»ö« | perl6.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot usage: 'perl6: say 3;' or rakudo:, alpha:, pugs:, std:, or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by lichtkind on 5 March 2010.
00:00 masak left, masak joined, hatsefla1s joined
masak blr, wobbly neighbournet... 00:00
lue blr? 00:01
00:01 hatsefla1s is now known as hatseflats
masak as in, 'yuck'. :) 00:01
lue does the tester have a built in ETOOLONG type situation? 00:02
masak no.
you'll have to ^C
lue will now hunt w/ masak for bug 00:03
masak might have just found it :/
yup.
it was yet another instance of RT #73034. 00:04
should have seen that one coming...
pushed. try pulling, rebuilding, and running the tests file again. 00:06
lue or just that one test. 00:07
masak yeah.
I tried that here, and it doesn't hang anymore.
now running all tests, to see what's the next interesting failure. 00:08
lue got a No Match. At least it didn't hang!
masak the `./test-regex 'a**bc' ''` should indeed give 'No match'. 00:09
it's trying to do '' ~~ /a ** b c/
00:09 justatheory left 00:11 dual left 00:13 dual joined 00:15 ash_ left
lue how's it going masak? 00:16
masak running the tests :) still not past 414. just passed 200, though.
whoa! Segmentation fault 00:17
that's been a while.
diakopter ltns, sigsegv
lue GPF (can you guess what it means _here_?) 00:18
masak the regex in question runs fine in isolation, so the segfault is due to general Parrot weariness.
masak tries the test suite again
getting to be time for me to hit the pods, too. 00:19
lue You have an ex-parrot!?
masak seems I do.
but on average, Parrot is much less dead nowadays than a year or two ago. 00:20
lue Then 2 things are true about you. 00:21
00:21 dual_ joined
lue 1) You don't have a Norweigan Blue 00:22
2) You didn't get it in Bolton
masak I find no fault in that logic.
lue Corrolary to 2) You didn't get it in Notlob 00:24
00:24 dual left
lue .mmorpg(The expansive bed chamber of RD* is ready to receive occupants) 00:24
masak yawns 00:25
just going to see if I can reproduce this segfault, and then go to bed.
if it persists, probably the way to run the other tests is to change the order in which the files are run. 00:26
lue So, today on my P6 checklist is a) begin a P6 created general-purpose utility for dealing with DNA/RNA 00:27
b)design the layout of the rakudo death star
masak 'dealing with'?
do you mean handling transcription/translation and their inverse operations?
lue "decoding/encoding" it, analysis, maybe even a Life Simulator (in the distant future of course) 00:28
masak hm, translation is proteins.
Tene 'night masak
masak Tene: 'night. if you're here during the weekend, I'll be happy to meet to discuss the Last Mile of Web.pm.
lue It's scary how DNA, out of all organic matter and processes, relates to programming. (IMO)
masak aye. DNA is digital.
Tene lue: www.guardian.co.uk/science/2010/may...-life-form 00:29
00:29 dual_ left
masak huh. this time, it didn't segfault. that's both good and bad news. 00:29
Tene masak: Yes, that would be great.
lue Tene: I know. It kills me
The Dark Side now can patent the Creation of Life 00:30
masak don't worry. I'm sure there's Prior Art.
lue c) Profit!
masak: the USPTO tends to patent discoveries. I'm scared.
Although, he does prove now that I could throw nucleotides into the proverbial blender and by chance make a TARDIS :D 00:32
masak it is in times like this that I would advice not forgetting to breathe properly.
lue *gurk* 00:33
00:33 dual joined
masak throwing nucleotides into the proverbial blender is half of how evolution works, so I can't claim the idea doesn't have some merit... 00:33
lue or, if you want to have a chance, change the Hox Genes, and change the proteins for brain-building into producing cells of Si (yada yada yada) 00:35
masak yay, the tests now run past 415. \o/ 00:38
lue \o/
masak test 440 fails. 00:40
that'll be tomorrows riddle to crack.
s/ws/w's/ 00:41
lue goodnight, masak 00:42
masak 'night
00:42 masak left 00:43 dual_ joined 00:45 dual left 00:48 dual_ left 00:49 alanhaggai left 00:51 dual joined 00:57 dual left 00:58 huf_ is now known as huf 00:59 justatheory joined
BrowserUk ? 01:07
01:07 BrowserUk left
lue ?? 01:10
01:17 justatheory left 01:19 rgrau` joined
pmichaud ?! 01:19
01:19 rgrau left 01:21 Patterner left, Psyche^ joined, Psyche^ is now known as Patterner
TimToady . 01:22
Trashlord yeah 01:24
lue hears crickets 01:29
01:30 plobsing joined, am0c left 01:40 seatek joined 01:48 eternaleye joined
colomon ?? !! 01:54
01:57 wknight8111 left 01:58 [mark] joined 01:59 meppl left
diakopter : 02:05
lue ~~ 02:10
rakudo: multi sub infix:<^.^>($a, $b){say "$a $b happy";}; 3 ^.^ 4
p6eval rakudo 6a7ded: OUTPUT«3 4 happy␤» 02:11
lue .u dna 02:15
phenny lue: Sorry, no results for 'dna'.
lue .u helix
phenny lue: Sorry, no results for 'helix'.
lue Unicode gives me an entire deck of cards, but no DNA symbol? Weird... 02:16
02:18 Exodist joined
diakopter .u pair 02:20
phenny U+21C8 UPWARDS PAIRED ARROWS (⇈)
diakopter .u diam 02:23
phenny U+2300 DIAMETER SIGN (⌀)
diakopter .u diamo
phenny U+25C6 BLACK DIAMOND (◆)
lue .u ace 02:24
phenny lue: Sorry, no results for 'ace'.
lue .u card
phenny lue: Sorry, no results for 'card'.
isBEKaml .u double 02:43
phenny U+29FA DOUBLE PLUS (⧺)
isBEKaml .u urn
phenny U+26B1 FUNERAL URN (⚱)
02:44 snarkyboojum joined
isBEKaml .u symbol 02:44
phenny U+2136 BET SYMBOL (ℶ)
isBEKaml .u beta
phenny U+03D0 GREEK BETA SYMBOL (ϐ)
isBEKaml .u bet
02:44 am0c joined
phenny U+2136 BET SYMBOL (ℶ) 02:44
02:48 alester joined 02:57 k23z__ left 02:59 jjore left 03:00 gbacon joined
dalek psi: 1e6efbb | masak++ | .gitignore:
[.gitignore] added
 03:09
yapsi: 6a5a68e | masak++ | (2 files):
yapsi: [Yapsi] blocks now return
03:09 dalek left 03:10 dalek joined
dalek november: d6cd2f8 | (Jason Felds)++ | skins/CleanAndSoft/register (3 files): 03:10
november: Start of registration pages. Copied from login.
03:10 dalek left 03:11 dalek joined
lue o no not again 03:11
dalek ok: 0bc215d | (Hinrik Örn Sigurðsson)++ | (4 files):
Fixed problem with -l/--locate, added test for it
 03:13
ok: c06849e | (Hinrik Örn Sigurðsson)++ | (4 files):
Look up the man pages. Depend on Perl6::Doc 0.42
ok: 8cf5c68 | (Hinrik Örn Sigurðsson)++ | (9 files):
Bump version to 0.16_02
ok: 3454ac2 | (Hinrik Örn Sigurðsson)++ | Changes:
Make the Changes file historically accurate
grok: 056121c | (Hinrik Örn Sigurðsson)++ | (4 files):
grok: Small Pod fixes
03:14 Chillance left
dalek ok: 8e36bf7 | moritz++ | src/classes-and-objects.pod:
grammar fix pointed out by diakopter++
 03:14
03:15 JimmyZ joined
lue Are we supposed to get grok? 03:15
dalek vember: 1c9a598 | (Jason Felds)++ | lib/November.pm:
Quickie one-liner from login useful in register.
 03:17
03:24 stephenlb left
pugssvn r30746 | svatsan++ | [u4x] render.pl - removed unnecessary module and minor cleanups 03:33
isBEKaml rakudo: my $sq = -> $x { $x * $x }; say $sq 22; 03:50
p6eval rakudo 6a7ded: OUTPUT«Confused at line 11, near "say $sq 22"␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
isBEKaml rakudo: my $sq = -> $x { $x * $x }; say $sq(22);
p6eval rakudo 6a7ded: OUTPUT«484␤»
isBEKaml rakudo: my $sq = -> $x { $x * $x }; say $sq(1..10); 04:01
p6eval rakudo 6a7ded: OUTPUT«100␤»
isBEKaml rakudo: my $sq = -> $x { $x * $x }; say $sq(1..^10);
p6eval rakudo 6a7ded: OUTPUT«81␤»
isBEKaml rakudo: my $sq = -> $x { $x * $x }; say $sq(^10);
p6eval rakudo 6a7ded: OUTPUT«100␤»
04:05 [particle] left 04:41 envi^home joined 04:52 alester left 05:11 am0c left 05:15 justatheory joined 05:36 am0c joined 05:37 [particle] joined 05:38 eternaleye left 05:40 lue left 05:41 lue joined 05:47 eternaleye joined 05:55 gbacon left 05:58 am0c left 05:59 Juerd left, Juerd joined 06:05 jakk joined, Targhan left 06:07 jakk left 06:33 justatheory left 06:43 TiMBuS joined
eternaleye (backlogging) bbkr, might Quill be parsing the Q as the quote meta? 06:56
06:59 Su-Shee joined
TiMBuS how do you check match results in current rakudo head? using $0 will fail if there was no match. 07:37
moritz_ if $/ { }
TiMBuS that makes sense
07:37 iblechbot joined
TiMBuS is it supposed to fail or is it just not implemented correctly yet? 07:38
moritz_ NYI
TiMBuS that's ok then. thanks
pugssvn r30747 | moritz++ | [S03] s/shaped/nested/ when talking about auto-recursive hyper ops 07:41
dalek kudo: 9a15b82 | moritz++ | src/old/ (87 files):
remove src/old/.
 07:43
isBEKaml [OT]: www.ibm.com/developerworks/library/...;S_CMP=HP. This is the most reasonable argument in favour of opensource I've ever read. :) 07:44
moritz_ rakudo: say $/.Bool 07:47
p6eval rakudo 6a7ded: OUTPUT«0␤»
moritz_ rakudo: $0 07:49
p6eval rakudo 6a7ded: OUTPUT«Method 'postcircumfix:<[ ]>' not found for invocant of class ''␤current instr.: '!postcircumfix:<[ ]>' pc 11523 (src/builtins/Any.pir:54)␤»
moritz_ rakudo: say $/.PARROT 07:50
p6eval rakudo 6a7ded: OUTPUT«␤»
moritz_ rakudo: say $/.perl
p6eval rakudo 6a7ded: OUTPUT«Any␤»
07:50 cls_bsd left, cls_bsd joined
moritz_ rakudo: say Any.Seq 07:50
p6eval rakudo 6a7ded: OUTPUT«Method 'Seq' not found for invocant of class ''␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
moritz_ rakudo: say Any.list
p6eval rakudo 6a7ded: OUTPUT«Any()␤»
moritz_ rakudo: say Any.list.perl 07:51
p6eval rakudo 6a7ded: OUTPUT«(Any, )␤»
07:51 cls_bsd left 07:55 cls_bsd joined
pugssvn r30748 | moritz++ | [u4x/web] fix a syntax error 07:55
07:56 cls_bsd left 08:00 cls_bsd joined 08:01 cls_bsd left 08:05 cls_bsd joined
moritz_ I havve a patch for $0... that now produces a different error :/ 08:05
rakudo: $/.list.[0]
p6eval rakudo 6a7ded: ( no output )
moritz_ rakudo: say $/.list.[0]
p6eval rakudo 6a7ded: OUTPUT«Any()␤» 08:06
08:06 cls_bsd left 08:10 cls_bsd joined 08:11 cls_bsd left
isBEKaml oops... :(, isBEKaml-- 08:12
moritz_ no problem
08:15 cls_bsd joined
moritz_ somebody please confirm my sanity: 3[0] should be 3, 3[1] should be Nil, right? 08:19
I'm about to implement that in Rakudo, and get lots of tests failing that seem to indicate otherwise
m6locks rakudo: say 3[1]
p6eval rakudo 6a7ded: OUTPUT«get_pmc_keyed() not implemented in class 'Integer'␤current instr.: 'perl6;Positional[::T];postcircumfix:<[ ]>' pc 11343 (src/gen/RoleToInstanceApplier.pir:170)␤»
08:22 nnhi joined
moritz_ ah well, my patch produces Any instead of Nil 08:24
still, it's a start 08:25
08:26 nnhi left
m6locks :3 08:29
spinclad i wouldn't expect 3 to have Positional parts (or Associative for that matter) 08:33
moritz_: ^ 08:34
moritz_ spinclad: well, it's TimToady's decree that everything should be usable like a List
spinclad right. singleton lists == their singlething. well, it doesn't DW_I_M, but it is well entrenched by now. 08:37
moritz_ I must confess I also feel a bit uneasy at the notion
spinclad so yes, 3[1] would be Nil to mark EOL. 08:38
(or is that only if you actually iterate it? dang, i'll have synopsize on it again.) 08:39
moritz_ problem is, it's probably not in any synopsis, but an uttering here in IRC
spinclad rakudo: say [3, Nil].perl; say [3, Nil, 2].perl
p6eval rakudo 6a7ded: OUTPUT«[3]␤[3, 2]␤»
spinclad rakudo: say [3, Nil][1].perl 08:40
p6eval rakudo 6a7ded: OUTPUT«Proxy.new()␤»
spinclad the .[1] element is off the end, so this Proxy (or what it turns into soon, iirc) makes the most sense to me... 08:42
moritz_ aye
uohm
my $thing = { a => 1} 08:43
say $thing[0]
what should that produce?
