»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.
00:01 rurban joined 00:04 boggard joined, Alina-malina left 00:08 skids joined 00:18 chenryn joined 00:21 chenryn left
[Coke] r: say (:::[]) 00:21
camelia rakudo-moar f55867: OUTPUT«===SORRY!===␤MVMArray: Index out of bounds␤»
..rakudo-jvm f55867: OUTPUT«===SORRY!===␤VMArray: Index out of bounds␤»
00:24 aindilis joined 00:26 chenryn joined
[Coke] anyone know who Pieter Zeemanweg is? 00:27
r: say ::::::::::::::[] 00:31
camelia rakudo-{moar,jvm} f55867: OUTPUT«5===SORRY!5=== Error while compiling /tmp/tmpfile␤Name component may not be null␤at /tmp/tmpfile:1␤------> 3say ::7⏏5::::::::::::[]␤»
00:36 telex joined 00:41 dpk joined 00:42 raiph joined 00:46 tokuhirom joined 00:51 tokuhirom left, aborazmeh joined, aborazmeh left, aborazmeh joined, colomon joined 00:53 ParsonsNose joined
ParsonsNose Perl6 just got me laid 00:53
Just sayin'
00:57 rurban left
tadzik congratulations 00:57
00:58 yeahnoob joined 01:00 aborazmeh left
ParsonsNose Now to disband this username... 01:05
And stop lying
01:05 ParsonsNose left 01:07 ParsonsNose joined, AndyDee joined 01:09 chenryn left
ParsonsNose I've heard that perl6 does graphemes correctly. So... If I did one of those classic evil multi-tiered cthulu letters, would it count as one per 'what we all think of as letter'? 01:10
Wow... My app didn't disguise me. I shoot myself.
01:17 Juerd joined 01:18 Actualeyes joined 01:19 aborazmeh joined, aborazmeh left, aborazmeh joined 01:40 Zoffix joined
dalek kudo-star-daily: d8b838d | coke++ | log/ (9 files):
today (automated commit)
 01:46
rl6-roast-data: e23f313 | coke++ | / (9 files):
today (automated commit)
01:51 rangerprice left, ShimmerFairy joined 01:53 SHODAN joined 02:13 cognominal joined 02:14 kid51 left 02:16 chenryn joined
dalek kudo/nom: 1fd9e8d | coke++ | src/Perl6/World.nqp:
Avoid internal compiler error for RT #115326
 02:16
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=115326
[Coke] RT: 1025 02:18
02:19 ParsonsNose left 02:20 raiph left, dayangkun_ joined 02:25 dj_goku joined 02:33 tokuhirom joined 02:42 Actualeyes left 02:52 Oatmeal joined
awwaiid Is there an IO String like thing? 02:52
TimToady see modules.perl6.org/ 02:54
awwaiid oh, found IO::String
TimToady++ # pushing me in the pond to catch some fish 02:55
02:59 BenGoldberg left 03:00 Ben_Goldberg joined
Ben_Goldberg m: sub bozosort ( @a is copy ) { until ([<=] @a) -> { my ($a, $b) = @a.keys.pick(2); @a[$a, $b] = @a[$b, $a] }; @a }; say bozosort( (1..5).pick(*) ); 03:01
camelia rakudo-moar 1fd9e8: OUTPUT«[1 2 3 4 5]␤»
Ben_Goldberg m: sub bozosort ( @a is copy ) { until ([<=] @a) -> { my ($a, $b) = @a.keys.pick(2); @a[$a, $b] = @a[$b, $a] }; @a }; say bozosort( (1..6).pick(*) );
camelia rakudo-moar 1fd9e8: OUTPUT«[1 2 3 4 5 6]␤»
Ben_Goldberg m: sub bozosort ( @a is copy ) { until ([le] @a) -> { my ($a, $b) = @a.keys.pick(2); @a[$a, $b] = @a[$b, $a] }; @a }; say bozosort( 'Just another bozo'.comb ); 03:03
camelia rakudo-moar 1fd9e8: OUTPUT«(timeout)» 03:04
03:15 Ben_Goldberg left 03:20 kaare_ joined 03:56 llfourn joined 04:01 rindolf joined 04:02 xenu joined
dalek osystem: 8561f3b | ugexe++ | META.list:
Net::HTTP github.com/ugexe/Perl6-Net--HTTP

yet another http client, now with connection caching/keep-alive		 04:18
04:30 xpen joined 04:48 rindolf left 04:58 Actualeyes joined 05:15 cognominal left 05:26 yeahnoob left, chenryn left 05:34 abraxxa joined 05:38 abraxxa left
dalek ast: 825dea9 | usev6++ | S03-operators/repeat.t:
Unfudge test for rakudo.jvm (RT #123830)
 05:44
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=123830
05:52 abraxxa joined 05:56 skids left 06:00 crux joined 06:04 cgfbee left 06:09 davido_ joined 06:11 cgfbee joined 06:12 Ch0c0late joined 06:17 mayuresh joined 06:25 dayangkun_ left 06:26 dayangkun joined 06:28 abraxxa left 06:30 chenryn joined 06:31 mayuresh left 06:37 bjz joined 06:41 bjz_ joined, bjz left 06:44 dayangkun left 06:48 bjz joined 06:50 bjz_ left 06:58 ChoHag joined 07:06 dayangkun joined 07:07 dayangkun left
[Tux] test 50000 39.079 38.959 07:07
test-t 50000 38.440 38.321
07:08 dayangkun joined 07:13 bjz_ joined, davido_ left 07:15 bjz left 07:16 dayangkun left 07:18 FROGGS joined
El_Che what's the header for [Tux]'s tests? 07:25
dalek ast: 8e47b67 | usev6++ | S02-literals/string-interpolation.t:
Add test for RT #120449
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=120449
07:27 leont joined
moritz \o 07:31
07:35 [Sno] joined 07:36 abraxxa joined 07:39 domidumont joined
[Tux] El_Che, you mean what the columns mean? 07:39
07:44 domidumont left 07:45 mr-foobar joined
El_Che [Tux]: yes 07:45
07:45 RabidGravy joined
[Tux] Text::CSV performance test: test is a reference script (no extra funtionality), test-t is full-featured 07:46
50000 is the number of fields parsed: 10_000 lines with 5 fields each
07:47 leont left
[Tux] next column is the time in seconds used to do the job 07:47
pink_mist so the full-featured script is faster than the one with no extra functionality? 0_o
[Tux] tux.nl/Talks/CSV6/speed4.html 07:48
pink_mist, the reference script has only been changed to work under GLR 07:49
the other script has been changed a lot more
so yes, that makes sense
07:49 rurban joined 07:53 domidumont joined 07:54 domidumont left
El_Che [Tux]: thx! 07:56
[Tux] I am open to other optimisations :) 07:57
07:57 Actualeyes left 07:58 darutoko joined 08:00 g5 joined
El_Che [Tux]: having trouble understanding the graph. X is time, Y is time (seconds to run test-t). Ok. So with what scenario correpond the 4 graphs? 08:04
[Tux] first graph is full range, the other three zoom in on a smaller speed range 08:05
the red parts do not fir the range
only vertical zooming, not horizontal zooming 08:06
El_Che [Tux]: I see
your module is a good candidate for speed testing 08:07
[Tux] more like the canary in the coalmine
08:09 azawawi joined
El_Che p5 text::cvs_xs is pretty fast 08:09
azawawi hi
El_Che hi azawawi
azawawi El_Che: :)
El_Che [Tux]: I was impressed after some superficial native call exploring. What's your experience coming from an XS background? 08:10
azawawi is working on his unzip.p6 script (see github.com/azawawi/scripts/blob/ma.../unzip.p6)
[Tux] for the same task, Text::CSV_XS takes 0.017 seconds: tux.nl/Talks/CSV6/speed3.html 08:11
azawawi and the p6 version?
[Tux] 39 seconds
perl6 + Inline::Perl5 + Text::CSV_XS = 16 seconds 08:12
dalek c/teodozjan-patch-1: 8b0e97e | (Kamil Kułaga)++ | doc/Language/classtut.pod:
Comment about new method to not confuse new users

Due to first example contains custom `new` it may give impression that writing `new` is something that programmer want to do. Maybe it would be good to mark it as non standard behavior and say that probably want to write build method?
azawawi cool
El_Che auth
c
08:13 g5 left 08:16 mr-fooba_ joined, mr-foobar left 08:20 mr-fooba_ left 08:21 mr-foobar joined 08:27 ely-se joined, pdcawley joined 08:33 FROGGS left
dalek p: e301575 | (Pawel Murias)++ | src/vm/js/bin/run_tests.pl:
[js] Run passing test 29 on run tests.
 08:45
p: fc91b03 | (Pawel Murias)++ | src/vm/js/nqp-runtime/bootstrap.js:
[js] Fix bug.
p: 4467abe | (Pawel Murias)++ | src/vm/js/QAST/Compiler.nqp:
[js] Fix bug.
08:46 pmurias joined
p: 11d9d69 | (Pawel Murias)++ | src/vm/js/ (2 files):
[js] We now pass test 43. Update make js-test.
08:49 virtualsue joined 08:56 Alina-malina joined
RabidGravy just has a cold call on my moobile of someone trying to sell me design and marketing services for my domain jeezus-knievel.com 09:02
he struggled to stay on his script shall we say
09:10 mr-foobar left, ely-se left
azawawi github.com/azawawi/scripts/blob/master/unzip.p6 # Proof of concept using Compress::Zlib worked on one file member :) 09:11
azawawi starts working on File::Zip
RabidGravy azawawi++ # go! go! go!
El_Che azawawi: what's the scope? 09:13
azawawi File::Zip.unzip(file-name => 'test.zip'); File::Zip.unzip(string => 'some zipped file content'); File::Zip.files
El_Che: deflate for now and uncompressed methods
El_Che: any ideas? 09:14
El_Che looking at compress zlib now
native call to zlib?
azawawi documentation should be better for uncompress... spent like 1 hour debugging it lol
that module is great but needs more examples 09:15
El_Che so by now, perl6 should have more libs than golang :) 09:16
RabidGravy At some point I want to add deflate/compress transfer encoding support to HTTP::UserAgent at some point so I'm glad someone is picking the bits out of this stuff first ;-) 09:18
El_Che, 419 modules as of this moment 09:19
09:19 dayangkun joined, ely-se joined
El_Che RabidGravy: impressive 09:19
RabidGravy I imagine at any given moment that some small percentage of them don't actually work but hey ;-) 09:20
azawawi RabidGravy: more tests => more working modules 09:21
09:28 rarara_ joined
rarara_ Hello, is it possible to expand the automatic "Usage:" output of "sub MAIN" with extra informations? 09:29
09:29 Psyche^ joined
jnthn morning, #perl6 09:30
dalek osystem: 56dd4f1 | azawawi++ | META.list:
Add File::Zip

Perl 6 interface to the Zip File format		 09:33
09:36 spider-mario joined 09:37 ely-se left 09:40 chenryn left 09:41 chenryn joined, espadrine joined, ely-se joined, kjs_ joined
moritz rarara_: I think if you add pod comments to your MAIN sub, they appear in the usage message 09:42
rarara_: example: github.com/tadzik/panda/blob/maste.../panda#L13
09:44 dayangkun left 09:45 dayangkun joined 09:46 dayangkun left
rarara_ moritz this is a great idea; would be cool to have it as flexible as possible 09:47
jnthn rarara_: Well, for ultimate flexibility you get to write your own USAGE sub, iirc :) 09:48
09:48 dayangkun joined
moritz rarara_: on the one hand, yes. On the other hand, there are limits to what a signature-based command line parser can do 09:48
rarara_ Yes, right: this is a compromize between fast coding and ultimate flexibility 09:49
dalek ast: cb2965a | jnthn++ | S06-multi/positional-vs-named.t:
Test codifying multi + optional named semantics.

Covers the ruling on RT #119929.		 09:52
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=119929
RabidGravy azawawi, re tests, yep :) That's why the Travis CI thing is good 09:53
every once in a while I select a module that appears to be failing and try to fix it
09:54 dayangkun left 09:55 dayangkun joined
jnthn xmas-- :) 09:55
jnthn at last slept decently and should probably seize the opportunity to work on one of the less easy xmas RTs today... :) 09:56
jnthn picks RT #125927 09:59
09:59 Actualeyes joined 10:02 dayangkun left 10:03 dayangkun joined
rarara_ Is there a way to get the number of consumed elements of a Seq? 10:04
10:04 FROGGS joined, kjs_ is now known as kjs
rarara_ I kind of feels like ".elems" should work this way 10:04
otherwisi it is just plain useless
jnthn .elems just counts the number of things that'd be produced 10:05
It's useful for things like @foo.split('bar').elems for example 10:06
10:06 kjs left
rarara_ jhthn I suspect that what [email@hidden.address] is doing is actually produce the things and then count them, isn't? 10:07
jnthn rarara_: I don't know; the iterator API is actually powerful enough to let us do better 10:08
Whether we are or not, I don't know without looking at the code :)
rarara_ no worry
jnthn Feels like an opt lizmat++ mighta already got to, though :)
.grep(...).elems is another quite common one 10:09
rarara_ I feel like after doing .grep, you might want to know how many iteration you took 10:10
even just for debugging purposes
let you want to know how % of things get grepped: how can you do it now?
10:12 Begi1115 joined
rarara_ gist.github.com/anonymous/a8c2b8324da798a4cef7 10:13
what is wrong in this?
thanks
10:20 FROGGS left
jnthn rarara_: Golfing it... 10:20
rarara_: I don't think anything is wrong in your code; I think start is at least somewhat to blame 10:21
10:21 tm joined
rarara_ thank you, a workaround is to add eager after astart 10:21
jnthn m: say await start "d\te\tf".chomp.split("\t");
camelia rakudo-moar 1fd9e8: OUTPUT«This Seq has already been iterated, and its values consumed␤ in block <unit> at /tmp/qaPKYuiRCa:1␤␤»
jnthn It golfs to that.
m: say (start "d\te\tf".chomp.split("\t")).result; 10:22
camelia rakudo-moar 1fd9e8: OUTPUT«This Seq has already been iterated, and its values consumed␤ in block <unit> at /tmp/YjlL4kgdyH:1␤␤»
jnthn And further to that
m: say (start (1,2,3).Seq).result;
camelia rakudo-moar 1fd9e8: OUTPUT«This Seq has already been iterated, and its values consumed␤ in block <unit> at /tmp/S30nON1Gzg:1␤␤»
jnthn And even that.
10:22 FROGGS joined
jnthn RT'able, anyways 10:22
rarara_ ok, I can write the mail 10:23
jnthn builds a possible fix 10:24
10:24 dakkar joined
jnthn Yeah, seems to do it 10:25
Oh, seems it maybe already was RT'd 10:26
RT #122715
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=122715
10:27 dayangkun left, Ch0c0late left
jnthn Though the tests are off 10:27
rarara_ with eager it works now
jnthn Yeah. With commit I'll make soon your original woulda been fine too 10:28
rarara_ ah ok, it is the same
jnthn I'm going to add the golf of yours as an additional test case.
10:28 chenryn left
rarara_ sure! 10:28
jnthn Did you file antoher RT? 10:29
10:29 Actualeyes left
jnthn Or did I catch the dupe in time? :) 10:29
rarara_ nono, I was just about to copy your golf 10:30
:)
jnthn OK, don't worry about filing it. :) 10:31
It's another case of that one I linked above. 10:32
azawawi re doc.perl6.org/routine/seek, is there constants that we can import from somewhere? 10:34
constants for $whence parameter that is
jnthn m: my @outer = 0..3; say await start { @outer.map: *+1 } 10:36
camelia rakudo-moar 1fd9e8: OUTPUT«This Seq has already been iterated, and its values consumed␤ in block <unit> at /tmp/of9ygj1KeG:1␤␤»
rarara_ Maybe eager should be added automagically by start in that case 10:38
but it would not work for infinite lists
jnthn rarara_: Well, ^^ works locally, so no eager needed :) 10:40
10:41 rindolf joined 10:42 ely-se left
dalek kudo/nom: 3c17701 | jnthn++ | src/core/Promise.pm:
Keeping a Promise must not sink the kept value.

