»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.
00:00 shinobicl___ joined 00:05 Leiting joined 00:06 shinobicl___ left 00:08 Leiting left
dalek p: 2246b9d | jnthn++ | src/HLL/Grammar.pm:
Fix preclim handling in operator precedence parser to be consistent with STD's semantics.
 00:09
jnthn pmichaud: Don't worry, 2246b9d helped. :) 00:10
dalek kudo/stdinit: 4f23c1a | jnthn++ | src/Perl6/Actions.pm:
Fix assign_op calls.
00:11 pernatiy joined
dalek p/bs: 19f1639 | moritz++ | / (2 files):
[bigint] fix detection of negative numbers from binary ops

This should fix RT #109740		 00:17
p/bs: 2246b9d | jnthn++ | src/HLL/Grammar.pm:
Fix preclim handling in operator precedence parser to be consistent with STD's semantics.
p/bs: 239b54f | jnthn++ | / (6 files):
Merge branch 'master' into bs
jnthn sleep & 00:31
00:33 lestrrat left, lestrrat joined 00:34 tty234_ joined 00:37 lestrrat left 00:38 lestrrat joined, tty234_ left 00:44 thou left 00:48 whiteknight joined 01:08 woosley joined 01:18 scott__ joined
sorear good * #perl6 01:23
eviltwin_b: AIUI, line buffering of stdin never(well ok there are corner cases) affects semantics 01:24
eviltwin_b: you may be thinking of stdout, which is line buffered if terminal, block buffered otherwise
TimToady getc works for single character input as long as you call 'stty raw -echo min 1 time 1' first (on *nix) 01:30
be sure to call 'stty sane' on the way back out though
01:39 woosley left 01:41 woosley joined 01:48 tty234 left
flussence
.oO( tty internals are the scariest thing I think I've ever encountered in *nix )
 02:02
eviltwin_b already pointed to "stty -g" which is a more accurate way to save and later restore tty settings 02:03
02:05 replore joined 02:12 sudokode left 02:15 Chillance joined 02:29 sudokode joined 02:43 mls_ left, mls joined, kmwallio joined 02:44 wolfman2000 joined 02:49 whiteknight left 03:19 shinobicl___ joined 03:20 carlin left 03:24 lestrrat left 03:25 lestrrat joined 03:27 DarthGandalf left 03:33 DarthGandalf joined, kmwallio left, wooden left 03:37 aloha left 03:38 unobe left, aloha joined 03:41 orafu left 03:42 orafu joined 03:50 unobe joined 04:18 [Coke] left 04:23 Su-Shee_ joined 04:25 wolfman2000 left 04:26 Su-Shee left
skids norole A [:$b] { has $.a = $b; }; my $a = 1; $a does A[:b(42)]; $a.a.say; my $b = 1; $b does A[:b(13)]; $b.a.say; $a.a.say;m: 04:35
erk.
nom: role A [:$b] { has $.a = $b; }; my $a = 1; $a does A[:b(42)]; $a.a.say; my $b = 1; $b does A[:b(13)]; $b.a.say; $a.a.say;
p6eval nom 2251cb: OUTPUT«42␤13␤13␤»
skids scratches head.
sorear you used 'does' on a variable of type 'Int' 04:36
you've modified the literal constant 1
nom: my $a = 1; $a does role { method Str { "Suprise!" } }; say 1
p6eval nom 2251cb: OUTPUT«Suprise!␤»
sorear did you really mean to modify a constant? 04:37
04:37 winnerdood joined
benabik I'm constantly surprised that but and does modify the variable in place rather than returning a new object. 04:37
sorear benabik: does modifies in place 04:38
benabik But I've been doing a _lot_ of functional programming.
sorear but returns a new object
that's why we have two operators!
skids Ah, well that explains why $a.a changed then :-)
benabik I thought it was because does took a role and but takes a value. 04:39
TimToady maybe we should remove does
except as a trait declartor 04:40
04:40 ponbiki joined
moritz o/ 04:41
sorear benabik: Both take both forms. 04:42
04:42 shinobicl___ left
benabik sorear: Oh! 04:42
I've never seen that, but neat. 04:43
04:48 winnerdood left 05:10 birdwindupbird joined 05:40 replore left 05:44 kaleem joined 06:16 thou joined 06:30 yertalert left 06:42 cognominal___ joined 06:44 cognominal_ left 06:47 overrosy left, Vlavv` left, ashleydev left, etneg left, kshannon left, Gothmog_ left, TimToady left, Yappoko___ left 06:50 kaleem left 06:51 overrosy joined, Vlavv` joined, ashleydev joined, etneg joined, kshannon joined, Gothmog_ joined, TimToady joined, Yappoko___ joined 06:54 kaleem joined, wtw joined 06:57 zby_home left 07:59 ponbiki left 08:04 mj41 joined 08:11 tarch joined 08:12 Khisanth left
dalek ecza: 419e3bd | sorear++ | src/niecza:
Change |$ to | in src/niecza
 08:14
ecza: 05e3637 | sorear++ | src/Niecza (2 files):
Allow defining extra operators in the setting, and implement an efficient way to load a bunch of operators into $~MAIN in one go
ecza: de81026 | sorear++ | lib/CORE.setting:
First cut at enabling Set operators
08:20 aindilis left 08:23 snearch joined 08:27 Khisanth joined
dalek ecza: 891be08 | sorear++ | lib/CORE.setting:
Switch Set and a few other things to at_key so that slicing works more sanely
 08:27
bbkr1 JSON::RPC is back on nom, including client and server 2.0 spec :) 08:29
sorear Yayness. 08:30
bbkr1 moritz: your fallback hint worked perfectly - github.com/bbkr/jsonrpc/blob/maste....pm#L10-35 - thanks! Now it's the easiest client to use ever.
08:31 Su-Shee_ is now known as Su-Shee
sorear std: constant Set term:<∅> = set(); 08:41
p6eval std 48335fc: OUTPUT«===SORRY!===␤Missing initializer on constant declaration at /tmp/BUnxDw3qZd line 1:␤------> constant Set ⏏term:<∅> = set();␤Confused at /tmp/BUnxDw3qZd line 1:␤------> constant Set ⏏term:<∅> = set();␤ expecti…
08:49 birdwind1pbird joined
masak g'morning, #perl6 08:53
dalek ecza: df96012 | sorear++ | lib/CORE.setting:
Eliminate is equiv on multi candidates; niecza should carp on this but doesn't yet. Also, forgot term:<∅>.
 08:54
sorear g'masak morning.
niecza: 1
p6eval niecza v14-23-g0699296: ( no output )
sorear evalbot rebuild niecza
p6eval OK (started asynchronously)
dalek Rebuild of niecza complete. 08:56
Timbus oh god they're like telepathically linked twins that complete each others sentences
and that is just creepy 08:57
sorear niecza: say set(2,3,4) ∩ set(3,4,5)
p6eval niecza v14-28-gdf96012: OUTPUT«set(3, 4)␤»
masak but creepy-good, right? :) 08:58
sorear shows off the new functionality ^^^
masak ooh, ∩!
sorear niecza: say ∅ 08:59
p6eval niecza v14-28-gdf96012: OUTPUT«set()␤»
masak niecza: say (2,3,4) ∩ (3,4,5)
p6eval niecza v14-28-gdf96012: OUTPUT«set(3, 4)␤»
masak \o/ 09:00
niecza: say (2,3,4) (&) (3,4,5)
p6eval niecza v14-28-gdf96012: OUTPUT«set(3, 4)␤»
masak (\o/)
sorear niecza: say (2,3,4) [&] (3,4,5)
Timbus oh good, i can type that
p6eval niecza v14-28-gdf96012: OUTPUT«Potential difficulties:␤ Useless use of [] around infix op at /tmp/4AwqQzX3B_ line 1:␤------> say (2,3,4) ⏏[&] (3,4,5)␤␤all((2, 3, 4), (3, 4, 5))␤»
Timbus why not allow 'n' and 'u' 09:01
i guess its not that common to use them 09:02
sorear thinks (&) should require more keypresses than ++
masak counts 4 for (&) and 3 for ++ 09:03
oh, but you meant as opposed to n which requires 1. I see.
