»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! | tinyurl.com/p6contest
Set by moritz_ on 28 December 2010.
00:02 mkramer1 left
dalek href="https://modules.perl6.org:">modules.perl6.org: 38300ab | snarkyboojum++ | web/ (4 files):
Fix camelia icon for IE

IE doesn't seem to be able to render .ico files. Create a local png version of the camelia icon.		 00:03
00:04 hanekomu_ joined, hanekomu left 00:07 hanekomu_ left 00:09 awwaiid_ joined
dalek href="https://modules.perl6.org:">modules.perl6.org: fa83aaf | snarkyboojum++ | web/ (2 files):
Use a better image filename
 00:10
ecza: 9347367 | sorear++ | FETCH_URL:
Update bootstrap for recompilation fixes
 00:15
sorear snarkyboojum: that is so ironic... .ico is a Windows-native format and IE was the first browser to render them 00:16
snarkyboojum sorear: weirdness.. doesn't render in HTML that I can see :|
00:17 mkramer joined, dorlamm left
snarkyboojum sorear: obviously renders favicon.ico ok... 00:17
00:23 hercynium left, pmurias left
snarkyboojum I suspect it was probably working, but was being blocked through the proxy I was using to check it (stupid corporate firewalls). Ugh. 00:25
00:29 plobsing joined 00:38 cosimo joined
snarkyboojum phenny: tell moritz_, whatever process copies files from the modules.perl6.org github repo to the web server needs to be updated to include two additional images web/proto.png and web/camelia.png :| 00:40
phenny snarkyboojum: I'll pass that on when moritz_ is around.
snarkyboojum ok I'm going to stop breaking stuff now ;)
Tene Hmm. I wonder why IO.dir doesn't return IO objects... 00:53
flussence well, you could do dir».IO ... 00:56
Tene Yeah, I guess. It just surprised me. 00:57
diakopter niecza: say 12
p6eval niecza v1-149-g56c80b0: OUTPUT«12␤»
Tene IO could use a nicer .Str, too
01:01 QinGW joined, klunky left 01:03 dju left, dju joined 01:07 sftp left
Tene flussence: not quite, actually, as the strings returned don't have path information 01:08
rakudo: dir('/tmp')>>.IO>>.created
p6eval rakudo 388eed: OUTPUT«Operation not permitted in safe mode␤ in 'Safe::forbidden' at line 2:/tmp/rFAkqEjEg7␤ in main program body at line 22:/tmp/rFAkqEjEg7␤»
01:10 sftp joined 01:11 drbean joined 01:16 [Coke] left
dalek ecza: e2d1ed2 | sorear++ | Makefile:
Write VERSION file
 01:19
01:20 QinGW left 01:21 QinGW joined
sorear I need some sort of way for people to give input on long-term priorities 01:25
MMD? Lexotic controls? Parallelism? etc
01:27 [Coke] joined
snarkyboojum MMD! 01:27
sorear: and slangs! :D 01:32
01:36 kst left, kst joined
snarkyboojum sorear: is there something like a github hosted wiki you could use? 01:36
01:40 QinGW left 01:42 QinGW joined, lichtkind_ left
diakopter sorear: I suggest fleshing out the various phasers, if they're not already 01:50
01:52 clkao joined 02:00 QinGW left 02:01 satyavvd joined, QinGW joined 02:04 noganex left 02:06 noganex joined 02:25 mtk left 02:26 pencilk joined
[Coke] does "needs tests" still mean "assign to moritz" ? 02:29
coldhead hahaha 02:31
02:31 mtk joined 02:37 cafesofie joined
colomon not at the moment... he has a lot on his mind. :) 02:41
snarkyboojum the perl 6 spec could use some pretty pictures :P
snarkyboojum hopes that's deemed heretical :P 02:43
oops *not* deemed...
colomon: enjoyed your last post explaining your stack machine in p1 :D 02:45
colomon snarkyboojum: thanks.
I love stack machines. :)
snarkyboojum colomon: that came across too :) 02:46
02:51 mkramer1 joined 02:52 mkramer left 02:54 lopaway is now known as lopnor
snarkyboojum I wonder if, generally, this "right of reply" will influence masak++'s decision at all 02:54
02:55 mkramer joined, mkramer1 left 03:01 mkramer1 joined, mkramer1 left, mkramer left 03:03 mkramer joined, mberends left 03:05 plobsing_ joined
colomon snarkyboojum: I don't think it should with that script. I mean, mine is pretty clearly #4. Maybe #5 if you're being mean. I think it's interesting and cute code, but it's clearly not a good solution to the problem. 03:05
03:08 plobsing left
snarkyboojum colomon: sure, I guess I meant in the final analysis, for an arbitrary definition of "final" :) 03:08
colomon On the other hand, p2 is going to be interesting.
snarkyboojum colomon: aye, at least a week of enjoyable analysis/reading ahead :) 03:11
03:15 justatheory left 03:16 envi joined 03:27 _jaldhar joined 03:35 QinGW left 03:37 mkramer1 joined 03:39 mkramer left 03:41 lopnor is now known as lopaway 03:42 satyavvd left 03:44 QinGW joined 03:46 mkramer joined 03:47 mkramer1 left 03:50 agentzh joined 04:02 QinGW left, QinGW joined 04:04 Grimnir_ left 04:06 Vlavv left 04:15 Grimnir_ joined 04:17 satyavvd joined 04:19 Vlavv joined 04:20 justatheory joined, gabiruh left 04:22 QinGW left 04:23 QinGW joined 04:25 awwaiid left 04:26 race6 joined 04:27 Su-Shee left, awwaiid joined 04:28 Su-Shee joined 04:29 drbean left 04:38 bacek_ left 04:41 race6 left 04:42 satyavvd left 04:56 orafu left, orafu joined 04:58 satyavvd joined 05:13 p6eval left 05:14 dalek left 05:15 diakopter left, TimToady left, sorear left 05:25 sftp left 05:26 sftp joined 05:28 dalek joined, ChanServ sets mode: +v dalek, sorear joined, p6eval joined, ChanServ sets mode: +v p6eval, orafu left, TimToady joined, diakopter joined, orafu joined
sorear interesting 05:35
p6eval and diakopter have completely different IPs but they lose connectivity at once
snarkyboojum ergo, diakopter is a bot 05:36
sorear that does not follow
05:39 dalek left 05:40 dalek joined, ChanServ sets mode: +v dalek 05:42 QinGW left, kst left, QinGW joined 05:43 Patterner left, kst joined
snarkyboojum sorear: what sort of qualitative evidence do you have for your original statement? 05:43
05:45 Psyche^ joined, Psyche^ is now known as Patterner, p6eval left
snarkyboojum diakopter: is still here 05:46
05:46 dalek left 05:51 sorear left
snarkyboojum rakudo: say "Yikes, I don't work" 05:52
05:52 sorear joined
sorear ~[6~[6~[6~just look at the logs 05:52
sorear, diakopter, TimToady, dalek, and p6eval all pinged out in the same 5 minute interval
also I am currently fighting 120+ second lag
snarkyboojum are they all IRC bouncing through the same server? 05:53
oh, diff ips :P
sorear no, that's my point
snarkyboojum well, the last disconnect for p6eval didn't result in a disconnect for diakopter or TimToady
sorear the humans are on host02.appflux.net, the bots are on host04
05:54 dukeleto left 05:55 dukeleto joined 06:04 kaare_ joined
sorear rakudo: (1..5).say 06:19
snarkyboojum p6eval is dieded 06:22
06:22 am0c joined
diakopter it may return 06:30
snarkyboojum sorear: are you using the t/run_spectests harness for Niecza?