Otherwise, a Seq handed back from a Promise will end up being consumed at the point of keeping the Promise, and thus the result will be useless.		 10:43
10:44 CIAvash joined
dalek ast: 98fd1d1 | jnthn++ | S17-promise/start.t:
Fix and unfudge RT #122715 tests; add one more.
 10:44
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=122715
10:47 tokuhirom left
jnthn The RT has a small write-up on the exact semantics also :) 10:47
RabidGravy azawawi, no but there is a comment in the code to make an enum, I was going to do it at one point but can't remember what stopped me
jnthn rarara_++
10:48 Actualeyes joined 10:49 brrt joined
jnthn gets back to his graphemes :) 10:49
10:51 tokuhirom joined
moritz "the grapheme and me" 10:52
10:56 zakharyas joined
azawawi is there a way in Perl 6 to unpack("L L L") into an object in one line? :) 10:57
10:57 weihan_ joined 10:58 chenryn joined 10:59 Oatmeal left
dalek kudo/bool_enum: 660d0e1 | (Stefan Seifert)++ | src/ (3 files):
Turn Bool into a proper enum
 11:01
kudo/bool_enum: 313ed64 | (Stefan Seifert)++ | src/core/operators.pm:
Fix SEQUENCE not expecting Bool ~~ Real to be true
kudo/bool_enum: 6c57f31 | (Stefan Seifert)++ | src/Perl6/Metamodel/EnumHOW.nqp:
Make Bool augmentable
kudo/bool_enum: 080f3ce | (Stefan Seifert)++ | src/Perl6/Metamodel/EnumHOW.nqp:
Fix type check for binding parameters

EnumHOW just needed an is_composed method like many other HOWs. Fixes regression in t/spec/S06-signature/optional.t
kudo/bool_enum: ff92838 | (Stefan Seifert)++ | src/Perl6/Metamodel/ (2 files):
Allow for mixing into enums.

Fixes Bool but role {} and t/spec/S14-roles/mixin.t regressions
nine jnthn: oops, seems like I just pushed an outdated bool_enum branch. Will have to correct that in the evening, when I'm on my desktop at home again. 11:02
RabidGravy azawawi, you mean something like : class M { has Int $!one; has Int $!two; has $!three }; my M $m = M.unpack-from-buf($buf); ?
nine jnthn: just in case you planned to have a look at it 11:03
RabidGravy then no, not currently but I am working on something like that on the back burner
can't remember how far I got
11:04 Oatmeal joined
azawawi Cool thanks 11:04
pink_mist suggestes "s/n/ /g" there, now all your code is in one line :P
*suggests
RabidGravy basically annotate the attributes with a "picture trait" which contains the unpack pattern and then provide a constructor which will pick the bits out of the Bug 11:05
Buf
need it for a couple of things I'm slowly working on
let me look to see how far I got 11:06
11:10 azawawi left
RabidGravy of course it would help if I could remember what I had called it 11:12
itz_stmuk is a rakudo star intended for niceville? 11:15
rarara_ m: say do for $*IN.lines -> $line {$line.say} 11:16
camelia rakudo-moar 1fd9e8: OUTPUT«Céad slán ag sléibhte maorga Chontae Dhún na nGall␤Agus dhá chéad slán ag an Eireagal ard ina stua os cionn caor is coll;␤Nuair a ghluais mise thart le Loch Dhún Lúich’ go ciúin sa ghleann ina luí␤I mo dhiaidh bhí gleanntáin ghlas’ G…»
rarara_ m: map $*IN.lines: -> $line {$line.say}
camelia rakudo-moar 1fd9e8: OUTPUT«Cannot call Numeric(Block: ); none of these signatures match:␤ (Mu:U \v: *%_)␤ in block <unit> at /tmp/m2G_AhFh_N:1␤␤»
rarara_ makes sense
itz_stmuk likes how the rakudo.org/rt/easy link displays no tickets 11:17
rarara_ but I am wondering what is the use of map? Autoparalleization?
RabidGravy that case probably not a good use: 11:21
m: $*IN.lines.map( {"> $_".say })
camelia rakudo-moar 1fd9e8: OUTPUT«> Céad slán ag sléibhte maorga Chontae Dhún na nGall␤> Agus dhá chéad slán ag an Eireagal ard ina stua os cionn caor is coll;␤> Nuair a ghluais mise thart le Loch Dhún Lúich’ go ciúin sa ghleann ina luí␤> I mo dhiaidh bhí gleanntáin gh…»
RabidGravy obviously, say there is stupid as there are better ways of doing it 11:23
11:23 rurban left
jnthn rarara_: map is useful when you want to turn one list into another by applying some mapping 11:26
m: $*IN.lines.map(*.uc).say
camelia rakudo-moar 3c1770: OUTPUT«(CÉAD SLÁN AG SLÉIBHTE MAORGA CHONTAE DHÚN NA NGALL AGUS DHÁ CHÉAD SLÁN AG AN EIREAGAL ARD INA STUA OS CIONN CAOR IS COLL; NUAIR A GHLUAIS MISE THART LE LOCH DHÚN LÚICH’ GO CIÚIN SA GHLEANN INA LUÍ I MO DHIAIDH BHÍ GLEANNTÁIN GHLAS’ GHAOTH…»
RabidGravy aye 11:27
.tell azawawi I may have temporarily mislaid that code, I'll find it eventually 11:28
yoleaux RabidGravy: I'll pass your message to azawawi.
mrf jnthn: is there an easy way of doing a hyper map on a list? 1..10.>>map(#some expensive iterator)?
jnthn Is .hyper.map(...) not easy enough? :) 11:29
mrf ah. I wasn't aware I could do .hyper on things. Though I needed to do >>
jnthn Note that .hyper and .race need to come before the operations that will be affected by them 11:30
11:30 andreoss joined
jnthn It actually gives you back a HyperSeq, which then can dispatch to parallel implementations of things 11:30
It's also a little under-implemented (as in, there's loose ends to tie up)
mrf ok 11:31
jnthn It is clever enough to realize that .hyper.map(...).grep(...).map(...) should batch things up and do the map/grep/map sequence on one thread for a given batch. 11:32
(For good CPU cache properties)
mrf what is the syntaxt for calling that on a seq.
jnthn .hyper is defined on any Iterable
(And Seq is Iterable)
mrf doing 1..10.hyper.map doesn't seem to be valid?
ilmari m: (1..10).hyper.map(*.say) 11:33
camelia rakudo-moar 3c1770: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
11:33 brrt left
mrf but then again I am still having trouble with the varous iterable stuff and flattening 11:33
jnthn Precedence :)
ilmari m: (1..10).race.map(*.say)
camelia rakudo-moar 3c1770: OUTPUT«1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤»
jnthn Ah, ilmari was faster :)
10 items will likely go into the same batch :)
mrf Ahh I have to do *.say in the map not .say
jnthn Yeah but...doing side-effects in a map block that you're using with .race/.hyper is asking for trouble :) 11:34
m: (1..10).race(batch => 2).map(*.say)
camelia rakudo-moar 3c1770: OUTPUT«1␤2␤3␤5␤4␤6␤7␤8␤9␤10␤»
jnthn m: (1..10).race(batch => 2).map(*.say)
camelia rakudo-moar 3c1770: OUTPUT«3␤4␤5␤6␤7␤8␤1␤2␤9␤10␤»
jnthn :) 11:35
mrf Yeah. I was exploring the example that itz_stmuk gave regarding loop speed differences in p6 vs p5 and was curious if more difference was seen using new options in p6 than were potentially lost over speed issues in p6 11:36
it was a simple enough example I though I could hyper map it rather than for loop. It turned out for me to be non trivial. 11:37
jnthn: ilmari: thank you both 11:38
11:46 ely-se joined 11:48 kjs_ joined
dalek c: 82a8e05 | (Steve Mynott)++ | doc/Language/5to6-perlvar.pod:
update version string
 11:52
c: c95dba0 | (Steve Mynott)++ | doc/ (3 files):
correct minor typos
11:53 kjs_ left 11:55 kjs_ joined 12:00 xpen left 12:01 xpen joined, azawawi joined
azawawi hi 12:01
yoleaux 11:28Z <RabidGravy> azawawi: I may have temporarily mislaid that code, I'll find it eventually
12:01 xpen left
azawawi github.com/azawawi/perl6-file-zip/...es/test.p6 # listing files in a ZIP files now works :) 12:03
s/ZIP file/ZIP archive/
jnthn azawawi++ 12:05
jnthn is enjoying watching the ecosystem grow :)
12:05 rurban joined
azawawi jnthn: :) 12:07
12:10 lucasb_ joined
RabidGravy is .set_rw supposed to have an effect on the behavour of a Code or is it just a mutator on some attribute that is used elsewhere (or is there a way I can make a non-rw Code object rw on the fly?) 12:12
not a biggie as I can deal with it otherwise, just yuck code 12:13
12:14 chenryn left
RabidGravy i.e. 12:15
m: my $A; my $s = sub () { $A }; $s.set_rw; say $s.rw; $s() = 1
camelia rakudo-moar 3c1770: OUTPUT«True␤Cannot modify an immutable Any␤ in block <unit> at /tmp/7XjQgdvj6a:1␤␤»
dalek ast: 2080809 | jnthn++ | S15-nfg/grapheme-break-test-gen.p6:
Add script to generate grapheme break tests.

Translates GraphemeBreakTest.txt from the Unicode Character Database into Perl 6 tests (for now, just for the .char counts).		 12:18
ast: bd7c065 | jnthn++ | S15-nfg/grapheme-break.t:
Add generated grapheme break tests.

