»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.
jlaire masak: the Knuth code you link to finds _strong_ addition chains, which have an extra requirement: you can only add a+b if |a-b| is also in the chain. This changes the optimal result for n=23: the regular addition chain [1 2 4 5 9 18 23] is not strong because 23=18+5 but 18-5=13 is not there. An optimal strong chain is one longer, [1 2 3 5 8 13 18 23] 03:22
however, Knuth also has some code for finding minimal addition chains at www-cs-faculty.stanford.edu/~uno/programs.html :)
jlaire n=83 is the first case where a strong addition chain requires 2 extra numbers: [1 2 4 8 16 17 33 66 83] vs. [1 2 3 5 8 13 21 26 31 57 83] 03:29
sorear good * #perl6 05:12
TimToady that's the first time I've ever used a hacksaw on a computer, I believe 07:45
sorear hi TimToady 07:46
TimToady we just replaced quietfanatics broken display on his laptop; the new one was really for a different computer, but we made it work with a hacksaw :) 07:47
(had two metal mounting brackets on top that wouldn't fit, but electronically it was compatible) 07:48
sorear I've used one of these on a computer en.wikipedia.org/wiki/Nibbler does that count?
TimToady we coulda used one of those, I expect 07:49
and with less chance of putting metal shavings where they don't belong
moritz r: our multi foo() { } 08:31
p6eval rakudo b4caaa: OUTPUT«===SORRY!===␤Cannot use 'our' with individual multi candidates. Please declare an our-scoped proto instead␤at /tmp/wu7LUbqx88:1␤»
moritz should put that into changelog
dalek kudo/nom: 8c7747b | moritz++ | docs/ChangeLog:
[ChangeLog]: note that scoped multis now die
 08:33
ayad02 hello team 09:11
can you help me to transform this command in a code perl => #ssh 192.168.0.02 rsync -averb '/var/www/html/devaxecam2011/images/' '/var/www/html/devaxecam2011/videos/' /home/TMP_ARCHIVES/DEVAXECAM/dev.`date +%d%m%y.%H%M` 09:14
moritz hello ayad02 09:15
this channel is for Perl 6 usage and development
are you sure you want to do it in Perl 6?
ayad02 no i m no sure 09:21
moritz then you are probably in the wrong channel.
of course you can do it in Perl 6, but the easiest way would be to spawn a shell that executes the command you already pasted 09:22
(as will be the case with perl 5)
dalek Heuristic branch merge: pushed 196 commits to perl6.org/archive_dir by ranguard 09:53
ranguard moritz (or anyone) please review github.com/perl6/perl6.org/tree/archive_dir - this is just migrating the content, I've tried to do it with git history but not done this before so a sanity check would be appreciated :) 09:59
ranguard heads off for a couple of hours so no rush :)
moritz ranguard: will do
ranguard thanks
moritz ranguard: thanks for doing all that work in the first place
Ulti was that big download button on perl6.org there before? 10:05
moritz yep, it's from pre-niecza times 10:07
feel free to replace it with something that makes sense these days :-)
grondilu perl6: sub foo { gather { take 1; say "not so lazy"; take 2 } }; my @a <== foo; 10:35
p6eval niecza v15-4-g1f35f89: OUTPUT«Potential difficulties:␤ @a is declared but not used at /tmp/xWFXIGB2yl line 1:␤------> ke 1; say "not so lazy"; take 2 } }; my ⏏@a <== foo;␤␤Unhandled exception: Feed ops NYI␤ at /home/p6eval/niecza/lib/CORE.setting line 1362 (die @ 3)…
..pugs b927740, rakudo 8c7747: OUTPUT«not so lazy␤»
grondilu isn't a gather-take structure supposed to be lazy? 10:36
moritz it is 10:37
perl6: sub foo { gather { take 1; say "not so lazy"; take 2 } }; my @a := foo;
p6eval niecza v15-4-g1f35f89: OUTPUT«Potential difficulties:␤ @a is declared but not used at /tmp/8FpLiiVtVC line 1:␤------> ke 1; say "not so lazy"; take 2 } }; my ⏏@a := foo;␤␤»