yeah, I feel 'n' and 'u' are "too cute", perhaps. 09:04
if you want them, they're two lines of code.
sorear 6 for ∩ ... I need a better input method
09:04 daxim joined
sorear ctrl shift 2 2 2 9, mnemonic is not 09:05
Timbus true enough, i just cant see any harm in allowing them
sorear --> sleep 09:06
09:08 mtths joined 09:09 [particle] left 09:14 snearch left
dalek ecs: 0c822b5 | masak++ | S32-setting-library/Exception.pod:
[S32/Exception] corrected typo

Spotted by Util++.		 09:16
masak I should make a Perl 6 input method for irssi, and then share it with everybody else. :) 09:21
[backlog] +1 on removing 'does' as an operator, especially since it's basically but= 09:25
tadzik hrm, I quite like does 09:31
09:35 flussence left, flussence joined 09:50 alvis joined 09:51 woosley left, dakkar joined
masak this course intrigues me: courses.csail.mit.edu/6.851/spring12/ -- and both notes and videos upcoming. nice! 09:53
jnthn morning, #perl6
masak love the colored notes!
jnthn! \o/
jnthn: what *day* is it today? :D
jnthn NS day!
er
masak :P
jnthn Thursday?
Perl 6 day! \o/
masak Perl 6 day! \o/
jnthn The memory hierarchy section of courses.csail.mit.edu/6.851/spring12/ is certainly relevant to anyone working in VM/language implementation space, I suspect. 09:55
jnthn will have to at least check that one out.
er, does is *not* but=
masak oh :/ 09:57
jnthn does modifies the *object* in place. Assignment is all about scalar containers. 09:58
09:59 thou left
masak hm, yes. 10:03
but if I do 'my $obj = MyClass.new; $obj but= MyRole', there's never more than one instance, is there? 10:04
moritz but clones first 10:05
jnthn masak: Yes, but in this case you immediately throw it away 10:06
my $obj = MyClass.new; my $orig = $obj; $obj but= MyRole;
In the above $orig doesn't have the role, but still exists 10:07
my $obj = MyClass.new; my $orig = $obj; $obj does MyRole;
Now $orig and $obj (which of course point to the same object) have the role
Well, one object has the role but two containers point to it
Grr, not enough coffee to explain stuff
masak right. I see it now.
while I appreciate the distinction, it might be that it's too confusing to have both behaviors. people seem to get 'does' and 'but' mixed up. 10:08
Su-Shee good morning everyone. 10:09
jnthn masak: I'd rather fix the naming that lose functionality
*tyhan
oh, screw it, I'm going to make some coffee
o/ Su-Shee
masak Su-Shee! \o/ 10:10
10:11 scott__ left
tadzik hello :) 10:11
10:12 overrosy left
masak tadzik! \o/ 10:14
10:15 overrosy joined 10:22 kaleem left
moritz bbkr1: good to hear that the fallback worked. 10:30
10:35 [particle] joined 10:39 Trashlord left 10:40 Trashlord joined
jnthn ok, now I'm more caffienated :) 10:42
moritz: stdinit may be worth a spectest run. As far as I know, only two test files are busted, and it's becuase somehow my $x = [=>] 1,2,3; style things are
(combo of [=>] reduction and assignment) 10:43
moritz: Also your $*HAS_SELF stuff should work on top of this branch :) 10:48
moritz jnthn: does it require a newer nqp? 10:55
10:55 kfo left
jnthn moritz: Requires the preclim patch I did yesterday, yeah 10:56
moritz: oh, but I forgot to bump NQP_REVISION
moritz ok 10:58
11:03 mtk left 11:14 birdwind1pbird left, lestrrat left 11:15 lestrrat joined 11:18 lestrrat left 11:19 lestrrat joined 11:25 jaldhar left, ab5tract joined 11:30 ab5tract left 11:31 ab5tract joined 11:34 au left 11:58 fasad joined
fasad std: sqrt( 3++ ); 11:59
p6eval std 48335fc: OUTPUT«ok 00:01 108m␤»
fasad rakudo: sqrt( 3++);
p6eval rakudo 2251cb: OUTPUT«Cannot assign to a non-container␤␤»
fasad :(
niecza: sqrt ( 3++ );
p6eval niecza v14-28-gdf96012: OUTPUT«Unhandled exception: Writing to readonly scalar␤ at /tmp/N1CYVd6L9q line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3746 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3747 (module-CORE @ 65) ␤ at /home/p6eval/niecza/lib…
fasad :( 12:00
pugs: sqrt( 3++ );
p6eval pugs b927740: OUTPUT«*** Can't modify constant item: VInt 3␤ at /tmp/CC0SR6oAV0 line 1, column 7-11␤»
jnthn fasad: There's this other operator that will work on literal values, "+1" :)
perl6: say sqrt(3+1)
p6eval pugs b927740, rakudo 2251cb, niecza v14-28-gdf96012: OUTPUT«2␤» 12:01
jnthn ;)
fasad what does the rakudo error mean ?
jnthn fasad: What you can't assign to something if it's not a container
*That
++ expects to get a variable
fasad jnthn: yeah, but ++ is supposed to work (?) 12:02
jnthn nom: my $a = 3; $a++; say $a
p6eval nom 2251cb: OUTPUT«4␤»
jnthn Works fine.
You can't do 3++ because you'd be trying to modify the number 3 12:03
Which is a constant/literal.
fasad std says the sqrt( 3++ ) is correct syntax
*that
jnthn Sure
It's fine *syntax*.
It's not semantically correct, though.
fasad oh ! duh. 12:04
jnthn std: class A { }; A.does_not_exist
p6eval std 48335fc: OUTPUT«ok 00:01 107m␤»
fasad !
jnthn That's another example of fine syntax. But won't work if you run it :)
12:05 bluescreen10 joined
fasad perl (v5) is an interpreter; perl6 (rakudo) is... ? 12:08
i mean there is a virtual machine in between rakudo and the machine, so what does it make rakudo? 12:10
jnthn A compiler.
It compiles your program into code for the virtual machine.
moritz fasad: both p5 and Rakudo (and basically all modern implementations of dynamic languages) are a hybrid between compiler and interpreter 12:11
fasad: the distinction blurs
fasad: though rakudo is maybe a bit more compiler-y. Hard to tell.
jnthn Yeah. It's not such a useful distinction once you start asking interesting questions about what's really going on :) 12:12
fasad hmm. Is gcc a "pure" compiler?
moritz in particular, both p5 and p6 need the facility to execute code at compile time (more than just constant folding)
jnthn The virtual machine itself may be doing either interpretation or (JIT/AOT) compilation. 12:13
moritz fasad: much more so than rakudo and parrot
fasad: but it also provides some runtime library stuff, so it's not 100% "pure" compiler
fasad jnthn: AOT ?
moritz ahead of time compilation 12:14
masak perl6: say exp(pi * sqrt 163) - 640320 ** 3
p6eval rakudo 2251cb: OUTPUT«-12032460128␤»
..pugs b927740, niecza v14-28-gdf96012: OUTPUT«256␤»
masak neither of those results correspond to en.wikipedia.org/wiki/Complex_multi...onsequence
moritz nom: say exp(pi * sqrt 163) 12:15
p6eval nom 2251cb: OUTPUT«2.62537400608308e+17␤»
masak loss of precision, most likely. 12:16
moritz perl6: 640320 ** 3
p6eval pugs b927740, rakudo 2251cb, niecza v14-28-gdf96012: ( no output )
moritz perl6: say 640320 ** 3
p6eval pugs b927740, rakudo 2251cb, niecza v14-28-gdf96012: OUTPUT«262537412640768000␤»
moritz masak: but it's worrysome that the numbers diverge so much
nom: say 262537412640768000.Num
p6eval nom 2251cb: OUTPUT«2.62537412640768e+17␤»
masak niecza: say exp(pi * sqrt 163)
p6eval niecza v14-28-gdf96012: OUTPUT«2.6253741264076826E+17␤»
masak Niecza carries a few more significant digits, is all. 12:17
so it fares better.
