»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
[Coke] yes, way too much. :) 00:04
in App::Uni, I currently have: 00:05
00:05 ZofBot left
[Coke] $ PERL6LIB=lib bin/uni -c ¶ 00:05
¶ - U+0000B6 - PILCROW SIGN [Po]
Just wondering if there is a precedent for displaying the [Po] like that. 00:06
AlexDaniel just yoleaux and unicodable, that's it I think 00:07
00:08 mcmillhj joined, ZofBot joined, ChanServ sets mode: +v ZofBot, kurahaupo_ left 00:09 ZofBot left 00:10 ZofBot joined, ChanServ sets mode: +v ZofBot, ZofBot left 00:14 mcmillhj left 00:15 mcmillhj joined 00:20 ZofBot joined, ChanServ sets mode: +v ZofBot, mcmillhj left 00:21 wamba left
[Coke] zef bug? I have an installed version of App::Uni - I added a sub in the package; when I run the test in my git checkout, it's fine. if I do "zef install ." with an updated .json file (new version to force an install), it complains it cannot find the new sub. 00:22
ZofBot [Coke], zef bug? I have an installed version of App::Uni - I added a sub in the package; when I run the test in my git checkout, it's fine. if I do "zef install ." with an updated .json file (new version to force an install), it complains it cannot find the new sub.
[Coke] zoffix, fix yer bot.
feels like it's running the tests during the install against the installed lib version, not the new one. 00:23
ZofBot [Coke], feels like it's running the tests during the install against the installed lib version, not the new one.
00:23 ZofBot left
[Coke] Are you just going to keep repeating everything I say? 00:23
00:25 ZofBot joined, ChanServ sets mode: +v ZofBot
[Coke] ... weird. 00:25
timotimo should we give +v to a bot that does nothing but echo? :)
[Coke] .seen ugexe 00:26
yoleaux I saw ugexe 1 Apr 2017 05:48Z in #perl6: <ugexe> openssl bumped to 0.1.8 with MasterDuke's leak fix
00:26 kurahaupo joined 00:27 ZofBot_ joined, ChanServ sets mode: +v ZofBot_, ZofBot_ left, ZofBot left, ZofBot joined, ChanServ sets mode: +v ZofBot 00:29 ZofBot left 00:30 ZofBot joined, ChanServ sets mode: +v ZofBot 00:31 ZofBot left, ZofBot joined, ChanServ sets mode: +v ZofBot
[Coke] perl6 -MApp::Uni -e 'say uni-gist(0x1234)' 00:31
ሴ - U+001234 - ETHIOPIC SYLLABLE SEE [Lo]
there, that's why I did the stupid thing. \o/
00:32 ZofBot left 00:33 kurahaupo_ joined, skink left
[Coke] (though having the uni command means I don't really need it) 00:33
00:34 kurahaupo left 00:35 mcmillhj joined, pelev joined 00:36 ZofBot joined, ChanServ sets mode: +v ZofBot
ugexe [Coke]: I dont think its possible for the tests to use a installed copy of something that is in a dist that will be tested 00:36
00:36 ZofBot left 00:37 ZofBot joined, ChanServ sets mode: +v ZofBot
ugexe because they all get ran with `perl6 -Ilib ...` essentially - so lib/ is always in front 00:37
that sounds like a precomp thing
00:37 ZofBot left 00:38 ZofBot joined, ChanServ sets mode: +v ZofBot 00:40 ZofBot left
ugexe the output with RAKUDO_MODULE_DEBUG=1 will show you where each module gets loaded from 00:40
00:40 kurahaupo_ left, pelev left, mcmillhj left 00:42 cpage left, cpage__ joined, ZofBot joined, ChanServ sets mode: +v ZofBot 00:43 ZofBot left 00:44 kyan left 00:46 zapwai joined 00:51 kyan joined, ZofBot joined, ChanServ sets mode: +v ZofBot 00:54 mcmillhj joined 00:56 cyphase left 00:59 mcmillhj left
[Coke] ugexe: I see "Late loading 'App::Uni'" - then it eventually loads $HOME/.perl6/precomp/UUID 01:00
ugexe can you send me the gist of all of it?
but looks like if you delete $HOME/.perl6 it should work? 01:01
[Coke] gist.github.com/coke/14ec1086f7624...d73dd440cf
Sure, but why should it be picking that up?
ugexe im not sure, i've had it happen before but never consistently 01:02
[Coke] ok. removing ~/.perl6 cache...
ugexe so its not a common state to get into, but somehow...
01:02 labster left
[Coke] same error. 01:03
01:03 kyan left
[Coke] now loading App::Uni from /Users/williamcoleda/.rakudobrew/moar-nom/install/share/perl6/site/precomp/UUID 01:04
too many precomp dirs.
and why it's not getting invalidated, Iunno 01:05
zef install --force App::Uni - letting it complete, it installs, but all the tests fail (because I installed an intermediate version); running the command again, it works. 01:07
(that is, all the tests pass the second time)
unfortunately, the git repo I had locally that allowed me to get into this state no longer exists, I rebased it. 01:08
but I can probably come up with a scenario that fails again.
ugexe all the tests fail for me too 01:12
[Coke] .tell rjbs github.com/coke/p6-uni
yoleaux [Coke]: I'll pass your message to rjbs.
ugexe ahh i see. this is like before i think... you have to pass in the includes to run/shell 01:13
01:13 mcmillhj joined
ugexe in your tests 01:13
[Coke] the PERL6LIB isn't passed through to the sub-perl6, ah. 01:15
oh. you're not using the env var. 01:16
ok. I'll see if I can figure out how to do that. Danke.
ugexe my @include = $*REPO.repo-chain.map(*.path-spec)) should work 01:17
01:18 mcmillhj left
[Coke] ugexe++ 01:22
ugexe github.com/gfldex/perl6-typesafe-x....pm#L9-L10 [Coke] this is a good example
[Coke] I will have to digest that. Thanks. 01:23
nite
ugexe later - also you know how to unrevert this now :) github.com/perl6/doc/commit/445dc7...eca924eb7b 01:24
01:29 mcmillhj joined 01:33 mcmillhj left 01:42 bdmatatu joined 01:43 bdmatatu left 01:45 ilbot3 left 01:47 xfix joined 01:48 ilbot3 joined, ChanServ sets mode: +v ilbot3 01:50 mcmillhj joined 01:55 mcmillhj left 02:16 pochi left 02:28 noganex_ joined 02:31 noganex left 02:33 gdonald left, gdonald joined 02:38 mcmillhj joined 02:42 vendethiel- left, mcmillhj left 02:43 vendethiel joined 03:06 aindilis joined 03:23 cyphase joined 03:24 sammers left 03:25 mcmillhj joined 03:29 nightfrog joined 03:30 mcmillhj left 03:35 drbojingle joined
samcv [Coke], i think i may want to use freebsd's memmem on macos as well 03:36
03:36 perlnecrohacker left 03:37 kyan joined
samcv [Coke], would you be able to bench something for me on mac? 03:38
03:38 perlnecrohacker joined, perlnecrohacker left
samcv well just run simple test and compare between two commits 03:38
03:40 perlnecrohacker joined, perlnecrohacker left 03:41 perlnecrohacker joined, perlnecrohacker left
llfourn samcv: I can bench again soon if you give me a gist. Can't do it right now though. 03:42
samcv yay!
