudo fcc46e: OUTPUT«65536␤»
dalek ast: aeb686f | moritz++ | S32-io/dir.t:
basic tests for dir()
 18:39
flussence_ rakudo: multi sub infix:<↑↑>(Int $a, Int $b) { [**] $a xx $b }; say 3 ↑↑ 4 18:40
p6eval rakudo fcc46e: OUTPUT«Inf␤»
flussence_ guess it doesn't like 7 trillion... 18:41
rakudo: multi sub infix:<↑↑>(Int $a, Int $b) { [**] $a xx $b }; say 3 ↑↑ 3
p6eval rakudo fcc46e: OUTPUT«7625597484987␤»
flussence_ or I just did it wrong
moritz_ I know $x ↑↑ $x as the Fuga function 18:42
not sure if that's correct or wildly known though
Tene flussence_: you need to define that as a meta-op, so you can handle any number of up arrows
3 ↑↑↑↑ 3
flussence_ moritz_: haven't heard of it as that name before - en.wikipedia.org/wiki/Knuth's_up-arrow_notation 18:43
dalek kudo: d3e9a36 | moritz++ | t/spectest.data:
run S32-io/dir.t test file
 18:44
flussence_ how do you define a meta-op anyway? I looked before but couldn't find anything
colomon You can't easily define one without modifying Rakudo's source, if that's what your'e asking 18:45
huh. 18:46
> use Math::BigInt; say reducewith(&infix:<**>, (3L, 3L, 3L, 3L))
7625597484987
rakudo: say 3 ** 3 ** 3
p6eval rakudo fcc46e: OUTPUT«7625597484987␤»
colomon where am I losing the last 3 there?
oh, right.
right-associative, that is. 18:47
errr, that's going to be a lot bigger than 2 ↑↑ 5 was....
colomon has finally given Math::BigInt a problem big enough to make it go very slow. 18:48
very slow indeed 18:49
flussence_ I don't think the answer will fit into RAM :) 18:50
takadonet flussence_: depend which machine :)
colomon is that why my MBP is starting to smoke? 18:51
killed it before answer, sorry.
takadonet colomon: running out of ram? 18:52
colomon I don't know, but I was trying to get work done, so letting a pointless math problem trample the machine's processing and RAM seemed a poor idea. 18:53
The cooling fan definitely kicked in.
dalek ast: 3a34f01 | (Fitz Elliott)++ | S04-declarations/my.t:
[S04/my.t] fix and unfudge passing test for RT #63588 (global vars visible inside class)
 18:54
felliott is there anything more to #63588?
otherwise, I'll close it 18:55
dukeleto colomon: nice work on Math::BigInt and friends
PerlJam colomon++ what dukeleto said. Though I have to tell you that I keep waiting for the "and now this is built-in" post :) 18:56
colomon PerlJam: last time I talked about it with pmichaud, he was worried (with some justification) about slowing down Rakudo further. 19:03
So I've got no plans to try to get Math::BigInt in place at this time. 19:04
but give me a few days, and I'll probably take a stab at implementing Math::FatRat... ;)
PerlJam colomon: Sure, it's just if life were like a movie when I'm guessing the future plot changes, that seemed like a natural progression :) 19:04
colomon PerlJam: believe me, I'd love to get it in Rakudo proper. It would certainly be more elegant... 19:07
pmichaud if adding BigInt doesn't slow down the existing parts of Rakudo, I have absolutely no problem with it. 19:09
colomon oh. 19:10
pmichaud I don't mind if BigInt is slow, I just mind if it slows down what we already have. :)
colomon understood.
well, I'll keep that in mind, and talk with more about it in a few days, then. 19:11
pmichaud sounds great
PerlJam What does python use for arbitrary precission arithmetic? The python is *fast*
jnthn We already have the checks in place, iirc. 19:12
Promoting to BigInt where we today promote to Num is likely not too big a deal.
colomon jnthn: not really. we do calculations in Num, and then switch back to Int if they fit.
jnthn Oh.
