»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
japhb__ test 02:55
sjohnson japhb__: heh 02:58
hi
japhb__
.oO( "On the Choice of Weatherizing Pigments for Storage Facilities Containing Wheeled Conveyances: A Consensus-Based Approach" )
03:52
geekosaur *snrk*
TimToady I think we can improve that title... 03:54
japhb__ :-D
TimToady japhb__: feel free to add an entry to rosettacode.org/wiki/Distributed_programming when your peanut butter is sufficiently indistinguishable from magic 03:58
lue "We will begin this discussion by first exploring the full implications of the partial removal of all follicular matter present on the being thus referred to as /Bos mutus/."
TimToady that should come before the discussion 03:59
lue I feel it best occurs right at the start, lest we need to release a "prerequiste" paper.
japhb__ lue: nice 04:00
japhb__ TimToady: re: rosettacode, not a bad idea, thank you. Trying to pull up the page so I can see what I would put there, but RC seems to be slow at the moment. Or my net connection is flaking. :-/ 04:02
I like the concept of "peanut butter sufficiently indistinguishable from magic"
japhb__ Sheesh, RC seems unhappy right now 04:15
TimToady seems fine to me
lue me as well. Might be just you, or a combination of you and RC. 04:16
TimToady blames those skinny pipes from Google 04:17
it's possible you're in an address block that cloudflare doesn't like 04:18
course, I'm not in the Bay Area at the moment, so maybe we can blame the NorCal intertoobs 04:22
you haven't perchance had The Big One while you weren't paying attention? 04:23
TimToady is assuming that japhb__++ is still in that neck-o'the-woods 04:24
japhb__ "The Big One"? 04:35
Seems to have finally come back OK 04:36
japhb__ suddenly realizes that TimToady was probably referring to an earthquake -- in which case, no. :-) 04:42
moritz good morning, #perl6 07:08
moritz nimrod-lang.org/talk01/slides.html 07:58
moritz looks like feather2 is unresponsive 08:12
or just very, very slow
can anybody remember when Rakudo gained the nice syntax errors /TTIAR detection? 08:15
moritz those with a p6advent account can now preview perl6advent.wordpress.com/2013/12/0...6-in-2013/ 08:26
feedback and fixes welcome :-) 08:27
moritz seems that feather2 isn't down or slow, the network connection is just very lagging 08:27
s/lagging/lossy/ 08:28
66% packet loss
Woodi hallo today :) 08:43
Woodi how about nqp compiler on moarvm ? mvm is quite new so maybe there is a chance to have sources of nqpc written in moarvmlang ? but looking in the repo I found exactly what someone say, nqp is written in nqplang :) with some parrot ? 08:46
moritz Woodi: I only understood some single words of what you wrote 08:47
not the whole thing
there is no moarvmlang
Woodi ok, it probably my mistake :)
moritz there is a bytecode format
and the nqp compiler is written in NQP. 08:48
Woodi so nqp emits bytecode of nqp compiler ?
moritz nqp emits bytecode for moarvm. 08:49
Woodi yes, I wanted to say that... so first nqp on moarvm was written in what language ? 08:50
moritz in NQP
and cross-compiled by one of the other backends 08:51
Woodi oki, let brain postprocess that...
masak good antenoon, good #perl6 people (and bots) 09:39
I am having trouble today connecting to feather via ssh.
masak this is following a surprise reboot late last evening. 09:40
I skimmed the logs but found nothing about this. anyone have any more details?
nwc10 masak: Juerd did announce in advance that he was going to reboot them 09:42
(at least hours, if not a day)
masak oki, carry on then.
still doesn't explain why I now can't connect through SSH, though.
more exactly, 50% of the time I reach feather, but then the connection is *veeeery* slow. 09:43
maybe some heavy process is running on feather1.perl6.nl?
