»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:00 [TuxCM] left 00:17 rindolf left 00:26 laouji joined 00:34 mprelude left 00:39 laouji left 00:40 laouji joined 00:44 llfourn joined, laouji left 00:49 llfourn left, laouji joined
japhb o/ 00:52
01:12 BenGoldberg left 01:13 BenGoldberg joined 01:20 vendethiel joined 01:36 BenGoldberg left, Fleurety left, ab5tract_ joined 01:37 ab5tract left, zacts left, zacts joined 01:38 ggoebel2 joined 01:40 ggoebel left, cgfbee left, Fleurety joined 01:41 cgfbee joined
thou ugexe: FYI: this works like I want it to. gist.github.com/softmoth/1da0f8e722e98c2363bf 02:13
02:13 noganex_ joined
thou ugexe: regarding tap/act, I still contend that you've got a race there. See commented-out .act() calls in this code ^; the note() calls for stderr and stdout will overwrite each other. 02:14
I have to use another channel to properly merge the lines. I believe your process code has the same issue. 02:15
02:16 noganex left
thou erm, sorry, nevermind! I do think it's still a potential problem, but this one is due to my running multiple threads a the same time 02:16
I'd need to do some random sleeping in the act() closures to test my hunch 02:17
jobs
02:22 vendethiel left 02:43 thou left 02:45 llfourn joined, rangerprice left 02:50 llfourn left
ugexe if i want the lines properly merged, i access stdout and stderr directly. i would not use $merge 03:01
but thats assuming im using one persons definition of proper
03:02 kaare_ joined, llfourn joined
ugexe your problem is because you are calling say/note whatever inside your tap/act 03:05
if multiple threads are printing out $*OUT, you will get what appears to be incorrectly merged lines 03:07
if you are really trying to merge the output to the terminal, that is a different thing. a blog post that shows the general idea ugexe.com/create-a-perl6-terminal-progress-bar/ 03:11
03:13 peteretep joined
peteretep homebrew wants to know if I want to install rakudo-star with pcre 03:13
ugexe but you have to make choices. if you receive a line of text without a line break at the end, and then another process fires its own tap to print, what do you do?
peteretep Is there a quick overview of the implications of that choice?
ugexe zef::cli::statusbar puts a lock inside $*OUT's print method 03:15
03:23 llfourn left 03:27 aborazmeh left 03:30 rurban joined 03:33 lizmat left 03:43 laouji left 03:44 laouji joined 03:46 _itz joined, prevost joined 03:49 _itz_ left 03:50 _itz left 03:51 _itz joined 03:52 thou joined 03:53 llfourn joined 03:55 TEttinger joined 04:02 _itz_ joined 04:04 _itz left 04:16 rurban left 04:54 [Sno] left
TimToady checking out breakfast at the Abba, then off to the Alhambra 05:00
we have one free ticket if anyone happens to be here and awake enough to leave 8:00 or so 05:01
05:03 khw left
ShimmerFairy TimToady: how many more stops until you're home (or if not that, your next destination)? A quick look for Alhambra places at least two in Spain and California, so it's not like I could guess :) 05:19
05:36 emilbarton joined
emilbarton hi, why does $str.eval returns "method not found for invocant of class string"? 05:36
05:36 cdc_ is now known as cdc
ShimmerFairy it's spelled EVAL in Perl 6 now. so it'd be EVAL($str) or $str.EVAL :) 05:39
emilbarton ok thanks
05:40 emilbarton left 05:42 prevost left 05:52 cognominal joined
TimToady ShimmerFairy: "The" Alhambra is in Granada, Spain 05:55
ShimmerFairy ah, I somehow missed the "the" there :)
05:56 skids left
ShimmerFairy By the way, I think I'll be ready to push my val() and allomorphic types as a branch tonight :) . The only main issue is that my implementation of IntStr dies on bigints, and I'm not sure how to fix that. 06:05
(The only thing I can think of is a fallback Int attribute, and then reimplementing all of Int's methods in IntStr, but ick) 06:06
06:09 lizmat joined
ShimmerFairy m: say "".Numeric # this is surprising, and feels wrong to me 06:09
camelia rakudo-moar 5ba44f: OUTPUT«0␤»
GLRelia rakudo-moar 36ea47: OUTPUT«0␤»
nine ShimmerFairy: that's probably Perl 5 heritage 06:12
ShimmerFairy nine: replacing .Numeric with a val() call in my code caught that in a test, and I'm quite OK with saying "but I don't see a number in that string" :) 06:13
(replacing the implementation of Str.Numeric, that is)
dalek kudo/glr: cb9b821 | FROGGS++ | t/01-sanity/19-say.t:
remove test of say() that does not work with .gist, laben++
 06:15
kudo/glr: cf3b24c | timotimo++ | t/01-sanity/19-say.t:
Revert "remove test of say() that does not work with .gist, laben++"

Was supposed to be on the glr branch
This reverts commit cb9b821b08095c6d4c1157085e7fc2084f2dd72d.
kudo/glr: 5ba44fc | hoelzro++ | tools/build/NQP_REVISION:
Bump NQP_REVISION

For close_stdin fix in MoarVM
kudo/glr: 78b8f29 | (Stefan Seifert)++ | tools/build/NQP_REVISION:
Merge branch 'nom' into glr
ShimmerFairy nine: I could see it might've been some sort of guarantee for ?$string === ?$string.Numeric.Stringy , back when "0" was False :)
nine Darn...really have to leave for the office now 06:16
06:28 FROGGS joined 06:30 thou left 06:32 [Sno] joined 06:37 [TuxCM] joined 06:49 aborazmeh joined, aborazmeh left, aborazmeh joined 06:50 xinming joined 06:52 rurban joined 06:54 xinming_ left 06:58 ChoHag joined
masak morning, #perl6 07:05
ShimmerFairy o/ 07:07
07:07 itz joined 07:10 laben joined
laben hello #perl6 07:13
07:13 abraxxa joined
ShimmerFairy o/ laben 07:19
07:24 Guest4404 left 07:27 xfix joined 07:28 aborazmeh left 07:31 Ven joined 07:34 rarara left 07:35 RabidGravy joined 07:36 Ven left 07:38 darutoko joined 07:42 Ven joined 07:46 itz left, Ven left 07:47 Ven joined, rurban left
laben as i feared, Template::Mustache doesn't work because the grammar or the actions fail to create the correct result. now, where do i sign up for a crash course in grammars and actions? 07:56
07:56 rurban joined
nine laben: doc.perl6.org/language/grammars is quite good 07:57
laben nine: do you know if Grammars and/or Actions internally use some sort of push that would need to be adjusted for glr? 07:59
DrForr laben: I've got a pretty extensive grammar/actions combo in the ANTLR module in the ecosystem.
08:04 [Tux] left
FROGGS laben: do you get error messages you could gist? 08:05
yoleaux 30 Aug 2015 21:00Z <laben> FROGGS: the LAST phaser doesnt seem to fire when the last execution is finished with next, see "for 1 {LAST {say "last"};next}"
FROGGS ohh, the LAST phaser thing might be some very easily fixable...
laben just a quick question, the TOP action is called last, after every other action has been called, right? 08:06
08:06 rarara joined
nine laben: my Apache-To-Nginx seems to parse just fine and that's a largish grammar. From what I can see, there's just a couple of GLR fixes necessary in the converter code itself. 08:08
FROGGS: do you by chance still have my short micro USB cable? 08:09
JimmyZ isn't the TOP first? 08:11
laben somehow an action doesnt get called, which means the correct ast is never made. let's see why... 08:12
08:15 espadrine joined 08:16 [Tux] joined
FROGGS nine: damn, it was yours? 08:18
ó.ò
I convinced tadzik to take it -.-
nine This...may even be an advantage. You aren't coming to YAPC are you? 08:19
FROGGS JimmyZ: the TOP token is called first, so the TOP action method is called last
nine: I'm not, correct
08:19 colomon joined
nine tadzik may be, so I may even have a chance of getting it back ;) 08:19
FROGGS :S
nine: Sorry
nine No worry.
tadzik nine: I'm not, sorry :( 08:20
nine: but I'm still in Zurich :)
if you want to catch me
nine I'm already back in Linz
tadzik ah :/
nine I should have charged my phone in Switzerland at least once. Then I'd have noticed the missing cable. 08:21
08:21 colomon left
nine Curse you large battery! 08:21
08:21 rindolf joined, colomon joined
tadzik :) 08:22
DrForr DAMN YOU FOR BEING SO EFFICIENT 08:23
nine Maybe I should just give in and register a Google account to suck my battery empty in a day ;) 08:24
laben gosh, the bottom action doesn't get called even once on glr, that's why it does not create any ast. at the same time, why the heck is nom calling the same action 2 times, with the same argument, when there's evidently only one occurrence?
nine laben: sounds like we overcompensated? 08:25
08:26 lizmat_ joined 08:28 domidumont joined, lizmat left
laben seems like it's called twice because it's used in an assertion. but the assertion totally doesn't get called in glr. what's going on? 08:30
need to go now, have a good day #perl6 08:42
08:42 laben left
lizmat_ good *, #perl6 08:51
08:51 lizmat_ is now known as lizmat
lizmat messages? 08:51
yoleaux 30 Aug 2015 10:39Z <nine> lizmat: I figured out what's wrong with zip-latest.t! It's just that every way we tried to fix the issue hit an underlying bug. Array.clone re-uses the same $!reified buffer for the clone and List.from-iterator creates a new $!reified but uses the same scalar containers for the values. So they are still linked.
lizmat .tell nine Good to know what caused the issue... :-) 08:52
yoleaux lizmat: I'll pass your message to nine.
08:52 telex left
nine lizmat: It's also fixed by now :) 08:52
yoleaux 08:52Z <lizmat> nine: Good to know what caused the issue... :-)
lizmat cool!
nine There's now only 2 GLR related spec test failures: Sequences do not yet throw the correct exceptions on invalid input and one of the advent tests is not properly GLRed yet. 08:53
08:54 telex joined
lizmat I saw 5: 1 of them is a known issue on OS X with IO::Socket::INET 08:54
the other two I'm not sure about
nine t/spec/S02-literals/numeric.t fails also on nom 08:55
ShimmerFairy++ fixed numeric.t to actually test what it's claiming to test and that's why it now shows failures. They should just be todo'd
cdc .tell [Coke] about irclog.perlgeek.de/perl6/2015-08-29#i_11138242, I use for each arguments of a Bash command-line: sub escape-arg(Str $arg is copy) { if $arg ~~ s:g/([ ' ' | \' | \\ | <:Control> ])/{ $0.ord.fmt('\\u%04x') }/ { [~] <$'>, $arg, <'> } else { $arg } } 08:56
yoleaux cdc: I'll pass your message to [Coke].
nine many-processes-no-close-stdin.t was also a nom issue that has since been fixed by hoelzro++ (NQP bump was missing)
08:56 Ven left 08:58 araujo joined, araujo left, araujo joined
ShimmerFairy sorry for not TODOing those tests, by the way; I think my idea was "oh, I messed around in that part of the grammar recently, I could fix that soon!" and then didn't :P 09:00
09:03 espadrine left, espadrine_ joined
ShimmerFairy m: say :10<-42> # this is how I guessed you negate adverbial numbers, but apparently I'm wrong :) (though to be fair the spec says nothing about a negative number of this form) 09:13
camelia rakudo-moar 5ba44f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/jP8rFHCpOz␤Malformed radix number␤at /tmp/jP8rFHCpOz:1␤------> 3say :10<7⏏5-42> # this is how I guessed you negate␤ expecting any of:␤ number in radix notation␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/zkxkvTeU3r␤Malformed radix number␤at /tmp/zkxkvTeU3r:1␤------> 3say :10<7⏏5-42> # this is how I guessed you negate␤ expecting any of:␤ number in radix notation␤»
RabidGravy If I need to reference a type 'B' in a role 'F' that may be consumed in term by 'F' is doing something like "self ~~ ::('B')" in the role valid way of not having to load the class (and get blown out for circular dependency)? 09:16
ShimmerFairy do you want ::?CLASS, per chance? 09:17
RabidGravy no I don't
the role may be consumed by a variety of things, it needs to check if something is one of them 09:18
(actually self was a bad example there as the actual code doesn't do that) 09:19
ShimmerFairy you can always stub the class(es) in question, and hopefully be able to use them well enough in the role. 09:20
RabidGravy the actual code is something more like: 09:23
m: role F { method b(Mu:D $f) { say $f ~~ ::("B"); } }; class B does F { }; B.new.b(B.new)
camelia rakudo-moar 5ba44f: OUTPUT«True␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«True␤»
RabidGravy where obviously the circumstances of the creation of the second B are somewhat more complicated 09:24
09:26 Ven joined
RabidGravy I'm going to go with it and see what happens 09:27
09:38 mprelude joined 09:43 lizmat left
nine ShimmerFairy: could you please todo them now? I'm not sure about the protocol there. For example, do we require RT tickets for todos? 09:45
09:47 rurban left
ShimmerFairy nah, I'll get to it. Just a case of focusing on other stuff, but it'd only take a second. :) 09:48
nine That's even better :)
Btw. sorry if I've been a bit grumpy about this. Past few weeks have been exhausting. 09:49
ShimmerFairy That's alright :) 09:51
dalek ast: 6b38ee7 | ShimmerFairy++ | S02-literals/numeric.t:
TODO a couple tests.