dalek p/bs: ffeec9c | jnthn++ | src/6model/serialization.c:
Set SC on objects as we deserialize them.
p/bs: 7275a66 | jnthn++ | src/6model/serialization.c:
Serialize/deserialize v-table and method cache from STable.
p/bs: 2ea450f | jnthn++ | src/6model/serialization.c:
Introduce concept of a cloned code reference; this will be used when implementing closure serialization.
masak still, 256 isn't so close to the expected 744, either... 12:18
moritz jnthn: I can confirm your results from spectesting stdinit 12:19
jnthn moritz: OK. So just the reduction operator thingy to fathom.
moritz jnthn: right, and bumping NQP_REVISION 12:21
jnthn moritz: Feel free to pull your $*HAS_SELF patch into that branch too, or some other way to try them in combination.
moritz: I'm pretty sure the "has $!x = $!y" issues are now fixed.
moritz jnthn: I'll try it
colomon niecza: say "a" ∈ <a b c d e>
p6eval niecza v14-28-gdf96012: OUTPUT«Unhandled exception: System.Collections.Generic.KeyNotFoundException: The given key was not present in the dictionary.␤ at System.Collections.Generic.Dictionary`2[System.String,Niecza.P6any].get_Item (System.String key) [0x00000] in <filename unknown>:0 ␤… 12:22
colomon !
jnthn If Pm shows up today, maybe he'll have some hints on the reduction issue.
(or feel free to poke if you've got any ideas)
I can't quite figure out how STD gets it right.
colomon p6eval: rebuild niecza 12:23
dalek kudo/nom: 828e4d6 | moritz++ | src/core/Exception.pm:
allow numification of X::AdHoc

This is a bit questionable, but a spectest requires it In the end we might want to make X::Adhoc a wrapper around a payload, and only use the stringification for HLL interop.		 12:24
moritz evalbot rebuild niecza
p6eval OK (started asynchronously)
dalek Rebuild of niecza complete.
moritz jnthn: it's a parse issue?
the error comes from REDUCE_METAOP_RIGHT where it tries to invoke infix:<=>> with one argument 12:25
jnthn moritz: Yes, precednece issue in parsing 12:26
my $x = [+] 1,2,3; say $x
That also shows it much more simply.
It stops parsing at the , for some, because of the precedence limit. 12:27
s/for some,//
masak because the = imposes scalar assignment? 12:30
moritz the precedence limit from infix:<=> ?
masak inneresting -- that's correct from the viewpoint of the infix:<=>, but the [+] creates a new parsing context "one level down", so to speak, which should lift that constraint. 12:32
jnthn moritz: Correct
masak: Yes, and I fail to see where STD does that.
masak std: my $x = [+] 1, 2, 3; 12:33
p6eval std 48335fc: OUTPUT«ok 00:01 112m␤»
12:33 pernatiy left
masak clearly it does that, though. 12:33
moritz well no
it wouldn't syntax-error
jnthn right
moritz (my $x = [+] 1), 2, 3 is also valid p6 12:34
jnthn std: (my $x = [+] 1), 2, 3
p6eval std 48335fc: OUTPUT«ok 00:01 113m␤»
jnthn yeah, and the 2 and 3 ain't warned about
masak huh.
indeed.
jnthn Well
I'm only guessing STD gets it right because Niecza does
It's possible Niecza has got something STD hasn't got here. 12:35
masak check how Niecza does it :)
jnthn wonders if :sub<e=> in %list_prefix would fix it
12:35 am0c joined
jnthn or in reduce at least 12:35
12:36 am0c left
jnthn alas, no 12:39
moritz do prefixes generally reset the prec limit?
jnthn Don't believe so 12:40
masak just had an insight: unary prefixes are very loose on the term level, and list prefixes are very loose on the expression level 12:41
12:53 tarch left 12:58 au joined
cognominal___ I was discussing with rgs and elbeho on #perlfr about Mathematica. I note something which is interesting for people fearing typing weird symbol. Expressions have a fullform and a short form. In the full form, everything is expressed like C functions except as bracket instead of parentheses. 13:02
The programmer types code in a form that is a mix of full and short form.
Example of short from : a*b + c 13:03
Coreesponding full form: Plus[Times[a,b],c] 13:04
Operators which are unicode can be typed in full form to avoid unicode but will eventually appear in short form 13:06
reference.wolfram.com/mathematica/t...guage.html
flussence I guess in vim you could write a bunch of ":imap <Leader>intersect ∩" lines too 13:12
masak not a bad idea.
and then have the same shortcuts in irssi... :) 13:13
flussence though with that you have to type the entire thing in one go without a single typo or it aborts, so maybe not for everyone :)
masak oh.
maybe something more like \cap in TeX? 13:14
cognominal___ You don't have to type: a*b+c in longform though
Before knowing about compiler internals, I probably got from Mathematica the idea that program are really not lines but trees. 13:16
moritz +1 to exposing all set operations as (ASCII-named) methods (or subs) 13:17
cognominal___ Chomsky would talk about surface syntax, but I am not sure TimToady likes Chomsky too much. 13:18
moritz: that could be a trait.
masak I never fully bought into the concept of surface syntax. it seems to have contradictory meanings in the literature. 13:19
moritz cognominal___: a trait? how so?
cognominal___ …specifying how to go form the long form to the short one. how so, not to sure yet. 13:21
flussence macros :) 13:22
cognominal___ In mathematica that's a tree pattern matching transformation.
flussence (and I think maybe it should be the other way around, latin-1 should be the canonical form and unicode should be the sugar on top) 13:23
cognominal___ With signatures, we do pattern matching already except they are expressed in a weird "short form". 13:24
13:25 skids left
moritz Mathematica is very Lisp-y under the hood 13:25
cognominal___ Well perl has always been about short forms except TimToady calls that huffman coding.
I think Mathematica as a cross of Lisp and Prolog. 13:26
moritz ProLisp :-)
cognominal___ Should go Back in 2 or 3 hours 13:27
masak there's two sides to Huffman coding. sometimes things get longer names because of it, too.
13:29 ab5tract left
arnsholt masak: As (more or less) a linguist, I don't really buy the Chomskian deep/surface distinction either =) 13:31
If it's any consolation
masak it is. it mostly serves to confuse me. 13:32
as a computer scientist, I don't see it having rigour enough.
dalek kudo/stdinit: 7b6cf5d | moritz++ | src/ (3 files):
Merge remote branch 'origin/has-self' into stdinit
kudo/stdinit: 48c50ff | moritz++ | tools/build/NQP_REVISION:
bump NQP revision to get preclim fix
arnsholt Yeah, that's my main beef with Chomskian linguistics in general
jnthn moritz: has $!y = $!x # works in stdinit? 13:33
arnsholt They make a great issue of economy and some vague notion of "possible to compute", but Chomskian linguistics are pretty much completely absent from NLP/compling
Transformational syntax is just too hairy to implement as a computer program 13:34
(Only counter-example I can think of is something called tree-adjoining grammar, but that's pretty old)
moritz jnthn: I haven't tested it specifically, but the merge only added passing TODOs to the spectest run, so I guess "yes" 13:35
13:35 pernatiy joined
jnthn moritz: yay 13:36
So, just need to fathom this reduce thingy and then it can be merged.
Am working on the serialization bits at the moment, though.
arnsholt And speaking of unicode operators and vim, it should be possible to set up imaps so that the texas operators are magically transformed into unicode ops
13:37 benabik left
moritz jnthn: yes, works 13:39
jnthn \o/ 13:41
masak arnsholt: that might go wrong, though. think of %h<bar>».foo -- I wouldn't want vim to turn that into %h<bar»>.foo 13:49
arnsholt Oooh, good point
moritz jnthn: yes, works 13:50
jnthn moritz: The same thing you told me works 11 minutes ago? ;) 13:51
moritz jnthn: now Ronja told you :-)
jnthn :D
masak she's getting much better at typing words. 13:52
13:52 ajox joined
moritz amazing, isn't it? she must have learned a lot from me :-) 13:56
13:58 [Coke] joined 14:00 mtk joined 14:03 ab5tract joined
dalek ecza: 6a63017 | (Solomon Foster)++ | lib/CORE.setting:
Restore the immutability of Set and Bag.