Rakudo on Moar currently fails 83 tests of 402; these capture the cases where the current NFG algorithm is insufficient.
jnthn RabidGravy: That's far too late
RabidGravy: It can only affect code-gen if you do it at BEGIN time and before the block is "composed"
RabidGravy: So doing it in a block trait is probably fine, for example.
RabidGravy ah, so in a "method trait" it should work? Let's see 12:20
azawawi hmmm unzip for all archive is now working... benchmarking :) 12:21
real0m21.838s user for Perl 6, 0m0.049s for unzip :) 12:24
RabidGravy It's always good to have room for improvement 12:25
;-)
azawawi 35 folders/files archive
dalek ecs: e1a1afe | (Steve Mynott)++ | S06-routines.pod:
minor typo fix
azawawi but it is cool to see how much more readable code is in Perl 6
RabidGravy if nothing else the lack of boiler plate around getting and checking arguments is a massive win 12:26
azawawi now that i was able to unzip the firefox webdriver extension archive, i can dump the need of the external unzip command in the firefox webdriver :) 12:31
12:32 darutoko left
timotimo azawawi: is "unzip" the one implemented in C, or a pure perl5 implementation? 12:32
RabidGravy jnthn, ah what was confusing me was that "nextsame" doesn't have the rw-ey-ness but "callsame" does
azawawi timotimo: the one with debian that is in c
timotimo OK
that's probably not easy to beat :) 12:33
especially as our nativecall implementation is yet to get a little rewrite for jittability 12:34
RabidGravy yeah, you could probably improve by making a C wrapper that does as much as possible in C code, but the marshalling of the bastarding great arrays is still going to be costly
timotimo we can directly access C arrays from perl6 12:35
azawawi Sometimes being able to finish your task without installing external dependencies is more important than speed
one less dependency to worry about :)
and makes your code more portable
RabidGravy timotimo, sure but if you want to do something with them later you have to copy them
timotimo hm?
if it's an array of ints, for example, you can just use their values as native ints and end up only copying registers around 12:36
RabidGravy yeah, but they CArray don't have an iterator, .elems or anything else a Array doth 12:37
12:37 CIAvash left
RabidGravy or has something changed since I last looked 12:38
timotimo probably not
there's still not a great way to differentiate a C array that we want to be able to grow with realloc vs one we've been passed from C space where we can set its size to anything we want
12:39 brrt joined
RabidGravy sure, bearing in mind that in the Audio::* modules it's all about the bastarding great CArrays 12:40
timotimo mhm 12:41
12:42 sufrostico joined
RabidGravy :) 12:42
timotimo you might be in the perfect position to suggest better API around this :)
RabidGravy I've been thinking about "special constructors" for Array and Blob which do stuff at the lowest possible level do reduce the cost of the Perl level copying of elements 12:44
moritz fwiw I'd love a kind of wrapper around a CArray that you can tell its length, and then use much more naturally in Perl 6
but when you pass it to a native sub, it unwraps automatically
RabidGravy even something like "Array[Int].new(CArray $a, Int $elems)" would be a win TBH 12:47
12:48 telex left
azawawi worked like a charm for Selenium::WebDriver::Firefox... File::Zip rocks :) 12:49
12:50 telex joined
RabidGravy azawawi++ # nice one 12:50
12:51 ely-se left
azawawi :) 12:51
btw im using the atom editor for all this work... one cool thing i noticed in it is that you could open a file which inside an archive.... but you cannot modify it unfortunately... 12:53
12:54 ely-se joined
timotimo azawawi: i'd suggest ""signature = " ~ $signature.fmt("%08x") over the printf version 12:56
isn't buf.unpack("A*") the same as buf.decode("ascii")? 12:57
m: Buf.new(22, 23, 24).decode("ascii").say
camelia rakudo-moar 3c1770: OUTPUT«^W␤»
timotimo m: Buf.new(0x22, 0x23, 0x24).decode("ascii").say
camelia rakudo-moar 3c1770: OUTPUT«"#$␤»
12:57 aborazmeh left
ShimmerFairy still thinks unpack and pack shouldn't exist in Perl 6 until we know how P6 should be doing binary parsing, fwiw 12:58
timotimo azawawi: how about a --profile of File::Zip unpacking something? 13:00
azawawi timotimo: sure but there is still debugging code... i will do it :) 13:01
timotimo hm, there's a bare .decode in EndOfCentralDirectoryHeader for the comment thing 13:02
13:02 xpen joined
timotimo might want to be .decode("ascii") perhaps? 13:02
13:04 ely-se left
RabidGravy ShimmerFairy, I'll leave the MIDI module to you then ;-p 13:05
ShimmerFairy ?
dalek p: ee00ba0 | (Pawel Murias)++ | src/vm/js/nqp-runtime/reprs.js:
[js] Remove some dead and/or commented out code.
 13:06
p: ca02098 | (Pawel Murias)++ | src/vm/js/nqp-runtime/reprs.js:
[js] Serialize and compose autoviv info for P6opaque.
13:06 xpen left
RabidGravy well out of .pack/.unpack and working software or nothing I'll take pack/unpack 13:07
ShimmerFairy I still don't follow. Are you suggesting that a MIDI thing would be ~utterly impossible~ without a Perl 5-ish pack and unpack? 13:08
RabidGravy and if some better idea for "binary parsing" comes along then they can be relegated as "low level" features but for now
I'm not saying it would be impossible, but for a large variety of packed binary formats one would have to invent something strangely like unpack 13:09
ShimmerFairy I'd rather have nothing than a sloppily thrown-together subset of Perl 5 features someone decided to support. I've never seen any evidence that the current impl. was part of a carefully considered "how do we do binary parsing?" decision 13:10
AFAICT, the current implementation is just a Perl 5 relic that shouldn't be there. It's like seeing someone define $^O somewhere in the rakudo code to me :P 13:12
RabidGravy I'm sure everyone will be delighted to receive your proposal for something better, but in the meantime I want to write software 13:13
ShimmerFairy m: say pack("A", "☃"); # not to mention that not supporting unicode for something that's purported to handle strings on a codepoint basis (so says the docs) is just unacceptable
camelia rakudo-moar 3c1770: OUTPUT«non-ASCII character '☃' while processing an 'A' template in pack␤ in block <unit> at /tmp/szFEvk2Y6a:1␤␤»
ShimmerFairy RabidGravy: for the record I did share what I had written so far about my thoughts on binary grammars some time ago, I guess it just didn't catch anyone's interest :) 13:14
13:14 kjs_ left
brrt ShimmerFairy: i do have some interest 13:14
ShimmerFairy brrt: gist.github.com/ShimmerFairy/72a25...9aab077158 (I stopped at the point where I would've written about the <?pos(42)> equivalent and other offset-related things) 13:15
13:18 ely-se joined 13:21 raiph joined
FROGGS ShimmerFairy: there is a decision about how pack/unpack should work btw 13:21
ShimmerFairy And I admit that some of the syntax choices I made in that gist are quite strange, but hopefully the underlying concepts still look reasonable :P
FROGGS: Really? I haven't seen it. (Certainly is missing from the synopses)
FROGGS ShimmerFairy: and that's not like how pack/unpack is implemented right now
brrt is reading 13:22
azawawi ShimmerFairy: unpack is at least working for me to implement actual software in Perl 6 :)
FROGGS ShimmerFairy: TimToady said that he wants it to be like we do it for NaticeCall... you declare a structure (like a CStruct), and unpack the data using that structure
RabidGravy FROGGS, fwiw that's *exactly* how I would like it to work 13:23
FROGGS that will work when we have sized arrays and stuff in attribute slots in CStructs or classes in general
ShimmerFairy FROGGS: I like that, though I hope it's not referred to in terms of CStructs exactly, of course :) 13:24
FROGGS dunno
RabidGravy in the meantime I was working on something that did similar but at a higher level which I appear to have lost
FROGGS but using CStructs might be the first usable implementation
ShimmerFairy (precisely, I hope it's not actually _called_ "CStruct" when I'm not working in Nativecall-land)
FROGGS: my real concern, fwiw, is P6 being forced to support the ad-hoc P5-era notation currently in place if it's still there by Christmas. ("Yes, you _could_ use some of the P5 stuff, but you really should instead-- oh, they're gone already" :P ) 13:26
FROGGS yeah 13:27
brrt the thing is that not every binary format maps neatly to structures 13:28
in fact, anything with variable-length fields doesn't neatly map
FROGGS brrt: aye
ShimmerFairy and it's the reason why I'd like to have pack/unpack ripped out now before Perl 6 is officially released and stability is suddenly a lot more important.
brrt and in that sense, a 'binary grammar' seems a viable idea
pmurias having something that is bidirectional would be great 13:29
brrt if we rip it out, and not have something comparable by christmas, is that a large problem?
RabidGravy yes
ShimmerFairy that's not to say that the existing pack/unpack can't be launched right into a module, it just shouldn't be in core :)
brrt what about a moon on a stick pmurias :-P
RabidGravy I have three modules that use them already
brrt (grammars aren't bidirectional). 13:30
ShimmerFairy pmurias: I think that's where having some kind of support for "serialization" comes in handy :)
brrt moving pack/unpack to a module seems like a *very* sane idea to /me 13:31
makes everybody get what they want
diakopter El_Che: I guess you're kidding, since godoc.org indexes 96,000 packages for golang
RabidGravy looks at the code 13:32
I'd be relaxed if pack/unpack went in a module right away
the code does nothing fancy with nqp and seems eminently improvable 13:33
13:33 darutoko joined
RabidGravy perhaps as an augment on Blob for least surprise 13:34
pmurias brrt: I'm pretty sure having bidirectional binary serialization isn't something new
lucasb_ gist.github.com/smls/bc5d0fb42f199574e339 <-- another relevant gist about binary parsing 13:35
ShimmerFairy I think module-space would also let pack/unpack avoid the spectre of "but should this pack directive from P5 be supported in core?", because it's a module and can do what it want :)
RabidGravy gwan admit it, you'd still complain about it
;-) 13:36
FROGGS ShimmerFairy: +1
diakopter gwan?
RabidGravy "go on"
in sarf lahnun
ShimmerFairy RabidGravy: nah, if it's a module then I can not use it all I want :P
diakopter that was a good show
RabidGravy :) 13:37
ShimmerFairy My current suspicion is that binary stuff wants binary grammars for parsing the more complex binary data types, Nativecall-like structures for the simpler stuff, and maybe even those said structures being a part of the actions side of binary grammars. 13:38
flussence one gripe I have with pack is that it embeds a second-class language (along with sprintf)... and it's like that outside perl too. I lack any ideas to make it better though :( 13:39
azawawi timotimo: done ... thanks for the code review :)
ShimmerFairy lucasb_: one thing to note on that gist is that the :bytes adverb in regexes comes from an older era of thinking on strings, and a binary regex would (in my mind) be separate from textual regexes in the first place :) 13:40
rarara_ m: my $lista = do for ("a~b","c~d").Seq -> $l { $l.split("~"); }; for |$lista -> $el ($first, $second) { say $first }
camelia rakudo-moar 3c1770: OUTPUT«Too few positionals passed; expected 2 arguments but got 0 in sub-signature of parameter $el␤ in block <unit> at /tmp/tLohMiQGgu:1␤␤»
flussence
.oO( I guess one place that handles binary data in the language well is... C )
rarara_ m: my $lista = do for ("a~b","c~d").Seq -> $l { eager $l.split("~"); }; for |$lista -> $el ($first, $second) { say $first }
camelia rakudo-moar 3c1770: OUTPUT«a␤c␤»
ShimmerFairy flussence: the one thing about sprintf is that it's the best at i18n (when it supports explicit arg indices). I'm not sure how you could accomplish it more easily with, say, C++'s streams 13:41
timotimo azawawi: you're welcome :) 13:42
rarara_ So Seq doesn't destructure
PerlJam good * #perl6
timotimo azawawi: also, you seem to be starting to search for a 4 byte long string at the last byte of the file
maybe you want to start earlier 13:43
rarara_ wouldn't be nice that it would?
azawawi timotimo: that's the eocd.. it is at the end lol
RabidGravy I do have small desire to expose the sprintf grammar as $~Sprintf or so for some nefariety
azawawi timotimo: but maybe i need to quit earlier for non-zip files
timotimo azawawi: well, you're not going to find a 4-byte-long string at file-length - 1
azawawi timotimo: yup :) 13:44
timotimo that's what i mean
azawawi is looking at the profile
flussence ShimmerFairy: agreed there. I think ICU has something fancier... but then you'd be using ICU. *shudder*
El_Che diakopter: I was tongue in cheek paraphrasing Ovid (I thought he was in the channel): twitter.com/OvidPerl/status/658915941112270848
ShimmerFairy flussence: P6 has %1$s and such, if I remember the syntax :) . Why don't you like ICU, ooc? I don't want to use it because, even though I can use an iterator to do operate on graphemes NFG-style, I don't like that I have to use an iterator. I'd much rather write my own C++ unicode library :P 13:46
13:46 skids joined
azawawi timotimo: raw.githubusercontent.com/azawawi/...44693.html 13:48
jnthn 'cus writing Unicode libraries is fun! :P
flussence ICU's like using a bucket-wheel excavator to dig a garden bed :)
RabidGravy is there any method trait out there that is known to cause other applied traits problems? 13:49
timotimo azawawi: holy F'ing what? 89% time spent in infix:<~> ?!?
ShimmerFairy jnthn: I may or may not be working on a small collection of P6 files to extract info from the UCD, so I may or may not be able to agree with that sentiment. :P
brrt pmurias: it isn't new, no. but it is quite different from parsing arbitrary binary formats :-) 13:50
ShimmerFairy flussence: from what little I saw of ICU when trying it out, it's set up more for someone who always wants to get down to the finest points of Unicode. I much prefer a more P6 style where things Just Work™ by default :)
zengargoyle how do people test a modules un-exported functions? 13:51
jnthn ShimmerFairy: I'm working on fixing the NFG stuff Right Now :)
ShimmerFairy: It's not too crazy, but... :)
13:51 n0tjack joined
ShimmerFairy (e.g. have a string type for NFG level, instead of having to setup a long-named iterator to get at graphemes) 13:51
timotimo azawawi: could you send over the output of MVM_SPESH_LOG=unzip_speshlog.txt ? (might want to compress that one before sending) 13:52
AFK for a little bit
RabidGravy zengargoyle, assuming they aren't "our" subs, by testing the code that uses them
if they aren't used then what are they doing there ;-)
13:53 _mg_ joined
ShimmerFairy jnthn: heh, it's kinda funny that Perl 6 tends to have things that I want when using C++ :P . I'd like to write a P6 regex engine someday (instead of using inferior syntaxes), and of course that depends on a nice P6-inspired unicode library :) 13:53
zengargoyle well, they're used internally and there may be many of them and they may be indidually testable but have no need to be exported to the end user.
RabidGravy so yeah, it's test the things that use them *or* inject methods via a role for instance that can exercise them directly 13:54
ShimmerFairy zengargoyle: have you tried perhaps accessing them as Module_Name::sub_name(), if it's an "our" sub? 13:55
RabidGravy zengargoyle, the latter has the advantage of being able to test behaviour that might not occur in the way they are being used in your methods 13:56
zengargoyle well, what i settled on for the momemnt is doing 'is export(:FOR-TESTING)' like and using that in tests.
RabidGravy normal methods
which works too, but special code to enable testing always feels icky
zengargoyle but it seems smelly to export something only so you can actually test it 13:57
13:57 tokuhirom left
RabidGravy yeah 13:57
zengargoyle yeah...
RabidGravy the role idea would work
so e.g.
azawawi timotimo: it is 7 seconds without inflate(...) and 21 seconds with inflate
zengargoyle and yeah, can mess with using 'our' but that seems a bit wonky as well.
flussence zengargoyle: could extend the module with another one that exports the subs...
ShimmerFairy That's actually how I'd do it, probably. "All is fair if you predeclare [wanting access to functions normally only needed in testing]" and all that :P 13:58
RabidGravy m: class F { sub wazz() { say "foo" };}; role Testy { method test-wazz() { wazz() }}; my $f = F.new; $f does Testy; $f.test-wazz 13:59
camelia rakudo-moar 3c1770: OUTPUT«5===SORRY!5=== Error while compiling /tmp/nYnVXvKy1c␤Undeclared routine:␤ wazz used at line 1␤␤»
zengargoyle didn't occur to me that a module could be extended like a class.
azawawi timotimo: i bet it is github.com/retupmoca/P6-Compress-Z...ib.pm6#L95 :)
RabidGravy okay that's a shit idea
zengargoyle still a bit fuzzy on module vs class vs old perl 5 package and namespaces and such 14:00
azawawi home & 14:01
n0tjack zengargoyle: the solution is just to write a big long unmodular script with a bunch of functions
14:01 rurban left
zengargoyle heh, other thought was a sub MAIN('test') and calling that as a subtest from a .t file. :) 14:01
14:02 cygx joined
cygx o/ 14:02
m: my @a; say @a[^3] = 1..*; .say for @a 14:03
camelia rakudo-moar 3c1770: OUTPUT«()␤»
cygx m: my @a; say @a[flat ^3] = 1..*; .say for @a
camelia rakudo-moar 3c1770: OUTPUT«(1 2 3)␤1␤2␤3␤»
cygx ^- bug or feature?
ShimmerFairy zengargoyle: another option is to have an exported sub test-internal that runs tests on the internal subs. Still a tiny bit leaky, but at least you're just exposing a test function :)
diakopter anyone have a link to a recent perl6-bench output run?
RabidGravy m: module F { sub wazz() { say "foo" };}; module F { our sub test-wazz() { wazz() }}; F::test-wazz()
camelia rakudo-moar 3c1770: OUTPUT«5===SORRY!5=== Error while compiling /tmp/V4B5zfSOrq␤Redeclaration of symbol F␤at /tmp/V4B5zfSOrq:1␤------> 3{ sub wazz() { say "foo" };}; module F7⏏5 { our sub test-wazz() { wazz() }}; F:␤ expecting any of:␤ generic role␤»
RabidGravy I'll hit the right thing at some point 14:04
14:05 azawawi left
pmurias brrt: it's different from parsing arbitrary ones, most of the binary formats are intended to be machine readable so they don't contain as much crazy stuff as the text things grammars parse 14:06
brrt yeah, i know. i'm just... sceptical it should be a primary goal 14:07
pmurias brrt: for the crazy bits it should be possible to provide both a hand written serializer/deserializer
14:08 xpen joined
brrt uhuh 14:08
14:08 softmoth joined, kjs_ joined
jnthn zengargoyle: I don't know the answer, but it is on the xmas list of thigns to review: rt.perl.org/Ticket/Display.html?id=124316 14:09
zengargoyle i guess i'd like a way to violate scoping during the 'create the module' phase, but keep things nice and locked away once it's compiled and stuffed into a repo somewhere.
n0tjack m: say $4; # why is this valid syntax, what does it mean? 14:10
camelia rakudo-moar 3c1770: OUTPUT«Nil␤»
pmurias brrt: for things like the moar variable length integer encoding I would find it acceptable to have to write imperative serialize/deserialize methods 14:11
14:11 ely-se left
zengargoyle doesn't grok jnthn's link. 14:11
n0tjack what does $4 mean? 14:12
jnthn zengargoyle: oh, sorry!
n0tjack m: say $(4+5);
camelia rakudo-moar 3c1770: OUTPUT«9␤»
jnthn cygx: I mean rt.perl.org/Ticket/Display.html?id=124316 for the answer to your question :)
zengargoyle :)
jnthn should learn to read :)
RabidGravy reading's over-rated 14:13
n0tjack m: say $.WHAT;
camelia rakudo-moar 3c1770: OUTPUT«5===SORRY!5=== Error while compiling /tmp/DGeOnvIUsZ␤Variable $.WHAT used where no 'self' is available␤at /tmp/DGeOnvIUsZ:1␤------> 3say $.WHAT7⏏5;␤ expecting any of:␤ argument list␤ term␤»
flussence n0tjack: «perl6 --target=ast -e '$4'» might explain 14:14
14:14 _mg_ left, xpen left
n0tjack flussence: that's a neat switch I didn't know about. but I'm not familiar enough with P6's grammar to make sense of that QAST 14:15
this seems relevant QAST::Var(local __args__ :decl(param))
ShimmerFairy m: "abcde" ~~ /(a) (b) (c) (d) (e)/; say $4
camelia rakudo-moar 3c1770: OUTPUT«｢e｣␤»
ShimmerFairy n0tjack: it's just a shorthand for accessing numbered captures in a regex match :)
n0tjack oh, it's sugar for $/[4]
it's the normal (in a p5 sense) $4 14:16
cygx jnthn: so it's on The List of Things to Fix Before Christmas
good
n0tjack ShimmerFairy: thanks
ShimmerFairy n0tjack: except zero-indexed, where (IIRC) P5 is 1-indexed
flussence that's the "&postcircumfix:<[ ]>, lexical $/, Int" part at the bottom
14:16 virtualsue left, andreoss left
n0tjack flussence: thanks, learning to read the AST will def. help me come up to speed with p6 14:17
flussence it doesn't always make sense but it's useful when it does
RabidGravy ShimmerFairy, yeah 'cause $0 is something else in P5 (which it got from the unix shell)
n0tjack and what is $ in $(4+5) ? 14:18
brrt pmurias: yes... ish. some of these varint things are really standard
flussence n0tjack: item context
ShimmerFairy itemization, e.g. $@list or $(Foo.I-return-something-that's-not-guaranteed-an-item) 14:19
n0tjack thanks flussence, ShimmerFairy
brrt on the other hand, you can always fall back to just reading the bytes out of a Buf[uint8]
timotimo retupmoca: hey, would you be interested in optimizing Compress::Zlib a little bit? github.com/retupmoca/P6-Compress-Z...ib.pm6#L95 - these calls to infix:<~> are hella expensive 14:20
retupmoca: so maybe it should build a list of bufs and join them at the end ... if we even have a join candidate for bufs? 14:21
14:22 bjz_ left 14:23 ely-se joined
timotimo retupmoca: from the implementation of infix:<~>, it seems like it's the typical problem where the beginning of the string - the part that keeps growing - gets copied over and over again 14:24
i would be interested to know how many chunks inflate spits out for the files azawawi's inflating 14:25
in total, there's 3155 calls to infix:<~>, and inflate is called 27 times 14:26
m: say "so the average file has { 3155 / 27 } chunks"
camelia rakudo-moar 3c1770: OUTPUT«so the average file has 116.851852 chunks␤»
timotimo so the first ~1024 bytes get copied 115 times, the second 114 times, the third 113 times etc etc 14:27
RabidGravy nasty
retupmoca oh, that's not nice
timotimo :) 14:28
at least the bottleneck was easy to spot!
retupmoca would [~] on a list be better if Buf.join doesn't exist?
timotimo yes
no
it wouldn't, i'm afraid
jnthn wonders how ~ on Bufs works also
retupmoca eh, that's what I suspected
jnthn Like, does it loop and push, or does it use splice
timotimo no, it uses loop and bindpos/atpos
jnthn eek
timotimo two loops
jnthn nqp::splice would be way faster
diakopter <- oops? 14:29
jnthn Not ideal, but.. :)
n0tjack m: say ~1e6; 14:30
camelia rakudo-moar 3c1770: OUTPUT«1000000␤»
n0tjack m: say ~1e66; 14:31
camelia rakudo-moar 3c1770: OUTPUT«1e+66␤»
n0tjack wonder where the cutover is
oh, maybe at the integer boundary
m: say ~((2**64)-1);
camelia rakudo-moar 3c1770: OUTPUT«18446744073709551615␤»
n0tjack m: say ~((2**64));
camelia rakudo-moar 3c1770: OUTPUT«18446744073709551616␤»
timotimo subbuf also uses a loop and bindpos/atpos, so could also use splice instead
RabidGravy make all things faster! 14:32
timotimo now about six people in here know exactly what to do about this; who is going to actually write the patch? :) 14:33
RabidGravy m: say (^6).pick 14:34
camelia rakudo-moar 3c1770: OUTPUT«0␤»
timotimo whoaaa, what are the odds!
RabidGravy WE HAVE A WINNER! 14:35
zengargoyle is not a number
zengargoyle is a free man
timotimo ahahahahahahahahahaha
tadzik :D
*guitars* 14:36
timotimo *drums*
diakopter tadzik: XD
RabidGravy *mad thing that makes a sound like a leprechaun being tortured*
tadzik doesn't actually remember if guitars come in first
tadzik also doesn't know where this quote actually comes from and blindly attributes it to Iron Maiden 14:37
timotimo iron maiden got it from the TV Show "The Prisoner"
that's also the name of the song, IIrC
tadzik quite appropriate, yes
timotimo i only saw a little bit of The Prisoner. i did not quite understand it
zengargoyle who are you? the new number 2. who is number 1? you are number 6. i am not a number, im a free man. 14:38
gfldex RabidGravy: you may be interested in this merchandice www.grand-illusions.com/acatalog/No...Set_1.html
timotimo we want information. information. information!
RabidGravy I've got the whole bunch of them on DVD and I still don't completely understand it
tadzik m: say "we want" ~ "information, " xx 3
camelia rakudo-moar 3c1770: OUTPUT«we wantinformation, information, information, ␤»
tadzik yeah, drums come in first 14:39
tadzik hands over his nerd badge
itz_stmuk m: macro a { 'foo' }; say a
camelia rakudo-moar 3c1770: OUTPUT«===SORRY!===␤Too few positionals passed; expected 3 arguments but got 2␤»
retupmoca timotimo: github.com/retupmoca/P6-Compress-Z...a519fb5299
14:39 chenryn joined
RabidGravy gfldex, :) 14:39
itz_stmuk never watch the last episode of The Prisoner
tadzik it breaks the 4th wall and makes yourself a prisoner? 14:40
pmurias jnthn: sometimes when compiling larger files with nqp-js (cross-compiling), parts of the output (JavaScript) go missing
jnthn: how should I debug such an error?
itz_stmuk its like Matrix 2 and 3 .. only far far worse
timotimo retupmoca: is that a lot better?
14:40 sufrostico left
timotimo i don't really know how good buf.list is 14:41
retupmoca idk, I didn't benchmark it
I just killed the '~'
RabidGravy itz_stmuk, possiblly what you are looking for was
m: macro a { quasi { "foo" } }; say a
camelia rakudo-moar 3c1770: OUTPUT«foo␤»
timotimo itz_stmuk: which episode is the one with the trial? where the jumpy guy comes out and bounces all around, annoying everybody?
jnthn pmurias: "Go missing"? o.O
retupmoca but it's a VMArray, so I expect .list to be decent 14:42
jnthn pmurias: Well-formed bits, or just random bits?
14:42 rindolf left
itz_stmuk timotimo: I think there is a computer in it but I blanked most of it 14:42
I've not watched for ages
pmurias jnthn: Parts of code that should be joined by a nqp::join just disappear 14:43
itz_stmuk RabidGravy: I was testing the error message for that macro which isn't as doc'd 14:44
timotimo fair enough :)
retupmoca: it still passes tests, though?
RabidGravy ah right
timotimo seems like you only have ::Raw in travis-ci 14:45
14:46 Oatmeal left
retupmoca yeah, it passed tests on my local machine 14:47
and also it's in travis-ci now
timotimo good & good :)
Compress::Zlib now handles Big Data™ 14:48
jnthn pmurias: No idea...I'd probably try to produce an isolated test case 14:52
retupmoca I'm not sure who was running the profile for Compress::Zlib before, but I'd be interested to see how the new version compares
masak greetings, #perl6 14:54
m: sub m() { return -> { return "foo" } }; say m()()
camelia rakudo-moar 3c1770: OUTPUT«Attempt to return outside of any Routine␤ in block <unit> at /tmp/FGmtrjjjyW:1␤␤»
masak I claim there's no returning outside of any Routine going on here, so it's not a good error message.
thoughts?
I mean, those `return`s are clearly inside a Routine... 14:55
brrt rakudo doesn't think so
14:55 rindolf joined
gfldex both returns should return from m. So your m can return twice. 14:55
masak does Rakudo not believe in lexical returns?
brrt m: sub m() { -> { return "foo"; } }; say m().perl; 14:56
camelia rakudo-moar 3c1770: OUTPUT«-> { #`(Block|59615944) ... }␤»
timotimo retupmoca: it was azeawawi; he's currently commuting towards home IIUC
brrt m: sub m() { -> { return "foo" } }; say m()();
camelia rakudo-moar 3c1770: OUTPUT«Attempt to return outside of any Routine␤ in block <unit> at /tmp/RYbK44r_8s:1␤␤»
timotimo or does "home &" mean that his bus reached his destination?
brrt the returned thing is clearly a closure
masak yes, it is. I agree.
brrt but rakudo thinks it is a block
jnthn m: sub m() { return -> { return "foo" } }; sub n() { say m()() }; n()
camelia rakudo-moar 3c1770: OUTPUT«Attempt to return outside of any Routine␤ in block <unit> at /tmp/FeOOSQqU8b:1␤␤»
masak nor it's inside even more routines :P
now* 14:57
diakopter neither
jnthn That clearly shows it's lexical, otherwise it might accientally work
brrt routinings
jnthn "work" :)
m: return
camelia ( no output )
masak jnthn: ok, so it's just the error message that's confused.
jnthn Yeah
Feel free to patch the error message if you've a better one.
14:57 tokuhirom joined
masak m: macro m { quasi { return } }; sub foo { say "A"; m(); say "B" }; foo() 14:57
camelia rakudo-moar 3c1770: OUTPUT«A␤Attempt to return outside of any Routine␤ in any at /tmp/unLtwpEl6E:1␤ in sub foo at /tmp/unLtwpEl6E:1␤ in block <unit> at /tmp/unLtwpEl6E:1␤␤»
gfldex if you add any more features rakudo might become self-aware 14:58
masak grins, madly
14:58 n0tjack left
masak that's actually completely the expected *behavior* 14:58
dalek kudo/nom: 1f81099 | (Vladimir Lettiev)++ | tools/build/Makefile-Moar.in:
fixed linking order

