»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.
00:12 Psyche^_ joined 00:15 Psyche^ left
dalek rl6-roast-data: eb7d578 | coke++ | / (5 files):
today (automated commit)
00:19
timotimo [Coke]: any way you can look into the cancellations of all those rakudo-jvm tests? 00:20
00:21 ajr_ left
[Coke] timotimo: aye. just coming off a narsty head cold. 00:21
this is the most conscious I've felt since thursday. :) 00:22
00:22 xenoterracide left
timotimo good to hear you're feeling better! 00:22
[Coke] danke. 00:23
I killed most of those manually. it went into the weeds when I commented earlier, and never came back, I manually killed all the subtests just now so I could get -a- run completed. 00:24
diakopter boggles at p6u 00:25
[Coke] fires up chrome... 00:26
timotimo oh
yikes
diakopter eiro: ping
timotimo easy solution: show him how string.split works 00:27
diakopter eiro: your mail to perl6-users made me boggle :) because rakudo doesn't use MoarVM yet...
timotimo oh, that's right. hehe 00:28
seems like there's people outside of our little bubble who are even more optimistic about our progress than us! 00:29
00:38 risou is now known as risou_awy 00:40 ssutch left
Mouq jnthn++ # RaNIW notes 00:41
flussence r: 'č' ~~ /<+:Print -:Space -:ID_Continue>/; # I'm trying to build a regex to filter out "weird" stuff like whitespace and combining chars, but all I can find for the latter is ID_Continue which is way too broad. Any ideas? 00:48
camelia ( no output )
flussence r: say ?('č' ~~ /<+:Print -:Space -:ID_Continue>/); # er... 00:49
camelia rakudo b78da4: OUTPUT«False␤»
flussence r: say ?('c' ~~ /<+:Print -:Space -:ID_Continue>/);
camelia rakudo b78da4: OUTPUT«False␤»
flussence yeah, that's *definitely* not what I wanted... 00:50
TimToady r: say ?('č' ~~ /<+:Print>/); 00:56
camelia rakudo b78da4: OUTPUT«True␤»
00:57 eternaleye left
TimToady 'course, the notion of combining characters will be relatively meaningless under NFG... 00:57
flussence (fwiw, I'm trying to do stuff with this list, which is full of fiddly unicode: www.whatwg.org/specs/web-apps/curre...references ) 00:58
TimToady oh, yes, ID_Continue is not what you want at all 00:59
Mouq r: say ?('č' ~~ /<+:Print -:Expands_On_NFC>/);
camelia rakudo b78da4: OUTPUT«Unicode property 'Expands_On_NFC' not found␤ in regex at /tmp/ync1WVOcfn:1␤ in method ACCEPTS at src/gen/CORE.setting:12080␤ in method ACCEPTS at src/gen/CORE.setting:742␤ in block at /tmp/ync1WVOcfn:1␤␤»
flussence (also I find it pretty amusing that my browser doesn't understand most of those :) 01:00
Mouq flussence: You can peruse www.unicode.org/reports/tr44/tr44-4.html for classes, I don't know if that helps you 01:02
r: say ?('č' ~~ /<+:Modifier_Letter>/);
camelia rakudo b78da4: OUTPUT«False␤»
flussence oh, that looks like a useful page 01:03
Mouq Oop, sorry, www.unicode.org/reports/tr44/ is up to date 01:07
[Coke] here's one NPE on rakudo.jvm: gist.github.com/coke/6679105 01:08
flussence r: say "\x2004" ~~ /<+:L +:N +:P>/; 01:09
.oO( freenode's having a bad day today )
camelia rakudo b78da4: OUTPUT«Nil␤»
01:10 [Sno] left 01:11 araujo left, araujo joined
Mouq TimToady: Where do I look to remind myself what NFG is? 01:13
01:14 pmurias left
TimToady it isn't very well documented yet. Basically it's NFC plus temporary characters generated for those graphemes that don't have a precomposed form. 01:14
timotimo it's a fixed-bytes-per-grapheme format that invents new codepoints for combinations of things in the negatives
TimToady well, negatives is an implementation detail, but yes 01:15
01:15 benabik joined
[Coke] weird. the NPE is reporting from nqp code, not java code. adding --ll-exception still only shows nqp lines. 01:15
TimToady for a string with no combining characters, the NFC form is identical to the NFG form 01:16
Mouq Ohhh, cool. So, Normalized Form Great
:)
timotimo Normalized For Graphemes?
01:25 EvanTeitelman left
diakopter eiro: u around? 01:26
01:26 Vlavv` left 01:30 EvanTeitelman joined 01:31 btyler joined, EvanTeit1lman joined 01:32 EvanTeit1lman left 01:34 xinming left 01:35 xinming joined
TimToady so, basically, we just need to write a bangup QAST to Perl 6 translator, and then anyone who ports their language to the QAST backend has already written the Foolang-to-Perl6 translator :) 01:38
with sufficient refactoring smarts, it can even translate bad Perl 6 to good Perl 6 :) 01:39
01:39 Vlavv` joined
diakopter heh. 01:39
use v6; use Moose;
timotimo oh lord, they were right ... you truly are stark ravin' mad!
diakopter also, stork craving mad 01:40
timotimo hmm, does stork taste good?
01:40 EvanTeitelman left 01:42 EvanTeitelman joined, EvanTeit1lman joined, EvanTeit1lman left
diakopter well, at least my runaway GC isn't consuming all memory anymore 01:43
now it's just spinning wildly
TimToady someday it will wise up and return home
diakopter "bad gc! you're supposed to collect the garbage, not multiply it!" 01:44
(I call it "my" gc since I've sufficiently corrupted it from jnthn's original design... ) 01:47
the parallel-ness of it, anyway
01:49 EvanTeitelman left
benabik Porting QAST to perl 6 isn't hard... It will just be Perl6 that generates a QAST tree and runs it. :-D 02:01
timotimo don't forget there's more to it than just QAST, there's also the declarative bits like classes that'd have to be translated ... but at least perl6's class system is pretty flexible 02:05
(though of course you'd probably have a sixmodel based class/object system in your QAST-targeting language anyway and in theory could load the objects directly from a SC if need be) 02:06
benabik Looks clever although I'll make no guesses at correctness b/c I'm a bit tired. 02:11
02:13 SamuraiJack left
diakopter benabik: the quicksort? 02:13
benabik diakopter: Guh, yes. Sorry about channel hopping. 02:14
diakopter careful, obfuscating your online communications will soon be classed as "evading arrest" ;) 02:16
benabik Vg'f nyevtug, vs gurl gel gb qrpbqr zl zrffntrf, V'yy uvg gurz jvgu gur QZPN.
diakopter I think there's a pony somewhere in there 02:17
02:18 SamuraiJack joined
diakopter boy, who wrote these threads.t tests; they're diabiological 02:20
benabik Okay, enough getting IRC wrong for me. Gotta get up in the morning. 'night all. 02:21
02:21 benabik left 02:22 SamuraiJack left 02:35 SamuraiJack joined 02:39 SamuraiJack left 02:40 SamuraiJack joined 02:51 PZt joined
[Coke] looks like multiple of the jvm failures are due to: java.lang.NullPointerException in block at src/gen/CORE.setting:1051 02:57
03:05 dmol joined 03:20 colomon left 03:24 dmol left 03:25 eternaleye joined 03:29 raiph joined 03:32 gtodd left 03:38 gtodd joined 03:59 preflex left 04:00 preflex joined, ChanServ sets mode: +v preflex 04:05 preflex_ joined, ChanServ sets mode: +v preflex_ 04:06 preflex left, preflex_ is now known as preflex 04:08 raiph left 04:14 ssutch joined 04:52 [Sno] joined 05:03 btyler left 05:05 BenGoldberg left 05:06 hummeleB1 joined 05:11 eternaleye left 05:14 eternaleye joined, eternaleye left, eternaleye joined 05:20 fridim__ left 05:26 zby_home joined 05:30 SamuraiJack left, kaleem joined 05:39 kaleem left 05:40 [Sno] left, SamuraiJack joined 05:42 denis_boyun joined 05:44 xinming left 05:46 xinming joined 05:51 PacoAir joined 05:53 denis_boyun left 05:56 mberends left, PacoAir left 06:12 iSlug joined
eiro hello |@perl6 06:17
06:17 iSlug left
eiro diakopter, now i am 06:17
diakopter eiro: hi :)
I'm glad Patrick replied with the link to the Synopses about the record separator spec'd feature; I had a tiny memory it was there but no idea how to find it 06:18
eiro just read pmichaud's answer, replied to him and i'm trying to figure out if i could be able to patch rakudo and MoarMV 06:20
diakopter well, rakudo doesn't use MoarVM yet, so it'd be no use to patch MoarVM :)
but patching the spec seems more important, since it seems inconsistent at least 06:21
(as you pointed out)
eiro i have to admit i felt dumb when i realized i haven't look at the S*
06:21 zby_home left
diakopter I feel dumb every time I'm reminded I've read something there but forgot it; don't feel bad :) 06:22
eiro :)
if a spec update can be done, it would be really cool to introduce IFS as well: it's usefull in many cases 06:23
diakopter what's IFS
eiro (posix DB, pure CSV or TSV, ISO2709 files)
input-field-separator 06:24
diakopter surely that shouldn't be in the core language..?