 14:23
14:23 chra left 14:24 chra joined
masak colomon: an interesting case of how easy it is to accidentally break encapsulation by handing out references through accessor methods. 14:26
14:27 tarch joined, birdwind1pbird joined
colomon masak: yup 14:27
14:30 etneg_ joined, Vlavv` left, Gothmog_ left, etneg left, Gothmog_ joined, Vlavv` joined
dalek ast: 8f304c2 | (Solomon Foster)++ | S02-types/key (2 files):
Unfudge tests that work after sorear++'s changes last night.
 14:31
14:36 birdwindupbird left 14:44 skids joined
skids which git branch to use for most functional REPR stuff, e.g. CStruct? 14:49
masak nom, I think. 14:50
jnthn Right.
Note the implementation of CStruct actually lives in the NQP repo 14:55
skids Will nom fetch the newest with --build-nqp? 14:56
erm, --gen-nqp?
14:57 PacoAir joined
jnthn skids: It'll fetch recent enough (more) 14:57
There hasn't been any additions to CStruct et al in the last couple of weeks 14:58
14:58 unobe left
skids Thanks, jnthn++ 14:59
14:59 birdwind1pbird left, unobe joined 15:10 tty234 joined 15:23 kaleem joined 15:26 GlitchMr joined
moritz but if you want to patch stuff, you should 'git checkout master; git pull' in nqp before you start 15:29
15:30 kaleem left
[Coke] # 02/09/2012 - niecza at 99.64% 15:30
"niecza", 19553, 150, 732, 1582, 22017, 22863
"rakudo", 19623, 32, 616, 1977, 22248, 22873
sorear, colomon: note that niecza has 150 failures atm. 15:31
moritz [Coke]++ # test statistics, feeding the friendly competition
masak nom: sub postfix:<%>($n) { $n / 100 }; say "Rakudo at {(1/99.64% * 100).fmt("%6.2f")}%" 15:32
p6eval nom 828e4d: OUTPUT«Rakudo at 100.36%␤»
moritz masak: could I add you as possible mentor for the custom sprintf implementation in wiki.enlightenedperl.org/gsoc2012/ideas ? 15:33
masak yes.
[Coke] the detailed reports are at feather.perl6.nl/~coke/ *.out 15:35
jnthn hmm, I need to review that multi dispatch one
I did a bunch of the work there already
moritz jnthn: btw rakudo commit 33fb020a12b6eccc2b1e0118d9c1a362ecd27d39 contains a commented-out multi method that causes circularity 15:37
I wonder if that's legit 15:38
jnthn Unlikely 15:39
(unlikely to be legi)
*legit
PerlJam There's only 3 gsoc ideas for Rakudo?
[Coke] rakudo is haaaaaaard. 15:40
PerlJam given the number of Perl projects under the TPF moniker, it seems unlikely that Perl 6 will get much love. 15:42
15:42 kaleem joined
moritz PerlJam: so far we've always got enough slots for Perl 6 projects, once we had interested and good students 15:43
15:44 Psyche^ joined
PerlJam well, that's really the trick anyway, isn't it? Finding good students. 15:44
dalek p/bs: b06857a | jnthn++ | src/6model/reprs/P6opaque.c:
P6opaque should ensure its REPR data is calculated before any serialization of it takes place.
 15:45
p/bs: f6f2e8a | jnthn++ | t/serialization/03-closures.t:
Add initial test for handling of cloned code refs (not testing outer handling yet, though - just very simple case of cloning).
p/bs: f31afe1 | jnthn++ | src/6model/serialization. (2 files):
Initial changes to add a closures table.
p/bs: 317ff89 | jnthn++ | src/6model/serialization. (2 files):
Start sketching out cloned code ref serializtion, in prep for really doing closure support.
p/bs: b81e767 | jnthn++ | src/6model/serialization.c:
Add initial code for deserializing a row from the closures table; this will eventually grow to be the full closure deserialization, but doing something much simpler for now.
p/bs: be303bd | jnthn++ | t/serialization/03-closures.t:
Update test case to actually really use the static code object.
p/bs: bf9b58e | jnthn++ | src/pmc/nqplexinfo.pmc:
Get NQPLexInfo to remember the Sub it is attached to.
p/bs: 664e87c | jnthn++ | src/6model/serialization.c:
Locate static code ref that a closure is based on, plus a few other fixes. First tests for cloned code objects now pass; this is the first pre-requisite for closure serialization done.
PerlJam starts crafting an appropriate email to the local CS club about the GSoC
[Coke] should really do better outreach to the local schools. (RPI is one of the locals, I should really be able to find someone there. ;) 15:46
15:47 Patterner left, Psyche^ is now known as Patterner
[Coke] what does the bs in nqp/bs stand for? 15:50
tadzik Boom! Serialization!
PerlJam "I can't be bothered to come up with a real name" :)
tadzik Multiple Dispatch Optimization in Rakudo -- isn't this already done by jnthn++? 15:51
15:51 kaare_ joined
flussence I guess the b is binary 15:51
tadzik bounded
flussence
.oO( as opposed to xml serialisation... ew. )
15:57 kaleem left 16:02 benabik joined
jnthn bounded serialization :) 16:04
16:04 yertalert joined
masak tadzik: yes, but the project suggestions are from last year. 16:07
tadzik yeah, I think so
well, sprintf is new
masak tadzik: it's likely that it can be updated to do even cooler optimizations.
tadzik sniffs for a nice task for this year
jnthn Yeah, I'm pondering the best way to update that task :) 16:08
colomon [Coke]: I already sent a list of failing test files to sorear++ via e-mail. :) 16:14
16:18 wtw left
TimToady Ooh, set ops, shiny! Used to be I couldn't spell mathmatishun, and now ir1. 16:20
16:25 mj41 left 16:32 Trashlord left
masak niecza: say "a" ∈ <a b c d> 16:36
p6eval niecza v14-29-g6a63017: OUTPUT«Unhandled exception: System.Collections.Generic.KeyNotFoundException: The given key was not present in the dictionary.␤ at System.Collections.Generic.Dictionary`2[System.String,Niecza.P6any].get_Item (System.String key) [0x00000] in <filename unknown>:0 ␤…
masak something the matter there, though.
16:37 Trashlord joined, benabik left 16:43 fasad left 16:46 kaleem joined
TimToady maybe we need a sqrt(Rat, :precision) that returns a Rat 16:47
.oO(square rats)
 16:48
or some variant of ** that returns a Rat 16:49
masak could anon slurpy arrays *@ be made lazy? 16:52
sorear good * #perl6
TimToady o/
masak sorear! \o/ 16:53
sorear TimToady: sqrt(2).Rat(1e-9) isn't enough?
TimToady irclog.perlgeek.de/perl6/2012-02-09#i_5123826
masak: maybe it's too early, or too late, but I can't picture what you're asking 16:54
jnthn masak: Lazy as in, just throw away the rest of the values?
masak: Rather than actually make an array?
