»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.
00:02 am0c joined 00:07 silent_h_ joined 00:16 ymasory left 00:18 ymasory joined 00:21 silent_h_ left 00:22 rgrau left 00:24 jevin left 00:25 jevin joined, silent_h_ joined 00:35 am0c_ joined, jevin left, am0c left 00:36 jevin joined 00:48 am0c_ is now known as am0c 00:54 jevin left, jevin joined 01:02 silent_h_ left 01:03 am0c left 01:19 am0c joined 01:30 whiteknight left, woosley joined 01:31 woosley left 01:32 woosley joined 01:44 nymacro joined 01:51 plobsing_ joined
sorear good * #perl6 01:53
01:54 plobsing left
colomon \o 01:55
Tene phenny: ask masak how that sliding window alg we discussed worked out for him. 01:56
phenny Tene: I'll pass that on when masak is around.
Tene phenny: thanks
PerlJam re the viking bear looking like masak ... years ago I took an art class where we were paired and how to sculpt our partner in clay. The teacher said that invariably the sculpture ended up looking like the sculptor rather than the subject. So perhaps there's some truth to "making things in our own image" :-) 02:16
s/how/had/
(the teachers words were borne out in our experience too)
02:20 am0c left 02:23 am0c joined 02:27 justatheory left 02:57 Transformer joined 02:58 Transformer left 03:00 bacek left 03:23 Su-Shee_ joined 03:26 Guest17428 joined 03:27 Su-Shee left, Guest17428 left
geekosaur it's certainly true from a cognitive point of view 03:29
03:40 bacek joined 03:44 sftp left 03:51 hudnix left 03:56 ymasory left 04:17 orafu left, orafu joined
sorear mm, niecza bootstrap failed 04:28
looks like there won't be a release this month
v5 will most likely come out May 1 or so after I have time to sort out what broke 04:29
04:31 TheD joined 04:53 kaare__ joined 05:12 Helios` left 05:13 Helios` joined 05:15 xinming_ joined 05:18 xinming left 05:25 birdwindupbird joined 05:34 daxim left 05:37 daxim joined 05:55 bacek left 05:58 Mowah joined 06:10 [Coke] left 06:11 bacek joined 06:26 sivoais joined 06:28 sivoais left, Su-Shee_ is now known as Su-Shee 06:29 sivoais joined 06:35 [Coke] joined, mj41 joined 06:45 dju left, mtk left 06:53 mtk joined 07:16 yegor joined 07:18 yegor left 07:19 elkng joined
moritz jnthn: I get two test failures in ctmo (t/setting/ {hash,resizablepmcarray}) - is that expected? 07:23
07:34 [particle]1 joined
tadzik same here 07:36
07:36 [particle] left 07:39 cosimo left 07:45 am0c_ joined 07:48 am0c left 07:52 am0c_ is now known as am0c 07:55 rhebus_ joined, rhebus left, rhebus_ is now known as rhebus 08:02 yegor joined
jnthn ohhai 08:05
moritz: We don't go fiddling with Parrot types in NQP's setting any more.
moritz: So the tests are no longer valid.
08:08 fhelmberger joined 08:09 snearch joined, fhelmberger left, fhelmberger joined
moritz ok 4 - Join elements 08:11
Could not find sub join
so sub join() is gone?
jnthn moritz: I'm fine with that being put back. 08:12
moritz: Didn't remove it intentionally.
Just needs to go in one of the files in core, or add one.
mathw mutters darkly about namespace pollution mumble mumble
jnthn The only intentional change is we don't monkey-patch Parrot's PMCs any more. :) 08:13
It was a little anti-social, after all. ;)
moritz a bit javascript-esque
jnthn mathw: Relax. It's a setting. It's just the outer lexical scope of your program. :)
mathw: Compile with --setting=MYAWESOMESETTING and you can define your own. ;) 08:14
mathw \o/
or just redefine join if I want to :D
I do like the idea that the setting is just another lexical scope
it makes so much sense
jnthn :)
08:15 am0c_ joined
mathw dives headlong into the dark and unhappy world of network packet captures 08:17
08:18 am0c left, tzhs joined, yegor left, am0c_ is now known as am0c 08:19 donri joined
rhebus \ 08:27
08:30 bacek left 08:33 yegor joined, yegor left 08:35 jjore left 08:39 jjore joined 08:41 pernatiy joined 08:46 bacek joined 08:52 daxim_ joined, daxim left 08:53 [particle]1 left 08:56 jjore left 08:57 cotto left
rhebus gah ssh is being totally weird with my github usernames 09:01
it keeps authenticating as the wrong user, even when i hide that user's id_rsa from ssh 09:02
i don't know how that's even possible
donri does perl6 have reserved keywords that can be syntax errors to use for identifiers?
or does TTIAR "solve" that
moritz donri: it has nothing to do with TTIAR 09:03
donri oh? :)
moritz rakudo: my %h = if => 3; say %h.perl
p6eval rakudo 8533c3: OUTPUT«{"if" => 3}␤»
moritz rakudo: sub if() { say "alive" }; if();
donri i suppose
p6eval rakudo 8533c3: OUTPUT«alive␤»
moritz rakudo: sub if() { say "alive" }; if;
p6eval rakudo 8533c3: OUTPUT«alive␤»
09:03 am0c left
donri i'm thinking in python, where i think TTIAR *would* be a solution if it was used 09:03
mathw What's TTIAR? 09:04
donri two terms in a row
moritz two terms in a row
mathw ah
that
donri nowait there are situations where ttiar isn't enough in python
recv = yield
moritz doesn't see two terms in a row 09:05
donri exactly :P 09:06
i was working under the assumption that all situations involving keywords are ttiar in python
it's an example to the contrary
also i might be misusing the term 'ttiar'
+ i'm probably thinking of 'statements' rather than keywords 09:07
moritz to answer your question, Perl 6 does not have reserved words
donri ok :)
they're rather annoying in python
moritz but it doesn't matter all that much, because variables and attributes have sigils anyway 09:08
donri i guess it makes sense in python's philosophy to keep the grammar simple and the rules obvious but "meh"
moritz: well, subs
moritz so even if perl had reserved words, $next wouldn't collide with next;
jnthn Recently when reading some C# code, I did a double take when I saw a variable whose name started with an @ 09:09
donri do you have to use () to disambiguate then?
jnthn I was like "huh, C# doesn't have sigils?!"
So I went and read the langauge grammar.
moritz rakudo: sub if() { say "alive" }; if
jnthn Turns out that if you want to use a reserved word as a variable name, you put an @ before it. :D
p6eval rakudo 8533c3: OUTPUT«alive␤»
donri rakudo: if 09:10
p6eval rakudo 8533c3: OUTPUT«Could not find sub &if␤ in main program body at line 22:/tmp/6UULO6r94d␤»
donri it's not really ambigous if it's not syntactically an if
moritz jnthn: is that allowed for non-reserved words as well?
09:10 bacek left, bacek joined
jnthn moritz: yes ;) 09:11
donri rakudo: sub next() { say "next!"; }; multi sub dostuff() { next; }; dostuff()
p6eval rakudo 8533c3: OUTPUT«next!␤»
jnthn moritz: So I could start using it on all my arrays :)
donri how do you get the 'next' statement instead?
or is it a sub?
jnthn I think it's a sub and you've just gone and overridden it :)
donri aye 09:12
jnthn Probably you can get the "real" one with SETTING::next();
In theory. Not in current Rakudo.
moritz jnthn: I should troll some C# boards on the internet suggesting to always use @ on variables as a "best" practice
or OUTER::next() or so
jnthn moritz: :D
mathw moritz: yes yes!
do it
donri but is there any situation such that i'm trying
mathw then they can come tell us we need to use SETTING::next() always in case someone overrides it
09:12 [particle] joined
mathw and then we can practise our defenestration 09:12
09:12 cotto joined
donri is @ part of the identifier in C#? 09:13
the convention in python is "append _" but it's part of the name so you have to handle that case everywhere
09:13 jjore joined
moritz "meh" 09:14
jnthn wrestles with PAST::Compiler
mathw thinks Python has a few too many underscores in some places
donri hehe 09:17
mathw I don't get why __init__ 09:18
it looks like some sort of dirty compiler internal to me
why not just reserve all uppercase identifiers as special and call it INIT?
oh wait we did that here already :)
dalek p/ctmo: 7aa35a5 | jonathan++ | src/ops/nqp.ops:
Add a way to take a type object and find out if its representation considers itself as representing a boxed primitive.