spinclad rakudo: say (3, Nil).perl
moritz_ a Pair?
p6eval rakudo 9a15b8: OUTPUT«(3, ())␤»
spinclad rakudo: say (3, Nil)[1].perl
p6eval rakudo 9a15b8: OUTPUT«()␤»
spinclad rakudo: say (3, Nil)[2].perl
p6eval rakudo 9a15b8: OUTPUT«Any␤»
spinclad (EOL is called Empty, i think?) 08:44
rakudo: my $thing = { a => 1 }; say $thing[0] 08:45
p6eval rakudo 9a15b8: OUTPUT«Method 'postcircumfix:<[ ]>' not found for invocant of class ''␤current instr.: '!postcircumfix:<[ ]>' pc 11523 (src/builtins/Any.pir:54)␤»
spinclad rakudo: my $thing = { a => 1 }; say $thing.pairs.[0]
p6eval rakudo 9a15b8: OUTPUT«a 1␤»
spinclad i like rakudo's present point of view here (the error message could get hoisted to Perl 6 terms instead of Parrot, but that's minor) 08:48
rakudo: for (3, Nil) { .say } 08:49
p6eval rakudo 9a15b8: OUTPUT«3␤»
spinclad rakudo: for (3, Nil, 2) { .say }
p6eval rakudo 9a15b8: OUTPUT«3␤2␤»
08:51 dalek left
spinclad -> bed o/ 08:52
08:52 meppl joined 08:54 dalek joined 09:04 silug left 09:25 kel__ joined 09:29 Jedai joined 09:48 wknight8111 joined, ethel left 09:59 agentzh joined 10:21 masak joined
masak it's weekend! \o/ 10:22
m6locks party tiem 10:24
colomon o/ 10:25
masak celebrates by finding the next bug in GGE
10:41 rgrau_ left 10:46 molaf joined 10:51 BrowserUk joined
BrowserUk ? 10:54
moritz_ gives match objects another shot 10:55
time to rebase another branch... 10:56
when I have a long-lived topic branch, should I merge from master from time to time?
currently I rebase, and then use a new branch name 10:57
which is why there are mob2, mob3 and mob4 on github...
... now only mob4 :-)
\o/ suddenly stuff works 10:58
the 'mob5' branch handles named captures without evil workarounds 10:59
... as long as they are not quantified :/
masak there's not much of 'should' in the possible git workflows themselves. 11:00
and I'm not sure I know how to argue for one model or the other.
moritz_ is there something like "best practise"
if somebody else tracks my branches, giving them new names every few weeks is certainly confusing
masak apart from "don't push/rebase/push", I don't know. 11:01
maybe ask on #git?
11:01 gbacon joined
moritz_ maybe googling a bit first :-) 11:01
masak does that
'Rebasing and mergin: some git best practices' lwn.net/Articles/328436/ 11:02
11:03 rgrau joined
masak 'A successful Git branching model' nvie.com/git-model 11:03
11:04 salmonix_ joined
moritz_ hates the "Method '$meth' not found for invocant of class ''" errors 11:07
11:08 agentzh left
masak they have something to do with the type objects. 11:08
moritz_ rakudo: say Array.new.perl
p6eval rakudo 9a15b8: OUTPUT«[]␤»
masak I'm currently trying to squash a bug in which an array of Ints suddenly magically contains a GGE::Match object. 11:09
11:10 salmonix_ left
masak I find the whole thing very worrying. 11:10
moritz_ understandable
colomon I'm currently trying to do a lexicographical compare on two arrays of numbers and failing badly. 11:11
oooo.... just occurred to me that I should probably prototype in normal code, considering how "say" no longer works in the setting...
moritz_ what's so hard about it?
rakudo: say [1, 2, 3] cmp [1, 0, 3]
p6eval rakudo 9a15b8: OUTPUT«1␤» 11:12
moritz_ rakudo: say [1, 2, 3] cmp [1, 4, 3]
p6eval rakudo 9a15b8: OUTPUT«-1␤»
colomon moritz_: I didn't think there was anything hard, but my code is failing badly.
moritz_ colomon: why not juse use cmp?
colomon orly?
no one ever told me cmp did that?
moritz_ rakudo: say [1, 2] cmp [1, 2, 3]
p6eval rakudo 9a15b8: OUTPUT«-1␤»
moritz_ rakudo: say [1, 2, 3] cmp [1, 2]
colomon are you sure it does, actually?
p6eval rakudo 9a15b8: OUTPUT«1␤»
Su-Shee I don't understand the output.
colomon rakudo: say [111, 2] cmp [2, 3, 4] 11:13
p6eval rakudo 9a15b8: OUTPUT«-1␤»
11:13 silug joined
moritz_ rakudo: say 111 cmp 2 11:13
p6eval rakudo 9a15b8: OUTPUT«1␤»
moritz_ rakudo: say 1 cmp 2
p6eval rakudo 9a15b8: OUTPUT«-1␤»
moritz_ cmp needs fixing
colomon moritz_: I don't think cmp does what you think it does....
moritz_ it should compare numbers numerically
colomon moritz_: it does
moritz_ Su-Shee: which part don't you understand?
colomon but it compares arrays by string. 11:14
moritz_ oh.
should it?
Su-Shee moritz_: does it return yes/no or the position of what's different or how many elements are the same or not..?
colomon well, you just now is the first I ever heard that it should compare arrays lexicographically.
but that doesn't mean you are wrong. 11:15
moritz_ Su-Shee: it's three-way comparsion
11:15 slavik joined
moritz_ Su-Shee: and returns smaller, same or greater 11:15
Su-Shee oh, I thought it compares element by element.
moritz_ rakudo: say (1, 2, 3) »cmp«(1, 0, 5) 11:16
p6eval rakudo 9a15b8: OUTPUT«01-1␤»
colomon moritz_: I don't see anything in the spec about cmp working intelligently on lists/arrays. 11:17
but I may be missing something.
moritz_ oh, it might be very well that I'm wrong, or the spec is incomplete, or both
TimToady: any thoughts? 11:18
quick question on src/core/Array.pm lines 20 to 23... why does it fill up to $i+1, not up to $i? 11:20
afk, lunch etc (will backlog) 11:21
colomon moritz_: I think because it doesn't matter if you fill extra?
which is a non-answer answer....
but then, I thought that code wasn't currently being used, and is scheduled to be rewritten. 11:22
11:25 FreeStorm joined
colomon moritz_: turns out the correct answer is that my lexicographical code was working fine, but its support code was broken. 11:28
11:29 wknight8111 left
colomon rakudo: say [!!] 0, 0, -1, 1, 0 11:31
p6eval rakudo 9a15b8: OUTPUT«Unable to parse postcircumfix:sym<[ ]>, couldn't find final ']' at line 11␤current instr.: 'perl6;Regex;Cursor;FAILGOAL' pc 1696 (ext/nqp-rx/src/stage0/Regex-s0.pir:932)␤»
colomon rakudo: say [||] 0, 0, -1, 1, 0
whoops
p6eval rakudo 9a15b8: OUTPUT«-1␤»
colomon ack, I love p6!
rakudo: say [||] 0, 0, -1, -3, 255 11:32
p6eval rakudo 9a15b8: OUTPUT«-1␤»
11:38 [Coke] left 11:42 BrowserUk left
dalek meta: r293 | pawelmurias++ | trunk/Sprixel/t/parse_bug2.t:
test for a parsing bug
 11:45
11:46 [mark] left
isBEKaml I don't understand self!fill in Array.pm - where's it defined? 11:46
is that a private sub in there? 11:47
colomon isBEKaml: it's defined in Array.pir, I believe. Or possibly Seq.pir
11:48 pmurias joined
isBEKaml colomon: I'm looking at Seq.pir, since I don't find it in Array.pir 11:48
pmurias diakopter: hi
colomon isBEKaml: Seq.pir, line 118. (ack is your friend. ;) 11:49
11:49 [Coke] joined
isBEKaml colomon: aye, sir. but I don't have ack, I use find and fgrep. :) 11:50
colomon: the problem is, even if I do find the sub/method, I won't know how that's related to what I'm looking for. For instance, I didn't know that Array is a Sequence. :) 11:56
colomon :)
pmurias masak: looking at your slides do you claim that if (my $foo = 42) {...} is an example of a bug sigmund finds 11:57
colomon isBEKaml: I believe Seq is on the chopping block at the moment, too.
pmurias masak: if $a = 42 {...}
isBEKaml colomon: I don't understand. Chopping block? You mean, it will be dropped? 11:58
masak pmurias: yes. it'd probably not merit a note on the error level, but on the warning level.
because in some rare cases it might be what the programmer intended. 11:59
colomon isBEKaml: there is supposed to be a major rewrite of list/Seq/Array in the very near future, and it seems most of us hate Seq.
masak colomon: what are the reasons for hating Seq? I don't think I've had much exposure to it. 12:00
but I definitely think there is an undue proliferation of almost-similar types in that area.
isBEKaml colomon: Oh, I see...
colomon masak: give me a moment, trying to finish a blog post before I have to take on the morning duties of a dad. 12:03
masak oh, no rush.
mathw Hey
pmurias masak: if my $a = 42 {...} is a thing i use regularly 12:06
isBEKaml masak: I have made that mistake before and I force myself a re-look every time I write code. I still make mistakes. 12:08
masak pmurias: even with the RHS being a literal?
isBEKaml: me too.
pmurias masak: literal no
masak pmurias: that's mainly the case I'm talking about. 12:09
isBEKaml masak: I keep forgetting that p6 doesn't need parens around if blocks, parens help even if that's a practice from elsewhere. :)
pmurias if'ing for an literal doesn't seem very sensible
masak pmurias: well, hence the high probability of a thinko. 12:10
slavik how can I make an iterator for a list?
colomon loliblogged: justrakudoit.wordpress.com/2010/05/...-numerics/
isBEKaml masak: if ( my $a == 42) # I prolly won't make that mistake if we need parens here..
slavik is it: $blah <== @list ?
masak pmurias: I sense some slight talking at cross purposes here. you do realize that sigmung finds possible thinkos?
isBEKaml masak: OTOH, using if as a postconditional does make me go off. say 'blah' if my $a = 42; ;0 12:11
masak isBEKaml: not sure what you mean. why would requiring parens help you not make a mistake? and what mistake are you referring to? using 'my' in the if statement?
isBEKaml masak: I was referring to '=' and '==' thinkos errors. 12:12
pmurias masak: using my in a if statement doesn't seem to be that likely a mistake
masak pmurias: indeed. but I only need to use 'my' there when I'm testing for definedness. otherwise I tend to use '->'. 12:13
isBEKaml masak: Now I see you were referring to the 'my' in an if conditional. Now that doesn't make sense to me.
masak isBEKaml: more cross purposes. this conversation needs to slow down a bit. :) 12:14
maybe we should talk in full sentences via nopaste instead.
isBEKaml :)
masak: maybe, we just found another error here. ;) 12:15
pmurias masak: so sigmund is intendend to be a "check for the suspisious stuff" debugging aid rather than a Devel::Critic for Perl6
masak pmurias: yeah.
granted, it could evolve into a Critic too.
pmurias * Perl::Critic
masak but it was born out of the need to find stupid thinko stuff early. 12:16
pmurias it seems like something that could benefit very much from editor/ide integration
masak I started making a list over a year ago of things that would be trivial or easy to catch, using only the parse tree or a slghtly higher-level AST.
pmurias: yes. I'll explicitly leave that to someone else, though. the Padre people, maybe.
I'll try to make such integration easy, of course. but I think that would detract too much from what I want to be working on, and probably someone else would wnjoy it more anyway. 12:17
pmurias masak: you are an emacs user?
masak aye. and a vim user. 12:18
I also use Perl 5, Perl 6, Python and Ruby. :)
pmurias masak: you use viper? 12:19
masak no.
I tend to avoid hybrids.
isBEKaml masak: if you're an emacs user, you could write up a plugin for that using elisp?
masak perhaps.
isBEKaml I don't use emacs, so I don't know. (vim user)
masak well, it's certainly possible.
I'm still not terribly good at Elisp. 12:20
and I don't use Emacs for programming much.
mostly for prose.
isBEKaml I'm not good at vimscript either.
slavik rakudo: say $*IN.WHAT
p6eval rakudo 9a15b8: OUTPUT«IO()␤»
slavik dang it 12:21
masak slavik: how so?
slavik rakudo: my $a = 1..10; say $a.WHAT;
p6eval rakudo 9a15b8: OUTPUT«Range()␤»
slavik masak: ?
masak slavik: why 'dang it' on that result?
seems perfectly fine to me.