ok llfourn github.com/samcv/MoarVM/tree/memmemindex_test
03:42 skids left
llfourn ^ let you know when I've done it :) 03:42
samcv new test branch. need benching between the most recent commit it memmemindex_test and the one before it
it's ready to clone
llfourn rgr that
03:42 Cabanossi left
samcv so most recent commit in that branch is "Commit to test speed between native and FreeBSD memmem" so need that compared to the previous commit (previous commit uses a system's native memmem) 03:43
i'm expecting the most recent commit to be faster
03:43 perlnecrohacker joined, perlnecrohacker left
samcv say 'a' x 1000000000 ~ 'b' ~~ /abcde/; say now - INIT now; and then also try say 'a' x 1000000000 ~ 'b' ~~ /a/; say now - INIT no 03:43
evalable6 samcv, Full output: gist.github.com/3d409ce26b1b5bd525...038840de23
(exit code 1) 04===SORRY!04=== Error while compiling /tmp/NfX2Ny1ToV
No vali…
samcv not for you evalable6 03:44
geekosaur heh
03:45 Cabanossi joined, perlnecrohacker joined, perlnecrohacker left 03:46 mcmillhj joined, perlnecrohacker joined, perlnecrohacker left
u-ou hello o/ 03:47
03:47 perlnecrohacker joined 03:48 perlnecrohacker left 03:49 labster joined 03:50 khw left 03:51 mcmillhj left
samcv glibc uses knuth-pratt-meyer string search. modern freebsd uses Two-way string-matching as described in this paper cry.nu/files/Two-way%20string-matc...201991.pdf Crochemore, perrin 1991 03:52
u-ou what's memmem? is there a memememe?
samcv memmem finds blocks of memory inside memory
ilbelkyr it's to strstr what memcpy is to strcpy I guess
samcv so it will make our regex and string indexing faster
strstr is just a defective memmem that cannot search things that contain null bytes :P 03:53
03:54 zapwai left
samcv Abstract. A new string-matching algorithm is presented, which can be viewed as an intermediate 03:55
between the classical algorithms of Knuth, Morris, and Pratt on the one hand and Boyer and Moore, on
the other hand. T
hmm
interesting :)
04:00 AlexDaniel left
samcv oh llfourn make sure to do a make clean before testing the other one 04:02
04:04 mcmillhj joined 04:06 Limit left 04:09 mcmillhj left
samcv timings for me: /abcde/ GLIBC 4.29s FBSD 3.78s 04:10
and /a/ GLIBC 3.43s FBSD 3.36 04:11
avg of 3-4 runs
will be cool to see what the macos results are :)
04:12 Cabanossi left 04:14 Cabanossi joined
samcv and percentages: glibc 13% faster with /abcde/; fbsd 2% faster with /a/ (probably statistically meaningless 04:14
geekosaur either I am not understanding your timings or one of those numbers is wrong? 04:15
samcv hmm? 04:21
TEttinger yeah, you gave numbers wehere FBSD took less time for /abcde/ 04:22
samcv oh
that's reversed
my bad
glibc was slightly faster with abcde
TEttinger how many runs are you doing? 04:23
samcv 4
very constant times btw
all very close
err tighty grouped that is
TEttinger I may have missed something, but that's not 4 string finds that take 3-4 seconds I hope :) 04:24
geekosaur generally it's fast enough that you do a run of 10000 or so
samcv well it's 'a' x 1000000000 ~ 'b' ~~ /abcde/
so it's a pretty long string
geekosaur ah
TEttinger I wonder how much of that is overhead from creating a billion-char string 04:25
04:26 gdonald left
samcv maybe a lot 04:26
04:26 gdonald joined
samcv err not that much i guess 04:27
i get the same time from my $var = 'a' x 1000000000 ~ 'b'; my $t1 = now; $var ~~ /b/; say now - $t1
gonna compare just the raw nqp::index function now 04:34
though i think the actual regex is probably better representative 04:35
fbsd looks a bit faster with just nqp::index starting at 0. but glibc def has an edge in actual perl6 regex 04:38
llfourn, you done tests yet?
04:42 mcmillhj joined, drbojingle left 04:47 mcmillhj left 04:53 mcmillhj joined 04:58 mcmillhj left, mcmillhj joined 05:00 rurban joined, rurban left
samcv llfourn, you still alive ;) 05:01
05:03 mcmillhj left
u-ou :) 05:09
05:10 mcmillhj joined 05:11 Cabanossi left
BenGoldberg The fastest libraries for doing regex are going to be dfa-based ones, like re2 and ragel. 05:11
u-ou I think perl 6 grammars are a really really good idea 05:12
there i said it
samcv yay! 05:13
and they are nice 05:14
u-ou I also like twigils
05:15 Cabanossi joined 05:16 mcmillhj left 05:24 mcmillhj joined 05:29 mcmillhj left 05:31 kyan left 05:35 mcmillhj joined 05:44 faraco joined 05:57 Cabanossi left 05:59 Cabanossi joined 06:05 xinming left 06:07 xinming joined 06:13 wamba joined 06:18 dwarring left 06:28 ChoHag joined 06:31 rurban joined 06:33 jdmmmmm left 06:34 wamba left
moritz good morning 06:37
(off-topic) I've cobbled together a cover image for my first book: leanpub.com/deploy feedback welcome!
my wife says it looks like a vacation guide, so I'm looking for something more abstract now... :-) 06:40
mst I actually quite like that cover 06:41
06:41 Cabanossi left
mst this almost certainly means you should replace it ;) 06:41
moritz :-) 06:42
mst: thanks
samcv good morning 06:43
moritz \o samcv
samcv your wife is not wrong
moritz she usually isn't, when it comes to that sort of thing :) 06:44
samcv maybe something in black and white. like archetecture of buildings or something
06:45 Cabanossi joined
samcv or even something like this www.thepluspaper.com/wp-content/upl...prints.png 06:47
the piece was supposed to have symbolic meaning moritz ?