06:31 justatheory left
diakopter host02 and host04 are in the same datacenter, ish 06:32
snarkyboojum host04 is looks to be currently down, but host02 isn't fwiw
diakopter it'll return 06:33
snarkyboojum aye :)
sorear diakopter: odd that they have completely different IPs
diakopter meh
sorear I thought IP routing was prefix-based
diakopter it can be, simplisticly 06:34
06:34 cafesofie left
diakopter incidentally, I can reach host04 from host02 :) 06:35
snarkyboojum diakopter: so it's a routing issue then... ?
sorear snarkyboojum: no
diakopter snarkyboojum: yeah
sorear snarkyboojum: I haven't tried run_spectests
diakopter well, a semi-planned issue, according to the vps host 06:36
sorear I know it won't work on Linux
snarkyboojum sorear: what that a moritz_ WIP to setup a test harness for Niecza then
was*
sorear if it works at all it's Windows-exclusive - it assumes you can just run a .exe program directly
snarkyboojum sorear: because it just needs some TLC to do mono run/Niecza.exe as p6eval etc to work again 06:37
06:37 c9s joined
snarkyboojum sorear: i.e. it's still referring to niecza_eval. Oh well :) 06:38
diakopter snarkyboojum: ?
oh, not the p6eval evalbot.pl
that doesn't do that
sorear snarkyboojum: moritz_ gave a pull request and we can't agree how it should work
snarkyboojum diakopter: no, the t/fudgeandrun script in the niecza reop 06:40
repo(
eek
sorear: fair enough :) was nice when it was working a little while ago.. will bide my time then :) 06:41
sorear: there should be a build system dependent invocation to run the "right thing" I guess re tests 06:44
06:47 pencilk left
snarkyboojum i.e. change the test target in the Makefile to do the right thing, or add a spec_test target 06:47
06:48 c9s left 06:50 cafesofie joined
diakopter blinks hard at mozillalabs.com/zaphod/ 06:51
snarkyboojum diakopter: did you read brendaneich.com/2011/01/harmony-of-my-dreams/ ? 06:53
diakopter most of it 06:54
till the blearies in my eyes pulled their lids stuck
06:54 dalek joined, ChanServ sets mode: +v dalek
diakopter bows 06:55
snarkyboojum diakopter: I enjoyed the focus on parens-free control structures etc
diakopter is always pleased at how robustly dalek (botnix) reconnects
I enjoyed the focus on #subs 06:56
coldhead new javascript features just make me sad because they're always years away from widespread adoption
there's nothing people won't update like their battered old internet explorer 6
06:57 dduncan_ joined
snarkyboojum and implicit return values 06:57
06:57 dduncan_ left, p6eval joined, ChanServ sets mode: +v p6eval
diakopter perl6: .die 06:58
p6eval rakudo 388eed: OUTPUT«Method 'die' not found for invocant of class ''␤ in main program body at line 22:/tmp/E8p6IjkE7a␤»
06:58 wtw joined
p6eval ..pugs: OUTPUT«*** Died␤ at /tmp/tEdUswben4 line 1, column 1 - line 2, column 1␤» 06:58
..niecza v1-149-g56c80b0: OUTPUT«Unhandled exception: Unable to resolve method statement_level in class Hash␤ at /home/p6eval/niecza/src/NieczaActions.pm6 line 2129 (NieczaActions C766_ANON @ 1)␤ at /home/p6eval/niecza/boot/lib/SAFE.setting line 73 (SAFE C36_ANON @ 1)␤ at
../home/p6eval/niecza/boot/lib/SAFE.setti…
06:58 drbean joined
diakopter pugs++ 06:58
niecza: {}.statement_level 07:00
p6eval niecza v1-149-g56c80b0: OUTPUT«Unhandled exception: Unable to resolve method statement_level in class Hash␤ at /tmp/z6L8K1obD7 line 1 (MAIN mainline @ 3)␤ at /home/p6eval/niecza/lib/SAFE.setting line 914 (SAFE C354_ANON @ 1)␤ at /home/p6eval/niecza/lib/SAFE.setting line 915 (SAFE module-SAFE @
..29)␤ at /home/…
07:01 cogno joined 07:02 justatheory joined 07:06 justatheory left 07:14 ascent__ left, puge left 07:15 ascent_ joined, puge joined 07:21 cogno left 07:24 am0c left
sorear no, Hash does *not* have statement_level 07:26
statement_level is defined on Op, anyway 07:27
niecza keeps the meta-levels rigidly separate; statement_level only exists in the compiler
07:28 cogno joined
diakopter I was just trying to reproduce the error it gave from .die 07:28
sorear btw while dalek was out I added Range
niecza: say 1..10
p6eval niecza v1-149-g56c80b0: OUTPUT«(timeout)» 07:29
sorear niecza: say 1..10
p6eval niecza v1-149-g56c80b0: OUTPUT«12345678910␤»
07:32 [particle] left 07:34 cogno left 07:36 hanekomu joined
sorear it looks like making .method; work has a dependency on the planned operator overhaul 07:38
sorear bumps that
07:48 masak joined
masak early morning wakeup attempt WIN \o/ 07:48
morning, #perl6
"we moved to a new development model" -- blog.whatwg.org/html-is-the-new-html5 (via @obra) 07:55
07:57 yarp joined
moritz_ phenny: tell snarkyboojum see web/build-project-list.pl in the moduels.perl6.org repo - it does the copying 08:05
phenny moritz_: I'll pass that on when snarkyboojum is around.
moritz_: 00:40Z <snarkyboojum> tell moritz_ whatever process copies files from the modules.perl6.org github repo to the web server needs to be updated to include two additional images web/proto.png and web/camelia.png :|
sorear hello masak. 08:06
although I'll be going to bed soon :p 08:07
snarkyboojum moritz_: oh my, ok - that's some hardcoded goodness :)
phenny snarkyboojum: 08:05Z <moritz_> tell snarkyboojum see web/build-project-list.pl in the moduels.perl6.org repo - it does the copying
sorear masak: it turned out all p6eval needed to fix "Array.keys" was a clean rebuild. You might try that.