With gnu linker option --as-needed (default in most linux distros) -lmoar discarded if object files that use it is not behind it
kudo/nom: f699831 | FROGGS++ | tools/build/Makefile-Moar.in:
Merge pull request #564 from vlet/fix

fixed linking order
masak but the error message needs work.
(007 gets that one wrong right now)
14:58 n0tjack joined
masak jnthn++ # was probably responsible for making returns so lexical :) 14:59
14:59 khw joined
diakopter m: sub m() { BEGIN return -> { -> { BEGIN return "foo" } } }; sub n() { say m()()() }; n() 14:59
camelia rakudo-moar 3c1770: OUTPUT«foo␤»
gfldex i believe to have found a nasty bug with rosettacode.org/wiki/Ordered_words#Perl_6 15:00
masak m: macro m { quasi { &COMPILING::return() } }; sub foo { say "A"; m(); say "B" }; foo()
camelia rakudo-moar 3c1770: OUTPUT«A␤Cannot find method 'Any'␤ in any at /tmp/8e4ts_lHpz:1␤ in sub foo at /tmp/8e4ts_lHpz:1␤ in block <unit> at /tmp/8e4ts_lHpz:1␤␤»
masak NYI, I guess
someone should tell the macro guy to hurry up
gfldex if i run that one locally it returnes 3 words where it should return 16
diakopter masak: ^ look I fixed it
RabidGravy when I first started actually writing Perl 6 I was fascinated by the idea of macros, but I haven't actually encountered something where I would sensibly use one
gfldex the perl5 version works as expectd
if i sullpy it with a shorter word list it starts to work
masak diakopter: you are truly helpful
lucasb_ wait, what? A plain 'return' outside subs not giving a warning is wrong, isn't it?
15:00 chenryn left
diakopter masak: I know 15:01
masak lucasb_: define "outside".
m: return
camelia ( no output )
gfldex the culprit seams to be 'grep { [le] .comb }' but only for long lists
diakopter what even does BEGIN return mean
masak lucasb_: I mean, that's clearly wrong, IMO.
gfldex would be nice if someone could confirm that
lucasb_ masak: ^^ ok, it's wrong to me too :)
masak diakopter: it means return, but really really early :P 15:02
diakopter TO WHERE
masak Kansas.
lucasb_ m: if 1 { say 'a'; return; say 'b' }; say 'c'
camelia rakudo-moar 3c1770: OUTPUT«a␤»
lucasb_ m: while 1 { say 'a'; return; say 'b' }; say 'c'
camelia rakudo-moar 3c1770: OUTPUT«a␤»
15:02 tokuhirom left
masak jnthn: I think "Routine already returned" or "Routine cannot return twice" is what I would expect to see. 15:03
jnthn: the "return outside of any Routine" message is fine when things are *lexically* outside of Routines.
diakopter m: say say say say -> { BEGIN return BEGIN return BEGIN return say return 333 }() 15:05
camelia rakudo-moar 3c1770: OUTPUT«333␤True␤True␤True␤»
15:05 brrt left
tadzik （　° ー°） 15:05
diakopter too few TRUEs
lucasb_ diakopter: the standard placeholder number is 2*333 :) 15:06
diakopter m: 2**333
camelia rakudo-moar 3c1770: OUTPUT«WARNINGS:␤Useless use of "**" in expression "2**333" in sink context (line 1)␤»
pink_mist 668 The Neighbour of the Beast
diakopter m: say 2**333 15:07
camelia rakudo-moar 3c1770: OUTPUT«17498005798264095394980017816940970922825355447145699491406164851279623993595007385788105416184430592␤»
diakopter m: say 2**2**333
camelia rakudo-moar 3c1770: OUTPUT«0␤»
masak 'night, #perl6
diakopter wait what
tadzik g'nite
15:07 CIAvash joined
diakopter how is that zero 15:08
p: say 2**2**333
rp: say 2**2**333
n: say 2**2**333 15:09
camelia niecza v24-109-g48a8de3: OUTPUT«sh: mono: command not found␤»
FROGGS diakopter: it overflows somewhere
timotimo gnite masak
diakopter panics
timotimo diakopter: we have a bug in there
FROGGS (it's already tickited)
timotimo is nobody else working on the splice patch for infix:<~>(Buf, Buf) and subbuf? otherwise, i'll start on it 15:10
jnthn timotimo: Not me :)
diakopter m: say 2**333333333
FROGGS me neither :o)
PerlJam anyone know right off where S/// is described in the specs?
camelia rakudo-moar 3c1770: OUTPUT«(timeout)»
lucasb_ PerlJam: has it ever got written down? I had to hunt the irclogs and changelogs to learn about it 15:11
PerlJam lucasb_: good question. I dunno
flussence there's only mention of an S metaop there... 15:12
FROGGS PerlJam: there you go :o) github.com/perl6/roast/commit/e82e...e05b5f05ae
lucasb_ just reminding that the S metaop is broken :)
diakopter it is?
FROGGS how so?
lucasb_ m: say so False S& False
camelia rakudo-moar 3c1770: OUTPUT«True␤»
PerlJam I'm interested in what the return value of "foo" ~~ S/a//; should be 15:13
FROGGS PerlJam: "foo"
PerlJam I would expect it to be "foo", but currently that's not the case
FROGGS O.o
m: say "foo" ~~ S/a//
camelia rakudo-moar 3c1770: OUTPUT«Potential difficulties:␤ Smartmatch with S/// can never succeed because the subsequent string match will fail␤ at /tmp/j5jD2MNZhi:1␤ ------> 3say "foo" ~~ S/a//7⏏5<EOL>␤False␤»
diakopter heh 15:14
FROGGS m: $_ = "foo"; say S/a//
camelia rakudo-moar 3c1770: OUTPUT«Nil␤»
FROGGS hmmmm
PerlJam well, since at least 2 of us agree on what it should be, I'll take the current behavior as a bug 15:15
15:15 pmurias_ joined
FROGGS IMO this: $_ = "foo"; say S/a// should translate more or less directly to: say "foo".subst(rx/a/, '') 15:16
flussence m: say S[a] = '' given 'foo' 15:17
camelia rakudo-moar 3c1770: OUTPUT«Nil␤»
flussence m: say s[a] = '' given 'foo'
camelia rakudo-moar 3c1770: OUTPUT«Nil␤»
flussence m: say 'foo' ~~ s/a//
camelia rakudo-moar 3c1770: OUTPUT«False␤»
FROGGS the smartmatch there returns the matched string 15:18
err, the match
flussence m: say $/ given ('foo' ~~ s/a//)
camelia rakudo-moar 3c1770: OUTPUT«Nil␤»
flussence oh I see
15:18 pmurias left
diakopter m: $_ = "foo"; S/o/f/; say $_; s/o/f/; say $_ 15:18
camelia rakudo-moar 3c1770: OUTPUT«foo␤ffo␤»
diakopter hrm 15:19
PerlJam m: $_ = "a b\nc\td"; say Ss/b c d/x y z/; 15:25
camelia rakudo-moar f69983: OUTPUT«a x␤y z␤»
PerlJam m: $_ = "a b\nc\td"; say SS/b c d/x y z/;
camelia rakudo-moar f69983: OUTPUT«5===SORRY!5=== Error while compiling /tmp/J6kGAuDlRq␤Missing required term after infix␤at /tmp/J6kGAuDlRq:1␤------> 3$_ = "a b\nc\td"; say SS/b c d/x y z/7⏏5;␤ expecting any of:␤ prefix␤ term␤»
15:26 rogu317 joined 15:28 sufrostico joined
timotimo oh, we can't make subbuf use splice - or at least not just a single splice - because splice would begin at the beginning of the "source" array 15:28
15:29 CIAvash left 15:31 pollei joined 15:33 rurban joined 15:34 rogu317 left 15:35 zakharyas left
dalek ast: 64f19c9 | (Steve Mynott)++ | S32-io/IO-Socket-INET.pl:
constants have worked for a while so use them
 15:35
pink_mist .pl not .p6? 15:36
15:37 cygx left 15:38 ely-se left
lucasb_ m: $_ = 'foo'; say S/a// 15:40
camelia rakudo-moar f69983: OUTPUT«Nil␤»
lucasb_ star: $_ = 'foo'; say S/a//
camelia star-m 2015.09: OUTPUT«foo␤»
lucasb_ ^^ worked in star? Is a regression?
in star, 'foo' ~~ S/// is a smartmatch too 15:41
FROGGS lucasb_: "worked" probably means that the behaviour was different 15:43
n0tjack m: $_ = 'whatevs'; say s[s]='z';
camelia rakudo-moar f69983: OUTPUT«｢s｣␤»
FROGGS lucasb_: TimToady decided that the old behaviour was not correct, and I agree
lucasb_ FROGGS: oh, ok. sorry for the noise 15:44
FROGGS np :o)
bbl &
15:44 FROGGS left 15:46 weihan_ left
n0tjack m: $_ = 'whatevs'; say S[s]='z'; 15:47
camelia rakudo-moar f69983: OUTPUT«whatevz␤»
n0tjack hee hee
15:47 mayuresh joined
n0tjack I honestly cannot wait for the first IOP6C 15:48
mayuresh hello :)
n0tjack heyo
mayuresh n0tjack, what the iop6c?
some kind-a conference for perl6?
n0tjack international obfuscated p6 contest
mayuresh whoa ... :)
n0tjack I mean look at that ^ ... that's /normal/ p6.
15:49 rurban left
gfldex m: say ++$; say ++$; say $+=2 for 1..10 15:50
camelia rakudo-moar f69983: OUTPUT«1␤1␤2␤4␤6␤8␤10␤12␤14␤16␤18␤20␤»
n0tjack m: say $+=2 for 1..3; 15:51
timotimo jnthn: using splice instead of the two loops makes time perl6 -e 'my $res = Buf.new; loop (my int $i = 0; $i < 10_000; $i++) { $res ~= Buf.new(0, 9, 9, 9, 9, 9, 9, 9, 9, 1) };' a little bit more than 2x faster
camelia rakudo-moar f69983: OUTPUT«2␤4␤6␤»
n0tjack why does that start at 2?
it should start at 3
assuming $+= ... is the same as $_ += ... is the same as $_ = $_ + ...
gfldex it starts at 0 before the first $ = $ + 2 15:52
n0tjack oh but it's not, it's ignoring $_
gfldex $ aint $_
n0tjack what is $ as a variable then?
m: say $; 15:53
camelia rakudo-moar f69983: OUTPUT«(Any)␤»
jnthn timotimo: Improvement :)
n0tjack m: $ += 99;
camelia ( no output )
n0tjack m: $ += 99 say $;
camelia rakudo-moar f69983: OUTPUT«5===SORRY!5=== Error while compiling /tmp/aj9Li73EYo␤Two terms in a row␤at /tmp/aj9Li73EYo:1␤------> 3$ += 997⏏5 say $;␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ statement modifier␤ …»
n0tjack m: $ += 99; say $;
camelia rakudo-moar f69983: OUTPUT«(Any)␤»
gfldex n0tjack: it's a (a!) anonymouse state variable 15:54
timotimo jnthn: do you think we'd want a bigger version of splice with an extra argument that'd let us splice a from another offset than 0 in the source array
gfldex n0tjack: see design.perl6.org/S02.html#Names
n0tjack gfldex: you realize I'm going to use these powers for evil, right? 15:55
gfldex n0tjack: i expect you to!
timotimo jnthn: actually, it seems to scale better
15:55 mayuresh left
jnthn timotimo: I've been pondering a general memcpy op 15:55
gfldex n0tjack: the devil is in the detail 'say $++; say $++' are actually two independent variables, that happen to have no name 15:56
timotimo at 20_000 it's 14s vs 41s, but at 10_000 it's only 3.9s vs 10.9s
jnthn: me, too
n0tjack gfldex: it's reset every statement?
no, then the accumulator thing you did wouldn't work...
timotimo i wonder if we have any tests about infix:<~> between Bufs of different int sizes
gfldex n0tjack: yes. I could not resist to add this to rosettacode.org/wiki/Variables#Perl_6
n0tjack m: say $++ for 1..3; for 1..3 { say $++; } 15:57
camelia rakudo-moar f69983: OUTPUT«0␤1␤2␤0␤1␤2␤»
15:57 jdv79 joined
dalek kudo/nom: a333147 | timotimo++ | src/core/Buf.pm:
infix:<~> on Buf can use nqp::splice for better speed
 15:57
itz_stmuk wonders if IO::Socket::INET::PIO::PF_INET etc. should be available to programmers 16:01
n0tjack gfldex: ha! before reading that, I hadn't realized subset notation worked on normal variable declarations too.
16:02 adu joined
gfldex rosettacode is full of wisdom :) 16:02
n0tjack m: my \naked = <n u d e>; say naked.WHAT; say naked[2];
camelia rakudo-moar f69983: OUTPUT«(List)␤d␤»
n0tjack is naked a scalar (reference) to an array there, or a proper @-like array?
16:02 crux left, mbrock joined 16:03 rurban joined
jnthn n0tjack: Proper @-like 16:03
n0tjack: Without sigil means "don't impose any context"
n0tjack what do you mean by "don't impose context"? if naked is @-like, then the compiler surely noticed <n u d e> is an array and (from context) decided naked's true sigil is @? 16:04
gfldex m: loop (;$++ < 4;){ loop (;$++ < 3;) { say $++ } } # this may or may not be a bug 16:05
camelia rakudo-moar f69983: OUTPUT«0␤1␤2␤»
jnthn m: my @a = 1; my \a = 1; say @a.perl; way a.perl 16:06
camelia rakudo-moar f69983: OUTPUT«5===SORRY!5=== Error while compiling /tmp/B4mJzGKvf9␤Undeclared routine:␤ way used at line 1. Did you mean 'say'?␤␤»
jnthn m: my @a = 1; my \a = 1; say @a.perl; say a.perl
camelia rakudo-moar f69983: OUTPUT«[1]␤1␤»
jnthn n0tjack: Assigning to something wiht an @ sigil is storing into an array
And note that 16:07
psch gfldex: not sure where you see a bug there
n0tjack jnthn: so \whatever creates the minimal possible container for the referent?
jnthn m: my @a = <n u d e>; say @a.WHAT
camelia rakudo-moar f69983: OUTPUT«(Array)␤»
jnthn m: my \a = <n u d e>; say a.WHAT
camelia rakudo-moar f69983: OUTPUT«(List)␤»
jnthn n0tjack: It creates *no* container
gfldex psch: all 3 $s should be different state variables, as it stands there are only 2
psch m: my $outer-state = 0; my $inner-state = 0; loop (;$outer-state++ < 4;) { loop (;$inner-state++ < 3;) { say $++ } }
camelia rakudo-moar f69983: OUTPUT«0␤1␤2␤»
psch gfldex: ^^^ that should clear it up 16:08
jnthn n0tjack: It's just a direct reference stored in the lexical scope (which may in turn even be optimized down to a VM-legel register if nothing closes over it)
psch gfldex: there is three, but the inner loop state isn't reset
n0tjack ah, so a sigily-variable has the powers of both its sigil and its type, but a bare-variable only has the powers of its type?
jnthn *level
gfldex psch: i stand corrected and confused
jnthn n0tjack: Something like that, yes :)
psch gfldex: state maintains it's state across reentry and all those things
gfldex: so the inner loop starts at 0 once, and never restarts 16:09
gfldex psch: that makes sense and $++ even more ebil!
n0tjack m: say \foo = my $bar = 'baz';
camelia rakudo-moar f69983: OUTPUT«5===SORRY!5=== Error while compiling /tmp/bUfsIH75Rh␤Preceding context expects a term, but found infix = instead␤at /tmp/bUfsIH75Rh:1␤------> 3say \foo =7⏏5 my $bar = 'baz';␤»
gfldex .oO( Stranger: I know Perl 6! Me: Good, explain { loop (;say $++ < 4;){ loop (;say $++ < 3;) { say $++ } } }. 16:10
n0tjack gfldex: He would have to be a very strange Stranger indeed to ask that question in the first place
oops, commute that pun. 16:11
dalek kudo/bool_enum: 49c8f9b | (Stefan Seifert)++ | src/ (3 files):
Turn Bool into a proper enum
 16:12
kudo/bool_enum: 8b0d845 | (Stefan Seifert)++ | src/core/operators.pm:
Fix SEQUENCE not expecting Bool ~~ Real to be true
kudo/bool_enum: 25fc19b | (Stefan Seifert)++ | src/Perl6/Metamodel/EnumHOW.nqp:
Fix type check for binding parameters