FROGGS[mobile] masak yesterday a simple rp: invocation took 3 minutes, and then it took seconds again 10:07
masak huh. 10:08
masak ok, I'm on feather now. but it's extremely laggy. 10:14
FROGGS :/ 10:15
masak surely someone else is noticing this, too? 10:19
masak ...and suddenly everything's fine again. never mind. 10:34
tadzik if feather a bit freenode-disconnecty to you too? 11:48
timotimo o/ 12:03
jnthn Looks like we're missing dalek after the feather stuffs too 12:08
timotimo i'd like a login for the perl6 advent calendar please 12:11
timotimo ah, there's a register button 12:11
jnthn I'm not sure who gives those out...
timotimo aaw, both timo and timotimo are taken already 12:12
jnthn Try timotimotimo?
masak tadzik: freenode-disconnecty just about sums it up. 12:14
tadzik: but it got a lot better for me about 2 hours ago.
moritz lots of packet loss 13:22
could anybody please review the advent calendar post for tomorrow? 13:23
jnthn moritz: link?
moritz jnthn: perl6advent.wordpress.com/2013/12/0...6-in-2013/
jnthn 404... 13:24
oh, maybe I need to log in...
moritz you do
'cause it's not yet published
isBEKaml It's already advent time? I'm looking forward to it!
moritz isBEKaml: well, tomorrow :-)
isBEKaml \o/
jnthn "There are still some bits missing, mostly notable support " 13:25
most *notably*
parrot probably wants capitalizing
On "much faster at runtime", could link a couple of the blog posts folks made talking about that. 13:26
moritz just remembers one 13:28
jnthn I thought there was the recent one on rendering stuff?
And the colomon one longer ago?
moritz jnthn: I remember the rendering one 13:29
justrakudoit.wordpress.com/2013/08...rformance/ also
jnthn So that's two, no? :) 13:30
moritz aye 13:31
ok, linked those two
jnthn May be worth a general mention that there's been loads of improvements to built-in types/functions too 13:38
moritz jnthn: I've added a sentence about those 14:49
though I kinda hope that lizmat++ will make a separate post about Bag/Set/* :-)
colomon moritz++: nice post. One tiny note: I think I'd be inclined to end the first sentence ("Welcome to…") with an exclamation point. 16:06
moritz changes it that way.
arnsholt 22.Clojure from the ground up: Macros (aphyr.com) 84 points by llambda 15 hours ago | flag | 14 comments 16:20
23.Pressure builds on Boehner for NSA vote (thehill.com) 3 points by 1337biz 13 minutes ago | flag | discuss
24.First Draft of a Report on the EDVAC (von Neumann, 1945) [scribd] (sites.google.com) 4 points by ptype 2 hours ago | flag | 1 comment
25.Ask HN: How do you come back from a failed acquisition? 58 points by droppedcothrow 8 hours ago | flag | 22 comments
26.Show HN: Probabilistically Generating HN Post Titles (grantslatton.com) 627 points by gamegoblin 1 day ago | flag | 260 comments
27.How Too Many SMS Messages Can Shut Down Your Nexus Phone (allthingsd.com) 37 points by dsr12 12 hours ago | flag | 16 comments
28.What I’ve learned in 5 years of running a SaaS (aculo.us) 236 points by rahulroy 1 day ago | flag | 65 comments
29.IRCCloud (irccloud.com) 94 points by jacobr 19 hours ago | flag | 72 comments
geekosaur o.O
arnsholt Whoops
sjn important news \o/
timotimo hehe. 16:21
Hater News ...
