»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.
00:01 Hor|zon joined, xenoterracide left, xenoterracide joined 00:02 BenGoldberg joined 00:06 Hor|zon left 00:10 xenoterracide left 00:13 xenoterracide joined 00:16 xenoterracide left 00:17 xenoterracide joined 00:22 xenoterracide left 00:23 ajr_ left 00:27 Mouq joined 00:28 xenoterracide joined 00:33 xenoterracide left 00:38 panchiniak joined 00:46 xenoterracide joined 00:51 xenoterracide left 00:59 atroxaper joined 01:00 Hor|zon joined, kbaker left 01:04 atroxaper left 01:06 Hor|zon left, xenoterracide joined 01:14 _slade_ left, Mso150_g_s_s_j joined 01:15 nbrown_ joined 01:26 _slade_ joined 01:30 panchiniak left
gtodd except ... the aliens abduct all the ways of building software: auto-%@$, CMake ,rake, scons, waf, ant, ninja ... Inf 01:42
01:42 gtodd left 01:46 ilbot3 joined 01:49 Mso150_g_s_s_j left 01:52 yeahnoob joined 02:01 Sqirrel left 02:07 noganex_ joined 02:09 noganex left 02:22 leont left 02:42 vukcrni left 02:43 vukcrni joined 02:44 simcop2387 left 02:45 simcop2387 joined 03:02 Hor|zon joined 03:04 erkan left, erkan joined 03:06 Hor|zon left 03:10 rindolf joined 03:11 fhelmberger joined 03:15 Mouq left 03:16 fhelmberger left 03:27 ChoHag left 03:44 ChoHag joined 03:54 ChoHag left 03:55 [Sno] left, Sqirrel joined 04:06 ChoHag joined 04:07 anaeem1_ joined 04:16 atroxaper joined 04:19 virtualsue joined 04:20 atroxaper left 04:22 BenGoldberg left 04:32 xenoterracide left 05:03 Hor|zon joined
firefish5000 m: class Person { has $.Is=$0 if "very sensable" ~~ m{(.+)}; has $.IsNot=$0 if "Insane" ~~ m{(.+)}; }; my $FireFish = Person.new(); say "I am {$FireFish.Is}"; 05:06
camelia rakudo-moar b5032f: OUTPUT«(timeout)» 05:07
05:07 Hor|zon left 05:11 Pleiades` joined, kaare_ joined 05:12 xenoterracide joined 05:17 firefish5000 left 05:20 [Sno] joined, SamuraiJack_ joined 05:25 bjz left 05:28 bjz joined 05:34 ecocode joined 05:37 bjz_ joined
vendethiel o/, #perl6 05:37
05:40 bjz left, xenoterracide left 05:43 odc` left
ingy \/\/\o vendethiel! 05:53
06:05 darutoko joined 06:28 jack_rabbit left 06:32 Mso150 joined 06:33 yeahnoob left 06:36 jack_rabbit joined 06:50 jack_rabbit left
moritz \o 06:51
06:51 ghostlines joined
_slade_ oハハ八八ノヽノヽノヽノ \ / \/ \ 06:52
06:57 jack_rabbit joined 06:58 FROGGS joined 07:04 Hor|zon joined 07:08 Hor|zon left 07:16 ab5tract joined 07:17 kjs_ joined 07:19 denis_boyun_ joined 07:20 _slade_ left 07:21 FROGGS left 07:30 pdcawley joined 07:32 FROGGS joined 07:35 atroxaper joined 07:37 kjs_ left 07:38 labster joined 07:39 atroxaper left 07:40 ab5tract left
dalek href="https://modules.perl6.org:">modules.perl6.org: 24e8143 | raydiak++ | web/index.tmpl:
Re-enable js search w/less truncation. See issue 5
07:47 zakharyas joined, fhelmberger joined 07:50 kaleem joined
moritz raydiak++ 07:51
raydiak :) thanks...doesn't fix the dt/dd issue, but I figured it's better than a broken search box (or none at all) 07:53
nwc10 timotimo: Elizabeth has just spotted the Camelia on the APW site - "wow - that's very nice" 07:55
TimToady: Elizabeth has just spotted the Camelia on the APW site - "wow - that's very nice" 07:56
timotimo: that might also be intersting to you, but it's probably more amusing to him
FROGGS nwc10: I like the shirt design very much btw :o) 07:59
nwc10 FROGGS: I don't think that the designer will be in Salzburg, but you'll probably see her friend/our contact at the same time as me, so you could pass that on directly. 08:01
FROGGS nwc10: will do that :o) 08:02
08:07 Hor|zon joined 08:09 jack_rabbit left 08:11 ab5tract joined
dalek href="https://modules.perl6.org:">modules.perl6.org: fd86dda | moritz++ | web/lib/P6Project/Hosts/Github.pm:
Fix error reporting on older Mojolicious versions
08:14 Ven joined
Ven what's the last date for registering a talk to apw2014? 08:16
08:16 yakudza joined
moritz modules.perl6.org now all fixed, and 90% awesome 08:18
FROGGS Ven: perhaps ask nwc10 directly 08:20
Ven nwc10: ping :)
08:20 virtualsue left
nwc10 Ven: hi. I might vanish without notice 08:21
Ven nwc10: hi. what's the last date for submitting a talk (apw2014)?