  nine++ for catching that they needed it.		 09:56
ast/glr: 8fe513c | ShimmerFairy++ | S02-literals/allomorphic.t:
Updating allomorphic tests.

In particular coercing the allomorphic types for 'is' tests, since allomorphic types probably shouldn't implement === for an allomorphic type and one of its component types.		 10:04
ast/glr: d47db92 | ShimmerFairy++ | S02-literals/listquote.t:
Fix accidental pure string expectation

Since the test in question isn't meant to test allomorphic types, I changed the tested key/value pair to not make an allomorphic type.
ast/glr: 6b38ee7 | ShimmerFairy++ | S02-literals/numeric.t:
TODO a couple tests.

  nine++ for catching that they needed it.
ast/glr: ba3393d | (Stefan Seifert)++ | S02-literals/ (3 files):
Merge branch 'master' into glr
10:06 Ven left 10:07 Ven joined 10:21 rurban joined 10:23 dayangkun left 10:27 Loren_ joined 10:33 [TuxCM] left 10:44 espadrine_ left 10:45 espadrine_ joined 10:53 ggoebel2 left 10:57 TEttinger left
nine m: (1, 2, 6 ... *)[5] 11:05
camelia rakudo-moar 5ba44f: OUTPUT«Unable to deduce arithmetic or geometric sequence from 1,2,6 (or did you really mean '..'?)␤ in block <unit> at /tmp/imuAJihaAN:1␤␤Actually thrown at:␤ in block <unit> at /tmp/imuAJihaAN:1␤␤»
GLRelia ( no output )
ShimmerFairy m: say (1, 2, 6 ... *)[5] 11:10
camelia rakudo-moar 5ba44f: OUTPUT«Unable to deduce arithmetic or geometric sequence from 1,2,6 (or did you really mean '..'?)␤ in block <unit> at /tmp/ZD7R2u6Sq4:1␤␤Actually thrown at:␤ in block <unit> at /tmp/ZD7R2u6Sq4:1␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Nil␤»
11:13 ggoebel joined 11:14 telex left 11:15 telex joined
dalek line-Perl6/glr: e508f48 | (Stefan Seifert)++ | t/0 (2 files):
Adopt GLR semantics in test suite
 11:18
ChoHag Is a capture consisting of all the arguments passed to a function available without explicitely capturing it? 11:20
timotimo no, for performance reasons 11:21
i think
hm, though perhas with the nqp ops
m: sub test($a, $b, $c?) { say nqp::usecapture(); } 11:22
camelia ( no output )
GLRelia rakudo-moar 78b8f2: OUTPUT«===============================================================================␤The use of nqp::operations has been deprecated for non-CORE code. Please␤change your code to not use these non-portable functions. If you really want␤to keep using nqp:…»
timotimo m: use nqp; sub test($a, $b, $c?) { say nqp::usecapture(); }; test(1, 2);
camelia rakudo-moar 5ba44f: OUTPUT«Cannot find method 'gist': no method cache and no .^find_method␤ in sub test at /tmp/7OQ5pmkqAg:1␤ in block <unit> at /tmp/7OQ5pmkqAg:1␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Cannot find method 'gist': no method cache and no .^find_method␤ in sub test at /tmp/XkS6IdVbWZ:1␤ in block <unit> at /tmp/XkS6IdVbWZ:1␤␤»
timotimo that gives you a very low-level capture
you'd inspect it with nqp::captureposarg and whatever lets you grab the named arguments 11:23
ChoHag Is there any other way to call another function with the same arguments as passed in to this one?
11:24 virtualsue joined
timotimo m: sub foo($a, $b) { say "$a $b" }; sub bar($a, $b) { &foo.callsame }; bar(1, 2) 11:25
camelia rakudo-moar 5ba44f: OUTPUT«Method 'callsame' not found for invocant of class 'Sub'␤ in sub bar at /tmp/m7g9cyiqGC:1␤ in block <unit> at /tmp/m7g9cyiqGC:1␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Method 'callsame' not found for invocant of class 'Sub'␤ in sub bar at /tmp/SGvANG8O3S:1␤ in block <unit> at /tmp/SGvANG8O3S:1␤␤»
timotimo hm
11:33 Ven left 11:35 virtualsue left, virtualsue joined
RabidGravy great, a bug report for a twenty year old module against perl 5.14.4 on cygwin 11:40
bless their little pointy heads
11:43 Ven joined
hoelzro o/ #perl6 11:47
nine Good morning, hoelzro! 11:48
hoelzro afternoon, nine!
Ven \o, hoelzro , nine , and #perl6
hoelzro ahoy Ven! 11:49
BooK just looked at jnthn's talk in yapc asia. pretty cool 11:53
but the code examples do not work for me (just did a `rakudobrew build moar`) 11:54
is there a place where I can peek at the complete code?
11:54 zakharyas joined
nine Doesn't his talk contain the features that are only implemented on the glr branch? 11:54
_itz_ I think race is only in glr 11:59
timotimo aye 12:00
BooK ah ok 12:01
switching branches then 12:02
meanwhile, doing my first PR to perl6/doc 12:03
nine \o/
hoelzro BooK++
Ven BooK++ :)
dalek c: c29b488 | (Philippe Bruhat (BooK))++ | lib/Type/Supply.pod:
Fix syntax error
c: 64d55fb | FROGGS++ | lib/Type/Supply.pod:
Merge pull request #120 from book/patch-1