arnsholt Laptop trackpad has buttons close to the edge, and sometimes they get pressed when my laptop is on, you know, my lap
sjn hey arnsholt, wanna join us wednesday and make some decision about organizing something #perl6-relevant in Oslo next year? :) 16:22
timotimo arnsholt: if the device is separate from the other stuff you may have luck with "xinput disable" 16:23
arnsholt timotimo: Ooh, good tip 16:24
I might just disable "both mouse buttons for middle click" outright, since I never use that particular feature
arnsholt sjn: I think wednesday evening is my one non-busy day next week, so I was sort of hoping to stay home that day 16:25
sjn arnsholt: don't worry, we don't have a tradition to stress a lot when Oslo.pm is out drinking beers and discussing things :) 16:27
jnthn But with so much good beer, choosing which one to have next is SO stressful! 16:28
sjn jnthn: ssh! 16:29
timotimo so i just had an idea for an optimization yet again. if we have a scan followed by a literal, we do boyer-moore, but if we have a scan followed by an alt/altseq of literals, we don't do that.
sjn jnthn: don't tell him about the Beer Selection Stress Syndrome
timotimo i wonder how hard it would be to make a boyer-moore derivative that works with multiple needles, matching the first one it finds
timotimo also, i think it would be nice if we had the same optimization for .*? that we have for scan, maybe for .* with a boyer-moore-from-the-end search, too. 16:30
arnsholt timotimo: I think it should be possible to merge two Boyer-Moore jump tables. For any character, the merged table has the minimum of the value in the two tables 16:35
timotimo i thought something about the same lines, yeah 16:36
arnsholt That increases my prior confidence that the reasoning is correct =)
timotimo hah, i only thought about it idly ;)
arnsholt Me too =) 16:37
timotimo now i'm thinking just merging the tables will cause a big performance loss. if instead you use a table per string and always advance the leftmost needle, you'll be correct 100% and you won't get a big penalty if your patterns are unlucky 16:41
arnsholt Yeah, that sound like it'll work as well 16:42
timotimo i think making use of .*? + literal should be closer to the top of my list, though 16:44
arnsholt Probably an easier optimisation, for sure 16:48
timotimo hopefully :) 16:49
i'm not yet sure how exactly our capture stack works, so i'll have to read the code for that first.
jnthn I think I may have described $!cstack in the workshop slides... 16:50
timotimo i probably won't have time for that today 16:52
timotimo is there a flag i can read out in the compiler that tells me if we're compiling to serialise a module or directly execute code? 17:12
jnthn yeah
timotimo that's good
jnthn But I dunno where.
Search for precomp_mode or precompilation_mode or so 17:13
timotimo now that i know it exists, i'll have an easier time looking for it :)
jnthn Where are you, btw?
(when you want to know this)
timotimo i'm thinking of either the regex compiler or the optimizer
jnthn If you've a World instance handy I think that knows.
timotimo i do in the optimizer. not sure about the compiler. 17:14
jnthn yeah, you have one in both of those scenarios.
timotimo good
another thing, is it feasible to build a hot-regex-detection in NQP?
so that we can do some extra-expensive optimisations only if the regex is sufficiently hot?
jnthn We don't statically tend to knwo about hotness 17:15
that's more a runtime-optimization thing.
timotimo yes, exactly
jnthn So best you have is in heuristics 17:16
timotimo i was hoping for run-time statistics gathering
keeping the qregex QAST tree around maybe.
jnthn Ah...that sounds like it'll get costly to keep 'em all around...
arnsholt We don't really do any run-time fiddling with the executed code ATM, do we?
timotimo i don't think we do
jnthn No, that's very much VM-space. 17:17
arnsholt And JIT-space, I guess? 17:17
jnthn That too 17:17
jnthn Though MoarVM will do type specialization stuff before it gets JIT. 17:18
That will kinda be the input to the JIT...
