»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
sorear there have been several aborted attempts to run Perl 6 on the p5vm 00:00
00:01 jfried left 00:12 lumi left 00:13 lumi joined
ruoso sorear: well, as I repeated several times, if you consider p5 a vm, smop is a vm too ;) 00:18
I think 6model is a very likely solution to port rakudo to Perl 5 00:19
00:27 whiteknight left 00:35 Serpen64 joined
Serpen64 pp 00:35
00:35 Serpen64 left 00:42 whiteknight joined 00:43 silug joined 00:48 whiteknight left 00:49 lumi left 00:50 lumi joined 00:53 silug left 00:54 silug joined 00:55 thou left 01:01 noganex_ joined, jaldhar joined, whiteknight joined 01:04 noganex left 01:05 silug left, thou joined 01:07 lumi left, lumi joined 01:08 whiteknight left 01:09 thou left 01:11 whiteknight joined 01:18 gbacon joined 01:24 pernatiy joined 01:31 thou joined
colomon sorear: how'd you implement the bigints? 01:32
01:39 drbean joined 01:42 whiteknight left
sorear colomon: I took System.Numerics/BigInteger.cs from Mono, confirmed the MIT license, and copied it into my project 01:48
colomon what version of mono does current niecza need?
ah, I see
(readme)
sorear it definitely still runs fine on 2.6.7 01:49
p6eval is using that
colomon my mono on OS X is still 2.0.1, so step two is going to be updating that.
01:53 lumi left 01:54 lumi joined 02:02 jaldhar left 02:06 yinyin joined
colomon is dreaming of full Int and FatRat... 02:06
02:13 jaldhar joined
colomon sorear: how are user grammars working at the moment in niecza? (He asked, while waiting for the mono install to finish.) 02:13
colomon is also wondering what the odds are that niecza will work on his Droid faster than Rakudo on his MacBook Pro... 02:14
02:15 isBEKaml left
colomon afk # bed, mono still installing 02:28
02:28 ymasory left 02:33 lumi left, jaldhar left 02:34 lumi joined 02:35 jaldhar joined 02:41 envi_laptop joined 02:48 f00li5h is now known as f00li5h[HireMe] 02:50 Sarten-X left 02:57 Sarten-X joined 03:12 arnsholt_ joined 03:13 Sarten-X left 03:14 arnsholt left 03:18 stephanepayrard_ left, stephanepayrard_ joined, lumi left 03:19 lumi joined 03:20 Sarten-X joined 03:21 ethan left 03:28 Sarten-X left, yinyin left 03:33 agentzh joined 03:35 yinyin joined 03:36 Sarten-X joined 03:40 yinyin left, _jaldhar joined, jaldhar left 03:41 _jaldhar left 03:42 _jaldhar joined 03:43 silug joined 03:46 yinyin joined 03:51 lumi left 03:52 lumi joined 03:54 silug left 03:56 gbacon left
sorear phenny: tell colomon user grammars work pretty well. Mono 2.0.1 isn't likely to work. 04:05
phenny sorear: I'll pass that on when colomon is around.
04:16 envi_laptop left 04:20 lumi left 04:22 lumi joined 04:39 benabik left, donri joined 04:41 birdwindupbird joined 04:42 birdwindupbird left 04:46 lumi left 04:47 birdwindupbird joined, lumi joined 04:52 baest_ is now known as baest 04:55 Sarten-X left
dalek ecza: be9e4f6 | sorear++ | src/STD.pm6:
First batch of post-v6 STD convergences
05:01
05:03 Sarten-X joined, hudnix left 05:06 jmp___ left 05:08 lumi left 05:09 lumi joined 05:15 sftp_ left 05:37 Sarten-X left 05:38 kjeldahl joined 05:42 Sarten-X joined, kjeldahl left 05:47 woosley joined, nymacro joined 05:53 ethan joined 05:57 lumi left 05:58 lumi joined 06:01 wtw joined 06:04 stephanepayrard_ left, stephanepayrard_ joined 06:07 thou left
dalek p: 5658834 | pmichaud++ | tools/ (3 files):
First version of tools/build/gen_parrot.pl and NQP::Config library.
06:23
p: c5c081b | pmichaud++ | .gitignore:
Update .gitignore for new install location, config.status.
p: f412826 | pmichaud++ | / (2 files):
Move Makefile.in to tools/build.
p: 290d603 | pmichaud++ | / (7 files):
Refactor build/gen_how_pm.pl and build/gen_setting.pl into tools/build/gen-cat.pl. Update to shiny new Configure.pl.
p: ac7720e | pmichaud++ | build/ (6 files):
Remove outdated build tools and directory.
p: 7e4531a | pmichaud++ | Configure.pl:
Restore checks for files needed to build NQP.
p: 6e0ebf3 | pmichaud++ | STATUS:
Remove way outdated STATUS file.
p: ba0fe9d | pmichaud++ | docs/6model/ (2 files):
Merge branch 'master' of github.com:perl6/nqp
06:24 koban` joined 06:26 fhelmberger joined 06:31 lumi left 06:32 lumi joined 06:35 noganex joined 06:36 noganex_ left, plobsing left 06:38 jfried joined 06:39 Shozan joined, thou joined 06:40 ethan left 06:42 amkrankruleuen left 06:48 ethan joined 06:51 satyavvd joined 06:53 amkrankruleuen joined 06:55 thou left 07:01 amkrankruleuen left
sorear hello pmichaud 07:03
07:03 silug joined 07:05 pernatiy left, amkrankruleuen joined, Katip joined 07:06 pernatiy joined 07:10 lumi left 07:11 lumi joined 07:14 Katip left, silug left 07:15 silug joined 07:16 mj41 joined, Sarten-X left 07:18 wamba joined 07:23 Sarten-X joined 07:27 envi_laptop joined
colomon sorear: FYI, just discovered 2.6.7 didn't build for me, and my first thought wasn't "Darn. No niecza for me." It was "Maybe it's time to finally install OS X 10.6...." 07:28
phenny colomon: 04:05Z <sorear> tell colomon user grammars work pretty well. Mono 2.0.1 isn't likely to work.
sorear ha 07:29
2.4 will _probably_ work, although I don't have continuous integration for that
sorear -> seelp 07:31
jnthn morning, #perl6 07:39
tadzik morning jnthn 07:41
heh, someone posted niecza on reddit and misspelled the name 07:42
moritz good morning
tadzik good morning moritz
07:43 Sarten-X left
moritz tadzik: my internet connection at home doesn't work atm, so I'm afraid I'll only be available for mentoring during the day (roughly 10am to 6pm) - I'll surely backlog #phasers, highlights for me in #perl6 and read mails 07:46
tadzik moritz: no worries
moritz sorear++ # niecza v6 07:47
yath moritz: still your dsl splitter? i meant this honest yesterday, just drill the wires together
07:48 lumi left 07:49 lumi joined
moritz yath: I replaced the splitter, still doesn't work. So could be either the router, or the telco company 07:49
yath ah, hm
07:50 Sarten-X joined 07:53 araujo left 07:59 lumi left, lumi joined 08:05 awoodland joined 08:08 Katibe left 08:09 Su-Shee joined
Su-Shee hi everyone. 08:10
sorear: niecza is on reddit. ;)
08:11 cibs left, Katibe joined 08:12 daxim joined
tadzik no it's not, it's some niezca compiler :) 08:13
moritz tried to use it now and then :-)
Su-Shee yeah well choose a name which is easier to write properly and that doesn't happen. ;) 08:15
08:16 lumi left, amkrankruleuen left 08:17 lumi joined, Katibe left
moritz people misspelle raduko... erm, rakudo all the time :-) 08:17
08:17 amkrankruleuen joined
Su-Shee actually I just thought that it might make sorear happy... 08:18
tadzik Su-Shee: the niecza on reddit? 08:20
Su-Shee yes.
tadzik oh, I'm sure it will
is "niecza" difficult to spell for english-speaking people? 08:21
jnthn tadzik: I don't find it so, but I'm probably a bad example. :) 08:23
Su-Shee tadzik: well I'm german speaking and I have the same problem with all eastern european languages as I have difficulties to remember cz, sc, sz, younameit properly. 08:26
but anyways. I don't think the spelling was the main point of the reddit posting.
jnthn Su-Shee: Living for a couple of yeras in East Europe helps. ;)
I'm not sure that solution scales very well though. ;)
arnsholt_ On a related note, there are a couple of Polish cities where I happily ignore the native names. Stetting for example 08:32
08:32 arnsholt_ is now known as arnsholt
arnsholt *Stettin 08:32
tadzik jnthn: yeah, you don't count
Su-Shee: you said "younameit", you mean "Tadeusz"? :) 08:33
arnsholt: do you mean Szczecin?
arnsholt (A Polish friend tried to teach me how to pronounce the Polish name, but I just couldn't do it)
tadzik I can understand that
arnsholt Yes. ETOOMANYFRICATIVES
Affricates even 08:34
tadzik I was quite suprised that folks on Holland Hackathon took a challenge of prononcing "tadzik", I thought they will stay on "Ted" :)
08:35 amkrankruleuen left 08:37 hanekomu joined
dalek p: af48f8f | pmichaud++ | src/HLL/Compiler.pm:
Switch local %config hash to be a package-scoped %parrot_config shared by all HLL::Compilers.