Fix syntax error
FROGGS BooK++
RabidGravy yay! BooK! 12:05
12:06 virtualsue left, virtualsue joined
BooK I fondly remember my first patch to libwww, it was marginally bigger than this one (swapping two characters) :-) 12:10
timotimo pretty cool, BooK :) 12:11
welcome to the doc! :)
BooK mmm, there are others of the same kind in that page 12:12
gotta catch them all!
Ven who documents the tests, who tests the documents
moritz BooK: I've invited you to the perl6 org on github 12:13
BooK: if you accept, you can push to the doc repo (and a few others) directly
nine And I was just gonna say that BooK needs a commit bit :)
moritz nine: you're too sloooow :-) 12:14
BooK thanks 12:15
I'll drop out of the github interface then :-)
RabidGravy BooK, as a complete aside I was looking for a picture of you and Eric in those little cars in Braga the other day. That was ten years ago :-O 12:18
12:21 rurban left
BooK oh I found that picture again recently 12:21
I think it was taken by DrHyde
cognominal BooK, time to invinte Ven for P3
Ven mmh? 12:22
12:22 rurban joined
BooK Ven: patch.pm/p3/ # liz & wendy included 12:22
cognominal maybe he can share a ride with liz and wendy back from Granada
BooK interesting: to filter a chunk of code through perl6. I need to do !perl6 - 12:23
otherwise I seem to see the output of the REPL
cognominal that is the picture with mongueurs on tricycle?
Ven I'm afraid I'm at school on the 12th
BooK it's a Saturday 12:24
Ven yes. I'm at school that day
BooK claim you're from a religion that doesn't allow you to work on inconvenient days 12:25
FROGGS Ven: but that's like... weekend :o)
cognominal and prescribes -O fun, that would be the Perl 6 religion.
Ven FROGGS: yes. my school often makes me work saturdays and sundays (so basically, 13 days in a row, from 9am to 8pm. I have to sacrifice a weekend every 3 weekends for school)
BooK I think the religion for Monday is Garfiedinity, but that might actually be just an instance of Lazyness (which prescribes working on Any day) 12:26
if you have to make ritual sacrifices, it is a religion 12:27
Ven I had an incredibly hard time to convince my school letting me go to SPW was a good idea, took me more than a week
dalek pan style="color: #395be5">perl6-examples: 45bcffa | (Steve Mynott)++ | categories/cookbook/02numbers/02-05-roman-numerals.pl:
import Roman numerals from Rosetta Code
 12:30
BooK my daughters look forward liz and wendy's arrival: klonk.bruhat.net/wendy+flore.jpg
12:31 laouji left
FROGGS Ven: 13 days in a row is horrible! 12:31
Ven BooK: your daughter seems to be pretty tall already
FROGGS: it gets long and tiring, yes :(
FROGGS .oO( 'oribleaux! )
:/
cognominal is it 42? 12:33
Ven no 12:34
not important anyways
BooK RabidGravy: not sure you can find the original on the interwebs, but my notes say it was taken by DrHyde
articles.mongueurs.net/comptes-rend...closer.png # this version is online, and you're on it too!
cognominal so you haet yve not found the answer to the big question :_
RabidGravy BooK :) I'd forgotten about that 12:35
poor Greggy 12:36
cognominal Guillame still got hair on this photo.
* Guillaume 12:37
Ven ha! that matz one is great
cognominal btw, it seems that Matz was in person in yapc asia. 12:38
Ven really? that's nice to hear
12:38 andreoss joined
BooK didn't 12:40
andreoss $ time perl6 prob060.pl
real 91m35.536s
$ time perl prob060.pl5
real 0m4.815s
timotimo oh my
12:40 virtualsue left
RabidGravy oops 12:40
timotimo andreoss: any chance you could try that on rakudo/glr? (and do the necessary modifications to the code, of course)
12:40 virtualsue joined
RabidGravy on the bright side there is room for improvement 12:40
timotimo yeah, but someone has to go and do that :P 12:41
cognominal ven, second guest speaker yapcasia.org/2015/ scroll down
andreoss m: say 91 * 60 / 4.8
camelia rakudo-moar 5ba44f: OUTPUT«1137.5␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1137.5␤»
BooK wonders if it would be possible to mark code examples and expected output in a special way, so that some test could check the examples work with the current perl6?
12:41 mprelude left
Ven cognominal: indeed! 12:42
BooK pugs had something like this, and executable spec
andreoss timotimo: ok
timotimo thank you very much!
andreoss can i install glr with rakudobrew? 12:43
_itz_ BooK: that would be very useful
BooK andreoss: I did a checkout glr in ~/.rakudobrew/moar-nom
_itz_: is the doc in perl5 pod or perl6 pod? 12:44
ShimmerFairy mostly POD 5, for now
timotimo with rakudobrew you can use "triple" 12:45
rakudobrew triple glr master master
or something like that
BooK ".'.IO.watch.act( { say "$^file changed" } ); 12:46
that looks weird
should I remove the ".'. at the beginning?
Ven BooK: the . at the beginning means "$_." 12:47
_itz_ m: say ".".IO.WHAT
camelia rakudo-moar 5ba44f: OUTPUT«(IO::Path)␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«(IO::Path)␤»
Ven m: $_ = 'abc'; say .uc 12:48
camelia rakudo-moar 5ba44f: OUTPUT«ABC␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«ABC␤»
BooK right, so one of the quotes is incorrect
m: say ".'.IO.WHAT
camelia rakudo-moar 5ba44f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/LTGGuk9CCD␤Unable to parse expression in double quotes; couldn't find final '"' ␤at /tmp/LTGGuk9CCD:1␤------> 3say ".'.IO.WHAT7⏏5<EOL>␤ expecting any of:␤ argument list␤ doubl…»
GLRelia rakudo-moar 78b8f2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/pe6pzycVId␤Unable to parse expression in double quotes; couldn't find final '"' ␤at /tmp/pe6pzycVId:1␤------> 3say ".'.IO.WHAT7⏏5<EOL>␤ expecting any of:␤ argument list␤ doubl…»
cognominal one typo here and there is a good pedagogical device to force the reader to think. 12:49
BooK I have no doubt the doc is perfect and the mistakes are there for the only purpose of my education :-) 12:50
dalek c: c45f662 | (Philippe Bruhat (BooK))++ | lib/Type/Supply.pod:
fix more syntax errors
BooK maybe it's time to read CONTRIBUTING.md 12:51
12:51 CQ joined
dalek pan style="color: #395be5">perl6-examples: e83ca50 | (Steve Mynott)++ | categories/cookbook/08-file-contents/08-0 (3 files):
three LHF file contents examples
 13:02
13:03 visitor1 joined 13:08 Ven left
andreoss m: say $*VM 13:11
camelia rakudo-moar 5ba44f: OUTPUT«moar (2015.7.108.g.7.e.9.f.29.e)␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«moar (2015.7.108.g.7.e.9.f.29.e)␤»
andreoss how do i distinguish glr from nom? 13:12
FROGGS andreoss: look at the nick of the bot
andreoss m: say $*PERL 13:13
camelia rakudo-moar 5ba44f: OUTPUT«Perl 6 (6.Advent)␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Perl 6 (6.Advent)␤»
FROGGS andreoss: glr itself is just a branch, so it won't show up in a version string 13:14
but...
m: say ::('Seq') ~~ Failure ?? "good ol' nom" !! 'GLR'
camelia rakudo-moar 5ba44f: OUTPUT«good ol' nom␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«GLR␤»
andreoss m: my @x := grep Int, 1..10; 13:17
camelia ( no output )
GLRelia rakudo-moar 78b8f2: OUTPUT«Type check failed in binding; expected 'Positional' but got 'Seq'␤ in block <unit> at /tmp/uN6Fhkj1do:1␤␤»
13:19 Ven joined
andreoss how do i port my code with := to GLR? I get a bunch of Seq errors 13:21
moritz andreoss: if you need to access @x by index, it's best to add .list to the expression on the right-hand side 13:22
andreoss: if you want to iterate over it, the nicest fix is
m: my \x := grep Int, 1..10; .say for x;
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
13:25 mprelude joined
andreoss >This Seq has already been iterated, and its values consumed 13:29
moritz if you need to iterate it twice or more often, use .list to get a proper List 13:30
13:30 muraiki joined
andreoss i'm actually iterating on copy 13:31
FROGGS andreoss: are you still using binding by any chance? 13:35
m: my @x = grep(Int, 1..10).list; .say for @x; 13:36
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
FROGGS m: my @x := grep(Int, 1..10).list; .say for @x;
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
FROGGS ahh, hmmm, so that does not trigger the said msg
m: my @x = grep(Int, 1..10); .say for @x;
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
FROGGS hmmm, no idea
moritz andreoss: if you want help, you need to show us some code that produces the error 13:37
FROGGS m: my @x = grep(Int, 1..10); .say for @x; .say for @x
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
timotimo m: my $x = grep(Int, 1..10); .say for $x.list; .say for $x.list;
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
FROGGS m: my @x := grep(Int, 1..10); .say for @x; .say for @x
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Type check failed in binding; expected 'Positional' but got 'Seq'␤ in block <unit> at /tmp/bG3kvFqvWg:1␤␤»
FROGGS there we go
moritz m: my $x = grep(Int, 1..10); .say for $x.List; .say for $x.List
camelia rakudo-moar 5ba44f: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤This Seq has already been iterated, and its values consumed␤ in block <unit> at /tmp/gFHUIlm6Tt:1␤␤»
FROGGS dont use assignment if you want to have a proper copy
timotimo oh, .List not .list
moritz .list caches, .List not 13:38
timotimo 'k
moritz so, depends on what you want
timotimo .so depends on what you want? but don't you always get a Bool from .so? :p
moritz but I haven't seen a good use case for either while still keeping the original Seq around
FROGGS I guess Seqs are like Junctions... you want to collapse them as early as possible 13:39
moritz or pass them around without storing them in a variable
timotimo i'd rather say "you don't want to let them loose on the program" than the other thing
13:40 zakharyas left
andreoss paste.debian.net/309804/ 13:40
nine You don't need .list if you assign to an @array anyway
andreoss feels like i have to unlearn some Perl 6 already 13:41
timotimo a tiny bit, yeah :| 13:42
moritz Variable '@x' is not declared │········
at glr-andreoss:60 │········
nine I think there's just fewer reasons to use := in the first place, which I see as a good thing.
timotimo agreed
andreoss paste.debian.net/309806/ 13:43
RabidGravy m: say "Foo".WHAT.^name; # is there a better way of doing this if I must get the name of the type as a string? 13:47
camelia rakudo-moar 5ba44f: OUTPUT«Str␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Str␤»
13:47 rurban left
tadzik I don't think so 13:47
nine m: say "Foo".^name;
camelia rakudo-moar 5ba44f: OUTPUT«Str␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Str␤»
tadzik but why do you?
oh :)
timotimo m: say Str.gist; say Str.Str; say Str.perl
camelia rakudo-moar 5ba44f: OUTPUT«(Str)␤Use of uninitialized value of type Str in string context in block <unit> at /tmp/zKBwRBUMzH:1␤␤Str␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«(Str)␤Use of uninitialized value of type Str in string context in block <unit> at /tmp/qlxNzApGbY:1␤␤Str␤»
timotimo m: say Str.perl
camelia rakudo-moar 5ba44f: OUTPUT«Str␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Str␤»
moritz likes $obj.^name best
timotimo m: my $foo = Str; say $foo.^name 13:48
GLRelia rakudo-moar 78b8f2: OUTPUT«Str␤»
camelia rakudo-moar 5ba44f: OUTPUT«Str␤»
timotimo ah
m: my $foo = Str; say $foo.VAR.^name
camelia rakudo-moar 5ba44f: OUTPUT«Scalar␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Scalar␤»
timotimo m: my $foo = Str; say $foo.VAR.name
camelia rakudo-moar 5ba44f: OUTPUT«$foo␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«$foo␤»
timotimo so that's how you get that
m: sub what's-my-name(\foo) { say foo.name }; my $foo = Str; what's-my-name($foo) 13:49
camelia rakudo-moar 5ba44f: OUTPUT«Method 'name' not found for invocant of class 'Str'␤ in sub what's-my-name at /tmp/K5zwpfvq_2:1␤ in block <unit> at /tmp/K5zwpfvq_2:1␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Method 'name' not found for invocant of class 'Str'␤ in sub what's-my-name at /tmp/XQHOFjqi3q:1␤ in block <unit> at /tmp/XQHOFjqi3q:1␤␤»
pink_mist foo.VAR.name surely?
nine andreoss: I'd just write for grep Prime, 1 .. 10000 -> $prime {
visitor1 Hi all!
timotimo yeah, pink_mist, i was just checking :) 13:50
pink_mist right =)
visitor1 Why this works:
r: class C { has Int $.v is rw = 0; multi method new(Int $v){ self.bless(v=>$v) }; multi method new(@vals){ self.bless(v=>[+](@vals)) } }; my @numlist = (10,20,30); my $c = C.new(@numlist); say $c.perl;
nine andreoss: oh, sorry, forget that
GLRelia rakudo-jvm 78b8f2: OUTPUT«Can't open perl script "/home/steve/rakudo-inst/bin/eval-client.pl": No such file or directory␤»
..rakudo-moar 78b8f2: OUTPUT«C.new(v => 60)␤»
camelia rakudo-{moar,jvm} 5ba44f: OUTPUT«C.new(v => 60)␤»
timotimo hi visitor1
visitor1 and this does not?
r: class C { has Int $.v is rw = 0; multi method new(Int $v){ self.bless(v=>$v) }; multi method new(Int @vals){ self.bless(v=>[+](@vals)) } }; my @numlist = (10,20,30); my $c = C.new(@numlist); say $c.perl;
13:50 khw joined
GLRelia rakudo-jvm 78b8f2: OUTPUT«Can't open perl script "/home/steve/rakudo-inst/bin/eval-client.pl": No such file or directory␤» 13:50
..rakudo-moar 78b8f2: OUTPUT«Default constructor for 'C' only takes named arguments␤ in block <unit> at /tmp/tmpfile:1␤␤»
camelia rakudo-{moar,jvm} 5ba44f: OUTPUT«Default constructor for 'C' only takes named arguments␤ in block <unit> at /tmp/tmpfile:1␤␤»
andreoss nine: i need @primes as stack, so it wont' check same prime number twice
nine: that's why .shift
visitor1 The difference is using (@vals) or (Int @vals) in an argument list for a new method 13:51
13:52 virtualsue_ joined
timotimo visitor1: if you have "Int @vals" as a parameter, you have to explicitly pass in an array you've manually declared to be "of Int" 13:52
13:52 virtualsue_ left
nine andreoss: so what's the remaining issues with your script? 13:52
visitor1 r: class C { has Int $.v is rw = 0; multi method new(Int $v){ self.bless(v=>$v) }; multi method new(Int @vals){ self.bless(v=>[+](@vals)) } }; my Int @numlist = (10,20,30); my $c = C.new(@numlist); say $c.perl; 13:53
GLRelia rakudo-moar 78b8f2: OUTPUT«C.new(v => 60)␤»
..rakudo-jvm 78b8f2: OUTPUT«Can't open perl script "/home/steve/rakudo-inst/bin/eval-client.pl": No such file or directory␤»
camelia rakudo-{moar,jvm} 5ba44f: OUTPUT«C.new(v => 60)␤»
moritz visitor1: see also my rant about this on p6u: www.nntp.perl.org/group/perl.perl6....g2155.html
visitor1 Ahhh ok... thanks timotimo!
timotimo seems like we want to prevent glrelia to try to run rakudo-jvm
well, that or build rakudo-jvm on it, too
13:53 virtualsue left 13:54 virtualsue joined