TimToady well, you can make a first-order guess that the inside of a loop is hotter than the outside :) 17:19
jnthn Aye. And loop in regex is quantifier, I guess... :) 17:20
timotimo and after scans
(though not necessarily if there's a literal following the scan directly)
TimToady but the basic problem of an interpreter is that anything could be hot in the right circumstances 17:21
timotimo yes 17:22
timotimo berekuk: i like questhub a lot :) 17:49
especially the way you can re-order your items by dragging and dropping is great 17:50
berekuk timotimo: thanks :D
timotimo i was also happy to see that it responds well to being displayed on a phone, or with varying browser width 17:51
but that's probably just bootstrap?
berekuk no, not just bootstrap, I wrote plenty of responsive media queries myself 17:52
actually, QH doesn't use bootstrap grid at this point
timotimo ah, great! :) 17:53
did you read what i wrote about the chrome developer tool audit?
berekuk no..
timotimo let me paste you something
gist.github.com/timo/e25319572c495761dd60 17:54
"d" seems to refer to a font from edgefonts.net 17:55
berekuk right. I'd like to get all static stuff on CDN, it'd speed things up significantly 17:58
TimToady moritz: s/lead to/led to/ 18:10
and after the "sad news" you might add that the new plan is to target .NET as an additional backend for rakudo (indeed, the possibility of this is one reason development of niecza has slowed) 18:19
TimToady nr: sleep 0; say "cool" 18:25
camelia rakudo-parrot 1d729f, rakudo-jvm 1d729f, niecza v24-98-g473bd20: OUTPUT«cool␤» 18:25
colomon Hasn't the biggest reason for Niecza development stopping simply that sorear++ and to a lesser extent I have been too busy to work on it? 18:35
timotimo maybe 18:35
colomon deeply regrets not having more time to work on p6 at the moment, and hopes things will be better in 2014. 18:39
TimToady wow, a year ago there were only 4 spec changes in December 18:43
colomon this year there haven't been any! ;)
TimToady things tend to freeze in December 18:44
colomon this year in November, round these parts.
TimToady they tender to freezer morer in Decemberer 18:45
FROGGS does somebody know what happened to dalek? 19:17
TimToady it's resting 19:18
FROGGS :/
lue
.oO(Its vision has been impaired!)
19:18
TimToady how do you cherrypick a single commit from a different branch? 19:23
FROGGS TimToady: ariejan.net/2010/06/10/cherry-picki...er-branch/ 19:25
git cherry-pick 62ecb3 # not more than that as it seems 19:26
TimToady thanks
FROGGS yw :o) 19:27
Mouq o/ 19:30
FROGGS o/
Mouq I'm not really sure what's actually supposed to happen, but there is an interesting difference here: 19:35
r: my ($a, :$named) = 1, :named<3>; say $named
camelia rakudo-parrot 1d729f, rakudo-jvm 1d729f: OUTPUT«"named" => "3"␤»
Mouq r: my ($a, :$named) := 1, :named<3>; say $named
camelia rakudo-parrot 1d729f, rakudo-jvm 1d729f: OUTPUT«3␤»
FROGGS I think binding and assignment have different precedence levels
or.... 19:36
Mouq r: my (:$named) = (:named<3>); say $named
camelia rakudo-parrot 1d729f, rakudo-jvm 1d729f: OUTPUT«"named" => "3"␤»
Mouq r: my (:$named) := (:named<3>); say $named
camelia rakudo-parrot 1d729f: OUTPUT«2 unexpected named parameters passed (key, value)␤ in block at /tmp/rZ3OY0l2IH:1␤ in any at /tmp/rZ3OY0l2IH:1␤ in any at gen/parrot/stage2/NQPHLL.nqp:1146␤ in any eval at gen/parrot/stage2/NQPHLL.nqp:1133␤ in any evalfiles at gen/parrot…»
..rakudo-jvm 1d729f: OUTPUT«2 unexpected named parameters passed (value, key)␤ in block at /tmp/xaF2NIo7Og:1␤ in any eval at gen/jvm/stage2/NQPHLL.nqp:1086␤ in any evalfiles at gen/jvm/stage2/NQPHLL.nqp:1292␤ in any command_eval at gen/jvm/stage2/NQPHLL.nqp:1196␤ in an…»
FROGGS what?