Though Rakudo may do that already...
masak jnthn: right.
so you can pass lazy, potentially infinite lists to the sub. 16:55
jnthn masak: But you should already be able to do that for the named ones...
masak I was looking at the first example of en.wikipedia.org/wiki/Corecursion#Discussion
and thinking how to port it to Perl 6.
nom: sub fibgen { $^x, fibgen $^y, $x + $y }; .say for fibgen 0, 1
p6eval nom 828e4d: OUTPUT«maximum recursion depth exceeded␤ in sub fibgen at /tmp/UI8CzVd3TC:1␤ in sub fibgen at /tmp/UI8CzVd3TC:1␤ in sub fibgen at /tmp/UI8CzVd3TC:1␤ in sub fibgen at /tmp/UI8CzVd3TC:1␤ in sub fibgen at /tmp/UI8CzVd3TC:1␤ in sub fibgen at /tmp/UI8CzVd3TC:1␤ in sub fibg…
sorear masak: (backlog) I'd be +1 to dehuffmanizing does, perhaps to .APPLY_ROLE_IN_PLACE (lc?) or something like that 16:56
masak sorear: sounds good.
sorear making it a (meta?) method would hint at the fact that it mutates the object
masak sorear: meta sounds about right for it.
jnthn Doesn't need to be uppercase. It's not *that* scary.
masak .^distort 16:57
TimToady I was completely unsuccessful in getting hamming recursion to work last night on either niecza or nom
.oO(use MONKEY_DOES)
jnthn masak: huh
.^mixin
Call it what it does!
masak ooh
jnthn We don't need to make new words for everything :)
PerlJam jnthn: +1
masak I didn't think of that you can verb 'mixin'.
jnthn Surely it's a nonification of "to mix in" ;) 16:58
16:58 thou joined
jnthn nom: class A { }; role R { method foo() { 42 } }; my $a = A.new(); $a.^mixin(R); say $a.foo 16:58
p6eval nom 828e4d: OUTPUT«42␤»
jnthn It even works already. :)
sorear masak: "interesting case" - red herring. Niecza still treats all methods as 'is rw', which is the problem here 16:59
16:59 fsergot joined
fsergot o/ 16:59
masak sorear: oh! I considered whether that might be part of it. 17:00
sorear o/
TimToady perl6: constant @foo = 1, @foo; say @foo[3] 17:01
p6eval pugs b927740: OUTPUT«pugs: *** Undeclared variable: ("@foo",MkPad (padToList []),[PCompiling {pc_pad = MkMPad {mp_id = 140635094937585, mp_pad = <ref:0x7fe828ae7d99>}}])␤ at /tmp/HXoh5Rds_j line 1, column 20-24␤»
..rakudo 828e4d: OUTPUT«===SORRY!===␤Variable @foo is not declared␤at /tmp/RRmhwfflvi:1␤»
..niecza v14-29-g6a63017: OUTPUT«Any()␤»
masak still, we've established that the *general* problem of DTRT with encapsulation of private data structures is unsolvable, or at least never solved satisfactorily in a language with assignment.
TimToady perl6: my @foo; @foo.push: 1, @foo; say @foo[3]
p6eval rakudo 828e4d, niecza v14-29-g6a63017: OUTPUT«Any()␤»
..pugs b927740: OUTPUT«␤»
masak decommutes
jnthn TimToady: iirc, we'd discussed ".plan" instead of ".push" to make cases like that work. 17:03
er, if I understand what you were after :)
TimToady er, right
perl6: my @foo; @foo.plan: 1, @foo; say @foo[3]
p6eval rakudo 828e4d: OUTPUT«Method 'plan' not found for invocant of class 'Array'␤␤»
..pugs b927740: OUTPUT«*** No such method in class Array: "&plan"␤ at /tmp/8X_UxfXLnX line 1, column 10-28␤»
..niecza v14-29-g6a63017: OUTPUT«Any()␤»
jnthn Hm. Seems we didn't implement .plan in Rakudo, though... 17:04
TimToady still don't work though
17:04 ajox left
PerlJam .plan ?? 17:05
daxim perlmonks.org/index.pl?node_id=952765
17:09 kaleem left 17:10 benabik joined 17:12 MayDaniel joined
[Coke] hopes kikuchiyo filed a bug report! 17:12
jnthn sighs about Rakudo's list handling
flussence nom: my $s; for 1..10 {$s+=1/$_**2}; say $s 17:13
p6eval nom 828e4d: OUTPUT«1.54976773116654␤»
flussence nom: my $s; for 1..10 {$s+=1/$_**2}; say $s.WHAT 17:15
p6eval nom 828e4d: OUTPUT«Rat()␤»
flussence nom: my $s; for 1..358 {$s+=1/$_**2}; say $s.denominator 17:16
p6eval nom 828e4d: OUTPUT«1449787577947121375205308609556609711535433536157696434138530380517766506486305567414633608794650140470225443769804050741336943542145049679650936206123191521214587589518173757956538010055687965209269408060586721810178407168365269586498897970105441515066452356030319783…
flussence nom: my $s; for 1..358 {$s+=1/$_**2}; say $s.denominator.chars
p6eval nom 828e4d: OUTPUT«304␤»
flussence 304 looks familiar... decimal exponent for the largest 32-bit float, I think 17:17
TimToady it works find it Num arithmetic, if you cange the 1/ to 1e0/
*change
flussence nom: my $s; for 1..359 {$s+=1e0/$_**2}; say $s.denominator.chars 17:18
p6eval nom 828e4d: OUTPUT«Method 'denominator' not found for invocant of class 'Num'␤␤»
flussence wha
TimToady Nums don't have denominators
perl6: my $s; for 1..1000 {$s += 1e0/$_**2};say $s 17:19
p6eval niecza v14-29-g6a63017: OUTPUT«Use of uninitialized value in numeric context␤ at /home/p6eval/niecza/lib/CORE.setting line 1221 (warn @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 255 (Any.Numeric @ 6) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/BEs41VM9GV line 0 (mainli…
..pugs b927740: OUTPUT«1.6439345666815598031390580238222155896521␤»
..rakudo 828e4d: OUTPUT«1.64393456668156␤»
TimToady niecza gets the correct answer, but you can't see it for the noise 17:20
PerlJam pugs++ too
TimToady perl6: my $s = 0; for 1..10000 {$s += 1e0/$_**2};say $s 17:21
17:21 jaldhar joined
p6eval rakudo 828e4d: OUTPUT«1.64483407184807␤» 17:21
..pugs b927740: OUTPUT«1.644834071848059769806081833310310903538␤»
..niecza v14-29-g6a63017: OUTPUT«1.6448340718480652␤»
dalek ecza: 7cb1d49 | sorear++ | lib/CORE.setting:
Fix (elem) private method issue
sorear out
TimToady nom: my $s; for 1..358 {$s+=1/$_**2}; say $s.WHAT 17:26
p6eval nom 828e4d: OUTPUT«Rat()␤»
moritz daxim: thanks for the link, I've written a reply
TimToady is it really a FatRat in disguise?
moritz yes 17:27
in rakudo, each Rat is a FatRat :/
jnthn moritz: Is there a (huge amount) more to a first crack at fixing that beyond "has Int $.numerator; has int $.denominator;" ? 17:28
moritz jnthn: well, handling overflow 17:29
jnthn ah, yes. 17:30
jnthn just checked the relevant part of S02 and is happy to see we don't auto-promote to FatRat :)
17:31 Tedd1 left
daxim "rancorousness", hah 17:31
TimToady and it's supposed to be Rat[Int,Uint64], actually 17:32
17:33 snearch joined
TimToady I don't know where moritz++ got 2**128 from... 17:33
but I like the way he replied better than the way I didn't :)
17:34 Tedd1 joined
jnthn oh...the O(N**2) is not due to the iteration. *phew* 17:35
[Coke] moritz++ 17:36
moritz TimToady: memory gone wrong 17:37
niecza: say ( 1 / 2**65).WHAT 17:41
p6eval niecza v14-29-g6a63017: OUTPUT«Num()␤»
moritz jnthn: maybe if we make infix:</> sufficiently magic...
and build the rest of the Rat arithmetic based on that instead of Rat.new 17:42
dalek ast: 355851a | (Solomon Foster)++ | S02-types/keybag.t:
Make the multiple-element access tests stronger.