 09:19
rhebus can someone add ppotter as a committer so i can update t/spec/S32-str/indent.t with a new plan? 09:47
bbkr_ tadzik: my Pies bootstrap problem occured on Debian 5.0 "Linux dev.v.l 2.6.29.2-vs2.3.0.36.12 #1 SMP Fri May 8 16:22:41 CEST 2009 i686 GNU/Linux", with libtest-harness-perl 3.12-1 installed 09:48
phenny bbkr_: 26 Apr 20:08Z <tadzik> tell bbkr_ what os is that? Is that possible that 0 is not a success on it?
09:48 yegor joined
bbkr_ tadzik: 0 means success, nothing exotic here 09:49
tadzik strange
you can reproduce that, righ?
09:50 birdwindupbird left
bbkr_ checking... 09:50
09:51 birdwindupbird joined
bbkr_ yes 09:52
tadzik oh, I has a bug
bbkr_ i'll check if it has something to do with TODO tests
tadzik my thought is maybe prove gives some non-zero exit status on such occasion
bugs! bugs everywhere! 09:54
09:54 woosley left
rhebus who handles pull requests for perl6/roast? 09:55
moritz rhebus: nobody, which is why we hand out commit bits instead :-) 09:56
rhebus: what's your github ID?
rhebus ppotter
moritz hugme: add ppotter to perl6
hugme hugs ppotter. Welcome to the perl6 github organization
rhebus nice one
dalek ast: 5f51953 | (Philip Potter)++ | S32-str/indent.t:
add plan to S32-str/indent.t
 09:57
moritz rhebus++
rhebus yeah, github-- or ssh-- or something. i have 2 github accounts; rhebus is already a perl6 committer but i just can't get rhebus's identity to auth through ssh atm :( 09:58
moritz you can configure the public ssh key that github uses for auth 10:01
github.com/account -> SSH Public Keys 10:02
tadzik what the hell, pies stopped to even build for me 10:05
zsh: segmentation fault panda install . 10:10
ohgreat
bbkr_ tadzik: confirmed, when prove has any todo() test passed it returns 1 on exit. 10:11
tadzik cool 10:12
what does it return on error?
bbkr_ tadzik: 1
tadzik supercool >{( 10:13
bbkr_ I removed todo() from JSON::Tiny 04-roundtrip.t and Pies installed
tadzik nopaste.snit.ch/42225 10:14
bbkr_ which is weird, because according to TAP::Harness todo test always pass
and rakudo.org is down (blank page). day for disasters :) 10:16
tadzik meh, turns out it's me who broke my parrot :) 10:19
10:27 yegor left 10:34 yegor joined 10:35 yegor left 10:37 elkng left
rhebus t/spec/S19-command-line/dash-e.t fails on my system; i suspect, though, it's my shell and not rakudo which is at fault here 10:42
smolder.parrot.org/app/projects/tap.../15626/389
other people seem to have the same issue, but it's not universal: smolder.parrot.org/app/projects/tes...ory/5/1015 10:43
perl6: say q{\c[LATIN SMALL LETTER A WITH DOT ABOVE]} 10:50
p6eval pugs, rakudo 8533c3, niecza v4-76-g7bf5ef9: OUTPUT«\c[LATIN SMALL LETTER A WITH DOT ABOVE]␤»
rhebus perl6: say q[\c[LATIN SMALL LETTER A WITH DOT ABOVE]]
p6eval niecza v4-76-g7bf5ef9: OUTPUT«Unhandled exception: Unable to resolve method ast in class List␤ at /home/p6eval/niecza/src/NieczaActions.pm6 line 944 (NieczaActions C384_ANON @ 1)␤ at /home/p6eval/niecza/src/NieczaActions.pm6 line 990 (NieczaActions NieczaActions.process_nibble @ 4)␤ at 10:51
../home/p6eval/niecza/sr…
..pugs: OUTPUT«\c[LATIN SMALL LETTER A WITH DOT ABOVE]␤»
..rakudo 8533c3: OUTPUT«===SORRY!===␤Confused at line 22, near "]"␤»
rhebus i suspect that the second test in t/spec/S19-command-line/dash-e.t is incorrect; it looks like a single-quotish context but it expects interpolation 10:52
oh wait, i'm getting confused 10:53
perl6: say q[ȧ]
p6eval rakudo 8533c3, niecza v4-76-g7bf5ef9: OUTPUT«ȧ␤»
..pugs: OUTPUT«decodeUTF8': bad data: '\551'␤ȧ␤»
10:55 cosimo joined
rhebus so yeah, here is the summary of the issue I'm seeing: scsys.co.uk:8002/99941 10:59
scsys.co.uk:8002/99942 shows that echo works fine, so i'm not sure where the encoding issue is creeping in 11:00
huh, --target=parse doesn't give any output for -e programs -- anyone else see this? 11:07
TiMBuS yep its a known issue 11:13
rhebus ok 11:14
is there a bug report system for smolder.parrot.org?
TiMBuS ¯\(°_o)/¯
rhebus :/ 11:15
11:17 ggoebel__ left 11:23 ggoebel__ joined
moritz rhebus: the software itself is on CPAN, so it has a queue in the CPAN rt 11:25
rhebus: as for this particular installation... ask in #parrot (on irc.perl.org)
rhebus ooh, thanks
*scurries*
11:30 jevin left 11:31 woosley joined 11:33 cosimo left 11:37 SHODAN joined 11:45 jevin joined 11:48 jevin left 11:49 jevin joined 11:53 Patterner left
colomon moritz, jnthn, masak: what are you guys using to make presentations? 11:53
moritz colomon: a custom perl script that generates HTML, and the 's5' html+js+CSS framework 11:56
github.com/moritz/perltalk/
I think jnthn uses powerpoint
11:56 pmurias joined
moritz and masak writes a new script for each presentation :-) 11:56
11:57 Psyche^ joined, Psyche^ is now known as Patterner
colomon So the physics file is the source for the talk? 11:58
and can embed p6 code? 11:59
moritz those slides marked with :perl6 are highlighted as Perl 6 12:02
with Text::VimColor, vim and perl6.vim
the file 'talk' contains another talk
yes, very clever file names, I know :-)
colomon moritz++ 12:03
Ack, I don't have HTML::Template::Compiled installed, and I just got a bunch of test failures when I tried to install it.
moritz oh 12:04
there's a new development version which fixes those on perl 5.12
HTML-Template-Compiled-0.94_002
cpanm HTML-Template-Compiled-0.94_002 # should help
colomon moritz++ 12:05
errr, that doesn't actually work? "Cannot make HTML-Template-Compiled-0.94_002, don't know what it is." 12:06
moritz cpanm search.cpan.org/CPAN/authors/id/T/T...002.tar.gz 12:07
moritz tries to get the maintainer to release the last version as non-devel
colomon is cpanm a typo or something different from cpan? 12:10
arnsholt CPAN minus IIRC
moritz App::cpanminus
arnsholt A simpler CPAN client
colomon ah
takadonet colomon: you will love it 12:11
moritz very convenient, asks far fewer (I think "none") questions
mathw it actually works
takadonet morning all
moritz and it Just Works[tm]
takadonet faster too
mathw which I find is a big plus point
colomon is now using cpan to get cpanm. ;)
moritz colomon: that's the last thing I used cpan for :-)
BinGOs uses cpanm to bootstrap cpanp when he needs to. 12:12
colomon ah, yes, that seems to be helping matters greatly. :)
BinGOs cpanm doesn't do test reporting so is of no use to me apart from bootstrapping something I can use to do test reporting. 12:13
12:19 mikemol left, mikemol joined 12:22 MayDaniel joined 12:28 sftp joined 12:37 ymasory joined 12:40 MayDaniel left 12:48 hudnix joined 13:03 ymasory left 13:06 eml joined 13:13 leprevost joined
dalek ar: 589a29d | pmichaud++ | skel/docs/announce/2011.04:
Add 2011.04 release announcement.
 13:16
PerlJam looks forward to the time when 6model and GSoC gives us some of those not-quite-there features. 13:18
dalek ar: 8c68a22 | pmichaud++ | / (2 files):
Bump Makefile and gen_parrot versions.