12:22 carlin joined
slavik masak: it return IO(), I expected it to say something that would indicate it as an iterator 12:22
isBEKaml slavik: that's an implicit file handle for reading used by p6. what's wrong?
slavik rakudo: my $a = 1..10; for =$a { say }; 12:23
p6eval rakudo 9a15b8: OUTPUT«Confused at line 11, near "for =$a { "␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
masak slavik: prefix:<=> is dead, long live .get 12:25
isBEKaml rakudo: my $a = 1..10; for $^a { .say }
p6eval rakudo 9a15b8: OUTPUT«12345678910␤»
masak o.O
masak submits rakudobug
slavik masak: so ... it's $*IN.get ?
masak slavik: for one line, yes.
slavik: but for many lines, you want .lines
slavik masak: just learning what the new (while (<>)) would look like ;) 12:26
isBEKaml masak: bug? :O
masak rakudo: my $a = 42; say $^a
p6eval rakudo 9a15b8: OUTPUT«42␤»
masak std: my $a = 42; say $^a
p6eval std 30748: OUTPUT«===SORRY!===␤Placeholder variable $^a cannot be used in this kind of block at /tmp/u3vXsAgXHU line 1:␤------> my $a = 42; say $^a⏏<EOL>␤Check failed␤FAILED 00:01 115m␤»
masak alpha: my $a = 42; say $^a
p6eval alpha 30e0ed: OUTPUT«42␤»
masak huh.
slavik: for lines($*ARGFILES) { ... } 12:27
isBEKaml masak: doesn't look like a bug to me. I think of ^ as a limiting operator here. 12:28
masak isBEKaml: sure, if you use it before the dollar. 12:29
isBEKaml: after, it's a twigil.
12:29 alanhaggai_ joined
isBEKaml rakudo: my $temp = 5; for $^temp { .say } 12:29
p6eval rakudo 9a15b8: OUTPUT«5␤»
masak rakudo: my $temp = 5; for ^$temp { .say }
p6eval rakudo 9a15b8: OUTPUT«0␤1␤2␤3␤4␤»
masak that's the difference.
isBEKaml masak: yes, I defined $temp to be a range here.
rakudo: my $temp = 0 .. 5; for $^temp {.say } 12:30
masak no, you didn't. it's an Int.
p6eval rakudo 9a15b8: OUTPUT«012345␤»
masak isBEKaml: do you believe me when I claim that you're slightly confused about this?
isBEKaml: I'll be happy to set you straight, though.
isBEKaml masak: I'm always confused. ;)
masak isBEKaml++ # excellent starting point 12:31
12:31 alanhaggai_ left
isBEKaml masak: Range first, Int second, Range again. # I don't know about twigils 12:31
masak isBEKaml: ok, so. in for loops, you'll be most likely to see `for ^$variable { ... }`
that's the range you're talking about. $variable is numified and used as the exclusive-upper point of a range starting at 0.
TimToady referred to prefix:<^> as 'upto' the other day. 12:32
a 'twigil', or a 'secondary sigil' is something that comes after a sigil (one of $ @ % & ::)
twigils usually denote different scoping of some kind. 12:33
for example $!attr and $.meth denote object has-scoping.
and $*contextual denote some dynamic scoping.
slavik $.meth makes no sense ... $.makemeth on the other hand ... does
:) 12:34
masak isBEKaml: so ^$var creates a range, whereas $^var denotes a variable which also is an auto-declared positional parameter in a sigil-less block. 12:35
isBEKaml: a bit like `sort { $a <=> $b }` in Perl 5.
isBEKaml: in Perl 6, you'd write `sort { $^a <=> $^b }`, which is sugar for `sort -> $a, $b { $a <=> $b }` 12:36
i.e. a block with two parameters $a and $b. but with $^a and $^b, you don't have to declare them as such.
isBEKaml masak: yeah, I tried -> this morning.closures. 12:37
masak now, it's CONCEIVABLE that one sometimes means `for $^range { ... }` in code, but it'd have to be very rare.
most likely it's a typo for `for ^$range { ... }`
12:38 Su-Shee left
isBEKaml masak++ # twigils and range limit operator 12:39
masak: Now, I definitely agree there has to be some sort of a gotcha here by the p6 compiler to notify the programmer.
masak isBEKaml: you're lucky to be here in 2010 so you get the long answer. :)
isBEKaml: in 2015, we'll have to dole out FAQ URLs instead.
isBEKaml masak: I'm lucky I got in early! ;) 12:40
masak yup.
isBEKaml rakudo: my $temp = { $^x * $^x }; say $temp(42); 12:41
p6eval rakudo 9a15b8: OUTPUT«1764␤»
isBEKaml yay for twigils! :)
masak S02 has the story.
and probably S06, a bit.
isBEKaml masak: I still haven't gone through it fully... too much detail, too much distraction :( 12:42
masak know the feeling. :/
you will, eventually.
rakudo: my $a = sub { say $^a + $^a }; $a(4)
p6eval rakudo 9a15b8: OUTPUT«8␤»
masak rakudo: my $a = sub { say $^a + $^a }; $a(4, 5)
p6eval rakudo 9a15b8: OUTPUT«Too many positional parameters passed; got 2 but expected 1␤current instr.: '_block48' pc 329 (EVAL_1:33593300)␤»
masak ah, closing rt.perl.org/rt3/Ticket/Display.html?id=73106 12:43
isBEKaml rakudo: my $temp = { $^b + $^c }; say $temp(23, 45);
p6eval rakudo 9a15b8: OUTPUT«68␤»
isBEKaml twigils can bind themselves to anything? this looks a lot like closures... 12:44
masak well, they are closures.
they bind themselves to the innermost surrounding block.
12:44 BrowserUk joined
masak which is a problem sometimes, because you might want to refer to a further-out surrounding block. 12:45
isBEKaml masak: AFAIK, closures can bind one local parameter and one free variable. Binding the free variable here is what constitutes a closure. #I might be wrong here...
masak which is why $a is a synonym for $^a after the latter has been used once.
isBEKaml: a closure is simply a block that retains information about its surrounding environment. 12:46
isBEKaml rakudo: my $temp = { say $^b; $^c ** 2 }; say $temp(23, 'String') 12:47
p6eval rakudo 9a15b8: OUTPUT«23␤0␤»
masak rakudo: sub iter-gen() { my $c = 0; return { $c++ } }; my $i = iter-gen(); .say for $i(), $i(), $i()
p6eval rakudo 9a15b8: OUTPUT«0␤1␤2␤»
isBEKaml rakudo: my $temp = { say $^b; say $^c ** 2 }; say $temp(23, 'String')
p6eval rakudo 9a15b8: OUTPUT«23␤0␤1␤»
isBEKaml rakudo: my $temp = { say $^b; say $^c ** 2 }; $temp(23, 'String') 12:48
p6eval rakudo 9a15b8: OUTPUT«23␤0␤»
masak here, the closure is { $c++ }. it retains a binding to the $c, declared in the 'iter-gen' sub.
you could store the closure from several different iter-gen calls into different variables, and they'd all count up their very own $c independently of each other.
I'd show it, but Rakudo is broken in that area right now, so one needs to manyally pir::clone() to make it really work. 12:49
but anyway, that's closures. they 'close' over a surrounding environment, keeping references to some variables which would otherwise have gone out of scope.
in many ways, they have the aspects of very light-weight objects. 12:50
isBEKaml masak++ # closures-in-detail 12:51
12:51 carlin left
masak I seem to be in Tutor Mode today. :P 12:51
maybe a good time to take a look at the book. 12:52
isBEKaml I only wish I were here early than having to worry about not being able to read specs... :(
yeah and maybe update the glossary 12:53
masak the specs are hard to read. I'm currently struggling through what the last section of S04 actually means.
isBEKaml masak: Very relevant. "When a closure is not a closure" 12:55
:)
masak yeah. I think I see the intent of that section, but my brain hasn't yet succeeded in enveloping the technical specifics.
which probably means I should keep chunking up until the concepts are few enough to fit. :) 12:56
(and no, 'chunking up' is not a drug euphemism, it's an NLP term. www.renewal.ca/nlp27.htm ) 12:57
isBEKaml that's what this says: ...ever use the current reference to the routine, it gets the current snapshot of its world in terms of the lexical symbols that are visible to it
:D
masak: don't bother explaining euphemisms. Non-native here. :) 13:02
masak hey, me too. I'm just a silly Swede.
isBEKaml masak: Btw, that quote from S04 says: "Keep it small, keep it close" :D
masak likes that quote 13:03
isBEKaml and I, a confused Indian. ;)
masak ah, Bharato. 13:04
isBEKaml Bharat, yeah. (internal name among ourselves, like Nippon to Japanese )
jnthn o/ 13:05
masak jnthn!
isBEKaml yo, jnthn !
jnthn мэсак!!
masak 'Bharato' is the name sometimes used instead of 'Hindio' by politically-aware Esperantists.
:)
I haven't heard anyone use 'Nipono', though. 13:06
though some say 'Suomo' rather than 'Finnlando'.
Suomio* 13:07
isBEKaml suomi!
jnthn I heard their language is hard to finnish learning.
masak -Opun 13:08
Finnish is said to be very grammatically pleasing, actually.
isBEKaml Finnish won't finish learning their language!
masak Perl6ers won't finnish their language... :)
jnthn masak: Yeah, I've heard that. Which makes me wonder...since I like grammar, would I actually find it not quite so terrifying. 13:09
iiuc, there's lots of cases. But, well, those are OK. :-)
masak jnthn: there's a prepositional noun case in Estonian (which is related to Finnish), meaning 'beside', derived from the word for 'ear'. the thought pleases me. 13:10
jnthn Wow. :-) 13:11
isBEKaml :) 13:12
masak [citation needed]
BrowserUk phenny tell ruoso: In your vision, can coderefs be passed between kthreads? Eg. my $x=123; sub ($a) thing{ return $a * $x } ...; my @a <== map{ thing( $_ ) }, 1 .. 10; That is, the piped map is run in a kthread, it uses thing() which is a first class function, defined in the current kthreads as a closure over $x, which is a lexical defined in the spawning kthread. If closures are done by CPS keeping around stack frames, how does the spawned
thread get to see the value of $x?
masak ETOOLONG
jnthn masak: It's the Lund Carnival today
masak ENOCOLON
jnthn masak: Feels like a bit of a DDoS of the city. :-)
masak :P
jnthn But kinda fun too. 13:13
pmichaud good morning, #perl6
jnthn :-)
masak pmichaud! \o/
jnthn moritz_, pmichaud
oh
*morning
pmichaud (closures): nqp has working closures, even if rakudo doesn't at the moment :-)
masak good moritz_ :)
jnthn wtf, I tried to tab-complete "morning"
BrowserUk phenny tell ruoso In your vision, can coderefs be passed between kthreads? Eg. my $x=123; sub ($a) thing{ return $a * $x } ...; my @a <== map{ thing( $_ ) }, 1 .. 10; That is, the piped map is run in a kthread, it uses thing() which is a first class function, defined in the current kthreads as a closure over $x, which is a lexical defined in the spawning kthread.
masak ESTILLNOCOLON
BrowserUk phenny: tell ruoso In your vision, can coderefs be passed between kthreads? Eg. my $x=123; sub ($a) thing{ return $a * $x } ...; my @a <== map{ thing( $_ ) }, 1 .. 10; That is, the piped map is run in a kthread, it uses thing() which is a first class function, defined in the current kthreads as a closure over $x, which is a lexical defined in the spawning kthread.
phenny BrowserUk: I'll pass that on when ruoso is around.
masak jnthn: clearly we need to invite someone called 'morning' to the channel. :P 13:14
13:14 isBEKaml is now known as morning
morning :D 13:14
jnthn \o/ 13:15
BrowserUk phenny: tell ruoso If closures are done by CPS keeping around stack frames, how does the spawned thread get to see the value of $x?
phenny BrowserUk: I'll pass that on when ruoso is around.
morning uhh...
morning pmichaud
ah, that's better! ;)
13:16 morning is now known as isBEKaml
masak morning++ 13:16
pmichaud nqp: sub iter-gen() { my $c := 0; return { $c++ } }; my $x := iter-gen(); my $y := iter-gen(); say $x(), $y(); say $x(); say $x()
p6eval nqp: OUTPUT«Routine declaration requires a signature at line 1, near "-gen() { m"␤current instr.: 'parrot;HLL;Grammar;panic' pc 501 (src/cheats/hll-grammar.pir:196)␤»
jnthn pmichaud: nqp-rx doesn't support - in identifiers. 13:17
pmichaud nqp: sub iter-gen() { my $c := 0; return { $c++ } }; my $x := iter-gen(); my $y := iter-gen(); say($x(), $y()); say($x()); say($x())
p6eval nqp: OUTPUT«Routine declaration requires a signature at line 1, near "-gen() { m"␤current instr.: 'parrot;HLL;Grammar;panic' pc 501 (src/cheats/hll-grammar.pir:196)␤»
jnthn pmichaud: I was tempted to fix it but then thought you may have a reason.
pmichaud nqp: sub itergen() { my $c := 0; return { $c++ } }; my $x := itergen(); my $y := itergen(); say($x(), $y()); say($x()); say($x())
p6eval nqp: OUTPUT«00␤1␤2␤»
jnthn nqp: my $oh-noes = "doesn't work";
p6eval nqp: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 1, near " \"doesn't "␤current instr.: 'parrot;HLL;Grammar;panic' pc 501 (src/cheats/hll-grammar.pir:196)␤»
jnthn nqp: my $oh-noes := "doesn't work";
p6eval nqp: OUTPUT«Multiple Dispatch: No suitable candidate found for 'subtract', with signature 'PPP->P'␤current instr.: '_block11' pc 0 (EVAL_1:4)␤»
pmichaud jnthn: yeah, I need to think about that (and what it might mean for translating to an underlying vm) 13:18
nqp: sub itergen() { my $c := 0; return { $c++ } }; my $x := itergen(); my $y := itergen(); say($x(), $y()); say($x()); say($x()); say($y());
p6eval nqp: OUTPUT«00␤1␤2␤1␤»
pmichaud yay for working closures (in nqp)
jnthn \o/
pmichaud: How's the context PAST node coming along?
pmichaud needs a good name :-) 13:19
jnthn pmichaud: I really haven't been able to pay much attention to developments over teh last few days.
pmichaud jnthn: I've primarily been working on REPL
I'm almost done
jnthn Oh, cool. \o/
pmichaud we now have this, though:
pmichaud@orange:~/nqp-rx$ ./nqp
> 3+4
7
>
(note: no "say()") 13:20
jnthn > say 3 + 4 # what does this do?
pmichaud however, if you explicitly print something, the return value isn't printed
jnthn OK, cool.
pmichaud pmichaud@orange:~/nqp-rx$ ./nqp
> say(3+4); 5+7;
7
>
masak someone should write up a HLL which can run the snippets in this old April Fools interview: www.perl.com/pub/a/2001/04/01/parrot.htm
jnthn > sub foo() { say 42 }; foo();
pmichaud > sub foo() { say(42); }; foo(); 13:21
42
jnthn
.oO( lol, using pmichaud++ as a repl )
masak :D
BrowserUk wonders if he's the only one who finds the lack of definitiveness in the SOs more than vaguely disconcerting?
pmichaud > sub foo() { 42 }; foo();
42
jnthn
.oO( Save Our Synopses )
masak 'SOs'?
ah, synopses, I presume. wasn't used to seeing that abbrv. 13:22
BrowserUk always takes 3 attempts to type Synopses.