moritz timotimo: I hope that GLRelia's life time is counted in days, anyway 13:54
timotimo sure
hm
RabidGravy merge imminent?
timotimo though i'd say it'd be interesting to keep nomelia around after glr becomes the main development branch
so that we can easily see differences on the channel if we want to 13:55
i wouldn't run it for every "m:", though
maybe "nom:"
moritz timotimo: shouldn't be so hard
timotimo aye
moritz just like we kept pre-nom rakudo ("Beijing") around for a while 13:56
13:56 cognominal left
timotimo ah 13:56
i don't remember that :)
moritz timotimo: and if you do a MoarVM release today, I can do a nom-based Rakudo release tonight
timotimo oh!
moritz and then we can use that release as the reference thingy for camelia 13:57
timotimo i should get going, then!
moritz no pressure, but yes, do it now! :-)
andreoss nine: it fails to work with glr 13:58
timotimo haha, that's "the" reason? :) 13:59
the remaining issue, i mean?
andreoss nine: i assume because of multiple copies of Seq
13:59 virtualsue left
timotimo and i assume getting a --profile of that 90 minute run thingie would result in a ridiculous amount of data 14:00
just because we record a bit of data for every single gc run 14:01
and in 90 minutes, the gc runs *a* *lot*
(unless you're only numbercrunching native ints/nums in a tight loop)
nine andreoss: fails in what way? 14:02
timotimo maybe we want a little bit of intelligence that'll start only recording major collections after a certain number of minor collections
andreoss nine: >This Seq has already been iterated, and its values consumed 14:04
timotimo i'm really looking forward to getting these two new monitors <3
nine andreoss: cannot reproduce that here 14:05
moritz has two glr test failures: t/spec/S03-sequence/misc.t and t/spec/integration/99problems-21-to-30.t 14:06
nine moritz: yes, that's the two remaining ones
14:06 thou joined
nine moritz: 99problems-21-to-30.t is a plain missing GLRification one and the sequence test fails because we do not detect invalid input early enough. 14:07
masak wow. two.
14:07 skids joined
moritz m: say combination(2, <a b c>) 14:08
camelia rakudo-moar 5ba44f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/HcDhsG_w2b␤Undeclared routine:␤ combination used at line 1. Did you mean 'combinations'?␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/tacMUsSW36␤Undeclared routine:␤ combination used at line 1. Did you mean 'combinations'?␤␤»
timotimo +s 14:09
andreoss nine: paste.debian.net/309810/ this fails for me
moritz timotimo: that test file has a custom sub combination
andreoss may be i have a broken installation or somethimg, sorry if i wasted your time
14:10 virtualsue joined
timotimo ah, i see 14:10
nine andreoss: add a .Array in line 45 14:11
andreoss: gather returns a Seq, so you created an Array containing Seqs and expected it to be an Array of Arrays 14:12
andreoss nine: thanks 14:13
14:14 cognominal joined 14:16 domidumont left
nine andreoss: FWIW I see quite some opportunity for optimization. If you change your implementation to not use .shift anymore and instead operate only on Lists, it may be quite a bit faster. 14:16
14:18 domidumont joined
andreoss but List cannot be used as stack 14:21
nine Exactly! Your implementation makes countless copies of the same list. If you use the same static list and use subscripting instead, it may be faster. 14:23
dalek kudo-star-daily: 023f6a2 | coke++ | log/ (9 files):
today (automated commit)
 14:24
nine m: my @list = grep Prime, 1 .. 100; sub foo(@primes) { say @primes.Str }; foo(@list[1 .. *];
camelia rakudo-moar 5ba44f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/R6cLoEJcGC␤Unable to parse expression in argument list; couldn't find final ')' ␤at /tmp/R6cLoEJcGC:1␤------> 3 { say @primes.Str }; foo(@list[1 .. *];7⏏5<EOL>␤ expecting any of:␤ argu…»
GLRelia rakudo-moar 78b8f2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/IKdXWb_OM1␤Unable to parse expression in argument list; couldn't find final ')' ␤at /tmp/IKdXWb_OM1:1␤------> 3 { say @primes.Str }; foo(@list[1 .. *];7⏏5<EOL>␤ expecting any of:␤ argu…»
nine m: my @list = grep Prime, 1 .. 100; sub foo(@primes) { say @primes.Str }; foo(@list[1 .. *]);
camelia rakudo-moar 5ba44f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/iUMEFrBsyZ␤Undeclared name:␤ Prime used at line 1␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/UAHvGFDUaF␤Undeclared name:␤ Prime used at line 1␤␤»
nine m: my @list = grep { $_.is-prime }, 1 .. 100; sub foo(@primes) { say @primes.Str }; foo(@list[1 .. *]); 14:25
camelia rakudo-moar 5ba44f: OUTPUT«3 5 7 11 13 17 19 23 29 31 37 41 43 47 53 59 61 67 71 73 79 83 89 97␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«3 5 7 11 13 17 19 23 29 31 37 41 43 47 53 59 61 67 71 73 79 83 89 97␤»
14:27 [TuxCM] joined
nine Meh...we'll still create a copy of the array in that case. 14:27
timotimo grep-index? :)
hm, won't really help, right? 14:28
nine I'd say a plain old get-remarkable(:@primes, :$start-index)
JimmyZ about parrot backend, I think the better ways is to disable it with one line code in configure.pl insteal of remove the code. 14:30
moritz if we disable it, we can't test it anymore 14:31
zostay m: my $x = 1, 2, 3; $x.WHAT.say 14:34
camelia rakudo-moar 5ba44f: OUTPUT«WARNINGS:␤Useless use of "," in expression "my $x = 1, 2, 3" in sink context (line 1)␤(Int)␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«(Int)␤»
zostay m: my $x := 1, 2, 3; $x.WHAT.say 14:35
camelia rakudo-moar 5ba44f: OUTPUT«(Parcel)␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«(List)␤»
JimmyZ moritz: the advice is to someone who want remove the code from git repo ;)
ShimmerFairy Is there a reason there's a 'say' multi that's just for Str just so you don't call .gist on it, even though Str.gist is just the Str itself? :) 14:36
moritz JimmyZ: but it's not a solution for anybody; either it's removed, or it stays. There is no middle ground.
ShimmerFairy: yes. Performance.
ShimmerFairy moritz: I figured, though gist(Str:D:) { self } seems so innocuous... :)
JimmyZ incase someone do Str.gist 14:37
ShimmerFairy (I ask because allomorphic types pick that multi accidentally, and "remove the Str multi" seemed like a potential solution)
timotimo ShimmerFairy: one of the most expensive things moar can do is invoke a sub, or so it seems :|
14:38 Khisanth left
ShimmerFairy interesting, is there any clue as to why? 14:38
.oO(The obvious solution is to make sure everything is a method, and thus everything is classes. Where have I heard that before...?)
 14:39
JimmyZ java?
timotimo running a script through perf shows that we spend a big amount of time in increasing the ref count of a frame
oh, i meant "invoking a code object"
not just subs
14:40 Mhoram joined
JimmyZ sometimes moarvm is fast enough, but the build plan in rakudo like gimme, very slow 14:40
timotimo it's so good we've gotten rid of gimme + reify :) 14:41
ShimmerFairy timotimo: to be fair, I'm sure handling what could in some contrived way be called "just advanced gotos" isn't easy :)
JimmyZ the build plan is as slow as gimme ...
timotimo :)
14:42 abaugher left 14:51 virtualsue_ joined
nine But surely we can inline a gist(Str:D:) { self }? 14:51
14:52 virtualsue left, virtualsue_ is now known as virtualsue 14:53 domidumont left 14:56 domidumont joined 14:58 cognominal left 15:03 MilkmanDan left
timotimo hopefully 15:05
15:05 perl6_newbee joined
masak perl6_newbee: hola. 15:06
15:09 Khisanth joined 15:12 laubster joined, telex left
[Coke] bots, wake up, get messages. 15:12
yoleaux 08:56Z <cdc> [Coke]: about irclog.perlgeek.de/perl6/2015-08-29#i_11138242, I use for each arguments of a Bash command-line: sub escape-arg(Str $arg is copy) { if $arg ~~ s:g/([ ' ' | \' | \\ | <:Control> ])/{ $0.ord.fmt('\\u%04x') }/ { [~] <$'>, $arg, <'> } else { $arg } }
15:14 telex joined 15:16 xinming_ joined 15:20 xinming left, MilkmanDan joined 15:25 Ven left
muraiki hi all. I have a script that creates a number of IO::Notification.watch-path Supplies. what I've noticed is that CPU usage tends to stay at a constant level regardless of how many directories are watched, but that the CPU usage also increases in proportion to the # cpu cores and their speed 15:26
for instance, on a dual core, core 2 duo machine I have 5% cpu usage. but on a 16 core machine (not sure of the exact cpu type, but obviously quite new) I see 10% cpu usage 15:27
hrm, well actually these cores on the 16 core machine aren't as fast as I thought. so maybe that's it
15:29 araujo left
muraiki is there a way I can get statistics on how many threads are being used across all these supplies? (note: I obviously don't really understand how this all works :) 15:30
15:30 cognominal joined
JimmyZ m: say $*SCHEDULER.loads 15:34
camelia rakudo-moar 5ba44f: OUTPUT«0␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«0␤»
15:34 virtualsue left 15:37 andreoss left
JimmyZ m: await do for ^5 { start { sleep 1 }; say $*SCHEDULER.loads } 15:37
camelia rakudo-moar 5ba44f: OUTPUT«0␤1␤2␤3␤4␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«0␤1␤2␤3␤4␤»
JimmyZ m: await do for ^5 { start { sleep 1 }; say $*SCHEDULER.loads }; sleep 2; say $*SCHEDULER.loads
camelia rakudo-moar 5ba44f: OUTPUT«0␤1␤2␤3␤4␤0␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«0␤1␤2␤3␤4␤0␤»
JimmyZ muraiki: ^^ maybe this is you want? 15:38
15:38 FROGGS left
muraiki JimmyZ: thanks, but I think the problem actually just comes down to "things aren't optimized yet." a perl6 -e 'sleep' showed similar cpu usage on these computers, heh 15:39
pflanze masak, was it junctions that you said in your talk you'd only use lexically? If I understand right, e.g. any(@foo) eq $x would be equivalent to @foo.any(* eq $x) assuming that there's an any method on arrays (I don't know if there is)?
muraiki yeah actually sleeping results in the same amount of cpu usage. :(
pflanze I just ran into those and was questioning sanity, then I remembered your talk.
masak pflanze: yes :) 15:41
pflanze: I do things like `any(@foo) eq $x` too sometimes
pflanze: the important thing to be aware of is that the result is not a *Bool*, but a *junction of Bools*
pflanze: you can amend this by doing `so any(@foo) eq $x`
JimmyZ muraiki: I am not sure, since sleep calls the system sleep 15:42
pflanze oi
masak pflanze: for more information, see perl6advent.wordpress.com/2014/12/...junctions/
muraiki JimmyZ: yeah, but what i mean is I'd have trouble believing system sleep results in the 5%-30% cpu usage I'm seeing
pflanze *Is* there a higher-order any method on sequences, too?
muraiki so I assumed it was overhead from moarvm or something
pflanze Which would simply return a boolean. 15:43
JimmyZ muraiki: maybe it is the parsing time? 15:44
pflanze goes trying 15:45
muraiki JimmyZ: the cpu usage initially starts at 100%+ for a few seconds, then dies down to 5-30%. so I figured that parsing is what's going on for those first few seconds
15:45 virtualsue joined
JimmyZ have no iden then :) 15:46
muraiki thanks anyways :)
it just stinks because I'm probably going to have to rewrite this in go, as 30% is a dealbreaker T_T
JimmyZ optimizations is hard
muraiki I want my supplies :(
masak jump-scares muraiki 15:47
muraiki: SUPPLIES!
muraiki ;)
15:48 ggoebel left 15:49 rurban joined 15:52 Ven joined
pflanze Ok, so there is no higher-order "any" method, or more to the point, the "any" method is the junction-building one. (I still don't get how/when "multi" methods and "method" methods are the same and when not.) 15:52
15:54 cognominal left 15:56 Ven left 15:59 abraxxa left 16:01 MadcapJake joined
perl6_newbee hi masak! 16:03
\o/
:-)
16:11 zakharyas joined 16:12 larion left, leont joined
timotimo muraiki: i can't reproduce such cpu usage for perl6 -e 'sleep 30' 16:15
it reports the cpu time the job got is about 0%
16:15 xinming joined
timotimo 0.07user 0.01system 0:30.08elapsed 0%CPU (0avgtext+0avgdata 61636maxresident)k 16:16
or what exactly are you doing?
wait, i forgot to sleep there, didn't i? 16:17
oh no, wrong workspace
dalek rl6-roast-data: 466bfa5 | coke++ | / (10 files):
today (automated commit)
 16:18
muraiki timotimo: I was just doing a perpetual sleep via `perl6 -e 'sleep'`
16:18 xinming_ left
muraiki I've reproduced decent cpu usage on 3 systems, one of which just built perl6 today 16:18
the other 2 have perl6 from sometime last week
er, rakudo 16:19
16:19 larion joined 16:24 larion left
JimmyZ muraiki: I found the bug, sleeps costs 5% bug 16:26
muraiki JimmyZ: oh nice! 16:27
JimmyZ well, the -e 'sleep'
16:27 vendethiel joined
JimmyZ sleep(10) works well 16:27
muraiki yeah I just discovered that perl6 -e 'while (1) { sleep(5) }' doesn't really use any cpu while perl6 -e 'sleep' does 16:28
JimmyZ muraiki: try s/5/1e16/
that's the bug
nine When should one use fail and when die? 16:29
JimmyZ m: say my uint64 $ = 1e16 * 1e9 16:30
camelia rakudo-moar 5ba44f: OUTPUT«This type cannot unbox to a native integer␤ in block <unit> at /tmp/FkGXkDkr3k:1␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«This type cannot unbox to a native integer␤ in block <unit> at /tmp/fyHQBk_bMV:1␤␤»
muraiki JimmyZ: thanks so much for figuring it out! in my script I'll use the while/sleep(5) form for now
JimmyZ m: say my uint64 $ = (1e16 * 1e9).Int 16:31
camelia rakudo-moar 5ba44f: OUTPUT«Cannot unbox 84 bit wide bigint into native integer␤ in block <unit> at /tmp/4pYkdgWzcN:1␤␤»
GLRelia rakudo-moar 78b8f2: OUTPUT«Cannot unbox 84 bit wide bigint into native integer␤ in block <unit> at /tmp/iReVR8XJrY:1␤␤»
JimmyZ yeah, the bug
leont Hmmm, using an enum as MAIN argument doesn't DWIM, bummer
JimmyZ muraiki: I will make a patch to rakudo 16:32
muraiki JimmyZ: you are awesome 16:33
vendethiel leont: how do you mean? enum A <a b c>; sub MAIN(A) {}? 16:34
leont I was hoping for the getops implementation to do something sensible if the argument is either a number(value) or a string (name)
16:36 domidumont left 16:37 gfldex left, gfldex joined 16:41 raiph joined
JimmyZ well , maybe to moarvm 16:44
16:46 FROGGS[mobile] joined
FROGGS[mobile] o/ 16:48
yoleaux 30 Aug 2015 16:55Z <nine> FROGGS[mobile]: do you by chance still have my micro USB cable? I seem to remember giving it to you on Wednesday
30 Aug 2015 17:37Z <laben> FROGGS[mobile]: the LAST phaser doesnt seem to fire when the last execution is finished with next, see "for ^3 {.say; next if $_ < 3; LAST {say "last"}}"
dalek kudo/glr: 8a9d288 | (Stefan Seifert)++ | t/01-sanity/99-test-basic.t:
s/done/done-testing in 01-sanity/99-test-basic.t