 13:21
13:26 maja joined 13:29 noganex joined 13:32 noganex_ left 13:34 maja left
jnthn back 13:36
takadonet jnthn: wb 13:39
mikemol is pleased the pi spigot sequence has been so interesting.
jnthn My talks from OSDC.TW are available as video! Includes Perl 6 Signatures: video.osdc.tw/1301385198.mp4 and Inside A Compiler: video.osdc.tw/1301386903.mp4 13:43
moritz is the signature talk the same as at YAPC::EU? 13:44
jnthn Yes 13:45
13:47 leprevost left 13:48 pochi left
colomon jnthn: do you have all your talks located in one convenient place on the web? I'd like to point people to them in my presentation. :) 13:51
jnthn colomon: www.jnthn.net/articles.shtml
colomon jnthn++
moritz wonders if that's linked to from perl6.org 13:52
arnsholt jnthn: The TPJ link at the bottom of that page is dead 13:54
Apparently tpj.com is for sale
jnthn arnsholt: Ah...
I don't think I've seen the bottom of that page for a loooong time :)
arnsholt Yeah, the length of that page is quite impressive 13:55
dalek href="https://perl6.org:">perl6.org: c64b77f | moritz++ | source/documentation/index.html:
[documentation] lik to the excellent presentations by jnhtn++
 14:03
14:04 Moukeddar joined 14:05 leprevost joined, leprevost left 14:06 leprevost joined 14:07 spq joined 14:12 am0c joined
tadzik oh oh, star is today? 14:13
colomon looks like. \o/ 14:14
tadzik dang, I wanted to push Panda there 14:15
moritz hurry up, maybe it's not too late
14:15 alester joined
tadzik pmichaud: how much time until the Star release? 14:15
14:15 leprevost left
pmichaud tadzik: I was wondering if Panda is ready for release. I'm thinking of including it 14:16
14:16 leprevost joined
pmichaud tadzik: if you think Panda should be included, I'm inclined to agree 14:17
right now I'm just doing some basic testing and building stuff, so I think there's time
moritz it's the only working module installer we have :-)
pmichaud (working installer) Yes, I know. :-)
PerlJam We need more animal named software so that we can have the "lions and tigers and bears" release of Rakudo Star :) 14:18
tadzik pmichaud: the problem is it started segfaulting for me today. Surely Parrot's fault, but disturbing anyway 14:20
see nopaste.snit.ch/42225
pmichaud tadzik: using Parrot 3.3.0 ?
tadzik aye 14:21
well, the git version from --gen-parrot
I'll to remove it completely, bootstrap and install a few things
pmichaud tadzik: I'll test it here also. But my plan was to try to include Panda in the release.
tadzik cool 14:22
pmichaud afk for a bit (errands)
tadzik oh, I was also thinking about removing some of the modules from the META.list, so the star-panda (oh the pun!) will contain only those installable and working (more or less)
moritz +1 from me 14:23
tadzik oh good, it didn't segfault now, while bootstraping 14:25
moritz: I'm going to remove all the modules from SHELTER from the META.list, they'll still live in the shelter and in the old projects.list, what do you think? 14:26
moritz tadzik: do they install? 14:27
tadzik moritz: not all of them
I can try some
I know a plenty of them have some missing dependencies
okay, after a full reinstalation of panda and deps it doesn't segfault anymore 14:28
14:34 wtw left, leprevost left
tadzik uh-oh, does LWP::Simple build for anyone? 14:34
moritz it tested fine last I tried 14:35
tadzik t/get-rakudo-org.t .............. Method 'match' not found for invocant of class ''
oh wait, isn't rakudo.org down?
I believe LWP::Simple needs a better testing target
moritz just a sec, I can fix it 14:36
14:37 snearch_ joined
moritz tadzik: fix pushed 14:38
tadzik moritz: pushed where? 14:39
14:39 am0c left
moritz tadzik: [email@hidden.address] 14:39
tadzik ok, works 14:40
14:41 twinshadow left, hercynium joined
pmichaud ...why is rakudo.org down? 14:41
moritz has no idea, but it has happened a few times recently 14:42
pmichaud That's... not good. 14:43
moritz agreed.
dalek osystem: 2d8ceac | tadzik++ | / (2 files):
Module::Tools has no tests and is broken, kill it with fire
pmichaud I think I'll set up a new site for rakudo on my server (WordPress)
tadzik colomon: Testing is broken, isn't it? 14:44
colomon tadzik: not so far as I know...
pmichaud Maybe we can get Andy to point rakudo.org to a new location
tadzik colomon: .notdef is used
alester pmichaud: No, it's not good.
tadzik rakudo: "dupa".notdef
p6eval rakudo 8533c3: OUTPUT«.notdef is deprecated, please use negated .defined instead␤ in 'Mu::notdef' at line 1293:CORE.setting␤ in main program body at line 22:/tmp/s2NfPypQS9␤»
alester I am going to point elsewhere. 14:45
I'm not at all in love with Drupal.
pmichaud I agree, Drupal has been painful.
alester s/going/willing/
colomon tadzik: looking into it now
pmichaud anyway, I'll set up wordpress on my server today and we can start populating it
alester pmichaud: What do you need form me?
pmichaud (I have wordpress installed already... I just need to set up the account(s))
alester If there's some convert-o-dump-o-tron that needs DB access or whatever, let me know. 14:46
pmichaud: I also have WordPress installed.
But if you wanna drive, that's fine.
pmichaud alester: I'll use mine for now... I think I'm a little less squeamish about giving out access to my server than you are :)
alester pmichaud: Probably ture. 14:47
Thanks for driving.
pmichaud anyway, yeah -- the biggest question is how much of the old content from rakudo.org we feel we need to preserve (and how to get it)
alester And then I can shoot Drupal. :-) 14:48
moritz it would be nice to preserve some URLs (not necessarily pages) that are linked everywhere
like how-to-get-rakudo
tadzik FWIW, I'm now installing every module in the ecosystem to see which are working to at least some extend
colomon tadzik: Testing fix pushed 14:49
moritz tadzik++
tadzik colomon++, great
pmichaud I'm certain we can preserve urls if rakudo.org is forwarded somewhere
14:50 Moukeddar left 14:51 MayDaniel joined 14:53 am0c joined, snearch left 14:55 tzhs left, Holy_Cow joined 14:59 Sarten-X left
tadzik also, Perl6-Term--ANSIColor is unmaintained since I remember, and all of my attempts to contact its author failed. Mind if I remove that? We have another Term::ANSIColor, tested and maintained 15:00
(and mine, but that's a coincidence)
pmichaud +1
moritz tadzik: go right ahead 15:02
tadzik: if the author complains, we can tell him to fix his module and re-add it to the index file :-)
tadzik :)
I hope masak doesn't mind if I remove pun 15:03
15:03 MayDaniel left, icwiener joined
tadzik ==> Succesfully installed Proto 15:04
nice
dalek osystem: ccc953a | tadzik++ | META.list:
Remove some obsolete modules
15:10 ymasory joined, Sarten-X joined
tadzik wow, perl6-ioc looks like Bread::Board for Perl 6. It seems a bit unmaintaned, but it builds and tests fine 15:11
this smokebuilding experience should be blagged about today 15:12
moritz it's rather new, iirc
tadzik hmm, right, January this year
seen jasonmay 15:13
aloha jasonmay was last seen in #perl6 12 days 22 hours ago saying ".oO( has $.x is service lifecycle Singleton depends-on <y z> )".
tadzik oh, a resident :)
phenny: tell jasonmay hey, your perl6-ioc looks quite nice, would you mind adding a META.info to it thus fulfilling the latest modules' spec? 15:14
phenny tadzik: I'll pass that on when jasonmay is around.
jasonmay sure
phenny jasonmay: 15:14Z <tadzik> tell jasonmay hey, your perl6-ioc looks quite nice, would you mind adding a META.info to it thus fulfilling the latest modules' spec?