..pugs b927740: OUTPUT«not so lazy␤»
..rakudo 8c7747: ( no output )
moritz perl6: sub foo { gather { take 1; say "not so lazy"; take 2 } }; my @a := foo; say @a[0]
p6eval rakudo 8c7747, niecza v15-4-g1f35f89: OUTPUT«1␤»
..pugs b927740: OUTPUT«not so lazy␤1␤»
moritz it seems that <== isn't lazy in rakudo and pugs; dunno if it should be 10:38
grondilu I'm pretty sure it is supposed to be stricly lazy
moritz that would surprise me 10:39
grondilu anyway I hadn't thought of using ':='. I'll do that. 10:40
moritz S06 talks about each part of the feed chain being in its own thread 10:41
which doesn't make any sense for strictly lazy evaluation
woah, S06 has a lot to say about feeds 10:44
and much of that is rather vague (in terms of the APIs used behind the scenes to make the stuff happen in the way the spec wants it to happen) 10:45
masak antenoon, #perl6. 10:50
colomon o\ 10:52
moritz good am, ma
perl6: say so 'am' ~~ /<*masak>/
p6eval rakudo 8c7747: OUTPUT«===SORRY!===␤Unrecognized regex metacharacter (must be quoted to match literally) at line 1, near "*masak>/"␤»
..niecza v15-4-g1f35f89: OUTPUT«===SORRY!===␤␤Action method assertion:sym<*> not yet implemented at /tmp/h91Opm_1WZ line 1:␤------> say so 'am' ~~ /<*masak⏏>/␤␤Unhandled exception: Unable to resolve method oplift in class Any␤ at /home/p6eval/niecza/s…
..pugs b927740: OUTPUT«*** No such subroutine: "&so"␤ at /tmp/GIL9OZsI_D line 1, column 5 - line 2, column 1␤»
moritz I figured that it might not be too hard to implement after all, once LTM is in place for all stuff (and not just protos) 10:54
because it really just marks als DFA or NFA as accepting
and LTM takes care that you get as much as possible
masak jlaire: aye. 10:55
er, moritz: aye.
jlaire: thanks for the Knuth link correction -- 'tis fixed now.
moritz phenny: tell ranguard just reviewed the archive_dir branch of archive_dir. Seems to be fine, everything interesting is there (including history) 11:05
phenny moritz: I'll pass that on when ranguard is around.
moritz phenny: tell ranguard I may move that stuff to a different repository eventually, but I'm not sure yet. Please let me know if you want any help setting up redirects or so 11:07
phenny moritz: I'll pass that on when ranguard is around.
moritz oh, and ranguard++
ranguard thanks moritz - I'll tidy a bit and then merge over the w/e 11:13
phenny ranguard: 11:05Z <moritz> tell ranguard just reviewed the archive_dir branch of archive_dir. Seems to be fine, everything interesting is there (including history)
ranguard: 11:07Z <moritz> tell ranguard I may move that stuff to a different repository eventually, but I'm not sure yet. Please let me know if you want any help setting up redirects or so
ranguard you can then do what you like with it from there, I'll add the redirects from dev.perl.org/perl6/ once this is live 11:14
moritz masak: just saw a nice autopun on #perlde: somenick-- # for spreading a negative base mood 11:21
masak :) 11:26
flussence wonders how a negative-base number system would work 11:30
moritz with great care :-) 11:31
masak flussence: nothing inherently contradictory in that.
moritz except that 10_{-10} > 20_{-10), but 10_{-10} < 11_{-10} 11:32
flussence I guess the sign of a number would be <- +>[$length % 2]
or something like that
moritz hey, it's 2012-04-01 soon :-) 11:33
masak flussence: let's say you want to express 8 in base -3. well, 8 == 1 * (-3) ** 2 + 1 * (-3) ** 1 + 2 * (-3) ** 0 == :-3<112>
the calculation is done just like with positive number bases.
moritz is that a uniq representation? 11:34
erm, is that representation uniq?
masak I think so.
moritz funny things you encounter on #perl6 :-)
masak though it isn't very lexicographically sane.