EnumHOW just needed an is_composed method like many other HOWs. Fixes regression in t/spec/S06-signature/optional.t
kudo/bool_enum: 0102ca6 | (Stefan Seifert)++ | src/ (2 files):
Bool improvements
gfldex is there any real and sensible usecase for anonymouse state variables? 16:15
16:16 kjs_ left
psch m: class Singleton { my $self; method new() { if !$++ { $self = self.bless } else { $self } } }; my $s = Singleton.new; my $ss = Singleton.new; say $s === $ss 16:16
camelia rakudo-moar f69983: OUTPUT«True␤»
psch heh
TimToady nope, we're all gonna use Turing machines cuz everything else can be reduce to them
timotimo m: class Test { submethod BUILD { say $++ } }; Test.new(); Test.new(); Test.new()
camelia rakudo-moar f69983: OUTPUT«0␤1␤2␤»
TimToady does not recall deciding that failing S/// should return Nil, and suspects it was just an accident of reimplementation 16:17
diakopter m: my \Colossus = my \Turing = return 16:18
camelia ( no output )
16:19 FROGGS joined
n0tjack m: constant \Does-this-work = do {42}; say Does-this-work; 16:19
camelia rakudo-moar f69983: OUTPUT«42␤»
gfldex m: my $a = $; my $b = $++ // 2; my $c = ++$ // 2; dd $a, $b, $c; 16:25
camelia rakudo-moar a33314: OUTPUT«Any $var = Any␤Int $var = 0␤Int $var = 1␤»
gfldex m: my $a = $ // 2; my $b = $++ // 2; my $c = ++$ // 2; dd $a, $b, $c;
camelia rakudo-moar a33314: OUTPUT«Int $var = 2␤Int $var = 0␤Int $var = 1␤»
gfldex why is b's $ turned into Int before the increment happend? 16:26
dalek osystem: 6411446 | RabidGravy++ | META.list:
Add Object::Permission

  github.com/jonathanstowe/Object-Permission		 16:27
psch gfldex: because postfix:<++> returns something
m: sub postfix:<plusplus>($a is rw) { my $b = $a; $a += 1; $b } # post increment as it's written, minus containering 16:28
camelia ( no output )
lucasb_ Doesn't it make sense if both s/// and S/// return a match object when success and Nil otherwise? Now, there should be a shortcut for both s/// and S/// to return the string (modified or not), otherwise the programmer would need 2 statements: one to change and the other get the actual string content 16:29
*to get
psch gfldex: actually, that was the wrong candidate i looked at :)
gfldex: multi postfix:<++>(Mu:U $a is rw) { $a = 1; 0 } # the right one
RabidGravy possibily the most fundamentally pointless and crack induced thing I have made yet
diakopter RabidGravy: crack is bad 16:30
itz_stmuk RabidGravy: not XAP then ;)
RabidGravy itz_stmuk, I was thinking of porting it, I actually have the source code here ;-) 16:31
16:31 Begi1115 left
n0tjack m: say @[1]; 16:31
camelia rakudo-moar a33314: OUTPUT«[1]␤»
n0tjack m: say +@;
camelia rakudo-moar a33314: OUTPUT«0␤»
itz_stmuk :) 16:32
n0tjack m: say %{'made-up'};
camelia rakudo-moar a33314: OUTPUT«5===SORRY!5=== Error while compiling /tmp/HKc0n167bN␤Unsupported use of %{'made-up'}; in Perl 6 please use %('made-up')␤at /tmp/HKc0n167bN:1␤------> 3say %{'made-up'}7⏏5;␤»
n0tjack m: say @[*-1];
camelia rakudo-moar a33314: OUTPUT«[WhateverCode.new]␤»
n0tjack oh, I see, @ isn't an anonymous array variable, it's a contextifier 16:33
gfldex n0tjack: not the sigil itself but @ followed by any bracket 16:35
psch m: sub f($a) { say (@ .push($a)).perl }; f 1; f 2; f 3 16:36
camelia rakudo-moar a33314: OUTPUT«[1]␤[1, 2]␤[1, 2, 3]␤»
n0tjack gfldex: how does that relate to the use of bracketless @ for typecasting, as in sub(Int @foo) {...}; foo(my Int @ = [1,2,3,4]); ? 16:37
psch i'm really starting to like the space-lenient method call infix
gfldex n0tjack: in that case i understand it to be some anonymouse state array 16:38
psch in "my Int @", @ is clearly not a state-scoped, but my-scoped
+variable 16:39
16:39 adu left, cschwenz joined
gfldex it's a mind-field! 16:39
n0tjack heh
16:39 cschwenz left
lucasb_ when p6 have a Cat object, it will run after the anonymouse variables 16:39
n0tjack lucasb_: going to be hard to catch him though
psch lucasb_: and eye the Rats suspicuously :)
16:40 raiph left
itz_stmuk the cat would like the FatRats 16:40
n0tjack m: say @[Inf]; 16:42
camelia rakudo-moar a33314: OUTPUT«[Inf]␤»
n0tjack heh
gfldex m: sub foo (@a) { dd @a }; foo((state Int @) = [1,2]); foo((state @) = [1,2]);
camelia rakudo-moar a33314: OUTPUT«Array[Int] $var = Array[Int].new(1, 2)␤Array $var = $[1, 2]␤»
n0tjack m: (constant @) = [1,2,3]; 16:43
camelia rakudo-moar a33314: OUTPUT«5===SORRY!5=== Error while compiling /tmp/zlKM508FvG␤Missing initializer on constant declaration␤at /tmp/zlKM508FvG:1␤------> 3(constant @7⏏5) = [1,2,3];␤»
n0tjack eh, worth a shot.
TimToady lucasb_: I don't understand why you want to make the match vs string decision separate from the s/// vs S/// notation, when the whole point of S/// is to return the string instead of the match
gfldex m: $.VAR.name.say; 16:45
camelia rakudo-moar a33314: OUTPUT«5===SORRY!5=== Error while compiling /tmp/NMgZcz_BGo␤Variable $.VAR used where no 'self' is available␤at /tmp/NMgZcz_BGo:1␤------> 3$.VAR7⏏5.name.say;␤ expecting any of:␤ term␤»
lucasb_ TimToady: I'm trying to clarify my thoughts in a example, just a second :)
gfldex m: $ .VAR.name.say; # not very introspective that fellow
camelia rakudo-moar a33314: OUTPUT«===SORRY!===␤Method call must either supply a name or have a child node that evaluates to the name␤»
psch r: A: { for 1 { for 1 { last A } } } 16:47
camelia ( no output )
..rakudo-moar a33314: OUTPUT«control exception without handler␤ in block at /tmp/tmpfile:1␤ in block <unit> at /tmp/tmpfile:1␤␤»
psch i wonder where moar gets its handler there
NQP doesn't install one via Grammar or Actions 16:48
nqp-m: A: { for [1] { for [1] { last A } } }
camelia nqp-moarvm: OUTPUT«No exception handler located for last_label␤ at gen/moar/stage2/NQPHLL.nqp:1416 (/home/camelia/rakudo-inst-2/share/nqp/lib/NQPHLL.moarvm::20)␤ from <unknown>:1 (/home/camelia/rakudo-inst-2/share/nqp/lib/NQPHLL.moarvm::23)␤ from <unknown>:1 (<ephemeral f…»
psch ah
hm
that at least shows i'm right with not seeing it on an nqp level :)
jnthn: should that work on nqp level? 'cause at the moment it looks to me like fixing it there is the easier way to fix the r-j bug 16:50
16:50 abraxxa left 16:59 tokuhirom joined
lucasb_ TimToady: ok, I was confused, nevermind. So a failed S/// should really return the original string, not Nil, right? I just thought it could be convenient to get a match object with .subst or S/// if the user ever needs it, or get a string from .subst-mutate and s/// if the user wants it, instead of the match object. But I understand, they are different operations. 17:02
17:04 tokuhirom left
TimToady is never confused, except when he is. 17:04
dalek kudo/nom: f5276ca | TimToady++ | src/core/Failure.pm:
handled failures are still failures