08:38
p: 8dd1944 | pmichaud++ | src/HLL/Compiler.pm:
Add per-compiler configuration hash, --show-config option.
p: 3599a33 | pmichaud++ | / (4 files):
Tools to add version configuration information (based on git-describe)
p: bc9df6c | pmichaud++ | src/HLL/Compiler.pm:
Move HLL::Compiler version string out of BUILD and into version() method. Eliminate $!version attribute.
pmichaud good morning, #perl6 08:40
jnthn morning, pmichaud
wow, early morning ;)
pmichaud late night, akshually
jnthn pmichaud: How will %parrot_config fit in with @other_vm? 08:41
pmichaud nqp has some new build tools :)
jnthn yay! :D
pmichaud jnthn: we'll rip/fix it as we get @other_vm in place
jnthn pmichaud: OK, wfm. :)
pmichaud jnthn: eventually it may just become %vm_config
jnthn *nod*
pmichaud mainly I wanted a way to get all of the configuration information I need from nqp directly
jnthn yeah!
pmichaud so.... nqp --show-config gives that to me
that will vastly simplify rakudo (and other languages) build configuration system 08:42
jnthn starts to see how this could get very nice :)
pmichaud++
Also, I'm happy you've been able to dig into new nqp after my hackfest. :)
pmichaud and it's build into nqp itself, instead of requiring a separate tool (e.g., the way that parrot_config is separate from parrot -- I decided that's really kinda icky.)
*built
szabgab hi 08:43
pmichaud now that I've had time to look at it in more detail, I'm really impressed with the nqp stuff
szabgab on github.com/rakudo/star/downloads I see that the lates binary for windows was in July 2010
pmichaud szabgab: we don't have a good setup for building windows installers
szabgab is there a technical reason for that or just lack of tuits to release it?
pmichaud so, both technical and lack of tuits to resolve the technical :) 08:44
szabgab is the rakudo start now relocatable?
pmichaud you mean the binary? no, not yet. That's a parrot limitation.
szabgab or do you have to install to the same path where you compiled?
so same path
moritz pmichaud: gist.github.com/1000178 08:45
pmichaud you have to install to the same path where it was built, yes.
szabgab ok
pmichaud moritz: looking
szabgab let's see if I can hold on to my tuits till the evening when I am infront of my own computer :)
pmichaud moritz: hmmm, I'm not getting that. 08:46
08:46 amkrankruleuen joined
pmichaud moritz: do you have a tools/build/PARROT_REVISION ? 08:46
jnthn pmichaud: btw I pushed some docs on 6model last night to nqp repo. Will continue working on those. 08:47
pmichaud jnthn: I saw that they were pushed; haven't looked at them yet.
jnthn pmichaud: I plan to do a demo of how to do class based OO, then another one of prototype OO.
pmichaud moritz: oh! maybe it's complaining because --gen-parrot wasn't specified or something
moritz pmichaud: yes
pmichaud let me try that.
moritz pmichaud: right, that was without --gen-parrot
jnthn With lots of notes on why stuff is being done. 08:48
tadzik pmichaud: re the Parrot limitation, what is the problem exactly? I tried to bring it up on #parrot some time ago, and I got persuaded that the problem doesn't exist 08:50
oh, I need to backlog more
dalek p: ed47aba | pmichaud++ | Configure.pl:
Fix output when Configure.pl invoked with no options and parrot_config
08:51
pmichaud tadzik: last time I checked, Parrot hardwires the paths where it looks for libraries into the parrot executable. 08:52
so, while you can move the parrot binary around, the library files have to remain in their original locations (or somewhere that parrot looks) 08:53
08:54 Mowah_ joined
pmichaud from a windows perspective, that means that if someone compiles parrot/rakudo into C:\RAKUDO, then that's the only place the files can be installed and have things work -- they aren't relocatable. 08:54
08:55 lumi left 08:56 lumi joined
pmichaud (nqp build) working on the build system was nice -- I got to git rid of a lot of fossils in this refactor :) 08:57
s/git/get/
and nqp --version now has up-to-date version information :) 08:58
08:58 frank left
pmichaud at some point we need to decide how (if) we'll institute a release cycle for nqp 08:58
oh, one question I ran into tonight -- in nqp is there a way for a subclass' method to invoke the superclass' method of the same name? 09:00
e.g., for NQP::Compiler.somemethod to invoke HLL::Compiler.somemethod on self?
("not yet" or "no" are perfectly acceptable answers :-)
jnthn Not yet. 09:01
Well, there is a way ;)
HLL::Compiler.HOW.find_method(HLL::Compiler, 'somemethod')(self)
Or some such.
There's no sugar 09:02
pmichaud that would actually be okay with me :-)
jnthn But you can do it through the mop
shorter is
pmichaud makes perfect sense... I should've thought of it :)
jnthn pir::find_method(HLL::Compiler, 'somemethod')(self)
One day that'd be nqp::...
It's potentially faster because that way will hit the method cache.
pmichaud for the rare occasions that we need it thus far -- having these two choices is fine. 09:03
jnthn oh, you may need an extra . in there before the arg list to get it by the parser...
pmichaud okay, time for sleep here (short nap before having to get up to get kids to school) 09:04
tuit supply for Tuesday (and possibly rest of week) is not looking promising... but I'll work next on updating nom's build system so that it no longer needs --gen-parrot 09:05
it'll just be --gen-nqp or --with-nqp
anyway, sleep time 09:07
bbl
jnthn night o/ 09:08
09:18 Katibe joined, dakkar joined 09:21 hanekomu left, Sarten-X left 09:22 hanekomu joined 09:30 cotto left, Sarten-X joined 09:32 snearch joined, awoodland left 09:34 lumi left 09:35 lumi joined 09:37 Sarten-X left 09:38 cotto joined 09:39 Sarten-X joined 09:42 woosley left 09:44 araujo joined, wamba left 09:48 silug left, silug joined 09:57 agentzh left 10:02 lumi left 10:03 lumi joined 10:04 silug left, MayDaniel joined, MayDaniel left, MayDaniel joined 10:05 pernatiy left 10:07 S0laris joined
S0laris dear gents i need some help. I am looking to persuade further my education in information security, and i am not sure which university in Europe is perhaps a top in this field. Please any help would be appreciated. thank you 10:07
moritz I don't think universities are leading in the field of information security 10:10
at least not when it comes to real-world aspects 10:11
10:14 Ard1T joined 10:24 nordicdy_ joined 10:25 nordicdy_ left 10:30 S0laris left 10:31 hanekomu_ joined, hanekomu left 10:33 daniel-s joined
daniel-s hi 10:33
10:37 lumi left, lumi joined 10:55 MayDaniel left 11:01 Mowah_ left
bbkr_ tadzik: latest panda bootstraps without required Shell::Command 11:04
11:05 cotto left 11:07 daniel-s left 11:09 lumi left 11:10 lumi joined 11:11 cotto joined 11:13 Moukeddar joined
bbkr_ is global match implemented in rakudo? m:g/a/ says "Adverb 'g' not allowed" (not "not implemented"). 11:15
11:16 amkrankruleuen left 11:20 pernatiy joined 11:22 amkrankruleuen joined 11:24 hanekomu_ left 11:26 Mowah left
colomon rakudo: say "thisi is a test".match(/i./, :global) 11:27
p6eval rakudo 048573: OUTPUT«isi is␤»
colomon rakudo: say "thisi is a test" ~~ m:g/i./
p6eval rakudo 048573: OUTPUT«===SORRY!===␤Adverb 'g' not allowed on m at line 22, near ""␤»
colomon rakudo: say "thisi is a test" ~~ m/:gi./
p6eval rakudo 048573: OUTPUT«===SORRY!===␤Confused at line 22, near "say \"thisi"␤»
colomon rakudo: say "thisi is a test" ~~ m/i./:g
p6eval rakudo 048573: OUTPUT«===SORRY!===␤Confused at line 22, near "say \"thisi"␤»
colomon weird, I thought we hooked that up ages ago
11:28 Moukeddar left 11:31 Moukeddar joined
moritz the modifiers go to the frong 11:40
but m:g// has issues with its return value
s:g/// OTOH works
11:43 satyavvd left
jnthn tadzik: Just heard from $dayjob that I may be in Poland sometime in August. \o/ Can probably find a way to drop by a PM group. :) 11:44
11:45 wamba joined, lumi left 11:46 lumi joined 11:52 Sarten-X left 11:53 nordicdyno joined, hanekomu joined, nordicdyno left 12:00 Sarten-X joined, Moukeddar_ joined 12:01 Moukeddar left 12:04 Sarten-X left 12:06 wknight8111 joined 12:08 bs338 left 12:09 Moukeddar joined, Sarten-X joined, lumi left 12:10 lumi joined 12:11 Moukeddar_ left
takadonet morning all 12:11
12:12 hanekomu left
pmichaud takadonet: morning 12:12
takadonet pmichaud: I thought you went to bed?