Obviously no-one is running the sanity tests regularily...
16:49 araujo joined, spider-mario joined 16:52 pullphinger joined
[Coke] still 10 failures left on glr. running again to see what I've missed. 16:55
dalek kudo/glr: d47a42f | (Stefan Seifert)++ | src/core/operators.pm:
Fix (1,2,6...*)[5] not failing as it should

Fixes the remaining sequence spec tests. Sequences are now pretty lazy, so even throwing the deduction errors lazily on top might just be too lazy.		 16:58
timotimo cool! :)
nine [Coke]: ^^^
[Coke] folks running "spectest" aren't running all the tests, which might explain the diff.
nine: will insure my current run includes that. 16:59
16:59 CQ left
[Coke] (I haven't been running the inline::perl5 tests in my summaries, but I assume nine has. :) 17:03
17:03 perl6_newbee left
nine Correct :) Intriguingly, they have been passing for weeks already 17:04
17:04 domidumont joined 17:05 Khisanth left
skids has trouble understanding why (1,2,6...*)[5] would not just iterate up to the 6th element and return it. 17:06
pink_mist iterate how?
skids Oh, because 1,2,6..* itself is a problem. I see now. 17:07
TimToady but it should still try, and fail, as it does with [0] 17:08
skids btw I have some 3/4th complete code for (someseq)[...] = someotherseq and to fix @a[...] = some_iterable. 17:11
But for some reason MMD seems to not be very happy in the mire of all the postcircumfix:<[ ]> candidates.
I don't know if that's a new GLR thing or whether it's always been iffy in that area. 17:12
TimToady um, assigning to elements of a list seems iffy, since lists are immutable and do not guarantee containers (plus they'd be anonymous there anyway) 17:13
skids Especially constraints on \SELF in those cnadidates seem to be ignored.
TimToady or was that a sequence of containers?
skids Sequence of containers.
(and get a cannot-assign error if you run into an immutable in ther) 17:14
A lingering question is what the eagerness of list assignment for the seq = iteratable case should be -- none? chained to the left? 17:16
17:18 mohij joined 17:20 FROGGS joined
FROGGS uhh, it is getting toady in here... 17:20
17:20 FROGGS[mobile] left
FROGGS ahh, better :o) 17:20
17:22 lizmat joined
nine On a make stresstest we're now failing t/spec/integration/99problems-21-to-30.t and t/spec/integration/advent2012-day04.t 17:23
FROGGS lizmat: btw, my plan is to make the CU.load stuff work (== spectest clean), and then hand it over to you for api refactoring if there is any need and if you feel like
lizmat FROGGS: sounds like a plan to me... 17:24
FROGGS :o)
lizmat will be driving for most of tomorrow anyways
17:24 spidermario joined
FROGGS though, will have to make sure Inline::Perl5 keeps working of course 17:24
17:24 spider-mario left
lizmat FROGGS: of course :-) 17:25
17:25 Khisanth joined 17:26 andreoss joined
dalek ast/glr: 0cfc360 | (Stefan Seifert)++ | integration/advent2012-day04.t:
Fix a source of "expected 'Positional' but got 'Int'" in advent2012-day04.t
 17:26
17:27 MadcapJake left
andreoss timotimo: with moar-glr it's about the same 17:27
92 min
17:28 cognominal joined
andreoss .is-prime is slower than Math::Prime::XS but not 1000 times 17:29
JimmyZ muraiki: github.com/MoarVM/MoarVM/commit/68...81897c168d 17:30
muraiki JimmyZ: awesome! thanks :) 17:31
timotimo JimmyZ: good catch! 17:34
17:35 pullphinger left
JimmyZ thank you 17:36
nine m: my @a = 1; say 0, @a Z+ @a, 0;
camelia rakudo-moar 5ba44f: OUTPUT«1 1␤»
GLRelia rakudo-moar d47a42: OUTPUT«(1 1)␤»
17:36 pullphinger joined
nine m: my @a = 1, 1; say 0, @a Z+ @a, 0; 17:36
camelia rakudo-moar 5ba44f: OUTPUT«1 2 1␤»
GLRelia rakudo-moar d47a42: OUTPUT«(2 2)␤»
dalek ast/glr: 91cd7dc | (Stefan Seifert)++ | integration/advent2012-day04.t:
Fix generating Pascal's triangle in GLR