06:48 ufobat joined
moritz samcv: in a very abstract/stretchy way, yes 06:49
samcv active8studios.co.uk/structures
maybe one of these
i like these bridges
moritz nice idea, thanks 06:50
afk, sorry&
TEttinger something continuous... stretchy... salep dondurma
samcv something about these bridges. active8studios.co.uk/wp-content/upl...ridge1.jpg they are nice
mst moritz: INVENT A MOEBIUS TRUCK
samcv and i guess you can buy them
TEttinger www.ediblegeography.com/wp-content/...ecream.jpg 06:51
continuous mustache deployment
samcv i want one of those bridges now...
damn
even got wireframe boards and ships active8studios.co.uk/boats-ships 06:53
damn it. people stole my bridge idea www.amazon.com/s/ref=nb_sb_noss_2?...evelopment 06:57
and i had thought about. well maybe a bridge with a boat to be more original. and there's some 99 cent kindle ebook with a bridge and a boat 06:58
06:58 faraco left
TEttinger continuous deployment... a picture of a map of iraq. 06:58
samcv and this one is def awesome www.amazon.com/Continuous-Architec...evelopment 06:59
though a bit ambitious with the cover
what about racks of servers. but photoshopped to go on forever ititch.com/wp-content/uploads/2012...server.jpg 07:00
continuous etc 07:01
just to infinity
that could be neat and original
(maybe)
as long as there aren't contiuous deployment books with that then maybe it would be ok... idk.
maybe just a still from the movie Hackers wallpapersafari.com/w/VKJpHR/ 07:02
why not
oh. what about. windmills 07:03
because sustainable power
also they look nice
huf "tilting at windmills" is perhaps not what you want to suggest
samcv and they spin and aren't cliche like gears or like. something else
that's from a made up story though 07:04
huf ducks dont often grow 7 feet tall.
TEttinger ooooh a space filling curve
samcv not like old windmills. but the new ones
huf oh 07:05
samcv www.crystalinks.com/windmills.jpg look how big the blades are
idk. windmills turning is soothing
TEttinger here's a space filling curve I had fun making www.dropbox.com/s/w6uvhtsz30d8msr/...e.png?dl=0
samcv how did you make it. program? 07:06
TEttinger magicavoxel, hand-placed voxels, it rendered it
samcv i don't understand what windmills. www.miltwright.com/products/images_...ls2016.jpg has to do
07:06 BenGoldberg left
samcv with this topic. 07:06
TEttinger don quixote is a well-known classic tale 07:07
windmills, specifically the mill part, are "legacy hardware" 07:08
since most milling is done by more efficient machines
samcv even this type of windmill. which is pretty awesome discovermagazine.com/2008/oct/24-hi...ed-cousins
yeah i've read it
though i've never heard that phrase used before. other than in the booj
*book
TEttinger hm
samcv wind turbine 07:09
TEttinger I think you're on the righ track though
samcv i guess i mean
TEttinger engine of progress is a term
wheels of progress
samcv how about
the international space station
TEttinger turbines are sorta engines
heh
samcv *mindblown*
that could be cool
though idk maybe could make people feel distant and alone 07:10
but it's very much continuous delivery
TEttinger that curve image is for a 5x5x5 curve, which is one of zillons possible, but the group has interesting properties if you need a tight mapping of points in a 3D cube to items in a collection that has a max size that's a poor fit for more normal curves like the Hilbert curve 07:11
07:11 ChoHag left
TEttinger since a 2x2x2 uses 8 points or a power of 8, and this uses 125 points or, oddly, 125 times a power of 8 (due to how you can lay them out next to each other. continuously.) 07:12
samcv just have a 7-Eleven store on the cover 07:14
24 hour pharmacy
TEttinger ha totally
07:14 TimToady left
TEttinger there's some fun math stuff in the space-filling curve area. it's pretty crazy to me how there's always a relationship between hilbert curves and gray codes, despite them being discovered in completely separate fields and decades apart 07:14
also, that applies in arbitrary dimensions, which is the real mind blowing part 07:15
07:16 TimToady joined 07:30 domidumont joined
samcv nice 07:31
that sounds neat
there are many things in the natural world that are discovered in mathematics first and then end up getting discovered in physics and in nature later. because there is a natural structure of the universe and the rules that govern it can be seen many places 07:32
07:36 domidumont left, domidumont joined
samcv i took a class in university called symmetry and aesthetics in contempory physics, which was very interesting, in addition to many times stumbling upon the same algorithm seen in multiple places by *chance*, mathematicians and physisitss have often seeked symmetry, order, and elegance in their studies 07:36
one of my favorite classes I ever took I think 07:37
llfourn samcv: going to test now if you still want it? :) 07:42
samcv yes very much
i missed you 07:43
llfourn ^_^
samcv llfourn, sec lemme get you what to dtest 07:45
llfourn, gonna make it easy for you 07:49
07:50 avalenn left
samcv ok llfourn gist.github.com/2aac7ec4ad19129365...bcb6a4aae4 you can just run this 07:55
llfourn samcv sure
samcv just run it as a script and it will label each time seperately from each other
prolly will take 1 min per run 07:56
oh uh 07:57
for i in $(seq 1 4); do perl6 stringregexbench.p6 > /dev/null; done
and it will only output the times and not the results of the regex
since times go to stderr 07:58
llfourn kk :)
btw I am recompiling moar and then doing perl Configure.pl --gen-nqp --backends=moar
samcv you don't have to recompile nqp or rakudo
llfourn i guess I should recompile nqp each time?
samcv just moar
llfourn oh ok 07:59
samcv yeah
recompiling nqp you need to compile rakudo. but moarvm is standalone
since it just reads bytecode
llfourn yeah I guessed what you were doing is just needed at runtime but wasn't sure 08:00
08:02 ChoHag joined 08:05 rindolf joined 08:08 robertle joined
llfourn samcv: preliminary results: gist.github.com/LLFourn/f39625d7c0...4eda7bc3b8 08:11
looks like FreeBSD wins? 08:12
08:12 domidumont left
samcv yeah i knew that for macos. but geekosaur maybe i did have those backwards 08:13
cause i'm getting faster for freebsd compared to glibc 08:14
which is fine with me actually because the freebsd one is more easily integratable abort from libc. glibc was gonna be annoying if i wanted to use their library across platforms 08:15
and take the code out and use it for mac or windows 08:16
08:16 rindolf left, RabidGravy joined
llfourn \o/ 08:17
geekosaur you'd have a license problem anyway, don't think you could do that and relicense as dual gpl/artistic 2.0 08:18
samcv ok maybe i was wrong. i think that test was off for whatever reason. dunno why all results were slower 08:20
though libc may have a little edge on glibc. 08:23
so maybe it would just be easier to just use it for all versions
08:24 rindolf joined
llfourn gist.github.com/LLFourn/f39625d7c0...4eda7bc3b8 # updated with seq 1 4 results 08:26
samcv overall they seem very close between glibc and libc.. i'm inclided to keep my pull request as it stands now. 08:30
Woodi_ hallo #perl6 :) 08:31
samcv hello Woodi_
since glibc seems to do better with more complex needles which may be more likely to happen in practice. 08:33
geekosaur so, uh, did you even think about the license situation I mentioned? 08:36
samcv the GPL one?