12:12 araujo left
pmichaud I'm up again... time to take the kids to school 12:12
takadonet how much sleep did you get then? 12:13
pmichaud maybe 1.5 hrs
moritz colomon: fwiw most other regex flags work on m// - those that are implemented and just return a single Match object
pmichaud my sleep schedule is totally messed up
12:14 Sarten-X left
pmichaud jnthn (and others): any thoughts about a release cycle for nqp ? 12:15
12:15 Moukeddar left
moritz we could make it on the day between the parrot and rakudo release 12:16
but that would make for an unpleasantly tight schedule
pmichaud I don't mind if we alter the rakudo compiler release schedule for nom 12:17
moritz I think I'd prefer to do the nqp release on the Thursday, and the rakudo release on Saturday or so
or Friday maybe
pmichaud maybe we just say "in the days following ..." without committing to specific days?
or "not later than..."
moritz +1 to "not later than" 12:18
colomon +1
pmichaud so, if things are ready sooner, we can publish sooner; if we have a bit of a delay, no biggie.
wfm 12:20
afk again ... bbl
12:22 araujo joined, Sarten-X joined
tadzik jnthn: cool. But I'm afraid the only semi-active pm group is ours, in Warsaw. Krakow has one, but I dunno if it's active, no updates on the website 12:24
bbkr_: but does it work without it?
bbkr_ tadzik: no, it doesn't. i added File::Tools, JSON and Mime::base64 by hand to .perl6/lib and it works. 12:26
tadzik I'm confused 12:27
12:27 mtk joined
tadzik why would you need MIME::Base64? 12:28
12:28 amkrankruleuen left 12:30 amkrankruleuen joined
bbkr_ I have no idea :) 12:31
clean install of 2011.05 compiler build, no .perl6/ directory. i just ran bootstrap. it says it was successful. 12:32
12:33 Sarten-X left
jnthn tadzik: I think it's Warsaw that I'll be in :) 12:35
12:35 Sarten-X joined, hudnix joined
jnthn pmichaud: I agree making actual releases of NQP is good, as would be a roadmap for it. 12:36
pmichaud: On schedule, I'm not sure if we'll want/need a release every month.
Maybe for now that's a reasonable thing though.
moritz btw for building nom you have to manually create a Perl6/ directory 12:37
adding a Perl6/.gitignore (or running mkdir somewhere) would help
12:37 lumi left
moritz can't commit the fix right now, not a machine that has a ssh key set up :/ 12:37
12:38 lumi joined
jnthn oh yeah, git doesn't handle empty directories :( 12:39
I need to twiddle with that somewhat anywya though, as it doesn't work out on OSX.
moritz right
tadzik jnthn: that's supercool! 12:40
jnthn tadzik: Will let you know more, when I know more. :)
Wodka. Can haz. ;)
tadzik jnthn: about a year ago I was giving a Perl6/RakudoStar talk, I was thinking about doing a "Rakudo Star, a year after" in some upcoming time, if you were here we could make a full-blown Perl 6 workshop 12:41
bbkr_: I see
jnthn tadzik: ooh, yes :) 12:47
12:50 Holy_Cow joined, Holy_Cow left 12:52 silug joined 12:55 wamba left 13:01 JimmyZ joined 13:02 JimmyZ_ joined 13:06 JimmyZ left, JimmyZ_ is now known as JimmyZ 13:11 wamba joined, lumi left 13:12 lumi joined 13:20 bluescreen10 joined 13:23 silug left 13:32 hanekomu joined 13:37 lumi left, mikehh joined 13:38 lumi joined 13:40 JimmyZ left 13:51 Shozan left 13:52 ethan left, ethan_ joined 13:55 _twitch joined 13:58 _twitch left 14:09 PacoLinux joined 14:12 MayDaniel joined 14:17 hercynium joined 14:21 _twitch joined 14:22 benabik joined 14:26 lumi left 14:27 lumi joined 14:28 Sarten-X left 14:29 Moukeddar joined
Moukeddar what's the bigg fuzz about erlang ? 14:29
jlaire Moukeddar: what big fuzz? :o 14:30
Moukeddar it's spreading :)
14:35 Sarten-X joined
Util Moukeddar: It is excellent for concurrent programming, like multiple TCP connections (as a client, like a Web spider, or as a server, like your own custom Apache), without having to do buggy threading. 14:40
Featured in recent book: Seven Languages in Seven Weeks: A Pragmatic Guide to Learning Programming Languages 14:41
14:41 Moukeddar left
Util Dang, just missed him. 14:41
14:41 dotconf joined 14:43 Moukeddar joined
Moukeddar what did i miss? 14:43
what aout perl ?
<Moukeddar> where does it spark ?
14:44 dotconf left 14:45 yinyin left
Util Moukeddar: backlog here about Erlang: irclog.perlgeek.de/perl6/2011-05-31#i_3834086 14:45
Moukeddar: Do you mean "what is cool about Perl 6"? Or maybe "what is cool about Perl 5"? 14:47
14:48 Moukeddar_ joined, birdwindupbird left
Moukeddar_ the connection is very unstable 14:49
i wanted to ask , where does Perl sparks ?
Util Moukeddar: backlog here about Erlang: irclog.perlgeek.de/perl6/2011-05-31#i_3834086
Moukeddar: Do you mean "what is cool about Perl 6"? Or are you asking about Perl 5 (the version that almost everyone in the world runs right now)? 14:50
14:50 Moukeddar left
tadzik Moukeddar_: expresiveness would be my vote 14:50
Moukeddar_ expressiveness?
tadzik yes. It makes easy to express what you want to do 14:51
I find it like this at least
Moukeddar_ design feature :)
i mean real world cases?
tadzik is that not real-world enough? :)
14:52 JimmyZ joined
Moukeddar_ idk , 14:52
tadzik writing in Perl is fun. You rarely get a feeling that the language makes something hard to do 14:53
that's my feelings
Moukeddar_ in my university , they use perl to parse data
Util Moukeddar_: From Larry Wall's talk: "Perl, the first postmodern computer language" www.perl.com/pub/1999/03/pm.html
...it puts the focus not so much onto the problem to be solved, but rather onto the person trying to solve the problem.
Moukeddar_ Util, elaborate please 14:54
Tene news.ycombinator.com/item?id=2602340 -- .Net to javascript compiler, written by an acquaintence of mine; thought a few people here would find it interesting
14:55 Moukeddar_ left 14:57 Moukeddar joined
Moukeddar i'm so sorry guys , this connection is driving me crazy 14:57
i can't keep up with you. 14:58
14:59 kaare_ joined
Util Moukeddar: Can you get to the web to see the backlog? irclog.perlgeek.de/perl6/2011-05-31#i_3834086 14:59
That way you need not have IRC connected every moment. 15:00
Moukeddar: Perl takes the good features of many languages, while leaving behind their painful parts, and integrates them into a real *language* that a programmer's brain can use fluidly. 15:03
Most other languages feel like they have speedbumps between all the good parts.
15:03 ethan_ left, Moukeddar left 15:04 ethan joined 15:06 koban` left 15:08 lumi left 15:09 lumi joined
sorear good * #perl6 15:09
Util \o
colomon o/
Util Regarding perlcabal.org/syn/S32/Containers.html#min 15:12
I have run across two RosettaCode tasks that would benefit from .max and .min returning a list instead of a single entry when unbroken ties occur.
For example, this code:
rakudo: my @a=[11],[13,13,13],[14,14],[15,15,15],[19,19]; my $max = @a.max({.elems}); my @maxes = @a.grep({$max==.elems}); @maxes.perl.say; 15:13
p6eval rakudo 048573: OUTPUT«[[13, 13, 13], [15, 15, 15]]␤»
Util versus this:
rakudo: my @a=[11],[13,13,13],[14,14],[15,15,15],[19,19]; my @maxes = @a.max({.elems}); @maxes.perl.say;
p6eval rakudo 048573: OUTPUT«[[13, 13, 13]]␤»
Util If .max returned a list, then the second version would produce the same output as the first.
A) Doesn't this seem like a generally useful change?
B) Would there be any downsides to the change?
C) Wouldn't this also apply to .minmax?
15:13 snearch left
Tene jnthn, sorear: I think both of you have done some .net/clr stuff: news.ycombinator.com/item?id=2602340 15:13
moritz Util: consider 1 + @array.max 15:14
Util: on @array = (1, 5, 2, 4, 5);
15:15 thou joined
moritz Util: I'd be very surprised if that expression returned 3, and not 6 15:15
Util: but if .max returned a list of all maxes, it would return 3
(at least if I understood your proposal correctly) 15:16
Util moritz: I see your point. What if we change .max to be context aware as well? 15:17
15:17 DarthGandalf left
moritz Util: context doesn't flow inwards in Perl 6 anymore 15:18
Util: and the problem is not tied to numeric context
Util: 'Hallo ' ~ @names.min ~ ' and up to ' ~ @names.max
Util moritz: right, thanks. I will ponder it further. 15:19
15:19 Moukeddar joined
Tene rakudo: sub foo { return 'a' but ['a','a','a']}; my $x = foo(); say $x.perl; my @a = foo(); say @a.perl; 15:19
p6eval rakudo 048573: OUTPUT«"a"␤["a"]␤»
moritz Util: the fact remains that .min and .max returning multiple elements just isn't intuitive, and will surprise too many programmers. If you make the return value more magic, you just create weird edge cases where the expectations breakds down
15:21 ethan left
Moukeddar GIT or SVN for private projects ? 15:22
Util Perhaps new methods, like .mins and .maxs?