nwc10 Ven: other than lightning talks, it may well have passed. The schedule is already very full 08:22
I have actually lost track of what the revised date was
I'm on holiday (or is that "holiday") and I'm really not in a position to work anything out 08:23
Ven ha, because the form is still available
nwc10 if you have sufficient abstract ready that it's a copy-paste job, then you're welcome to submit a talk, but the others might have to reject it because it won't fit 08:24
Ven: that wouldn't surprise me
Ven eh
raydiak moritz++ # un-breaking things
nwc10 Ven: I've not been looking after the site, nor dealing with the talks 08:25
moritz raydiak: I also broke it in the first place :-)
Ven nwc10: alright. Next one: do you need to submit your real name in order to do a talk?
nwc10 Ven: I don't think so. Zefram often doesn't use his real name, nor does chromatic 08:26
raydiak moritz: you wouldn't have broken it if you weren't improving it in the zeroth place :)
moritz I think of "chromatic" as chromatic's real name. It's very real to me, even it's not his legal name
08:27 bartolin joined
Ven Well, I guess I'll submit "Perl 6 and the other cheek" 08:27
raydiak sleep time...good morning and good night #perl6 o/ 08:31
08:34 slavik left 08:35 pmqs joined
Ven well, submitted a 20min talk, I hope it'll go in :( 08:36
08:37 pmqs left 08:46 dakkar joined 08:47 Mso150 left, slavik joined 08:55 mberends joined 08:59 ab5tract left 09:00 kjs_ joined 09:08 virtualsue joined 09:10 Hor|zon left
timotimo is it that x creates strings and xx creates lists because a string is one thing and a list is many things? 09:20
(very roughly)
moritz yes 09:21
timotimo cute 09:22
09:24 zakharyas left 09:27 zakharyas joined, rurban joined 09:30 rurban left 09:32 Hor|zon joined, mberends left 09:34 mberends joined, slavik left 09:39 Hor|zon left, lsm-desktop joined 09:42 pepl joined 09:45 Hor|zon joined 09:46 slavik joined 10:11 pmqs joined 10:14 mberends left 10:21 Hor|zon left, xinming_ joined 10:25 xinming left 10:27 lizmat joined, anaeem1_ left 10:29 Hor|zon joined, anaeem1_ joined 10:35 nbrown_ left 10:45 anaeem1__ joined 10:46 anaeem1__ left 10:47 anaeem1 joined, anaeem1_ left 10:55 atroxaper joined 10:58 prevost joined 11:00 atroxaper left 11:05 nbrown_ joined 11:13 ghostlines left 11:15 nbrown_ left 11:16 ghostlines joined
dalek kudo/nom: e4adb29 | (Elizabeth Mattijsen)++ | src/core/IO/Spec/Unix.pm:
Remove my @ sillyness, any Iterative will do
ast: d8a8632 | (Elizabeth Mattijsen)++ | S32-io/io-spec-unix.t:
Make IO::Spec::Unix.path tests more sensible
11:23 anaeem1 left, anaeem1_ joined
dalek ecs: b89831b | (L. Grondin)++ | S32-setting-library/IO.pod:
fix minor syntax error
11:31 SamuraiJack_ left
dalek ecs: 66edfbe | (L. Grondin)++ | S32-setting-library/IO.pod:
adding missing =over
ast: fc2dbbe | usev6++ | S04-declarations/multiple.t:
Add ticket number RT #83430 to existing test
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=83430
11:49 leont joined
lizmat errands& 12:02
12:04 pepl left 12:07 Ven left 12:08 Hor|zon left 12:10 nbrown_ joined 12:14 yeahnoob joined 12:15 nbrown_ left 12:22 pepl joined 12:23 __rubio__ joined 12:26 FROGGS left
timotimo it's been a month since the first batch of videos of yapceu has been released; i wonder what's keeping them and if we can do anything to help? 12:30
.o( or even obtain raw videos for the talks i'm interested in :P )
12:31 kjs_ left 12:32 __rubio__ left
dalek kudo-star-daily: b746522 | coke++ | log/MoarVM-version.log:
today (automated commit)
12:38 SamuraiJack_ joined 12:41 kbaker joined, darutoko- joined 12:43 darutoko left 12:44 __rubio__ joined
dalek rl6-roast-data: c8f369f | coke++ | / (5 files):
today (automated commit)
rl6-roast-data: b535377 | coke++ | bin/ (2 files):
Remove rakudo.jvm from the daily runs

it's taking over 24 hours to run rakudo with moar,moar-jit,parrot & jvm. Nearly all the time is spent on rakudo.jvm. - this is a performance regression in the past 2-3 days.