0, 1, 2, 120, 121, 122, 110, 111, 112, 100, 101, 102... 11:36
masak hm, it seems uniq up to negative signs, or something like that. 3 can be written both as 120 and as -10. 11:37
moritz well, let's say it has it's own (in)sanity :-) 11:39
masak that seems to capture it, yes. 11:40
flussence I'm reminded of balanced ternary, which has slightly more practical use :) 11:51
icwiener Hello.
Is there a user channel somewhere or is the mailing list the place for that? 11:52
masak icwiener: perl6-users, or here.
we're very fine with taking user questions here on #perl6.
flussence: have you seen the 2i-base number system?
icwiener Oh, that's cool. :)
masak flussence: any complex integer (aka "Gaussian integer") can be written as a normal integer. 11:53
and with just the digits 0..3, at that!
no minus signs required. 11:54
icwiener Haha, happy as I was about being allowed to ask questions here, I found the answer by accident. :D 11:55
.. I think.
Will try it.
masak ask the question anyway. maybe we know a better answer :P
icwiener I was parsing json with JSON::Tiny and got the error message "no ICU lib loaded ...". So I searched a bit and found the hint to install libicu ... but it was installed here. Then I came here and asked to ask (*coughs*). Then I found the additional hint to rebuild rakudo after installing libice. So I figured it must be the dev files that are needed so I now installed libicu-dev and rakudo is currently rebuilding. Let's see if it works. :) 11:59
masak sounds about right. 12:02
I think it's Parrot that needs a rebuild after libicu-dev is installed. but rebuilding Rakudo might take care of that.
[Coke] yawns and drinks some narsty coffee. 12:03
icwiener And I so hoped for a better answer now. :D
[Coke] I just realized I let the daily test runner go for 2 days without even thinking about it. go to check... results for the last 3 days are all identical. Do I have to LOOK for it to work? ;) 12:04
[Coke] yes. parrot must be Configure'd with the libraries present, or it generates code that doesn't use the libraries. 12:06
it doesn't auto-detect at runtime.
# 03/15/2012 - rakudo++ ; niecza (96.22%); pugs (38.97%)
icwiener Jus treconfiguring and rebuilding did not work. I will now start a clean build. 12:07
flussence
.oO( I thought ICU was enabled-by-default? )
 12:08
[Coke] reconfiguring... rakudo? 12:09
no, because it's going to see that you have a working parrot and use it.
Sorry for the pain. 12:10
flussence: if it's not THERE, how can it be enabled?
icwiener It's no pain for me. :)
I have been watching the perl6 development from a distance for quite some time now. I was doing some perl5 about 10 years ago. Unfortunately I am one of those people who let themselves be held back by the fact that "the next version" would be "coming soon". So I never dug in deeper into perl5. Once in a while I start doing some smaller stuff in perl6, like now in Rakudo perl. Now is the first time I do something that is supposed to work when it 12:15
is finished though. So I am basically just starting to get a bit more serious. :)
masak cool. keep us posted. 12:18
icwiener Hmm, I am still getting that error message. 12:21
paste.kde.org/441200/33190053/ Do you maybe see more in it than me? 12:22
gfldex icwiener: try aptitude install libicu-dev 12:25
you will have to rebuild rakudo
icwiener I just did both.
Rebuilt Parrot and Rakudo.
gfldex you may have to do a make distclean for parrot, rakudo and nqp 12:27
[Coke] what command line are you using for parrot and/or rakudo build?