masak ok. 08:08
so just nuke niecza/ and clone it again?
sorear git clean -dfx 08:09
git pull
moritz_ snarkyboojum: you can also add globs to the script
snarkyboojum moritz_: goody - will have a look. thanks. 08:10
08:15 daxim joined
snarkyboojum moritz_: how does this perl script get copied to the server? 08:17
08:18 am0c joined
masak sorear: build trouble -- se bottom of gist.github.com/789410 08:19
08:23 am0c left 08:25 am0c joined 08:45 jjore left 08:46 jjore joined 08:50 yarp left
snarkyboojum rakudo: say [*] Nil 08:51
p6eval rakudo 388eed: OUTPUT«1␤»
snarkyboojum why is that the case?
moritz_ snarkyboojum: via github
snarkyboojum moritz_: i.e. there is a cronjob or something with copies from github? 08:52
moritz_ snarkyboojum: 1 is the neutral element to multiplication
snarkyboojum: it git pulls and then runs the script
yes, cronjob
snarkyboojum moritz_: ah, ok - to both answers :) thanks
moritz_ rakudo: say prefix:<[+]>() 08:53
p6eval rakudo 388eed: OUTPUT«0␤»
moritz_ rakudo: say prefix:<[*]>()
p6eval rakudo 388eed: OUTPUT«Could not find sub &prefix:<[*]>␤ in main program body at line 22:/tmp/AbpbTNq0Ka␤»
moritz_ rakudo: say infix:<*>() 08:54
p6eval rakudo 388eed: OUTPUT«1␤»
moritz_ rakudo: say infix:<+>()
p6eval rakudo 388eed: OUTPUT«0␤»
moritz_ rakudo: say infix:<&&>()
masak \o/
p6eval rakudo 388eed: OUTPUT«Bool::True␤»
snarkyboojum that's pretty cool
masak I wonder how to specify that for a user-defined chaining op. 08:55
moritz_ masak: you just define a nullary infix that returns the default element
masak maybe the way to define chaining ops will always be in terms of all their operands as a slurpy.
moritz_ yes
masak then no nullary case is needed. 08:56
moritz_ though it doesn't account for short-circuiting operators
08:59 hanekomu left
masak true. 08:59
are those chaining, though?
masak czechs
moritz_ sure 09:00
a() && b() && c()
masak no; they're list associative.
that's different.
moritz_ hm
masak "chaining" is more of a precedence level with certain special rules. 09:01
moritz_ 1 < 2 < 3 could be both chaining and short-circuiting
masak right, but that's handled very differently.
moritz_ in rakudo's Configure.pl, line 179, where does $1 get set? 09:02
ah
moritz_ dumb
09:05 kensanata joined 09:08 am0c_ joined
masak in the case of a tie between a multi method in a Parent class and a multi method in a Child class (belonging to the same proto method), does the Child multi always win? 09:08
09:10 am0c left, am0c_ is now known as am0c
masak rakudo: class Wide {}; class Narrow is Wide {}; class Parent { multi method foo(Narrow) { say "Parent" } }; class Child is Parent { multi method foo(Wide) { say "Child" } }; Child.new.foo # tie, or Child.foo wins? 09:11
p6eval rakudo 388eed: OUTPUT«No candidates found to invoke for method 'foo' on object of type 'Child'; available candidates have signatures:␤:(Mu : Wide ;; *%_)␤␤ in main program body at line 22:/tmp/d596Oc_qYy␤» 09:12
masak er
moritz_ I thought the child method is always considered first anyway?
masak rakudo: class Wide {}; class Narrow is Wide {}; class Parent { multi method foo(Narrow) { say "Parent" } }; class Child is Parent { multi method foo(Wide) { say "Child" } }; Child.new.foo(Wide) # tie, or Child.foo wins?
p6eval rakudo 388eed: OUTPUT«Child␤»
masak moritz_: that is my question. I know it is for only methods, but...
09:13 kst left
moritz_ -> afk 09:13
09:13 kst joined
masak moritz_: here, in essence, the method is *located* in Parent, with one multi definition in Parent and one in Child. 09:13
jnthn masak: ask nqp-rx/nom or nqpclr
masak: Answer is a tie. 09:14
By current spec.
And I expect nqp-rx/nom and nqpclr get it right.
Actually mutli-dispatch doesn't care about where the method is located. It does care about the type of the invocant, which as far as its concerned is just the first argument. 09:15
Like anywhere in Perl 6.
masak that's what I suspected/hoped. 09:16
jnthn Anyway, Rakudo is lagging current spec.
But I know how to implement current spec on Parrot now as I've done the hard bits in nqp-rx already. :)
09:18 jjore left
jnthn It'll land in the branch to get Rakudo onto 6model. 09:18
09:20 jjore joined
masak \o/ 09:26
09:29 am0c left 09:34 jlaire left, am0c joined 09:36 am0c left 09:38 REPLeffect left, jlaire joined 09:39 REPLeffect joined 09:43 kst left, kst joined, am0c joined
masak phenny: tell sorear that I ran into some build trouble -- see bottom of gist.github.com/789410 09:47
phenny masak: I'll pass that on when sorear is around.
09:47 masak left 09:51 tzhs joined 09:55 am0c left 10:06 QinGW left 10:11 mtk left 10:13 kst left, kst joined 10:18 Grimnir_ left 10:19 dakkar joined, mtk joined 10:29 cjk101010 joined 10:36 daxim left 10:47 Grimnir_ joined 11:17 VXZ joined
flussence argh, CPAN is annoying. It seems like half the interesting-sounding module names link straight to a "No matches" error page. 11:22
11:26 hcchien left
mathw yes, CPAN is very annoying 11:28
11:29 am0c joined 11:31 JimmyZ joined 11:35 Chillance joined 11:37 Bzek joined 11:43 Lorn left 11:44 agentzh left 11:46 Lorn joined 11:47 satyavvd left 11:58 cogno joined 12:10 hcchien joined 12:15 Bzek left 12:18 daxim joined 12:27 am0c left 12:31 cogno left 12:35 gimix left 12:43 am0c joined 12:45 drbean left 12:46 jferrero joined, JimmyZ left 12:50 jarfhy left 12:51 drbean joined
takadonet morning all 13:00
13:08 MayDaniel joined, MayDaniel left, MayDaniel joined 13:09 Kodi joined
Kodi rakudo: my $x = 2; say "$x =~ b"; 13:10
p6eval rakudo 388eed: OUTPUT«===SORRY!===␤Unsupported use of =~ to do pattern matching; in Perl 6 please use ~~ at line 22, near " b\";"␤»
Kodi rakudo: my $x = 2; say "$x << b"; 13:11
p6eval rakudo 388eed: OUTPUT«===SORRY!===␤Unsupported use of << to do left shift; in Perl 6 please use +< or ~< at line 22, near "b\";"␤»
Kodi rakudo: my $x = 2; say "$x +< b";
p6eval rakudo 388eed: OUTPUT«2 +< b␤»
Kodi Ah, RT #76878. 13:14
13:18 cafesofie left
arnsholt I'm gonna write a Trie module 13:26
13:26 MayDaniel left
arnsholt Should I call it Data::Trie, or Algorithm::Trie? 13:26
Kodi I vote for the former. 13:27
The word "data" suggests "data structure".