12:45 yeahnoob_ joined
dalek rl6-roast-data: 800882f | coke++ | perl6_pass_rates.csv:
annotate last 2 jvm runs.
12:47 kbaker left 12:48 yeahnoob left
grondilu timotimo: no P6-related videos yet indeed. It's almost as if they were censored or something. 12:50
timotimo i find that highly unlikely
12:51 gtodd joined
grondilu what's the explanation then? It's not like posting video on YouTube takes a month, is it? 12:51
timotimo could be anything, really 12:52
the person who was supposed to cut the videos may be on vacation/sick
computer dead
12:52 yeahnoob_ is now known as yeahnoob
timotimo trouble extracting the raw video from the cameras? 12:52
backup tapes can't be located?
grondilu if there was a technical incident it would be mentioned somewhere, wouldn't it? 12:53
12:53 atroxaper joined
jkg I tend to assume "volunteers running out of spare time" 12:53
yeahnoob where's the video? 12:54
timotimo www.youtube.com/user/yapceu/videos 12:55
[Coke] never assume malice where life is a probably cause.
*probable. (arglebargle!)
yeahnoob Maybe i can Translate&Add Chinese language subtitle for it. if it's not too long. 12:56
12:58 darutoko joined 12:59 darutoko- left
timotimo is someone in a yapceu channel or something and can ask what's up? 12:59
13:02 Hor|zon joined
grondilu can tweet them 13:02
timotimo that would be interesting
13:03 anaeem1_ left
grondilu not sure I still quite understand how twitter works but here is my attempt: twitter.com/grondilu/status/517298929965477888 13:06
timotimo that sounds a bit more like "fuck you, give me the videos", whereas i would have suggested something more like "can we help get this out faster?"
grondilu yeah with hindsight that was not very diplomatic 13:08
timotimo they are not our enemy, using the word "diplomatic" is also undiplomatic in this situation :P
yeahnoob " 25/08/14 Lost laptop Guys, we have lost one of the conference laptops. A metallic Asus Zenbook. If anyone have found it. Please contact me."
grondilu it's a good thing I can remove tweets
yeahnoob from act.yapc.eu/ye2014/. Is that the lost video in the ladtop? 13:09
timotimo yeahnoob: look a line above that: NB: talks from the main room Vitosha will take a little longer, as they had a much more complicated setup. We'll post here and on twitter as soon as they're ready!