moritz Moukeddar: git 15:23
Moukeddar Gracias
15:23 hanekomu left
Util Moukeddar: Whichever you are proficient in. If neither yet, then Git, definitely. 15:23
moritz Util: that would be better than changing the existing ones, but I'm not sure their use is sufficiently wide to warrant inclusion in the core language
Moukeddar neither , so Git :)
moritz Util: they are not even in p5's List::MoreUtils 15:24
Util moritz: good point
moritz usually can't be bothered to set up a local SVN server
15:26 Sarten-X left 15:27 leprevost joined
sorear jnthn: ping 15:34
15:35 Sarten-X joined
tadzik orka 15:36
15:41 noganex_ joined 15:42 bluescreen10 left, noganex left
tadzik :) 15:44
15:49 isBEKaml joined, _twitch left 15:50 lumi left, pernatiy left 15:51 lumi joined, silug joined
sorear tadzik: hi 15:52
15:53 DarthGandalf joined 15:55 Moukeddar_ joined, ymasory joined, Moukeddar left
sorear tadzik: ping. 15:56
15:57 REPLeffect joined 16:00 Moukeddar joined
jnthn home from $dayjob :) 16:01
16:02 Moukeddar_ left
jnthn sorear: pong 16:03
sorear jnthn: How well does/will ctmo support precompiling user modules?
colomon Wynne:docs colomon$ mono --version
Mono JIT compiler version 2.10.2 (tarball Mon Apr 18 09:14:01 MDT 2011)
\o/
sorear \o/ 16:04
colomon and niecza-6 works on my Mac! 16:05
jnthn sorear: At least as well as master.
sorear: Tehre's one more big refactor, post-nom, to take it the "rest of the way".
e.g. to have proper BEGIN support.
sorear yesss I was going to ask about that. 16:06
basically I still don't really get how BEGIN is supposed to work in combination with serialization
jnthn sorear: Alas, I put off dealing with it Right Now because I suspected it'd be hard and I have enough hard things to think about already with 6model, package refactor, etc. 16:07
sorear how do you stop BEGIN-time code from breaking all the invariants that make precomp work?
or is this just a case where we tell people "Here's a gun. Use it responsibly."
jnthn I suspect the latter.
sorear colomon: Did you have to upgrade the OS after all? 16:08
jnthn I mean, provided the result is odd behavior rather than segfaults, it's OK.
16:08 Su-Shee left
jnthn sorear: I suspect we need an implementation of it to experiment with before we'll really know. 16:08
16:08 Moukeddar left
sorear I'm more worried by the possibility of unpredictable behavior 16:08
jnthn *nod*
sorear A reliable segfault is better than randomly getting the wrong answer... I hate heisenbugs 16:09
16:09 wamba left
jnthn True. I'm not sure I'm expecting unrepeatable behavior unless people go out of their way to get it. 16:09
sorear well, FSVO unrepeatable 16:10
tadzik sorear: pong
sorear niecza has a compiled-modules cache that's supposed to be invisible to normal users
jnthn Right, that's what I expect Rakudo to also do.
sorear in a BEGINny situation, "what modules were compiled this run" can have big effects
jnthn In theory, separate compilation is meant to avoid that issue. 16:11
The GLOBAL that BEGIN sees isn't really global. 16:12
As far as I understand it, anyway.
So every compilation unit (e.g. module) gets an empty GLOBALish which becomes the GLOBAL as far as its BEGIN blocks care.
PROCESS is, admittedly, still a point of leakage.
And there may well be others. 16:13
16:14 mattp_ left, mattp_ joined, thou left
sorear tadzik: repong. 16:17
colomon sorear: I didn't have to upgrade the OS after all. Turned out all I needed to do was google "mono download" and download a pre-compiled package rather than trying to get it with the stupid MacPorts package manager. 16:18
sorear: fast and simple, once I tried to do it the Mac way rather than the Linux way. :)
16:18 Sarten-X left
sorear tadzik: I'm thinking now about how I can get into the Perl 6 Ecosystem ... stuff 16:18
16:20 Ard1T left
sorear tadzik: wondering what form that should take. Panda port? Incorporate module fetching into own build system? 16:20
colomon Panda port, I reckon.
16:21 nymacro left
pmichaud re-good morning, #perl6 16:22
sorear hello pmichaud
16:23 ethan joined
jnthn o/ pmichaud 16:24
pmichaud: Should be able to get some basic multi-dispatch back in this evening :) 16:25
pmichaud: When you have a moment, would be good to discuss containers a bit. 16:26
sorear What kind of containers?
pmichaud I have a moment now 16:27
jnthn pmichaud: OK.
16:27 nymacro joined
jnthn pmichaud: I've realized that we've been getting away with a load of stuff with our current container model that becomes...trickier...with 6model. 16:27
16:27 REPLeffect left
jnthn pmichaud: Of note, we now have a bunch of ops that do stuff with objects. .HOW, .WHAT and so on are macros. 16:28
pmichaud container meaning "variables" or meaning "aggregate containers"?
jnthn pmichaud: Variables, but...well...I kinda want to figure out the two. :)
pmichaud okay
sorear I have choice words for the ng variables model...
16:29 Sarten-X joined
jnthn pmichaud: The whole "v-table delegation on scalar decontainerizes" is probably not very sustainable. 16:29
pmichaud: I keep finding that the 6model ops expect to really have been given a pre-decontainerized thingy.
And am worried that explicitly doing that will result us explicit dereferencing bits all over the codebase. 16:30
I can make variants of the ops that do the dereferencing...which will then break when containerized stuff gets passed into NQP code.
pmichaud if a 6model op expects a decontainerized thingy, can it just decontainerize?
jnthn Yeah...but it's an NQP interop issue.
16:30 noganex joined
pmichaud I don't understand NQP interop 16:31
jnthn Meta-objects are written in NQP
I already ran into this one
multi trait_mod:<is>($child, $parent) { $child.HOW.add_parent($parent) }
This failed because ClassHOW (written in NQP) tried to get_how on $parent, but it was wrapped up in a ObjectRef. 16:32
16:32 wtw left
sorear It will make your lives much simpler if you rigorously separate ObjectRef from Object 16:32
jnthn Of more concern though, we don't currently treat scalar containers as real objects.
16:33 noganex_ left
jnthn So we don't have a good path to implement my $foo is OtherThing; 16:33
pmichaud ...why not make NQP aware of containers?
16:33 lumi left
pmichaud more to the point, I'm guessing containers should be implemented in NQP, not in Rakudo. 16:33
jnthn pmichaud: Well, and the compiler toolkit generally perhaps... :)
16:33 nymacro left
jnthn sorear: Yes, we have too much magic right now. 16:33
And as per usual, the magic is coming to hurt.
sorear Explicit (de)containerization has been very good to me.
pmichaud I don't understand "rigorously separate ObjectRef from Object" 16:34
16:34 lumi joined
jnthn pmichaud: I think sorear really means, don't try and pretend a containerized object and a decontainerized object can be used equivalently. 16:34
pmichaud: Which is, largely, what we've been doing so far.
pmichaud well, that's what Perl 6 does.
jnthn From the user's POV, yes.
But I'm not entirely sure we can go on that way. 16:35
Especially since we need to support user-defined containers.
and .VAR should really give back a real object. 16:36
sorear: Can you describe the explicit model a little more?
sorear: How do you decide when to decontainerize?
sorear remove the container when you need to operate on the object itself 16:37
pmichaud explain "need to operate" here
aloha positive: nothing; negative: nothing; overall: 0.
pmichaud you mean... invoke a method?
sorear I'm not quite sure how the explicit model would interact with NQP though
jnthn sorear: What pmichaud asked. :) That was what I really meant to ask. :) 16:38
sorear invoking a method is an odd grey area since it needs both the object and the variable
pmichaud if it's not "invoke a method" ... what other operations are you referring to in "operate on the object"?
sorear .HOW is the most basic example of "operating on the object"
jnthn Yes, .HOW is one I ran into also.
sorear: How many operations (roughly) do you have where you explicitly decontainerize? 16:39
sorear $obj.foo; -> $obj.FETCH.HOW.find_method("foo").invoke($obj)
jnthn: almost all of them
16:39 REPLeffect joined
jnthn sorear: Where "them" are...? 16:40
pmichaud
.oO( "Q: How many states are in the U.S.? A: All of them." )
jnthn sorear: All the "low level" operations, essentially? 16:41
sorear jnthn: dozens, maybe a hundred
all the low level operations
jnthn ok.
sorear: Do you do .FETCH explicitly as a method call or is it more "low level" than that? 16:42
pmichaud is ".FETCH" just a way of saying "decontainerize" ?
sorear jnthn: it's .Fetch() in the C# code
pmichaud: yes
jnthn pmichaud: fwiw, I'm leaning more towards the sort of model sorear++ is suggesting than what we're doing now. 16:43
sorear jnthn: niecza variables are not real Perl 6 objects (I still haven't made up my mind whether to fix this, or just to make VAR cheat)
pmichaud jnthn: your statements seem to imply I have some sort of attachment to what we're doing now 16:44
jnthn pmichaud: That wasn't my implication.
pmichaud: I was more implying that I think Rakudo should change here. 16:45
sorear actually in the distant past niecza used real objects with a "FETCH" method, but it turned out that running multiple inferior runloops from every op was ... slow would be an understatement
pmichaud so far I'm not following (1) what model sorear++ is suggesting or (2) what's really the issue with the current model
I totally agree Rakudo should change its model
that goes along with what I said yesterday about "rethinking things from the ground up"
16:46 wamba joined
jnthn At the moment, ObjectRef just delegates all vtable methods to the vtable method of the value it holds. 16:46
sorear the crux of what I'm proposing is that you should be explicit in saying "containers are allowed here, they aren't here"
pmichaud jnthn: I have no problem with saying that we have to decontainerize before doing any vtable methods (more)
jnthn And then it a few places we explicitly decontainerize. 16:47
pmichaud as I mentioned yesterday, we're really reaching a point where we shouldn't be doing much vtable access at all
jnthn What I think sorear is suggesting is that our expectation should more be that we decontainerize explicitly where we need to.
sorear why not do vtable access?
pmichaud because parrot vtables are horribly slow for custom vtables
sorear ... 16:48
pmichaud I'm not saying don't do them... I'm saying they don't fit the Perl 6 model
jnthn pmichaud: I still don't know that argument holds up with 6model.