(but a module instead?)
eiro \t , ",", ":"
diakopter oh
eiro well. the point of having a builtin for this so common operation is performance: we are always reading records then splitting them with the IFS 06:25
perl -F: -lanE 'say $F[0]' /etc/passwd 06:27
06:29 darutoko joined 06:30 logie left 06:35 FROGGS joined 06:41 denisboyun joined 06:42 brrt joined 06:47 kst left
FROGGS o/ 06:53
06:57 kaleem joined
brrt hi 07:00
hoelzro morning #perl6 07:05
moritz \o hoelzro, brrt, FROGGS, eiro
bonsaikitten bpaste.net/show/134925/ # y'all want a bugreport for that test failure in nqp-2013.09 ?
brrt hi moritz :-) 07:06
07:07 raiph joined
hoelzro in the JVM backend, does ((SixModelObject) smo).st.HOW == xmo? 07:07
s/xmo/smo/
FROGGS bonsaikitten: yes, please
bonsaikitten alrighty then
FROGGS hoelzro: as far as I understand everything on the jvm backend is a smo 07:08
same on moarvm, but not on parrot :o)
bonsaikitten urgh 07:09
some bugtrackers just exceed my tolerance limit in the first minute
well, at least it ain't githurp ... 07:10
moritz huh?
the nqp bugtracker is the one on github
bonsaikitten oh? that's mildly confusing
I'd have wrongly thought nqp is "part of" rakudo as far as bugtracking goes 07:11
oh well, then I'll just have to ignore this bugreport :\
moritz or you just submit the bug on the nqp bugtracker 07:12
bonsaikitten which apparently implies githerpes, which I don't deal with 07:13
hrm, my self-censoring seems to fail today
moritz ok, rephrase
you don't *have* to ignore this bug 07:14
bonsaikitten if the nqp bugtracker lives on github I won't be able to use it
moritz why?
bonsaikitten too sober at this time of the day
FROGGS I'll submit it 07:15
bonsaikitten FROGGS: ok. need any more information?
FROGGS bonsaikitten: many thanks for the report anyway :o)
err
nqp --version 07:16
and your uname -a
that should be alright
arnsholt hoelzro: No, that doesn't sound right
I think the only object whose HOW is itself is KnowHOW (the HOW which ends the recursion in 6model) 07:17
FROGGS .u 01F4A4 07:18
yoleaux U+1F4A4 SLEEPING SYMBOL [So] (💤)
bonsaikitten FROGGS: that's while building 2013.09, and uname -a is Linux localhost 3.11.0-gentoo #1 SMP Wed Sep 4 10:27:45 CST 2013 x86_64 AMD Phenom(tm) II X4 965 Processor AuthenticAMD GNU/Linux 07:19
FROGGS bonsaikitten: I'd still like to get nqp --version, because it includes the parrot version too
bonsaikitten This is nqp version 2013.09 built on parrot 5.7.0 revision 0 # there 07:20
FROGGS awesome, thank you!
07:20 domidumont joined 07:22 Mouq left
eiro FROGGS, maybe i missed a point: is IO.slurp on demand ? (like the haskell readFile ?) 07:23
FROGGS eiro: I don't know any haskell, sorry 07:24
eiro: but yes, Str.IO makes an iohandle, which can be slurped then
ohh 07:25
now I see the mails 07:26
eiro the question is: "foo".IO.slurp.split("\n")[^3]
does it read the whole file to get the first 4 lines 07:27
FROGGS I dont think slurp is lazy
eiro ?
FROGGS we really need to handle open( :nl ) I think, and pass that to nqp::readlinefh() 07:28
eiro so your solution isn't acceptable for me. anyway: i noted the >>.split syntax which is cool :)
FROGGS yeah, my solution isnt one actually ó.ò 07:29
eiro i need rakudo to run on moarvm if i want to patch myself
FROGGS yes, for now 07:30
we should patch moarvm, and then port the readlinefh over to parrot 07:31
07:31 flussence left, flussence joined
eiro an chance to see S32 fixed first ? 07:32
moritz fwiw IO::Socket::INET already has an .input-line-separator
FROGGS eiro: you mean the naming of :nl ?
eiro yep
moritz, so i can netcat my file :) this is a great idea! 07:33
lizmat .tell [Coke] the method2sub branch started using .can in SLICE_MORE in Any.pm, maybe .can is borked on jakudo ? 07:34
yoleaux lizmat: I'll pass your message to [Coke].
eiro FROGGS, not only :nl but .input-line-separator. RS is well known and used for a while, i really don't understand why introducing new conventions
moritz RS = Result Set? 07:35
arnsholt Record separator
eiro moritz, Record Separator
moritz favors :separator with a :sep shortcut
arnsholt Same terms as AWK uses, if that's of any use to you =)
eiro IRS, IFS (shell, ISO charsets, awk, perl5, ...)
FROGGS .lines( :sep("\n") )>>.words( :sep<:> ) # I think I prefer :sep/:separator too 07:38
07:38 logie joined
FROGGS :ifs() is rather cryptic 07:38
FROGGS .oO( insensitive file system? )
eiro unless you have an unix background or an experience as ISO spec reader 07:39
07:39 sqirrel joined
eiro but you're right: it's not a lot of people nowaday 07:39
07:39 hummeleB1 left
FROGGS moritz: do you think we break much when replacing IO::Socket::INET.input-line-separator? 07:41
(if we are going to do that )
07:42 woosley left
eiro also i liked the naming scheme : [IO][RFS]S 07:42
hoelzro oh, I see
yeah, that makes sense
I'm just trying to figure out how to get the "name" of an SMO on the JVM
masak morning, #perl6
hoelzro morning masak
but maybe that operation doesn't even make sense
FROGGS morning masak
eiro 'lo masak
07:43 logie left, woosley joined
hoelzro I've been hacking on tijmp (a object profiler for Java) to try and get it to give more data on a running Perl6 program 07:44
so far, results are good
FROGGS r: my $a = "abc"; say $a.HOW.name($a) # hoelzro, do you mean this? 07:45
camelia rakudo b78da4: OUTPUT«Str␤»
hoelzro FROGGS: right, but from the JVM 07:47
I mean to say
from within Java
FROGGS ahh, hmmm, no idea then :o)
hoelzro hmm 07:48
I'm guessing I have to look up the name method for the HOW object
FROGGS yeah, it must compile to something
hoelzro ugh 07:50
this will be fun
oh well; it's a good chance to learn about the JVM backend =) 07:51
07:52 raiph left 07:54 fhelmberger joined
arnsholt hoelzro: Thing is, I don't think types are explicitly named 07:56
hoelzro I figured
hmm
hoelzro tries an experiment
arnsholt OTOH, the ones that -are- named get names on the Perl 6 side 07:57
07:58 denisboyun left
jnthn morning, #perl6 07:58
yoleaux 23 Sep 2013 23:27Z <TimToady> jnthn: Also, I don't think Trigometric is a word. (Slide 173)
hoelzro welcome back jnthn
jnthn ...so what do you call 'em? :)
hoelzro arnsholt: if I can just catch the ones that are named, that's fine
arnsholt Oh, jnthn probably knows how/if this can be done
hoelzro indeed =)
jnthn What're you trying to do? :) 07:59
hoelzro well
two things
both of which are on the JVM backend
and unrelated to each other =)
the first is a bug I found: rt.perl.org/rt3/Public/Bug/Display...?id=119877 08:00
I tried fixing it, but no dice
the second is probably more interesting
I'm working on modifying a JVM object profiler to work with Perl 6
jnthn hands hoelzro a d20 or two...
08:00 woosley left
hoelzro here's the initial result: hoelz.ro/files/perl6-objects.png 08:01
08:01 flussence left
hoelzro r: (1..20).pick(1) 08:01
camelia ( no output )
hoelzro r: say (1..20).pick(1)
camelia rakudo b78da4: OUTPUT«13␤»
hoelzro =/
anywa 08:02
08:02 flussence joined
hoelzro I would like to display the name of the appropriate SMO in that table 08:02
instead of "org.nqp...SixModelObject"
jnthn I've for a while suspected that we might want to keep a "debugging name" for a type on the STable as a simple string. 08:04
And provide a way for the meta-object to set it
But that's Not Done Yet 08:05
hoelzro of course =)
I was thinking just doing that for P6Opaques would help
jnthn For now your best bet is to arrange to call obj.HOW.name(obj)
You can use Ops.how to get the HOW from an object, or just pull it right out of the ST
There are examples of calling methods in various places. 08:06
It's a little setup work and it won't be fast, but it'll get you back a string type name.