tadzik oh, cool :)
do we really need perl6-examples in the ecosystem? I don't remember when that was buildable 15:15
moritz no, IMHO 15:16
tadzik by "in the ecosystem" I really mean "available for installation"
dalek osystem: 6e6bc12 | tadzik++ | META.list:
Remove perl6-examples
tadzik okay, tried all of them
results: nopaste.snit.ch/42233
the notworking ones are not so common thing to do any mass removal imho 15:17
maybe there will appear someone who will want to fix them even
jasonmay tadzik: pastie.org/1839606 -- would this be sufficient 15:18
I basically cargo culted from one of the other files in the ecosystem
html-template I think
tadzik that's a good way to do it
jasonmay: that should be source-type and source-url, the former being optional. Otherwise it's perfect 15:19
I don't think otherwise is the word I'm looking for
moritz we're switching from 'repo' to 'source' because tarballs are not repos
jasonmay pushed 15:20
tadzik jasonmay++
15:21 Holy_Cow left
dalek osystem: ce89679 | tadzik++ | META.list:
IoC doesn't need no SHELTER
 15:21
pmichaud the latest rakudo star test build runs significantly slower on my system than the 2011.01 build 15:23
like, about 40% slower
moritz wow
pmichaud I'll run again in case I was doing something weird
moritz what does parrot_config gc say ?
tadzik wasn't ms2 default for 3.3? 15:24
moritz oh, the config key doesn't exist
yes
pmichaud: make sure to compile with --gc=gms
pmichaud but the rakudo configure script assumes --gc=gms
moritz s/compile/configure parrot/
pmichaud unless the star confi..... right
Star doesn't use Rakudo's configure to build parrot
okay, fixing. 15:25
tadzik okay, I declare panda star-ready
it has some LTA things, but it's usable more than annoying
pmichaud tadzik: could you update the star makefile(s) as appropriate?
tadzik: or do you need me to do that?
tadzik pmichaud: I think I can do that, yes
pmichaud: I can also fill in some of the NEWS, if that's ok for you
pmichaud please
tadzik++ 15:26
tadzik no problem with that
pmichaud: mind if I also remove those modules which proved to be not working? 15:27
pmichaud tadzik: please do that. make sure we note the removals in the announcement
tadzik pmichaud: I'd take care of it 15:28
dalek ar: f742a45 | pmichaud++ | skel/Configure.pl:
Add --gc=gms to Parrot configuration options.
 15:33
colomon pmichaud: Current slowness is consistent with my benchmarks. Alas, still haven't had a proper chance to look into it. 15:34
pmichaud I'm trying again now with --gc=gms 15:35
jnthn *drum roll* 15:36
dalek p/ctmo: 0ff3ff4 | jonathan++ | src/ (2 files):
First cut of natively typed attributes (int/num/str). Code gen could be better, but it works.
colomon jnthn++: \o/
pmichaud parrot_config option is "gc_type", fwiw. 15:37
(I'm continually amazed at how many times Parrot uses multiple names for the same option.)
and its value is "GMS" instead of "gms" 15:38
alester pmichaud: WHY NO HAVE WORDPRESS DONE YET
HURRY UP, QUICKER
pmichaud (I'm continually amazed at Parrot's inconsistent use of capitals.)
dalek ar: c38a2a6 | tadzik++ | / (3 files):
Remove some of the not working modules, add some new. Note the changes in the announcement
alester IT BEEN WHOLE 49 MINUTES NOW
jnthn Uniquitous language fail...
15:39 Moukeddar_ joined
jnthn colomon: I suspect some other leak is going on somewhere, but initial testing suggests that if you make a complex type with objects attributes and one with two native num attributes, then half a million instances of complex come out at 20 meg less or so. 15:40
colomon jnthn: !!!!!!!
tadzik oh my 15:41
that's super-cool
15:41 Moukeddar_ left
jnthn colomon: Note that this still has 500,000 objects hanging around in memory. Once we do packed arrays, my complex @foo; can pack them in so it's one big chunk. 15:41
But even so, this is a nice bit of progress :)
colomon wants 6model to land in Rakudo very badly... 15:42
jnthn colomon: My NQP pre-req task list is shrinking rapidly.
tadzik jnthn: if you try very hard, would you be able to ship it before Star? :P
15:42 yegor joined
tadzik I mean, you know, everybody's waiting and stuff ;) 15:43
jnthn Sure. :P
So long as I don't have to say *which* Star release. :P
tadzik aw, snap!
dalek ar: 227e50e | tadzik++ | skel/docs/announce/2011.04:
Add some advertising to the announcement
jnthn So, what's next... 15:45
jnthn glances his todo list
eww :)
Whenever my next task has "circularity" in it, I know it's gonna hurt some... 15:46
15:46 toebu left 15:58 yegor left 16:01 SHODAN left
pmichaud anyone who wants to author/admin on the new rakudo.org, please go to rakudoperl.org/wp/ and click the "Register" link at the bottom-right 16:01
16:02 mj41 left
jnthn pmichaud: registered (jnthn) 16:03
pmichaud jnthn/colomon: you now have admin privs 16:05
colomon thanks
jnthn yay 16:06
pmichaud++
PerlJam has registered also 16:07
(does this mean that rakudo.org is going away or what?)
16:08 Su-Shee left
pmichaud rakudoperl.org is a domain I already own; we'll hopefully get rakudo.org to forward to rakudoperl.org or something like that 16:08
16:09 molaf_ left 16:12 ymasory left, Su-Shee joined
pmichaud PerlJam/Util: updated 16:12
I have a lunch meeting I have to go to... I'll bbi120. Feel free to update things on the rakudoperl site as you see fit. 16:13
16:19 risou joined, risou left 16:20 risou joined 16:24 yegor joined 16:29 ymasory joined 16:30 risou_ joined, woosley left 16:31 cdarroch joined, cdarroch left, cdarroch joined 16:32 noganex_ joined 16:33 noganex left, risou left 16:45 birdwindupbird left 16:50 PerlJam left, PerlJam joined
dalek p/ctmo: cc013b6 | jonathan++ | src/ (5 files):
Setup the very basic KnowHOWAttribute in 6model core rather than leaving every HLL the problem of setting up an initial one. It's especially a pain if you're going to have a compile time meta-model and need the attribute meta-object available to the compiler. (FWIW, nqpclr ended up with this in core too, though with a different initial motivation.)
 16:53
p/ctmo: 2af2457 | jonathan++ | src/NQP/ (2 files):
Start using KnowHOWAttribute from the 6model core instead. Means we now create all meta-attributes during the compile. Toss the knowhow-specific workarounds.
16:53 nymacro left 16:57 Sarten-X left
sunnavy date 16:59
oops 17:00
17:00 bacek left 17:01 bacek joined 17:04 Chillance joined, leprevost joined 17:06 pmurias left 17:07 bacek left 17:09 wallberg joined 17:10 molaf joined 17:13 rhebus left 17:15 rhebus joined 17:20 Axius joined
dalek p/ctmo: 28dbe5e | jonathan++ | src/stage0/ (6 files):
Update bootstrap with various attribute changes.
 17:20
p/ctmo: 74ac547 | jonathan++ | / (2 files):
Toss the now-unused KnowHOWAttribute.pm.
p/ctmo: 835b51e | jonathan++ | src/Regex/ (3 files):
Fix lookups of meta-objects in regex library.
17:24 yegor_ joined 17:26 yegor left 17:30 takadonet left 17:32 yegor_ left 17:33 rhebus left, fhelmberger left 17:35 rhebus joined 17:51 masak joined
masak oh hai, #perl6 17:51
phenny masak: 01:56Z <Tene> ask masak how that sliding window alg we discussed worked out for him.
tadzik oh hi masak 17:53
masak Tene: heh, I already wrote a solution for that problem -- was curious how to do it "right" for next time. :) right now I've moved on to other problems at $work... but I'll keep on the lookout for similar problems.
hai tadzik
17:54 leprevost left
sorear hi masak 17:54
17:54 justatheory joined
masak Tene: I like the syntax I arrived at at the bottom of gist.github.com/942868 -- it exposes the essential complexity and hides the accidental complexity. the remaining problems I have it are miniscule. 17:54
sorear: hi there :)
sorear: if I were to toy around with the grammar extensions and AST transformations necessary to make something like that gist DTRT, should I try doing it in Niecza, or in the new 6model-enhanced nqp? 17:57
masak keeps tugging at loose end of the future :) 17:59
18:03 rhebus left 18:07 rhebus joined
sorear I don't totally understand what OVER is supposed to do here 18:07
also I don't know what new-nqp is capable of
18:08 MayDaniel joined
dalek p/ctmo: 177dadd | jonathan++ | src/HLL/SerializationContextBuilder.pm:
Toss a bunch of legacy package installations that we no longer need (this means that all type lookups are certainly working through the new packages).