[Coke] tries to remember the parrot invocation for "am I compiled to use icu" 12:28
gfldex i had the same problem a while a go and IIRC i had to rm -r rakudo/ to get it fixed
icwiener gfldex: Because it was too complex for me, I removed the whole source folder, recloned and then ran perl Configure.pl --gen-parrot --prefix=/home/motoko/usr/
[Coke] what does ./install/bin/parrot_config icu_shared say ? 12:30
what does ./install/bin/parrot_config has_icu
(that second one is more useful)
that'll tell us if parrot even has ICU.
icwiener 0 12:31
Hmm ...
[Coke] so, back to parrot, it's not finding your icu.
lemme see...
icwiener: what does "icu_config" say ? 12:32
no pathing, just the raw command.
icwiener parrot_config [ <config-key> [ <config-key> ... ] | --dump | --help ] 12:33
:D
[Coke] icwiener: 08:32 < [Coke]> icwiener: what does "icu_config" say ? 12:34
gfldex isnt it icu-config? 12:34
[Coke] wrong program.
gfldex: whoops! 12:35
[Coke] parrot runs "icu-config" (gfldex++) to see if you have icu. 12:35
if it can't find it, you (apparently) don't.
icwiener Ah, wrong progam, sorry.
There is no icu_config. icu-config gives usage info. 12:36
Do you mean icuinfo?
gfldex icu-config --ldflags 12:38
check if it does something reasonable
icwiener -ldl -lm -L/usr/lib -licui18n -licuuc -licudata -ldl -lm 12:40
icwiener parrot_config --dump lists "libparrot_linkflags => '-L"/home/motoko/compile/rakudo/parrot/blib/lib" -lparrot'" but parrot is not directly inside rakudo/. Is that something that could cause problems? 12:42
moritz the question is if there's a libparrot.so in /home/motoko/compile/rakudo/parrot/blib/lib 12:44
icwiener No, there isn't. 12:46
The folder "parrot/" is not even there.
The file is in ~/usr/lib/libparrot.so here 12:47
moritz did you configure parrot with --prefix=$HOME/usr ? 12:51
icwiener Yep. 12:52
perl Configure.pl --gen-parrot --prefix=/home/motoko/usr/
moritz uhm.
--gen-parrot assumes (or creates) a copy of parrot below the rakudo dir
did you remove that? 12:53
icwiener No. 12:54
It does not do that with my command.
If I remove the --prefix= it fetches it.
moritz then there's something rather fishy going on
was there an older parrot installed in ~/usr ?
icwiener Ah, might be, yes. 12:55
I mean, yes it was. The one I installed yesterday.
moritz ah 12:56
then rakudo's Configure.pl picked it up, and didn't try to build a new one
icwiener And because I did not know how to cleanly rebuild all the parts of the stack, I removed the source folder earlier.
daxim www.infoworld.com/d/application-dev...ail-188648 # hit /. earlier 12:57
moritz so, either remove the install dir too
or force rakudo to build a new parrot 12:58
for example with --gen-parrot=master --prefix=...
"Similarly, when Ruby began it was essentially a more powerful Perl." 13:00
was it? more powerful, I mean
tadzik maybe it's about all this "We have OOP and unicorns and everything" 13:01
moritz ah, right. "everyting is an object", so it must be more powerful.
tadzik right 13:02
it even sounds like power verbalized
colomon Objects: real ultimate power! 13:03
(not)
moritz speaking of objects, I need a bit help with naming something... 13:07
moritz I want to build a system that can notify you when... stuff happens 13:07
for example, if an artist releases a new CD, or goes on tour
usually I'd call that an 'event', but in this context 'event' could also be an artist performing a concert 13:08
so, what's more general than 'event', for when stuff happens?
or should I name it 'event', and find something other for the other event? 13:09
"performance" maybe?
[Coke] +1 on that. 13:10
also, names for programmers don't have to == names for users.
moritz right 13:11
thought it's easier for the programmer if he doesn't have to translate all the time
*though
icwiener Ok, parrot is now there. ICU still not. 13:12
moritz :/ 13:12
icwiener: if you go into the parrot dir, and run 'perl Configure.pl', there's line about icu
would be interesting if that ends with 'yes' or 'no'
not_gerd moritz: NQP build just failed here - see github.com/perl6/nqp/pull/31 13:13
[Coke] moritz: yes, but that just will say the same thing that ./parrot_config has_icu says.
icwiener auto::icu - Is ICU installed....................................yes.