08:06 FROGGS left
hoelzro that's fine 08:07
fast isn't a big deal =)
08:17 baest left, domidumont left 08:18 domidumont joined 08:21 FROGGS joined
dalek kudo/nom: 367963a | (Elizabeth Mattijsen)++ | src/ (20 files):
Sync class attributes comments between BOOTSTRAP and P6 modules for convenience
08:22
08:23 baest joined, woosley joined 08:28 Mouq joined
jnthn morning, lizmat :) 08:29
masak suddenly, commi! 08:30
commit*
jnthn Was gonna say, I didn't spot the communist :P 08:31
masak .oO( an Abelian group is a group where all the elements are communists ) 08:32
08:32 daxim joined
dalek rlito: cac6146 | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar - number special cases
08:34
daxim news.perlfoundation.org/2013/09/hag...ccept.html 08:36
congratulations
moritz \o/ 08:37
diakopter++
jnthn \o/
diakopter++
And, ++diakopter # need to do the grant now :D
arnsholt Oh, awesome! 08:38
diakopter heh.
masak diakopter++ ++diakopter
diakopter jnthn: fsdo "now"... 08:39
dalek rlito: 2a9824c | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar - number special cases
08:41
FROGGS diakopter++ \o/ 08:42
08:45 woosley left
tadzik finally, leopard 08:46
diakopter++
diakopter well, seeing as how she said I could start weeks ago... I already have anyway 08:47
hoelzro diakopter++ 08:48
08:48 sqirrel left
hoelzro just saw the news! 08:48
diakopter
.oO( report: 3 hours last Tuesday night solving coding problems in my sleep..; 2 hours daydreaming in $meetings about project )
08:49
(kidding entirely)
masak well. been there, done that. 08:53
diakopter meanwhile I need to write slides for pghpw.org/ppw2013/schedule
masak though I tend to daydream about p6cc problems and solutions :)
diakopter and actually good, interesting ones at that
08:57 dakkar joined 09:00 donaldh joined
donaldh jnthn: hoelzro: re. debugging names in jvm, I was thinking along same lines. Would like to figure out a way of naming P6Opaques that at least relates them to comp unit. 09:03
hoelzro indeed
that would be a big +
for now, I'll just check for __P6Opaque and convert appropriately
jnthn Well, apart from I was pondering we could cache those by shape...
09:03 FROGGS_ joined
FROGGS_ bonsaikitten: about the fileops.t fails: <arnsholt> It seems to work for me, on a Red Hat 6 machine. Could you perhaps try blowing away the install/ directory and recompiling everything? 09:03
bonsaikitten FROGGS_: ergh, that sounds quite silly 09:04
FROGGS_ bonsaikitten: why?
bonsaikitten because that was on initial compile
jnthn I think it belongs on the st...
donaldh jnthn: oh, true.
bonsaikitten and manually interfering with the build/test process is nasty
FROGGS_ bonsaikitten: ahh, oaky
donaldh jnthn: perhaps a build option. The benefit when using std debug and profiling tools would be immense. 09:05
09:05 FROGGS left
donaldh On a related note, top 20 heap dump after parsing CORE.setting: gist.github.com/donaldh/6676995 09:06
09:06 pmurias joined
donaldh Suggests the average HashMap fill ratio is about 1.3 09:07
pmurias cxreg: ping
donaldh I'm guessing that it's quite wasteful for every QASTNode to have a hash. Maybe we can allocate hashes on demand here. 09:08
Maybe also true for the QASTNode arrays. 09:09
diakopter donaldh: agree++
09:09 FROGGS_ is now known as FROGGS 09:12 darutoko- joined
donaldh For fun, I implemented VMHash with a TreeMap instead of a HashMap. There was some memory saving, but not substantial. Turns out that TreeMap instances are quite beefy too. 09:13
09:13 darutoko left, woosley joined, darutoko joined
masak donaldh: maybe a LinkedList will save both memory and be faster? it's been known to happen. 09:16
dalek rlito: c55f350 | (Flavio S. Glock)++ | / (4 files):
Perlito5 - grammar - vstring placeholder
09:16 darutoko- left
tadzik faster? 09:16
donaldh masak: performance gain was negligible if any. I would need to run both several times over to get a more accurate measure. 09:17
And I'm finding I get quite a lot of variation on my poor macbook air. 09:18
09:18 darutoko left
donaldh didn't realize that --stagestats had a wait feature. 09:20
09:23 iSlug joined, darutoko joined
pmurias draft state of the JavaScript backend gsoc project: pastie.org/pastes/8351242/text 09:26
masak pmurias: wow. nice. 09:29
pmurias: s/so to/to/
diakopter pmurias: did you find a js bigint library? there are bunches 09:30
masak pmurias: how feasible is this: page with big textarea -- I type some nqp into it, hit "Run" -- it runs in the browser. my guess is: pretty darn feasible.
pmurias: (my point is, it would be a pretty compelling demo) 09:31
diakopter s/type some/paste many MB of/
masak an nqptidy sounds like a very fun weekend hackathon.
pmurias nqptidy is badly needed to enforce a coding standard without doing boring cleanup ;) 09:32
masak: re textarea demo it shouldn't be very hard
09:33 colomon joined
dalek rlito: f624047 | (Flavio S. Glock)++ | t5/01-perlito/420-vstring.t:
Perlito5 - tests - vstring
09:34
09:35 domidumont left
jnthn pmurias++ 09:36
pmurias masak: thanks for the suggestion, I'll work on it during today
diakopter pmurias: did you find a js bigint library? there are bunches 09:37
masak pmurias: I would happily be a devoted early adopter :)
pmurias diakopter: there seem to be tons of js bigint libraries, I would have to find one with a similiar API (or at least a similiar set of support operations) 09:38
diakopter: and I haven't really searched for that yet, it will not be required for the web demo 09:39
food&
diakopter npmjs.org/package/big-integer # seem ok?
FROGGS pmurias++ 09:40
diakopter pmurias: also this one looks quite hefty github.com/jtobey/javascript-bignum 09:41
(that page also has lots of links to others)
09:44 flussence left 09:45 flussence joined, jonathanmarvens joined 09:48 jonathanmarvens left, jonathanmarvens joined, ksh left 09:49 ksh joined
arnsholt bonsaikitten: You're the one having trouble with github.com/perl6/nqp/issues/133 ? 09:51
09:52 denis_boyun joined
bonsaikitten arnsholt: I am 09:52
arnsholt I couldn't reproduce it on my machine, sadly
Are you building Parrot with --gen-parrot?
bonsaikitten not explicitly 09:54
arnsholt Huh?
FROGGS can it be a file system issue? 09:55
bonsaikitten arnsholt: I don't see any explicit reference to --gen-parrot in the script that builds parrot
so if it's not enabled by default then no
FROGGS it would be "perl Configure.pl --gen-parrot" in nqp's folder 09:56
arnsholt So you just "perl Configure.pl && make test"?
09:56 daxim_ joined, daxim left
bonsaikitten arnsholt: with a plain make call in between 09:56
arnsholt Right 09:57
FROGGS bonsaikitten: and you are building nqp directly or via rakudo?
bonsaikitten FROGGS: I can try adding --gen-parrot to nqp build
FROGGS: directly
FROGGS k
bonsaikitten waaat
who made the makefile call "git pull" 09:58
arnsholt Looks like it won't build Parrot without gen-parrot, so I assume you have a system Parrot install?
bonsaikitten o_O This demands some angry words!
arnsholt: yes
FROGGS if you can configure nqp without --gen-parrot, it means it finds a parrot in path (or it uses the one you provided with --with-parrot in an earlier configure run)
arnsholt Not the Makefile, but gen-parrot clones the Parrot repo and builds it. Done that since forever
It's not a bug, it's a feature. Thought you knew about it =) 09:59
bonsaikitten ah, --gen-parrot is pure evil and must be ignored
Thou Shalt Not Ignore The Package Manager!
arnsholt Works for most of us =) 10:01
FROGGS git is my package manager :P 10:02
arnsholt Given that we're mostly on non-rolling release systems, it's a lot simpler
bonsaikitten yeah, I'm too lazy for that :D 10:03
FROGGS bonsaikitten: what filesystem is that? 10:04
bonsaikitten FROGGS: XFS 10:05
10:05 sqirrel joined
bonsaikitten *poof* for a few minutes 10:05
10:06 iSlug left 10:15 dayangkun joined
FROGGS has somebody else here an XFS volume? 10:16
I can't reproduce it either
:o(
nwc10 pmurias: really really cool 10:29
small grammar nit - "nqp concatenates it's files" - *that* it's should be its
(the later "it's" is correct, as the later one is short for "it is")
so, is JS going to self-host and run Rakudo before MoarVM? :-) 10:30
jnthn The race is on!