or MIT one which freebsd has
i was not aware GPL was incompatible with Artistic 2.0 08:37
though i knew MIT was definetly ok
geekosaur gpl coe forces moarvm to be gpl instead of artistic 2.0
github.com/MoarVM/MoarVM/blob/master/LICENSE 08:38
Woodi_ moritz: #deploy picture is realy nice but looks dreamy/impresionistic. a lot of "sharpness" would be much more "classic" for IT book :) but "rack of servers" sounds realy nice for a topic. anyway, cover does not matter if book is good, could be plain black or white with some strings :) 08:39
samcv yep reading gpl/artistic texts makes that clear to me geekosaur. 08:41
i think i misthought cause i think perl is under artistic and GPL 08:42
so thought they were compatible mistakingly
geekosaur even that's not good enough because any contributed code that is gpl license can't be used under artistic license 08:44
the creator of a work can choose to dual-license. but since you took that code from glibc, you're bound by glibc's license
samcv yes i know
i mistakenly though tthey were compatible becuase perl was gpl and artistic. thank you for helping point that out to me 08:45
moritz mst: I like the idea of using a möbius strip, maybe something like this: www.flickr.com/photos/maxf/7692061042 08:48
RabidGravy Boom!
samcv i like that moritz
can it be a boat sailing on a mobius strip? heh 08:49
or this s-media-cache-ak0.pinimg.com/736x/...4b792b.jpg 08:51
"A Computer Mystery Adventure" 08:52
08:52 camelia left 08:53 camelia joined
samcv what's the craziest programming/computer book cover you guys know of? 08:53
RabidGravy what was that series of books with the people in traditional costumes? 08:54
08:54 ChanServ sets mode: +v camelia
samcv this one is not bad s-media-cache-ak0.pinimg.com/236x/...ac4d97.jpg 08:54
Woodi_ samcv: all that animals on O'Raily books ? :)
samcv i would buy that book
they're just animals 08:55
there's gotta be some super wierd cover
RabidGravy visualising hell sounds like my day job ;-)
samcv upload.wikimedia.org/wikipedia/en/...mcover.jpg
08:55 nine left
samcv oh no... 08:55
08:57 cognominal joined
RabidGravy images-na.ssl-images-amazon.com/im...3,200_.jpg - these ones, what is that all about? 08:58
also the "Beginning $X" series with really appalling pictures of the authors on the front
08:58 nine joined, camelia left 08:59 camelia joined 09:00 ChanServ sets mode: +v camelia, cog_ left 09:02 cognominal left 09:04 ChoHag left 09:10 AlexDaniel joined 09:11 ChoHag joined, ChoHag left 09:15 sammers joined 09:17 ufobat left, mxco86 left 09:21 kurahaupo joined, darutoko joined
gfldex samcv: i2.wp.com/gfldex.files.wordpress.c....jpg?ssl=1 09:24
moritz yes, that *is* weird
09:26 ChoHag joined
RabidGravy I think we have a winner 09:26
somewhere in the house I have a book like "learning FORTH and COBOL in parallel" which is a weird concept in itself 09:27
and in that genre of totally not right on and inappropriate computer covers flashbak.com/yugoslavian-computer-m...0s-370271/ 09:29
09:29 ChoHag left 09:34 ChoHag joined 09:36 AlexDaniel left 09:37 kurahaupo left 09:41 kurahaupo joined 09:42 ChoHag left 09:46 kurahaupo left 09:52 domidumont joined, andrzejk_ joined 09:53 ChoHag joined 10:02 ChoHag left 10:03 ChoHag joined 10:12 Cabanossi left 10:14 Cabanossi joined
andrzejk_ hey 10:17
:)
who is here?;s I need help much 10:18
10:24 TEttinger left
RabidGravy you just think you do 10:32
andrzejk_ help me please 10:37
RabidGravy ? 10:39
RabidGravy ASK THE QUESTION
andrzejk_ RabidGravy man don't be anger I just thought that noone here today 10:40
RabidGravy I am looking for details or history how rakudo was built
RabidGravy I'm not angry :) 10:41
it's just better to ask the question then anyone who comes along can see it and answer
anyhow I'm off out shopping :) 10:42
lizmat
.oO( it all started at a CamelHerders meeting long ago...)
10:45 kurahaupo joined
andrzejk_ lizmat is it some kind of council of the oldest perlers? 10:45
lizmat it used to be the name of the yearly meeting of p5p in person at OSCON (well, it was at least the one time I attended in 2000) 10:46
andrzejk_ ohh
lizmat and jonathan was there?
10:47 AlexDaniel joined, labster left
lizmat Jonathan wasn't in the picture yet on account of being too young, I think :-) 10:48
10:48 ufobat joined
andrzejk_ lizmat ohhh how that happend that Jonathan take a care of perl6 vm? 10:48
lizmat phew... that's a long stpry 10:49
one can write a book about that, and maybe someone will :-)
well, I hope someone will... :-)
timotimo jntn's bgrphy?
andrzejk_ huh 10:50
lizmat are there some documents about perl6 vm how to start or smth?
lizmat hehe... no, more like en.wikipedia.org/wiki/The_Soul_of_a_New_Machine
timotimo you should read 6guts.wordpress.com, it's where jnthn blogs
andrzejk_ huh he is doing it from 6 years :S 10:52
timotimo he had a blog on use.perl.org, but the link is dead
andrzejk_ humm 10:53
what about language design documents
design.perl6.org/
these looks very experimental
sammers hi all 10:54
timotimo at the beginning there were only the synopses
no, wait, the apocalypses might have been first?
lizmat yeah, wass about to say
timotimo right. first larry wrote the apocalypses, then damian conway wrote the exegesis 10:55
and from that combination of work the synopses were born
andrzejk_ what happend next? 10:58
lizmat many things were tried 11:00
11:00 ChoHag left
lizmat most failed, but we learned from it, and some left a nice legacy 11:00
such as the pugs project, that gave us roast 11:01
www.perlfoundation.org/perl6/index.cgi?pugs
www.perlfoundation.org/perl6/index.cgi?timeline # a timeline to 2010 11:02
andrzejk_ omg 11:03
17years
timotimo you didn't know that? :) 11:05
andrzejk_ no
timotimo but it looks like all vms were created by 1 person in most 11:06
mainly( 11:07
lizmat I guess that proves TimToady's point that it doesn't matter how long it took once it is there :-)
timotimo parrot was started by two
the beginning of moarvm was mostly jnthn, then diakopter came in 11:08
then a bit of japhb 11:09
and then jimmyz 11:10
11:10 pmurias joined, setty1 joined
timotimo more and more names appeared in the git log "4 years ago" 11:10
gotta go run an errand
andrzejk_ what about ideas 11:11
did they took ideas from old vms
and pick up them to new?
timotimo well, 6model was taken from parrot
and it has a GC, you can't have any new ideas in a GC
every single idea in GC has already been had in the late 1980 11:12
(well, not 100%, but at least 90%)
lizmat www.amazon.com/Garbage-Collection-...1420082795 # fresh GC ideas 11:13
and inspiration to MoarVM
andrzejk_ ye I heard about this book
timotimo "The Garbage Collection Handbook: The Art of Automatic Memory Management brings together a wealth of knowledge gathered by automatic memory management researchers and developers over the past fifty years."
not sure how fresh that makes it :D
andrzejk_ what about llvm? 11:14
why perl vm developers didn't took it?