13:22 protorom joined
masak BrowserUk: what helps most is if you air your disconcertedness in the form of questions here on-channel. 13:23
13:23 protorom left, protorom joined
masak some types of lack of definitiveness are actually the Sxx documents just being scattered due to the way they were written. 13:23
some things are indeed inconsistent, contradictory or not-specified-enough. we're always on the lookout for those things. 13:24
S09 was brought up as something the other day that'll probably undergo review as soon as someone starts implementing it.
also, S06 and S19 are known to contradict each other a bit. 13:25
pmichaud some things are not-definitive because we really need an implementation to work before we can be definite
13:29 protorom left
BrowserUk masak: It's quite hard to put into words with specific examples. I've been reading SO9 over and over, and there seem to be a lot of "maybe"s and "or shoudl it be?"s and "somehow"s ... more generally, just a lack of "this is how it will work". 13:29
masak BrowserUk: nod. 13:30
BrowserUk: what would be really useful, but will probably be quite hard, is if someone wrote "real code" on the current S09.
to sort of shake out the uncertainties.
colomon jnthn: (backlogging) Seems like once or twice a day I try to tab-complete a normal (but long) word here. :)
masak the problem with that, of course, is that there's nowhere to run that code right now. 13:31
BrowserUk I understand the reasoning for them starting out that way, but ... without joining any bandwagons... it feels like it should be somewhat more defined by now.
I'm guessing (hoping) that SO9 is probably one of the worst in this respect due to the nature of what it covers. 13:32
isBEKaml colomon: I was surprised to find my name in the release notes to Erlangen in your blog... What did I do wrong? ;) 13:34
13:34 redicaps joined
BrowserUk Normally, with less ambitious projects, you either define them in terms of what is known to be possible--and then implement. Or define what is required and then work out how to implement. BUt here both ends are speculating upon what the other will want/will be able to do. That makes progress from either end hard to reason about. 13:35
masak BrowserUk: IO (S16) is still weak. concurrency (S07) is weak. command line options (S19) are weak. Pod (S26) is weak. there are a couple of fairly solid synopses, notably S02, S03, S04, S05, S06 and S12.
isBEKaml Those solid synopses are what I'm trying to read, unsuccessfully, over the past couple of weeks... 13:36
isBEKaml thinks maybe he will get it in, one day... #just waits for the epiphany moment... ;) 13:37
BrowserUk doesn't have the requisite savvy for most of those, hence trying to stay out of them and stick to something I know a little about.
masak isBEKaml: what really helped for me was to use Rakudo and actually write stuff.
isBEKaml masak: I am, too. Trying to get existing examples in pugs repo working with Rakudo. Since, we no longer use pugs, I guess we're okay with porting them to rakudo. 13:39
masak isBEKaml: yes, that sounds like a worthy idea.
isBEKaml I have just temporarily put an hold on that pursuit until I gain more into S02 until S04... 13:40
13:41 FreeStorm is now known as FreeStorm_wTV
isBEKaml You see, coding up those in confused state is the last thing we all want... :) 13:42
pmichaud S08 is wrong. :-)
isBEKaml pmichaud++ ## he's definite, here.. :-)
BrowserUk maasak: with regard to writing real code. I've been trying. By utilising another language with sufficient existing working code to give a firm underpinning upon which to try things out. I even posted some Perl 5 code that emulated ruoso's test example, but it got no reaction at all. 13:44
13:45 molaf left
masak BrowserUk: sorry you feel you've been overlooked. I for one am excited to see discussion flowing on the concurrency topic, even if I'm too inexperienced to contribute any to it myself. 13:45
BrowserUk I'm playing with Erlang because it has both types of threading and a flexible syntax to allow me to "mock up" scenarios, but it slow going--the Erlang manuals are thorough, but tough to find your way around. 13:46
masak heh. I installed Erlang today, and plan to play around with it a little in my copious spare time. 13:47
BrowserUk doesn't feel "overlooked", just walking on eggshells most of the time.
masak good. I think we're all walking on eggshells. :)
moritz_ tramples over broken eggshells
pmichaud plasters over the broken eggshels
isBEKaml looks for eggshells! :-) 13:48
pmichaud "new coat of paint, and everything is fine!"
BrowserUk looks at the bulge in the carpet and and wonders if anyone else notices it :) 13:49
isBEKaml .u phone 13:50
phenny isBEKaml: Sorry, no results for 'phone'.
isBEKaml .u ☎ 13:51
phenny U+260E BLACK TELEPHONE (☎)
isBEKaml there... :|
.u tele 14:00
phenny U+2121 TELEPHONE SIGN (℡)
isBEKaml .u bl
phenny U+2712 BLACK NIB (✒)
masak phenny: tell TimToady -- I'm starting my Buf grant, and I have a question about the spec's use of the term 'character'; is it a contextually settable value rather than a Unicode abstraction level in itself, like 'byte' and 'graph'? 14:06
phenny masak: I'll pass that on when TimToady is around.
moritz_ hugme: add freestorm to irclog 14:09
hugme moritz_: sorry, I don't know anything about project 'irclog'
moritz_ hugme: add freestorm to ilbot
hugme hugs freestorm. Welcome to ilbot!
14:15 jakk joined 14:16 finanalyst joined
finanalyst hello everyone 14:24
moritz_ hi
pmurias BrowserUk: i use the book "Erlang Programming" for learning erlang
finanalyst i am having problems indexing an array inside a loop because the index is type Num, and rakudo wants type Int 14:25
moritz_ then .Int it
finanalyst thanx that works 14:26
is it a known bug?
isBEKaml finanalyst: what?
masak finanalyst: yes. 14:27
a quite quirky one.
isBEKaml rakudo: my @list = <tom dicken harry>; say @list[$_] for 0..+@list;
p6eval rakudo 9a15b8: OUTPUT«tom␤dicken␤harry␤␤»
finanalyst isBEMaml: inside a loop the index is considered type Num, but it should be Int.
isBEKaml: does not always happen 14:28
rakudo: for 1 .. 4 { say $_.WHAT }
p6eval rakudo 9a15b8: OUTPUT«Int()␤Int()␤Int()␤Int()␤»
masak rakudo: say <a b c d>[2.0]
p6eval rakudo 9a15b8: OUTPUT«c␤» 14:29
masak and it doesn't always produce an error.
that's the quirky part.
isBEKaml o.O
lue goood morning! 14:30
isBEKaml btw, why does +@list append a new line to the end of output? 14:31
rakudo: say $_ for 0..3;
p6eval rakudo 9a15b8: OUTPUT«0␤1␤2␤3␤»
isBEKaml rakudo: my @list = <tom dicken harry>; say $_ for 0..+@list; say @list[$_] for 0..+@list; 14:32
p6eval rakudo 9a15b8: OUTPUT«0␤1␤2␤3␤tom␤dicken␤harry␤␤»
moritz_ isBEKaml: you have an offbyone error
there are only 0..(@list-1) items in your array
or ^@list or @list.keys 14:33
isBEKaml moritz_++ #noticed it just now when I did $_
masak lue: \o 14:34
isBEKaml rakudo: say <a b c d>[2.00001]
p6eval rakudo 9a15b8: OUTPUT«c␤»
isBEKaml rakudo: say <a b c d>[2.554] 14:35
p6eval rakudo 9a15b8: OUTPUT«c␤»
finanalyst rakudo: class A { has $.a; method m { for $.a .. 5 { say $_.WHAT }}}; my A $x.=new(:a(2.0)); $x.m
p6eval rakudo 9a15b8: OUTPUT«Rat()␤Rat()␤Rat()␤Rat()␤»
isBEKaml maybe it does an equivalent of ceil() ?
d'oh, floor() ? 14:36
colomon isBEKaml: probably truncate
moritz_ rakudo: say (Any .. 5).perl
p6eval rakudo 9a15b8: OUTPUT«Any..5␤»
moritz_ rakudo: say (Any .. 5).Seq.perl
p6eval rakudo 9a15b8: OUTPUT«()␤»
isBEKaml but doesn't *always* show an error is really quirky..
masak here's the ticket I was thinking of: rt.perl.org/rt3/Ticket/Display.html?id=74430 14:38
finanalyst rakudo: class A {has $.a; has @.m; method foo { for $.a .. 5 { say @.m[$_] }}}; my A $x.=new(:a(2.0),:m(1,2,3,4,5));$x.foo
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'. Available candidates are:␤:(Mu : Int $i;; *%_)␤:(Mu : Block $b;; *%_)␤:(Mu : !whatever_dispatch_helper ;; *%_)␤␤current instr.: '!postcircumfix:<[ ]>' pc 11523 (src/builtins/Any.pir:54)␤»
masak it's array indexing and 'is copy' interacting badly with each other. 14:39
finanalyst isBEKaml: thats the error
masak hm, but that one seems to be attr arrays and indexing.
rakudo: class A { has @.m; method foo { @.m[2.0] } }; A.new(:m(1,2,3)).foo 14:40
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'. Available candidates are:␤:(Mu : Int $i;; *%_)␤:(Mu : Block $b;; *%_)␤:(Mu : !whatever_dispatch_helper ;; *%_)␤␤current instr.: '!postcircumfix:<[ ]>' pc 11523 (src/builtins/Any.pir:54)␤»
masak submits rakudobug
finanalyst++
rakudo: my @m = 1,2,3; say @m[2.0] 14:41
finanalyst masak: somehow the type of the attribute is being 'imported' during the call to new
p6eval rakudo 9a15b8: OUTPUT«3␤»
colomon it would be sooooo awesome if the first Mu there in all of those was the actual type involved...
masak finanalyst: you mean of $.a? but that's not it, my example didn't use $.a
isBEKaml rakudo: class A { has @.m; method foo { @.m[2] } }; A.new(:m(1,2,3)).foo 14:42
p6eval rakudo 9a15b8: ( no output )
isBEKaml rakudo: class A { has @.m; method foo { @.m[2.5] } }; A.new(:m(1,2,3)).foo.say 14:43
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'. Available candidates are:␤:(Mu : Int $i;; *%_)␤:(Mu : Block $b;; *%_)␤:(Mu : !whatever_dispatch_helper ;; *%_)␤␤current instr.: '!postcircumfix:<[ ]>' pc 11523 (src/builtins/Any.pir:54)␤»
finanalyst masak: I ran across a bug where arrays inside a method were being treated differently than outside. is this the same thing?
we discussed it about two weeks ago
masak rakudo: class A { has @.m; }; A.new(:m(1,2,3)).m[2.0]
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'postcircumfix:<[ ]>'. Available candidates are:␤:(Mu : Int $i;; *%_)␤:(Mu : Block $b;; *%_)␤:(Mu : !whatever_dispatch_helper ;; *%_)␤␤current instr.: '!postcircumfix:<[ ]>' pc 11523 (src/builtins/Any.pir:54)␤»
colomon finanalyst: probably is the same thing.
masak it's not even methods.
it's just the attribute array.
14:56 rv2733 joined
BrowserUk pmurias: Is "Erlang Programming" is a real, go out and buy book only, rather than an ebook? 14:57
pmurias BrowserUk: i have the ebook 15:02
BrowserUk: but i guess you could buy a printed copy if you want
15:07 orafu left, orafu joined 15:13 rokoteko joined
isBEKaml masak: an update on the rakudo bug around $^ variables. The usage appears to be correct since we can't use it for looping. 15:35
rakudo: my $temp = 1..10; for $^temp { say "hi"; }
p6eval rakudo 9a15b8: OUTPUT«hi␤»
masak isBEKaml: 'the usage appears to be correct since we can't use it for looping'. I do not understand this phrase. to me, it's clearly a bug.
isBEKaml gack, s/usage/logic/
masak what logic? the one behind the argument that it's a bug? :) 15:36
TimToady std: my $temp = 1..10; for $^temp { say "hi"; }
phenny TimToady: 14:06Z <masak> tell TimToady -- I'm starting my Buf grant, and I have a question about the spec's use of the term 'character'; is it a contextually settable value rather than a Unicode abstraction level in itself, like 'byte' and 'graph'?
isBEKaml masak: the logic behind the for loop, though there should be checks against using $^ variables in a loop.
p6eval std 30748: OUTPUT«===SORRY!===␤Placeholder variable $^temp cannot be used in this kind of block at /tmp/ENxwcIDYNz line 1:␤------> my $temp = 1..10; for $^temp⏏ { say "hi"; }␤Check failed␤FAILED 00:01 115m␤»
masak isBEKaml: no, I'm not sure there should. 15:37
isBEKaml: as I said, it's not always wrong.
there might be very unusual but completely admissible situations where it's right.