so don't be quite so quick to coerce them to 0 or ''		 17:06
kudo/nom: ad08a1d | TimToady++ | src/core/Failure.pm:
Failure should also not return useful integer
17:19 rangerprice joined
jnthn psch: Don't think labeled loops were implemented in NQP, no 17:19
n0tjack m: goto comefrom; 17:21
camelia rakudo-moar a33314: OUTPUT«5===SORRY!5=== Error while compiling /tmp/LMZMQwBqk5␤Undeclared routine:␤ comefrom used at line 1␤␤»
17:21 sufrostico left
rangerprice hi guyz ! 17:21
dalek ast: 79e6d16 | TimToady++ | S05-substitution/subst.t:
non-matching S/// should return original string
 17:22
psch m: A: for 1 { for 1 { last A }; CONTROL { default { .perl.say } } } # oh boy 17:29
camelia rakudo-moar ad08a1: OUTPUT«(signal SEGV)»
psch that seems to confirm that mvm is cheating somehow with it's control handlers..? :S
j: A: for 1 { for 1 { last A }; CONTROL { default { .perl.say } } }
camelia rakudo-jvm ad08a1: OUTPUT«X::AdHoc.new(payload => "Died")␤»
zengargoyle m: sub a(:$count){$count} %eg = 1 => ( count => 1 ), 2 => ( count => 2; say a(%eg<1>)
camelia rakudo-moar ad08a1: OUTPUT«5===SORRY!5=== Error while compiling /tmp/uUtpkvbgq8␤Preceding context expects a term, but found infix = instead␤at /tmp/uUtpkvbgq8:1␤------> 3sub a(:$count){$count} %eg =7⏏5 1 => ( count => 1 ), 2 => ( count => 2;␤»
RabidGravy segfault no good
zengargoyle m: sub a(:$count){$count} my %eg = 1 => ( count => 1 ), 2 => ( count => 2; say a(%eg<1>)
camelia rakudo-moar ad08a1: OUTPUT«5===SORRY!5=== Error while compiling /tmp/jkWJ_yH0ZL␤Strange text after block (missing semicolon or comma?)␤at /tmp/jkWJ_yH0ZL:1␤------> 3sub a(:$count){$count}7⏏5 my %eg = 1 => ( count => 1 ), 2 => ( co␤ expecting any of:␤ in…»
zengargoyle m: sub a(:$count){$count} my %eg = 1 => ( count => 1 ), 2 => ( count => 2 ); say a(%eg<1>); 17:30
camelia rakudo-moar ad08a1: OUTPUT«5===SORRY!5=== Error while compiling /tmp/xknOHfw4GE␤Strange text after block (missing semicolon or comma?)␤at /tmp/xknOHfw4GE:1␤------> 3sub a(:$count){$count}7⏏5 my %eg = 1 => ( count => 1 ), 2 => ( co␤ expecting any of:␤ in…»
zengargoyle m: sub a(:$count){$count}; my %eg = 1 => ( count => 1 ), 2 => ( count => 2 ); say a(%eg<1>);
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in sub a at /tmp/iCar9EaOtk:1␤ in block <unit> at /tmp/iCar9EaOtk:1␤␤»
zengargoyle m: sub a(:$count){$count}; my %eg = 1 => ( count => 1 ), 2 => ( count => 2 ); say a(|%eg<1>); 17:31
camelia rakudo-moar ad08a1: OUTPUT«1␤»
17:31 llfourn left
zengargoyle m: sub a(:$count,:$len){$count*$len}; my %eg = 1 => ( count => 1, len => 2 ), 2 => ( count => 2, len => 2 ); say a(|%eg<1>); 17:32
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 0 arguments but got 2␤ in sub a at /tmp/P0lP6pbZ_8:1␤ in block <unit> at /tmp/P0lP6pbZ_8:1␤␤»
zengargoyle m: sub a(:$count,:$len){$count*$len}; my %eg = 1 => ( count => 1, len => 2 ), 2 => ( count => 2, len => 2 ); say a(||%eg<1>); 17:33
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 0 arguments but got 2␤ in sub a at /tmp/C1qzw38FGo:1␤ in block <unit> at /tmp/C1qzw38FGo:1␤␤»
zengargoyle m: sub a(:$count,:$len){$count*$len}; my %eg = 1 => ( count => 1, len => 2 ), 2 => ( count => 2, len => 2 ); say a(|(%eg<1>.flat));
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 0 arguments but got 2␤ in sub a at /tmp/1L1k72M6yt:1␤ in block <unit> at /tmp/1L1k72M6yt:1␤␤»
zengargoyle m: sub a(:$count,:$len){$count*$len}; my %eg = 1 => ( count => 1, len => 2 ), 2 => ( count => 2, len => 2 ); say a(|%eg<1>);
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 0 arguments but got 2␤ in sub a at /tmp/ptBwoyFzwp:1␤ in block <unit> at /tmp/ptBwoyFzwp:1␤␤»
moritz : sub a(:$count,:$len){$count*$len}; my %eg = 1 => ( count => 1, len => 2 ), 2 => ( count => 2, len => 2 ); say a(|%eg<1>.hash);
m: sub a(:$count,:$len){$count*$len}; my %eg = 1 => ( count => 1, len => 2 ), 2 => ( count => 2, len => 2 ); say a(|%eg<1>.hash);
camelia rakudo-moar ad08a1: OUTPUT«2␤»
zengargoyle ah 17:34
moritz zengargoyle: | on a list produces positional elements, | on a hash or pair produces named
zengargoyle thanks
17:38 ^elyse^ joined 17:39 ^elyse^ left, sufrostico joined 17:41 Actualeyes left, MadcapJake joined
n0tjack m: my Int $i where * > 0; $i = 3; $i = 1; $i = -99; 17:43
camelia rakudo-moar ad08a1: OUTPUT«Type check failed in assignment to $i; expected <anon> but got Int␤ in block <unit> at /tmp/UPKiZyfGz9:1␤␤»
17:44 CIAvash joined
dalek c: e135e36 | (Steve Mynott)++ | doc/Language/5to6 (2 files):
rename Perl 5 to Perl 6 Translation to 5to6-nutshell GH #158
 17:44
c: cd52552 | (Steve Mynott)++ | doc/ (8 files):
Merge branch 'master' of github.com:perl6/doc
FROGGS n0tjack: LTA for sure
MadcapJake trying to submit an exercise for exercism's perl6 exercises and I can't figure out how to use modules. The first exercise `Bob` , I'm getting `could not find Bob in any of:` 17:45
and I have this at the top of `bob.pm6`: `unit module Bob;`
n0tjack FROGGS: I think it's amazing that you can have anonymous conditional types at all!
FROGGS: before, I thought the where-clause was just a feature of function signatures. 17:46
FROGGS MadcapJake: you'd have to perl6 -I. -e 'use bob'
MadcapJake what do those flags mean?
17:47 Oatmeal joined
n0tjack MadcapJake: -I. means "include the current directory (.) in the `classpath`". 17:47
MadcapJake: The -e means "execute this string as if it were a script" aka (e)val.
IOW, just make sure when you invoke perl, the Bob module is on your module search path. 17:48
psch (and preferably decide whether you want to call the module bob or Bob, and use the same in the filename *and* module declaration)
MadcapJake the test file has this: `my $module = %*ENV{'EXERCISM'} ?? 'Example' !! 'Bob'; \n EVAL("use $module"); 17:49
is that what you're talking about?
psch MadcapJake: if it runs "use Bob" and you have bob.pm, it won't fit it even if bob.pm is in the libpath 17:50
s/fit/find/
n0tjack MadcapJake: that question comes after the first: are either Example or Bob or both on your search path?
MadcapJake yeah that's what i'm seeing, why is that?
well if I use `-I.` it still doesn't find it 17:51
psch MadcapJake: because "use Bob;" looks for Bob.pm, not bob.pm
pink_mist because Bob is not bob?
MadcapJake ohhh, the filename
n0tjack Bob prefers Robert.
MadcapJake lol
dalek ast: af3bcda | TimToady++ | S05-substitution/subst.t:
add rt annotation
 17:55
17:56 sufrostico left 17:57 mbrock left, azawawi joined
azawawi hi 17:57
retupmoca: ping
17:58 pmurias_ left 18:00 dakkar left
retupmoca azawawi: pong 18:00
18:00 ambs joined
azawawi retupmoca: re latest patch, it made performance worse unfortunately... 18:01
18:01 AndyDee left
azawawi retupmoca: see github.com/azawawi/perl6-file-zip/...es/test.p6 18:02
retupmoca huh
I didn't expect that
azawawi retupmoca: this is my home machine (amd fx-8150) which is slower than my work machine in single core performance (i7 2600) 18:03
retupmoca: this machine is better in I/O though it has an SSD
18:04 kjs_ joined
azawawi retupmoca: with latest patch, real0m46.618s 18:06
retupmoca: when reverted to 40b310a, i get real0m16.582s 18:07
retupmoca oh, that's a lot worse
18:08 mayuresh joined
azawawi let me calculate avg file size 18:08
mayuresh hello :) 18:09
azawawi retupmoca: ~20k avg file size 18:10
18:11 xpen joined 18:12 dha joined 18:13 mayuresh left
azawawi retupmoca: buffer is 1K, increased it to 8K, got it down to 11 secs :) 18:14
retupmoca: more than 8K is minimal performance gain
18:15 leont joined 18:16 xpen left
psch well, the bit that segfaults on moar above is another 5 test file FAILs on j-spectest that i'd say need more than just a jvm-specific fix 18:16
azawawi retupmoca: my guess is that .list has a hidden cost attached to it 18:17
psch the "expected Positional but got Seq" might be related to the workaround for anon subs disappearing in sink context... 18:18
azawawi retupmoca: to my surprise, ~ is rather fast :)
psch which is something i looked at for quite a bit already but didn't make any headway with :/
r: sub f { sub { "foo" } }; f()()
MadcapJake line 16 of `bob.t` reads: `ok Bob.can('hey'), 'Class Bob has hey() method';` gives me this error: `Too many positionals passed; expected 2 arguments but got 3 in block <unit> at bob.t:16` Why would there be possible arguments in this situation?
camelia rakudo-jvm ad08a1: OUTPUT«java.lang.NullPointerException␤ in block <unit> at /tmp/tmpfile:1␤␤»
( no output )
jnthn azawawi: If you're on latest Rakudo, that's maybe 'cus timotimo++ patched it to be 2x faster earlier today :)
retupmoca azawawi: it looks like both .subbuf and .list do a slow-copy (not using splice), plus we do another when we call the buf8.new 18:19
azawawi: I just pushed another version - can you let me know if it's any faster this time?
MadcapJake my `hey` method is just ` method hey ($input) { ... }`
psch m: use Test; class Bob { method hey { } }; ok Bob.can('hey'), "Bob can hey"
camelia rakudo-moar ad08a1: OUTPUT«ok 1 - Bob can hey␤»
retupmoca (the only potentially slow bit is the list slice now)
psch m: use Test; class Bob { method hey($) { } }; ok Bob.can('hey'), "Bob can hey" 18:20
camelia rakudo-moar ad08a1: OUTPUT«ok 1 - Bob can hey␤»
MadcapJake so why's mine failing :P
psch MadcapJake: no idea, i'd need more context :)
azawawi retupmoca: still running lol 18:21
MadcapJake gist.github.com/MadcapJake/a4584ed1048783680d17 18:22
there's both files and I'm running `perl6 -I. bob.t` in the cmdline 18:23
azawawi retupmoca: this version is way too slow
psch m: use Test; module Bob { class Bob { method hey { } } }; ok Bob.can('hey'), "Bob can hey" 18:24
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 2 arguments but got 3␤ in block <unit> at /tmp/4NSWec4bCM:1␤␤»
psch MadcapJake: you're declaring a class Bob inside a module Bob
MadcapJake: so, Bob::Bob
m: module Bob { class Bob { method hey { } } }; say Bob.can('hey');
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 2 arguments but got 3␤ in block <unit> at /tmp/2DRgjI2wwR:1␤␤»
retupmoca azawawi: right. I'm going to revert back to the original, and tonight I'll set up a proper test and figure out how to stick 2 bufs together quickly
MadcapJake ohh, well the test file makes it seem like i needed to do it as a class method instead of a su
sub*
psch MadcapJake: it does, but you can declare a class as a package in perl 6 18:25
FROGGS m: module Bob { class Bob { method hey { } } }; say Bob.^can('hey');
camelia rakudo-moar ad08a1: OUTPUT«Too many positionals passed; expected 2 arguments but got 3␤ in block <unit> at /tmp/tWOW2_cyM2:1␤␤»
azawawi retupmoca: am i using the API in a correct way github.com/azawawi/perl6-file-zip/...ip.pm6#L93 ?
psch MadcapJake: as in, that's actually what happens when a file contains a "class" statement in the outermost scope
18:25 ambs left
MadcapJake so the outermost scope becomes a class? 18:25
azawawi retupmoca: im feeding it the compressed stream as a whole 18:26
retupmoca azawawi: yeah, that should be fine
azawawi retupmoca: real3m30.320s ....
18:27 ambs joined
azawawi retupmoca: so basically your test script is github.com/azawawi/perl6-file-zip/...es/test.p6 18:27
retupmoca: the webdriver.xpi file size is like 600K
MadcapJake i'm not sure i understand what I need to do, I tried removing the class dec and changing the `hey` method to a sub with `is export`, but still getting the same exact error
psch MadcapJake: that's because Bob is still a package, not a class 18:28
MadcapJake so how do i make it a class?
psch (or module)
MadcapJake: just remove the "unit module Bob;" at the top
and possibly the "is export"
MadcapJake oh :P confusing
psch not sure about the latter, but i'm pretty sure you don't need it
MadcapJake: yeah, a bit
MadcapJake: in my mind, modules and classes are both packages
azawawi jnthn: 2015.10-51-gad08a1d is the perl6 im using 18:29
psch MadcapJake: a package is one of those things that you talk about with ::, e.g. Friends::Alice
MadcapJake: and any level in such a package can be a module or a class
m: class A { class B { } }; A::B.new.perl.say
camelia rakudo-moar ad08a1: OUTPUT«B.new␤»
MadcapJake but they're different in Perl6 right?
timotimo azawawi: did you get a speed improvement in the version with ~ by upgrading your rakudo? 18:30
psch m: module A { class B { } }; A::B.new.perl.say
camelia rakudo-moar ad08a1: OUTPUT«B.new␤»
18:30 kjs_ left
psch MadcapJake: under the hood, yes, but not really from the outside 18:30
azawawi brb let me access my $work machine... and test :)
psch MadcapJake: which is why i think the error you got is definitely LTA
TimToady packages are really very much like packages in P5, except we don't put anything into them by default
psch MadcapJake: also of note is that my definition of package and module isn't quite synopses-conform
the synopses talk about module and package as synonyms, which i don't 18:31
(at least during my last reading, not sure if that changed)
so maybe i shouldn't have explained it like that because it might be confusing down the line :S
MadcapJake TimToady: I've never used P5 so that's probably why this is a bit confusing for me
18:32 sufrostico joined
psch FROGGS: does the NPE for sub f { sub { } } seem approachable to you? 'cause i don't seem to get it figured out 18:32
TimToady packages are simply a hierarchical system of named global namespaces 18:33
FROGGS psch: the what?
psch j: sub f { sub { "foo" } }; f()()
18:33 ambs left
camelia rakudo-jvm ad08a1: OUTPUT«java.lang.NullPointerException␤ in block <unit> at /tmp/ggXpHEOfsY:1␤␤» 18:33
FROGGS ahh, that one
psch FROGGS: ISTR that you and jnthn pointed me at compile_all_the_stmts
FROGGS: and i tried adapting the moar logic, which got hella confusing because of differences in stack handling and such 18:34
18:34 azawawi_ joined
FROGGS psch: yes, jnthn said something about a sinking feeling... 18:34
18:34 azawawi left
psch FROGGS: fwiw, there is differences between the backends, but i'm unsure in how far those differences are architectural 18:34
MadcapJake so you would use `unit module <module_name>` when you wanted a higher level modjule containing sub-packages? 18:35
psch MadcapJake: "module" and "package" both give you a class-less namespace
MadcapJake: they can also straight up export subs (or even methods) 18:36
TimToady having subpackages comes automatically with any kind of package
18:36 xfix joined
MadcapJake the module page in the docs never really explains when to us e`unit module` or what it is really 18:36
TimToady modules have extra stuff for exporting
classes have extra stuff for, er, classing...
MadcapJake oh so they're similar albeit one is export-focused while the other is inheritance-focused, you could say? 18:37
TimToady to the first approximation, modules are for providing a set of functions, while classes are for providing a set of methods
and packages don't care about supplying anything except global name lookups
but a class can also export, and any of them can supply global names 18:38
so it's just various degrees of specialization; a grammar is a more specialized class 18:39
MadcapJake so classes can provide all the exportational goodness of a module but you need to explicility specifiy it?
n0tjack Is package ⊃ module ⊃ class ⊃ grammar on the right rack?
TimToady it should work the same as a module, just add 'is export'
18:40 espadrine left
TimToady n0tjack: and roles fit in there somewhere too, but of course it doesn't have to be linear 18:40
n0tjack TimToady: except when using a class, you must either instantiate it or call static methods, no?
TimToady you can export functions too
FROGGS nqp-j: sub f { sub { "foo" } }; f()() 18:41
camelia nqp-jvm: OUTPUT«(signal ABRT)#␤# There is insufficient memory for the Java Runtime Environment to continue.␤# pthread_getattr_np␤# An error report file with more information is saved as:␤# /tmp/jvm-5288/hs_error.log␤»
TimToady you can probably change any module definition to declare it with 'class' and it'd still work as a module
n0tjack TimToady: so from an importer's perspective the difference between class Foo is export {} and unit module Foo is in the former you have to either use Foo.bar or my $f = Foo.new(); $f.bar whereas in the latter you can just say bar() with no rigamarole?
18:42 ambs joined
TimToady Foo.bar is never going to call a function 18:42
18:42 sufrostico left
n0tjack sorry, Foo.bar() and $f.bar() 18:42
TimToady Foo.bar() is never going to call a function 18:43
n0tjack no?
TimToady we keep them entirely distinct, when it comes to dispatchers
and bar() will never call a method
bar() only calls things in the current lexical scope (including things that were imported) 18:44
FROGGS psch: funnily, that works: nqp-j -e 'sub f() { sub i() { "foo" } }; say(f()())'
TimToady Foo.bar only calls things that are registered as methods in the metaclass object
n0tjack TimToady: I'm trying to picture how a class could export a function (not method) which could be invoked by an importer without mentioning Foo in some way
TimToady just declare it as a sub instead of as a method, and export it, just as you would in a module 18:45
n0tjack I see.
18:45 raiph joined
n0tjack Feels weird that classes are allowed to export unclassy things. 18:45
18:45 rurban left 18:47 rurban joined
TimToady m: class Foo { sub bar () is export { say "HERE" } }; import Foo; bar() 18:48
camelia rakudo-moar ad08a1: OUTPUT«HERE␤»
n0tjack wow, that's going to kill me when I forget actual class methods are spelled "method" and not "sub"
FROGGS heh 18:49
TimToady n0tjack: Perl 6 is mostly not about telling people what they can't do. As with the color wheel of light, we try to make things additive rather than subtractive. 18:50
18:51 tokuhirom joined
n0tjack zens out for a moment 18:51
18:55 tokuhirom left
n0tjack the prospect of having to learn git (YAVCS) is harshing my zen 18:56
RabidGravy YOU'RE HARSHING MY MELLOW MAN! 18:57
n0tjack I feel like mellow is an underappreciated noun 18:58
RabidGravy oh wait, nah it's not too much of a problem, I was a bit of a git-resister a few years back but really it's just another vcs but with a distinction between local and remote 18:59
jdv79 encino? 19:01
blast from the past
19:03 rurban left, virtualsue joined
flussence no need to learn another VCS, just forget all the old ones and use git's frontend to them instead :) 19:03
19:03 rurban joined
azawawi_ jnthn: confirmed with the timotimo++ performance fix and revert to old ~ buffer in Compress::Zlib, im seeing a drop from ~20 sec to ~12 secs 19:03
psch FROGGS: that's weird :/
FROGGS: oh, no, it isn't
azawawi_ retupmoca: also with increasing the buffer size from 1K to 8K, 12 seconds becomes 8 seconds :)
psch j: sub f { sub g { "foo" } }; f()() 19:04
camelia rakudo-jvm ad08a1: OUTPUT«java.lang.NullPointerException␤ in block <unit> at /tmp/Y84AtXxDFZ:1␤␤»
psch wait, it is
what
i thought that worked
FROGGS: turns out it *is* weird :P
n0tjack bbl &
psch FROGGS: i had thought i had maintained that it's the anon sub that gets swallowed, but apparently is any non-cont sub
j: sub f { my & = sub g { "foo" } }; f()() 19:05
camelia ( no output )
psch s/maintained/figured out/
FROGGS I'm trying to get behind it... 19:06
TimToady j: sub f { sub g { "foo" } }; f().WHAT.say
camelia rakudo-jvm ad08a1: OUTPUT«java.lang.NullPointerException␤ in block <unit> at /tmp/502V0csSne:1␤␤»
retupmoca timotimo, azawawi_: github.com/retupmoca/P6-Compress-Z...0ae16bbfe1 seems to speed it up 5x (using rakudo before the ~ patch)
TimToady it's not just .()
retupmoca azawawi_: yeah, I know, but I wanted to get this part sorted first
FROGGS psch / TimToady: it is an old standing bug...
psch / TimToady: remember that we had some "#?if jvm\n my $ = ...\n#?endif" in the setting before GLR? it is most likely the same issue 19:07
TimToady nod 19:08
MadcapJake is there a quick way to check if a list is empty or do I just compare length to 0?
TimToady a list is false if it's empty
no comparison necessary
timotimo azawawi_: aaw, not even 2x faster :( 19:09
19:09 n0tjack left
TimToady m: say ().so; say (0).so 19:09
camelia rakudo-moar ad08a1: OUTPUT«False␤False␤»
FROGGS MadcapJake: if @empty-list { } <--- it wont make it in the if's body
TimToady m: say ().so; say (0,).so
camelia rakudo-moar ad08a1: OUTPUT«False␤True␤»
MadcapJake m: my $s = "4?"; $s.comb(/<:L>/).defined
camelia ( no output )
FROGGS m: my $s = "4?"; say $s.comb(/<:L>/).defined
camelia rakudo-moar ad08a1: OUTPUT«True␤»
azawawi_ timotimo: again im testing on different CPU platforms :)
FROGGS m: my $s = "4?"; say $s.comb(/<:L>/)
camelia rakudo-moar ad08a1: OUTPUT«()␤»
FROGGS m: my $s = "4?"; say $s.comb(/<:L>/).so
camelia rakudo-moar ad08a1: OUTPUT«False␤»
MadcapJake what is the `so` method?
FROGGS MadcapJake: a list is defined, but false if it is empty 19:10
azawawi_ timotimo: AMD FX-xxxx is known for slow single core performance
timotimo azawawi_: i meant the 20sec to 12sec thing you said; i thought that's on a single box?
FROGGS MadcapJake: like asking for a Bool
TimToady MadcapJake: the opposite of .not
timotimo azawawi_: even if that processor is slow, the factor should be the same, no?
MadcapJake oh ok cool, thanks!
FROGGS MadcapJake: like in English, so?
azawawi_ timotimo: SSD on the slower machine. go figure :)
19:11 sufrostico joined
MadcapJake never heard it like that, maybe it's my generation but `so` is usually just for when you didn't bring your homework to class and you just didn't care :P 19:11
flussence has a SSD in a netbook and spinning rust on desktop. Great for noise levels but not much else... 19:12
19:13 rangerprice left, espadrine joined
psch FROGGS: those #if jvm bits are still around 19:14
19:15 sufrosti1o joined
psch FROGGS: core/metaops.pm 19:15
19:15 _dolmen_ joined
PerlJam MadcapJake: "make it so" :) 19:15
psch FROGGS: it's exactly this kind of issue, METAOP_REDUCE_* returns a sub
MadcapJake or the valley girls saying "soooo true!" xD
PerlJam MadcapJake: "That care is *so* blue" 19:16
s/care/car/
19:16 rarara__ joined
psch plays A7 F7 G7 C7 19:16
MadcapJake so really is a bizarre word if you think about it too much :S
psch well, Amin7 probably
or a7, depending on your notational preferences
azawawi_ timotimo: to be clear that benchmark was done on the same machine that i use @ work so yes 20 -> 12 after building a new rakudo
rarara__ Is there in Perl 6 an easy way to transpose a list of list ?
retupmoca azawawi_: I just bumped the buffer up to 8k in my repo as well
azawawi_: hopefully it's now "fast-ish" 19:17
MadcapJake if you're on exercism: give me some tips! exercism.io/submissions/4bbfaf53485...b49e6b41a9
retupmoca (my simplestic test went from ~38 seconds to ~2 seconds with both changes)
azawawi_ timotimo: and that's perfectly normal, i still didnt optimize I/O operations when writing files and recreating folders 19:18
dha I'm in favor of not making the case of the note significant, so I'd go with Amin7
19:18 boggard left
azawawi_ retupmoca: cool.. thanks :) 19:18
rarara__ something like ((A B C) (D E F) (G H I)) -> ((A D G) (B E H) (C F I)) ?? 19:19
MadcapJake looking at other implementations, I should've used `~~` xD
azawawi_ retupmoca: now down to 10 seconds on AMD machine :)
PerlJam rarara__: See rosettacode.org/wiki/Matrix_transposition#Perl_6 for some ideas :)
azawawi_ retupmoca++, timotimo++, jnthn++ 19:20
azawawi_ sleepy & 19:21
good night everyone
PerlJam azawawi_: g'night!
19:22 AlexDaniel joined
RabidGravy the problem with Perl 6 is that it *MAKES* me write things like "for $!handle.lines.map({Group.new(line => $_) }).sort({$^a.gid})", it's like demonic posession 19:22
lucasb_ m: say [Z] (<A B C>,<D E F>,<G H I>) 19:23
camelia rakudo-moar ad08a1: OUTPUT«((A D G) (B E H) (C F I))␤»
PerlJam at least it's concise possession
lucasb_ rarara__: ^^ rosettacode didn't had this example
rarara__ Yep
PerlJam lucasb_: add it! :)
rarara__ I was just figuring out :D
m: my $g=(("A","B","C"), ("D", "E", "F"), ("G", "H", "I")); say [Z] $g;
camelia rakudo-moar ad08a1: OUTPUT«((A D G) (B E H) (C F I))␤»
rarara__ much better than rosettacode 19:24
dha RabidGravy - you say that like it's a *bad* thing. 19:25
TimToady we should probably require @$g there
19:25 azawawi_ left
RabidGravy dha, >:) 19:25
and if I make "$!handle handles <lines>" then I can lose a few chars so it fits nicely in 80 chars with some space 19:27
19:28 bjz joined
psch that's actually interesting 19:28
i mean, the chainy stuff
as in, java gets flak for imposing similar programming style
i guess it comes down to huffmanization
RabidGravy I don't think I feel imposed upon to do that, it just seems natural to my poor brane 19:29
psch oh, right. i guess "i *can* write it like that" instead of "i either write it like that or add a bunch of variables for each call" is the better argument as for why it seems fine 19:31
19:31 ollej joined, ollej left, ollej joined
RabidGravy I'd never have done "for (sort { $a.gid <=> $b.gid } map { Group->new(line => $_ } <HANDLE>)" in P5 19:32
19:33 bjz left, kjs_ joined
RabidGravy or whatever 19:33
psch those seperately parsed block forms in perl5 looks so wrong to me by now
timotimo i would have really hoped for a much better improvement in run times for azawawi's unzip task :(
psch but yes, i agree that in p5 that's probably considered somewhat bad 19:34
RabidGravy I know people who would commonly write stuff like that 19:35
I'd probably abstract it with further layers of methods myself
geekosaur would write that but put each chunk on its own line 19:36
retupmoca timotimo: yeah, me too
geekosaur sort BLOCK \n map BLOCK [\n if EXPR is long enough] EXPR
retupmoca but at least we killed one bottleneck in Compress::Zlib
or improved at least
lucasb_ gfldex++ # added it to rosettacode :) 19:39
19:40 kjs_ left 19:42 Hor|zon joined 19:47 Hor|zon left
timotimo maybe someone wants to have a closer look at Buf.list to make it better? 19:48
retupmoca well, I'm not even using Buf.list anymore
I'm just directly calling nqp::splice
19:51 _dolmen_ left 19:52 Hor|zon joined 19:54 bjz joined 19:55 vendethiel joined 20:00 bjz_ joined, bjz left 20:06 CIAvash left
timotimo right, but if it's so costly, perhaps it'll benefit someone in the future 20:07
20:08 raiph left 20:09 raiph joined 20:12 bjz_ left, xpen joined 20:13 bjz joined 20:14 colomon left 20:18 xpen left 20:20 kjs_ joined 20:22 darutoko left 20:23 kjs_ left
gfldex m: sub foo(@a is copy, :$flag){ say @a }; my @a = 1,2; (&foo.assuming(:flag))(@a); 20:24
camelia rakudo-moar ad08a1: OUTPUT«replace this Array is copy logic␤ in sub __PRIMED_ANON at EVAL_2:2␤ in sub __PRIMED_ANON at EVAL_2:1␤ in block <unit> at /tmp/GhmXZoyt9p:1␤␤»
nine timotimo: Buf.list definitely looks like it can be improved
gfldex is that LTA, gfldex-bug or rakudobug?
nine timotimo: also it returns an Array instead of a List which is something I really don't like. 20:26
20:27 pdcawley left 20:29 Oatmeal left 20:31 Oatmeal joined 20:34 colomon joined, s_kilk joined, rurban left
FROGGS gfldex: that's a GLR leftover... a todo for us 20:34
psch: compiling on jvm is sooo damn sloooow :/ 20:35
20:37 rurban joined
psch FROGGS: yes, i know :l 20:38
FROGGS: i guess donaldh++'s suggestion of serializing annotations might help 20:39
FROGGS: but well, no idea how that'd look
FROGGS hmmm, me neither
I seem to lack imagination when it comes to jvm related stuff
leont FROGGS: I'm doing exactly the same right now :-/ 20:40
yoleaux 27 Oct 2015 19:24Z <FROGGS> leont: is there anything wrong with my patch here? github.com/rakudo/rakudo/commit/6063886343
FROGGS I currently try to port the logic of compile_all_the_stmts from moar to jvm... without much luck so far, but we'll see 20:41
hmpf, SOE again
leont No idea, can't really open firefox with perl6-j compiling, that'd be too much for my laptop surely
20:43 lizmat joined
lizmat good *, #perl6! 20:43
psch FROGGS: i mostly got ArrayIndexOutOfBoundsExceptions when i tried, which leaks from deep in objectweb.asm
20:43 [Sno] left
psch FROGGS: i'm slightly annoyed they'd leak that and not wrap it somewhat meaningful... 20:44
FROGGS leont: yes, thunderbird and firefox shut down here too :o) 20:45
leont: that patch just removes an exit(0) in Test.pm in an END block, which I think was a thinko in the first place
leont Test::Builder sets the exit value to min($failed, 254) 20:46
I'm not sure you can do that in Perl 6 without actually exiting
20:47 geraud joined
FROGGS leont: we do an exit(min($failed, 254)) conditionally, yes 20:48
the exit(0) was in an all-is-fine else branch
rarara__ Is there a way for plugging in C function which return a List in an efficient way? 20:49
leont in p5, you can set $? in and END block, and it will be used when perl is exited, even when other code runts after it (unless that also sets $? obviously)
FROGGS rarara_: efficient? in most cases we need to copy that returned list to create something more high level...
20:49 adu joined
FROGGS leont: we don't have that feature... 20:50
rarara__ C doesn't have lists. What I mean is to create it in C, ike you do it in XS
leont exit(0) if all is fine sounds not very useful at first sight
rarara__ so that you can do the cycles in C
FROGGS leont: okay, that's good enough for me :o)
rarara_: you can do stuff in C using Inline::C or NativeCall, and return stuff from C 20:51
leont wants to give Test.pm some love when the harness is done. It's rather primitive for my tastes.
FROGGS hmmm, primitive is sometime worth a lot
leont In particular, it's very hard to test itself, because it's outputting text instead of having abstractions in the back
FROGGS not necessarily here though 20:52
20:52 adu left
FROGGS ahh, yes 20:52
rarara__ FROGGS: with which data structure would you return a list of char* from C, for example? That would be converted in a list of string
leont Also, it's not very composible. The Test::Builder ecosystem proved how useful that can be, even if it the architecture proves to be limiting nowadays
FROGGS rarara_: you can return a CArray[uint8]
leont: I'm lacking experience there 20:53
20:53 tokuhirom joined
leont The idea is that you can build lots of tests modules that can be used together 20:54
For example Test::Deep, Test::Exception and Test::Class
rarara__ FROGGS is CArray a C type? 20:55
leont Or Test::Differences, which is one of my favorites (will actually diff in case of a mismatch)
20:55 samb1 joined
FROGGS rarara_: it maps directory to the memory which is potentially managed by C, so yes 20:55
leont Also, we kind of want extended diagnostics, but there are pretty much still in development in TAP-land :-/ 20:56
FROGGS leont: yes, I've used Test::Difference in the past
20:57 tokuhirom left
leont Also, being able to output something other than TAP may be useful in some environments. 20:58
20:58 dha left
leont I'm wondering what a testing slang would look like. Haven't thought out that idea yet though. 20:58
FROGGS yes, also true
flussence I've tried writing a module that does TAPv13, but it got kinda held up by the fact those diagnostics are YAML and the YAML module is seriously broken 20:59
leont is working on a YAML module for exactly this reason 21:00
flussence and I saw recently TAPv14 may have been started on, so I dunno if my code's much use
leont The TAP process is awfully slow :-/ 21:01
21:01 rurban left
RabidGravy I wrote a YAML parser in c# once, utter cack but it worked 21:02
rarara__ so char * is CArray[int8] 21:03
what about char ** ?
CArray[CArray[int8]]
leont I wrote on for p6, and it's reasonably close to being conformant, though edge-cases in YAML are always plenty 21:04
rarara__ then you would still have to loop over the whole external CArray, no?
FROGGS you can also have e.g. Pointer[CArray[uint8]]
lizmat rarara__: re irclog.perlgeek.de/perl6/2015-10-28#i_11443994 , .elems calls Iterator.count-only 21:05
which may be optimized, but if not, will indeed generate the entire Seq first
FROGGS rarara_: you have to access the array if you want to access the array... or maybe I misunderstand your question?
rarara__ would be nice to have access to an equivalent to nqp::p6box_s from C-code, no? 21:06
lizmat thanks
lizmat rarara_: re number of values seen, a simple counter within the callable would do that for you
rarara__ lizmat yea!
lizmat I don't think it's wise to let all grep's suffer in performance for that feature 21:07
my int $seen; foo.grep: { $seen++; actual check }
FROGGS rarara_: you can also return a Str from C, then it will be turned into a proper Perl 6 string 21:08
jdv79 why can't a type object be does'ed?
FROGGS rarara_: though it must be encodable properly of course
jdv79: you can 21:09
jdv79 i was looking for a poorman's way of marking a type object
FROGGS jdv79: that's how you do roles
ohh, be does'ed
rarara__ lizmat: I am figthing problems with Str.split
FROGGS like, class Foo { }; Foo does Bar; ?
flussence m: say ?(Int but False)
camelia rakudo-moar ad08a1: OUTPUT«False␤»
flussence m: say ?(Int but True) 21:10
camelia rakudo-moar ad08a1: OUTPUT«True␤»
lizmat rarara__: what iw the problem ?
flussence m: say ?+(Int but True)
camelia rakudo-moar ad08a1: OUTPUT«Use of uninitialized value of type Int+{<anon|69543024>} in numeric context in block <unit> at /tmp/ye2VPULbUv:1␤False␤»
jdv79 m: my $a = Any; role Foo {}; $a does Foo 21:11
camelia rakudo-moar ad08a1: OUTPUT«Cannot use 'does' operator with a type object.␤ in block <unit> at /tmp/4Zf64jDaOA:1␤␤»
jdv79 is there way to "mark" a type obj without altering it significantly? 21:12
rarara__ lizmat: now that I reason of it, there is no problem with the iterator. Indeed you can add a counter in yourself.
FROGGS m: my $a = Any; role Foo {}; $a = $a but Foo; say $a
camelia rakudo-moar ad08a1: OUTPUT«(Any+{Foo})␤»
moritz jdv79: my Mu %marks; %marks{Int} = True
rarara__ lizmat: it is the performance which is bad, and I dubt it is optimizable at all
lizmat the performance of grep or split ? 21:13
rarara__ split
moritz there, I just marked the Int type object without permanent damage :-)
lizmat ah, yeah, I was just looking at it
jdv79 moritz: thnaks ^H
forgot about but
FROGGS: thanks
leont Finally, the CORE.setting is compiled :-D
lizmat rarara__: I assume you're splitting on a string, right ?
not on a regex?
21:13 colomon left
rarara__ Right, in fact a "\t", but I checked the code and it is seems pretty good 21:13
leont Took "only" 50 minutes -_-
FROGGS leont: ewww 21:14
leont: Celeron with 466MHz?
21:14 CJKinni joined
lucasb_ FROGGS: omg, I had a celeron in the past 21:14
lizmat rarara__: I think I can speed up that case significantly
FROGGS lucasb_: I had a dual cpu mainboard with said cpu :o) 21:15
rarara__ that would be great
21:15 skids left
leont Awesome. It seems perl6-j doesn't define signal handling -_- 21:15
FROGGS aye
leont Not sure how to work around this cleanly :-/ 21:16
jobs
rarara__ FROGGS: the problem I see with the need of p6box_6 C api, is not that of returning a Str from C. It is the one of creating a whole List of strings and avoiding a Perl loop 21:17
This problem is general
FROGGS rarara_: you can also return a CArray[Str], and there would not be a Perl loop™ 21:18
21:18 lucasb_ left
FROGGS rarara_: because it happens in MoarVM 21:18
21:19 raiph left
rarara__ FROGGS thank you, seems great 21:19
FROGGS psch: still no luck :o(
21:21 telex left, raiph joined
FROGGS .tell psch That's my patch that does not help anything: gist.github.com/FROGGS/6f11befd8ca37b91fb4d 21:22
yoleaux FROGGS: I'll pass your message to psch.
21:22 telex joined
psch FROGGS: that looks close-ish to what i tried, from a first glance 21:23
yoleaux 21:22Z <FROGGS> psch: That's my patch that does not help anything: gist.github.com/FROGGS/6f11befd8ca37b91fb4d
FROGGS k
psch FROGGS: to me this strongly hints at "we're not quite looking at the right spot", fwiw
FROGGS yeah
psch FROGGS: as in, the differences in compile_all_the_stmts is architectural and *needs* to be there 21:24
FROGGS: the scary bit is the remaining questions, which is "why does it happen"... :P
[Coke] psch: use Bob on bob.pm will work on some filesystems.
FROGGS [Coke]: unless we enforce otherwise
psch i don't know if we can..?
FROGGS [Coke]: but it is usually not a good idea to be sloppy in that regard 21:25
psch but yes, i know that CI filesystems exist
FROGGS we can potentially check the filename we got back and bail out 21:26
leont Awesome, after all this effort it completely refuses to run my harness in any sensible way :-s 21:28
dalek kudo/nom: 07f186d | (Stefan Seifert)++ | src/core/Buf.pm:
Speedup Blob.list and Buf.list by 26x