11:15 AlexDaniel left
timotimo if you want to use llvm you have to design your program around llvm 11:16
you have no freedom any more, so to speak 11:17
and it was deemed much too low level for us
andrzejk_ what do you mean? 11:20
timotimo brrt-to-the-future.blogspot.de/2015...e.html?m=1
this post discussed llvm a bit
11:23 mr-foobar left
timotimo hope that helps at all? 11:24
11:24 mr-foobar joined
El_Che andrzejk_: are you writing an article? 11:26
andrzejk_ no
11:29 mingdao left 11:38 kurahaupo left
timotimo i might have actually been wrong about "everything was already invented in the 1980", and some interesting new stuff may exist 11:41
El_Che timotimo: yes, grandad :) 11:42
11:42 Cabanossi left
timotimo grand dad?! grand dad?!?!? what the *** ;) 11:43
andrzejk_ how to get perl6 language specification 11:44
anyone know it?
11:45 mscha joined, Cabanossi joined
mscha This may be a dumb question, but it seems like class Foo { has $bar } is valid. What's that supposed to do? It seems to be equivalent to class Foo { has $!bar }. 11:46
timotimo andrzejk_: the language is specified by the test suite 11:47
git clone [email@hidden.address]
jnthn mscha: It's a sop to people who don't like sigils. ;) It installs a lexical alias, meaning that $x will compile into $!x within the body of the class (unless you hide it) 11:48
lucs And before the beginning, the RFCs.
jnthn Uh, don't like *twigils*
:)
sammers can someone confirm a zef install error for me?
mscha OK, thanks.
timotimo we can try. what's wrong?
jnthn It hasn't been widely used, though :)
andrzejk_ what some ebnf definitions?
sammers I am getting a few failures on several modules, Bailador is one... 11:49
timotimo andrzejk_: don't have 'em
lucs (oops, a bit late -- note to self: make sure view is current when posting)
timotimo we do have the std grammar, though. which hasn't been updated in multiple years
(rakudo has instead been brought much closer to the std grammar)
sammers I just want to see if anyone else sees the same issue
andrzejk_ OK I NEED TO GO KIDS ;S
jnthn andrzejk_: If you want a formal specification of the syntax, then Grammar.nqp in Rakudo is as close as you'll get :)
11:49 kurahaupo joined 11:50 andrzejk_ left
timotimo sammers: can you give us all your versions first? 11:50
sammers also these three,
jnthn lunch &
sammers HTTP::MultiPartParser
Template6
yYTY
11:51 andrzejk_ joined
timotimo aha! 11:51
sammers timotimo: zef 0.1.5
timotimo Template6 relies on hash ordering in one of its tests
sammers ah, that changed recently? 11:52
timotimo no
it was never reliable
sammers ah
timotimo it was always a very bad idea to rely on hash ordering in anything
it's just not very likely you'll get caught, because we don't actively randomize order
which is something we might want to start doing soon 11:53
11:53 mr-foobar left
timotimo because attacks on hash keys is a Real Thing™ 11:53
let's see about making that test better 11:54
11:56 andrzejk_ left
sammers will hash ordering be removed eventually? 11:56
12:00 mr-foobar joined
timotimo hash ordering is not something you can remove 12:00
it's purely an observational thing
sammers so they are expecting the pairs to be in order, but there is no guarantee 12:01
timotimo that's right
sammers so lists are ordered, is there a way to make an unordered-list? shuffle? 12:02
timotimo you can use .pick(*) when you're iterating
sammers m: say ['a', 'b', 'c'].pick(*); 12:04
camelia (a b c)
sammers m: say ('a', 'b', 'c').pick(*);
camelia (b a c)
timotimo made a pullrequest
github.com/supernovus/template6/pull/18 12:05
okay, let's look at the MultiPartParser 12:07
i'd really like it if "zef look" had an option (or maybe a different command) that'd git clone the repository so i could immediately "hub fork", hack on stuff, then "hub pullrequest" 12:08
ugexe: ^ is that in scope for zef?
sammers yeah, there is an an open issue for it... 12:09
timotimo uh oh, that uses JSON::Fast ... i hope i didn't b0rk it
gfldex timotimo: it might be in scope for META6::bin
timotimo geez, the output for the multipartparser's tests are gigantic 12:11
12:12 mr-foobar left 12:13 mr-foobar joined
timotimo dangit 12:15
sammers: HTTP::MultiPartParser being broken is my fault, it's some problem inside JSON::Fast, which i just rewrote a part of
sammers what! 12:16
timotimo it's not backslashing correctly
gfldex always considered JSON as root of some evil 12:17
sammers ok, np, I am just looking for issues related to bailador some of the other web-related modules and assist where I can. thanks for jumping on these and fixing all the stuff you are breaking. 12:18
timotimo haha
sammers ha
timotimo yeah, how do you even properly \\ 12:19
probably not without .trans actually
Geth ecosystem: gfldex++ created pull request #314:
add Slippy::Semilist
timotimo i.e. not by calling .subst multiple times
sammers just add another .subst, it should fix it. 12:20
timotimo no, i don't think so
sammers :)
timotimo if i turn \\ into \ first, then \n into newline, it'll turn \\n into a newline
if i first turn \n into newline, then \\ into \, it'll turn \\n into \ and a newline 12:21
12:21 mscha left
timotimo there goes my hopes and dreams of being fast 12:21
12:22 eroux joined 12:27 espadrine joined, Cabanossi left 12:29 Cabanossi joined
timotimo the worst thing is my tests not catching any problems 12:29
12:32 domidumont left
timotimo bah, backslashes. it's a real hassle to get them right when you've got either a shell or perl6 strings 12:36
12:37 mr-foobar left 12:38 mr-foobar joined 12:43 mscha joined
mscha m: my %foo = 42=>(1,2); say %foo{42}; my @f = %foo{42}; say @f; # huh? 12:43
camelia (1 2)
[(1 2)]
mscha m: my %foo = 42=>[1,2]; say %foo{42}; my @f = %foo{42}; say @f; # huh? 12:44
camelia [1 2]
[[1 2]]
timotimo grr. is-deeply is giving me a fail, but visual inspection says it should pass 12:45
ah, is-deeply is "just" eqv 12:46
lizmat timotimo: I just had that, turned out a Map !eqv Hash :-) 12:48
timotimo interesting. 12:49
i wonder if that applies here.
but i have something mysterious here it seems like
12:49 Actualeyes left
timotimo i'm getting "cannot modify an immuatble Pair" from code like for @(@got[$i]) >>,<< @($exp[$i]) 12:49
12:50 mr-foobar left
timotimo oh, why not use Data::Dump::Tree 12:52
Geth ecosystem: dcb615b605 | (Wenzel P. P. Peppmeyer)++ | META.list
add Slippy::Semilist (#314)

  github.com/gfldex/perl6-slippy-semilist
timotimo huh 12:54
i thought Data::Dump::Tree had a diff functionality
oh yeah 12:55
wow, that module is very unhappy right now
method !set_lookup_table(%lookup, $ansi_code) { for %lookup.kv -> $k, $v { %!color_lookup{$k} = $ansi_code($v) } } 12:56
this line gets:
Cannot find method 'Any' on object of type Str
what the ...? :)
well, the module needs changed for the lexical require stuff anyway 12:57
12:57 mxco86 joined 12:59 mr-foobar joined 13:00 andrzejk_ joined 13:08 ZofBot left
timotimo :\ 13:08
Terminal::ANSIColor isn't cooperating with me here
13:08 ZofBot joined, ChanServ sets mode: +v ZofBot
timotimo timo@schmand ~> perl6 -e '(try require ::("Terminal::ANSIColor")); say ::("Terminal::ANSIColor")::("EXPORT").keys' 13:09
()
why is there nothing in the whole package?
oh, because "is export" only happens when you "use"?
well, wtf is this 13:11
i can ::("&Terminal::ANSIColor::color"), but ::("Terminal::ANSIColor")::.keys gives me () ... ?!?