sounds like the videos wouldn't be on that laptop
13:10 nbrown_ joined
nine It's quite common for converence videos to take even a couple of months to show up online. 13:12
grondilu That's quite difficult to believe, but I guess in these matters there are factors I'm totally oblivious of. 13:14
timotimo unless you've contracted NextDayVideo
13:15 nbrown_ left 13:17 kjs_ joined 13:18 Hor|zon left, prevost left
PerlJam wonders how complicated the setup was for Vitosha that it's taking so long ... the videos for the other rooms were up in just a couple of weeks. 13:20
13:24 Hor|zon joined 13:30 prevost joined
nine Volunteer organized conference: lots of work, getting more and more stressful the closer the conference is, followed by a couple of extremely intense days with even more work and stress. Then it's finally over. You need some time to relax. Then some time to follow up on lots of stuff you pushed out because you were busy with the conference. Good that the conference is over. Oh I forgot, there's some videos to upload. Not much motivation because there's no deadline 13:30
13:31 prevost left, prevost joined
nine Also the complicated setup could be that video and audio were not recorded on the same device. So they'll have to cut and sync them again. Tedious work. 13:33
PerlJam aye. 13:36
13:41 yeahnoob left 13:50 molaf joined
gtodd nine: hmm if only ACT included a builtin perl/jscript based video web app that could grab footage from laptop onboard cams .. :-) 13:55
13:57 kbaker joined
moritz that would be aweful quality 13:57
gtodd heh 13:58
nine Better wait for a few months than hardly understanding a word due to low audio quality ;) 13:59
gtodd hehe yeah but in those scifi futures in the movies automagically streaming recording, editing, subtitlling, translating and mixing randomly errm "supplied" video streams is trivial .... 14:00
because in the future they have p6 :-)
moritz I'm mostly amazed that in scifi movies, they can immediately decode any incoming video stream 14:04
nine moritz: they obviously just use mplayer
moritz no missing codecs or unknown streaming formats etc.
nine: :-)
gtodd nine: :-D 14:08
in the future you never have upgrade "Adobe Flash Player" 14:09
(which always makes me think "Adobe bought Flash? Oh yeah ...")
14:11 xenoterracide joined 14:16 nbrown_ joined 14:17 kjs_ left 14:21 kjs_ joined, nbrown_ left 14:24 spider-mario joined 14:26 ecocode left, ecocode joined 14:27 xenoterracide left 14:29 yeahnoob joined 14:30 treehug88 joined, ecocode left 14:40 Akagi201 left, Akagi201 joined 14:41 FROGGS joined
gtodd camelia: m: say "jubilation"; 14:47
camelia gtodd: rakudo-moar e4adb2: OUTPUT«jubilation␤»
gtodd speedy
14:51 arcterus joined 14:52 pepl left 14:54 pepl joined 14:55 arcterus left 15:00 virtualsue left 15:14 yeahnoob left 15:18 nbrown_ joined 15:22 nbrown_ left
japhb botsnack 15:32
yoleaux 30 Sep 2014 19:08Z <timotimo> japhb: the only relevant thing i see is --skip-incomplete, but it doesn't change anything :(
15:32 rindolf left
japhb timotimo: OK, back. So ... are you running `bench compare` or `bench history`? 15:33
timotimo: Also, pretty please p6weekly? It's may main source of #perl6 goings on these days. :-)
15:34 ghostlines left 15:38 denis_boyun_ left 15:43 Hor|zon left, pmurias joined 15:45 virtualsue joined 15:48 Hor|zon joined 15:53 kaleem left 15:58 Akagi201_ joined 16:00 virtualsue left 16:01 Akagi201 left, Hor|zon left 16:09 zakharyas left 16:11 atroxaper left 16:12 colomon_ joined 16:14 [Sno]_ joined 16:15 [Coke]_ joined, retupmoca joined, daxim_ joined, itz_ joined 16:16 avuserow left, rubio_terra joined 16:17 kaleem joined, [Sno] left, daxim left, retupmoc1 left, [Coke] left, itz left, colomon left, spider-mario left, __rubio__ left, araujo left, odc left, Tene left, PerlJam left, perlpilot joined, colomon_ is now known as colomon, robins joined, [Sno]_ is now known as [Sno], Tene joined, Tene left, Tene joined 16:18 atroxaper joined, nbrown_ joined 16:19 cotto joined, odc joined, araujo joined 16:21 atroxaper left, spider-mario joined, Mso150 joined 16:23 nbrown_ left 16:25 Hor|zon joined 16:28 avuserow joined 16:33 kaleem left 16:41 Mso150 left 16:42 dragon_d joined, dragon_d left 16:44 Mso150 joined 16:46 FROGGS left
timotimo japhb: this was bench history 16:47
also ... oh yeah the weekly :S
gtodd heh 16:50
timotimo my energy levels seem to be far from over 9000 these days
gtodd could change the name to p6timely
and or weekly could be read as a range ... 16:51
38 .. 41
16:53 _slade_ joined
dalek rlito: b0e7ab7 | (Flavio S. Glock)++ | Makefile:
Perlito5 - use 'nodejs' in Makefile
16:57 Akagi201_ left
dalek rlito: 6f48b7a | (Flavio S. Glock)++ | / (3 files):
Perlito5 - fix associativity of && and ||
grondilu m: multi f($a, $b where $a|$b == 0) { 0 }; multi f($a, $b where $a == $b) { "==" }; say f(0, 0); # can I rely on this behavior? (higher priority for junctions or something? 17:02
17:02 dakkar left
camelia rakudo-moar e4adb2: OUTPUT«0␤» 17:02
timotimo that junction isn't visible to the multi dispatch
it's part of an opaque "where block" type of thing
grondilu doesn't understand 17:03
timotimo the multiple dispatch code doesn't see a difference between the $a|$b == 0 and the $a == $b things 17:04
grondilu then why does he pick one?