[Coke] what you want more is the verbose output fromthat configure step, I think.
moritz wait, that's *not* the same as has_icu
[Coke] icwiener: and does ./parrot_config has_icu say "1" ?
icwiener Wee ... 13:14
dalek p: 07ce402 | (Gerhard R)++ | src/6model/base64. (2 files):
Make base64.[ch] compile with new error levels
p: ffd778a | moritz++ | src/6model/base64. (2 files):
Merge pull request #31 from gerdr/gerdr/cleanup

Make base64.[ch] compile with new error levels
icwiener Meh.
moritz not_gerd: thanks
not_gerd++
icwiener The parrot in ~/usr says 0, the one in rakudo/ says 1. No idea why it did not install it again. 13:15
moritz not_gerd: I've also given you commit access to nqp
[Coke] ok. so you've built a local parrot that has ICU. is your rakudo usin that parrot? 13:17
if you are building rakudo and telling rakudo to build parrot, it will install it in ./install, not in ~/usr 13:18
icwiener I guess it is using the one in ~/usr since that is in the PATH.
[Coke] (unless you're overriding that too.)
icwiener Even if I use --prefix=?
moritz no, prefix works (I just tried it) 13:19
but if there's an old parrot in the --prefix dir, rakudo doesn't bother to build a new one
so rm ~/usr/bin/parrot*
icwiener Yep, I will do that now. 13:20
moritz and then try again with perl Configure --gen-parrot --prefix=$HOME/usr
icwiener I messed up with the prefix once, so now I have both versions configured with different prefixes.
Is there a make clean command that cleans all the parts? 13:21
[Coke] moritz: seems like we'd want to be able to force a build of anew parrot.
(without having to uninstall it the hard way first.) 13:22
icwiener He mentioned --gen-parrot=master does that?
icwiener I just learned why it is better to keep modules in .perl6/lib ;) 13:45
And it works now. :D 13:46
Thanks all of you.
And sorry for messing up my install.
moritz [Coke]: I think --gen-parrot=SOMETHING always forces a rebuild, but yes, something more convenient wouldn't hurt 14:01
[Coke] works? \o/ 14:02
moritz: as long as there's a way to do it.
dalek p: fdec156 | moritz++ | tools/build/PARROT_REVISION:
bump parrot revision

just to get some better testing of newer birds, no specific reason		 14:05
kudo/nom: 40b20ca | moritz++ | tools/build/NQP_REVISION:
bump NQP revision to get NQP build fixes by not_gerd++
 14:07
dalek kudo/nom: 16bf0f4 | moritz++ | src/ops/perl6.ops:
unbreak the build.
 14:27
dalek ast: e79501d | moritz++ | S05-mass/rx.t:
fudge rakudo regression (ticket already filed)
 14:54
flussence my rakudo compile keeps exploding at the same place :( 15:43
gist.github.com/2050618
oh, problem on my end I think 15:45
yeah it was, I had a one-character local edit in a nqp file and that got git into a wedged state 15:46
flussence nope, failed again in the same place :/ 15:59
sorear goood * #perl6 16:09
moritz flussence: can you nopaste the error? 16:10
good * sorear 16:11
flussence moritz: retried after making sure all my git clones were in order, no change: gist.github.com/2050618/fecbcd12 16:56
sorear o/ fglock 17:16
fglock o/ sorear 17:17
moritz flussence: eeks, it fails in dyncall :/ 17:18
fglock I'm now adding wantarray() to the p5 compiler - it took me a while to get how it works
sjohnson hi 17:21
not_gerd flussence: parrot recently added -Werror=strict-prototypes to CFLAGS, which apparently breaks dyncall 17:24
moritz sigh
moritz I thought it was now parrot policy to test stuff downstream first 17:26
[Coke] moritz: yes, well, not everything is obviously something that needs testing. 17:29
not_gerd also, the build isn't broken everywhere - works fine for me (might be related to gcc version)
moritz not_gerd: point taken 17:30
flussence: what's your GCC version?