10:33 brrt left 10:34 zakharyas joined
FROGGS hehe 10:34
jnthn: you are therefor not allowed to give pmurias++ any hints! ò.ó
tadzik haha, what a face 10:35
like ( `ー´)
FROGGS that looks like a fat pidgeon
10:43 iSlug joined
dalek rlito/replito: cac6146 | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar - number special cases
10:45
rlito/replito: 2a9824c | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar - number special cases
rlito/replito: c55f350 | (Flavio S. Glock)++ | / (4 files):
Perlito5 - grammar - vstring placeholder
rlito/replito: f624047 | (Flavio S. Glock)++ | t5/01-perlito/420-vstring.t:
Perlito5 - tests - vstring
rlito/replito: ac9f6e1 | (Stanislaw Pusep)++ | / (5 files):
Merge branch 'master' into replito
10:53 pernatiy left 10:56 sqirrel left 11:01 iSlug left 11:02 iSlug joined, [Sno] joined
dalek rlito: 3857c53 | (Flavio S. Glock)++ | / (4 files):
Perlito5 - grammar - vstring parser
11:04
lizmat r: class A { method foo { say "in foo"} }; my $r = A.can("foo")[0]; $r.add_phaser("ENTER", { say "bye" }); .perl.say for $r.phasers("ENTER"); A.new.foo # why doesn't the ENTER phaser not fire 11:05
camelia rakudo 367963: OUTPUT«Block.new()␤in foo␤» 11:06
lizmat s/not//
jnthn: apparently the phaser *is* added to the list of phasers for that method, but it isn't called
ah, it needs to have a phaser of that type to begin with 11:07
r: class A { method foo { ENTER {}; say "in foo"} }; my $r = A.can("foo")[0]; $r.add_phaser("ENTER", { say "hello" }); .perl.say for $r.phasers("ENTER"); A.new.foo 11:08
camelia rakudo 367963: OUTPUT«Block.new()␤Block.new()␤hello␤in foo␤»
lizmat hmmm… nothing in the ast that points to any additional actions being done 11:09
bonsaikitten FROGGS: it might be related to any of gcc-4.8, glibc-2.17, perl-5.18.1 ... subtle changes can often cause "unrelated" failures 11:12
11:14 [Sno] left
FROGGS bonsaikitten: true 11:15
bonsaikitten is currently triaging build failures with bison-3 11:16
jnthn lizmat: It needs to be added during the compilation of the method.
bonsaikitten it's quite amusing how fragile things are :(
jnthn lizmat: That is, by code object "compose" times 11:17
lizmat: That's why a trait can do it just fine.
lizmat except nothing is there yet to do it at that point in the core setting
jnthn lizmat: Reason being that at that point, it looks at what traits we have, and augments the QAST tree for the method to do calls as needed.
lizmat yes, I get that now… 11:18
oddly enough, if there is an X phaser already, you *can* add more phasers at runtime
jnthn Yes :)
lizmat more X phasers, that is
jnthn 'cus the QAST just iterates the list of ENTER phasers
lizmat indeed 11:19
moritz cross phaser! 11:20
lizmat jnthn: was thinking that maybe it would be useful to have a Phaser is Block object, that would have a $!once attribute 11:26
so that after execution the first time, it would be thrown out of the list of phasers
jnthn But once is once per closure
lizmat or is there a better way to inject a once {} block into a sub with is DEPRECATED trait ? 11:27
jnthn Well, we could just fix add_phaser to not depend on stuff we didn't define yet... :)
lizmat looking at that, actually 11:28
so,seems once blocks already work early in the setting
so, all "is DEPRECATED" traits in the setting, would become once blocks
(and not have the trait formally added)
11:29 zakharyas left
lizmat the trait would add an ENTER phaser with a once block 11:29
preferably a ONCE block, as it were
except that once {} is no longer a phaser
FROGGS moritz: DONT CROSS THE BEAMS!!11 11:31
dalek rlito: 48ef9a7 | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar - vstring/bareword disambiguation
lizmat
.oO( a X phaser would execute ENTER and LEAVE at the same time, thereby executing any code in the block in constant time )
11:32
or no time ?
something like: 11:35
$ perl6 -e 'say "abc".ucfirst; say "def".ucfirst'
'ucfirst has been deprecated, please use 'tc' instead in block at -e:1
Abc
Def
dalek rlito: 880ba6e | (Flavio S. Glock)++ | TODO-perlito5:
Perlito5 - TODO update
11:37
kudo/nom: 36343bc | (Elizabeth Mattijsen)++ | src/core/Str.pm:
We shouldn't be using deprecated methods in the settings ourselves
11:39
11:44 ssutch left 11:53 pernatiy joined 11:56 ssutch joined, fhelmberger_ joined 11:58 fhelmberger_ left
dalek rlito: ad9e15f | (Flavio S. Glock)++ | / (3 files):
Perlito5 - js - object stringification
11:59
11:59 fhelmberger left 12:01 darutoko left 12:06 sqirrel joined 12:10 denis_boyun left 12:11 denis_boyun joined
dalek rlito: add49ad | (Flavio S. Glock)++ | / (3 files):
Perlito5 - js - object stringification
12:15
moritz is anybody working on star-on-jvm? 12:20
jnthn donahldh++ has done work on Panda towards that, and I believe arnsholt is working on Zavolaj, but not aware anybody is doing the stuff needed in the star repo... 12:21
I plan to give some tuits towards triaging spectests that fail on JVM in the coming days.
lizmat jnthn: the star release mentions "postcircumfix:<[ ]> and postcircumfix:<{ }> will become multi-subs rather than multi-methods. Both at_pos and at_key will remain methods." 12:23
so I guess we cannot use all capitals for at_key/at_pos
moritz I'll see how the modules-test for a potential star 2013.09 look
jnthn moritz++ 12:24
moritz lizmat: we can change the deprecation notice for the 2013.09 star release
lizmat also: for 2013.09, we might want to mention that that change is going to be happening for the 2013.10 release
moritz: indeed :-)
also, for the star release, maybe menttion that Sets/Bags and the like are now completely up to spec with exception of the empty set symbol 12:25
moritz lizmat: I think it's easiest if you mention that yourself :-)
12:28 Teratogen joined
Teratogen I am watching Windy by The Association music video on youtube and I swear Larry Wall is in the band! 12:28
moritz lizmat: I've given you a rakudo/star commit bit
anyone else need one? 12:29
12:30 ajr joined, ajr is now known as Guest9681 12:36 Guest9681 is now known as ajr_, jnap joined
dalek ar: 0e02471 | (Elizabeth Mattijsen)++ | docs/announce/2013.09.md:
Preliminary work on Rakudo Star 2013.09 announcement
12:37
moritz lizmat++
lizmat moritz: handing it back to you again :-)
dalek ar: 50bb81d | moritz++ | / (9 files):
bump version in preparation of 2013.09
12:38
moritz if anybody wants to help with the R* release: go through the modules, search for .bless(*, and remove the *, 12:40
masak idea for a module: ack through the source files of the entire ecosystem with a single command. 12:41
12:42 dmol joined
dalek c: b3c7bd2 | moritz++ | lib/ (3 files):
remove * from bless
12:43
masak I don't have time to do it myself, but I would happily award an Amazon book (60 USD/45 EUR) to whoever does.
12:43 colomon left
tadzik hmmmm 12:43
masak the module needs to meet my standards of code and tests for the prize to be handed out.
12:43 colomon joined
masak if you're unsure what those standards are, ask. :) 12:44
tadzik I'll let someone do it, I already got one book from you :)
jnthn Book? Meh. Now if it was beer... :P
masak jnthn gets a special expemtion clause of getting to claim 400 SEK worth of beer as a prize. 12:45
moritz use of uninitialized value of type Nil in numeric context in method serialize at /home/mlenz/p6/star/install/lib/parrot/5.5.0/languages/perl6/lib/XML/Writer.pm:9
masak (the expemption clause is jnthn-specific. everyone else gets a book.)
moritz URI has several of the .bless(*,) cases 12:46
12:46 donaldh left, revdiablo left, tokuhirom left, JimmyZ left, masak left, bloonix left, takesako left, jnthn left, moritz left, c1sung left, moritz joined, jnthn joined, JimmyZ joined, bloonix joined, revdiablo joined, revdiablo left, revdiablo joined, takesako joined, tokuhirom joined 12:47 masak joined
masak ouch. :/ 12:47
jnthn netsplit?
masak I've been having lags on feather in the past few days.
is feather having trouble in some way?
12:47 [Sno] joined, dmol left
jnthn masak: I just got reconnected to... 12:47
tadzik I didn't experience anything
jnthn And I'm not on feather
*too
moritz fwiw URI.pm seems cause the most of the .bless warnings in the star modules-test 12:48
*seems to
lizmat fwiw, I see many people from other places go and come back
moritz X::JSON::RPC has a few too
12:49 jonathanmarvens left 12:50 donaldh joined
dalek ar: a551b8f | moritz++ | modules/ (2 files):
update modules
12:51
12:52 c1sung joined 12:53 jonathanmarvens joined 12:57 darutoko joined, skids left
dalek osystem: 6626554 | moritz++ | META.list:
uri by colomon++ seems more maintained than the one by ihrd
12:58
colomon doesn't even really remember having a uri fork....
moritz colomon: now if you give me commit access to the uri fork you don't remember having, I can remove the .bless(*) warnings 12:59
:-)
tadzik MOAR FOARKS
hoelzro "Declaration from most-derived grammar beats less-derived" <- I take it that this deals with inheritance chains of grammars? 13:01
or am I missing something?
colomon moritz: you've got a commitbit
moritz hoelzro: it does
hoelzro ok, cool
jnthn Is that in the context of LTM tie-breaks?
jnthn knows we don't do all of them, but things we do do that one. 13:02
hoelzro jnthn: this is from S05, establishing the "pecking order"
#5 is "Declaration with least number of 'uses' wins"
and I'm not too clear on its meaning
jnthn Me either when you consider separate compilation :) 13:03
hoelzro ah, the following line...kinda clears it up.
a little.
13:06 donaldh left 13:08 benabik joined
dalek ar: d149f5d | moritz++ | modules/uri:
update module URI
13:09
colomon moritz++ 13:10
dalek Heuristic branch merge: pushed 23 commits to rakudo-star-daily by coke 13:11
[Coke] .to lizmat that commit is in the range I am bisecting, yes. 13:12
yoleaux 07:34Z <lizmat> [Coke]: the method2sub branch started using .can in SLICE_MORE in Any.pm, maybe .can is borked on jakudo ?
[Coke]: I'll pass your message to lizmat.