13:27 cogno joined
arnsholt Yeah, I think I agree 13:28
dalek ast: cdf3825 | (Kodi Arfer)++ | S03-operators/brainos.t:
[brainos.t] Added tests for RT #76878.
 13:31
13:32 awwaiid_ left 13:34 drbean left 13:38 Kodi left 13:39 cogno left 13:45 am0c left 13:47 dju left 13:48 dju joined 13:49 simcop2387 left 14:02 simcop2387 joined 14:13 plobsing_ left 14:31 [particle] joined, kaare_ left 14:34 JimmyZ joined 14:44 orafu left 14:47 orafu joined, mtk left, molaf joined 14:50 hanekomu joined 14:52 mtk joined 14:57 mtk left 15:02 mtk joined 15:13 Alias_ joined, kaare_ joined 15:15 Alias left, Alias__ left, Alias joined 15:19 dakkar_ joined, dakkar left, wtw left, dakkar_ is now known as dakkar 15:25 plobsing joined 15:33 plobsing left 15:37 cognominal joined 15:42 plobsing joined 15:45 estrabd_afk is now known as estrabd
flussence I thought I'd be clever at $p5_dayjob and use `xxd -r` for binary patching - it turns out xxd itself is *really* unforgiving with the format of lines... 15:45
dalek ok: a3f9258 | (Greg Donald)++ | README:
mention Perl 5 Template module dependency
ok: a8a5a89 | (Greg Donald)++ | src/operators.pod:
how to type hyper operator character in Emacs
15:46 szabgab left 15:47 kaare_ left 15:49 kaare_ joined
sorear phenny: tell masak wtf. What files are in boot/obj? What does MONO_LOG_LEVEL=debug mono boot/obj/CLRBackend.exe boot/obj MAIN.nam MAIN.exe 1 say? 15:49
phenny sorear: I'll pass that on when masak is around.
sorear: 09:47Z <masak> tell sorear that I ran into some build trouble -- see bottom of gist.github.com/789410
16:00 szabgab joined, cognominal left 16:05 JimmyZ left 16:07 kaare_ left
sorear good * #perl6 16:10
tadzik hello zebras 16:13
16:21 tzhs left 16:24 slavik joined 16:32 bluescreen joined 16:33 justatheory joined
TimToady sorear: still can't quite get niecza built: gist.github.com/789935 16:35
does it need some kind of a library path? 16:36
16:36 snearch joined
[Coke] niecza? 16:36
[Coke] wonders how to activate the bots.
TimToady (trying again after git clean) 16:38
[Coke] where is niecza?
16:38 kensanata left
sorear www.github.com/sorear/niecza 16:38
TimToady: ok, so you can reproduce masak's issue
16:39 dipthegeezer left
TimToady yes, same error after the clean 16:39
sorear TimToady: what files are in boot/obj? (I'm especially looking for SAFE.dll, Kernel.dll, SAFE.nam) 16:40
niecza: say "compilation worked"
p6eval niecza v1-149-g56c80b0: OUTPUT«compilation worked␤» 16:41
TimToady there's no SAFE.dll
sorear mm, 29 revisions behind
TimToady the other two are there
sorear erp. I think I just found a bug in p6eval 16:45
phenny: tell moritz_ Is it deliberate that build.pl only saves STDOUT to the log file?
phenny sorear: I'll pass that on when moritz_ is around.
sorear in any event I can now reproduce it 16:46
16:48 daxim left, MayDaniel joined 16:50 hanekomu left
jnthn evening :) 16:54
takadonet jnthn: hey 16:55
jnthn o/ takadonet
TimToady sorear: do you still want the log level 1 output? 16:56
or is that what you were referring to with "can reproduce" 16:57
sorear TimToady: I've tracked down the problem
16:57 bluescreen left
sorear /bin/sh in p6eval's chroot has a [ builtin that is not quite compatible with the [ builtin in my dev env 16:57
TimToady: Why are < and lt in Perl reversed from their sense in test(1)? 17:04
sbp test(1) has to be better at something 17:08
17:10 wamba joined
sorear niecza: say "did it work?" 17:12
p6eval niecza v1-178-gd94bb5e: OUTPUT«did it work?␤»
17:14 kst left
TimToady obviously, because test got it wrong :) 17:14
17:14 kst joined
TimToady maybe they liked fortran or somethin' 17:15
and because < looks lke math, and because lt looks like a string 17:17
[Coke] TimToady++
TimToady and because of that Hubris thang...
TimToady was not totally wowwed with the Wonderfully Perfect Design of All Things Unix 17:18
which is fine, given that others have felt the same way about Perl :) 17:20
17:20 slavik left, slavik joined
dalek ecza: bd091ab | sorear++ | Makefile:
Change Makefile to support dash builtin [

It's still not a POSIX usage pattern though.		 17:22
17:22 GinoMan left
sorear phenny: tell masak nevermind, found the problem and fixed it 17:23
phenny sorear: I'll pass that on when masak is around.
17:24 timbunce joined
[Coke] installs mono on darwin and gets another copy of perl (5.8.9) for free! 17:26
TimToady 'tworks!! \o/ 17:27
17:28 kaare_ joined, cogno joined
TimToady though the knapsack packer stil chews through memory like no one's business 17:30
a newer GC might or might not fix that... 17:32
sbp "Hey, I had to let awk be better at *something*... :-)"
— groups.google.com/group/comp.lang.p...4ae85fcbe4
17:36 c9s joined
diakopter shortcircuit: has someone added the new (allegedly) polynomial 3-SAT solution algorithm to your site? 17:39
if not, do so and I'm sure some of us around here would work on a port :)
17:39 MayDaniel left 17:40 pyrimidine joined
shortcircuit diakopter: Not AFAIK. It'd have to be vetted against license restrictions, though. (Whoever added it has to be able to license it to RosettaCode.org under the GFDL 1.2) 17:40
diakopter oh, I just mean a placeholder page
shortcircuit Also, I don't have the time to do it myself right, now, but anyone _could_.