(the speed one)
sorear I thought the point of vtables was that they were faster than methods
jnthn The "don't fit" one, OTOH, I agree with.
pmichaud sorear: for PMCs, yes. For custom HLL object types, they're a lot slower.
sorear (when they are implemented in C, anyway)
pmichaud (decontainerize explicitly where we need to) I have no problem with that. 16:49
jnthn OK.
pmichaud I do think that containers should be implemented/provided in NQP, not in Rakudo. 16:50
unelss there's something incredibly specific about Rakudo containers versus what other HLLs might want/need
sorear pmichaud: should NQP be magically container-aware? ($x + $y versus $x.FETCH + $y.FETHC)
pmichaud we do containers in Rakudo now simply because Parrot doesn't provide a usable container (or even lvalue) model
I don't understand "magically" 16:51
jnthn pmichaud: I suspect that the notion of containers and some primitives should go there, but they should work with higher level container types, so we can support custom containers.
pmichaud jnthn: yes, I'm thinking a base level container class
sorear jnthn: how hard would it be in 6model to make a cross-level hook thing where you can implement, say, Bool.get_bool in C?
jnthn sorear: If you add_parrot_vtable_override an NCI PMC in the meta-object, you can do that today. 16:52
pmichaud let me back the discussion up just a bit
jnthn sorear: That's know the KnowHOW and its circularity are implemented. 16:53
pmichaud is it possible in 6model for us to create a Scalar data type?
colomon sorear: I'm very impressed by your BigInt handling so far. Everything I've tried has been scary fast.
jnthn sorear: Well, the NCI bit.
pmichaud i.e., something that acts like an ObjectRef ?
jnthn pmichaud: We can do something like that.
pmichaud That's what we should do, then.
16:53 cdarroch joined, cdarroch left, cdarroch joined
jnthn pmichaud: I'm curious what it is that makes a type "be" a container. 16:53
sorear colomon: if you check the copyright statement on the file I checked in... send your thanks to kumpera in #mono 16:54
pmichaud jnthn: the way it responds to .FETCH and .STORE
.FETCH on a non-container returns self
.FETCH on a container returns the thing it contains
jnthn pmichaud: I more meant, how does .FETCH know that :)
16:55 noganex left
sorear niecza's current model: "Variable x" refers to a container, "P6any x" doesn't. If a read-only value needs to be put in a "Variable x" slot, create a rw = false container. 16:55
16:55 daxim left, mj41 left
sorear disclaimer: Kernel.NewROScalar is consistantly one of the top allocators and I'm not convinced I'll keep this model much longer 16:55
jnthn sorear: Essentially, it goes on type?
pmichaud .FETCH isn't just like a normal method or custom op?
colomon sorear: If I'm reading this right (and it's approximate for sure), it looks like it takes about a tenth of a second to calculate and print 2 ** 3 through 2 ** 200. That's just crazy. 16:56
jnthn pmichaud: I expect it'd be op-ish, like .HOW
pmichaud opish is fine
sorear colomon: 2**200 is only a 25 byte value
16:56 Sarten-X left, noganex joined
sorear operating on 25 or fewer bytes 197 times shouldn't take a modern computer more than a few hundred ns 16:57
16:57 [hercynium] joined, [hercynium] left, [hercynium] joined, ymasory left
jnthn pmichaud: What you say put containers into NQP, does this imply that we may add them to NQP as a language? Or are you thinking more that we just put some container support into the compiler toolkit? 16:58
pmichaud not in the language
jnthn So it's there for Rakudo and other HLLs?
OK
pmichaud NQP is lower-level than that
jnthn +1
colomon sorear: that may be so, but I reckon it's competitive with the speed Rakudo has handling 32-bit integers, at a guess...
pmichaud NQP doesn't need containers because it doesn't have assignment
but it can provide support for containers because HLLs do have assignment
jnthn Well, that's what I was really getting at. :)
pmichaud to justify it a bit further
jnthn (e.g. were you pondering adding assignment to NQP) 16:59
pmichaud noway
jnthn OK, that's my feeling too. :)
16:59 hercynium left, [hercynium] is now known as hercynium
pmichaud to justify it a bit further, I think one of the huge glaring holes of Parrot's design is that it doesn't support scalar containers "out of the box" 16:59
it also doesn't have any speedy support for any sort of object reference
i.e., there's no way to reference a reference 17:00
references in parrot aren't a built-in type
(whether that's as a PMC object or as a register)
NQP probably needs to overcome that limitation, at least for Parrot 17:01
jnthn Well, really it needs to provide a consistent view of it over multiple VMs.
pmichaud and that is the other reason for doing it there :)
jnthn I'm very much more pushing container support down into the compiler toolkit though. 17:02
pmichaud I guess to be more precise...
(more)
jnthn BTW, when you say "into NQP" these days, should I be reading that as "into the compiler toolchain" rather than just "into the NQP language"?
pmichaud I'm not committed one way or another about where container support lives (NQP versus Rakudo) 17:03
I'm very committed that our containers in Rakudo should be 6model based thingies and not PMCs
jnthn OK, that's where I want to go with this too.
pmichaud I think that NQP (the language) has to be aware somehow of the possible existence of container types and be able to work with them 17:04
colomon sorear: Confirmed: calculating and printing 2 ** 1 through 2 ** 30 seven times in Rakudo takes twice as long as calculating and printing 2 ** 1 through 2 ** 200 in niecza. sorear++
pmichaud I don't think that NQP needs to use containers for its own variable bindings.
17:04 [hercynium] joined 17:05 Sarten-X joined
pmichaud when I say "into NQP", whether that means the language or toolchain will probably depend on context. 17:05
17:05 hercynium left, [hercynium] is now known as hercynium
jnthn OK, makes sense now. :) 17:05
pmichaud I'll try to be explicit in the future... but just like "Perl 6" implies a language and a runtime environment, so does "NQP" these days :-)
jnthn I see what you meant by "into NQP".
Yeah, I'm tending to draw out 6model vs compiler toolkit vs NQP language a bit these days. 17:06
moritz \o/ I can haz internetz again!
jnthn I'm aware some in Parrot land want to use 6model without using NQP.
moritz: yay! \o/
Of course, I'm all for making the three together integrate really nicely so folks will really want to use them all together. :) 17:07
pmichaud afk for a bit
colomon moritz: \o/ 17:08
17:09 JimmyZ left
pmichaud back 17:11
17:13 lumi left 17:14 lumi joined
jnthn pmichaud: I plan to focus on multi-dispatch things this evening, once I've had some nom. After that, I'll trun my attention to containers and maybe try to come up with some suggestion or prototype. 17:15
*turn 17:16
pmichaud: I probably have most of the next two days available for Rakudo hacking.
colomon sorear: Complex.re isn't there in niecza yet? Also, seems like test.pl has many holes? (Or am I missing something about how the tests are supposed to work?)
17:17 masak joined
masak bonsoir, gezebroj. 17:17
colomon o/ 17:18
jnthn lolitsmasak!
masak \o/
jnthn masak: yay 1.5 red days ;)
masak yay 17:19
quick question: so... can I just nuke S32/Str:549..557?
no objections? :) 17:20
colomon what is it?
masak lvalue substr
colomon looks to pmichaud ...
masak thing that bugs me is this: either substr'd have to be magical for that to work.
pmichaud I think substr is supposed to be magical 17:21
masak a special syntactic form, akin to $str ~~ s[] = ...
pmichaud at least, it was last I checked.
(per the spec)
jnthn I'm not convinced it'd have to be a special syntactic form, fwiw.
masak is there any place OTHER than S32/Str:549 that mentions substr being magical in this way?
pmichaud no, I don't mean syntactically magical 17:22
colomon sorear: @*ARGS NYI? 17:23
jnthn bbs, making nom 17:24
pmichaud masak: I think it may fall into the category of "perl 6 has it because it exists in p5" or something like that. 17:25
masak pmichaud: is the Str returned from substr a special container with an overridden STORE method? 17:26
colomon just created a simple mandlebrot set using niecza
pmichaud masak: conceptually, yes.
masak pmichaud: so, essentially we're not using anything that's not available on the user level already to have this work?
I'm more fine with it staying, then.