not_gerd we probably shouldn't pass parrot's CFLAGS to 3rdparty libs... 17:31
moritz well, we should, to ensure binary compatiblity 17:32
or at least some them
it also doesn't make much sense to write our own configure system to detect all the necessary options a second time
not_gerd flussence: could you try gist.github.com/2051384 as a quick fix (untested) 17:44
actually, it might be the make version and not gcc -- according to the manual, GNU make used to export all variables by default, whereas new version do not 18:05
not_gerd moritz: I just verified thatGNU make 3.82.90 does not pass CFLAGS to the dyncall build system 18:10
fglock pmurias: o/ 18:17
pmurias fglock: hi 18:20
pmurias fglock: how will we mangle intentifiers javascript can't handle? 18:23
like @#?
[Coke] will be sure to get up to date copies of the big three on his laptop before hitting the planes tomorrow. 18:29
pmurias fglock: @# is not a special variable, it just has a wierd name 18:31
fglock pmurias: that's currently not possible, only globals can have special names - but mangling of lexicals can be added at the Var emitter 18:37
flussence undecommutes 18:38
moritz: GCC 4.5.3
moritz flussence: and make? 18:39
flussence make -v says 3.82
moritz flussence: thanks 18:40
not_gerd flussence: gmake -v hopefully says the same thing? 18:41
flussence yep
pmurias fglock: is there a way to pretty print the AST? 18:46
fglock yes, you can use the p5 backend, in the end of ./perlito5.pl add: use Data::Dumper / warn Dumper $ast 18:51
not_gerd flussence: did you see gist.github.com/2051384 , which should work for now until we come up with a proper fix... 18:52
flussence I'll give it a try 18:54
pmurias fglock: should the p5 backend attempt to insert 'our' for no strict code, or just disable strict in the emitted code? 18:58
pmurias explicitly qualifing the variables seems much better then inserting our 18:59
fglock it should emit "plain" perl5, so emit "no strict"/"our" as in the ast 19:01
pmurias "plain" meaning keep to the orginal code as close as possible? 19:02
geekosaur meaning don't assume the existence of a setting that provides a known environment, I think? 19:04
moritz oops, it looks like TPF hasn't been approved for this year's GSOC :( 19:05
rafl: do you know what went wrong? 19:06
rafl i have no idea
rafl the only things i can think are 19:06
a weak-ish ideas page
and us being officially under new management for the first time - last year duke sent in our application 19:07
moritz :(
[Coke] :(
rafl parrot is in, fwiw 19:08
moritz I know
rafl so you guys might get some stuff done with that, i guess
[Coke] well, rakudo projects at least are probably still cool over at parrot.
PerlJam Is it too late for TPF ?
moritz PerlJam: yes, I think so
rafl well - there's always next year and GCI, i suppose
PerlJam :(
rafl but otherwise yes
moritz still, rafl++ for doing all the work
rafl @us++ # so many helped out 19:09
i honestly felt we had the best org application since.. well.. ever
sorear o/ rafl
rafl hey
PerlJam Is there any way we can find out what happened so that it doesn't happen again? 19:10
[Coke] rafl++
yes, there's an IRC meeting later.
moritz and parrot doesn't even have an ideas page listed at www.google-melange.com/gsoc/org/goo...012/parrot
pmurias rafl: :(
rafl moritz: that's the org profile - that's different from the org application and orgs only had one hour before now to fill the profile out 19:11
PerlJam: yes. there's an irc meeting for rejected orgs next week
moritz rafl: oh, ok 19:12
masak rafl: :(
odd that TPF didn't get in.