PerlJam ah, gmp of course
colomon that's not at all the same as doing them in proper Int form internally. 19:13
jnthn I'm still not really too taken with those being the same type.
But I guess people can use int if they want speed.
Trouble is that then we force people into type annotations. 19:14
colomon though I guess we could do what we do now, and then just redo the calculation if the result would overflow an int.
pmichaud I think the Perl 6 vision was that Int would have two reprs behind it (one fast, one big) and would internally switch as needed
not that there'd be a single BigInt repr
colomon Ack, it needs clever scheme, and right now I should be worrying about $work.
+1
dalek ast: e6a9f12 | (Fitz Elliott)++ | S03-operators/overflow.t:
[overflow.t] add missing utility functions so tests pass; update plan;
 19:15
jnthn pmichaud: Ooh, REPR switching...
...that could be, er, evil. :)
pmichaud basically, Int is a type wrapper around multiple "integer" types
perhaps like Numeric is a type wrapper around Int + Rat + Num
jnthn Yes, but "type wrapper" is clearly defined there. Numeric is a role and the others do it. :) 19:16
That could be a viable solution here though, perhaps.
pmichaud Right, I mean it in a handwavy sort of sense. 19:17
jnthn :)
Yes, the trouble is that the hands have never not waved on this issue yet. ;)
I suspect we can come up with some sane REPR-y thing though.
pmichaud well, maybe *your* hands haven't :-)
jnthn :P 19:18
pmichaud but istr that timtoady++'s hands have
PerlJam Is there some sort of XS <-> Rakudo bridge somewhere?
jnthn pmichaud: Maybe I should search the ir clogs... :) 19:19
PerlJam: Blizkost is...sorta...ish. :)
PerlJam sortaish is good enough for me for now :) 19:19
jnthn I mean, you can use Perl 5 modules that are XS ones using it.
I dunno if that's what you were thinking of though. :) 19:20
PerlJam This talk of "fast/big ints" has me wondering how to use PDL directly
jnthn goes to prep for tomorrow's class a bit 19:21
Tene PerlJam: if you're just looking for C interop, look at Zavolaj. 19:24
PerlJam I don't know that PDL has an api exactly. In fact, I think it does not. 19:25
dalek ast: 32dd626 | moritz++ | S03-operators/short-circuit.t:
unfudge passing tests
PerlJam But I haven't looked at PDLs insides in a long while. 19:26
dukeleto PerlJam: TimToady was talking about how "subrectangles" in Perl 6 replace piddles 19:44
dalek ast: 5b5d98a | moritz++ | S03-metaops/reduce.t:
[reduce.t] tests for RT #76110
 19:46
moritz_ rakudo: my $c = 0; sub called($ignored) { $c = 1 }; use Test; dies_ok { called pi = 4 }; say $c 19:58
p6eval rakudo fcc46e: OUTPUT«ok 1 - ␤0␤»
moritz_ why the heck do I get Bool::True for this last output? 19:59
locally, that is
colomon :\ 20:00
moritz_ a, I was stupid
I had a , after dies_ok, not a ;
now both tests fail. Huh. 20:01
dalek ast: 2d2be1c | moritz++ | S03-operators/assign.t:
[assign.t] two (failing) tests for RT #77142
 20:03
moritz_ somebody please check this commit to roast
jdhore out of curiousity, can perl6/rakudo-star do most of the stuff Perl 5 can do? 20:05
Su-Shee much, muche more. or the same stuff more easily. 20:06
jdhore Well, I was more meaning, besides missing most of CPAN, is there anything from Perl5 that i'd currently miss in Perl6? 20:08
moritz_ speed
saaki speed =(
moritz_ and robust IO
colomon speed
jdhore ah :(
colomon errr, heredocs and forms, too.
moritz_ never used forms in p5 either 20:09
jdhore I've never heard of forms and heredocs (unless that's POD docs inside actual perl modules), but i'm still kind of a Perl noob. 20:11
moritz_ heredocs start with <<MARKER in p5, and end with MARKER on a line of its own 20:12
they are just string literals with a custom delimiter
jdhore ah
[Coke] wonders if anyone has ported damian's perl6::format to perl 6.