 17:53
17:56 MayDaniel left
moritz niecza: my $f = FatRat.new(1, 1); my $r = 1.0; say ($f * $r).WHAT; say ($r * $f).WHAT 17:56
p6eval niecza v14-29-g6a63017: OUTPUT«FatRat()␤FatRat()␤»
flussence perl6: say (1..'potato')[^10]; # www.reddit.com/r/programming/commen...to_potato/
moritz so FatRat is contagious? 17:57
p6eval rakudo 828e4d: OUTPUT«(timeout)»
..niecza v14-29-g6a63017: OUTPUT«Unhandled exception: Cannot parse number: potato␤ at /home/p6eval/niecza/lib/CORE.setting line 1359 (die @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3318 (ANON @ 10) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3320 (NumSyntax.str2num @ 4) ␤ …
..pugs b927740: OUTPUT«␤»
TimToady "The best approach to Perl 6 is to ignore it." --educated_foo
.oO(if only he would...)
jnthn No, no. "troll about" is a synonym of "ignore" these days :P 17:58
17:58 pernatiy left
TimToady yes, FatRat is contagious, which is why we try never to make one implicitly 17:59
17:59 fasad joined 18:02 retup-work joined 18:03 Teratogen left 18:09 dakkar left
moritz jnthn: is there a way to write a sub or block in the setting which is inlined? 18:20
jnthn moritz: What do you mean by "inlined"? In the optimization sense? 18:22
moritz jnthn: yes
jnthn moritz: The optimizer already takes care of some basic cases of that, but we're a bit limited in what we can do until we can serialize ASTs properly. 18:23
moritz: It already will inline things it knows how to when compiling CORE.setting, though.
moritz that's... vague 18:24
jnthn moritz: For something more concrete, see add_inlining_info_if_possible in Actions.pm, and inline_call in Optimizer.pm. 18:26
18:27 tarch_ joined
moritz only single statements. That won't do for my purpose :( 18:28
18:29 tarch left
jnthn moritz: Yeah, it's very limited right now. 18:29
moritz: I can extend it a bit, I suspect. But given it's mostly going to get a lot easier to do when we have serialization (which I'm working on now) and AST being 6model objects (which is up afterwards) I don't want to write too much to throw away. 18:30
moritz jnthn: 's ok 18:32
jnthn moritz: What's the use case, ooc?
18:33 fasad left
moritz jnthn: I thought about a sub that takes two Rational-ish types, a numerator and a denominator, and automagically makes a FatRat, Rat or Num 18:33
jnthn moritz: Do you expect a serious performance regression if you do it that way and we don't have inlining of it just yet? 18:34
moritz jnthn: well, one additional call per arithmetic operation
18:35 preflex_ joined, preflex left 18:36 preflex_ is now known as preflex
jnthn moritz: We can probably stomach that for the time being, in return for corrector semantics. 18:37
18:40 tarch__ joined 18:44 tarch_ left
dalek p/bs: e4f5e80 | jnthn++ | t/serialization/03-closures.t:
Add a failing test for closure serialization.
 18:45
p/bs: 40269bf | jnthn++ | src/6model/serialization.c:
Add a column to the closures table for the outer context.
p/bs: 9bdbaa1 | jnthn++ | src/6model/serialization. (2 files):
Update serialize/deserialize code to set up for contexts table and data segment.
p/bs: 5ee3a41 | jnthn++ | t/serialization/03-closures.t:
Correct test count.
jnthn Time for dinner break.
colomon moritz: what would be the purpose of such a sub? shouldn't it just be Rat or Num? 18:48
or I guess my question is, what are the circumstances in which you'd want to choose between those three choices, given two Ints?
moritz colomon: I consider how to write all the numeric ops 18:49
my idea ist
multi sub infix:<+>(Rational $a, Rational $b) { 18:50
my Int $nu = ...;
my Int $de = ... ;
MAKE_NUMBER($nu, $de, $a.WHAT, $b.WHAT)
}
colomon ah
moritz and depending on the actual types of $a and $b, that can return Rat, FatRat or Num
colomon though if you're going as far as Rational, there are a lot more choices than Rat and Num, right? 18:53
moritz yes
which means that we just need to make MAKE_NUMBER more general in the long run
18:55 PZt left
colomon bonus points if you make the conversion to Num smarter than just $nu.Num / $de.Num.... 18:55
moritz assuming that $nu and $de are already shortened, is there a smarter way? 18:56
18:56 PZt joined
moritz (or how do you call that? cancled out? /me bad at math English) 18:57
ah, cancled down probably
18:58 MayDaniel joined
eviltwin_b mormalized? 19:02
er
normalized
19:06 birdwindupbird joined 19:13 cogno joined 19:15 zby_home_ joined
colomon moritz: sorry I got distracted. The case I'm worrying about is that $nu.Num or $de.Num maybe be Inf, but the ratio of the two parts would make a perfectly good Num. 19:18
moritz colomon: oh right. Then it might make sense to divide them both by an appropriately large integer first 19:20
colomon really, this seems like the sort of thing the bigint library ought to already be smart enough to do. :) 19:21
but yeah, dividing by an appropriately large integer would probably be an okay approach
19:26 cogno left 19:27 pernatiy joined
moritz colomon: can you think of a different appraoch than dividing by a large int? 19:32
colomon I was thinking more like $nu div $de, and only looking at ($nu % $de) / $de if needed 19:36
but you're likely to end up dividing the last expression by a large int if $de.Num == Inf, it's true... 19:37
moritz that sounds a bit cleverer, actually :-)
19:37 bluescreen10 left
PerlJam I don't know what else the Swedes have contributed to the world, but right now I'm very thankful for swedish meatballs. 19:39
colomon swedish meatballs++
tadzik is that called 'shotbullars' or something like this? 19:40
moritz tadzik: that's how it's pronounced, but it's spelled differently, iirc
with kj in the beginning, or so
tadzik yeah, I don't recall seeing it written anywhere :)
moritz erm, just k 19:41
"köttbullar" it seems
PerlJam the only problem is that I don't have enough of them. 19:42
(But I guess that's really a blessing so I don't get fat from eating too much) 19:43
moritz PerlJam, colomon, tadzik: can I motivate you to write some strategies for the prisoner's dilemma?
(see perlgeek.de/blog-en/perl-6/iterated...lemma.html for the details)
colomon moritz: I've been meaning to, but my tuits are completely messed up at the moment
I'll make no promises, but I'll see what I can do.
tadzik moritz: certainly, counting from tomorrow :)
(I have one last exam tomorrow on 8AM 19:44
)
moritz \o/
tadzik then I should be back to the usual slacking o^W^Whacking routine
colomon is the only healthy person in his immediate family, meaning he is on childcare duty, supposedly also working a full time job, and is superhumanly tired 19:45
tadzik ouch 19:46
colomon.energy++
moritz colomon: it's similar here, though I can reduce the "full time job" part as needed
colomon admittedly I'm messing around with Bags at the moment, because I don't have enough brainpower handy to do $work
PerlJam moritz: maybe 19:47
Though to do anything but someple simple I'd need time to play with it and time always seems to be in short supply 19:48
s/someple/something/ # wow.
19:49 daxim left, bluescreen10 joined
PerlJam moritz: here's one now though: sub (:@theirs, *%) { +@theirs %% 2; } # :-) 19:52
pyrimidine moritz: it's best not to feed the _foo, he never comes around 20:01
moritz pyrimidine: I fear I can't resist at the moment :( 20:02
20:08 tarch__ left
pyrimidine moritz: I think he misses a major point. Where would Perl 5 be if Perl 6 never existed? Would perl 5.10 have happened? Would there be a Moose? 20:09
moritz pyrimidine: there are a whole lot of points he is missing 20:10
pyrimidine yup. he likes strawmen
jnthn back from le dinner 20:11
moritz jnthn: was it French? :-) 20:12
jnthn No, I just wanted to sound classy and stuff. 20:13
dalek ecza: b1a1194 | (Solomon Foster)++ | lib/CORE.setting:
Port KeyBag improvements back to Bag.
 20:20
ast: 8d18a30 | (Solomon Foster)++ | S02-types/bag.t:
70 more tests for Bag.
 20:21
tadzik wonders if *Bags are Rakudo-stealable 20:23
jnthn Should be fairly possible :) 20:24
PerlJam wonders what it would take to silence the perl6-grouches of the world.
masak I hereby dub the person in any post about Perl 6 who writes the inevitable comment saying "Perl 6 is failing/a farce/irrelevant/worthless", or something to that effect, causing deeply nested comments from insiders and outsiders alike that never lead anywhere, "the sponge". 20:25
please consider carefully how much time you want to waste on the sponge. ;) 20:26
jnthn Is that the sponge I wash up with, or sponge pudding?