TimToady std: { my $temp = 1..10; for $^temp { say "hi"; } } 15:38
p6eval std 30748: OUTPUT«Potential difficulties:␤ Useless redeclaration of variable $temp (see line 1) at /tmp/EpKKXBb4Lo line 1:␤------> { my $temp = 1..10; for $^temp⏏ { say "hi"; } }␤ok 00:01 112m␤»
15:39 elmex left
isBEKaml points at std... 15:39
masak TimToady: hunh. I thought that'd actually be illegal.
TimToady it shouldn't allow a $^temp after 'my $temp' 15:40
masak phew. 15:42
isBEKaml rakudo: my $somevar= { say $_ for $^temp; }; $somevar(1..10);
p6eval rakudo 9a15b8: OUTPUT«1 2 3 4 5 6 7 8 9 10␤»
15:42 elmex joined
masak isBEKaml: yeah, things like that. 15:43
isBEKaml ouch*
masak there's always the balance between helping people who don't know what they're doing, and getting out of the way of those who do, of course.
TimToady my first test was illegal because you can't use a placeholder in mainline code, though "this kind of block" is LTA 15:44
"in this lexical context" would be more accurate 15:45
15:49 TiMBuS left
isBEKaml wonders if he exhausted his mistake quota for the day.... ;) 15:49
masak isBEKaml++ # trial and error
TimToady: about the Spec/character question: I'm trying to figure out what a 'character' really is in the context of Perl 6. whether it is a Unicode level (and if so, which one) or some kind of parameter that you can set to a Unicode level (and if so, how). 15:50
pmurias TimToady: wouldn't "outside a block" be a better error message here
isBEKaml TimToady++, masak++, #perl6++ ## awesome community to help ease into trial & error 15:51
pmurias std: sub ($foo) {$^bar}
p6eval std 30748: OUTPUT«===SORRY!===␤Placeholder variable $^bar cannot override existing signature ($foo) at /tmp/zkqefQuVs6 line 1:␤------> sub ($foo) {$^bar⏏}␤Check failed␤FAILED 00:01 112m␤»
pmurias std: sub {$^bar}
p6eval std 30748: OUTPUT«ok 00:01 111m␤»
pmurias std: if 1 {$^bar}
p6eval std 30748: OUTPUT«ok 00:01 111m␤»
TimToady pmurias: yes, but all it knows is that the current lexpad doesn't take a signature; it might be in a block 15:52
15:52 tedv joined
TimToady std: class Foo { $^nope } 15:52
p6eval std 30748: OUTPUT«===SORRY!===␤Placeholder variable $^nope cannot be used in this kind of block at /tmp/5drlz4r1UQ line 1:␤------> class Foo { $^nope⏏ }␤Check failed␤FAILED 00:01 111m␤» 15:53
jnthn std: role Foo { $^nope }
p6eval std 30748: OUTPUT«===SORRY!===␤Placeholder variable $^nope cannot be used in this kind of block at /tmp/rwi1Z77Bag line 1:␤------> role Foo { $^nope⏏ }␤Check failed␤FAILED 00:01 111m␤»
TimToady let's not allow that :) 15:54
jnthn rakudo: role Foo { $^nope; method bar() { say $nope } }; class lol does Foo['omg'] { }; lol.new.bar
p6eval rakudo 9a15b8: OUTPUT«Symbol '$nope' not predeclared in bar␤current instr.: 'perl6;PCT;HLLCompiler;panic' pc 152 (compilers/pct/src/PCT/HLLCompiler.pir:108)␤»
jnthn Phew. :-)
TimToady: I could half see Rakudo acidentally making it work. :-)
pmurias TimToady: Placeholder variable $^nope cannot be used in a block which can't have a signature.? 15:56
maybe s/can't/is not allowed to/
15:57 redicaps left 16:00 FreeStorm_wTV is now known as FreeStorm
masak and shouldn't be allowed to :P 16:01
TimToady my current code say "may not"
and gives a different message if $*CURPAD === $*UNIT 16:02
making it check for an existing my is a bit harder though
pmurias std: {my $foo;say $^foo} 16:03
p6eval std 30748: OUTPUT«Potential difficulties:␤ Useless redeclaration of variable $foo (see line 1) at /tmp/AoqTit87g8 line 1:␤------> {my $foo;say $^foo⏏}␤ok 00:01 114m␤»
TimToady retroactively changing a non-parameter to a parameter is not nice 16:04
16:10 pmurias left
masak that's my point. 16:10
tedv By "not nice" do you mean "swears at the police officer" not nice or "breaks the speed of light" not nice? 16:11
cognominal \o 16:12
masak o/
slavik \o/ 16:13
masak a multi-hour bug hunt in GGE finally ended happily. 16:14
16:14 FreeStorm left
masak ultimate cause: `while index(...)` instead of `while defined(index(...))` 16:14
so, don't do that.
lue Well, the index(...) could have been 0! :) 16:18
masak right.
it was, in the cases where the bug manifested. 16:19
lue: for example in test 440, where I left off yesterday.
that one should pass now.
TimToady By "not nice" I mean "Does a bait-and-switch on the customer."
the customer here being the *reader* of the code, not the writer
masak the writer isn't being nice to the reader. 16:20
lue
.oO(Not Nice as in "Buying ebooks on the Kindle and then having Amazon take them back for no good reason.")
 16:21
TimToady though I suppose the compiler is also a reader
masak and, by transitivity, a programming language which allows such a bait-and-switch isn't being nice to the reader wither.
s/wither/either/
diakopter compilers emit tears too
masak diakopter++ # that should be a tutorial of some sort 16:22
16:22 iblechbot left
cognominal TimToady++ # "bait ans switch on the customer", so true for me, these last few days in meat space. But working actively in Internet space to be sure this will be fixed. 16:23
diakopter "how to break your compiler until it says uncle and cries"
16:23 JimmyZ left
lue diakopter: what if it tries to escape? 16:25
masak .oO( Compile-or-flight response ) 16:26
awk is known for trying to bail out when things get tricky.
lue I heard a rumor that '' keeps them from escaping (but that seems odd, because "" seems stronger) :) 16:27
masak lue: I think you have it too literally :P 16:28
nom & 16:29
16:29 masak left 16:37 Exodist left, Exodist joined 16:39 PacoLinux left 16:41 rgrau left 16:42 PacoLinux joined
moritz_ rakudo: say Array.new eqv [] 16:47
p6eval rakudo 9a15b8: OUTPUT«1␤»
cognominal rakudo: say True 16:50
p6eval rakudo 9a15b8: OUTPUT«1␤»
cognominal rakudo: say True.perl
p6eval rakudo 9a15b8: OUTPUT«Bool::True␤»
cognominal (.perl)++ 16:51
snarkyboojum rakudo: my @a = 1 .. Inf;
p6eval rakudo 9a15b8: ( no output )
cognominal beats me why this is not the default
snarkyboojum I take it that's timing out because lists aren't lazy yet
16:55 rindolf joined
rindolf Hi all. 16:55
16:58 JimmyZ joined
pugssvn r30749 | snarkyboojum++ | [t/spec] Added initial tests for Perl 6 Advent Calendar Day 23: Lazy fruits from the gather of Eden 17:05
rokoteko snarkyboojum: is there something like take(5, 1..Inf) in perl6? .. what Im expecting it to return is 1, 2, 3, 4, 5 17:12
snarkyboojum rokoteko: nope - not that I know of anyway 17:14
take(5, some list), will return 5 and then the list 17:15
rokoteko so how is one supposed to refer to the first 5 elements of an infinite list if perl6 provide lazy evaluation? 17:16
@arr = (1..Inf); @arr[0..4] ? when is @arr evaluated?
17:17 JimmyZ left
lue rakudo: my @a = (1..Inf); say @[0..4] 17:20
p6eval rakudo 9a15b8: ( no output )
lue rakudo: my @a = (1...Inf); say @[0..4]
snarkyboojum rakudo: my @a = 1..10; say @a[^5]
p6eval rakudo 9a15b8: ( no output )
rakudo 9a15b8: OUTPUT«12345␤»
lue I could have sworn we had lazy lists all figured out...
TimToady roketeko, use .batch(5) 17:22
snarkyboojum lue: I don't think Rakudo has properly lazy lists (?) 17:23
TimToady rakudo: say (1...*).batch(5)
p6eval rakudo 9a15b8: OUTPUT«12345␤»
rokoteko TimToady: (1..Inf).batch(5) ?
ah ok.
cant you have it on the LHS?
TimToady 'take' means something else in p6
I dunno 17:24
rakudo: say batch 5, 1..*
p6eval rakudo 9a15b8: OUTPUT«Could not find sub &batch␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
TimToady could easily be defined
rokoteko rakduo: say batch(5).(1..Inf) 17:25
rakudo: say batch(5).(1..Inf)
p6eval rakudo 9a15b8: OUTPUT«Could not find sub &batch␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
TimToady . is not what it is in Haskell
lue std: my @a = 1..Inf
p6eval std 30748: OUTPUT«ok 00:01 114m␤»
TimToady it's just method call, not function compositoin
so that doesn't make sense
rokoteko TimToady: how would you chain it then perl6? 17:26
+way
TimToady well, there's probably some way to curry it
hmm
rokoteko currying more than once is what I missed in perl5 :) 17:27
TimToady rakudo: my $x = *.batch(5); say $x(1...*)
p6eval rakudo 9a15b8: OUTPUT«12345␤»
TimToady there's one way
rokoteko what is *?
TimToady it's a "DON'T curry this argument"
rokoteko refers to the last result?
TimToady so *.batch(5) really means { $^arg.batch(5) } 17:28
or -> $arg { $arg.batch(5) } in a more lambdaish notation
rokoteko TimToady: Im not a haskell programmer :)
but well put words anyway. :) 17:29
TimToady why not? haskell is easy to learn; I've learned it five or six times now
but yeah, everyone is coming from a different background
so it behooves us to say things several different ways
in the hopes that one will stick
rokoteko well, Im planning to be, but it just for the cause of understanding something different from imperiative languages.
TimToady what is your background? 17:30
rokoteko as a programmer? ppl -> basic -> perl5 -> php -> c -> java -> haskell ... but currently at perl5. 17:31
ppl is pcobard programming language.
pcboard.
snarkyboojum rakudo: my @x = <a b c d e>; say @x.map: * ~ 'A'; 17:32
p6eval rakudo 9a15b8: OUTPUT«No candidates found to invoke␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
TimToady not all * forms are correctly curried yet in rakudo
snarkyboojum TimToady: but that should work eventually? 17:33
17:34 szabgab_ joined
snarkyboojum and build ["aA", "bA", "cA", "dA", "eA"]? 17:34
rokoteko rakduo: my $x = *.**2; say $x(1...5)
rakudo: my $x = *.**2; say $x(1...5)
p6eval rakudo 9a15b8: OUTPUT«Confused at line 11, near "my $x = *."␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
rokoteko rakudo: my $x = *.(**2); say $x(1...5) 17:35
p6eval rakudo 9a15b8: OUTPUT«Method 'Num' not found for invocant of class 'Block'␤current instr.: 'perl6;Mu;' pc -1 ((unknown file):-1)␤»
rokoteko intuitively its not clear for me. :/
oh wait. 17:36
rakudo: my $x = *.(**2); say $x(1..5)
p6eval rakudo 9a15b8: OUTPUT«Method 'Num' not found for invocant of class 'Block'␤current instr.: 'perl6;Mu;' pc -1 ((unknown file):-1)␤»
pugssvn r30750 | lwall++ | [STD] awesomize placeholder messages
r30750 | disallow use of placeholder after non-placeholder
r30750 | [viv] suppress recursion warning
TimToady snarkyboojum: correct 17:37
well, except for the [...] part
it would just be a parcel of parcels
snarkyboojum ok
TimToady std: my $x = *.(**2); say $x(1..5) 17:38
p6eval std 30749: OUTPUT«===SORRY!===␤Unable to parse argument list at /tmp/lJAZkrxa73 line 1:␤------> my $x = *.(⏏**2); say $x(1..5)␤Couldn't find final ')'; gave up at /tmp/lJAZkrxa73 line 1:␤------> my $x = *.(**⏏2); say $x(1..5)␤ expecting any
..…
TimToady std: my $x = *.(* * 2); say $x(1..5)
p6eval std 30749: OUTPUT«ok 00:01 112m␤»
TimToady ** is a valid term
17:39 ajs joined
moritz_ that breaks my old starry obfu :-) 17:39
TimToady we've broken one or two other things as well :P
moritz_ rakudo: **
p6eval rakudo 9a15b8: OUTPUT«Confused at line 11, near "**"␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
isBEKaml rakudo: say 2** 5;
p6eval rakudo 9a15b8: OUTPUT«32␤»
TimToady it's also a valid infix 17:40
ajs rakudo: toekn isegment { blah }
p6eval rakudo 9a15b8: OUTPUT«Could not find sub &isegment␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
ajs that one took me a while to see ;-)
Anyway, my URI.pm now parses... I'll debug it after I do some chores. 17:41
TimToady std: toekn isegment { blah } 17:42
p6eval std 30749: OUTPUT«===SORRY!===␤Undeclared routines:␤ 'blah' used at line 1␤ 'isegment' used at line 1␤ 'toekn' used at line 1␤Check failed␤FAILED 00:01 111m␤»
lue Is there *anything* for Rakudo/P6 that I can do? I'm extremely bored [checking ROADMAP and various] 17:43
TimToady here, wave these pompoms and yell :)
snarkyboojum lue: write tests for the advent calendar :)
TimToady std: {my $foo;say $^foo}
p6eval std 30749: OUTPUT«Potential difficulties:␤ Useless redeclaration of variable $foo (see line 1) at /tmp/avPDaQnzgt line 1:␤------> {my $foo;say $^foo⏏}␤ok 00:01 111m␤»
TimToady well, that should fix itself while I take a walk :) 17:44
bbl &
lue *o* 17:45
isBEKaml lue: what's wrong? 17:48
diakopter phenny: tell pmurias I fixed the problem(s) there, mostly.
phenny diakopter: I'll pass that on when pmurias is around.
lue No; that was a cheerleader (see TimToady ^^^) 17:49
isBEKaml :)
17:50 szabgab_ left
dalek meta: r294 | diakopter++ | trunk/Sprixel/ (12 files):
[perlesque] add more (commented) debug for some grammar combinators, make State

be fully debugged, fix parse_bug2.t since it tried to augment a class already defined, which can't occur.		 17:51
lue goes forth to have an advent-ure in P6 17:52
snarkyboojum lue: I'm having a go at day 19 atm fyi 17:53
dalek ok: a30e807 | jonathan++ | src/operators.pod:
Some initial explanations of smart matching.