moritz iirc last year some pretty big and respected orgs didn't get in either 19:13
rafl i just hope it isn't because they wanted to prefer smaller orgs this year
sorear twitter.com/#!/raimue/status/180732762540621824
rafl cause i told a few smallish perl projects specifically to not apply as chances seemed better with the umbrella org 19:13
[Coke] um, the number of orgs is increasing slowly.
rafl sorear: i think that's just melange being shit. the list is growing
[Coke] it was at 53. now at 54. that tweet says 46 19:14
rafl there's an upside, though - not having to deal with melange for half a year :)
[Coke] rafl;did you get a rejection notice? or are you basing it just on the list?
rafl that, and not being burned out at the end of the summer :)
i did get a rejection notice about an hour ago
[Coke] up to 55... 19:15
rafl anyway - pre social of a linux conference coming up
rafl i'm gonna drown my disappointment 19:15
[Coke] ah. The list won't show orgs that have not yet filled our their profile.
rafl sorry everyone for not getting things to work out for us this time around
[Coke] rafl++ for the work. 19:16
rafl: you going to yapc::na?
rafl yes
[Coke] I will happily buy you a drink there. ;)
rafl hopes to be over this by then, not needing a drink anymore :) 19:17
sjohnson mmmm.. drinks 19:21
sjohnson could use some doghair 19:22
sorear o/ sjohnson
fglock wonders how to debug javascript 19:24
sjohnson sorear: hi :3 19:25
fglock: firebug ?
dalek p: 99f34ab | (Gerhard R)++ | 3rdparty/dyncall/GNUmakefile:
Temporary fix for dyncall build failure

See github.com/parrot/parrot/issues/735		 19:26
not_gerd moritz, flussence: ^^
fglock my patch is misbehaving, I'd need to compare the trace with the unpatched version 19:27
fglock (adding contexts to the perl5 compiler) 19:27
felher I hope i'm not indiscreet with that but i wonder: where did pmichaud++ go and is he coming back? :) 19:43
[Coke] he's taking care of family issues. He (and we) hopes he'll get back soon, but no timeline. 19:44
plenty of progress in his absence, though, which is good.
felher [Coke]: yeah, i was just wondering. Thnx ;) 19:45
* :)
sorear felher: 19:48
(spurious send)
felher :) 19:49
PerlJam I think there's an install problem with nqp. After a fresh install of parrot, nqp, and rakudo, when I run rakudo, I get "Missing or wrong version of dependency 'src/stage2/QRegex.nqp'" 19:59
pyrimidine that GSoC melange showing is misleading. The org has to have a completed profile to show up, so more orgs are accepted that just those on that page 20:00
pyrimidine rafl: did you get an email saying Perl Foundation didn't get in? 20:00
rafl: nm, saw that you did 20:01
re: pmichaud's absence, would love to see him back, but I am also glad there is still progress being made 20:04
flussence not_gerd++, patch wfm 20:11
felher If i write a literal zero ('0'), does this create a new scalar container containing zero? (/me is wondering how 'my $x := 0' works) 20:13
pmurias fglock: in the p5 backend do we want the eval to pass the code through perlito? 20:14
not_gerd flussence: the gist one, the NQP commit or both? 20:15
sorear felher: yes, but it's a read-only container 20:16
in Rakudo read-only containers are punned to the corresponding objects to save memory; this seems nontrivial to copy in niecza/C# but if I ever do niecza/C I'll copy it
flussence not_gerd: the gist one 20:17
fglock hmm - maybe not, because in p5-to-p5 we don't keep the compiler state at emit-time, so we can't give eval the right env
flussence (I'm not sure how to build all of the rakudo toolchain manually, so I guess I'll just wait it out) 20:18
not_gerd flussence: the gist as-is breaks the windows build, so I committed a slightly different patch to the NQP repo
pmurias fglock: right env = lexicals + ?