PerlJam jnthn: it's a dirty, nasty used sponge with bacteria oozing from it's insides. 20:27
jnthn Eww. OK, 'cus wasting time on sponge pudding woulda been tempting. :)
PerlJam masak: I wasted just enough time to read the thread(s), but that was time I was going to waste doing something frivolous and non-coding anyway 20:28
masak nodnod
well, the time that sponges mostly suck up is if you enter a conversation with them.
PerlJam starts down voting moritz's node because of the sponge he helped inflict upon us ;) 20:30
s/node/nodes/
tadzik colomon: why is %elems in github.com/sorear/niecza/blob/b1a1...ting#L2128 Bool? 20:32
Su-Shee PerlJam: it would take a couple of cool projects out there for everybody to use. some must-haves. 20:33
tadzik without the constraint the implementation passes tests on Rakudo
colomon tadzik: because I copied and pasted without noticing that, and apparently niecza doesn't check
tadzik++
tadzik :)
applies for a nieczabug I suppose 20:34
jnthn tadzik++
masak this is why I don't type my attributes. :)
tadzik if anyone wants an LHF, gist.github.com/1782889 passes a similar set of tests that niecza does, on the first glance 20:35
I should probably be learning for my exam instead :)
masak nom: class A { has Int @.a }; my $a = A.new(:a(1, 2, "OH NOES")); say "alive"
p6eval nom 828e4d: OUTPUT«alive␤»
tadzik works for hashes
jnthn Works for individual array alements
But not assignment.
masak I don't like to write code that could fail in the future when the implementation gets better.
moritz: the technical comment about why Rat is slow is great. the personal comment about the 'educated_' prefix is not over-the-top in any way, and it's your pen, but... it's feeding the sponge, and nothing more. 20:39
dalek ecza: 760a302 | (Solomon Foster)++ | lib/CORE.setting:
Fix embarrassing mistake noticed by tadzik++.
 20:40
tadzik that may have been my first niecza contribution of some sort :) 20:41
masak moritz: if nothing else, it's evident from the followups that those two comments led to. the one with the high blood pressure in it even attracted chromatic. 20:43
tadzik are we talking about some perlmonks discussion?
masak yeah.
colomon geez, my head is so fuzzy I just completely botched a simple Thomas the Tank engine layout.
pyrimidine tadzik: perlmonks.org/index.pl?node_id=952765 20:44
colomon is probably going to regret looking...
masak nom: .say for "", { $_ ~ "potato".substr(.chars, 1) } ... "potato" 20:47
p6eval nom 828e4d: OUTPUT«␤p␤po␤pot␤pota␤potat␤potato␤»
masak Rakudo can count to potato :D
nom: .say for "", { $_ ~ "potato".substr(.chars, 1) // last } ... * 20:48
p6eval nom 828e4d: OUTPUT«(timeout)␤p␤po␤pot␤pota␤potat␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤…
masak nom: .say for "", { $_ ~ ("potato".substr(.chars, 1) // last) } ... *
hm.
p6eval nom 828e4d: OUTPUT«(timeout)␤p␤po␤pot␤pota␤potat␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤potato␤…
masak .oO( "I know how to count to potato, but I don't know when to stop..." ) 20:49
tadzik banananananananana
masak nom: say "ba" ~ "na" x (2..20).roll
p6eval nom 828e4d: OUTPUT«banananananananananana␤»
20:55 pernatiy left 20:56 pernatiy joined 20:57 bluescreen10 left 20:58 birdwindupbird left 21:04 lestrrat left 21:05 lestrrat joined 21:08 GlitchMr left
skids specwise, if two roles/classes with a same-named attribute are composed, does the overridden one still exist enough to, well I guess, "take up space?" 21:09
jnthn If you compose two roles with the same named attribute, it's a conflict. I'm not sure what you mean about composing two classes. 21:10
skids OK good. 21:14
21:15 bluescreen10 joined
masak perl6: role A { has $!x }; role B { has $!x }; class C does A does B {}; say "alive" 21:16
p6eval rakudo 828e4d: OUTPUT«===SORRY!===␤Attribute '$!x' conflicts in role composition␤»
..pugs b927740: OUTPUT«alive␤»
..niecza v14-32-g760a302: OUTPUT«===SORRY!===␤␤Attribute '$!x' is provided by both 'A[curried][...]' and 'B[curried][...]' at /tmp/fnAJyZgvin line 1:␤------> B { has $!x }; class C does A does B {}⏏; say "alive"␤␤Unhandled exception: Check failed␤␤ a…
masak Niecza likes curry.
jnthn LTA that Rakudo doesn't tell you which roles.
Could optimize the currying out in that case. 21:17
21:27 fsergot left
skids nom: class A { has $!a; }; class B is A { has $!a }; my B $a .= new(); $a.HOW.attributes($a).say 21:28
p6eval nom 828e4d: OUTPUT«$!a $!a␤»
colomon tadzik: I should warn you that I plan significant changes still for Set and KeySet. 21:32
tadzik okay :) 21:33
I'm not in a hurry
the more situations like this appear the more I think about something along STD.pm6, but for CORE.setting
colomon you mean a standard setting for more than one implementation? 21:35
tadzik aye
look at what we're doing now
copying This and That from Rakudo to Niecza, then other This and That from Niecza to Rakudo
colomon true
but usually there are at least minor changes when we do that
tadzik now it was adding : before %elems in BUILDs, and changing the %e{$a}:delete expressions to the method call equivalents 21:37
dalek p/bs: 44d674d | jnthn++ | src/6model/serialization. (2 files):
First crack at one-level-deep closure serialization; survives, though not done deserialization yet.
 21:40
21:48 kaare_ left
[Coke] I think that's a very reasonable goal to have that shared, but I think we're going to get there organically, rather than by attempting to force it. 21:51
21:53 MayDaniel left 21:54 Chillance left
colomon what [Coke] said. 21:54
[Coke] ponders graduate studies. 21:57
and perl6.
22:07 skids left 22:12 preflex_ joined, preflex left
colomon moritz: so, is that prisoner's dilemma harness *the* harness for the competition, or not? 22:12
22:12 preflex_ is now known as preflex
cognominal___ jonathan: you meant "le dîner"? :) 22:22
[Coke] that question makes me think you're trying to game the system, sirrah!
22:22 benabik left
cognominal___ or may be "le diner". There has been a spelling reform that is ignored both by the litterate or the sms generation, that invents its own language. 22:24
masak [Coke]: no, it makes sense to know how the competition will be judged. 22:26
cognominal___ it is funny to see that the spelling of words like jeanfoutre (jerker) is regulated by law : www.academie-francaise.fr/langue/or...phies.html
someday, we will have legal trouble in France, if we try to verb names. :) 22:28
Apparently, with antidote, the software, one can chose the spelling, mordern, ancient or both 22:31
indeed, the new spelling drop the accent. 22:32
*modern *drops
antidote++
www.druide.com/antidote.html
*literate (but from litteratus) 22:34
masak cognominal___: I would be sorry to see the accents go. I like most of them. 22:46
jnthn tries not to make a joke about ze French accent 22:48
cognominal___ the new rule specifies dropping or changing accents where French people are confused to start with so in practice it does not change anything. 22:49
masak right. people who were confused prior to the reform will remain so :)
22:49 yertalert left, nimiezko joined
masak rule of thumb: don't reform for the confused. :) 22:49
bonne nuit, #perl6 22:50
cognominal___ but I remember there have been an outcry when some conporary book recent publication has been rectified.