 17:55
ok: e86ce00 | jonathan++ | src/subs-n-sigs.pod:
Examples of where without a block, which just smart matches.
meta: r295 | diakopter++ | trunk/Sprixel/Main.cs:
[perlesque] once again, remove debug flag
 17:56
lue Where exactly are the advent tests?
isBEKaml lue: in the pugs repo. t/spec/integration
lue ah, thank you. 17:58
isBEKaml yw.
17:59 rurban joined
lue I'm confused. Do you write/change the tests, or make them the implementation run them correctly? 17:59
17:59 rurban left
snarkyboojum lue: just write tests where they're missing I'd say 18:00
isBEKaml rakudo: any(1,2,3).WHAT.perl.say; 18:02
p6eval rakudo 9a15b8: OUTPUT«Junction␤»
diakopter phenny: tell pmurias well, I guess I broke other things. :/
phenny diakopter: I'll pass that on when pmurias is around.
isBEKaml rakudo: my $tt = "yah"; say "booyah!" if $tt ~~ any("boo", "yah"); 18:03
p6eval rakudo 9a15b8: OUTPUT«booyah!␤»
pugssvn r30751 | snarkyboojum++ | [t/spec] Added initial tests for Perl 6 Advent Calendar Day 19: Whatever 18:04
isBEKaml snarkyboojum: I'm currently trying to come up with tests for day 13 junctions. I don't know how to go about them. :O 18:05
snarkyboojum: I'm reading the blog post over and over trying to comprehend it...
snarkyboojum isBEKaml: yeah, I took and look and passed over that one :) 18:06
TimToady std: {my $foo;say $^foo}
p6eval std 30750: OUTPUT«Use of uninitialized value $decl in concatenation (.) or string at STD.pm line 101334.␤===SORRY!===␤$foo has already been used as a non-placeholder in the surrounding block,␤ so you will confuse the reader if you suddenly declare $^foo here at /tmp/iHQIFAYRNC
..lin…
snarkyboojum maybe moritz_++ can give us an idea :)
18:07 whiteknight joined
snarkyboojum isBEKaml: how are junctions tested in the current spec test suite? 18:07
pugssvn r30752 | lwall++ | [STD] suppress uninit warning
isBEKaml snarkyboojum: that's what I was thinking. Maybe, I could take some ideas from the tests in the current spectest suite.. 18:08
TimToady std: class { my $foo; say $^foo }
p6eval std 30750: OUTPUT«===SORRY!===␤Placeholder variable $^foo may not be used here because the surrounding package block takes no signature at /tmp/67fwmKu8AF line 1:␤------> class { my $foo; say $^foo⏏ }␤Check failed␤FAILED 00:01 111m␤»
snarkyboojum isBEKaml: might be worth looking in there for inspiration yeah
TimToady std: sub { my $foo; say $^foo }
p6eval std 30750: OUTPUT«===SORRY!===␤$foo has already been used as a non-placeholder in the surrounding sub block,␤ so you will confuse the reader if you suddenly declare $^foo here at /tmp/CeYwMi4Qgs line 1:␤------> sub { my $foo; say $^foo⏏ }␤Check failed␤FAILED
..00:…
TimToady std: sub { my $foo; { say $^foo } }
p6eval std 30750: OUTPUT«ok 00:01 111m␤»
TimToady std: sub { my $foo; { $foo; say $^foo } } 18:09
p6eval std 30750: OUTPUT«Use of uninitialized value $decl in concatenation (.) or string at STD.pm line 101334.␤===SORRY!===␤$foo has already been used as a non-placeholder in the surrounding block,␤ so you will confuse the reader if you suddenly declare $^foo here at /tmp/hU9iBub6KH
..lin…
isBEKaml TimToady: message is part funny.. TimToady++
lue TimToady: short walk
diakopter long pier?
TimToady ENOTYET
isBEKaml TimToady: "so you will confuse the reader if you suddenly declare $^foo here"
TimToady Secundus called upon the matriarch
TimToady curtsies 18:10
diakopter hopes for a green All Tests Successful
TimToady std: sub { -> $^foo { } } 18:11
p6eval std 30750: OUTPUT«===SORRY!===␤You may not use the ^ twigil in a signature at /tmp/LY51rGuDDX line 1:␤------> sub { -> $⏏^foo { } }␤ expecting twigil␤Parse failed␤FAILED 00:01 112m␤»
dalek meta: r296 | diakopter++ | trunk/Sprixel/ (3 files):
[perlesque] temp bandaids (aren't they all)
isBEKaml snarkyboojum: I passed over Day 10 too. it was too broad..
TimToady std: sub { -> $x where $^foo { } }
p6eval std 30750: OUTPUT«Use of uninitialized value $decl in concatenation (.) or string at STD.pm line 101310.␤===SORRY!===␤Placeholder variable $^foo is not allowed in the signature at /tmp/t3wqVJSH8B line 1:␤------> sub { -> $x where $^foo⏏ { } }␤Unexpected block in
..…
isBEKaml snarkyboojum: in scope...
TimToady std: sub { method ($x where $^foo) { } }
p6eval std 30750: OUTPUT«===SORRY!===␤Placeholder variable $^foo is not allowed in the method signature at /tmp/SyG9C8trvj line 1:␤------> sub { method ($x where $^foo⏏) { } }␤Check failed␤FAILED 00:01 112m␤»
TimToady std: class { $^huh } 18:12
p6eval std 30750: OUTPUT«===SORRY!===␤Placeholder variable $^huh may not be used here because the surrounding package block takes no signature at /tmp/IIiNWgFR0T line 1:␤------> class { $^huh⏏ }␤Check failed␤FAILED 00:01 111m␤»
rindolf TimToady: BTW, do you recall saying "It's easier to port a shell than a shell script."? 18:13
TimToady yes
rindolf TimToady: ah.
TimToady: was it an original sentiment? 18:14
TimToady afaik
rindolf TimToady: OK, thanks.
rokoteko what does $^var mean in perl6?
TimToady a self-declaring parameter
{ $^x + $^y } is the same as -> $x, $y { $x + $y } 18:15
rokoteko what does that mean? like $foo = "bar" without my in perl5?
TimToady it's more like a 'my-but-really-I'm-a-parameter-in-the-surrounding-block'
std: -> $x { $^y } # should fail 18:16
p6eval std 30750: OUTPUT«===SORRY!===␤Placeholder variable $^y cannot override existing signature ( $x ) at /tmp/deGfaJ2hA3 line 1:␤------> -> $x { $^y⏏ } # should fail␤Check failed␤FAILED 00:01 112m␤»
rokoteko what is its scope then?
rindolf en.wikiquote.org/wiki/Larry_Wall - hmmm... this shell script quote was removed from here.
TimToady attaches to the surrounding curlies
18:16 Su-Shee joined
rokoteko TimToady: ahh.. so it is OUTSIDE the {}-block.. hmm. weird. 18:17
TimToady well, if they attribute it to someone else, they were probably quoting me
well, the weird thing is that parameters in general may be declared outside their lexical scope 18:18
but as far as the scope is concerned, they're inside the curlies
rokoteko { local ($x, $y) { $x + $y } } .. trying to translate it to perl5.
TimToady and the scope just thinks that parameters are 'my' vars
rokoteko TimToady: are you always correct? 18:19
TimToady no, more like sub { my ($x,$y) = @_; ... }
I'm always right, even when I'm not correct :)
rokoteko Im usually left.
TimToady sometimes I get over there too
but generally I'm a rabid moderate... 18:20
rokoteko it's the dude who can switch the north and south pole?
I dont completely get the point of $^x .. can anyone illustrate it more? 18:21
isBEKaml rakudo: my $tt = any(1|2|3); $tt!eigenstates.perl.say;
p6eval rakudo 9a15b8: OUTPUT«Method '!eigenstates' not found for invocant of class 'Integer'␤current instr.: '!DISPATCH_JUNCTION_METHOD' pc 17399 (src/builtins/Junction.pir:30)␤»
TimToady I think it's actually a girl superhero named Fluid Dynamo
are you familiar with Perl 5's sort? 18:22
sort { $a <=> $b } @list
for instance
rokoteko yes I am 18:23
TimToady in p5, that's hardwired behavior
diakopter phenny: tell pmurias I fixed the problems
phenny diakopter: I'll pass that on when pmurias is around.
TimToady in p6, it falls out naturally
std: sort { $^a <=> $^b }, 1,2,3
p6eval std 30750: OUTPUT«ok 00:01 112m␤»
TimToady which really means
rokoteko TimToady: sort { $^var1 cmp $var^2 } is valid?
ah. $^var2
TimToady std: sort { $^var1 <=> $^var2 }, 1,2,3 18:24
p6eval std 30750: OUTPUT«ok 00:01 112m␤»
TimToady placeholders are always assumed to be alphabetically ordered
so that's the same as
rokoteko I have no idea why on the eart the p6eval bot outputs that mucs utf8.
makes it difficult to read for humans.
TimToady std: sort -> $var1, $var2 { $var1 <=> $var2 }, 1,2,3 18:25
p6eval std 30750: OUTPUT«ok 00:01 113m␤»
TimToady rokoteko: most of us have irc clients that handle utf8 just fine
which client are you using?
isBEKaml what are eigenstates in here? 18:26
TimToady (one of our lesser goals is to drag people kicking and screaming into the age of Unicode :)
rokoteko irssi, but Im finnish. we're kinda stuck with latin1, or weird.
TimToady latin1 is a subset of unicode
rindolf rokoteko: you're still using Latin1 in .fi?
rokoteko rindolf: yes I am.
rindolf rokoteko: ah.
rokoteko: I hope not on IRC.
isBEKaml rakudo: sub foo(Any $temp) { $tt!eigenstates.perl.say; } foo(any(1,2,3)); 18:27
p6eval rakudo 9a15b8: OUTPUT«Confused at line 11, near "sub foo(An"␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
TimToady maybe start an alternate irssi with unicode
rokoteko rindolf: I type in ascii where the charset matters.
TimToady we do a *lot* of unicode in here
rindolf rokoteko: www.shlomifish.org/art/slogans/give...-death.svg
isBEKaml rakudo: sub foo(Any $temp) { $temp!eigenstates.perl.say; } foo(any(1,2,3)); 18:28
p6eval rakudo 9a15b8: OUTPUT«Confused at line 11, near "sub foo(An"␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
lue rakudo: say "hello" for ^3
p6eval rakudo 9a15b8: OUTPUT«hello␤hello␤hello␤»
dalek ok: eb8fadc | jonathan++ | src/subs-n-sigs.pod:
Describe captures.
 18:29
TimToady including using the replacement char for newline :)
lue rakudo: say ("hello" for ^3) ~~ "hello\nhello\nhello\n";
p6eval rakudo 9a15b8: OUTPUT«get_string() not implemented in class 'ArrayIterator'␤current instr.: 'perl6;Str;ACCEPTS' pc 13125 (src/builtins/Parcel.pir:161)␤»
Su-Shee rokoteko: if I can unicode you can. those handful of special chars in finnish.. ;)
rokoteko TimToady: do you know Archibald Higgings?
lue rakudo: say Str("hello" for ^3) ~~ "hello\nhello\nhello\n"; 18:30
p6eval rakudo 9a15b8: OUTPUT«Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 11␤current instr.: 'perl6;Regex;Cursor;FAILGOAL' pc 1696 (ext/nqp-rx/src/stage0/Regex-s0.pir:932)␤»
rokoteko Su-Shee: uni is dream in finnish. :)
or vice versa.