felher sorear: thanks :) 20:19
pmurias fglock: eval(compile_perlito5_to_perl5($code)) would have access to lexicals 20:20
felher sorear: ah, so if i do 'my $x := 0' and then '$x := $x + 1', then '$x + 1' creates a new container containing '$x + 1' and then in bind '$x' to that new container. Is that about right? 20:22
fglock pmurias: no, I was thinking about access to declarations that the compiler might need to know about - but I think the p5 emitter rely on "perl" to figure those out; and there is enough info in global vars too 20:24
so it might just work
sorear felher: yes 20:25
masak right. native $i++ operations will probably have to do a similar re-bind. 20:28
felher sorear: oh, and everytime i do a bind on a non-container (like $x + 1), it's read-only? As opposed to a bind like '$x := y'? 20:31
o/ masak: may i ask: have you already startet working on d2? :) 20:33
felher suddenly remembers something about new-ast-something? 20:34
sorear felher: yes, it's read only
felher sorear++: great, that helped a lot. thnx :)
sorear yay 20:35
masak felher: not really -- I'm blocking on QAST, so I'll want to look at that next. 20:40
felher: presently I'm attacking t4 and surroundings :) 20:41
felher masak: oh, yeah, the p6cc. great :)
masak .oO( ah, letter-digit codes for the initiated )
arnsholt Arglebargle. Why do supposely serious things like PayPal insist on this harebrained "security question" thing? 21:24
Bah! Humbug!
flussence arnsholt: they're running a web2.0, crowdsourced simulation of the "find the resistance between these two nodes of an infinite grid of 1Ω resistors" puzzle... 21:40
arnsholt Heh. Thankfully I've not done any electronics, so I can't fall prey to that =) 21:41
moritz luckily that thing can be solved analytically with discrete Green's Function theory 22:00
it's not easy to find the solution, but when you read the papers it's no rocket science either
jnthn Buenos tardes, #perl6 :) 22:26
*buenas 22:27
tssk
jnthn ain't staying long, just enough to note he's alive after $long-trip :)
fglock o/ jnthn 22:28
jnthn hi fglock
masak jnthn! \o/ 22:31
Buenos Aires, oh tardy one! 22:32
jnthn masak: :P 22:38
jnthn meant to have a relaxing first day after all the flying, then accidentally walked 10km or so
masak I hate when that accidentally happens.
felher Does NQP support typed arrays? I'm wondering because i saw something like: 'has @!CODES;' and and comment above that line 'a stack of code objects'. 22:44
jnthn felher: Not at the moment, no
felher jnthn: ah, okay :) 22:45
jnthn felher: It'd be very hard to statically typed that case anyway 22:46
felher jnthn: 'that case' means that particular array? 22:47
jnthn felher: yeah
That come is compiled before we get around to defining the Perl 6 metamodel, let alone the Code type :)
s/come/code/
felher jnthn: what is in that array anyway? src/core/Code.pm-code-classes? I don't see how they do implement add_phaser. Seems more like src/core/Block.pm-block-classes, maybe? :) 22:50
jnthn felher: Most likely, yeah
I guess a thunk could in theory end up on there 22:51
oh, no, I think that can't happen actually
so yeah, it's probably always Block and subclasses of that
(Block is a direct descendent of code, though, so the comment ain't wrong as such)
felher jnthn: how does that work? src/code/Block.pm doesn't do 'is Code' or something like that? 22:53
jnthn felher: No, that's all wired up in BOOTSTRAP.pm
felher: Some bits of the object system need to be wired up before we start parsing the setting 22:54
BOOTSTRAP.pm does that work, mostly in a BEGIN block these days so we serialize it rather than doing it every startup
felher jnthn: oh, Block.HOW.add_parent(Block, Code) i guess? 22:56
jnthn Right
Time to see if I can find a snack and a drink before taking an early night...back tomorrow o/ 23:01
felher jnthn: sleep well. And thanks for your explanations. :)
masak 'night, jnthn. dream of a relaxing second day, with only 20km of accidental walking :) 23:20

Powered by Cro and the Raku® Programming Language.

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

Thank you!