13:13 PacoAir joined
[Coke] .to diakopter Congrats! 13:14
yoleaux [Coke]: I'll pass your message to diakopter.
dalek kudo/nom: f4a21d6 | (Elizabeth Mattijsen)++ | src/core/ (4 files):
Proof of concept for is DEPRECATED handling, in settings and outside of it
13:18
lizmat afk for the rest of the day&
yoleaux 13:12Z <[Coke]> lizmat: that commit is in the range I am bisecting, yes.
[Coke] botches the bisect and starts over. arglebargle. 13:19
moritz [Coke]: I've done that a good many times too
lizmat [Coke]: please check of .can works on jakudo
method2sub branch uses that for <a b c> slices 13:20
if that doesn't work, all slices are borked
really afk now&
jnthn notes to himself to look at that code, since .can is almost certainly not the most efficient way to write it...
corecatcher Do we have a need for mirroring stuff? my employee update his mirror hardware and has lots of space over (already mirroring cpan) 13:22
[Coke] jnthn: are you back from vacation already? 13:23
jnthn [Coke]: Yeah, was just a long weekend break :)
moritz corecatcher: interesting question
pmichaud: do you think we need a mirror for rakudo.org/downloads/ ? corecatcher offers some
[Coke] jnthn: ah, welcome back.
13:23 sqirrel left
jnthn had his laptop fail while on vacation, which made it even more of one 13:24
tadzik win-win
13:24 lowpro30 joined
tadzik well, sort-of 13:24
jnthn An the even-more-win is that the replacement laptop is already ordered. \o/
13:24 jonathanmarvens left
jnthn With a nice i7 processor, SSD, and 16 gigs of RAM. That should make me more productive at confs/hackathons. :) 13:25
tadzik 530?
FROGGS jnthn: but you still have to care about performance then, because not everyone has such a nice hardware :o) 13:26
jnthn T430s
FROGGS has an i5, SSD and only 6gig ram
err, 8gig
jnthn Should be lighter too
tadzik heh, same on my 430
jnthn (lighter than my current one)
13:26 jonathanmarvens joined
JimmyZ has only 2gigs 13:27
tadzik it's like 2.4 kg or so
FROGGS poor JimmyZ
JimmyZ :(
tadzik I was thinking about this cdrom battery, but it's crazy expensive, at least here in PL
jnthn yeah, I didn't go for that...
FROGGS why do you guys have Lenovos? DELL is soo awesome
13:28 iSlug left
JimmyZ my one was old X200T 13:28
tadzik s stands for slimmer?
FROGGS: clitmouse. Case closed
jnthn Yeah :)
tadzik err, I mean nipp^Wtrackpoint
FROGGS "clit"-mouse?
tadzik I'm sorry
FROGGS what the
tadzik but look at it
jnthn wtf... :)
13:29 PacoAir left
jnthn hopes the keyboard is still good on thinkpads... 13:29
tadzik it is
FROGGS tadzik: you mean that red nipple? www.notebookcheck.com/uploads/tx_nb...T430_3.jpg
tadzik it looks different, but works exactly the same
jnthn ah, good
tadzik well, the key placement is different
you don't have this 'block' on top, but pgup and pgdown is next to the arrow keys 13:30
I find that a lot more convenient
13:30 PacoAir joined
tadzik FROGGS: yep 13:30
FROGGS well, I hate these trackpoints anyway... and I prefer a full keyboard
JimmyZ how much is T430s?
13:31 lowpro30 left
tadzik I wonder how much slimmer s is 13:31
FROGGS tadzik: this is how it has to look like: www.notebookcheck.com/uploads/tx_nb...6530_2.jpg
:o)
tadzik seems to be 1.02" vs 1.18"
13:32 lowpro30 joined
tadzik FROGGS: the only reason to have a numpad, for me, is playing THPS games :) 13:32
FROGGS hmmm, I really tend to type numbers there
and I dont play games
tadzik I actually got a ThinkPad USB keyboard for $work 13:33
it also has no numpad
JimmyZ T430s is about $2450 in china
FROGGS ohh wow
hoelzro " @substrings = slice @()" # umm...what's @()?
13:33 pmurias left
FROGGS r: say @().perl 13:33
camelia rakudo 36343b: OUTPUT«().list␤»
tadzik some say I need a small keyboard so I have enough space for all the rubbish I keep on my desk
hoelzro FROGGS: but what *is* it? 13:34
I see it on S05
*in
colomon r: say (@()).perl 13:35
camelia rakudo 36343b: OUTPUT«().list␤» 13:36
FROGGS I don't know what it is called but AFAIK it just calls .list on the things inside of it
hoelzro right, that much I knew =)
but since there's nothing in between @( and )...
jnthn I think @() itself does it to $/ though
hoelzro oooooh
ok, that makes sense
FROGGS r: $/ = 42; say @().perl 13:37
camelia rakudo 36343b: OUTPUT«().list␤»
FROGGS r: $/ = 42; say $().perl
camelia rakudo 36343b: OUTPUT«No such method 'ast' for invocant of type 'Int'␤ in block at /tmp/lQymOniDLD:1␤␤»
hoelzro r: ($/) = 42; say @().perl;
camelia rakudo 36343b: OUTPUT«().list␤»
hoelzro ...
FROGGS $() is a shortcut for $/.ast, which gives you the thing you explicitly made within the last regex match 13:38
but I don't think that @() or %() do the same, I believe $() is the only one that does that
jnthn Oh... 13:39
FROGGS (because I made it to do that)
jnthn No, S05 wants @() to be the positional matches
"Note that, as a scalar variable, $/ doesn't automatically flatten in list context. Use @() as a shorthand for @($/) to flatten the positional captures under list context." 13:40
FROGGS r: "ab" ~~ /(.)(.)/; say @().perl
camelia rakudo 36343b: OUTPUT«().list␤»
jnthn So if it don't do it today, we should make it do so
FROGGS which is LHF
jnthn aye
pick it!
13:40 FROGGS[mobile] joined
FROGGS I will if hoelzro doesn't want to 13:41
see ya in a minute
13:41 FROGGS left
hoelzro what don't I want to? 13:45
FROGGS can undoubtedly finish it faster than I =) 13:46
13:48 xenoterracide joined 13:58 FROGGS joined 14:00 kaleem left
FROGGS hoelzro: it is not just about speed, it is about learning too :o) 14:02
hoelzro point taken =)
I have a few things I'm hoping to fix 14:03
jnthn
.oO( me too )) )
FROGGS k, then I'll do this one
hoelzro but I'm working on improving p6 docs atm
FROGGS awesome!
jnthn \o/
hoelzro getting it to complete without crashing my machine has been fun =/
FROGGS jnthn: btw, my birthday is in a few weeks, and nqp@moarvm bootstrapping would be an appropriate gift *g* 14:04
(but no pressure :P)
hoelzro mine too, actually =)
you can double gift 14:05
we don't mind ;)
FROGGS *g*
awesome!!
jnthn Well, I've no teaching until October, so... :)
hoelzro \o/ 14:06
14:10 jonathanmarvens left 14:11 btyler joined, logie joined
FROGGS \o/ 14:13
hoelzro FROGGS: out of curiosity, when's your birthday? 14:14
14:14 GlitchMr joined 14:17 bluescreen10 joined 14:18 raiph joined
PerlJam hoelzro: when's yours? :) (I'm Oct 28) 14:19
hoelzro Oct 8
PerlJam (btw, Oct 28 is also Bill Gates' birthday)
colomon Oct 6 14:20
hoelzro damn, I just have Matt Damon =/
lots of Oct birthdays here it seems
FROGGS oct 16 14:21
timotimo i have albert einstein
JimmyZ Oct 15 :P
jnthn Wow... :)
JimmyZ but mine is lunar calendar
PerlJam We should start an October.pm so that a future Rakudo release can use it ;) 14:22
FROGGS *g*
colomon www.takipiblog.com/2013/09/24/how-t...-debugger/ # may be interesting for us? 14:28
14:29 ajr_ left
hoelzro I saw that! 14:29
14:29 skids joined
hoelzro I was thinking about that during my weekend hacking on TIJMP 14:29
14:29 ajr joined 14:30 ajr is now known as ajr_
jnthn TIJMP? 14:32
14:32 jonathanmarvens joined
jnthn ah, found it 14:32
hoelzro I used it on Perl 6 over the weekend
I added an "Export to CSV" button
and started on translating __P6Opaque to Perl 6 class names 14:33
I've been thumbing through last weekend's course notes a lot =)
14:41 Teratogen left 14:43 bluescreen10 left 14:45 bluescreen10 joined
arnsholt PerlJam: That's a pretty good birthday. Mine is the 29th =) 14:51
14:54 spider-mario joined
PerlJam arnsholt: you're just reinforcing the idea of an October.pm ;) 14:54
tadzik Fest.pm 14:56
sjn Beer.pm 15:00
No need to limit consumption to one month of the year 15:01
arnsholt I think jnthn wouldn't mind a Beer.pm
[Coke] as part of August.pm, I support this trend.
sjn certainly wouldn't
arnsholt: btw, you planning on dropping by Oslo.pm next week?
jnthn Beer.pm \o/
tadzik heh, I just thought about Python.pm, a support group 15:03
sjn hehehehe
Ruby.pm
tadzik we could gather, drink, cry, sob and pat each other on our backs
us, sad refugees
sjn INTERCAL.pm
[Coke] oooh. Can I be ColdFusion.pm ?
tadzik can I be Iron Man? :) 15:04
PerlJam notes the appropriate singularity in Coke's sentence ;) 15:05
[Coke] bite me, professor.