"Cannot resolve caller AUTOGEN(AnsiColor: ); none of these signatures match" ugh, what now? 13:14
13:16 Actualeyes joined
timotimo ;_; 13:16
the first test checks if the number of lines output is equal to a number
13:17 jdmmmmm joined
timotimo these tests make me thoroughly unhappy 13:18
graaaaaargh 13:25
the .perl outputs for the strings eqv, but not the strings themselves
what is this absolute madness 13:26
13:26 kurahaupo left
timotimo when i .encode('utf8') them, it's also eqv. and .WHAT is eqv between both, too 13:26
13:28 jdmmmmm left
timotimo they even eq to False 13:29
i'll leave this rabbit hole for someone else to figure out, i'd say >:( 13:30
llfourn timotimo: probably a bug in require not replacing the stub package properly 13:31
timotimo but in a one-liner i can use that sub for stuff? 13:32
llfourn hmm actually require Terminal::ANSIColor; say ::("Terminal::ANSIColor").WHO.keys works for me
timotimo that's the other way to require, yeah 13:33
is .WHO.keys different from ::.keys?
llfourn apparently after an indirect lookup it is :D
:: is compile time? 13:34
timotimo m: say "foo"::.keys
camelia 5===SORRY!5=== Error while compiling <tmp>
Confused
at <tmp>:1
------> 3say "foo":7⏏5:.keys
expecting any of:
colon pair
timotimo i wouldn't expect it to be, tbh 13:35
llfourn hmm yeah I don't think it is or at least it doesn't have to be
m: require Test; say Test::.keys
camelia (&todo_output EXPORT &failure_output &output)
13:36 eroux left
timotimo btw 13:37
m: require ::"Terminal::AnsiColor"
camelia ===SORRY!===
MVMArray: Can't shift from an empty array
llfourn m: say ::"Terminal::Ansifoo"
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3say ::7⏏5"Terminal::Ansifoo"
expecting any of:
infix
infix stopper
postfix
statement end
statement m…
llfourn That's weird 13:38
timotimo OK, i got it to work now
do you think you could help me figure out the string eqv thing? 13:39
um, what the heck ... 13:40
llfourn what's the string eqv thing exactly?
timotimo well, i have two strings that are equal in all respects i could find, but they are not eqv
their .WHAT are eqv, encoding them to utf8 gives me equal utf8 arrays, their .perl are also eq, and the strings themselves are also eq 13:41
llfourn how do we produce such strings?
timotimo it happens in the HTTP multipart parser test suite 13:42
wow, i'm a dummy 13:44
there was a method "color" and i added an attribute "color"
llfourn well, eqv for Str looks like it tests .WHERE essentially
timotimo oh? that's not right 13:46
Str is a value type?
llfourn oh no I misread it
it tests the address first then then iseq_s
timotimo OK 13:47
absolutely incredible!
D::D::T now passes its tests!
llfourn \o/ # tests passing for things
timotimo okay, got a pull request 13:48
llfourn goes to start reading moritz++'s devops book 13:49
timotimo i'm glad nkh just touched DDT in the last 2 weeks 13:50
13:54 BenGoldberg joined 13:58 Cabanossi left
masak timotimo: I read "D::D::T" and for one horrible moment I thought type definedness smiley syntax had evolved without me noticing 13:59
14:00 Cabanossi joined
moritz llfourn: thanks for buying! Feedback is always very welcome 14:03
14:07 dmaestro joined 14:08 bjz left
El_Che moritz: GoCD is confusing. I guess it has nothing to do with Go(lang)? 14:08
timotimo masak: :) 14:15
fwiw, Data::Dump::Tree is very cool and you should all check it out
(may need to use my pullrequest to make it work, though) 14:16
(and that pullrequest depends on AnsiColor's pullrequest ... maybe?)
hm. it actually doesn't
14:18 Rawriful joined 14:24 skids joined 14:28 Khisanth left
timotimo huh. one seems to have actually \r\n, the other seems to have a synthetic there 14:32
how the eff did we end up with that?
gfldex github.com/supernovus/exemel is broken too
likely IO related 14:33
timotimo gist.github.com/timo/94421784d6471...9e8d514814 - here's the strings as seen by gdb 14:34
gfldex perl6 -e 'say so IO::Handle.new ~~ IO' 14:36
that's the culprit
timotimo ah, yes
good catch
anyway 14:38
14:38 KongWubba joined
timotimo these two strings have different nqp::chars, but they render out to the same utf8 14:38
gfldex IO is a role, should be not problem to pull that in
timotimo IO will get/has been thrown out
it was useless
gfldex XML.pm6 does use it in a multi as a type check 14:39
github.com/supernovus/exemel/blob/...ib/XML.pm6
line 36 fails now because from-xml-stream does a check against role IO 14:40
timotimo The changes that may affect existing code are:
role IO together with its IO.umask method have been removed without any replacement
that's from the latest rakudo.org post
not latest
second-to-latest
14:41 Khisanth joined
gfldex is there a role now that says 'a IO::Path or IO::Handle i don't care'? 14:42
Geth doc: 460f1e6ba3 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
Remove trailing whitespace
 14:44
doc: 65cc372243 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path.concat-with

  - Does what .child used to do
  - .child will be changed to throw for non-child paths
Rakudo impl: github.com/rakudo/rakudo/commit/966a7e3a10 Tests: github.com/perl6/roast/commit/0e47f25e07		 14:45
timotimo gfldex: github.com/rakudo/rakudo/blob/nom/...discussion 14:47
moritz El_Che: GoCD and golang are not related
timotimo gfldex: github.com/supernovus/exemel/pull/43 14:48
gfldex ok 14:50
4 days ago :-/
15:06 mcmillhj left 15:13 mcmillhj joined
masak m: multi sub B() { say "multi" }; &B.wrap(sub {say before"; callsame; say "after"}); B() 15:18
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3{ say "multi" }; &B.wrap(sub {say before7⏏5"; callsame; say "after"}); B()
expecting any of:
infix
infix stopper
postfix…