 18:10
p/ctmo: 5d65440 | jonathan++ | src/NQP/Actions.pm:
Update some package lookups.
p/ctmo: c408fe7 | jonathan++ | src/Regex/ (3 files):
More legacy package lookup/installation cleanups.
sorear jnthn: grr, I'm gonna have so much catching up to do when I get my tuits back :D
jnthn sorear: :) 18:11
18:15 risou_ left 18:16 newbee joined 18:17 Sarten-X joined
masak sorear: OVER is supposed to hide the bookkeeping that the upper, uncommented code exposes. the two pieces of code are meant to do the same thing. 18:17
jnthn: how difficult would it be to introduce something like OVER in nqp at this point? just as an experiment. 18:18
18:18 Moukeddar joined
masak Moukeddar: arf! 18:18
tadzik was that an evil panda sound? 18:19
Moukeddar arf!?
masak keeps silent and practices a piercing look
Moukeddar what happened to ROAAAR!
18:19 molaf left
TimToady perl6: my @fib := 0, 1, @fib Z+ @fib[1..*]; say @fib[10]; 18:20
masak Moukeddar: in here, we tend to yap rather than roar :)
Moukeddar shell pirecing look ?
p6eval rakudo 8533c3, niecza v4-76-g7bf5ef9: OUTPUT«(timeout)»
..pugs: OUTPUT«␤»
TimToady pugs: my @fib := (0, 1, @fib Z+ @fib[1..*]); say @fib[10];
p6eval pugs: OUTPUT«␤»
masak TimToady: I've been told that that's never going to work, because there are no closures in there that will delay evaluation the way @fib needs to.
TimToady: I think sorear told me that. 18:21
TimToady um, the lists are lazy
newbee hi... what are the datatypes for date, time, and timedate in perl6? I know that Date and Duration works, bit Time don't: use v6; class Task; has Str $.command is rw; has Time $.exec_time is rw; has Date $.exec_date is rw; has Duration $.exec_timeout is rw;
It says that Time is not a valid datatype
TimToady and binding is not eager 18:22
masak newbee: Time does make much sense on its own. Date does.
jnthn masak: use NQP; class MyNQPGrammar is NQP::Grammar { ... }; class MyNQPActions is NQP::Actions { ... }; sub MAIN(*@ARGS) { ...same as in src/NQP/Compiler.pm but put your subclassed grammar/actions in instead... }
TimToady and the problem in rakudo is just that subscripting isn't lazy, I think
masak newbee: if you want, you could declare 'class Time { has $.hour; has $.minute; has $.second }' 18:23
newbee: but we don't do that for you because you might have different requirements beyond that.
newbee My intention was to use a "native" datatype for Time
masak newbee: maybe you're looking for Instant?
newbee is there any?
TimToady rakudo: say now.WHAT 18:24
tadzik Sure. Like counterspell, or terror :>
p6eval rakudo 8533c3: OUTPUT«Instant()␤»
tadzik I wonder if that joke isn't too hermetic
sorear TimToady: lists may be lazy but function calls aren't
newbee mmmm i see... Instant does that.. Much better than the generic "Time".. Thanks :D
sorear &infix:<,> has to evaluate both arguments right away
masak newbee: if it's any lead, "Instant" is for atomic time, whereas "DateTime" is for civil time. 18:25
TimToady sorear: fine, but Z+ is returning a lazy list that commits to nothing until munched, supposedly 18:26
er, hmm, precedence problem
perl6: my @fib := 0, 1, (@fib Z+ @fib[1..*]); say @fib[10];
masak the precedence with X and Z often surprises. but it's one of those can't-please-everyone problems.
p6eval rakudo 8533c3, niecza v4-76-g7bf5ef9: OUTPUT«(timeout)» 18:27
..pugs: OUTPUT«␤»
TimToady that's what I meant
18:27 aindilis joined
newbee i just need a Time without relation to a date.. i want to do a "schedule" system. For example, run this task at "14:30" for certain dates. 18:27
masak newbee: then I advise you to create a custom Time class.
sorear if we allow general functions like &zipwith to delay evaluation, we're opening a can of worms
masak or TimeOfDay or something.
sorear all the performance problems of Haskell without the elegance :)
pmichaud good afternoon, #perl6 18:28
masak pmichaud! \o/
jnthn o/ again, pmichaud :)
tadzik welcome back :)
sorear masak: I don't even understand what either version of the gist is supposed to do. 18:29
jnthn masak: fwiw, if you play with new nqp, I suggest doing so against the master branch. ctmo is where I feel free to break stuff. :)
TimToady sorear: I don't know what you mean by "delay evaluation"; the zipwith is called, but returns a lazy list that commits to nothing until munched
jnthn masak: When something becomes "good" I merge it to master. The native attrs are just about to get merged, for example.
TimToady I see no performance problems there
pmichaud hmm, R* with the new garbage collector doesn't seem to be significantly faster than 2011.01 :/
jnthn pmichaud: :/
masak jnthn: understood. 18:30
sorear: oh, sorry. :/ I'll attempt an explanation.
sorear TimToady: the zipwith is called, but it receives the *old* @fib as an argument
jnthn masak: In theory you should have no problems at all subclassing grammar and actions and setting up NQP::Compiler to use your new ones.
sorear @fib isn't rebound until zipwith and , return
tadzik I wonder what did slow Rakudo down 18:31
masak sorear: the general problem is this: given an incoming stream of ordered records-with-fields, where fields to the right "vary" faster than fields to the left (odometer-style), devise an algorithm that does something within each contigous chunk of field values.
pmichaud I'll run some tests.
18:31 xinming_ left
masak sorear: in this particular case, the incoming fields are person, shop, and item. for each person, print (in the correct order) that person's shops. print as soon as possible. 18:32
TimToady sorear: okay I see your point
masak sorear: thing is, everything will lag until the iteration after, when we know that a new person was found.
sorear: I want to avoid having to think about that in the code. 18:33
dalek Heuristic branch merge: pushed 21 commits to nqp by jnthn
TimToady perl6: my @fib.plan: 0, 1, (@fib Z+ @fib[1..*]); say @fib[10]; 18:34
that's more like
p6eval niecza v4-76-g7bf5ef9: OUTPUT«(timeout)»
..pugs: OUTPUT«*** ␤ Unexpected "@fib"␤ expecting "=", ".=", context, ":" or "("␤ at /tmp/Vu1zPIfhS1 line 1, column 4␤»
..rakudo 8533c3: OUTPUT«===SORRY!===␤Confused at line 22, near "my @fib.pl"␤»
dalek p/ctmo: 0a24347 | jonathan++ | src/stage0/ (7 files):
Update bootstrap.
masak .plan ? 18:35
TimToady see Containers.pod
sorear for .lines.map({[.split(',')]}).group(*.[0]) -> $person, @tuples { say ($person => @tuples.map(*.[1]).uniq) } # I would write it like this
pmichaud alester: is there a way to export the article contents from the rakudo.org drupal instance? I'm not too concerned with preserving comments 18:36
TimToady perl6: constant @fib = 0, 1, (@fib Z+ @fib[1..*]); say @fib[10]; # what we really want here 18:37
alester pmichaud: haven't looked
p6eval rakudo 8533c3: OUTPUT«===SORRY!===␤Constant type declarator not yet implemented at line 22, near "= 0, 1, (@"␤»
..niecza v4-76-g7bf5ef9: OUTPUT«(timeout)»
..pugs: OUTPUT«pugs: *** Undeclared variable: ("@fib",MkPad (padToList []),[PCompiling {pc_pad = MkMPad {mp_id = 140387682278377, mp_pad = <ref:0x7fae8de4a431>}}])␤ at /tmp/ahoZrnsUjj line 1, column 23-43␤»
alester Have you googled for a drupal-to-wordpress conversion?
pmichaud yes
alester I imagine it's a solved problem.
18:37 xinming joined
pmichaud no 18:37
alester No exiting tools?
Wow.