Fixes integration/advent2012-day04.t		 17:39
nine So there's really only a single spec test file failing now :) 17:40
[ptc] \o/
nine And that will not require any changes to rakudo. It's just a basic GLRification of a single test. 17:41
TimToady m: my @a = 1, 1; say 0, |@a Z+ |@a, 0; 17:43
camelia rakudo-moar 5ba44f: OUTPUT«5===SORRY!5===␤Variable '&prefix:<|>' is not declared␤at /tmp/Wi9cs20cEx:1␤------> 3my @a = 1, 1; say 0, 7⏏5|@a Z+ |@a, 0;␤Variable '&prefix:<|>' is not declared␤at /tmp/Wi9cs20cEx:1␤------> 3my @a = 1, 1; say 0, |@a Z+ 7⏏5|@a, 0…»
GLRelia rakudo-moar d47a42: OUTPUT«(1 2 1)␤»
nine Oh, so the parens would not have been necessary
TimToady list infixes are, like, list infixes, and stuff 17:44
that's why hypers aren't :) 17:45
17:46 woolfy joined
nine Anyway, why am I fixing spec tests? I should be a) making dinner b) work on my YAPC talk or c) both 17:47
TimToady & Glo --> Restaurante Páprika 17:48
dalek ast/glr: 34116e7 | (Stefan Seifert)++ | integration/99problems-21-to-30.t:
GLRify test in integration/99problems-21-to-30.t

This is it! The final spec test!		 17:50
nine It is...done.
lizmat nine++ !!!!
nine [Coke]: ^^^
lizmat congratulations!!!
woolfy nine++ woohooo!!! 17:51
cdc nine++
muraiki w00t 17:52
nine++
17:52 andreoss left
nine Yesterday on the train I wasted half an hour on this test. Today I just discarded that, went in, added a couple of characters and it just worked on the very first try :) 17:52
vendethiel nine: AMAZING :D 17:53
muraiki the joy of programming
nine And I still don't know what this thing is actually doing *g*
woolfy Magical programming! 17:54
FROGGS nine++ # \o/
17:54 telex left
nine All tests successful. 17:55
Files=1093, Tests=121114, 289 wallclock secs (17.66 usr 2.29 sys + 1969.59 cusr 73.32 csys = 2062.86 CPU)
Result: PASS
17:55 telex joined 17:56 mprelude left
dalek ast/glr: b458004 | TimToady++ | integration/advent2012-day04.t:
remove superstitious parens
 17:56
lizmat nine: I was wondering why the $ was needed on the return $[] 17:59
perhaps that's a sig issue with "sub" return?
18:00 larion joined
lizmat Files=1082, Tests=121039, 355 wallclock secs (22.69 usr 4.15 sys + 2341.63 cusr 135.15 csys = 2503.62 CPU) 18:02
1 fail: IO::Socket::INET (OS X specific fail)
18:03 zakharyas left
nine lizmat: it's needed because group is a recursive sub and its result is used in a map call 18:04
dalek rl6-roast-data: 6879af9 | coke++ | / (2 files):
another glr run for 'today'
 18:09
18:10 rurban left
[Coke] shapecatcher.com/ - can't find a unicode glyph? draw it. 18:11
18:18 larion left
mohij Rakudos build process is targeting perl 5.8.0, right? 18:22
18:24 llfourn left
geekosaur 5.10.0 I think 18:32
there's been discussion about wanting stuff from 5.10.1 but not wanting to bump the dependency
FROGGS m: require Flubber:from<Foo>
camelia rakudo-moar 5ba44f: OUTPUT«Could not find Flubber in any of:␤ file#/home/camelia/.perl6/2015.07.1-182-g5ba44fc/lib␤ inst#/home/camelia/.perl6/2015.07.1-182-g5ba44fc␤ file#/home/camelia/rakudo-inst-1/share/perl6/lib␤ file#/home/camelia/rakudo-inst-1/share/perl6/vendor/lib…»
GLRelia rakudo-moar d47a42: OUTPUT«Could not find Flubber in any of:␤ file#/home/steve/.perl6/2015.07.1-679-gd47a42f/lib␤ inst#/home/steve/.perl6/2015.07.1-679-gd47a42f␤ file#/home/steve/sandbox/perl6/rakudo/install/share/perl6/lib␤ file#/home/steve/sandbox/perl6/rakudo/install/…»
FROGGS m: require "Flubber":from<Foo>
camelia rakudo-moar 5ba44f: OUTPUT«Could not find file 'Flubber' for module Flubber␤ in any load_module at src/gen/m-ModuleLoader.nqp:200␤ in block <unit> at /tmp/bIJCLEJZrx:1␤␤»
GLRelia rakudo-moar d47a42: OUTPUT«Could not find file 'Flubber' for module Flubber␤ in any load_module at src/gen/m-ModuleLoader.nqp:200␤ in block <unit> at /tmp/K3X3oSaLPd:1␤␤»
FROGGS bug™
lizmat :-) 18:34
[ptc] nine++ # final spec test! 18:38
mohij geekosaur: thanks! 18:42
leont Depending on 5.10.0 is a bit silly, it's not very commonly used in the wild, and generally considered a flawed release. Might as well depend on 5.10.1 18:44
lizmat dinner& 18:47
18:47 lizmat left
pyrimidi_ leont: IIRC most of the 5.x.0 releases kinda fall into that 'flawed' category (same was said of 5.8.0) 18:47
18:47 blackcat_ joined
leont The same is true for 5.8.0 actually, for different reasons 18:47
pyrimidi_: yeah, before 5.12 at least
18:48 Sgeo_ joined
leont Nowadays it's just ".1 has a few bugs fewer" instead of "you've got to be insane to use this in production" 18:48
18:48 cschwenz joined, woolfy left
masak on some days I feel like you've got to be insane to use any software in production. 18:49
18:50 Loren_ left, Sgeo left
nine Or hardware 18:50
huf or people, really.
masak nothing works 18:51
leont L'enfer, c'est les autres!
huf masak: so how come a lot of things work?
how do things get better? that's the great mystery
masak huf: measurement error.
or dumb luck
pyrimidi_ chaos 18:52
moritz it's delusion to think lot of things work
speaking of not working
18:53 domidumont left
moritz I can't make a rakudo release because there's no moarvm release 18:53
pyrimidi_ www.youtube.com/watch?v=JmzuRXLzqKk
moritz and I can't make a maorvm release, because I don't understand moarvm well enough to write a sensible changelog file
so, no rakudo release this month, it seems
masak moritz: well, you tried
moritz well, I failed. 18:54
masak at least until someone who knows how makes a moarvm release.
moritz where I live, the month has only about 3 more hours, of which I don't plan to be awake more than 2 or so 18:55
so... unlikely
japhb dangitall 18:57
vendethiel masak: it's fine if it's only some days!
18:57 rurban joined
nine [ptc]: couldn't you write the MoarVM ChangeLog? 18:58
18:59 xinming_ joined
japhb moritz: At any given moment, there are an uncountable infinity of things not working, but thanks to the wonders of quantum mechanics, we don't notice them. On the other hand, this says absolutely nothing about the number of things that *are* working, which anecdotally I would say is a lot. Where by 'anecdote', I mean "apparent continuing existence of Earth, the Solar System, our galaxy, and in fact (as far as we can tell) the entire visible universe". 18:59
vendethiel japhb: I'm not sure it's quantum mechanics, I think it's rather my terrible eyesight 19:00
japhb vendethiel: Additive effects.
vendethiel *g*
japhb Nothing like philosophising on a full stomach and an empty brain.
[ptc] nine: I could give it a go. When does it need to be done by? 19:01
19:01 xinming left
nine [ptc]: well moritz's still awake for some 2 hours ;) 19:02
japhb [ptc]: Unless you can convince moritz to stay awake later, I'd say "quite soon"
nine I'm packing for a conference. Seems like I've only been at one yesterday...
Oh wait, I was!
[ptc] nine: I'm also packing for a conference ;-P
I'm giving it a go, however hoelzro, timotimo and JimmyZ have done a lot of stuff that I don't know about (or yet understand) 19:06
moritz japhb: there you have it, just anecdotes, no data :-) 19:07
19:10 pyrimidi_ is now known as pyrimidine 19:14 rindolf left
RabidGravy m: class C:ver<v0.0.1> { }; say C.^ver ; # anyone know where I can start looking to fix this? 19:14
camelia rakudo-moar 5ba44f: OUTPUT«Method 'gist' not found for invocant of class 'NQPMu'␤ in block <unit> at /tmp/WuBp2uGt4Y:1␤␤»
GLRelia rakudo-moar d47a42: OUTPUT«Method 'gist' not found for invocant of class 'NQPMu'␤ in block <unit> at /tmp/KDzGgWyHem:1␤␤»
19:15 yqt joined, larion joined
FROGGS RabidGravy: perhaps look at Perl6::Grammar and Perl6::Actions and check what the package_declarator does with colonpairs 19:15
RabidGravy: (colonpair via longname) 19:16
[ptc] wow, timotimo used "fabbergasted" in a commit message. I'm impressed! Haven't seen that used in long time :-)
Ulti has a go at getting his modules to run from the GLR branch...
moritz is that a misspelling of "flabbergasted"? 19:17
19:19 ChoHag left, rindolf joined 19:20 mprelude joined
zostay just published a blog post regarding P6SGI and the latest draft i finished this morning after having an epiphany last night, which probably should have come sooner 19:20
blogs.perl.org/users/sterling_hanen...ewing.html
for anyone interested
dalek kudo/glr: 0f8cdfb | moritz++ | src/core/operators.pm:
Simplify usage of die().

When it was still fail(), it needed to be nested in a function call, because fail() exits the current function and returns the failure.
die() needs no such sophistication
moritz m: await start { say 42 } 19:21
camelia rakudo-moar 5ba44f: OUTPUT«42␤»
GLRelia rakudo-moar d47a42: OUTPUT«42␤»
moritz zostay: start { } has been deprecated? really?
why is there no deprecation warning?
how did I miss this? 19:22
RabidGravy it hasn't
zostay it says it has in asyncops
as of 2015.06
RabidGravy no, thats some other usage 19:23
ugexe proc::async, and thats in some other branch
zostay okay, then i misread the source... i'll clear that out and adjust the spec then 19:24
RabidGravy m: sub foo {}; start(&foo)
camelia ( no output )
GLRelia rakudo-moar 0f8cdf: OUTPUT«Saw 1 occurrence of deprecated code.␤================================================================================␤start({ ... }) seen at:␤ /tmp/dD8_FHcXNS, line 1␤Deprecated since v2015.6, will be removed with release v2015.9!␤Please use sta…»
[ptc] timotimo, JimmyZ, hoelzro, jnthn, FROGGS: just updated the MoarVM ChangeLog and pushed a PR. If I've missed something or got something wrong, please update. 19:25
19:25 cschwenz left
RabidGravy so start as a function has been deprecated basically 19:26
zostay that makes me happier
19:27 larion left
hoelzro [ptc]++ 19:28
19:29 mohij left
nine moritz: ah, I've been wondering what the whole take sub { fail business was about 19:29
19:29 mohij joined 19:30 mj41 joined 19:31 [Sno] left
dalek rl6-roast-data: b8c3976 | coke++ | / (2 files):
another glr run for 'today'
 19:33
19:33 larion joined
moritz nine: so, are we waiting for something in particular before merging? 19:33
[Coke] note I still have one failure, but it's probably a flapper. 19:34
nine moritz: I'd actually rather rename glr to master... But really we're just waiting for someone to say, that it's ok.
moritz nine: I've had the same thought (renmaing to mater) 19:35
[Coke] S17-promise/stress.t aborted 1 test(s)
19:35 rangerprice joined
moritz we've been wanting to switch back to a 'master' branch for a while now, but didn't have a good trigger 19:35
nine moritz: I don't think I remember anyone at the SPW being against it.
[Coke] +1 from me on going to master... but also +1 from me on merging glr back to nom first before we decide on that. 19:36
rangerprice Where i can find a good book reference to create my own interpreter/compiler for a programming language ?
hoelzro I second [Coke]'s suggestion
nine [Coke]: S17-promise/stress.t is passing here. Considering that it's testing thready features, I really wouldn't be surprised if it was a flapper.
rangerprice hi moritz :)
[Coke] nine: my thoughts exactly. it did fail here twice in a row, and I only run one test at a time. 19:37
Ulti does Positional still exist?
nine Ulti: yes
Ulti hmmm
nine Ulti: though in most cases, you can probably require Iterable instead and be more flexible
Ulti and is Seq a Positional?
nine Ulti: no 19:38
Seq does Iterable
If you get some "cannot bind to Positional" error, just use assignment instead
dalek kudo/cu_load: 7262b98 | FROGGS++ | src/ (7 files):
move module loading code to CompUnitRepo