masak m: multi sub B() { say "multi" }; &B.wrap(sub {say "before"; callsame; say "after"}); B()
camelia before
multi
after
15:18 mcmillhj left
masak w00t 15:18
(RT #77472)
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...l?id=77472
15:20 zakharyas joined
BenGoldberg m: multi sub B() { say "multi" }; &B.wrap(sub {say "before"; {*}; say "after"}); B() 15:21
camelia 5===SORRY!5=== Error while compiling <tmp>
{*} may only appear in proto
at <tmp>:1
------> 3multi" }; &B.wrap(sub {say "before"; {*}7⏏5; say "after"}); B()
expecting any of:
horizontal whitespace
term
BenGoldberg m: multi sub B() { "multi" }; &B.wrap(sub {"before" ~ callsame ~ "after"}); say B() 15:23
camelia 5===SORRY!5=== Error while compiling <tmp>
Calling callsame(Str) will never work with declared signature ()
at <tmp>:1
------> 3() { "multi" }; &B.wrap(sub {"before" ~ 7⏏5callsame ~ "after"}); say B()
BenGoldberg m: multi sub B() { "multi" }; &B.wrap(sub {"before" ~ callsame() ~ "after"}); say B()
camelia beforemultiafter
BenGoldberg *almost* wishes perl6 did the perl5 parsing thing, where if the parser knows that some function has no arguments, code after it is not processed as a list. But that's one of the things which makes perl5 parsing evil, so ... 15:25
15:27 Cabanossi left 15:29 Cabanossi joined
Geth doc: 24a6ea9a47 | (Zoffix Znet)++ | 5 files
[io grant] Toss all of the TODO methods in IO::Spec*

Don't know who or what added these, but many (most?) of them are wrong (type doesn't provide them) and those that aren't list implementation-specific signatures.		 15:30
15:31 mcmillhj joined
pmurias if there is a test in roast that's fudged for rakudo it means it's part of 6.c but rakudo has a bug? 15:32
Geth doc: 1f75ddcb81 | (Zoffix Znet)++ | 2 files
[io grant] Document IO::Spec*.abs2rel
 15:33
15:34 drbojingle joined 15:36 mcmillhj left, khw joined 15:42 mcmillhj joined 15:47 mcmillhj left
TimToady m: multi sub term:<B>() { "multi" }; &term:<B>.wrap(sub {"before" ~ callsame() ~ "after"}); say B 15:54
camelia beforemultiafter
TimToady BenGoldberg: ^^^ you can, it just spelled differently 15:55
*it's
15:58 kurahaupo joined 16:01 wamba joined 16:03 tomboy64 joined
tomboy64 are there any news regarding rakudo + jvm? 16:04
e.g. is it still broken?
(doesn't build for me since about a year)
BenGoldberg m: multi sub term:<B>() { "multi" }; &term:<B>.wrap(sub {"before" ~ callsame ~ "after"}); say B 16:08
camelia 5===SORRY!5=== Error while compiling <tmp>
Calling callsame(Str) will never work with declared signature ()
at <tmp>:1
------> 3ulti" }; &term:<B>.wrap(sub {"before" ~ 7⏏5callsame ~ "after"}); say B
timotimo BenGoldberg: callsame isn't a term, so it's nomming ~"after" as one argument 16:17
Geth doc: cc62dd240c | (Zoffix Znet)++ | 2 files
[io grant] Kill IO::Path.abspath

Use IO::Path.absolute, which is the same thing
Rakudo removal: github.com/rakudo/rakudo/commit/cb323d5dd3		 16:19
16:21 drbojingle left
BenGoldberg .seen ioninja 16:22
yoleaux I saw IOninja 22 Mar 2017 06:33 EDT in #perl6: <IOninja> But if that's not up to par to the standards of this channel, I can leave.
BenGoldberg .seen zoffix
yoleaux I saw Zoffix 7 Apr 2017 14:12 EDT in #perl6: <Zoffix> BOTS WERE MEANT TO BE FREE!
simcop2387 yoleaux: help 16:23
BenGoldberg yoleaux's commands start with .
timotimo BenGoldberg: find zoffix in #perl6-dev
#perl6 is too noisy for 'em, which i can totally understand 16:24
16:24 noize joined
TimToady m: my &term:<same> = &callsame; multi sub term:<B>() { "multi" }; &term:<B>.wrap(sub {"before" ~ same ~ "after"}); say B 16:30
camelia beforemultiafter
TimToady though, given it's a term, and can have no args, one wonders why it's a multi...
if you were hoping to have different syntax for different multi candidates, that's insane :) 16:31
16:38 espadrine left 16:40 leont left 16:44 skids left 16:51 mscha left 16:55 MilkmanDan left, noize left, MilkmanDan joined 16:58 Cabanossi left, Rawriful left
ugexe timotimo: theoretically yes. i'm just not sure how to generalize that functionality to something that non-git modules could also use. Maybe its a good use for the difference of source-url and `support : { source: "..." }` 16:59
17:00 Cabanossi joined
ugexe curt tilmes says he is close to releasing his perl6 GraphQL module 17:02
Geth doc: 1973010455 | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Document IO::Path.ACCEPTS

N.B. currently refers to undocumented `:completely` arg to .resolve. It's not a mistake; the argument will be implemented and documented as part of IO grant work.		 17:03
ugexe (which is affected by current MultipartParser failure)
17:06 pochi joined 17:13 rrrr_ joined 17:15 domidumont joined 17:21 rrrr_ left 17:22 zapwai joined 17:24 zakharyas left 17:26 Cabanossi left 17:29 Cabanossi joined 17:31 andrzejk_ left, v1_ joined, v1_ left
timotimo ugexe: aye, it could very well be 17:33
Geth doc: b3a93247bf | (Zoffix Znet)++ | doc/Type/IO/Path.pod6
[io grant] Expand/fix up IO::Path.accessed

This isn't the doc about working with Instants, so toss all that prose		 17:39
timotimo sammers: with my latest JSON::Fast commits, HTTP::MultiPartParser installs fine 17:45
sammers: and i was able to install Bailador just now 17:48
ufobat \o/
Geth doc: 1cd7de018f | (Zoffix Znet)++ | type-graph.txt
[io grant] Fix up type graph

I'm assuming this file is maintained by hand and not a script.		 17:51
timotimo ugexe: what was the problem with zef and testers.p6c.org again? 17:52
17:52 espadrine joined 17:59 mcmillhj joined
timotimo Failed running /home/travis/.rakudobrew/bin/../moar-nom/install/bin/perl6 -Ilib bin/zef test . at /home/travis/.rakudobrew/bin/rakudobrew line 58 18:01
?!?!?