17:27 Holy_Cow joined 17:28 dakkar left 17:29 Mowah joined 17:30 Sarten-X left 17:35 sftp joined
colomon however, a 1001 x 1001 Mandelbrot is slow even in niecza.... 17:36
17:36 ymasory joined
colomon 4m, 15s, counting startup time. 17:37
bad news is, it looks like the math may be a little bit wonky? 17:38
17:39 Sarten-X joined
colomon niecza: say (1 + 1i).abc 17:39
p6eval niecza v6-1-gbe9e4f6: OUTPUT«Unhandled exception: Unable to resolve method abc in class Complex␤ at /tmp/9__2Lao8SY line 1 (MAIN mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 1438 (CORE C638_ANON @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 1439 (CORE module-CORE @ 42) ␤
..at /home/p6ev…
colomon niecza: say (1 + 1i).aba
p6eval niecza v6-1-gbe9e4f6: OUTPUT«Unhandled exception: Unable to resolve method aba in class Complex␤ at /tmp/XM2Pw6Amvl line 1 (MAIN mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 1438 (CORE C638_ANON @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 1439 (CORE module-CORE @ 42) ␤
..at /home/p6ev…
colomon niecza: say (1 + 1i).abs
p6eval niecza v6-1-gbe9e4f6: OUTPUT«1+1i␤»
colomon well, that explains the error, I reckon
ping sorear 17:40
17:40 donri left 17:43 sail0r joined
sail0r Am curious...how much of the test suite does the latest rakudo star release pass? 17:43
17:45 impious joined
colomon sail0r: I don't know if someone handy knows the answer, but I'm not sure the question really makes much sense. (after all, the test suite is also incomplete!) what are you curious about? 17:46
17:48 thou_ joined
colomon niecza: say (1 - 1i).abs 17:53
p6eval niecza v6-1-gbe9e4f6: OUTPUT«1-1i␤»
colomon niecza: say (-1 - 1i).abs
p6eval niecza v6-1-gbe9e4f6: OUTPUT«1+1i␤»
17:53 silug left, silug joined
moritz ouch 17:54
17:55 kaare__ joined, kaare_ left
sail0r niecza: say puke 17:56
p6eval niecza v6-1-gbe9e4f6: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'puke' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 469 (CORE die @ 2) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1146 (STD P6.comp_unit @ 31) ␤ at
../home/p6eval/niecza/sr…
sail0r niecza: say "puke"
p6eval niecza v6-1-gbe9e4f6: OUTPUT«puke␤»
17:59 kaare__ left 18:02 kaare__ joined
sail0r std: say "puke" 18:02
p6eval std 4b1b100: OUTPUT«ok 00:01 113m␤»
sail0r rakudo: say "puke" 18:03
p6eval rakudo 048573: OUTPUT«puke␤»
sail0r niecza: 2-2
p6eval niecza v6-1-gbe9e4f6: ( no output )
sail0r niecza: say 2-2
p6eval niecza v6-1-gbe9e4f6: OUTPUT«0␤»
sail0r rakudo: while(1){ say "puke} 18:04
p6eval rakudo 048573: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<{ }>, couldn't find final '}' at line 22␤»
sail0r haw haw haw
18:04 sail0r left
moritz rakudo: while 1 { say "puke" } 18:04
p6eval rakudo 048573: 18:05
..OUTPUT«(timeout)␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤pu…
18:05 arnsholt left
masak ...happy now? :P 18:05
colomon rakudo: loop { say "puke" }
masak or, should I say, "feeling better now?"
p6eval rakudo 048573:
..OUTPUT«(timeout)␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤puke␤pu…
18:08 kaare_ joined
moritz feels puked 18:08
18:08 kaare__ left 18:12 kaare__ joined, kaare_ left 18:14 hercynium left 18:17 Sarten-X left
colomon loliblogged: justrakudoit.wordpress.com/2011/05/...rakudo-it/ 18:18
masak colomon++ # title win
pmichaud ...leading contender? really? 18:19
masak yeah, I thought that was Yapsi. :P 18:20
18:21 kaare__ left
colomon well, I don't have a hard feeling for how much is missing in niecza. but there's a lot there right now, and sorear++ is making crazily impressive progress. 18:21
and if he can get the goods in place without losing that speed advantage... 18:22
pmichaud that's a little.... odd. one of the reasons rakudo's mathematics is slow is because of code you added, colomon. 18:24
sorry, that was low on my part. I think I need a break again for a while. 18:25
18:25 pmichaud left, Sarten-X joined
masak :/ 18:26
jnthn ...huh?!
masak I... I'm a bit stumped. 18:27
sbp too many loop { say "puke" } constructs in the maths modules
jnthn The way Num and Rat constants are done in Rakudo is not at all optimal, but it's also one of the things being fixed in nom, and I'm certainly happy to have had it working in some way rather than not up until this point. 18:28
18:28 impious left
masak colomon: fwiw, I agree about sorear++'s crazily impressive progress. but the features missing become evident as soon as one tries to run any Rakudo-based p6 code straight off. cf the contest entries. 18:30
szabgab hi, I built a windows installer for Rakudo Starr but it is 20 Mb
what parts of the original directory can I leave out? 18:31
jnthn Rakudo Starr....the drummers edition of Rakudo!
;)
moritz all .git dirs
jnthn szabgab: Did you make install? 18:32
szabgab yes make install
jnthn szabgab: OK, what did you bundle already?
szabgab and I don't think I have .git as it was built from the release
downloaded the released source
built it
masak #phasers in 29?
szabgab and the whole directory
colomon masak: yes, but there's also an increasing list of features that niecza supports that rakudo does not.
masak colomon: that's indeed worth pointing out. 18:33
jnthn szabgab: I suspect just the install directory rather than the entire source tree is needed.
e.g. if you specify a prefix when building like c:\Rakudo, then it's the stuff in there that needs installing. 18:34
masak someone(TM) should decide to build a feature matrix for various features and Pugs/Rakudo/Niecza.
jnthn masak: That'd be interesting to see.
szabgab oh I can specify prefix ?
I did not see that in the readme 18:35
let me try it now
colomon masak: and none of those features are ones which the contest programs will test, because they were supposed to work on Rakudo (in practice, if not in the rules).
masak colomon: troo
pugs: class A {}; class B { trusts A } 18:36
p6eval pugs: ( no output )
masak pugs: class A {}; class B { trusts A }; say "alive"
p6eval pugs: OUTPUT«alive␤»
masak Pugs has 'trusts'. no other implementation does.
jnthn szabgab: alas, no
masak niecza: class A {}; class B { trusts A }; say "alive"
p6eval niecza v6-1-gbe9e4f6: OUTPUT«===SORRY!===␤␤Action method package_declarator:trusts not yet implemented at /tmp/VtpAYZBAD4 line 1:␤------> class A {}; class B { trusts A⏏ }; say "alive"␤␤Unhandled exception: Unable to resolve method statement_level in class Str␤
..at /home/p6…
szabgab the thing works if I only have rakudo\install and nothing else 18:38
but I wonder what else should I try beside 42.say
18:38 cooper__ joined
jnthn Maybe try one of the built-in modules 18:38
rakudo: use JSON::Tiny;
p6eval rakudo 048573: OUTPUT«===SORRY!===␤Unable to find module 'JSON::Tiny' in the @*INC directories.␤(@*INC contains:␤ lib␤ /home/p6eval/.perl6/lib␤ /home/p6eval//p2/lib/parrot/3.4.0-devel/languages/perl6/lib␤ .)␤» 18:39
jnthn star: use JSON::Tiny;
p6eval star 2011.04: ( no output )
jnthn star: use JSON::Tiny; say "ok"
p6eval star 2011.04: OUTPUT«ok␤»
jnthn szabgab: ^
szabgab it told me ok 18:40
is that a p6 implementation? 18:41
masak it's Rakudo Star.
18:41 cooper__ is now known as cooper, Holy_Cow left
szabgab but the JSON::Tiny is I guess a p6 module that comes with r* 18:42
right?
moritz star: use JSON::Tiny; say to-json([1, 'abc'])
p6eval star 2011.04: OUTPUT«[ 1, "abc" ]␤»
szabgab oh to-json :)
I was trying to_josn 18:43
masak that's wrong, twice :P
flussence
.oO( javascript object syntax-notation? )
masak we should really implement that method name typo detector... :P
szabgab oh and if I do that on separat lines it does not work 18:44
it even forgets I loaded a module ?
masak star: use JSON::Tiny;␤say to-json([1, 'abc'])
p6eval star 2011.04: OUTPUT«[ 1, "abc" ]␤»
masak wfm
szabgab no 18:45
jnthn masak: I think he meant in the repl, where it should work, but it can be a little quirky.
18:45 pmichaud joined
pmichaud jnthn: ping 18:45
masak oh, the REPL!
jnthn pmichaud: pong
szabgab just restaretd rakudo between the use statement and the call to-json
pmichaud how hard would it be to get nqp to load the trans_ops dynoplib
szabgab I can forgive that it is not persitent between restarts :)
18:45 arnsholt joined
pmichaud for some reason Parrot thinks that the <pow> opcode ought to be a dynop. 18:45
jnthn pmichaud: --vm-libs=trans_ops 18:46
18:46 hercynium joined
jnthn If you want it one-off 18:46
szabgab so is there prefix flag for Configure.pl ?
pmichaud I'm thinking more permanent
jnthn oh
that's easy too, one sec
pmichaud more to the point, infix:<**> doesn't work in NQP now
because it depends on <pow> which has to be loaded.
(notably: <sqrt> is considered useful enough to be core.) 18:47
18:47 lumi left
jnthn pmichaud: src/NQP/Actions.pm, line 58 18:47
pmichaud: Add it to that array.
18:47 lumi joined
pmichaud line 72 in my copy...? 18:47
actually, 70 ? 18:48
jnthn oh wtf
the github line number column isn't aligned with the actual lines!
my @loadlibs := ['nqp_group', 'nqp_ops'];
That line
:)
pmichaud looks like line 70 to me on github, too :) 18:49
jnthn wonders how he's managed to get that broken locally :)
18:50 Ali_h left 18:52 silug left, Chillance joined, silug joined 18:54 kjeldahl joined
pmichaud It bugs me that we have a core Perl 6 contributor saying "Don't use Rakudo." 18:55
I kind of figured we were above such things. 18:56
masak pmichaud: I don't believe colomon meant it in such a black-and-white manner. 18:57
pmichaud masak: But I'm thinking of how others will read it.
colomon what masak said.
pmichaud furthermore, at the time the math code went in I said (1) it'll kill Rakudo performance and (2) it's the wrong way to do it.