jnthn branoc, masak
nimiezko rakudo: qx/ls/ 22:51
p6eval rakudo 828e4d: OUTPUT«qx, qqx is disallowed in safe mode␤ in sub restricted at src/SAFE.setting:2␤ in sub QX at src/SAFE.setting:9␤␤»
nimiezko rakudo: q:x/ls/
p6eval rakudo 828e4d: OUTPUT«===SORRY!===␤Colons may not be used to delimit quoting constructs at line 1, near ":x/ls/"␤»
22:52 snearch left
cognominal___ that's very French to publish laws that are totally ignored. I think one third of them don't have a application decret signed by the president meaning that le legislative process is ignored by him even when it has been intiated by him at a time it was mediatically opportune to do so. 22:52
s/le/the/
nimiezko nieczca: qx/ls/ 22:53
What is the status on quoting forms (Q with adverbs) from S02 in rakudo ? 22:54
cognominal___ aren't they macros? 22:55
22:55 cognominal___ is now known as cognominal
jnthn nimiezko: The general Q with adverbs form isn't implemented in Rakudo yet, as far as I'm aware. 22:55
22:56 REPLeffect joined
jnthn Don't think we're missing anything especially major to implement it, just didn't happen yet. 22:56
nimiezko so only abreviated forms works for now ? 22:57
jnthn Correct.
nimiezko thanks a lot
23:09 Gothmog_ left, chra left, sudokode left, awwaiid left, y3llow left, Gothmog_ joined, awwaiid joined, chra joined, sudokode joined, y3llow joined 23:11 y3llow left 23:12 y3llow joined 23:13 y3llow left 23:14 y3llow joined 23:16 y3llow left, y3llow joined
dalek p/bs: 0505b08 | jnthn++ | src/6model/serialization.c:
Don't attempt to serialize an outer that'll simply never work out.
 23:16
p/bs: cf55512 | jnthn++ | src/6model/serialization.c:
Start to sketch out overall closure deserialization process.
p/bs: 3b9c867 | jnthn++ | src/6model/serialization.c:
First crack at closure deserialization. Passes one of the tests, fails another...something still a bit off.
p/bs: 84c2256 | jnthn++ | src/6model/serialization.c:
Fix a couple of silly thinkos; basic closure serialization tests now pass.
23:17 leontopod joined
leontopod If you guys can come out with Perl 6.0.0, I will start coding in it. =) 23:17
cognominal we don't believe in numerology. :) 23:18
leontopod you know, Perl 6, Version 1 23:19
tadzik leontopod: what is that you don't like about the current version? 23:20
leontopod I don't think it is ready to produce production code yet. 23:22
Is anybody using Perl 6 to do anything Real? 23:23
tadzik masak does
leontopod Like, avionics, or controlling a nuclear power plant
I know, I talked to masak.
He nice guy! =)
I promised him I wouldn't troll, so I better stop.
flussence "$language doesn't do $completely-unrealistic-requirement, $language sucks!" 23:26
so what language is your nuclear reactor and/or air control tower currently using?
timotimo cobol? 23:27
hopefully ada or something similar
tadzik cobol I guess 23:32
23:37 whiteknight joined
timotimo did you come here to protect poor perl6-in-distress, whiteknight? :) 23:38
whiteknight timotimo: is perl6 in distress?
I came online to look up funny pictures of cats, but I'll help if help is needed
timotimo a troll just came here who's all like "but i cannot use perl6 in a nuclear power plant yet! so clearly it's not ready to be used!" 23:39
but i was mostly kidding, as was he
whiteknight oh, then I won't say anything mean about him :)
tadzik oh, I wouldn't exaggarate and call that trolling
it's more like the usual "when will you hit 1.0"
23:40 damsel_in_distre joined, damsel_in_distre is now known as DamselInDistress, nimiezko left
DamselInDistress Help me, whiteknight, I need to learn perl6 to please a nice guy who could probably be more than just a friend, if you catch my drift ^^ but I need to learn perl6 to help him code a server for a supercool RPGgame that I need to write ;-) it has everyting a traditional RPG has and will also have awesome features and allow much better rpggeneration. Canyouhelpmeplz? ^^'; 23:44
flussence
.oO( didn't we hit 1.0 some time around 1985? )
whiteknight DamselInDistress: ha, that's the coolest request for help I've ever heard. Unfortunately, I'm probably not best for the job
my knowledge of P6 is not great
DamselInDistress shucks 23:45
23:45 DamselInDistress is now known as lisa_schwarzhaar
lisa_schwarzhaar perl6 sounds cool though, I shall hang out in here for a while 23:45
leontopod flussence, probably ADA
timotimo is it ADA, not Ada? 23:46
leontopod I know a lot of air traffic control systems and even some avionics use Jovial
tadzik lisa_schwarzhaar: what do you need help with?
leontopod which is a really really old multithreaded/tasking version of Algol
tadzik lisa_schwarzhaar: do you have previous Perl/programming experience?
whiteknight lisa_schwarzhaar: now, if you want to learn C, and have some extra hair that you don't mind ripping out, I can help :)
lisa_schwarzhaar whiteknight, tadzik I can code c/c++ already and my hair is sacred 23:47
As for learning perl6, I guess I'll read a book / the internet and nag some of the local nerds
tadzik and is schwarz
whiteknight timotimo: Probably Ada. It's named in honor of Ada Lovelace
tadzik lisa_schwarzhaar: try the Using Perl 6 Book
timotimo yup, wikipedia claims so as well
whiteknight lisa_schwarzhaar: excellent strategy
lisa_schwarzhaar tadzik: sounds good
tadzik lisa_schwarzhaar: perl6.org/documentation/ generally 23:48
lisa_schwarzhaar I can still nag you with general questions, right?
23:48 nimiezko joined
tadzik sure thing 23:48
timotimo "so, how do i computer?", lisa_schwarzhaar?
whiteknight I can program some NQP, which is not quite what you want :)
timotimo what exactly is nqp? it seems to have something to do with native calls and belongs to Parrot VM? 23:49
lisa_schwarzhaar timotimo: Like I told you before, you gotta make sure your computer is feeling well. Treat it / him / her to some nice juicy electricity and generally treat it well...
tadzik timotimo: it's a subset of Perl 6 in which the core compiler is written
flussence it's perl6 without any syntactic sugar at all
timotimo oh, interesting/terrible
whiteknight timotimo: yeah, NQP is a subset of perl used for bootstrapping the compiler 23:50
so Rakudo perl6 is written in perl6
timotimo cool, i totally didn't know that
jnthn NQP is also self-hosting. :)
leontopod Rakudo runs Parrot Assembly Language I thought 23:51
timotimo so it's a little bit like pypy in that regard?
wait. it's a subset of perl? or perl6?
jnthn Subset of Perl 6.
whiteknight jnthn: yes, it's turtles all the way down
Until you finally get down to a single exhausted bit, flipping on and off extremely fast 23:52
timotimo until you reach elephants?
leontopod you know what was cool about the Atari ST, is that video memory was just part of main memory, and you had the ability with a small C program to reset the video pointer to anywhere in memory you wanted 23:53
so if you set it to low memory
each pixel represents one bit
so you can see counters counting up, bits flipping on and off, chunks of disk being read in 23:54
it was pretty neat!
oh right, and what made this all possible is the video chip ran asynchronously to the main cpu (68000) chip
you can't do that easily with 80x86 architectures because video memory is separate from main memory
lisa_schwarzhaar smiles happily while daydreaming of better times 23:55
leontopod really it hearkened back to the days when computers had lights on the console, and gleaming rows of toggle switches!
Toggle switches so you could toggle in bootstrap code. 23:56
timotimo those were better times
how did we get to this topic?
lisa_schwarzhaar I heard they had hybrid computers too.. I even got to touch one..
leontopod yes, the computing power in an android phone used to take up a large room in the 1960s 23:57
I programmed on some of those beasts.
timotimo uh
an android phone would take more than just one large room
leontopod well, ok!
=)
an entire fricking building!
timotimo those phones usually host dual-core processors and sometimes dedicated graphics processors? not sure.
lisa_schwarzhaar just clean out all the bloat and it would run on a decent old computer... 23:58
timotimo and then there's the GSM chip that runs its own realtime-OS
leontopod I am seriously looking into the android scripting environment
timotimo i would suppose you couldn't even get to the performance due to bandwidth limitations in the whole thing and the problem of parallelizing things
leontopod I'd like to write an android app in Perl
timotimo you mean SL4A? 23:59
leontopod I don't think it is QUITE there yet

Powered by Cro and the Raku® Programming Language.

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

Thank you!