TimToady er, not sure...I'm terrible with names, even among friends. :)
lue argh.. what be wrong?
rakudo: say ("hello" for ^3).Str ~~ "hello\nhello\nhello\n";
TimToady every time I remember a name I forget a Japanese verb...
p6eval rakudo 9a15b8: OUTPUT«Method 'Str' not found for invocant of class 'ArrayIterator'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
rindolf TimToady: heh. 18:31
lue then stop remembering names :)
rindolf TimToady++
lue [o wait, wrong priorities]
rindolf TimToady: are Japanese verbs inflected?
jnthn Is Japanese anything inflected? 18:32
TimToady rindolf: yes, but in a much regular way than most Indo-European languages
18:32 Exodist left
rokoteko TimToady: you were referring to Fluid Dynamo. to me that is the chick in : www.savoir-sans-frontieres.com/JPP/...ld_eng.pdf ( a comic about topology ) 18:32
TimToady s/much/much more/
lue rakudo: say ("hello" for ^3).perl
jnthn TimToady: Do you count "add this particle" as an inflection?
p6eval rakudo 9a15b8: OUTPUT«Method 'perl' not found for invocant of class 'ArrayIterator'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
lue rakudo: say ("hello" for ^3).WHAT
p6eval rakudo 9a15b8: OUTPUT«Method 'WHAT' not found for invocant of class 'ArrayIterator'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
rindolf TimToady: I see.
jnthn TimToady: Are are there "real" inflections that morph stuff?
lue Now _that's_ ironic.
jnthn *or are 18:33
lue the above code I entered (the .WHAT) should work, right?
jnthn Oh, wait...I guess Japanese does have real inflections on verbs, looking it up.
:-)
D'oh.
lue [last I heard, tone is important in 日本語] 18:34
rokoteko rakudo: my $x = *.(**2); say $x(1...5) # so how should this be done?
p6eval rakudo 9a15b8: OUTPUT«Method 'Num' not found for invocant of class 'Block'␤current instr.: 'perl6;Mu;' pc -1 ((unknown file):-1)␤»
moritz_ jnthn: did you get around to fix the multi/candidates bugs 18:35
rokoteko or it just a bug ?
lue rakudo: my @a = "hello" for ^3; say @a.WHAT
p6eval rakudo 9a15b8: OUTPUT«Array()␤»
lue rakudo: my @a = "hello" for ^3; say @a
TimToady lue: yes, in the sense that Japanese has "pitch accent", but it is not a tonal language
moritz_ general Whatever-currying is not yet implemented in Rakuo
p6eval rakudo 9a15b8: OUTPUT«hello␤»
lue rakudo: my @a = ("hello" for ^3); say @a 18:36
p6eval rakudo 9a15b8: ( no output )
lue ō.o
rindolf rakudo: my @a = "hello" for ^3; say @a.perl
TimToady basically, instead of accenting a syllable by stressing it, they accent it by dropping the pitch on the next syllable
p6eval rakudo 9a15b8: OUTPUT«["hello"]␤»
Su-Shee lue: in mandarin or cantonese you really have to learn "tones".
TimToady or between syllables, if you think of it that way
lue so, in Japanese, monotone would be very... interesting [based on my understanding] 18:37
TimToady rokoteko: I have no idea what you're trying to do
jnthn moritz_: No, been writing book instead. :-)
moritz_ jnthn: ah, I've just seen the commits on #perl6book
jnthn: btw I've made progress on match objects 18:38
TimToady lue: there are 25 or so different pitch accent systems in use in different parts of Japan, and "none" is one of them :)
moritz_ I've got named (non-quantified) captures working, without evil workarounds
lue rakudo: say ((say "hello") ~~ "hello")
p6eval rakudo 9a15b8: OUTPUT«hello␤0␤»
lue rakudo: say ((say "hello") ~~ "hello\n")
p6eval rakudo 9a15b8: OUTPUT«hello␤0␤» 18:39
isBEKaml moritz_: can you throw some light on what eigenstates are when it comes to junctions? I saw that on the spectest suite and tried searching up/down on S03, but for some reason my FF just hangs.. :(
rokoteko TimToady: I was expecting it to be the same as: @res = map { $_**2 } (1..5); # oh wait, I again put the three dots instead of 1..5
moritz_ isBEKaml: eigenstates are the values that make up junctions 18:40
TimToady you don't get a map for free like that, and 3 dots should also work
moritz_ in (1|2|3) the eigenstates are 1, 2 and 3
rokoteko three dots why? is "===" also perl6 syntax?
isBEKaml rakudo: my $foo = any(1,2,3); say $foo!eigenstates.perl;
p6eval rakudo 9a15b8: OUTPUT«Method '!eigenstates' not found for invocant of class 'Integer'␤current instr.: '!DISPATCH_JUNCTION_METHOD' pc 17399 (src/builtins/Junction.pir:30)␤»
isBEKaml rakudo: my $foo = any(1,2,3); say $foo.eigenstates.perl; 18:41
TimToady and in none(1,2,3), they're also 1,2,3, and have an exclusive relationship with the actual set involved
p6eval rakudo 9a15b8: OUTPUT«(1, 2, 3)␤»
moritz_ it shouldn't be a public method
isBEKaml moritz_: well, it looks like it..
jnthn moritz_: Ah, great.
TimToady the only junction that is anything like a set of its eigenstates is all(1,2,3)
jnthn moritz_: I need to go buy / eat noms now, but back later on. :-)
moritz_: Does the mob4 branch build against the same Parrot as current master? 18:42
lue sub marine($parameter) { say $parameter; }; marine("h3llo"); marine()
rakudo: sub marine($parameter) { say $parameter; }; marine("h3llo"); marine()
moritz_ jnthn: nope; I've rebased, and pushed as 'mob5'
p6eval rakudo 9a15b8: OUTPUT«h3llo␤Not enough positional parameters passed; got 0 but expected 1␤current instr.: 'marine' pc 194 (EVAL_1:79)␤»
TimToady any(1,2,3) is really *six* different sets
isBEKaml TimToady: so any(1,2,3) or none(1,2,3) don't take on all of their eigen states?
jnthn moritz_: ah, cool. I needn't go building an older Parrot hen. :-) 18:43
*then
TimToady and none(1,2,3) is an an infinite number of sets (as defined defined in the current universe of discourse) - 6
which is why we will allow only all() junctions to autoconvert to sets, if any do 18:44
18:44 rokoteko left
isBEKaml Ah, I see... 18:44
lue Would that go under aleph-zero or similar?
[none(1,2,3)]
TimToady well... 18:45
lue ∞-6=∞
TimToady rakudo: say "what's my universe?" if pi === none(1,2,3)
p6eval rakudo 9a15b8: ( no output )
TimToady hmm
rindolf rakudo: pi 18:46
p6eval rakudo 9a15b8: ( no output )
rindolf rakudo: say pi
p6eval rakudo 9a15b8: OUTPUT«3.14159265358979␤»
TimToady rakudo: say "what's my universe?" if pi == none(1,2,3)
p6eval rakudo 9a15b8: OUTPUT«what's my universe?␤»
TimToady right, rakudo is still busted in autothreading ===
lue so === is "too strict for true to be true"? :)
isBEKaml === is too strict against types. 18:47
pugssvn r30753 | snarkyboojum++ | [t/spec] Added initial tests for Perl 6 Advent Calendar Day 15: .pick your game
isBEKaml rakudo: say non(1,2,3).WHAT;
p6eval rakudo 9a15b8: OUTPUT«Could not find sub &non␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
isBEKaml rakudo: say none(1,2,3).WHAT;
p6eval rakudo 9a15b8: OUTPUT«Junction()␤»
TimToady but to answer your question, pi is irrational, so I think it's at least א₁
lue I think I ought to set up a custom keymap for third/fourth/etc levels and 3button emulation (so on) 18:49
isBEKaml TimToady: I haven't got my head around autothreading though. And, it's atleast - sorry, I can't read that character. :(
lue Because I want to be able to type both an aleph and an o w/ umlaut :)
Su-Shee lue: just input the unicode code point.
isBEKaml .u aleph
phenny isBEKaml: Sorry, no results for 'aleph'.
isBEKaml .u umlaut
phenny isBEKaml: Sorry, no results for 'umlaut'.
isBEKaml :( 18:50
Su-Shee I mean in your term/irc client/editor and so on. :)
lue hah! If only KDE would do that.... :(
.u alef
phenny U+2135 ALEF SYMBOL (ℵ)
lue .u diaresis
phenny lue: Sorry, no results for 'diaresis'.
TimToady rakudo: my $x = { map { $^x * 2}, @_ }; say $x(1..5); 18:51
p6eval rakudo 9a15b8: OUTPUT«246810␤»
TimToady rakudo: my $x = { map { $^x * 2}, @_ }; say $x(1...5);
p6eval rakudo 9a15b8: OUTPUT«246810␤»
BrowserUk -?
moritz_ wonders if BrowserUk knows greetings that contain letters
isBEKaml is glad BrowserUk typed two chars instead of one. ;) 18:52
lue !(~~>] 18:53
TimToady phenny: tell rokoteko you want something like: my $x = { map { $^x * 2}, @_ }; say $x(1..5);
phenny TimToady: I'll pass that on when rokoteko is around.
isBEKaml moritz_: I find Junctions only take eigenstates as a private method. But I find them on p6eval to be publc. Was that changed recently? 18:55
TimToady phenny: tell rokoteko Yes, ... is a dwimmy "series" operator, while .. is just interval ranges; === is exact type and value equivalence, while == is forced numeric as in p5 18:56
phenny TimToady: I'll pass that on when rokoteko is around.
Juerd .u diaeresis
phenny U+00A8 DIAERESIS (¨)
Juerd lue: Correct spelling helps :)
lue thank you. I can never remember that. Is it supposed to be english?
moritz_ isBEKaml: "recently", yes
like, within in the last two years :-)
Juerd lue: As English as most of the language... that is: not really 18:57
isBEKaml moritz_: wow, spec test methods need to reflect that, then.. :-)
lue I think it's weird. I prefer "umlaut" :)
moritz_ isBEKaml: don't they?
Juerd lue: Umlaut is a diaeresis used for a specific purpose.
isBEKaml moritz_: No..
moritz_: t/spec/S03-junctions/eigenstates.t 18:58
moritz_ this test... sucks.
Juerd lue: Umlaut is not correct if the ¨ is used to separate syllables, breaking what would otherwise be a diphthong 18:59
TimToady you aren't supposed to be able to use the bare !method syntax outside of the class in question
18:59 pmurias joined
TimToady you have to say $x!ThatClass::method 19:00
diakopter pmurias: say something
isBEKaml TimToady: yes, the spec test suite actually violated them.
pmurias diakopter: hi
phenny pmurias: 17:48Z <diakopter> tell pmurias I fixed the problem(s) there, mostly.
pmurias: 18:02Z <diakopter> tell pmurias well, I guess I broke other things. :/
pmurias: 18:23Z <diakopter> tell pmurias I fixed the problems
diakopter :)
TimToady walking to lunch on Castro &
pugssvn r30754 | moritz++ | remove a test file that is mostly wrong, isBEKaml++ for pointint it out
pmurias diakopter: what was the source of the bug? 19:01
isBEKaml moritz_: fgrepping revealed one more file in the same directory. misc.t 19:02
rindolf How would you write a cumulative average of fields one-liner in p6? In Perl 5 we got it down to «(echo '1 1 1 2' ; echo '3 3 3 4') | perl -lape '$f=0;$s[$f++]+=$_ for@F;$_="@{[map$_/$.,@s]}"'» 19:03
isBEKaml moritz_: lines 391, 392
diakopter pmurias: well, you can't declare a class P6object in that test when one already exists 19:04
19:04 envi^home left
diakopter pmurias: (the one you wrote in C#) 19:04
TimToady Amarin Thai, probably & # really gone 19:05
isBEKaml moritz_: :391,404 d
lue well, I've seen an "umlaut" used like that, but I believe you.
pugssvn r30755 | moritz++ | [t/spec] remove another wrong/confusing test
pmurias diakopter: i'll remove the C# versions of the p6 classes then 19:07
diakopter pmurias: ok :)
pmurias diakopter: and why did the test pass when commenting out some of the methods?
diakopter I didn't know it did
moritz_ rakudo: class A is Array is Hash { method push($x) { self.Array::push($x) } }; my $x = A.new(); A.push(3); say A 19:08
p6eval rakudo 9a15b8: OUTPUT«Type objects are abstract and have no attributes, but you tried to access ␤current instr.: 'perl6;Seq;!fill' pc 14888 (src/builtins/Code.pir:161)␤»
diakopter pmurias: which methods
moritz_ rakudo: class A is Array is Hash { method push($x) { self.Array::push($x) } }; my $x = A.new(); $x.push(3); say $x.[0]
p6eval rakudo 9a15b8: OUTPUT«3␤»
pmurias diakopter: the last one
diakopter I don't know; it shouldn't have
pmurias diakopter: but i don't think it's important to understand that 19:09
diakopter didn't work for me when I did that just now
pmurias why does class Z {};class X is Z { has List<Z> $positionals} fail? 19:11
diakopter I dunno; i'll try it
pmurias List[Z]
fixes 19:12
diakopter oh, yeah :D
pmurias: oh btw, should I make an Array type?
moritz_ mixing syntaxes considered harmful
diakopter which were mixed?
oh, parametric roles syntax for parametric types 19:13
pmurias patametric roles vs templates
moritz_ C++ templates
pmurias C++ considered evil
diakopter that's just C#
different from C++ templates
moritz_ C# cripped it from C++ 19:14
mathw C# generics are rather different to C++ templates
diakopter (ditto)
moritz_ talked only about the syntax
diakopter but still, p6 doesn't have parametric/generic classes (booo)
pmurias it has parametric roles 19:15
diakopter right, but that doesn't help :P since perlesque won't have roles
too complex for that layer.