18:38 Axius left
pmichaud there are tools for older versions of wordpress and drupal 18:38
I don't find anything current
masak sorear: .group ?
pmichaud and I find lots of articles that say the tools don't really work
sorear masak: a hypothetical Perl 6 version of the Haskell groupBy
masak: sort of like .classify, but lazy and expects sorted input
TimToady sorear: I believe "constant @fib =" should bind the lazy list at compile time 18:39
pmichaud so, I'll be happy to just get the article text somewhere and re-build the articles manually from that
there aren't that many, iirc
18:40 Moukeddar left, yegor joined
masak TimToady: so 'constant @fib =' and 'constant @fib :=' are semantically identical? 18:44
TimToady um
std: constant @fib := 42; 18:45
p6eval std 3468e14: OUTPUT«===SORRY!===␤Malformed constant at /tmp/CojEtiERau line 1:␤------> constant @fib :⏏= 42;␤ expecting any of:␤ coloncircumfix␤ signature␤Parse failed␤FAILED 00:01 118m␤»
tadzik arnsholt, ingy, moritz and quietfanatic are people with 2 or more modules in the SHELTER
arnsholt SHELTER?
tadzik aye, where the not-updated-to-the-new-spec modules live 18:46
mentioned in ttjjss.wordpress.com/2011/04/18/new...les-world/
18:47 MayDaniel left
arnsholt SQL::Grammar can be deleted 18:47
I never got anywhere with that, and it can be safely deleted 18:48
tadzik alright
masak TimToady: so binding and constants don't mix?
arnsholt Gimme a sec and I'll update the Viterbi one
tadzik alright. Familiar with META.info's?
arnsholt Not yet =) 18:49
18:49 leprevost joined
tadzik :) 18:49
You can either read the spec, or just cargocult it from some panda-badged module on modules.perl6.org 18:50
arnsholt Looking at the spec now, and then I'll probably cargo-cult =)
sorear the way constants are parsed is a tad dubious 18:52
tadzik I have a feeling that only the neutro's META.info has been actually written from scratch. I should've put some funky whitespace in it, and now give a special prize for everyone without that character in their META.infos :) 18:53
18:53 risou joined
sorear '5 * constant $x = 2' parses as (5 * constant $x) = 2, but the latter isn't valid syntax 18:53
masak it isn't? 18:54
std: 5 * constant $c = 2
p6eval std 3468e14: OUTPUT«ok 00:01 120m␤»
masak std: (5 * constant $x) = 2
p6eval std 3468e14: OUTPUT«===SORRY!===␤Malformed constant at /tmp/LdYyerpRjm line 1:␤------> (5 * constant $x⏏) = 2␤ expecting any of:␤ desigilname␤ trait␤ twigil␤Parse failed␤FAILED 00:01 121m␤»
masak huh.
something's odd there, fershure. 18:55
18:56 impious joined
arnsholt How do Panda and ufo interact? 18:58
tadzik they don't
arnsholt Right. No deps for my module then
18:58 impious left
masak right. ufo was never meant to be a dep of anything. 18:59
I seem to gravitate to writing that kind of module, for some reason :)
arnsholt There. New version pushed to github 19:01
tadzik arnsholt++
TimToady re pseudo-assignment to declarators, see S03:4802
pseudo assignment has nothing to do with normal assignment, even wrt precedence 19:02
tadzik arnsholt: fwiw, source-type is optional, Panda is smart enought to figure that out from the url
TimToady declarators all look for '=' as part of their "term"
arnsholt Yeah, I noticed. But the files I cargo-culted from had it, so y'know... =) 19:03
tadzik :)
I'know :)
TimToady (that is, declarators that take initializers, not things like 'sub')
arnsholt So now all that's missing is that I remove the module from SHELTER and update META.list in perl6/ecosystem? 19:04
dalek osystem: 1ba36bf | tadzik++ | / (5 files):
Algorithm::Viterbi does not live in a SHELTER anymore. Clean the SHELTER from some (re)moved modules
 19:05
arnsholt Or that. Cheers!
newbee hi again :) I was creating the TimeOfDay class: use v6; class TimeOfDay { has $.hour; has $.minute; has $.seconds; } 19:06
pmichaud alester: if you can put rakudo.org up long enough for me to do a webscrape of its contents, that would be okay
alester pmichaud: Oh, shoot, I'm sorry, forgot about you. Real Job got in the way. 19:07
let me see what I can do re: export
pmichaud np, just letting you know options
no real hurry
newbee now, i know i can put constraints to the data using where, for example, 0 >= $hour <= 23... but i dont wan't the program to fail if i assing a 61 to $.hour
how can i make "accesors" for those data members?
19:08 am0c left
tadzik newbee: you already did 19:08
19:08 rhebus_ joined
tadzik $.foo is an attribute foo, with a public accessor 19:08
rakudo: class Foo { has $.bar = 5 }; say Foo.new.bar
p6eval rakudo 8533c3: OUTPUT«5␤»
19:08 rhebus left, rhebus_ is now known as rhebus
sorear pmichaud: what's up with rakudo.org? 19:08
pmichaud sorear: apparently it's down, so I'm moving it to a wordpress-based site on my server 19:09
alester "What's up" is a rare condition. :-)
tadzik TiMBuS! TiMBuS! Your META.info is broken!
alester is it down NOW?
dammit
pmichaud it was down earlier when I tried it 19:10
sorear Why does it keep going down?
alester sorear: You are far too late to this party. :-)
If I knew, I'd fix it.
sorear alester: I know, hence the questions :)
:/ 19:11
alester oh man, I might be able to kill my mysql instance, too.
pmichaud: want a mysqldump?
I'll paste it into the channel.
sorear I hope you mean a link 19:12
alester sorear: I kid.
Wuss.
Can't handle thousands of lines of SQL INSERTs?
newbee so, i did accesors just by declaring the variable... but, how can i send a warning if i pass 61 to $.hour ??
pmichaud alester: a mysqldump would work okay, yes 19:13
tadzik newbee: you can supply your own BUILD submethod which will validate the attributes
alester pmichaud: OK, can't look right this second. 19:14
but I will
tadzik rakudo: class Mytime { has $.hour; submethod BUILD { callsame; die "how many? Are you nuts?" unless 0 < $.hour < 24; } }; Mytime.new(hour => 77)
p6eval rakudo 8533c3: OUTPUT«how many? Are you nuts?␤ in 'Mytime::BUILD' at line 22:/tmp/2JjrX55tCA␤ in main program body at line 22:/tmp/2JjrX55tCA␤»
tadzik rakudo: class Mytime { has $.hour; submethod BUILD { callsame; die "how many? Are you nuts?" unless 0 < $.hour < 24; } }; Mytime.new(hour => 17)
p6eval rakudo 8533c3: ( no output )
pmichaud alester: great, no rush 19:15
alester pmichaud: I just saw them on April 12th.
tadzik newbee: keep in mind that is only during the object creation
github++ # "fork and edit this file" 19:16
alester github++ # Just in general.
masak github++ # happily paying customer 19:18
19:18 rhebus_ joined 19:19 rhebus left, rhebus_ is now known as rhebus, mj41 joined 19:21 birdwindupbird joined, am0c joined
newbee but what happens if i change $.hour after the creation? I understand that BUILD runs only once... 19:22
19:23 leprevost left
tadzik newbee: if you don't add "is rw" to "has" you can't change it 19:23
and Time sounds like something I'd prefer immutable
I believe they're immutable in Perl 6 to
19:23 risou left
masak DateTime is mutable; Date is immutable. 19:24
19:24 yegor_ joined
masak but yes, I agree with tadzik that Time sounds like it could well be immutable. 19:24
19:25 yegor left
pmichaud ...I thought the new garbage collector was to give a significant speed improvement, but for some reason I'm not seeing it here. 19:26
at least, I'm not seeing it in the spectests. 19:27
newbee ok, lets put this example then: class Soldier { has $.health is rw } and i want to perform some action if health goes beyond 100, all this without a special method. Just by saying $Asoldier.health = 120;
tadzik newbee: that's either hard or impossible in the current state of Rakudo 19:28
pmichaud in Rakudo? what about in Perl 6 in general?
colomon pmichaud: did you get any speed improvement with that Parrot flag you were trying to set earlier today?
pmichaud colomon: by changing the flag I eliminated the 40% penalty I was seeing 19:29
colomon pmichaud: that's something, anyway!