18:05 noganex_ left 18:06 mcmillhj left 18:07 noganex joined
ugexe why not moar-blead ? 18:08
timotimo i'unno, is that the reason? 18:11
ugexe there are not any methods for sending data. almost everything using the Fetcher role has the capability to send data though, so just adding some method to a few of those (curl/wget) and then adding a Zef::Report phase (just the boilerplate from Zef::Fetch/Build/etc) module and entry into config.json 18:12
the phases Zef::Fetch/Build/etc are just aggregators of Zef::Service::* that fullfill some combination of Fetch Build Test etc roles - so that boilerplate already handles sending reports to multiple reports (if they existed) 18:14
18:15 zakharyas joined
ugexe timotimo: Zef::Client also has a $.logger that you can tap (probably some more stuff should use these events internally) github.com/ugexe/zef/blob/master/l...t.pm6#L339 18:21
18:22 Rawriful joined
Geth doc: 56256d03fd | (Zoffix Znet)++ | doc/Type/IO/Special.pod6
[io grant] Minor formatting improvements in IO::Special
 18:26
timotimo oof, so i'm supposed to implement it? :) :) 18:27
18:27 jdmmmmm joined
ugexe really you should only have to implement some generic send method for wget/curl (or any other Fetcher) like here: github.com/ugexe/zef/blob/master/l...rl.pm6#L18 18:33
mst moritz: I was also thinking about en.wikipedia.org/wiki/Penrose_stairs
ugexe the other part is easy too but i'd probably get around to doing it if that much was done
Geth ecosystem: 7d36674a83 | (Nick Logan)++ | META.list
META.info -> META6.json
 18:44
18:46 mcmillhj joined
Geth doc: 184342c79e | (Zoffix Znet)++ | doc/Type/IO/Special.pod6
[io grant] Document IO::Special.what

Split off part of .new prose and list .what as a separate section, since it is a public attribute.		 18:48
18:49 Ven joined 18:50 Ven is now known as Guest22771 18:52 mcmillhj left 18:55 domidumont left, ChoHag joined 18:58 mcmillhj joined 19:03 AlexDaniel joined, mcmillhj left 19:08 darutoko left 19:09 Guest22771 left 19:16 wamba left 19:26 BenGoldberg left 19:29 ChoHag left, zapwai left 19:34 yoleaux joined, ChanServ sets mode: +v yoleaux, kent\n left 19:36 kent\n joined 19:51 BenGoldberg joined, Ven joined 19:52 Ven is now known as Guest89504
gfldex m: class C {}; dd $?PACKAGE, $?PACKAGE::.keys, GLOBAL::.keys; 19:52
camelia GLOBAL
()
("C",).Seq
gfldex if $?PACKAGE points to GLOBAL, shouldn't $?PACKAGE::.keys say the same then GLOBAL::.keys? 19:53
19:53 Guest89504 left 20:00 mcmillhj joined 20:01 labster joined 20:04 jdmmmmm left 20:09 mcmillhj left 20:12 cognominal joined 20:18 zakharyas left 20:26 mcmillhj joined 20:30 mcmillhj left 20:32 Peter_R left 20:34 Peter_R joined 20:36 titsuki left 20:37 titsuki joined 20:49 AlexDaniel left 20:53 perigrin left 21:12 Cabanossi left 21:15 Cabanossi joined 21:18 labster left 21:21 TEttinger joined 21:28 labster joined 21:35 pytuger joined 21:42 rindolf left 21:44 lowbro joined, lowbro left, lowbro joined 21:52 rurban left
Voldenet is there any use case (performance-related use case) to use Array[Int]? 22:06
And is there some good way to convert a range or iterable into an Array?
timotimo you can just @foo = @$the-thing 22:09
Voldenet m: sub b(Array[Int] $x) { $x.say }; b(^10.Array); 22:10
camelia Potential difficulties:
Precedence of ^ is looser than method call; please parenthesize
at <tmp>:1
------> 3sub b(Array[Int] $x) { $x.say }; b(^107⏏5.Array);
Type check failed in binding to parameter '$x'; expected Array[Int] bu…
Voldenet not as expected
m: sub b(Array[Int] $x) { $x.say }; b(Array[int](^10));
camelia Type check failed in assignment to ; expected int but got Int (0)
in block <unit> at <tmp> line 1
Voldenet m: sub b(Array[Int] $x) { $x.say }; b(Array[Int](^10));
camelia [0 1 2 3 4 5 6 7 8 9]
Voldenet but is Array[Int] a /performant way of doing calculations/ (as in: can I use SSE anyhow?) 22:11
22:11 Cabanossi left
timotimo it can not yet do sse or some such 22:11
for very performant stuff you want "my int @foobar" 22:12
which will give you 64bit ints
as opposed to Int which is arbitrary-size
Voldenet ah, it's best as I can get now
timotimo yup, just be careful to not go back and forth between int and Int too much; values would have to be boxed and unboxed in that case
22:12 RabidGravy left
timotimo i'm not sure if it's still the case, but i think $foo++ is still a lot slower than $foo = $foo + 1 with native ints 22:12
Voldenet huh, that's very odd 22:13
but thanks for the info, well for SSE i can use native calls, I guess 22:14
22:14 Cabanossi joined
timotimo yeah 22:16
we're working on a new JIT that can be a whole lot smarter about things, but i have no clue how hard it'll be to teach it sse instructions
Voldenet well yeah, most languages would fail in SSE situation 22:17
22:18 bjz joined
timotimo we do already have nice syntax for vectorized operations, like >>+<< for example 22:19
Voldenet uh, >>+<<... never heard of that one 22:25
ah, it's a union 22:26
or not
geekosaur docs.perl6.org/language/operators#..._Operators 22:27
Voldenet yeah, I'm reading about it now, I see how it works
it's quite fancy :D
m: say ^10 >>->> ^2 22:28
camelia (0 0 2 2 4 4 6 6 8 8)
geekosaur winces at "...full reign over..."
Voldenet ;-) 22:30
22:35 mr-foobar left 22:37 mr-foobar joined
gfldex i just found a usecase for UNIT:: :) 22:38
22:43 SCHAPiE left 22:57 Cabanossi left
samcv hello beautiful people 22:58
hope you all are doing great today
23:00 mr-foobar left, SCHAPiE joined, Cabanossi joined 23:01 AlexDaniel joined
u-ou o/ 23:01
23:02 mr-foobar joined 23:06 SCHAPiE left, SCHAPiE joined, mcmillhj joined 23:08 bjz left 23:10 mcmillhj left 23:11 bjz joined
Geth doc: dmaestro++ created pull request #1273:
Document the infix method call operators		 23:12
tbrowder hey samcv, u-ou, and rest of #perl6 gang! 23:15
samcv hello
u-ou :) 23:16
23:17 shmibs left 23:20 mr-foobar left 23:24 shmibs joined, Rawriful left 23:27 Cabanossi left 23:29 Cabanossi joined 23:33 titsuki left 23:34 titsuki joined 23:35 titsuki left, titsuki joined 23:40 jdmmmmm joined 23:44 Sgeo left 23:46 ufobat left 23:47 labster left 23:53 KongWubba left 23:54 redfd joined
redfd hi 23:55
i use rakudo
how install zef?
timotimo you can git clone [email@hidden.address] 23:56
and then cd in there and then: perl6 -Ilib bin/zef install . 23:57
23:57 mcmillhj joined
redfd perl6 -Ilib bin/zef install . All candidates are currently installed No reason to proceed. Use --force to continue anyway 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!