This means we are calling CompUnitRepo.load_module for every use/require/need we hit. Later we move the CU loading code to CompUnit.load, and will work with CompUnit objects in CUR instead. This patch fails one test of integration/precompiled.t		 19:39
Ulti yeah Iterable looks like the right change here thanks nine 19:40
mj41 Can anybody try to reproduce glr vs. nom speed regression of "say $_ for 1..2_000_000;" gist.github.com/mj41/6761c021135ae...1-txt-L129 ? Seems like rakudo issue (3m25s instead of 0m10s). nqp-m is a bit faster.
Ulti at least the only thing being relied upon is .elems
zostay type coercion is possible when a type has a method named for the target type, right? 19:41
moritz mj41: it's probably an optimization that wasn't reinstanted for GLR yet 19:42
nine time perl6 -e 'use nqp; nqp::say($_.Str) for 1..2_000_000;' > /dev/null takes only 2.4 seconds instead 19:43
mj41 moritz: merge blocker? as perl5 -MMoose takes 0m0.356s
so pre glr 10s was slow and now it is super slow :-) 19:44
19:46 AlexDaniel joined 19:47 telex left
moritz mj41: I don't quite see the connection between a "for"/say benchmark and -MMoose 19:47
_itz_ m: say Parcel.new.perl 19:48
camelia rakudo-moar 5ba44f: OUTPUT«()␤»
GLRelia rakudo-moar 0f8cdf: OUTPUT«5===SORRY!5=== Error while compiling /tmp/uZHw8pltfX␤Undeclared name:␤ Parcel used at line 1␤␤»
mj41 moritz: only startup time penalty :-)
19:48 telex joined
AlexDaniel Hi! 19:48
Ulti weird I am getting a SORRY! without any line number
"Method 'iterator' must be implemented by Iterable because it is required by a role"
hoelzro AlexDaniel: greetings!
_itz_ I did wonder if a clearer deprecation message for Parcel use might help .. the one from instally Test::Mock is particularly odd 19:49
mj41 moritz: gist.github.com/mj41/6761c021135ae...e1-txt-L23 vs. gist.github.com/mj41/6761c021135ae...e1-txt-L28
moritz Ulti: do you have any class that implements the Iterable role?
Ulti: if yes, it needs updating
AlexDaniel It seems like a lot of reports were merged into rt.perl.org/Public/Bug/Display.html?id=61602 and so I thought, maybe this report should be merged as well rt.perl.org/Public/Bug/Display.html?id=125336 ?
Ulti nope or at least not that I'm aware of
_itz_ s/instally/installing 19:50
Ulti just a package with a tonne of exported subs no OO at all
AlexDaniel it looks like it is related. However, I have no idea what should these infinite ranges actually mean, so perhaps it is a separate issue 19:51
nine Iterable used to be a class. s/is Iterable/does Iterable/ can help wonders
AlexDaniel m: .say for Inf..0 19:52
GLRelia ( no output )
Ulti I'm going to just assume its a List being passed in I only really had Positional because Parcel would turn up
19:52 blackcat_ left
AlexDaniel oh wow 19:52
camelia rakudo-moar 5ba44f: OUTPUT«(signal XFSZ)-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-9223372036854775808␤-922337203685…»
AlexDaniel m: .say for NaN..0
camelia rakudo-moar 5ba44f: OUTPUT«(signal XFSZ)NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN…»
GLRelia rakudo-moar 0f8cdf: OUTPUT«(timeout)NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤NaN␤N…»
zostay is there a brief syntax for performing coercion that doesn't involve a signature or assignment? 19:53
Ulti is there a nice picture of what the types look like post GLR anywhere?
so I can investigate without diving into Rakudo source
19:54 baest_ joined
AlexDaniel Ulti: postcircumfix.com/media/typegraph_a...kudo.svg?2 ? 19:54
Ulti: (from postcircumfix.com/2015-08-29-untang...hierarchy)
moritz m: say Int("42").perl; # for zostay
camelia rakudo-moar 5ba44f: OUTPUT«42␤»
GLRelia rakudo-moar 0f8cdf: OUTPUT«42␤»
zostay m: say Promise(Supply.new).perl; 19:55
camelia rakudo-moar 5ba44f: OUTPUT«Promise.new(scheduler => ThreadPoolScheduler.new(initial_threads => 0, max_threads => 16, uncaught_handler => Callable), status => PromiseStatus::Planned)␤»
GLRelia rakudo-moar 0f8cdf: OUTPUT«Promise.new(scheduler => ThreadPoolScheduler.new(initial_threads => 0, max_threads => 16, uncaught_handler => Callable), status => PromiseStatus::Planned)␤»
zostay m: say Promise(Promise.new).perl;
camelia rakudo-moar 5ba44f: OUTPUT«Cannot find method 'Promise'␤ in block <unit> at /tmp/apKIDc5HYd:1␤␤»
GLRelia rakudo-moar 0f8cdf: OUTPUT«Cannot find method 'Promise'␤ in block <unit> at /tmp/dwC9KYs57w:1␤␤»
Ulti thanks AlexDaniel though that's kind of what I imagined
zostay that doesn't always work
moritz m: say Promise.new.Promise 19:57
camelia rakudo-moar 5ba44f: OUTPUT«Method 'Promise' not found for invocant of class 'Promise'␤ in block <unit> at /tmp/XtqRIWial_:1␤␤»
GLRelia rakudo-moar 0f8cdf: OUTPUT«Method 'Promise' not found for invocant of class 'Promise'␤ in block <unit> at /tmp/18NsrjiTqg:1␤␤»
moritz zostay: just like oridinary coercion doesn't always work
zostay well, Promise(Any) $p as a parameter always works... seems like there should be a syntax that always works without having to say $p ~~ Promise orself $p.=Promise or something similar 19:59
*orself
*orelse
timotimo [ptc]: these are not all things i would have put into the changelog
zostay dang
timotimo like the parse_jitgraph thing just came out of nowhere, because the tool it refers to wasn't mentioned in the changelog; and it's only for internal use anyway 20:00
and the "jit graph is now an expression tree" thing refers to brrt's excellent work that - alas - hasn't reached the master branch yet 20:01
20:02 laubster left
timotimo but there's good stuff in there, too :) 20:03
[ptc] timotimo: feel free to update :-) 20:04
timotimo: I made lots of guesses as to what I should add while doing some repo archaeology
[Coke] enjoys his actual day off before going back to work.
[ptc] should one be spectesting nom or glr for the current release? 20:05
t/04-nativecall/16-rt125408.t still fails for me
RabidGravy yeah
grr
it keeps getting fixed "by accident" 20:06
Ulti is there a new and wonderful way to create sequences?
RabidGravy that's why I put the test in
dalek kudo/glr: 994fb0a | (Stefan Seifert)++ | src/core/io_operators.pm:
Add generic single arg candiate of say back in

Speeds up say $_ for 1..2_000_000; by about 10x		 20:07
nine mj41: ^^^
moritz: ^^^ 20:08
RabidGravy :)
moritz nine++
RabidGravy wtawtaw
moritz nine: is the slurpy still called?
nine moritz: passes all spec tests
mj41 nine++ ... anybody to try another 10x ? :-) 20:09
nine yep, seems to be doing just fine
[Coke] current release has to be nom, not glr. 20:11
20:11 bronco_creek joined
bronco_creek .tell lizmat Saw Yes in concert last night. For some reason thought of you when they played some Tormato songs. (Sorry to be far off topic.) 20:12
yoleaux bronco_creek: I'll pass your message to lizmat.
mj41 0m16s .. much better 20:13
20:13 diana_olhovik_ joined
RabidGravy m4lj, 94% of that time is in the "say" 20:14
that is that loop calling an empty function is ~15x faster 20:16
moritz RabidGravy: with or without nine's last patch?
RabidGravy without 20:18
let me check again
mj41 nqp is 4x faster gist.github.com/mj41/6761c021135ae...e1-txt-L38 now
colomon nine++ 20:19
20:21 llfourn joined 20:22 rindolf left
RabidGravy moritz, the last was against nom 20:22
nom is twice as fast for that as glr with nine's patch 20:23
20:24 cschwenz joined
RabidGravy calling an empty function in that loop is 23x faster on glr with nine's patch 20:24
so it's say still slo(er)
slow
nine I think we're still missing the for -> while loop optimization 20:25
mj41 nope gist.github.com/mj41/6761c021135ae...e1-txt-L38
line 53 line 58
gist.github.com/mj41/6761c021135ae...1-txt-L137
20:26 llfourn left, darutoko left
nine Please note that the version in glr will probably produce nicer output in threaded programs that just say without any locking 20:27
RabidGravy also 'say join "\n", 1..2_000_000' is 4x faster than the loop version (there's a moral in there) 20:30
moritz well yes, "do less" 20:31
RabidGravy aye
mj41 Ok. Good too know before I go to bed. Tiaveling to Granada.
err, Tired traveling to Granada. 20:32
good night
RabidGravy have fun
20:33 mj41 left
dalek kudo/glr: 57b7ebc | (Stefan Seifert)++ | src/core/io_operators.pm:
Add print multi candidates for single arguments