pmichaud which means rakudo is only as fast as the january release, not faster
colomon pmichaud: are we going to make setting the flag that way the default for Rakudo?
pmichaud colomon: it's been the default already 19:30
it was the default in the march release
and in the april release
--gc=gms
sorear niecza: class Soldier { has $!health; method health() { _newtiedscalar(Any, Any, { $!health }, { die "Oops too big" if $_ > 100; $!health = $_ }); } }; my Soldier $Asoldier .= new; $Asolder.health = 5; say $Asoldier.health; $Asoldier.health = 120
p6eval niecza v4-76-g7bf5ef9: OUTPUT«===SORRY!===␤␤Variable $Asolder is not predeclared at /tmp/haIrhJxpoe line 1:␤------> }; my Soldier $Asoldier .= new; $Asolder⏏.health = 5; say $Asoldier.health; $Asol␤␤Unhandled exception: Check failed␤␤ at
../home/p6eval/niecza/boot/lib/CORE.se…
sorear niecza: class Soldier { has $!health; method health() { _newtiedscalar(Any, Any, { $!health }, { die "Oops too big" if $_ > 100; $!health = $_ }); } }; my Soldier $Asoldier .= new; $Asoldier.health = 5; say $Asoldier.health; $Asoldier.health = 120
p6eval niecza v4-76-g7bf5ef9: OUTPUT«5␤Unhandled exception: Oops too big␤ at /home/p6eval/niecza/lib/CORE.setting line 413 (CORE die @ 2)␤ at /tmp/D2WE45_RKj line 1 (MAIN C6_ANON @ 2)␤ at line 0 (ExitRunloop @ 0)␤ at /tmp/D2WE45_RKj line 1 (MAIN mainline @ 9)␤ at /home/p6eval/niecza/lib/CORE.setting
..line 1311 (C…
sorear \o/
19:32 Rotwang1 joined
pmichaud which means that either (a) the GC didn't provide nearly as much improvement as claimed, or (b) something has happened to Rakudo to cause it to become much slower than it was in January (and thus lose any gains provided by the GC) 19:32
or (c) other
newbee hehe, great, thanks sorear :) now, ¿what is _newtiedscalar? looks like a hack, ¿it is? 19:33
sorear _newtiedscalar is how niecza makes special lvalues 19:34
pmichaud I think the standard Perl 6 mechanism would be to use WHENCE
sorear no good decision has been made on how to do that
newbee oh well... in the meantime niecza's method is what i will use :) 19:35
sorear pmichaud: WHENCE only lets you run code the first time a variable is bound or assigned to
pmichaud method health() is rw { $!health WHENCE { die "Oops too big" if $_ > 100 } }
or something like that
oh yes, only for vivification
hmm
sorear there's something in the spec about Proxy.new:, but I find that interface vaguely disturbing 19:36
pmichaud that pre-dates WHENCE
it's probably fossil
afk, kid pickup
alester Today I learned that pmichaud has a kid. 19:40
sbp he didn't say he has one 19:42
could be picking anybody's kid up!
masak newbee: subset Health of Real where { 0 <= $_ < 100 }; class Soldier { has Health $.health is rw }
colomon sbp: he's also quite knowledgable about Thomas the Tank Engine. ;) 19:44
sbp proof at last
tadzik masak: that's nice
TimToady std: subset Health of Real where 0..^100
p6eval std 3468e14: OUTPUT«ok 00:01 120m␤»
tadzik also, that's Perl6-y;
masak TimToady++
tadzik: yes; problem is it dies if you try to make the assignment. 19:45
maybe you just wanted a warning?
newbee i do want a warning.. :) 19:46
masak there you go :) 19:47
TimToady subset Health of Real preferring 0..^100 :)
masak TimToady: now you're just making stuff up :P
sorear he's TimToady. he can do that. 19:48
masak rakudo: subset Health of Real where 0 ..^ 100; my Health $h = 47; say $h; $h = -20; say 'alive'
p6eval rakudo 8533c3: OUTPUT«Null PMC access in invoke()␤ in 'ACCEPTS' at line 1␤ in '&infix:<=>' at line 1␤ in main program body at line 22:/tmp/NXqWXvAzfS␤»
masak huh? 19:49
masak submits rakudobug
sbp subset Health of Real where Maybe 0..^100
moritz tadzik: huh? which of my modules miss a meta info thingy file?
masak neither 'preferring' or 'Maybe' have anything to do with giving a warning instead of an error. 19:50
colomon does it make sense to have a subset of a role?
rakudo: subset Health of Int where 0 ..^ 100; my Health $h = 47; say $h; $h = -20; say 'alive'
p6eval rakudo 8533c3: OUTPUT«47␤Type check failed for assignment␤ Container type: Int␤ Got: Int␤ in '&infix:<=>' at line 1␤ in main program body at line 22:/tmp/SaP_Ta4xRj␤»
TimToady std: subset Health of Real where (0..^100) S| warn("Health out of range")
p6eval std 3468e14: OUTPUT«ok 00:01 121m␤»
masak colomon: sure, since you can instantiate (classes snapshotted from) roles. 19:51
TimToady: ooh.
tadzik moritz: CGI::Application, Nonogram and Module::Starter
sbp Perl6: Our operators are from Sendak
masak so, S| is basically || but without the need to put things in a block? :P
colomon TimToady: warn returns something which is True? 19:52
TimToady if it succeeds
masak what kind of object does S| create? a Junction-but-serial?
TimToady supposedly
S is vaguely specced to mean several things
masak glares irritatedly at S
colomon Yes, as I recall, S made me cross last time I looked at the spec for it. :) 19:53
masak to me, S is the "one metaop too many" metaop.
TimToady masak: it also differs from || in not making you compare explicitly with the value, since it's still a junction 19:54
moritz tadzik: I think Module::Starter can be deleted
masak TimToady: the &warn call is part of the 'where' clause. surely its return value will be matched against the subtyped variable? 19:55
19:56 jaldhar left
TimToady not if it's a Boolean... 19:56
colomon masak: it's not that S is "one metaop too many"... it's that it's "one metaop, four different meanings, more or less magically determined".
pmichaud rakudo 2011.01 spectest: 29m45s
masak TimToady: oh!
sbp S is for polySemy 19:57
pmichaud rakudo 2011.03 spectest: 30m15s
colomon randomly picked "four", been too long since he looked at the spec for S
pmichaud rakudo 2011.04 spectest: 28m59s
I see no significant speedup here. :-(
TimToady that's okay, since Unicode has 98 things that match /« S »/
jnthn pmichaud: No, not much... :(
TimToady .u ∾ 19:58
phenny U+223E INVERTED LAZY S (∾)
jnthn pmichaud: I shoved nqp through the Visual Studio profiler recently. 19:59
pmichaud: 6model guts were...a long way down the profiling list...
19:59 Moukeddar joined
jnthn pmichaud: And that's before we're even really using it optimally... 19:59
alester Phooey, the durpal dump is >1.5G 20:00
but I don't know hy
moritz lots of spam?
alester i'm sure
i don't know which tables to dump.
masak .oO( this would all have been easier with event sourcing... ) 20:01
20:01 Moukeddar left
tadzik moritz: I'm on it 20:01
moritz: now I think about writing one anyway :) 20:02
dalek osystem: f981b4e | tadzik++ | / (2 files):
Remove Module::Starter
 20:03
sorear S is the "I have no idea how it's expected to work" metaop 20:04
pmichaud alester: I've scraped the site
I'll grab what I need from the html
I'm guessing the drupal dump is 1.5G owing to the spam comments 20:05
alester I have raw stuff for you, pmichaud
only two tables you need
once I figure how to get mysqldump to dump only certain tables. 20:06
but if you're happy w/the scraping, that's fine, too, and I'll stop
20:06 Rotwang1 is now known as Rotwang, ymasory left
pmichaud let me try with the scrapes first... if I need the dumps I'll let you know :) 20:06
moritz tadzik: I've added META.info to CGI::Application and Nonogram 20:07
20:08 stkowski joined
moritz tadzik: which ecosystem repo is the "authoritative" one right now? yours or the one in perl6? 20:09
tadzik moritz: perl6. I should remove mine, doing that now 20:10
done
moritz tadzik: do we still need projects.list for anything? 20:13
tadzik moritz: for backwards compatibility and historical reasons I think
20:13 pmurias joined
pmichaud channel poll: is it important to re-post the various Rakudo compiler announcements to rakudoperl.org ? 20:14
pmurias sorear: what's wrong with Proxy?