vendethiel first one?
tries to match it, it does, yay
grondilu I would have expected an "ambiguous" error or something 17:05
vendethiel it's not ambigous
ambiguous*. Not really.
It's very useful to have "cascading matches"
grondilu m: multi f($ where * %% 2) { 2 }; multi f($a where * %% 3) { 3 }; say f(6); # indeed it picks the first one 17:06
camelia rakudo-moar e4adb2: OUTPUT«2␤»
17:07 Hor|zon left
grondilu I don't quite understand how I managed to get an "ambiguous call" error once. (in a very different code, though) 17:07
grondilu doesn't remember the exact phrasing of the error
grondilu looks in rakudo's code 17:08
grondilu finds X::Multi::Ambiguous
grondilu then wonders when exactly X::Multi::Ambiguous is raised 17:09
timotimo m: multi f(Int $a, Cool $b) { 1 }; multi f (Cool $a, Int $b) { 2 }; f(1, 1);
camelia rakudo-moar e4adb2: OUTPUT«Ambiguous call to 'f'; these signatures all match:␤:(Int $a, Cool $b)␤:(Cool $a, Int $b)␤ in sub f at /tmp/1MmXiAF4nW:1␤ in block <unit> at /tmp/1MmXiAF4nW:1␤␤»
grondilu ok
17:11 fhelmberger left, pmqs left 17:12 chipotle joined 17:19 nbrown_ joined 17:20 FROGGS joined, kjs_ left 17:24 nbrown_ left 17:30 PZt joined 17:32 Hor|zon joined
grondilu m: grep { ($_ == 5) ff ($_ == 7) }, 3, 4, 5, 2, 7, 9, 5; # is there a more concise way to write this? 17:33
camelia ( no output )
grondilu m: say grep { ($_ == 5) ff ($_ == 7) }, 3, 4, 5, 2, 7, 9, 5; # is there a more concise way to write this?
camelia rakudo-moar e4adb2: OUTPUT«5 2 7 5␤»
timotimo m: say grep { 5 ff 7 }, 3, 4, 5, 2, 7, 9, 5 17:34
camelia rakudo-moar e4adb2: OUTPUT«3 4 5 2 7 9 5␤»
timotimo ah, not quite
m: say grep { $_ ~~ (5 ff 7) }, 3, 4, 5, 2, 7, 9, 5 17:35
camelia rakudo-moar e4adb2: OUTPUT«␤»
timotimo also not quite that
how do i ff.
grondilu m: say grep { /5/ ff /7/ }, 3, 4, 5, 2, 7, 9, 5
camelia rakudo-moar e4adb2: OUTPUT«5 2 7 5␤»
grondilu ^this would work but I'd like to keep it numeric
17:35 bloomberg joined 17:37 Hor|zon left
dalek href="https://perl6.org:">perl6.org: ab6916b | raydiak++ | source/compilers/index.html:
Un-break box styling on the Compilers page
timotimo yay
spider-mario: do you have time to build a rakudo-star 2014.09 package on aur soon? 17:41
raydiak o/ afternoon all 17:42
17:45 gfldex joined
timotimo how are you doing, raydiak? :) 17:45
spider-mario oh, sorry, I was going to but I forgot
(school, etc.)
thanks for reminding me 17:46
timotimo that's all right :)
it's just i talked about perl6 with an arch user yesterday
so i noticed the version is behind a month
which is pretty fantastic compared to some other packagings we've had in the past
spider-mario I also provide a binary repo, but I’m not sure many people know about it 17:47
it has little visibility
wiki.archlinux.org/index.php/Unoff...ies#rakudo 17:48
timotimo that looks pretty hidden, aye
raydiak timotimo: eh, this is not a good year for me so far...but un-breaking things is theraputic :) how are you? 17:49
timotimo i'm okay. currently in a writer's block / low energy mode kind of phase :\ 17:50
raydiak do season changes do that to you? 17:51
timotimo i'm not sure
17:52 xfix joined
timotimo i don't remember such a phase from last year 17:52
xfix Hello.
raydiak o/ xfix
timotimo ohai xfix!
dalek href="https://perl6.org:">perl6.org: d2b09dd | (Konrad Borowski)++ | README.md:
Testing site under IE6 and IE7 is unnecessary.