Pushing directly into a List's $!reified buffer instead of going through Array's ASSIGN-POS to avoid the overhead.
  .list now also returns a List instead of an Array
21:28 rindolf left
lizmat nine++ :) 21:29
FROGGS leont: perl6-j only passes about 32% of the spectests, mind
26x??
nine++
leont It seems it doesn't do run() :-/ 21:30
nine Sometimes I kinda miss the GLR time :) The path was so clear then
jdv79 so i got it to work but its pretty gross
leont Or at least not output redirection on run()
jdv79 $.obj.set-once-attr = $obj.set-once-attr but Attribute::SetOnce::IsSet # to set the attr but have it remain "unset" or something:) 21:31
nine I also tried making Blob.list O(1) which should work because Lists and Blobs are both immutable and a List's $!reified buffer is a VMArray just like Blob is. Unfortunately those native types get in a way because List uses for example nqp::atpos while Blob uses nqp::atpos_i
jdv79 hopefully that can be cleaned up with a better design someday 21:32
21:32 colomon joined 21:34 ^elyse^ joined 21:35 xfix left
dalek ast: a254a27 | usev6++ | S04-declarations/my.t:
Add test for RT #121807
 21:36
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=121807
21:37 rurban joined
Hotkeys welp 21:39
tried to install Task::Star
Template::Mustache failed the tests
Why not install all the other modules after that panda
Why stop
FROGGS these are just dependencies of Task::Star 21:40
21:40 kmel joined
FROGGS and panda does not know that you can live without Task::Star itself, but with most if its deps 21:40
Hotkeys Ah 21:41
21:42 TEttinger joined
FROGGS dunno if the --force option would change anything here 21:43
21:45 zakharyas joined
Hotkeys I amrunning it with --notests 21:46
probably not a great idea
but I'll deal with the consequences
21:46 f33r3x1 joined 21:48 rarara__ left 21:49 kjs_ joined 21:50 rarara__ joined
Hotkeys it still got angry 21:50
different reason this time
21:53 sufrostico left, sufrosti1o left 21:54 Ven joined 21:56 Ven left 21:59 Alina-malina left
dalek osystem: 808d856 | RabidGravy++ | META.list:
Add Unix::Groups