It seems hardly fitting to blame Rakudo for it now. 18:58
masak pmichaud: if Rakudo has any prestige to lose because Niecza is 50 times faster, it will no matter what colomon writes.
tadzik sorear: I'd start from porting/running ufo
colomon pmichaud: if you're talking about numeric literals, that has absolutely NOTHING to do with the performance difference I'm seeing
tadzik running Pies should be quite straightforward, then you have running ufobuilder, which is a Panda implementation. The long-term goal will be porting Panda itself 18:59
pmichaud not just literals, but the whole way the math classes and operators have been structured
colomon About the only real world code that will ever slow down is the trig tests, and they needed it, because what you had before was too broken to use.
tadzik sorry for the absence, my isp is nuts 19:00
szabgab Argument "RELEASE_3_3_0" isn't numeric in addition (+) at Configure.pl line 25, <$REQ> line 1.
just sayin'
masak szabgab: on Windows?
szabgab that's when I run perl Configure.pl ..
yes
5.12.3
masak goes to look at that line in Configure.pl 19:01
jnthn Is the Configure.pl in star the same as the one in Rakudo, ooc?
pmichaud no
jnthn doesn't see that warning on Windows on the Rakudo one.
Ah, OK
19:01 Sarten-X left
pmichaud gist.github.com/1001064 # jnthn, this program gives me a "Cannot take substr outside string" error when run in nqp 19:02
moritz is it #phasers time?
masak $reqsvn += 0;
colomon moritz: yes
masak moritz: yes.
szabgab: that's the line. 19:03
szabgab is there a rakudo logo somewher in the r* that could be used for a menu item in Windows?
pmichaud I think that line can be removed.
masak szabgab: feels like the wrong way to convert something to a number in the first place...
moritz never thought of updating star's Configure.pl to git revisions
masak that might be why :P
jnthn pmichaud: Oddness
pmichaud jnthn: you might need the push I just made to get the trans_ops dynoplib loaded. 19:04
jnthn pmichaud: Got a backtrace?
pmichaud: Going to try and reproduce it now.
pmichaud gist.github.com/1001069 # with backtrace
19:05 mj41 joined, thou_ left
jnthn aarrrgh, that. :| 19:05
masak++ also hit and filed that one. 19:06
Platform specific. :|
Will fix.
pmichaud I've always hated that Parrot throws an exception there anyway. 19:07
masak right, that one. 19:08
yeah, one is bound to run into that bug if one plays around with big numbers for long enough...
pmichaud ...what does it have to do with big numbers, ooc? 19:10
19:10 Sarten-X joined
pmichaud it happens even if I only do the loop once 19:10
masak huh. 19:11
pmichaud and it says the error is in set_mainline_module
masak oh, this is nqp!
I had already dived into RT looking for the original ticket :) 19:12
but, no, here it is: github.com/perl6/nqp/issues/13
nothing to do with big numbers.
19:12 envi_laptop left
masak to be perfectly fair, I didn't even discover it. someone++ at the YAPC::Russia hackathon did. 19:13
jnthn masak: embarassingly, it also goes away if the filename is too long 19:14
er
is longer
masak good luck with that one!
jnthn I didn't expect an exception there :P
meh
pmichaud I don't understand what set_mainline_module is trying to do anyway
19:14 Sarten-X left, lumi left
masak jnthn: btw, both times I've been building rakudo/nom, I've had to manually mkdir 'Perl6' to get the 'make' to go all the way through. is that a known defect? 19:15
19:15 lumi joined
jnthn masak: Somebody else mentioned that today. 19:15
moritz
jnthn masak: That whole area needs a twiddle to fix the OSX breakage though.
pmichaud that Perl6 dir is just asking for trouble 19:16
masak oh, so it's actually supposed to be there?
jnthn masak: yeah
masak: Well..."yeah" :)
masak my guess was that those modules ended up being generated to the wrong dir.
jnthn Well, they need to be where "use" will see them.
masak there are ways to commit directories, even in git.
moritz but on OSX it conflicts with ./perl6, so the location of these modules needs to change
masak well, on any default-configured hard drive on OS X. 19:17
first thing I do when I get an Apple laptop is to reformat the hard drive. :/
19:18 wamba left
pmichaud "use" needs to be smart enough to look in places other than the current dir 19:18
or, we need to change the name of the executable
jnthn pmichaud: Indeed.
pmichaud (we need to do the first, regardless)
moritz any objections to calling the binary 'rakudo'?
jnthn Patches welcome if you have tuits, otherwise I'll try and get to it in the next couple of days.
pmichaud I'm working on the build system already, I'll put it on my list. 19:19
if someone else beats me to it, that's fine also.
19:21 Sarten-X joined, kjeldahl left 19:23 Mowah_ joined
pmichaud ...trying to build niecza-6, "make" gives me 'make: *** No rule to make target `FETCH_URL', needed by `.fetch-stamp'. Stop. 19:24
what am I forgetting/missing?
nm, there's already a .exe for me 19:25
gist.github.com/1001108 # there's hope for Rakudo speed yet. 19:27
tadzik pmichaud: isn't that mostly mono startup cost? 19:29
pmichaud tadzik: maybe, but we also have Parrot startup cost
tadzik sure
pmichaud: how does that test scale?
19:29 birdwindupbird joined
moritz pmichaud: last I looked niecza had rather slow startup, but faster run time 19:30
19:31 whiteknight joined, whiteknight left
colomon errr... why aren't the numbers the same? 19:31
oh, second one goes one further, is all
pmichaud one has < and the other has <=
19:32 Sarten-X left
pmichaud Yes, niecza has a 1.39 second startup on my machine 19:34
afk, kid pickup 19:35
19:40 Sarten-X joined 19:49 _jaldhar left 19:52 Mowah left 19:53 silug left, silug joined 19:57 birdwindupbird left
masak lol! I blogged! strangelyconsistent.org/blog/a-mont...ndamentals 19:59
szabgab so I have win32 binary with aan instalelr.exe
where shall I put it?
pmichaud I can put it in the star directory with the others 20:02
moritz szabgab: do you have commit access to star?
pmichaud and I can give commit access to star
szabgab I don't think I have commit bit there 20:03
let me put it on my server now
and one you takes it from there and adds to the server ? 20:04
to the repo
I should go to sleep ASAP :)
20:05 pjcj left, pjcj joined
szabgab perlide.org/download/rakudo-star-2011.04.exe 20:05
pls don't advertise this link 20:06
colomon oh, now that's interesting! on a stupid string benchmark, rakudo actually beats niecza -- even factoring out the startup times.
szabgab it is about 6.5 Mb
moritz downloads it, and will upload to github when done 20:08
szabgab thanks
night&
20:08 silug left 20:09 silug joined
masak 'night, buttercamels. 20:09
colomon \o
20:10 masak left, lumi left 20:11 lumi joined 20:15 silug left
sorear good * #perl6 20:18
lots of nice highlighted backlog
tadzik oh, hello sorear 20:19
20:20 shachaf_ joined
sorear colomon: what do you mean by "holes in test.pl"? And yes, no .re 20:21
20:21 pjcj left, pjcj joined
sorear colomon: @*ARGS definitely works with the precompiler. I might have broken it again in compile-and-run-mode 20:22
colomon sorear: I believe you said (and maybe I misunderstood) that test.pl was the best place to check to see what was implemented? I seem to keep finding stuff that works but isn't in test.pl as far as I can tell.
@*ARGS definitely wasn't working for me, but it was compile-and-run-mode, and you say. 20:23
moritz maybe that was from before I added some spectests :-)
colomon oooo, spectests
sorear colomon: complex numbers are untested and probably don't work
20:24 shachaf_ left
colomon sorear: for sure .re and .abs don't work, but it seems like basic complex math works. 20:24
20:24 pjcj left 20:25 pjcj joined
colomon either that or it was one hell of a coincidence I got the image of a mandelbrot set! 20:25
sorear yikes, it looks like I'm starting to sow strife
colomon yeah, sorry about that.
for what it's worth, I think you're doing an amazing job 20:26
pmichaud I think sorear++ is doing an amazing job also
there's no question about that.
20:28 Sarten-X left
pmichaud nor do I see it as "niecza is sowing strife." 20:29
20:32 molaf_ left
moritz just sees it as a tongue in cheek comment that didn't arrive as such 20:34
20:36 Sarten-X joined
sorear colomon: what's the really bad string case? 20:44
rakudo: (3+4i).abs.WHAT.say
p6eval rakudo 048573: OUTPUT«Num()␤»
sorear rakudo: 5.re.perl.say 20:45
p6eval rakudo 048573: OUTPUT«Method 're' not found for invocant of class 'Int'␤ in main program body at line 22:/tmp/h9DzxiSskU␤»
20:48 Sarten-X left 20:49 lumi left
colomon sorear: for 1..20000 { $a ~= "b"; } 20:49
niecza: my @a = 1..10; say @a[*-1] 20:50
p6eval niecza v6-1-gbe9e4f6: OUTPUT«10␤»
colomon niecza: my @a = 1..10; say @a[*/2]
p6eval niecza v6-1-gbe9e4f6: OUTPUT«6␤»
20:50 lumi joined
sorear not suprising that you got the same result in two imlementaions with that 20:50
it spends most of its time in memcpy() 20:51
20:57 ymasory left 20:58 Sarten-X joined
colomon niecza: say 5 %% 4 21:00
p6eval niecza v6-1-gbe9e4f6: OUTPUT«Bool::False␤»
21:02 thou joined
moritz fwiw I can't upload the win32 star installer to github 21:02
colomon niecza: say [||] (4, 5, 0)
p6eval niecza v6-1-gbe9e4f6: OUTPUT«===SORRY!===␤␤This macro cannot be used as a function at /tmp/pEPpJtjYIQ line 1:␤------> say [||]⏏ (4, 5, 0)␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 469 (CORE die @ 2) ␤ at
../home/p6eval/niecza/src…
moritz "Something went wrong that shouldn't have. Please try again or contact support if the problem persists."
and I'm too tired to contact the support
if soembody else manages to upload, that would be nice
moritz -> sleep
sorear I had the same problem when I was trying to upload niecza-v5 21:04
21:09 hercynium left, wknight8111 left 21:11 shachaf left
dalek nch-scripts: c55c266 | (Solomon Foster)++ | prime-3.pl:
Rework the prime benchmark so it works on niecza.