Who uses that, anyway? XP is not supported anymore, anyway.
xfix How is Perl 6 going? The Blogs link on Perl 6 page is dead. 17:55
17:55 molaf left
timotimo ah, yes, planetaria :( 17:55
planetaria kinda died
xfix I see.
timotimo pl6anet.org/ 17:56
when did you last check in?
xfix Long time ago.
Perhaps a few months before.
timotimo so you missed moarvm getting a jit?
xfix No.
I come here very rarely, however.
timotimo OK
jnthn grondilu: "where" clauses in multi-dispatch work like smart-match (they actually do .ACCEPTS on the thing you supply, even), meaning whichever matches first wins. I did try and get it to be ambiguous instead, but TimToady won that one, and these days I think it was probably the right call.
xfix timotimo: Do you think perl6.org should be changed to include pl6anet instead?
timotimo did you notice Inline::Perl5 getting up? 17:57
xfix Yes.
jnthn made it home
And oh wow...it's SO quiet.
timotimo hey jnthn!
glad to see you :)
jnthn Almost unnervingly quiet.
jnthn notes he has been awake > 20 hours by now, so should not be asked anything important this evening ;) 17:58
timotimo is P == NP?
jnthn I did read lots of papers on the plane, and get some decent thinking/design time, though :)
timotimo oh awesome
i guess it was a high percentage of escape analysis related stuff?
xfix timotimo: That still wasn't solved (unless Perl 6 solved it while I was away) :-). 17:59
jnthn timotimo: Yes, if the constant factors are infinite... :P
vendethiel xfix: xp is actually still supported ;-)
timotimo xfix: what is "that"?
xfix Yeah, I know, this Microsoft update for security issue.
timotimo: P == NP.
However, moritz pushed the version to care about from IE6 to IE8, but didn't update the IE5.5 below. 18:00
jnthn timotimo: No, though masak and I did a couple of esc patches at BJ airport today...will push them from my laptop later...or tomorrow...
timotimo oh, cool :) 18:01
xfix And to be honest, I don't think this site works under IE6 anyway. Not that it has to. If somebody still uses IE6, the chance that this someone is interested in Perl 6 is close to 0 (if you still keep this old program...).
jnthn timotimo: Was more pondering native arrays, and parametricity in general...
timotimo that sounds great
jnthn Well, I might have come up with something that solves a bunch of problems at once... :) We'll see. I'll need a day or two to get it down coherently. :)
raydiak is the moar jit enabled by default yet? 18:02
nine jnthn: and probably a night or two to recover ;)
timotimo raydiak: we want that, but i don't know if anyone patched it yet
raydiak so it's pretty much stable? 18:04
timotimo we hope so :) 18:05
worst case we turn the default around before the next release
we want to at least get it tested more heavily from now on
nine Seems like parsing an 11000 lines file with a grammar is a bit too ambitious
jnthn nine: We parse a 20,000+ line one every single Rakudo build... 18:06
timotimo did you see how big the core setting is? we parse that in 25 seconds on moar nowadays
however, that's with a grammar built in nqp
raydiak cool, thanks for the info timotimo++
timotimo and action methods in nqp
which is known to be faster than perl6 in many cases
.oO( nqp-ing is overrated )
jnthn True, there's still some gap to close...
18:06 virtualsue joined
jnthn On JIT by default, I think I said I was fine wiht it, somebody could patch it, or I'd do it once I was back in action :) 18:07
timotimo gist.github.com/timo/fa6d113e897c9f1e1905 - does this seem sane?
xfix My personal opinion is that P == NP, however, you need infinite-length data structures to actually solve NP problems in P time. Not that this is useful in practice, however. 18:09
nine 7 minutes and 2.6 GiB RAM and still running. Is there anything obvious to keep in mind/avoid when writing such a grammar? 18:10
moritz nine: don't use left-recursion
timotimo do you have only a grammar or are you also building stuff with your actions?
moritz nine: also, don't quantify a rule that might match zero characters 18:11
timotimo (like <ws>)
nine timotimo: no actions yet
timotimo hm, interesting 18:12
xfix / ''* / # will always cause an infinite loop with current implementations (in future, Perl 6 implementation may notice this, but don't expect more than exception)
timotimo can you get a shorter input file and run a --profile with your moarvm?