Mouq vOv
timotimo why what? 19:39
EZ. on the left side of the = you have a signature with a named parameter
on the right you have an object that unpacks to .value and .key
Mouq So that's... correct behaviour? 19:40
timotimo i think so
r: my (:$named) := \(:named<3>); say $named
camelia rakudo-parrot 1d729f, rakudo-jvm 1d729f: OUTPUT«3␤»
timotimo this is how you do what you meant
Mouq Ahhh. timotimo++
timotimo yw :) 19:42
Mouq That also makes the fact that `my ($a, *@a) := 1;` doesn't work more sensible. It's fine with `\1`
timotimo yup. 19:43
sometimes i amaze myself with what cool stuff i know about perl6 :)
moritz maybe it should just die with "Can only bind Capture objects to a Signature, but you passed in a $.whatitstype" 20:14
ajr_ There are 2 perl6-p in 2013.11:rakudo-star-2013.11/perl6-p and rakudo-star-2013.11/rakudo/perl6-p 20:18
preflex ajr_: you have 1 new message. '/msg preflex messages' to read it.
ajr_ Is the first one the definitive one?
moritz ajr_: the definitive one is the one in install/bin/ 20:19
TimToady are those just shell scripts? 20:19
moritz no, perl6-p are fakexecutables
TimToady oh, -j and -m are shell scripts 20:20
moritz aye, and .bat files on weird OSes 20:24
:-)
ajr_ Thanks, moritz. 20:25
lue r: say "abc \x[FFFE] def" 20:27
camelia rakudo-parrot 1d729f: OUTPUT«===SORRY!===␤Invalid character for UTF-8 encoding␤␤»
..rakudo-jvm 1d729f: OUTPUT«===SORRY!===␤Invalid code-point U+0FFFE␤»
lue If I understand this correctly, that shouldn't be happening: www.unicode.org/versions/corrigendum9.html 20:28
moritz lue: yes, that's how it sounds to me too 20:29
raydiak moritz: I've been mulling over my entry for the advent calendar, and am wondering if there are any (beginner to intermediate) specific topics you were hoping to see which aren't yet covered. 20:53
TimToady n: say "abc \x[FFFE] def" 20:55
camelia niecza v24-98-g473bd20: OUTPUT«abc \x{FFFE} def␤»
TimToady ⊙⨀ 20:56
⧁°⧀
funny that niecza uses a p5-ism to indicate that 20:57
TimToady pugs: say "abc \x[FFFE] def" 21:00
camelia pugs: OUTPUT«abc \x{FFFE} def␤»
TimToady well, there's prior art :)
lue FROGGS: too bad slangs themself don't work in Rakudo yet, or else someone (perhaps I) could have written a day on them, before your v5 slang. :) 21:04
FROGGS :o)
lue r: my int8 $a = 0xFE; say ++$a; # I can guess why this fails, but it still un-DWIMmy to me. 21:08
camelia rakudo-parrot 1d729f: OUTPUT«Cannot modify an immutable value␤ in sub prefix:<++> at gen/parrot/CORE.setting:4270␤ in sub prefix:<++> at gen/parrot/CORE.setting:1732␤ in block at /tmp/pPGJwLwjlm:1␤ in any at /tmp/pPGJwLwjlm:1␤ in any at gen/parrot/stage2/NQPHLL.nqp…»
..rakudo-jvm 1d729f: OUTPUT«Cannot modify an immutable value␤ in sub prefix:<++> at gen/jvm/CORE.setting:4259␤ in block at /tmp/Dq17aeGLLN:1␤ in any eval at gen/jvm/stage2/NQPHLL.nqp:1086␤ in any evalfiles at gen/jvm/stage2/NQPHLL.nqp:1292␤ in any command_eval at gen/j…»
TimToady that should fix itself when 6model learns about C pointers as a kind of container 21:09
timotimo lue: NYI :( 21:10
TimToady so we can have a C-style lvalue, basically
lue timotimo: also NYI,
r: my int8 $a = 0x100;
camelia ( no output )
TimToady the problem being that a C-style pointer is itself a kind of native type 21:10
timotimo i'm looking forward to that. i think that'll also almost immediately give t^Hcompact arrays and such
TimToady lunch & 21:11
lue one of the brainstorm ideas for the advent describes "unary sort", but I'm never able to actually find out what this means. 21:11
timotimo oh? 21:12
colomon r: say (1, 4, 323, -1, 0).sort(+*)
timotimo it's easy
camelia rakudo-parrot 1d729f, rakudo-jvm 1d729f: OUTPUT«-1 0 1 4 323␤»
colomon lue: ^^
timotimo yeah, give sort a callable that only takes one argument
and it does the swartzian transformation
also, make that callable return a list of comparables to implement sub-sorting
colomon or we hope it does the swartzian, anyway. ;)
(sp?)