17:40 cdarroch joined, cdarroch left, cdarroch joined
shortcircuit diakopter: Ah. Well, on that note...There's no task-creators' club. See blog.rosettacode.org/2011/01/gettin...lling.html for an easy intro on how to create the task. 17:41
diakopter so, a draft task 17:42
shortcircuit Yup; the community on the wiki will help the creator through the rest of the process.
17:46 plobsing left
sorear is the 3-SAT thing for real? 17:51
diakopter in order to prove his algorithm is 100% successful, he would need to prove that the 3rd output message code is unreachable :) 17:52
"failure of classication—in case when HSS forming was completed, but a satisfying set couldn’t be found." 17:53
in *his* tests, he never got that message
sorear I'd like to see him crack RSA-100 with his NP solver, then I'll care
diakopter :) 17:54
shortcircuit diakopter: If you haven't already, go ahead and create the task page. It'll be interesting to see what happens. 17:57
17:57 mtk left
diakopter has to master vmware esxi this morning 17:57
17:58 mtk joined, mtk left 17:59 mtk joined, masak joined
shortcircuit I expect someone at Google is already planning a run of the code through their server farm. 18:01
Or a reasonable translation of it.
masak hola, zebras! 18:02
phenny masak: 15:49Z <sorear> tell masak wtf. What files are in boot/obj? What does MONO_LOG_LEVEL=debug mono boot/obj/CLRBackend.exe boot/obj MAIN.nam MAIN.exe 1 say?
masak: 17:23Z <sorear> tell masak nevermind, found the problem and fixed it
masak sorear++
whatever it was. :)
masak cleans, pulls and tries again
sorear it seems that -nt isn't POSIX and bash and dash interpret it slightly differently 18:03
masak sorear: was the gist I posted of any use in debugging that?
sorear no 18:04
masak also, what's dash? I know that's probably what I'm using, but I only recently saw the name.
is it an alternative sh interpreter?
sorear Debian Almquist SHell
yes
masak googles
smaller, more streamlined, faster. nice. 18:05
Tene A friend and I were talking about ideas for an alternative setting for Perl 6 with changes to make it nicer for usage as an interactive shell. 18:08
There's some interesting possibilities there.
[Coke] (man does mono have a lot of prereqs on darwin)
18:11 mberends joined 18:12 dakkar left, impious joined, impious left, timbunce left 18:14 mtk left 18:15 mtk joined
masak Tene: tell us more. 18:18
18:19 timbunce joined 18:22 cogno left 18:23 cogno joined, rhr left, slavik left 18:24 slavik joined 18:25 masonkramer joined 18:26 cjk101010 left
moritz_ sorear: not deliberate - feel free to fix 18:30
phenny moritz_: 16:45Z <sorear> tell moritz_ Is it deliberate that build.pl only saves STDOUT to the log file?
18:34 cogno left
sorear moritz_: I ran 'apt-get install strace' in the p6eval chroot; is this proper form? 18:40
masak now has a built niecza again o/ 18:46
sorear: autoprinting of expressions went away completely in the REPL? 18:47
18:48 cogno joined 18:49 wamba left
dalek albot: 52b1dc9 | sorear++ | build.pl:
Redirect STDERR when autobuilding, too.
 18:49
sorear erm. yes
(oops?)
do you want it back? 18:50
18:50 Util left, Juerd left, Juerd joined, Util joined
masak I think so. preferably with .perl applied, if possible. 18:50
PerlJam It would be nice to have a REPL where you could build up a lexical environment piece-wise and choose when to execute it. 19:00
dalek ecza: 9311ad1 | sorear++ | lib/SAFE.setting:
dump => perl for consistency with Rakudo
ecza: 98d0cc5 | sorear++ | src/niecza:
Restore auto-say in REPL
tadzik my niecza still fails to run
masak tadzik: did you clean -dfx? 19:01
19:01 rhr joined
tadzik maybe not, let me see 19:01
building again
Tene masak: I ran into a couple of awkward issues with the current setting. Consider this: dir('/tmp/')>>.IO>>.created 19:04
19:04 MayDaniel joined
Tene That currently dies, because the strings returned from dir() don't have path information. 19:05
I'm considering pushing to get IO objects returned instead.
masak Tene: just the other day I was thinking of that!
Tene: I'm not sure &dir should strip the path information.
Tene Well, it's not stripped.
it's not there in the first place.
TimToady dir doesn't know the path in the first pal... what he said 19:06
the path is '.'
masak I say return the full (relative) path, and provide a simple way to strip off the path.
Tene TimToady: *dir* does know the path... it's passed as an argument! :P
masak: There are some nice things, though. dir ==> grep /foo/
TimToady well, okay, but it's not guaranteed to be an absolute path
masak then people who want the IO can do .IO, and the people who want only the file name can do... something.
TimToady: not needed.
TimToady trooo 19:07
masak just that it leads to the file.
Tene TimToady: It's guaranteed to be a path that can be used to make an IO object, though.
TimToady I wish I could say that was my first mistake this year...
masak .split($dirsep)[*-1] would give the filename again.
TimToady well, I could say it, but it would be a mistake... 19:08
masak modulo nasty escaped $dirseps etc
maybe there should be a convenience method for that.
oh, oh!
.IO.filename :)
19:09 hercynium joined
masak or .IO.name, doesn't matter. 19:09
TimToady well, dir() was never really designed, it just sorta popped up existing
masak let's pop it sideways a bit to give full paths.
before it's too late. :)
tadzik you can use File::Find and get the paths 19:10
TimToady or have a different function that does that
masak tadzik: that's nice to know, but I think &dir would be better if it gave the paths.
tadzik maybe 19:11
TimToady or have a dir($dir, :bare) option
masak here's my full argumentation: people might want three things from &dir: just the filename, the full (relative) path, or the IO object. if we give the full (relative) path, the other two are easy to reconstruct.
TimToady rawdir()
Tene masak: you could always return the string name, but mix in the IO object with but
;)
hercynium was going to try to make a glib remark about dir being designed vs. evolved but decided against :)
19:11 envi left
wolverian masak: why not IO object? 19:12
masak Tene: :P
TimToady an IO object only reprents a file, it doesn't have to have opened it yet
masak wolverian: hm, maybe. if it stringifies to the full (relative) path.
TimToady *represents
unless you're worried about races
hercynium of course, there's no reason for it's evolution to be a result of the design on a lower level ;)
19:12 Tedd1 left
hercynium um, I mean *not* to be 19:13
Tene IO objects as they're written now, btw, breaks if you chdir between instantiation and use.
hercynium too many negatives
TimToady certainly the ingyesque view of IO objects is lazy about deciding how to open it
==> IO vs IO ==>
Tene I don't understand "==> IO vs IO ==>" 19:14
allbery_b what does an io object stringify to?