18:12 ghostlines joined
timotimo also, putting some debug say statements at the end of some complex rules or something ... 18:14
or maybe you're trying to print the result of the matching, which may take a loooong time trying to concatenate lots of strings ...
xfix I think that Rakudo debugger supports regexes. 18:15
jnthn timotimo: Yes, though it's more about whether to build the JIT than enable/disable
nine oh indeed
I am printing the results
timotimo jnthn: oh? 18:16
jnthn --no-jit would match better with existing options we have
timotimo: You can disable it with MVM_JIT_DISABLE even if it's built
timotimo oh
well, i can rename them
jnthn timotimo: The command line flag means "don't even try building it"
timotimo but the change makes the previous behavior of --enable-jit the default 18:17
jnthn --no-jit works for me
Right, agree it does the right thing
timotimo or --without-jit
jnthn But no other option we have is called --without
We should at least pretend to believe in consistency... :)
timotimo OK
i must be AFK soon, so not sure if i'll fix up the patch soon 18:18
jnthn should go buy some food :)
timotimo and i still have to write the weekly ...
bon appetit :)
jnthn I'll probably eat something, write a "yes I got home safely don't worry about me" mail, and sleep... :)
18:18 perlpilot is now known as PerlJam
jnthn bbl, or bbt... :) 18:19
18:20 nbrown_ joined
xfix On sidenote, how fast Perl 6 boot time is now (on MoarVM, let's say)? 18:23
Is it faster than Perl 5 with Moose?
18:24 brrt joined, rubio_terra left
grondilu it was said to be so indeed 18:24
timotimo 15% less time than perl5 with moose
xfix Sounds great.
Last time I was here, it was slower than Moose.
18:24 rubio_terra joined
PerlJam Though they're both moving targets as Moose gets improved too 18:25
18:25 nbrown_ left, ghostlines left
xfix Moose was there for a while, it doesn't change that much, does it? 18:25
18:26 psch joined, psch left, psch joined
psch hi #perl6 18:27
xfix Hi, psch.
psch tr/// adverbs are confusing
(also i was apparently too slow to quit irssi and stick it into a screen)
bloomberg timotimo: yeah, the weekly. I've been waiting for that
psch Str.trans adverb are similarly confusing, of course
especially as neither 1) are implemented or 2) specced 18:28
xfix Well, tr/// works like tr in UNIX shell (and y/// in sed, of course).
psch xfix: neither of those have :squash or :complement, afaik
PerlJam xfix: I dunno ... ether did a Moose release 5 days ago.
psch i guess i'll spec them according to what i think the spectest means they mean
xfix Oh right. Well, I personally don't use those flags. 18:29
18:29 pmqs joined
psch although short forms should probably not be confuseable with other short forms for s/// 18:29
xfix (they exist for compatibility with Perl 5, I believe)
perldoc.perl.org/perlop.html#tr/SEA...TLIST/cdsr 18:30
Perl 5 has documentation for these, but it's rather confusing.
PerlJam what's confusing about it? :)
18:31 rubio_terra left
psch xfix: thanks, i hadn't looked there. i guess i don't need to spec them, iff our spec still is "if it's not described, it's as it is in p5" 18:31
xfix The behavior of /c, /s, and /d.
psch where the last not-quoted 'is' is "behaves as", or some such :)
xfix I can see how tr may provide performance enhancements in rare situations, but that doesn't change that its behaviour is rather strange, especially with options. 18:33
flussence for the sake of people's sanity, the spec should probably specify "same as p5" means v5.8.8 (if it doesn't already)
PerlJam xfix: How is its behavior strange? It's always seemed straight-forward to me. 18:34
xfix I just don't get how it works. And why actually use those options.
Sure, you can write tr/a-zA-Z/ /cs. But so you can write s/[^a-zA-Z]+/ /g, which is more readable in my opinion (but to be honest, you should use Unicode letters meta character). 18:35
18:35 [Coke]_ is now known as [Coke]
dalek rl6-roast-data: 07d545b | coke++ | / (4 files):
today (automated commit)
[Coke] ^^ no jvm runs because too slow.