moritz_ [Coke]: mathw has started, I think 20:12
[Coke] yay. 20:13
colomon heredocs are awesome, and p6 heredocs are supposed to be even better, but they're not supported in Rakudo yet. 20:14
moritz_ I think niecza does them so 20:15
niecza: say Q:to<foo>␤some stuff␤foo
p6eval niecza v1-209-ga7cd075: OUTPUT«some stuff␤␤»
moritz_ \o/ 20:16
sorear++
moritz_ and TimToady++ 20:16
flussence_ niecza: my $x = Q:to<foo>;␤ some stuff␤ foo␤say $x.perl
p6eval niecza v1-209-ga7cd075: OUTPUT«" some stuff␤"␤» 20:17
flussence_ hasn't got the whitespace magic stuff, but it's still impressive :)
masak submits nieczabug 20:20
:)
dalek ast: 2be4bce | moritz++ | S32-str/split.t:
[split.t] tests for RT #63066
 20:22
colomon You know, I thought the whitespace magic was p6's killer feature. (back in 2003, that is. ;)
PerlJam I thought p6 had two killer features: grammars and a nice OOP system. The first one is still a killer feature but the second one seems to have had enough time to leak out into the world 20:25
flussence_ (I've "almost" got .indent done btw, "just" need to figure out how to handle \t properly :)
PerlJam flussence++
masak flussence_: cool! that's great news. 20:26
if you feel the need to air ideas for how to implement that (or how to make S32/Str clearer), we're here for you.
PerlJam flussence_: What's the problem with \t exactly? And how can we help? :)
masak the problem is that there's a complicated invariant to be upheld between a \t and a certain number of spaces. 20:27
"\t".indent(4) eq "\t ", but "\t".indent(8) eq "\t\t" 20:28
and outdenting is tricky, too.
PerlJam whoa, is is a requirement that .indent generate tabs? 20:29
flussence_ I look a look at the Text::Tabs takadonet++ wrote, it's a bit verbose but I get the general idea of how to do it
pmichaud PerlJam: if tabs are present, I think .indent is supposed to preserve/extend them 20:30
takadonet flussence_: I never did finish it and I was simply doing line by line converting from the p5 version 20:31
i'm happy someone else is finishing it
PerlJam preservation I get, but extending I'm not so sure about. 20:32
And I just checked my copy of the spec and it only says "C<$steps> number of spaces have been added to each line"
though, I guess "consistent with leading whitespace" could be interpretted to mean "add tabs if the string starts with tab" 20:33
okay, never mind me :) 20:34
PerlJam wonders about other leading whitespace though 20:35
if the line starts with a unicode non-breaking space, does that mean that .indent will add NBSs rather than space characters? 20:36
pmichaud in the past, I've found .indent to conceptually be very much a dwimmy sort of construct 20:41
so I'd think "yes"
i.e., it's past descriptions imply that the dwimminess knob is turned up fairly high
PerlJam I guess the requirement that they all be the same is a concession to implementors for making that easy :) 20:43
otherwise you'd have to have crazy logic for all of en.wikipedia.org/wiki/Space_(punctu...in_Unicode
pmichaud "Perl exists to torment the implementors on behalf of the users." 20:44
or something like that.