moritz pmurias: no, but I'd love to preserve the R* release
pmurias moritz: meant pmichaud? 20:15
moritz yes, sorry
pmichaud: see above :-)
pmichaud yes, I'm planning to keep the R* release announcements
here's a preview: rakudoperl.org/?p=7&preview=true
tadzik pmichaud: notfound
moritz Apologies, but the page you requested could not be found. Perhaps searching will help.
masak pmichaud: I think it's important to show something each month. I remember when I wasn't too involved with Rakudo. the release announcements served as a reminder that the project still existed and was alive.
pmichaud: posting things is cheap.
pmichaud masak: I mean the historical postings 20:16
masak oh!
pmichaud masak: we're moving rakudo.org to a wordpress site
rakudoperl.org
masak ah. sorry. :) I'm not keeping up :P
pmichaud and I'm wondering how much content from the old rakudo.org site I need to keep
dalek osystem: 72f62a5 | moritz++ | / (3 files):
remove two modules from SHELTER, and add META.info links instead
pmichaud thus: do I need to transfer all of the monthly release announcements for rakudo compiler
preview post might only be available if you're logged in/registered 20:17
moritz no
pmichaud I'll go ahead and publish it
pmurias sorear: the interface to Proxy seems perfect to me
pmichaud rakudoperl.org/2010/06/19/rakudo-st...y-july-29/
moritz (and if yes, it would probably be less work to just take the announcements from git)
20:17 rhebus left, rhebus joined
pmichaud it seems to be really easy to copy-n-paste from the drupal html into wordpress 20:18
pmurias sorear: with a possible BIND for bvalues 20:25
20:25 yegor__ joined 20:26 Moukeddar joined 20:28 yegor_ left 20:29 molaf joined 20:38 Moukeddar left, MayDaniel joined, ymasory joined 20:40 birdwindupbird left
tadzik heh, I need Test::Simple for every language I use. Did anyone ever wrote Test::Simple for Java? :) 20:42
20:43 Moukeddar joined, Sarten-X left
masak yeah. but I never published it. 20:50
it was before I embraced JUnit a bit more.
20:51 molaf left 20:52 Sarten-X joined 20:54 snearch_ left
tadzik yeah, I guess I should look into this jUnit 20:55
I was trying to limit the "new things to learn" to just java
tadzik writing Battleship game for the university project
arnsholt tadzik: Yeah, I have really come to like TAP 20:56
20:57 TheD left
masak TAP, like POD, should really be more widespread than just Perl. 21:01
21:02 yegor_ joined 21:03 kaare__ left 21:05 Mowah left, yegor__ left 21:09 MayDaniel left 21:10 yegor_ left 21:16 Moukeddar left
pmichaud jnthn: how many of your articles from rakudo.org would you want to have migrated to rakudoperl.org? ("all" is an acceptable answer) 21:17
(other opinions welcome also)
21:18 Rotwang left
pmurias masak: what does jUnit use instead of TAP? 21:18
jnthn pmichaud: How many are there? 21:19
pmichaud: And what were they about?
pmichaud: I don't ever remember writing anything exclusively for rakudo.org... 21:20
pmurias pmichaud: why is there a migration from rakudo.org to rakudoperl.org
jnthn Oh, maybe the odd thing...
pmichaud I'm not sure they were "exclusive"
jnthn Yeah
I suspect they were cross-posts. 21:21
pmichaud rakudo.org/node/55
21:19 <jnthn> pmichaud: And what were they about?
oops
alester pmichaud: why the domain name change?
pmichaud rakudo.org/node/52
rakudo.org/node/51
etc
jnthn Oh, I'm sure they were cross-posted
pmichaud alester: I'm only using "rakudoperl.org" to distinguish new versus old 21:22
jnthn I did them on use.perl.org
pmichaud alester: when things are moved across it'll just be "rakudo.org" again
alester pmichaud: ok.
jnthn So in terms of "not losing them for good" they won't be.
21:22 leprevost joined
jnthn So I've no issue with them not being migrated. 21:22
alester My A-record-modifying fingers stand at the ready.
21:22 mj41 left
alester jnthn: Or rather, is there any reason NOT to migrate them? 21:22
pmichaud well, I don't know how much I trust use.perl.org to stay around, either :) 21:23
I've been thinking of migrating my use.perl.org posts to my blog
21:23 hercynium left
jnthn alester: No, there's no reason not to either. 21:24
alester: I don't feel strongly either way. :)
alester Is there a way you could be even more ambivalent? 21:25
More extreme in your ambivalence?
TRUE NEUTRAL, BABY.
pmichaud: HURRY I YEARN TO KILL THE DRUPAL. 21:26
"He took his Drupal sword in hand...." 21:27
frettled alester++ 21:30
Is WordPress on rakudoperl.org updated to 3.1.2 yet? 21:31
alester I'm sure Drupal is really awesome and powerful and all, but damn, I can't maintain the thing.
frettled WP 3.x is easy to upgrade, at least. 21:32
But there has been a spate of vulnerability fixes recently, more than Drupal, and a whole bunch of automated attacks.
PerlJam frettled: Drupal is easy to upgrade too ... it's just dealing with the fall-out that's the problem. :-) 21:33
pmichaud frettled: I'm at 3.1.0 atm. I can upgrade to 3.1.2
one moment
Done. 21:34
21:35 frettled joined
frettled Sigh, irssi froze. : 21:35
hugme: hug me 21:36
hugme hugs frettled
frettled (:
masak hugme: hug hugme
hugme hugs hugme
frettled masak: hug masak
masak hugs masak
frettled \o/
masak I'm gonna hug myself into bed now. 21:37
'night, #perl6.
TimToady o/
frettled Splendid idea, masak++
21:37 masak left
frettled TimToady: golfing? :) 21:39
21:40 wallberg_ joined, justatheory left 21:41 rhebus left
tadzik arnsholt! arnsholt! 21:43
arnsholt: you missed a comma in your META.info
I should make the ecosystem more typoproof 21:44
21:44 rhebus joined
arnsholt OHNOES! 21:52
There. Fixed 21:53
tadzik thx
21:53 ymasory left
frettled If only other software developers could be half as quick. 21:53
arnsholt I can either write my thesis, which is due on monday, or procrastinate. Not a difficult choice ;) 21:55
dalek osystem: e2cebef | tadzik++ | META.list:
Remove links to nonexisting META.infos
 21:57
p/ctmo: 6435cc3 | jonathan++ | src/ (3 files):
Simply multi method construction and dispatcher testing.
 22:00
p/ctmo: 245026d | jonathan++ | src/ (2 files):
Shuffle method-adding code into the SC builder.
p/ctmo: fef0b59 | jonathan++ | src/HLL/SerializationContextBuilder.pm:
First steps at getting method definitions to go through the compile time meta-model. Some regressions.
ar: 0601403 | pmichaud++ | skel/docs/announce/2011.04:
More announcement updates.
 22:03
22:13 bacek joined 22:14 rhebus left 22:19 karamorf joined 22:26 wallberg_ left 22:32 wallberg left, whiteknight joined 22:38 icwiener left 22:46 alester left
dalek p/ctmo: 6ca730b | jonathan++ | src/ (3 files):
Partial fix for multi-method dispatch regression.
 22:53
p/ctmo: 5b9b016 | jonathan++ | src/NQP/Actions.pm:
Get adding of regex declarations done through the compile time meta-object. Well, mostly...there's a horrible issue with the way prefixes are stored to deal with first...
22:53 donri left
jnthn pmichaud: ping 22:54
22:55 stkowski left 22:56 pmurias left
dukeleto if anybody has any lingering commits to rakudo.git, please push. I am testing a github post-receive hook that runs and times the test suite 23:02
23:12 justatheory joined 23:14 justatheory_ joined 23:16 justatheory left, justatheory_ is now known as justatheory 23:25 karamorf left 23:38 cosimo joined
jnthn sleep & 23:39
23:39 Chillance left

Powered by Cro and the Raku® Programming Language.

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

Thank you!