. o O (the truth hoits)
FROGGS Lol.pm
and ProjectEuler.pm perhaps? 15:06
tadzik NO LOL 15:07
NoLol.pm
15:08 lowpro30 left 15:09 benabik left
FROGGS r: "abcdef" ~~ /(.(.))(.)/; say @($/) # jnthn, is that what we want? 15:10
camelia rakudo f4a21d: OUTPUT«「ab」␤ 0 => 「b」␤ 「c」␤␤»
jnthn FROGGS: Yes. 15:11
r: "abcdef" ~~ m:P5/(.(.))(.)/; say @($/) # vive le difference
camelia rakudo f4a21d: OUTPUT«「ab」␤ 「b」␤ 「c」␤␤»
15:11 spider-mario left
FROGGS r: "abcdef" ~~ /(.(.))(.)/; say $/ 15:12
camelia rakudo f4a21d: OUTPUT«「abc」␤ 0 => 「ab」␤ 0 => 「b」␤ 1 => 「c」␤␤»
FROGGS hmmm, I think I understand 15:13
r: "abcdef" ~~ /(.(.))(.)/; say @($/).perl
camelia rakudo f4a21d: OUTPUT«(Match.new(orig => "abcdef", from => 0, to => 2, ast => Any, list => (Match.new(orig => "abcdef", from => 1, to => 2, ast => Any, list => ().list, hash => EnumMap.new()),).list, hash => EnumMap.new()), Match.new(orig => "abcdef", from => 2, to => 3, ast => Any,…»
jnthn Match objects nest in Perl 6 15:14
Which is why we get match trees
FROGGS yes, that is not the point
jnthn ok, what is the point? :)
FROGGS r: "abcdef" ~~ /(.(.))(.)/; say @($/) # I find the "0 => " irritating
camelia rakudo f4a21d: OUTPUT«「ab」␤ 0 => 「b」␤ 「c」␤␤»
15:15 spider-mario joined
jnthn That's how match objects gist 15:15
FROGGS r: "abcdef" ~~ /(.(.))(.)/; say @($/).elems
camelia rakudo f4a21d: OUTPUT«2␤»
FROGGS ahh
okay
jnthn The 0 => ... comes from the inner (.) of the first match
FROGGS right 15:16
15:17 jonathanmarvens left 15:18 iSlug joined, silug_ left
FROGGS n: "abcdef" ~~ /(.(.))(.)/; say @() 15:20
camelia niecza v24-95-ga6d4c5f: OUTPUT«「ab」␤ 0 => 「b」␤ 「c」␤␤»
FROGGS cool
niecza++
15:21 denis_boyun left
jnthn FROGGS: Do %() while you're at it, btw ;-) 15:22
FROGGS jnthn: sure
jnthn FROGGS++ 15:23
FROGGS nr: say "abracadabra" ~~ m:overlap/ a (.*) a / 15:24
camelia rakudo f4a21d: OUTPUT«False␤»
..niecza v24-95-ga6d4c5f: OUTPUT«===SORRY!===␤␤Regex modifier overlap not yet implemented at /tmp/uYgcVluC8P line 1:␤------> say "abracadabra" ~~ m:overlap⏏/ a (.*) a /␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.sett…»
jnthn r: "abracadabra" ~~ m:overlap/ a (.*) a /; say $/ 15:25
camelia rakudo f4a21d: OUTPUT«「abra」␤ 0 => 「br」␤␤»
jnthn r: .say for "abracadabra".match(/ a (.*) a /, :overlap)
camelia rakudo f4a21d: OUTPUT«「abracadabra」␤ 0 => 「bracadabr」␤␤「acadabra」␤ 0 => 「cadabr」␤␤「adabra」␤ 0 => 「dabr」␤␤「abra」␤ 0 => 「br」␤␤»
FROGGS why is it falseish? 15:26
jnthn I'm not sure 15:27
m:overlap/.../ matches immediately though
15:27 lowpro30 joined
jnthn "foo" ~~ m/.../ normally works out becuase Match.ACCEPTS(...) returns self. 15:28
And because "foo" is put into $_ before the m/.../ runs
r: "abracadabra".match(/ a (.*) a /, :overlap).WHAT.say
camelia rakudo f4a21d: OUTPUT«(List)␤»
jnthn But that gives a List.
TimToady spec says that should fake up a Match object to indicate success 15:29
jnthn OK 15:30
15:30 donaldh joined
jnthn Should we find the exhaustives lazily? 15:30
'cus at the moment Match is (for performance) designed to have a reified thingy there, I think...
15:31 silug_ joined
colomon lazy++ 15:31
jnthn Though maybe we just sneak a lazy list into the first slot of that and hope normal flattening semantics take care of it... :P 15:32
15:32 rindolf joined 15:34 FROGGS[mobile] left
donaldh On the subject of optimisation on JVM, lazy creation of VMHash storage maybe gives a ~5 % performance improvement. Not got enough measurements to be conclusive tho. 15:35
jnthn That sounds worthwhile. We may also want to look at doing something special for very small hashes (which most of them probably are) 15:37
Like, up to 4 elements is just a key/value list that we search through. 15:38
donaldh Yeah, I was thinking along those lines too.
jnthn We can make it a single allocation by just doing an 8 element Object[] even. 15:39
arnsholt Probably a good idea. I read an interesting blog post once about how the core list and map classes on OS X have varying asymptotic behaviour, depending on the number of elements
15:39 jonathanmarvens joined
arnsholt Since the classes use the appropriate data structure variants behind the scenes 15:39
jnthn Don't even need a count if we use null as a sentinel. 15:40
15:40 kbaker_ joined 15:42 btyler_ joined
TimToady though then you can't use null to indicate deletion 15:43
probably okay to just pull the rest down one 15:44
jnthn yeah, true
15:44 btyler left
jnthn delete is comparatively rare. 15:45
TimToady that's what I thought, but neglected to say :) 15:46
15:49 perlo66 joined 15:50 perlo66 left 15:51 rhr joined
diakopter can just use some other sentinel for deleted 15:51
15:58 lowpro30 left
FROGGS either my machine has serious problems or or nqp/rakudo is in trouble 16:05
Stage parse : 386.691
16:05 daxim_ left
FROGGS it usually takes 150s these days 16:05
jnthn o.O 16:06
FROGGS: Which backend, ooc?
FROGGS parrot
16:07 gtodd1 joined
FROGGS I can't imagine that this is due to my @() changes... 16:07
16:08 gtodd left
diakopter I can 16:10
16:16 risou_awy is now known as risou, lowpro30 joined 16:19 zby_home joined, zby_home left, djanatyn left 16:23 Tene left 16:25 djanatyn joined, iSlug left 16:26 Tene joined 16:31 ajr_ left, dylanwh left 16:32 dylanwh joined, geekosaur left, ajr joined, ajr is now known as Guest20152 16:33 Guest20152 is now known as ajr_, jonathanmarvens left, donaldh left
jnthn FROGGS: JVM build time doesn't look out of the ordinary, fwiw. A little longer than I remember it, but only a few seconds. 16:35
16:36 bluescreen10 left 16:37 geekosaur joined 16:39 bluescreen10 joined
FROGGS jnthn: now both repo's are clean from any local changes, and it still takes ages... 16:47
I'll reboot, and if this doesn't fix it I'll bisect it 16:48
16:48 FROGGS left 16:50 FROGGS joined 17:02 ajr_ left 17:08 dakkar left 17:10 risou is now known as risou_awy 17:14 pernatiy left 17:22 risou_awy is now known as risou, risou is now known as risou_awy 17:27 jonathanmarvens joined
jnthn did did a make spectest on JVM and it got all the way through 17:31
So I ain't sure what's happening with the daily run 17:32
17:35 risou_awy is now known as risou 17:38 spider-mario left 17:39 spider-mario joined
masak could be RAM-related? 17:39
jnthn looks sheepish 17:42
17:43 spider-mario left, spider-mario joined
masak .oO( who is incorrigible when it comes to puns, jnthn? ewe are. ) 17:46
17:47 ajr joined
nwc10 wonders wether there are many sheep puns left 17:47
17:47 kbaker_ left, ajr is now known as ajr_
masak I'm pretty sure. if people do too many of them in a row, they fall asleep. 17:47
nwc10 can we count on that? :-) 17:49
jnthn What compsci course do sheep like best?
Lambda calculus
masak .oO( do androids dream of lambda calculus? ) 17:50
17:50 kurahaupo joined, araujo left
FROGGS ha, @() and %() is just one extra line... 17:51
17:52 sqirrel joined 18:00 kurahaupo_ joined, kurahaupo left 18:03 kurahaupo_ left 18:13 rindolf left 18:16 dmol joined, kurahaupo joined
dalek kudo/nom: 7e2d8bc | (Tobias Leich)++ | src/Perl6/Actions.nqp:
support @() and %() shortcuts
18:16
ast: eee19ce | (Tobias Leich)++ | S03-operators/context.t:
unfudged for rakudo

Means we can run it now unmodified.