TimToady whether it's opened for reading or writing depends on whether you read or write it 19:15
19:15 cogno left
Tene Ah. 19:15
TimToady see the IO module in P5
Tene One other thing that makes me nervous is the pun between .say on strings and .say on IO objects. 19:16
TimToady so other than possible race conditions and a bit more overhead (maybe), I don't see much problem with dir returning IO objects directly
assuming they can be used like strings
but I guess there was some heartburn a few days ago about IO being in Cool 19:17
Tene I've seen plenty of places where people iterate over a collection calling .say on everything to see it printed.
TimToady I guess that was over +IO rather than ~IO though
Tene You pass IO objects into that, and suddenly your debug or logging method or whatever is writing to the files you're looking at.
What was the problem with +IO?
TimToady what would it mean? 19:18
sorear I think Rakudo is wrong to have IO and FileHandle be the same type
masak TimToady: no, I was just wondering *why* an IO was a Cool. we didn't seem to use it for anything.
sorear TimToady: What does +IO and ~IO mean?
I see IO as being a kind of Stringy and thus Cool; FileHandle is a separate type in Niecza and not Cool 19:19
TimToady I mean what would an IO object numify/stringify to
masak TimToady: also IO can have nice stringifications/numifications without being in Cool.
not saying it should/shouldn't. :)
19:19 Tedd1 joined
sorear IO being Cool allows .comb, etc to work 19:20
Tene sorear: but is .comb over the contents of the file, the filename, what?
sorear Tene: the filename 19:21
my IO identifies a file, not a file handle
Tene sorear: so you're saying that you can't do IO on your IO objects? 19:22
TimToady it might possibly contain a filehandle, and delegate in some circumstances
it represents the identity of a file
Tene ;)
sorear Tene: correct.
Tene My inclination would be towards having a File class oslt, and use the name "IO" for handles, fds, etc. 19:23
sorear masak: have you read the latest Parrot release message? I think cotto_work is trying to steal your schtick
Tene: we could even call it Path, for consistency with qp// (which presumably returns them)
Tene sorear: Yes, I like that. 19:24
sorear Tene: What happens with $filename.IO.e ? Does it become $filename.Path.e ?
Tene sorear: I'd prefer .File.e, but IMO yes, that's better. 19:25
masak sorear: now I've got to more'n skim it :P
Tene I certainly don't think that I've given sufficient thought to the entire problem space of IO, sockets, fds, IPC, etc, but the current architecture seems rather suboptimal. 19:26
TimToady or we rename qp to qio
Tene Bleh.
TimToady well, "path" is arguably too overloaded 19:27
Tene There are plenty of files that aren't meaningful to read or write to, and you can do IO on plenty of non-file things.
PerlJam So ... a Path is something like "perl6.org/community", right?
Tene TimToady: Yes, which is why I prefer File
TimToady File is wrong, and PerlJam++ sez 19:28
s/and/as/
[particle] java must have a terminology for URIs and non-uniform resources too. 19:29
but is it any good?
cotto_work sorear: masak's release announcements were where I got the idea.
Tene File is *Wrong*, or file is *incomplete*?
TimToady a socket is not a file
masak sorear: inclined to agree. cotto_work++. well, I'll have to up the ante, then. :)
[particle] a link is not a file (well, slow links are)
TimToady and I'm fine with a qp// returning an IO 19:30
Tene TimToady: what does .stat or .e or .w mean on a socket? 19:31
TimToady depends on whether it's a Unix-domain socket :P
Tene TimToady: IO as it is right now is exclusively about Files, and not about sockets or other IO-like concepts.
TimToady: Not all unix sockets are associated with a file. 19:32
TimToady so some io operations can fail; nothing new here
[particle] rudy can't fail.
PerlJam failure is not an option? 19:33
TimToady trying to prevent IO failures using type theory is an exercise in exercise
Tene TimToady: That's very true, but that certainly doesn't mean that no IO type hierarchies are less useful than others.
TimToady just trying to guard the huffman coding from the user's perspective here 19:35
IO is short, so should be public facing
PerlJam Tene: When you're best candidates are all heavily overloaded terms, it's time to change the lexicon a bit I think :)
s/you're/your/
Tene Maybe. 19:36
I certainly don't have well-supported conclusions here, just questions.
TimToady perhaps qp[] returns an IO::Path
masak eyes qp[] warily 19:37
TimToady but I'd like to get somewhere close to ingy's ==> io("filename") smarts
reading and writing a list to a file seems very huffman-worthy 19:38
as it is, we're still a long way from >filename
Tene Is there a common role for IO-related types? 19:39
masak Tene: sure, but it's all out in the land of sludge known as S16. 19:40
Tene I'm also a bit skeptical about attempts to unify http, local filenames, sockets, and kitchen sinks in the short default API. I'd be a bit grumpy if the default-recommended function for reading files can suddenly make network requests if a user happens to pass a specially-formatted filename. 19:42
PerlJam one person's dwim is another person's accident, eh? 19:44
TimToady so io() is the general one, and we also have file(), http(), etc for the paranoid :)
or more likely, uri() 19:45
19:45 rgrau joined, timbunce left
Tene TimToady: If you plan for io() to be so general, why have IO be so file-specific? 19:47
TimToady it's not, in my mind
Tene maybe io() could return an IO::File object instead, which does the IO role?
19:47 klunky joined
Tene Ahh, I misunderstood then. 19:47
TimToady to me, io() is the most generic and dwimmy, and there are more specific functions to be more, er, specific 19:48
likewise with IO
it's got some defaultness as the top of the IO hierarchy
flussence all I see io() doing is delegating to a more specific thing based on the ^(.*):// part.
TimToady and tries to be generic across all registered IO types
Tene Okay, I'm a lot happier with that.