Have an optimized version of print to gain the same speed advantages as say. For print we need to have separate candidates for Iterables and the rest because @array.Str would include spaces we don't want with print. Unless @array is itemized of course (GLRR).		 20:35
kudo-star-daily: 79609fe | coke++ | log/ (9 files):
today (automated commit)
 20:38
[Coke] GLR R? 20:40
20:41 xfix left, rurban left
[ptc] I'm getting "MVMArray: Can't shift from an empty array" in t/spec/S17-promise/start.rakudo.moar in the spectests on nom. Dunno if that's useful to anyone... 20:42
20:43 rurban joined 20:47 lizmat joined
nine [Coke]: GLR Revised ;) Means that we don't flatten a single arg if it's itemized 20:49
[Coke] only if it's a release blocker, I guess. 20:52
(to ptc) 20:53
_itz_ m: say 'GLR'~ ('R' x 1..Inf)
camelia rakudo-moar 5ba44f: OUTPUT«GLRR S T U ...␤»
GLRelia rakudo-moar 57b7eb: OUTPUT«GLR...␤»
TimToady useless use of x 1 ? 20:55
m: say 'GLR' ~ ('R' .. Inf) 20:56
camelia rakudo-moar 5ba44f: OUTPUT«GLRR S T U ...␤»
GLRelia rakudo-moar 57b7eb: OUTPUT«GLR...␤»
_itz_ oh yes
TimToady m: say 'GLR' ~ ('R' Xx (1 .. Inf))
camelia rakudo-moar 5ba44f: OUTPUT«(timeout)» 20:57
GLRelia rakudo-moar 57b7eb: OUTPUT«(timeout)»
TimToady ~ forces eager, I guess
[Coke] we've got a bunch of tests marked "submix behavior still under discussion" that are passing in glr. ok to unfudge? 20:58
20:59 skids left
[Coke] seems like testing that current functionality stays working is good. 20:59
moritz +1 21:00
lizmat [Coke] +1
yoleaux 20:12Z <bronco_creek> lizmat: Saw Yes in concert last night. For some reason thought of you when they played some Tormato songs. (Sorry to be far off topic.)
lizmat bronco_creek: I'm jealous
[Coke] ah... 15 tests in a row; every other test passes.
(previously all 15 were skipped)
That is slightly more dubious.
(going for it anyway.) 21:01
lizmat bronco_creek: On the other hand, I will be able to see Yes in the Cruise To The Edge in November :-)
and yes, "Tormato" is one of my favourite Yes albums :-)
alas, I will not be able to see the Rickenbacker Master in action :-( 21:02
bronco_creek Lizmat: Vary sad to loose Chris. Sherwood filled in remarkably well on both bass and vocal IMO. 21:04
lizmat looking forward to that :)
Yes will go on, as Chris wanted it :-)
_itz_ . o O ( they should get Trevor Horn back ) 21:05
lizmat bronco_creek: did Sherwood do The Fish ?
bronco_creek No. Only bass solo was on the encore rendition of Starship Trooper. 21:06
virtualsue watches _itz_ 's troll go unanswered 21:07
bronco_creek Also, Jon D. is excellent. Close your eyes and it could be Jon A. Open them, and the same holds true...
21:11 diana_olhovik_ left
lizmat can't wait for november: cruisetotheedge.com 21:11
bronco_creek _itz_: ;-) Yes line-up permutations should be a rosetta code problem. 21:12
21:12 kaare_ left 21:13 muraiki left
_itz_ using the Pete::Frame::Tree module 21:14
[Coke] 1,031 tickets 21:20
21:21 llfourn joined
jdv79 i don't think i've ever done a 8h layover. should have checked for more alternative routes maybe:( 21:22
timotimo so ... what do i do about the test for #125408 staying fixed? 21:24
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=125408
timotimo just release anyway?
21:25 cognominal left
[Coke] jdv79: I had a 9 hour layover scheduled on the way back but managed to rebook . 21:25
21:26 llfourn left
dalek ast/glr: 3d283da | TimToady++ | S06-signature/unpack-array.t:
only needs .List, not .list, doesn't cache
 21:26
[ptc] 'night everyone! 21:28
timotimo gnite [ptc]!
21:28 ggoebel joined 21:29 xnrand is now known as tumblr
lizmat gnight [ptc] 21:30
lizmat also gets some sleep&
timotimo * [new tag] 2015.08 -> 2015.08 21:32
hoelzro night [ptc], lizmat
timotimo maybe i should have waited for the spec test to run through, though?
dalek ast/glr: cdab420 | coke++ | S0 (10 files):
(auto) unfudge several tests.

  ... and close some RTs.		 21:36
ast/glr: 97b6b02 | coke++ | S05-capture/match-object.t:
Comparing types doesn't need .gist

Update comment for RT #124998
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=124998
colomon t/spec/S17-procasync/many-processes-no-close-stdin.t (Wstat: 65280 Tests: 0 Failed: 0) 21:38
timotimo t.h8.lv/MoarVM-2015.08.tar.gz
colomon Non-zero exit status: 255
Parse errors: Bad plan. You planned 1 tests but ran 0.
Ulti well the Stats module is now GLR ready :3
timotimo .tell jnthn when you're back, could you upload t.h8.lv/MoarVM-2015.08.tar.gz to moarvm.org?
yoleaux timotimo: I'll pass your message to jnthn.
hoelzro colomon: ='( 21:40
timotimo moritz: we have a moarvm release, at least in theory
Ulti Bag.kxxv.flat.list to get back to a list you might call bag on is a bit weird though
hoelzro colomon: is that on GLR? have you pulled recently?
colomon hoelzro: I pulled earlier this afternoon
I know I didn’t get cdab420 above 21:41
hoelzro colomon: if you run that test with just ./perl6-m, do you get any output?
colomon might be a flapper or something
hoelzro (it will probably throw up a bunch of output about /tmp/test-file missing)
it shouldn't flap
21:42 bronco_creek left 21:44 vendethiel left
[Coke] someone opened a twitter account with [email@hidden.address] 21:46
japhb .tell bronco_creek Years ago I came across a Santana album in which the liner notes consisted of a complex infographic of the changing band lineup, just so fans could get it straight.
yoleaux japhb: I'll pass your message to bronco_creek.
[Coke] just stole it back. weird.
21:52 mrocks77 joined, mrocks77 left
colomon hoelzro: 21:53
cat: /tmp/test-file: No such file or directory
too many open files
in block <unit> at t/spec/S17-procasync/many-processes-no-close-stdin.t:7
hoelzro =(
colomon: what do moar/nqp-m/perl6 --version all output?
I just bumped the NQP_REVISION file to use the fix this morning
colomon hoelzro: oh, I’m probably missing that. 21:54
hoelzro if you've pulled but not re-Configure.pl'd, that would explain it
colomon give me a minute
masak BooK: the supply/whenever stuff only works in the glr branch, IIUC. 21:55
21:55 [Sno] joined
masak oh, someone already told you. nvm. 21:55
colomon hoelzro: right, that’s probably it. In middle of full rebuild at the moment. 21:57
hoelzro whew
masak 'night, #perl6 22:00
colomon hoelzro++ # works after more complete rebuild 22:01
hoelzro \o/
night masak
colomon o/ masak
with complete update, I get a completely clean spectest from GLR \o/ 22:06
awwaiid hops up and down
22:07 mohij left, tumblr is now known as xnrand
colomon bad news (for the moment) is that Audio::Taglib::Simple won’t install on GLR. guess I need to switch back to nom for a minute. 22:08
_itz_ there is some ecosystem breakage as would be expected 22:09
22:09 bbkr left, bbkr joined 22:11 skids joined
colomon _itz_: of course. I’m hoping to get a decent measure of how much breakage and where later today or tomorrow. 22:15
_itz_ some of the errors look a little odd 22:17
22:18 [TuxCM] left
_itz_ the text of the messages I mean 22:18
22:19 [TuxCM] joined 22:20 pullphinger left 22:23 TEttinger joined 22:24 rurban left 22:31 bbkr left 22:36 espadrine_ left
RabidGravy how soon is the glr going to become the mainstream? I've been holding off testing my modules on the grounds I couldn't be arsed to do it multiple times. 22:41
BooK masak: yes, I've been told that and the examples from jnthn talk worked better on it when I tried :-) 22:42
22:42 muraiki joined
BooK mmm, ok so I was reading my backlog too :-) 22:43
22:44 rangerprice left 22:45 rangerprice joined
RabidGravy of course until the rt#125408 is fixed I'm not sure what the point would be 22:45
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=125408
22:54 spidermario left
muraiki m: say chrs(0x1D047, 0x1D0F0) 22:56
camelia rakudo-moar 5ba44f: OUTPUT«𝁇𝃰␤»
GLRelia rakudo-moar 57b7eb: OUTPUT«𝁇𝃰␤»
22:56 thou left
muraiki hrm. I don't suppose anyone can recommend a decent intro to unicode? I thought that 0x1D0F0 was a combining character, but it seems like it's not behaving that way 22:56
23:01 RabidGravy left
muraiki darn, maybe it's not a combining character. :( 23:07
geekosaur .u U+1D0F0 23:10
yoleaux U+1D0F0 BYZANTINE MUSICAL SYMBOL KENTIMATA NEO MESO [So] (𝃰)
muraiki yeah, I think I can't use this the way I hoped to
oh well
23:11 aborazmeh joined, aborazmeh left, aborazmeh joined
muraiki I guess it was too much to expect unicode to not only have byzantine notation characters but also to work as a layout engine for byzantine chant :) 23:11
japhb Given the machine-native guts of github.com/rakudo/rakudo/commit/994fb0aa3f , why doesn't IO::Handle.print have a str multi to avoid the box and unbox of $str in the call to $out.print($str)? 23:17
I would think that might actually chew up a fair amount of GC and performance during heavy output, especially when you consider that multi sub print() has the same problem. 23:19
23:21 cognominal joined 23:22 llfourn joined
TEttinger muraiki: I'm curious what kind of project actually uses the really rare parts of unicode 23:23
(I guess they were added because they thought they would be used, but Byzantine is starting to describe some parts of implementing unicode it seems, heh) 23:24
23:27 llfourn left
muraiki TEttinger: hehe. I actually sing byzantine chant using byzantine notation. I'd like to write a FLOSS editor for it. it's actually the project that led me to switching careers to become a programmer 23:27
TEttinger wow! 23:28
muraiki but now that I'm a professional programmer, I haven't actually done much work on that project. seeing additional byzantine chant code points added in unicode 8 got my hopes up
at the very least I now have a way to encode neumes using an international standard
instead of "some weird font, character 7"
I wrote the initial version in racket, which turned a type of byzantine chant lisp into graphical output 23:29
TEttinger nice
muraiki there's a picture here github.com/muraiki/byzscribe
TEttinger I mean I have no idea if it's right or not, but that does look like a fair amount of work to input each character. hopefully perl6 can be easier to write this with? 23:33
muraiki yeah, I was hoping to use p6's great unicode support and grammars
and now I would be able to store the actual characters with some additional info, instead of long english names 23:34
I suppose I will have to pick up The Haskell School of Music again... it goes through learning music theory + haskell by developing a way to both encode music and render it. I've never done any kind of real programming that had to deal with laying out complex things graphically 23:35
23:36 aborazmeh left 23:39 virtualsue left 23:41 leont left 23:43 yqt left 23:48 aborazmeh joined, aborazmeh left, aborazmeh joined 23:59 thou joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!