18:37 xenoterracide joined, denis_boyun joined
PerlJam xfix: sounds like you're saying, because this more powerful thing exists that can do the same job, we don't need the special-purpose tool. 18:37
[Coke] Please let me know if moar-jit becomes the default. 18:38
moritz [Coke]: jnthn++ plans to make it the default, if nobody objects 18:39
xfix I just think that tr/// looks like an old feature that maps just exists here to make moving from sed to Perl easier.
[Coke] I assume that some platforms JIT will not occur, so I'll still need to test the non jit version with an env. variable...
xfix that just exists here*
PerlJam moar-jit default +1
xfix: maybe I've just got a soft spot for tr/// since I've used it so long :) 18:40
xfix That makes sense.
[Coke] does anyone have bench runs with recent commits of rakudo.jvm ? 18:41
xfix I just don't see how it's useful, unless you do something really special. I cannot think of examples that wouldn't be contrived, however.
[Coke] btw, rakudo.moar & parrot all test clean.
moritz [Coke]: fwiw I'm installing the oracle java 8 package right now 18:42
PerlJam xfix: flip side ... I just don't see why I should fire up the regex engine just to do some simple character replacements. :) 18:43
xfix That makes sense.
But well, I personally don't care about micro-optimisation. 18:44
18:44 ghostlines joined
PerlJam xfix: sometimes it's about optimization, but often it's also about saying what you mean. 18:44
vendethiel a.replace(from.split X=> to.split) :D
xfix One of contrived examples is tr/234/122/s which replaces all instances of 2 with 1, and all instances of 3 and 4 with 2 while squashing replacements.
I also don't really like thinking about characters. Usually a single character in a string doesn't matter. Entire strings are manipulated. substr is very rare. 18:46
brrt wonders if we're going to see a new flurry of JIT bugs if it will become default 18:48
timotimo m: token word { \w+ }; rule phrase { <word> [ \, <word> ]* \. }; if "hey, there\n" ~~ / <phrase> \n / { say "yay" } 18:49
camelia rakudo-moar e4adb2: OUTPUT«Useless declaration of a has-scoped method in mainline (did you mean 'my token word'?)␤Useless declaration of a has-scoped method in mainline (did you mean 'my rule phrase'?)␤No such method 'phrase' for invocant of type 'Cursor'␤ in method ACCEPTS a…»
vendethiel brrt: computers might break down,yes 18:50
timotimo m: my token word { \w+ }; my rule phrase { <word> [ \, <word> ]* \. }; if "hey, there\n" ~~ / <phrase> \n / { say "yay" }
camelia ( no output )
timotimo m: my token word { \w+ }; my rule phrase { <word> [ \, <word> ]* \. }; if "hey, there.\n" ~~ / <phrase> \n / { say "yay" }
camelia ( no output )
.oO( that's why we need continuations : ATNs ! )
18:52 Hor|zon joined
timotimo what are ATNs? 18:53
also, who knows what's wrong with my matching up there?
it seems to match fine up until the . and then maybe gets confused by the \n? 18:54
vendethiel timotimo: augmented transition network
timotimo m: my token word { \w+ }; my rule phrase { <word> [ \, <word> ]* \. }; if "hey, there." ~~ / <phrase> / { say "yay" } 18:55
vendethiel (tended to be used for sentence parsing)
camelia rakudo-moar e4adb2: OUTPUT«yay␤»
18:55 prevost left, xenoterracide left
vendethiel \n is "n"? 18:55
18:55 xenoterracide joined
vendethiel (in the regexp, that is.) 18:55
timotimo i don't think so :S 18:56
m: say "\n" ~~ / \n /
camelia rakudo-moar e4adb2: OUTPUT«「␤」␤␤»
vendethiel fair enough
timotimo m: say "hello\n" ~~ / \w+ \n /
camelia rakudo-moar e4adb2: OUTPUT«「hello␤」␤␤»
timotimo m: say "hello.\n" ~~ / \w+ \. \n /
camelia rakudo-moar e4adb2: OUTPUT«「hello.␤」␤␤»
18:57 [Sno] left
grondilu good to know : using a double entry cache is much slower than concatenating the keys: 18:59
(my %){rand}{rand} = rand for ^100_000; # ~ 22s on my machine
18:59 Hor|zon left
grondilu (my %){rand ~ "|" ~ rand} = rand for ^100_000; # less than 10s 19:00