21:12
21:14 shachaf joined
tadzik sleepytime, see you folks 21:19
sorear night! 21:20
jnthn Dammit, weather was hot today and now my appartment is also too hot. 21:24
ahh...a cold beer :) 21:26
dalek kudo/nom: 5a3602a | jnthn++ | / (2 files):
Start to get the multi dispatcher core ported over the 6model.
21:27
kudo/nom: abedc89 | jnthn++ | src/pmc/perl6multisub.pmc:
Remove old Perl 6 multisub PMC, which we'll no longer use.
kudo/nom: 56ee0f5 | jnthn++ | .gitignore:
Update .gitignore.
21:32 mj41 left, lumi left 21:33 ruoso left, lumi joined
sorear jnthn: around now? 21:44
jnthn sorear: around but trying to concentrate on multi-dispatch :)
sorear jnthn: how does the ctmo bootstrap work? 21:45
jnthn I'm not sure exactly what bootstrap you're talking about.
sorear suppose you decide sometime in Sep to start a Javascript ctmo backend
21:45 Sarten-X left
jnthn ah, nqp? 21:45
21:45 Patterner left
sorear can ctmo cross-compile to another backend? 21:46
jnthn The trick is mostly to write a PAST compiler for the target backend and implement the 6model primitives there.
21:46 Psyche^ joined, Psyche^ is now known as Patterner
jnthn And then eventually get nqp to spit itself out on another platform. 21:46
sorear jnthn: but don't you also need some kind of serialized metaobject stuff?
jnthn Well, at the moment it's not really serialized per se. 21:47
But yes
It'll depend on that format being portable, at least during the initial bootstrap.
sorear I thought ctmo was the serialization branch
jnthn No
Not the full-blown serialization.
It's the cheating serialization :) 21:48
It was mostly about doing all the meta-object setup during the compile.
For now it records enough to re-do that at startup, but it has already statically resolved a bunch of symbol cross-references.
It's kinda half way to what I want. 21:49
The actions and grammar won't have to change beyond this, the refactor from here is pertty much all in HLL::SerializationContext and NQP::SymbolTable and those kinds of places. 21:50
I coulda taken it on but...priorities. 21:51
21:54 Sarten-X joined
sorear jnthn: does it record while it builds, or does it generate the redo info based on the metaobjects? 21:56
sorear is currently trying to debug why eval() always returns Any on niecza 21:57
also, reddit thinks I need a better name.
jnthn sorear: While it builds. 21:58
sorear: Of course, the eventual serialization would just look at the metaobjects.
sorear: My roadmap is something like, get nom good enough to be master, serialization, do BEGIN fully, PAST uses 6model, dig into writing an optimizer. 22:01
Though it doesn't matter if I do all of those personally. :)
sorear You plan to have a master without working BEGIN?
jnthn (as in, I'm hoping not to :))
Well, it'll be no less broken than current master has it. 22:02
sorear How broken is current master?
jnthn In fact, it'll be better than current master
Very.
sorear more broken than niecza? :)
jnthn Dunno. :) 22:03
niecza: BEGIN { say 42 }
p6eval niecza v6-1-gbe9e4f6: OUTPUT«42␤»
sorear protolexpads help.
jnthn Yeah
Got them in nom.
sorear but it's being treated as INIT, so, that runs after the parse finishes
jnthn ah, ok
sorear if in a module, perhaps long after 22:04
jnthn glances NQP to see how he did multi subs there, and realizes it cheats more than Rakudo can get away with 22:06
sorear niecza: say module Foo { 2 + 2 } 22:07
p6eval niecza v6-1-gbe9e4f6: OUTPUT«Any()␤»
sorear this is the crux of the Niecza eval->Any bug
except it's "my module CORE; ...; {YOU_ARE_HERE}"
jnthn hmm
jnthn somewhat expects that to output Foo() (e.g. the type object for the module) 22:08
sorear as would I, except modules don't have type objects yet 22:09
on reflection, the real issue is that evals are being installed in the setting at all 22:10
they should be getting installed at the call site!
jnthn gah, internet fail... 22:13
jnthn back
22:16 phenny left, sbp left, nsh left 22:17 crux_ joined 22:24 leprevost left
Util Does Niecza use the Setting from Rakudo? 22:27
tylercurtis Util: no.
Util Why not?
22:33 lumi left
tylercurtis Util: partly because Rakudo's setting is somewhat wrong due to limitations in Rakudo that aren't present in Niecza. 22:33
22:33 lumi joined
tylercurtis I expect. 22:33
sorear hah 22:34
jnthn heh
sjohnson any of you guys excited for the NHL games tomorrow?
Util tylercurtis: thanks, that helps
tylercurtis The reverse also likely applies.
sorear it's really more the other way aound :p
jnthn I suspect that for many low-level things, the setting is just too innevitably tied to the guts of an implementation also.
Also, different implementations may have different most efficient ways of implemetning the same built-in. 22:35
tylercurtis sorear: well, I was mostly thinking about Rakudo's setting being full of our-scoped things. 22:36
sorear large portions of niecza's setting are, however, more or less directly cloned from Rakudo
one thing worth noting in niecza's setting is that it has no multis
partly because for a long time niecza didn't support multis, and partly because they're still disgustingly slow 22:37
22:39 crux_ left
dalek kudo/nom: 9052157 | jnthn++ | src/ (3 files):
First untested cut of op to enter the multi-dispatcher and a very basic, non-caching multi dispatcher entry point.
22:40
kudo/nom: 865919b | jnthn++ | src/ops/perl6.ops:
Oops, copy-pasto.
kudo/nom: f9a8b45 | jnthn++ | src/Perl6/ (2 files):
Start to get handling of multi and proto in place. Dispatch doesn't actually work just yet, though.
22:40 noganex_ joined 22:43 noganex left
sorear colomon: What was pmichaud referring to when he said that complex number speed was your fault? 22:46
22:49 Sarten-X left 22:50 silug joined 22:58 Sarten-X joined
sorear hrm. I guess -n and -p should only affect MAIN's setting, not the setting of every single module :) 22:59
23:03 cooper left 23:04 cooper joined 23:09 lumi left, lumi joined 23:12 Sarten-X left 23:15 MayDaniel left
dalek ecza: b8c3c97 | sorear++ | / (5 files):
Improve handling of free variables with settings, make {YOU_ARE_HERE} in a loop work
23:24
ecza: 49b11af | sorear++ | / (5 files):
Add $*ARGFILES, -n and -p option handling
ecza: b23212e | sorear++ | src/NieczaBackendDotnet.pm6:
Fix @*ARGS accidentally not being passed to compile-and-run code
ecza: 2bacd41 | sorear++ | lib/COR (2 files):
Add CORP and CORN setting files, -n & -p are now fully functional
tylercurtis sorear++
Does Rakudo have those options yet?
sorear I think so 23:25
jnthn Yeah, masak++ added -n and -p a while ago, though the Niecza factoring of it is preferable.
Well, I *guess* it is :) 23:26
23:28 Chillance left
sorear Why do you say you guess it is? 23:29
jnthn sorear: Because I said it was preferable before I'd read the patch. 23:30
sorear Ah
jnthn sorear: Just read it, and it's what I expected. So, it certainly is. :)
23:30 Sarten-X joined
dalek p: 970ebe2 | jonathan++ | src/pmc/sixmodelobject.pmc:
Add a (hopefully quite temporary) workaround for Parrot's Parrot_pcc_invoke_from_sig_object sucks.
23:33
23:33 Vlavv left
dalek kudo/nom: 518fec3 | jnthn++ | src/ (3 files):
Needed to mark parameters that the multi-dispatcher should consider; also make ;; work again.
23:37
kudo/nom: c40fc7e | jnthn++ | src/binder/multidispatch.c:
Fix a silly inverted condition, and we have basically functioning, if incomplete and unoptimized, multiple dispatch.
23:40 silug left 23:45 Vlavv joined, thou left, kst` is now known as kst 23:48 cdarroch left 23:50 lumi left 23:51 jql left, stephanepayrard_ left, dayangkun joined, lumi joined 23:54 sirmacik left, Sarten-X left, sirmacik joined, sirmacik left, sirmacik joined, baest_ joined 23:55 ribasushi left, ribasushi joined 23:56 baest left
dalek ecza: 29be4bb | sorear++ | lib/ (2 files):
On-same-runloop eval, simplify somewhat
23:56
23:58 frettled left