lue Ah. Here I was thinking it was a type of sorting algorithm, like "bubble sort" :) 21:13
lue (Maybe call it "Unary &sort" instead?) 21:13
geekosaur Schwartzian 21:14
colomon geekosaur++ 21:15
lue Also, the advent reminded me I should make some updates to S32::Temporal. Is linking to Wikipedia an acceptable "see details" link? (There's an RFC too, though it apparently doesn't *exactly* mirror ISO 8601) 21:16
retupmoca nativecall issue: if I do "is native('libz')" I get "Cannot locate native library 'libz.so'" 21:27
if I do "is native('libz.so.1')" it works 21:28
this may be related to a gentoo linux weirdness as well: gist.github.com/retupmoca/7724686
retupmoca (or rather, "is native('libz.so.1')" works with a bugfix to NativeCall: github.com/jnthn/zavolaj/pull/30) 21:29
so, long story short, how does NQP currently handle a .so that is a linker script (like in the gist above)? 21:31
lue retupmoca: it doesn't, from what I recall. 21:32
retupmoca ok 21:34
are there plans to handle it, ooc?
jnthn retupmoca: At present, we (some extension-adding aside) hand things off to dlopen or the OS equivalent fairly directly. I'm guessing those functions are not too aware of linker scripts... 21:36
preflex jnthn: you have 3 new messages. '/msg preflex messages' to read them.
jnthn retupmoca: Merged your PR; thanks. 21:38
retupmoca jnthn: tyvm - I'll just load .so.1 directly then 21:41
jnthn arnsholt++ may have some more thoughts on this. I'm the original NativeCall guy, but he's done all the hard work of late on it. :) 21:42
jnthn I'm guessing either somewhere it's spec'd how dynamic loading should interact with linker scripts, or it's not and they're just an annoying mess. 21:42
FROGGS the problem with linker scripts is that you need to locate them
FROGGS like you would need to locate a common lib when not using dlopen for that 21:43
jnthn TimToady: re 73b2fe31, I'd rather we don't leave commented out code hanging around in Rakudo. There's always version control if we want to find what it used to be. 21:55
timotimo: The setautothreader patch looks good 21:56
moritz retupmoca: re advent, nothing in particular, but there are lots of ideas in github.com/perl6/mu/blob/master/mi...instorming
(and it's actually PerlJam who seems to manage it this year)
lue wonders if there isn't a program or something that can handle the linker scripts without too much issue. 22:00
.oO[ Doing work on git-repo'd stuff feels a lot more dangerous without dalek around to tell me if commits have happened since the last time. :) ] 22:01
retupmoca it looks like the error message from dlopen will give you the full path to the file, at which point we could run a regex or something to try and get the real .so 22:02
raydiak moritz: thanks, that list is the spark of creativity I was looking for :) 22:06
timotimo jnthn: thanks :) 22:09
took me a while to notice $Binder down there ;)
lue Since dalek's not around, here are some specs commits I just pushed: github.com/perl6/specs/commit/c323...09310cb98a and /97c5bc4e47b819b640fc5f815b861d753ed8aa9c 22:46