18:18
FROGGS will add more tests in a bit 18:19
18:21 raiph left 18:27 zwut00 joined, zwut00 left 18:28 SamuraiJack left, darutoko left, zwut00 joined 18:29 sqirrel left 18:33 denis_boyun joined 18:36 raiph joined 18:37 spider-mario left, risou is now known as risou_awy, risou_awy is now known as risou 18:38 spider-mario joined
[Coke] has a rakudo night tonight for a few hours. 18:41
[Coke] will get the (*&#@$ bisect to finish, anyway.
[Coke] will also get back to nqp op docs.
and maybe some more partcl-nqp 18:42
18:42 spider-mario left
jnthn [Coke]: Bisect will be useful, as I couldn't reproduce the issue on my box. 18:42
[Coke] -> dayjob
18:42 spider-mario joined
[Coke] jnthn: ahgood. (if I were nwc10 I would have scripted this bisect!) 18:43
"oh, it's too much work to script this oneshot"
(someone remind of this the next time I complain of a long running manual bisect, please. ;) 18:44
-> dayjob furrealzies.
18:47 diakopter left
raiph masak, diakopter, anyone else who's taken a close look at ometa (jnthn?): i'd love to see your comments on "ometa vs nqp" (ignoring any aspect that references Perl 6)... 18:47
yoleaux 9 Sep 2013 08:07Z <diakopter> raiph: www.reddit.com/r/perl/comments/1lum...ed/cc3swfq
18:47 dmol left
masak raiph: I don't feel I know enough about ometa. but from a high level, it feels like the goals of the projects are different. 18:49
ometa wants to "remake computing". nqp is a compiler toolkit.
raiph masak: where's that quote from? my understanding of ometa is like "a generalisation of parsing expression grammars which allow arbitrary input and can be extended in an object-oriented way" 18:52
(and that it's a metacompiler)
i'm currently seeing it as similar, in principle, to a piece of what nqp is 18:53
masak I don't know where I read it.
seems you know more about ometa than I do :)
jnthn I suspect "a piece of" is accurate
raiph i searched the #perl6 irclogs for ometa and you were exploring it's World :)
jnthn Perl 6 grammars are PEG-y
raiph s/it's/its 18:54
jnthn But I don't think there's any 6model equivalent
raiph jnthn: right
jnthn Or code-gen part.
18:54 kst joined
jnthn One of the nice things about the NQP toolchain is how much you can build before having to dig into VM specifics. 18:55
18:56 kurahaupo left
raiph jnthn: right. nqp is the whole enchilada for a comprehensive compiler toolkit (that "just happens" to include a metacompiler bit, though that does itself have big benefits long term that will become increasingly important as the ecosystem matures over the next decade) 18:56
(and that deals nicely with backend abstraction)
(and, and, and ...) 18:57
jnthn I've talked about plenty of meta things but I'm not sure I've ever used the word metacompiler... :)
raiph i'm trying to place NQP in the computing world
en.wikipedia.org/wiki/Metacompiler 18:58
jnthn raiph: It's the thing that we've built to help build Rakudo in a sane, manageable way.
raiph: And seeing it as a product in its own right means that (a) with care/guidance it can become one, but (more important right now) (b) it gives us clear architectural boundaries. 18:59
FROGGS .oO( Maybe we should create ads that say: Perl 6 - Programming for the Meta-Cloud(tm) )
18:59 kst left
FROGGS nobody knows what that means, so everybody will buy it 18:59
18:59 kst joined
timotimo but there is nothing to buy :o 19:03
arnsholt jnthn: Speaking of NQP as a product and such, how interesting are compilers for other languages ATM?
FROGGS timotimo: LIAR!
raiph ah, so that's why development of metacompilers paused for a few decades...
FROGGS timotimo: it is just like a cloud, you can't touch it, you can't see it, but you know that you need it, for years now without recognizing 19:04
raiph "TREE-META was the last of a line of metacompilers, starting with META II, right before subsequent versions were designated classified technology by the U.S. military and government agencies."
timotimo %)
raiph "NQP -- try it before the US gov bans it"
19:05 risou is now known as risou_awy
jnthn "a metacompiler is written in its own language and translates itself"...yeah, that's NQP :) 19:05
FROGGS somebody should add a link :o)
jnthn And of course, it notes cross-compiling (which is how we reach new platforms) and bootstrapping (what we do once we get there) :)
masak will probably like the sourdough comparison 19:06
masak FROGGS: well volunteered! :)
raiph :)
masak jnthn: I do like it.
FROGGS so nqp proves that the 60's and 70's are back? :P
19:06 sqirrel joined
FROGGS masak: :P 19:06
FROGGS goes back to do webstuff 19:07
masak FROGGS: no, we won't have another 60's and 70's in ~47 years.
FROGGS awww
raiph .oO ( we won't in another 23 years either... )
19:09 kaare_ joined, btyler_ left 19:11 btyler joined, jonathanmarvens left
masak pmurias: aww, you didn't fix the s/so to/to/ typo in your report :/ 19:14
FROGGS and the "(url and url)" 19:15
masak pmurias: what's the use of sending it out for proof-reading if you don't act on the fixes? :)
19:19 pmurias joined 19:20 kst left 19:21 kst joined
FROGGS beetlejuice beetlejuice beetlejuice 19:21
Dancer is soo cool! 19:23
19:24 iSlug joined
pmurias masak: sorry :( I copy pasted the pastebin instead of the fixed copy in my vim :( 19:29
masak pro tip: don't do that :P 19:30
19:31 araujo joined 19:32 EvanTeitelman joined
sjohnson does Perl 6 support ENUM's? I seem to remember it supporting that or something like that 19:38
PerlJam sjohnson: yes
sjohnson kewl.
lue hello world o/
PerlJam sjohnson: See S12:1785 19:39
synopsebot Link: perlcabal.org/syn/S12.html#line_1785
19:39 diakopter joined
jnthn r: enum SupportsEnums <No HellYeah>; say HellYeah 19:39
camelia rakudo 7e2d8b: OUTPUT«HellYeah␤»
jnthn r: enum SupportsEnums <No HellYeah>; say +HellYeah
camelia rakudo 7e2d8b: OUTPUT«1␤»
sjohnson i've been using constants with integers in perl5 to sort of get around it 19:40
19:40 bluescreen10 left, diakopter left 19:41 sqirrel left 19:42 jonathanmarvens joined
FROGGS sjohnson: which are just subroutines under the hood, right? 19:43
sjohnson integers
19:43 donaldh joined
sjohnson FROGGS: use constant MODE_SOMETHING => $i++; use constant MODE_OTHER => $i++; 19:43
etc
FROGGS yeah, and MODE_SOMETHING is a sub 19:44
p5eval: 1
sjohnson FROGGS: so i can do: $mode == MODE_BLAH instead of: $mode == 'blah' with strings;
p5eval FROGGS: 1
FROGGS p5eval: use constant MODE_SOMETHING => 42; print MODE_SOMETHING()
p5eval FROGGS: 421
FROGGS p5eval: use constant MODE_SOMETHING => 42; print ref &MODE_SOMETHING 19:45
p5eval FROGGS: 1
moritz p5eval: use constant MODE_SOMETHING => 42; ref \&MODE_SOMETHING
p5eval moritz: CODE
FROGGS moritz++ :o)
moritz still uses p5 a lot
sjohnson if constants are code blocks under the hood, then so be it. :3 19:46
FROGGS moritz: me too, but not the scary parts *g*
(joke)
sjohnson: that is how Perl 5 is, it is hard to invent new stuff
(and not needed for all cases of course) 19:47
moritz fwiw in p5 constants are single-statement subroutines with empty prototypes, and p5 constant-folds them :-) 19:48
sjohnson so long as it works, I can go home happy as a clam :) 19:49
moritz sure it works
but it has implications you might care to know about 19:50
FROGGS sjohnson: dont get me wrong, I love Perl 5 even when I sometimes rant about its warts
moritz for example you can write $hash{ MYCONSTANT() } = $value
to disambiguate between constant and literal, auto-quoted strings
sjohnson i have a few complaints of my own.
moritz and it means you can access them as methods
YourPackage->MYCONSTANT 19:51
sjohnson readpipe bug, as well as given/when being pretty easy to shoot oneself in their foot due to some scoping that's hard to remember.
sjohnson squishes the readpipe bug 19:52
19:52 SmokeMachine joined
sjohnson moritz: i didn't know that. 19:52
moritz.smart
19:53 jonathanmarvens left
sjohnson hmm, one more thing is the & in p5 i don't really like... 19:54
i just explicitly call my own subs with & now.
i move some function else where, and stuff starts breaking without it.
my $animal = &getAnimal(); 19:55
moritz what? stuff starts breaking when you remove the & ?