PerlJam flussence_: where is your implementation of .indent? 20:46
flussence_ right now, here: github.com/flussence/p6-misc/blob/.../indent.p6 20:47
dalek ast: f45e8fe | (Fitz Elliott)++ | S32-str/split.t:
[split.t] fix typo in test
 21:01
mkramer is fascinated by the unicode spaces table
moritz_ felliott++ 21:09
PerlJam flussence_: gist.github.com/804840 I updated a comment and added a case for indenting where the string starts with a sequence consisting only of some other space character. No tests though. 21:23
flussence_ wait, is that default{} inside when{} correct? 21:28
masak flussence_: you can put a default {} anywhere. 21:31
flussence_ huh, I never knew that!
masak 'given' and 'when/default' are completely decoupled.
moritz_ you can also do for @list { when ... { }; default { } } 21:32
masak 'given' is just 'OK, we'll use this value as $_' (or whatever signature you pass into the pblick)
PerlJam flussence_: actually that was a cut-n-paste error :)
masak 'when' is just 'smartmatch $_ against this expression'
as moritz_ points out, 'given' is a kind of loop. 21:33
(that only iterates once)
PerlJam flussence_: fixed. But I'm glad you learned something from my error :)
masak rakudo: say "$_ is odd" when * % 2 for 15..20 21:34
p6eval rakudo d3e9a3: OUTPUT«15 is odd␤17 is odd␤19 is odd␤»
masak \o/
flussence_ either way, adding that line causes test #2 to infinite loop. weird :/ 21:35
flussence_ oh, got it 21:37
V15170R rakudo: sub prefix:< > ($val is rw) { $val = Mu }; my $a = 0; say $a; 21:45
p6eval rakudo d3e9a3: OUTPUT«0␤»
V15170R rakudo: sub prefix:<_> ($val is rw) { $val = Mu }; my $a = 0; say _$a; 21:46
p6eval rakudo d3e9a3: OUTPUT«Mu()␤»
flussence_ .u 2000 21:48
phenny U+2000 EN QUAD (􏿽xE2􏿽x80􏿽x80)
flussence_ .u 2001
phenny U+2001 EM QUAD (􏿽xE2􏿽x80􏿽x81)
flussence_ not ok 12 - Default to prepending ASCII spaces for mixed, non-tab leading whitespace characters # :(
snarkyboojum ni hao :) 23:07
masak 你好, snarkyboojum. 23:12
snarkyboojum oh hej, masak :) 23:13
masak :) 23:14
diakopter sighs at owning a Sandy Bridge CPU in a mobo with 4 defective 3Gb/s SATA ports. I'm assuming it will be 4-6 months before Asus/Intel fix it :( 23:17
*effectively* defective ports (I do realize the problem is with the chipset)
jdhore :( 23:19
masak pmichaud: is pmichaud.com/sandbox/rakpar.txt still open for spelling suggestions? if so, s/noticably/noticeably/ 23:26
pmichaud masak: it's already been posted, but thanks :)
I'll fix it when it goes to pmthium, though. 23:27
(if it gets posted there)
masak pmichaud: I'm not done reading yet, but it's a very good read so far. 23:30
pmichaud there are some responses already on parrot-dev
masak ooh 23:31
pmichaud lists.parrot.org/pipermail/parrot-d...05422.html is the start of the archive thread, if you want it :) 23:34
pmichaud oops 23:34
lists.parrot.org/pipermail/parrot-d...05410.html
(grabbed wrong message) 23:35
masak thanks -- I'm reading it right now, in my email client.
'there must be something wrong with "now"' -- yes, it creates objects, and that's very expensive. 23:36
oh, pmichaud++ has a response outlining that further down :) 23:38
sorear good * #perl6 23:41
masak very interesting thread. I hope to see more of it.
good *, sorear.
good night, #perl6 :) 23:42
sorear pmichaud: Woah, am I to understand Parrot is seeping into curricula? 23:45
sorear flussence_: I hear you've thought a lot about whitesace in here docs. 23:50
flussence_ as I understand the spec for heredocs, the whitespace rule looks identical to .indent(*) 23:53
sorear I've heard some people say that dedenting is sensitive to how the string is built 23:54
e.g. in my $foo = ' bar' ; say qq:to(pie)\n$foo\n pie 23:55
flussence_ hm, I see 23:57
IMO the whitespace trimming should happen before any interpolation 23:59
I can see how it might complicate things either way, though
sorear how do you suggest to accomplish that?

Powered by Cro and the Raku® Programming Language.

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

Thank you!