  github.com/jonathanstowe/Unix-Groups		 21:59
21:59 s_kilk left 22:01 Ven joined 22:03 sufrostico joined, Alina-malina joined 22:04 kmel left 22:05 [Sno] joined 22:06 zakharyas left 22:10 virtualsue left
RabidGravy Well that's so dumb I could have had a five year old make it, but I couldn't wait until I found one 22:10
Hotkeys Anyone have any idea what's going on here pastebin.com/UqQWwPX4 22:11
I'm on windows 10 if that helps
this is with panda & rakudo installed via rakudobrew
jdv79 a 5 yo that knows perl6?
seems rarer than a you 22:12
22:13 FROGGS left 22:14 ^elyse^ left 22:15 Oatmeal left
raiph HotKeys: mebbe github.com/MoarVM/MoarVM/commit/808b772e66 and cousins? 22:16
Hotkeys mebbe 22:17
so should I just wait a little built and rebuild and try again 22:18
22:19 kjs_ left
Hotkeys stupid hexchat is refusing to let me just click that link 22:21
had to copy and paste it 22:22
so much work
(/s)
22:24 kjs_ joined 22:25 Ven left 22:26 Oatmeal joined 22:27 bjz left
RabidGravy 101 distinct authors in the ecosystem now, looking good 22:28
Hotkeys I need to contribute so I can be #102 22:32
RabidGravy better hurry up as it's definitely accelerating 22:33
Hotkeys I have no idea what to contribute 22:34
I am but a babby
to perl6
raiph HotKeys: also see irclog.perlgeek.de/moarvm/2015-10-28#i_11446544 -- mebbe pop on #moarvm and ask jnthn if your issue is related to his work? 22:40
22:41 _dolmen_ joined
Hotkeys will do 22:41
thanks
22:43 bacek joined
bacek aloha 22:43
Hotkeys what up
dalek rl6-most-wanted: e82d954 | RabidGravy++ | most-wanted/modules.md:
Add a couple more WIPs
 22:44
bacek p6: say 3;
camelia rakudo-moar 07f186: OUTPUT«3␤»
RabidGravy Hotkeys, if you're looking for ideas then there's a load of things there ^ (the ones with WIP are being worked on)
psch m: say ++$ + [*] $++ + ($++ + ($++ ** $++) xx (++$ + ++$)) 22:46
camelia rakudo-moar 07f186: OUTPUT«3␤»
psch whoo line noise \o/
RabidGravy the horror 22:47
22:47 kaare_ left
RabidGravy y'see the obfu in Perl 6 is even more bonkers than Perl 6 as there are more operators ;-) 22:48
psch yes, Perl 6 obfu is worse than Perl 6 obfu
RabidGravy than Perl 5
psch :D
RabidGravy: actually, i think this one works almost the same in perl5, you'd just have to use $| instead of $ 22:51
at least if i remember $| correctly 22:52
oh, and use x instead of xx
leont $| has a meaning, but you can abuse most magical variables for such a purpose
psch err, and [*] is obviously not there
so i guess that's the one "more operator[...]"
RabidGravy so when the least strong beer in the house is "Gothic Imperial Stout" you just know it's going to go pear-shaped 22:57
psch is cheap 22:58
d9gkwhfwam31p.cloudfront.net/00170...ghtbox.png is my current beverage
but hey, it's obvious what the ABV is at least
leont: other magicals also share the "can't increment beyond 1" of $|? 22:59
leont: i expect at least $[ to not work there 23:00
23:02 skids joined
Hotkeys I've been thinking about brewing a mead 23:02
although I've never had mead 23:03
how bad of an idea is that
mead just seems really easy to make
takes a long time
but not much work
RabidGravy I've got some weird Polish mead somewhere in the house, it scares me 23:04
Hotkeys it's basically honey, water, yeast, and then stuff for flavor if you want
stick it all in a carboy
wait a month or so
move it to another container
age to preference 23:05
drink
23:06 f33r3x1 left
RabidGravy I haven't the patience for home brewing 23:07
I made wine while I was at college thirty odd years ago, but we ran out of booze and started drinking it before it was ready 23:08
Hotkeys lol
i've heard mead is hard to drink before 6 months and is best after a year or more
RabidGravy these were in the days before you could get booze any time you want 23:09
Hotkeys I wasn't even alive thirty odd years ago lol 23:10
leont Yeah, mead is much better if you're patient 23:11
23:11 ChoHag left
leont Given good honey, you don't need stuff for flavor, really, but good honey can be disappointingly expensive 23:12
RabidGravy or get your own bees
leont That helps, though it still matters where you put them
RabidGravy and a good garden
Hotkeys I kind of want to try making an acerglyn mead 23:13
23:13 ChoHag joined
Hotkeys (mead w maplesyrup) 23:13
RabidGravy we've been growing lavender, the bees love that (we don't keep bees but there are lots of them around here)
Hotkeys to be patriotic and all
23:16 raiph left 23:17 CJKinni_ joined 23:20 sufrostico left 23:21 CJKinni left, BenGoldberg joined 23:34 kjs_ left 23:35 vendethiel left 23:40 MadcapJake left 23:41 rarara_ left
zengargoyle doc.perl6.org/type/Int .lsb and .msb seem to be a bit confused / confusing 23:44
both cound from the right, and it's last/first 1 in the binary number. correct? 23:45
*count 23:46
23:46 rarara__ left
RabidGravy dunno 23:49
TEttinger do right and left mean anything in binary? 23:51
it's just least significant and most significant
the explanation makes sense 23:52
Hotkeys it seems like they just return the position of the least and mostsignificant bits
TEttinger err, the demo code
Hotkeys as lsb and msb imply
yeah the demo code does 23:53
the explanations given seem a bit wonky though
zengargoyle .lsb Otherwise returns the zero-based index from the right of the first 1 in the binary representation of the number. 23:54
it's the last 1 in the binary representation
.msb Otherwise returns the zero-based index from the left of the first 1 in the binary representation of the number.
Hotkeys not a fan of the terms first and last
zengargoyle the index is from the right just like .lsb
Hotkeys lsb gives you the (zero-based) position, from the right, of the least significant bit 23:55
msb does the same for the most significant
as shown by demo code
those explanations are bad
in the doc
it seems like whoever wrote them was trying to avoid the terms most and least significant 23:56
but did so poorly
23:57 n0tjack joined 23:58 Celelibi joined
zengargoyle i'd fix it, but i'm not sure of the desired terminology, so just pointing it out. :) 23:58
Hotkeys yeah I'm not sure how to word it well either
as long as one looks at the demo code they should understand it :p

Powered by Cro and the Raku® Programming Language.

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

Thank you!