19:50 molaf left
TimToady including (I hope) the ability to lazily determine its own actual type according to usage 19:50
19:50 timbunce joined
Tene Certainly would be nice to do something like: class IO::Frob { .. }; and then have support for frob:// urls. 19:50
TimToady but maybe there's some way that can be determined statically rather than dynamically
yes, which is why I said "registered" 19:51
19:51 am0c^ joined
masak .oO( sounds a bit like Dependency Injection... ) 19:51
TimToady don't it? 19:52
IO = "Injectable Objects" :)
almost sounds like objects with a different dispatch mechanism 19:53
delegation to registrants
flussence (sounds like KDE's IO stuff...) 19:54
[Coke] whoa. macports not only installed perl 5, it complained that I needed to upgrade my java (which on os x, requires I agree to a dev license agreement...) what the hell was I installing again? ;)
hercynium uri isn't paranoid! they really *are* out to get him!
masak hercynium: or PUT, or POST, or DELETE... 19:55
TimToady maybe if he changes his name to url, he can put them off his trail
hercynium masak++
duke of url 19:58
hercynium needs to call uri, I'm almost out of cocoa :) 19:59
20:03 [particle]1 joined 20:04 [particle] left 20:05 timbunce left, [particle]1 is now known as [particle] 20:06 kst left, kst joined 20:07 cognominal joined 20:10 timbunce joined 20:11 plobsing joined 20:15 dukeleto left 20:16 dukeleto joined
moritz_ hello zebras! 20:20
tadzik o/ 20:21
moritz_ sorear: re apt-get install in chroot, yes that's the correct way
masak moritz_! \o/
20:21 cognominal left
masak dedicates himself to pushing out a p2 blog post 20:22
jnthn o/ moritz_ 20:24
colomon \o 20:32
moritz_ how are things in the Perl 6 world? 20:33
masak IO-y :)
moritz_ \IO/ 20:34
sjohnson yo friends of perl 20:36
20:36 MayDaniel left
masak sjohnson! \o/ 20:36
20:37 kst left, kst joined 20:40 plobsing left
dukeleto would like to announce that Parrot now supports unicode filename, thanks to nwellnhof++ 20:42
unicode filenames for .pir, .pbc and .pasm, that is
jnthn perl6 пиво.п6 20:43
dukeleto I trust that the #perl6 folks will have the appropriate amount of fun with this :)
masak you bet :) 20:44
TimToady wonders how parrot supports unicode filenames when the OS doesn't... 20:48
20:58 c9s left
shortcircuit are any of .pbc, .pir andd .pir generated by Parrot? 20:59
er.
.pbc, .pir and .pasm
masak all of them, likely as not. 21:00
some people insist on writing PIR by hand... :P 21:01
shortcircuit Remind me to finish that GeSHi file I was working on for PIR and PASM. 21:02
moritz_ IMCC (which is part of parrot) compilers PIR to PBC
masak the compiler that sounds like a weapon of mass destruction... 21:03
moritz_ IMCC vs. ICBM? 21:04
masak yes, that's the association I'm making when I see it.
TimToady last I heard, POSIX filesystems have no clue what their filename encodings might be
is Parrot just guessing? 21:05
moritz_ afaict they just assume the file name to be binary
TimToady one can, of course, treat filenames like blobs, but at some point you might want to print a filename in an error message 21:07
not to mention reading filenames from the command line
moritz_ error blobs! 21:08
TimToady there is no sane way forward; the least insane way forward is to assume UTF-8 unless instructed otherwise, for some insane definition of "instructed" 21:09
moritz_ looking at the local might be good first guess
TimToady as I said, "insane" 21:10
moritz_ the real trouble is that linux et. al. just don't make assumptions
all "system" file names are ASCII only
and what the user does and what the userspace programs do is left to them
aka "I'm looking in the other direction, don't bother me" 21:11
TimToady forcing the metadata out of band without giving a different band for metadata
shortcircuit You might probe during initialization; try creating a file with a unicode name, test that it was created, then remove it.
TimToady "insane"
shortcircuit I don't have an answer as to _where_ to create the file, though.
masak /usr/bin :P 21:12
what could possibly go wrong?
moritz_ /proc :-)
masak *lol*
TimToady the least insane place is '.'
shortcircuit What about TMP? 21:13
TimToady then at least you're probably in the right filesystem
shortcircuit mm. Point.
TimToady wrong filesystem
shortcircuit Ech. It hit sme that you'd have to probe for each location you might try. The best you might be able to assume is that a location will be safe once shown to be safe once.
shortcircuit wonders what a sme would be, and what hit it. 21:14
moritz_ the whole idea of creating a file at startup is completely insane
flussence on linux, couldn't you just check that %*ENV<LC_ALL> ~~ /.UTF-8$/? 21:15
TimToady there should be a way to ask the fs (or better, directory) what encoding its filenames should be in
moritz_ people will start to wonder why a program behaves differently on a read-only mounted file system, even if the program nominally doesn't write a file
or why the modification date of the directory changes whenever you run that program
TimToady either that, or require all non-utf-8 filenames to be explicitly marked with their encoding as part of the filename
moritz_ let's just use the same encoding that MIME uses for email subjects 21:16
TimToady didn't want to barf today...
moritz_ barks 21:17
TimToady just so you don't byte
flussence I say assume utf-8 and let the minority of people with oddball fs encodings jump through hoops when they want to create filenames with squiggly characters, because any other default would lead to the opposite situation and that's crazy. 21:22
shortcircuit Isn't that what Parrot did? 21:23
(more or less analogously, anyway)
21:25 plobsing joined
shortcircuit moritz_: So, Rakudo can run on a read-only filesystem? (i.e. no .pir, .pasm or .pbc files will be created) 21:25
shortcircuit needs to learn how Unicode works on Linux. He's spent too much time coding on Windows. 21:26
flussence basically, you install the OS then find out 6-8 months later you forgot to put LC_ALL=en_US.UTF-8 somewhere because all your files are showing up as ??????? :) 21:28
shortcircuit Been there, done that. Too many anime soundtracks... 21:33
flussence hehe
21:36 plobsing left, plobsing joined 21:37 stkowski joined 21:43 Br joined 21:44 klunky left 21:45 timbunce left 21:46 Br left
[Coke] bah. mono install finally failed, after several hours. 21:48
moritz_ shortcircuit: I think so, never tried 21:50
21:50 MayDaniel joined 21:51 timbunce joined 21:57 hercynium left 21:58 klunky joined 22:01 klunky left
diakopter [Coke]: did you try ftp.novell.com/pub/mono/archive/2.8...ll.x86.dmg 22:02
colomon Is that the latest stable mono for OS X? (he asked hopefully) 22:04
diakopter yeah
x86 22:05
there's also one for ppc
www.go-mono.com/mono-downloads/download.html
snarkyboojum FWIW, that's what I installed and am using and it works a charm 22:06
22:08 mtk left 22:09 klunky joined
[Coke] diakopter: <homer>noooo</homer> 22:17
diakopter: I found a resolution to the dependency failure that tripped me up earlier.
(so it's going again) 22:18
diakopter [Coke]: use the .dmg :)
[Coke] but I'M SO CLOSE!
22:30 slavik left 22:33 kaare_ left 22:35 plobsing left 22:38 awwaiid_ joined 22:40 timbunce left
masak hugs [Coke] 22:43
22:45 awwaiid_ left 22:46 timbunce joined 22:49 klunky left, awwaiid_ joined 22:50 bacek_ joined 22:54 klunky joined
szabgab crap, I cannot build parrot on win32 22:54
I just sent the output to the perl6-compiler list
[Coke] szabgab: did you use the wiki page?
szabgab nope 22:55
jnthn szabgab: More specifically, you can't build Parort on Win32 with Strawberry + its toolchain "out of the box".
szabgab which wiki page?
oh that's bad
[Coke] szabgab: trac.parrot.org/parrot/wiki/Platforms/Windows
jnthn szabgab: With ActivePerl + MSVC toolchain, it "just works". Ideally it should work for both though.