I have trouble believing that.
masak that sounds odd, yes. 19:56
unless you leave out the parens :)
sjohnson hmm, it's the parens then
that's the one!
thanks mask.
masak* 19:57
moritz use parens, or pre-declare
sjohnson ahh, parens...
moritz the & is quite ugly, and does funny things with @_ if you remove the parens
sjohnson masak++ # brains
masak sjohnson: the parens are the "call" part, you know.
sjohnson: without it, '&getAnimal' is just talking about the sub itself. 19:58
sjohnson: with it, you're going "good, now take that sub, and call it"
moritz masak: in Perl 6, yes
sjohnson indeed.
masak moritz: oh, we're not in Perl 6 here?
sjohnson haha
masak I'm momentarily confused. which language are we programming? :)
moritz masak: no. In Perl 6 you don't even need the parens for post-declared subs
we had a brief 5y digression 19:59
masak the perils of jumping in sideways into a discussion... :)
sorry, for confusingly assuming p6 ;)
moritz :-)
masak: and yet your remark about the () being important, not the &, was spot-on
sjohnson well, i've learned a lot. thanks guys. (moritz FROGGS masak)++ 20:00
FROGGS pleasure :o)
masak moritz: apparently I helped \o/ 20:01
20:01 Tene left
masak .oO( rent masak at an affordable hourly rate. he helps even when he doesn't know what language the problem is in! ) 20:02
20:02 djanatyn left
FROGGS masak: I'll all you when we have some PHP issues at $work :P 20:03
moritz wonders if there is a demand for Perl mentors
FROGGS call*
damn, I really need a new keyboard
sjohnson i have my fire share of global $PHP_ISSUE's 20:04
20:05 kaare_ left
masak FROGGS: I won't know it's PHP, but I'll wonder why I suddenly have a headache and a sudden depression :) 20:06
FROGGS .oO( There is something entirely wrong... Ohh... ?> ) 20:07
20:08 djanatyn joined
masak the '?>' smiley is there to conveys existential terror. 20:08
convey*
geekosaur by the pricking of my thumb...
20:09 colomon left, Tene joined, colomon joined
masak geekosaur: someone's wiki this way comes? :P 20:10
20:10 diakopter joined
masak dang, shoulda made it "MediaWiki this way comes"... 20:10
donaldh There's 8 file-ops tests failing in nqp on jvm. Does anyone know how far back I might need to bisect it ? 20:16
masak donaldh: not sure. but da7b7b1f67685f920e50c4e5f14fcf603fe60b18 looks like a candidate for the culprit. 20:19
(or rather the things it merges in)
20:20 spider-mario left
donaldh Hmm. That fixed a bunch of tests iirc 20:21
FROGGS donaldh: if the last test of t/nqp/19 fail, that it might be due to the added nqp::readlinefh tests 20:28
20:29 colomon left 20:36 iSlug left 20:39 jnap left 20:41 colomon joined 20:42 jnap joined
lue idly wonders when we can expect to see t4 results... 20:45
20:47 geekosaur left, dylanwh left
donaldh FROGGS: I think you're right, these tests may never have passed on jvm. 20:51
masak lue: I can guarantee you won't see any before 2013-10-02 :/ 20:57
($dayjob sucking up all available time again)
lue as long as it's within two months, you'll beat the previous time :)
masak nods, sobbing 20:58
hopefully I'll be done before tadzik announces p6cc 2013, at least :)
lue oh, he's running it this year? 20:59
21:00 colomon left, cognominal left
masak it's not a complete certainty. but I'm pretty sure I'm *not* running it. 21:00
tadzik I was supposed to run one during summer too :P
but yes, there will be p6cc 2013
masak I'm willing to play a fairly big part in the background if someone else leads it.
and tadzik++ stepped forward last time we talked about this :)
lue & 21:01
21:02 skids left, dayangkun left 21:05 EvanTeitelman left 21:09 denis_boyun left 21:14 colomon joined 21:17 dayangkun joined 21:20 REPLeffect left, EvanTeitelman joined
masak 'night, #perl6 21:20
FROGGS gnight
21:22 FROGGS left 21:23 jonathanmarvens joined, geekosaur joined, jonathanmarvens left 21:24 dylanwh joined 21:32 REPLeffect joined 21:35 jnap left 21:36 jnap joined 21:37 BenGoldberg joined 21:43 raiph left 21:52 lowpro30 left
dalek p/lazy-vmhash: f7c396a | (Donald Hunter)++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/VMHash (2 files):
Lazy allocation of HashMap for VMHash storage on JVM.
21:54
p: f7c396a | (Donald Hunter)++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/VMHash (2 files):
Lazy allocation of HashMap for VMHash storage on JVM.
21:56
p: ec8898f | donaldh++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/VMHash (2 files):
Merge pull request #134 from perl6/lazy-vmhash

Lazy allocation of HashMap for VMHash storage on JVM.
jnthn That was a nice short diff. :) 22:00
timotimo does that give a big reduction in memory usage? 22:04
jnthn mits
22:05 zacts joined
donaldh timotimo: it gives a reduction. I haven't reliably measured how much. Need a robust way of forcing GC and then running jmap -histo at specific times. 22:13
s/times/points/
timotimo the perl token in the power plug now also has a "via 9gag.com" branded version of itself on the internets
corecatcher \o/ 22:23
22:24 skids joined
geekosaur now someone has to photoshop that imnto 6gag 22:27
22:38 jonathanmarvens joined
jnthn has All The Tomorrow to do Perl 6 things o/ 22:43
22:43 Guest595 left
corecatcher needs more time 22:44
22:44 btyler left
jnthn whoa 22:44
jnthn just avoids having an IPA explode over this keyboard
corecatcher you what?! 22:45
jnthn I opened the can and it was more alive than expected...
timotimo jnthn: a can full of International Phonetic Alphabet keyboard stickers? 22:46
corecatcher you mean really "alive"?
timotimo corecatcher: the next karlsruhe.pm meeting is the evening before my exam :|
corecatcher timotimo: its okay, we just going to make 4hrs instead of 8
:>
jnthn
.oO( "What's a linguist's favorite kind of beer?" :D )
timotimo and no beer
corecatcher ppl who dont rely on alc are very welcome 22:47
timotimo i don't rely on alc to get my work done ;)
jnthn doesn't rely on it
corecatcher jnthn: A good St(r)out? :)
timotimo it rarely does anything good for me
corecatcher agreed, programming under beer makes me open up steam and play for hrs ;) 22:49
timotimo let's be friends on steam
see query on other 'net 22:50
jnthn: what's your reaction to my findings that turning lexicals into locals has apparently zero effect on performance? 22:51
jnthn timotimo: On which backend? Though, I'm a little surprised it had no effect... 22:52
timotimo jvm and parrot
(i made sure that the vars would actually get changed in the microbenchmarks i ran, there was no noticable change at all) 22:53
jnthn I'd have expected at least some win on both...
22:53 kay_ joined
jnthn You checked the QAST post-optimize to make sure they'd really been lowered too? 22:53
timotimo t.h8.lv/p6bench/lexlocal.html
yes, that's what i checked
7d... is the one with the change
(and don't ask me why half of the benchmarks have been skipped, i ought to look into that tomorrow or something) 22:54
timotimo disappears into the night 22:56
22:58 jonathanmarvens left 23:04 ajr_ left, jonathanmarvens joined 23:05 pmurias left
jnthn 'night, #perl6 23:10
corecatcher nn
23:17 jonathanmarvens left, donaldh left 23:19 jonathanmarvens joined
dalek rl6-roast-data: f3b271d | coke++ | / (3 files):
today (automated commit)
23:20
rl6-roast-data: 3beff1a | coke++ | perl6_pass_rates.csv:
fix yesterday's date (run took too long)
23:39 araujo left 23:42 zacts left, EvanTeitelman left 23:45 jnap left 23:46 jnap joined 23:47 camelia left, woolfy left, prammer left, bonsaikitten left, ponbiki left, integral left, gabriel_ left, krunen left, jonathanmarvens left, SmokeMachine left, ssutch left, telex left, huf left, Util left, BenGoldberg left, kst left, cibs left, jnap left, REPLeffect left, GlitchMr left, [Sno] left, masak left, xinming left, stevan_ left, d^_^b left, frettled left, lestrrat left, awwaiid left, sergot left, dayangkun left, djanatyn left, Woodi left, pmichaud left, Juerd left, p5eval left, sjohnson left, LordV left, PerlJam left, bakedb left, jercos left, Celelibi left, felher left, Yappo__________ left, Hor|zon left, jtpalmer left, mls left, silug_ left, PZt left, betterwo1ld left, jdv79_ left, itz_ left, sunnavy left, BooK left, groky left, bonsaikitten joined, gabriel joined, integral joined, prammer joined, integral left, integral joined, krunen joined, woolfy joined 23:51 araujo joined 23:53 ponbiki joined, ponbiki is now known as Guest37284, gabriel is now known as 20WAC7D55 23:54 groky joined, BooK joined, sunnavy joined, itz_ joined, jdv79_ joined, betterwo1ld joined, PZt joined, silug_ joined, camelia joined, jnap joined, jonathanmarvens joined, BenGoldberg joined, REPLeffect joined, dayangkun joined, djanatyn joined, SmokeMachine joined, kst joined, GlitchMr joined, [Sno] joined, masak joined, ssutch joined, xinming joined, stevan_ joined, telex joined, d^_^b joined, Woodi joined, pmichaud joined, Juerd joined, huf joined, frettled joined, Util joined, p5eval joined, lestrrat joined, sjohnson joined, LordV joined, cibs joined, PerlJam joined, sergot joined, awwaiid joined, mls joined, jtpalmer joined, Hor|zon joined, Yappo__________ joined, felher joined, Celelibi joined, jercos joined, bakedb joined, zacts joined, zacts is now known as Guest32443, telex left 23:55 zwut00 left, telex joined