[Coke] and it used to. someone broke it.
jnthn [Coke]: *nod*
szabgab it is something "new" it worked last time I checked on this machine 22:56
[Coke] I hope to have some time tomorrow to hack on that.
22:56 shi joined
jnthn [Coke]: Suspect it's that nobody doing regular Parrot or Rakudo dev is building on that platform. 22:56
szabgab I'll be in the Alps by that time without network and I wanted to have it working
I shoudl have stayed with the one from 2 months ago :(
[Coke] jnthn: often enough that we have a workaround, though. ;)
jnthn [Coke]: Aye.
22:58 visitor_ joined 22:59 klunky left 23:00 timbunce left
szabgab gmake is not recognized as an internal or external command 23:00
but that's what the wiki said 23:01
mingw32-make works, I wonder if it will be good
oh there is a whole list at the bottom of that page 23:02
[Coke] yay, I can run niecza on my OS X box.
Is it currently slow?
(the initial say "hi" is slow, but future say invocations seem zippy.) 23:03
23:04 klunky joined
sorear It's currently slower than Perl 5 in most cases 23:06
I consider this a flaw
Parse time is a LOT slower than Perl 5 because it's using STD.pm6, not yacc 23:07
niecza treats CORE.setting as just another module to be compiled on first use, which is why the first say is so slow
visitor_ wer i can see example on array of arry slicing under perl6? 23:09
masak here! :) 23:10
23:10 hugme left
masak rakudo: my @array = 0..10; say @array[4..7] 23:11
p6eval rakudo 388eed: OUTPUT«4567␤»
masak rakudo: my @array = 0..10; say @array[4..7].perl
p6eval rakudo 388eed: OUTPUT«(4, 5, 6, 7)␤»
masak visitor_: any questions?
23:13 kfo joined
visitor_ masak: Ah, I mean if i wants to get at a slice (part of a row) in multidimensional array. 23:15
masak: same as i did under perl5? 23:16
23:16 kfo_ left
huf @array[range;range;range;.....] i guess, tho iirc the @array[range][range]... syntax works too 23:17
but i'm not sure
masak rakudo: my @matrix = [1, 2, 3], [4, 5, 6]; say @matrix[1][1..2].perl
p6eval rakudo 388eed: OUTPUT«(5, 6)␤»
masak huf: hum, I've never considered many ranges in a multidim indexing... should work, though...
23:17 justatheory left
masak ES09TOOSLUDGY 23:17
huf masak: i seem to recall seeing this in whatever i read to learn about the array syntax 23:18
but i'll be damned if i know what 23:19
masak S09.
probably.
visitor_ masak: let us say I need 1 and 2 from first array and 5 and 6 from second array, how i put that?
sorear [Coke]: need anything else? 23:21
masak rakudo: my @matrix = [1, 2, 3], [4, 5, 6]; say (@matrix[0][0..1], @matrix[1][1..2]).perl 23:26
p6eval rakudo 388eed: OUTPUT«((1, 2), (5, 6))␤»
masak visitor_: like that? :)
it's possible to do shorter, but you didn't say whether there was a pattern.
visitor_ masak: yes, it is clear and nice to look at! 23:27
masak \o/ 23:28
visitor_: as huf++ mentions, there's also a semicolon syntax, but no implementation has it yet.
visitor_ reference based method for dealing with complex data structure is no longer used un perl6 ? 23:29
masak TimToady: would the .[ $a..$b ; $c..$d ] be able to pull out a sub-rectangle of an array of arrays?
visitor_: the references are there. you just don't see them anymore. 23:30
like in Java.
dukeleto Anybody considering writing a Perl 6 version of JSYNC? github.com/ingydotnet/jsync-pm
visitor_ thank god.. it was a nightmare..
masak dukeleto: I think ingy was at some point... 23:31
dukeleto masak: i heard rumors, but I can't find any code...
masak probably best to ask ingy. :) 23:32
sorear niecza: use JSYNC; my @array; push @array, $(@array); say to-jsync(@array) 23:34
TimToady masak: yes, subrectangles are supported by S09 slicing
p6eval niecza v1-181-g98d0cc5: OUTPUT«(timeout)»
sorear niecza: use JSYNC; my @array; push @array, $(@array); say to-jsync(@array)
masak TimToady: \o/ 23:35
p6eval niecza v1-181-g98d0cc5: OUTPUT«===SORRY!===␤␤Q:CgOp not allowed in safe mode at /home/p6eval/niecza/lib/JSYNC.pm6 line 5:␤------> :CgOp { (box Str (to_jsync (@ {$obj}))) ⏏} }␤␤Q:CgOp not allowed in safe mode at /home/p6eval/niecza/lib/JSYNC.pm6 line 6:␤------>
..ns (from_j…
sorear bah
stupid safe mode
TimToady: so qp{foo} and $*IN have the same type?
TimToady the intent is to do nearly everything you can do today in PDL via a native syntax
sorear dukeleto: niecza has native JSYNC support via C# code, does that count? 23:36
TimToady I would say, more like, different types that can be used as the same type under some operations
and IO is that umbrella type
but $*IN's identity is just a file descriptor 23:37
and qp{foo}'s identity comes via a filename
dukeleto sorear: sure! That is cool.
23:38 visitor_ left
dukeleto TimToady: native PDL-like syntax is very awesome 23:38
TimToady only if it ever gets implemented :) 23:39
dukeleto TimToady: sounds fun to my sick sick mind :)
sorear TimToady: Is IO a mutable type? If $f = qp{foo}; can $f's identity change to "descriptor"?
TimToady I suspect IO is mutable in that way 23:40
though perhaps it's more like "but" mixins 23:41
sorear "but" or "does"?
TimToady does would probably be harder
but maybe is necessary 23:42
certainly filehandles have always been "mutable" with respect to, say, file pointer 23:43
23:43 am0c^ left
TimToady so maybe IO is a kind of container 23:43
masak blig blag blog. strangelyconsistent.org/blog/p2-is-...he-polygon
23:45 snearch left 23:50 vmspb joined
sorear masak++ 23:54
23:56 slavik joined
TimToady masak: in strangelyconsistent.org/p6cc2010/p2-matthias/ you missed the idiomatic use of: if $bx - $ax -> $dx { 23:56
masak that's true. adding that. 23:57
TimToady++
23:57 shi left 23:58 cafesofie joined 23:59 slavik left, slavik joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!