pmurias perlesque shouldn't even have syntax 19:16
diakopter ?
it shouldn't be a language?
pmurias IMHO it should accept AST in JSON 19:17
diakopter I thought we talked through this and you agreed that was impossible?
besides, how would it translate anything? 19:18
who will write the "translate this type of AST node to this" translation routines 19:19
show me [an example of or the spec of] the AST API the JSON would match against 19:20
VAST?
if it's VAST, that's nowhere close to low-level enough to translate trivially to CIL. 19:21
ajs Does rakudo just not support BUILD submethods yet, or is that no longer how we define signature-specific initializers for classes?
moritz_ ajs: it never was
ajs oh, I must have misread 19:22
moritz_ ajs: perlgeek.de/blog-en/perl-6/object-c...ation.html
ajs So I should be defining my own new, then?
moritz_ yes
ajs hmm... and here I thought blessing was so Perl 5 ;-) 19:23
moritz_ ajs: Perl 6 still uses some of the Perl 5 ideas under the hood
19:24 rindolf left
lue rakudo: %nonfoos{"a" => 1, "b" => 2}; say (%nonfoos.keys.fmt("'%s'", ', ')).WHAT 19:25
p6eval rakudo 9a15b8: OUTPUT«Symbol '%nonfoos' not predeclared in <anonymous>␤current instr.: 'perl6;PCT;HLLCompiler;panic' pc 152 (compilers/pct/src/PCT/HLLCompiler.pir:108)␤»
19:25 rindolf joined
lue rakudo: my %nonfoos{"a" => 1, "b" => 2}; say (%nonfoos.keys.fmt("'%s'", ', ')).WHAT 19:25
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'fmt'. Available candidates are:␤:(Mu : Str $format = { ... };; *%_)␤␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
lue ō.o
Why did that fail? 19:26
moritz_ lue: did you mean to assign to that variable?
lue I meant to put something in the hash, yes.
moritz_ then do it
you know, assignment with the = operator 19:27
lue I thought hashes were special (don't use them often)
rakudo: my %nonfoos = {"a" => 1, "b" => 2}; say (%nonfoos.keys.fmt("'%s'", ', ')).WHAT
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'fmt'. Available candidates are:␤:(Mu : Str $format = { ... };; *%_)␤␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
lue ō.ō 19:28
moritz_ rakudo: (1, 2).fmt('a', 'b')
p6eval rakudo 9a15b8: ( no output )
moritz_ rakudo: my %h; say %h.keys.WHAT
p6eval rakudo 9a15b8: OUTPUT«MapIterator()␤»
moritz_ ah
that's why
lue It's part of the advent calendar (2009 Day 9) 19:29
capturing named arguments (arbitrary amount)
19:31 Ross^ joined
lue rakudo: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys.fmt("'%s'", ', ') = ("a","b") 19:31
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'fmt'. Available candidates are:␤:(Mu : Str $format = { ... };; *%_)␤␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
lue alpha: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys.fmt("'%s'", ', ') = ("a","b")
p6eval alpha 30e0ed: ( no output )
moritz_ alpha: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys.fmt("'%s'", ', ')
p6eval alpha 30e0ed: ( no output ) 19:32
moritz_ alpha: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys
alpha: say 2
p6eval alpha 30e0ed: ( no output )
alpha 30e0ed: OUTPUT«2␤»
19:33 rindolf left
isBEKaml rakudo: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys; 19:33
p6eval rakudo 9a15b8: OUTPUT«ab␤»
lue old advent code then; should it still work? 19:34
19:34 finanalyst left 19:35 finanalyst joined, Ross^ left, Ross^ joined
moritz_ i think so 19:35
isBEKaml rakudo: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys>>.Str.fmt(;"'%s'", ', '); 19:36
p6eval rakudo 9a15b8: OUTPUT«Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 11␤current instr.: 'perl6;Regex;Cursor;FAILGOAL' pc 1696 (ext/nqp-rx/src/stage0/Regex-s0.pir:932)␤»
moritz_ lue: don't you want to submit a patch for rakudo that implements it?
isBEKaml rakudo: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys>>.Str.fmt("'%s'", ', ');
p6eval rakudo 9a15b8: OUTPUT«'a', 'b'␤»
moritz_ lue: likely it's straight forward to backport it from alpha
lue I can, but I'm writing the test file for advent day 9, maybe afterwards
isBEKaml lue: ^^ 19:37
19:37 jakk left, Ross^ left
lue I'll just put it in as-is. After all, it is a *test* I'm writing. 19:37
19:38 Ross^ joined
isBEKaml moritz_: does this still need a patch? >>.Str seems to do it... 19:38
moritz_ isBEKaml: hash .fmt'ing shoul still work 19:39
lue alpha: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys.fmt("'%s'", ', ').WHAT
p6eval alpha 30e0ed: ( no output )
isBEKaml alpha: my %nonfoos = {"a" => 1, "b" => 2}; say %nonfoos.keys.fmt("'%s'", ', ');
lue I'll assume it returns an array
19:40 finanalyst left
p6eval alpha 30e0ed: ( no output ) 19:40
isBEKaml huh? no output?
moritz_: you mean, .fmt should be able to stringify a MapIterator into a flat string? 19:41
moritz_ aye
19:43 Ross^ left
isBEKaml moritz_: .fmt currently seems to take a string and format it into the specified format. So, don't know how I'd do that otherwise. :) 19:45
moritz_ on everything listy it should take a separator too, I think 19:46
rakudo: say <1 2 3>.fmt('%02d', 'a') 19:47
p6eval rakudo 9a15b8: OUTPUT«01a02a03␤»
pmurias diakopter: re taking JSON i agreed that's propably a lot of work, and there are propably a lot of things in perlesque that are more worth doing
isBEKaml moritz_: I'm currently looking at EnumMap.pm to see how .fmt's implemented. It maps out the pairs to format according to the %format% given and joins them with the separator. 19:48
moritz_ snarkyboojum: is (@deck .= pick(*); @deck.elems), 4 * 13, 'Shuffled card deck'; 19:49
snarkyboojum: that's not working how you think it does 19:50
snarkyboojum moritz_: ok :)
19:50 jakk joined
snarkyboojum moritz_: I thought it was a dodgey way to test it anyway :) 19:50
lisppaste3 pmurias pasted "diakopter: why doesn't that work?" at paste.lisp.org/display/100307 19:51
pugssvn r30756 | snarkyboojum++ | [t/spec] Added initial tests for Perl 6 Advent Calendar
pmurias diakopter: why does that cause an error
?
moritz_ snarkyboojum: it's better to do the @deck .= pick(*) before the line with the is() 19:52
snarkyboojum moritz_: ok - will fix
moritz_ snarkyboojum++
diakopter pmurias: b/c the expression parser cheats a lot
as in, doesn't exist
you need parens
(self.positionals).Add($pos); 19:53
snarkyboojum moritz_++ # thanks for reviewing the tests :)
diakopter pmurias: at first glance.
19:53 Ross^ joined
lue hi snarky! Back from the Perl 6 golf course, eh? 19:54
diakopter pmurias: oh, I'm wrong; that's not it.
snarkyboojum moritz_: is that was you had in mind - gist.github.com/410321 ? 19:55
s/was/what/
moritz_ snarkyboojum: aye
lue ( must've seen a lot of .t's :) )
snarkyboojum k
lue: o/ 19:56
moritz_ rakudo: DEBUG(3)
p6eval rakudo 9a15b8: OUTPUT«Could not find sub &DEBUG␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
diakopter pmurias: hrm. grammar engine buglet. it'll take me awhile to trace it down 19:57
pugssvn r30757 | snarkyboojum++ | [t/spec] Cleaned up a test for Advent Calendar Day 5 in response to feedback from moritz_++ 19:59
isBEKaml moritz_: Alright, I'm tempted. How do I go about implementing this functionality ? ## Or, what should I know b4 I do this? 20:00
20:00 Ross^ left, Ross^ joined
diakopter pmurias: well, it works on its own.... odd. 20:01
oh
pmurias hm?
moritz_ isBEKaml: the sanest thing would be to convert it to an ordinary list (or maybe Seq) first, and then hand it over to the fmt method of that new list 20:02
rakudo: say %*ENV.keys.WHAT
p6eval rakudo 9a15b8: OUTPUT«MapIterator()␤» 20:03
moritz_ rakudo: say list(%*ENV.keys).WHAT
p6eval rakudo 9a15b8: OUTPUT«List()␤»
moritz_ rakudo: say list(%*ENV.keys).fmt('%s%s', '|-|')
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'fmt'. Available candidates are:␤:(Mu : Str $format = { ... };; *%_)␤␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
isBEKaml rakudo: say %*ENV.keys.List.WHAt;
p6eval rakudo 9a15b8: OUTPUT«Method 'List' not found for invocant of class 'MapIterator'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
isBEKaml rakudo: say %*ENV.keys.List.WHAT;
p6eval rakudo 9a15b8: OUTPUT«Method 'List' not found for invocant of class 'MapIterator'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
isBEKaml rakudo: my %nonfoos = {"a" => 1, "b" => 2}; say list(%nonfoos.keys).fmt("'%s'", ', '); 20:04
diakopter moritz_: to clarify earlier, perlesque uses the [], not the <>
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'fmt'. Available candidates are:␤:(Mu : Str $format = { ... };; *%_)␤␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
moritz_ diakopter: that's how I understood it, yes
diakopter oh ok
isBEKaml rakudo: my %nonfoos = {"a" => 1, "b" => 2}; say list(%nonfoos.keys)>>.fmt("'%s'", ', ');
p6eval rakudo 9a15b8: OUTPUT«No applicable candidates found to dispatch to for 'fmt'. Available candidates are:␤:(Mu : Str $format = { ... };; *%_)␤␤current instr.: '!dispatch_method_parallel' pc 357 (src/glue/dispatch.pir:71)␤»
isBEKaml moritz_: EnumMap.pm ?
pmurias diakopter: other annoying thing why do i have to declare the return type of a constructor? 20:05
moritz_ huh?
lue How do I evaluate a piece of code as being ok in a test?
moritz_ isBEKaml: sorry, I'm a bit distracted at the moment
what do you mean with "as ok"?
diakopter pmurias: it's on the list to fix.. 20:06
the mental list
afk&
lue test to see if it works, such as this [which should fail]: 20:07
sub f(Array @a){...}
moritz_ dies_ok { code }, 'description';
like this?
lue And then for a success, it's something like eval_ok ?
moritz_ lives_ok
isBEKaml lue: for more, look at Test.pm. Easy to read (I don't know p6, either) 20:08
:)
lue thank you.
20:09 Ross^ left
snarkyboojum shuteye& 20:09
20:09 Ross^ joined
lue rakudo: sub f(Array @a) { say @a;}; say f((1,2,3)) 20:10
p6eval rakudo 9a15b8: OUTPUT«Nominal type check failed for parameter '@a'; expected Positional[Array] but got Parcel instead␤current instr.: 'f' pc 203 (EVAL_1:78)␤»
lue rakudo: sub f(Array @a) { say @a;}; my @b = 1,2,3; f(@b)
p6eval rakudo 9a15b8: OUTPUT«Nominal type check failed for parameter '@a'; expected Positional[Array] but got Array instead␤current instr.: 'f' pc 242 (EVAL_1:22837788)␤»
moritz_ that's correct 20:11
isBEKaml moritz_: is there a way to know where some method comes from ? I tried deliberately introducing an error for a stacktrace, but the unknown file always gets me off.. :( 20:12
20:12 Ross^ left
BrowserUk ? 20:12
moritz_ isBEKaml: grep/ack
20:13 Ross^ joined
BrowserUk Hey moritz 20:13
moritz_ hi BrowserUk
Ross^ Toallchannels:Sorry for the annoying reconnect messages, my net connection is through a server which is currently having it's file server module configured... and it's messing stuff up.
BrowserUk phenny: tell pmurias, do you have a link to the ebook version of Erlang Programming? 20:14
phenny BrowserUk: I'll pass that on when pmurias is around.
isBEKaml moritz_: and some fore-knowlege about how such classes are related. :) This got me pretty recently, when I tried to search for Array's private Fill method. Turned out in Seq, but didn't know that Arrays are essentially Seqs. (colomon++) 20:15
moritz_ ack '(sub|method).*!fill\b' 20:16
finds src/builtins/Seq.pir pretty quickly
isBEKaml moritz_: I'm using find . -type f -exec fgrep fmt {} + 20:17
:)
20:17 kcwu left
moritz_ isBEKaml: cpan -i App::Ack # welcome to the 21st century 20:17
use perl 5 regexes for searching, get decent hilighitng, automatic exclusion of .svn files etc. 20:18
lue Oi! Has the spec changed on this bit of code (from Day 9):
sub f(:$a, :$b) {...};
when the advent was written, this automagically makes every forced named parameter optional 20:19
jnthn That's still troo.
pmurias BrowserUk: a legal one?
phenny pmurias: 20:14Z <BrowserUk> tell pmurias do you have a link to the ebook version of Erlang Programming?
jnthn rakudo: sub f(:$a, :$b) { say "look I'm ok!!!" }; f
p6eval rakudo 9a15b8: OUTPUT«look I'm ok!!!␤»
[Coke] ʘ‿ʘ 20:20
20:20 meteorja1 left
lue AAH! 20:20
20:21 pmurias left, pmurias_ joined, pmurias_ is now known as pmurias
lue Oh, wrong test I thought keeled :) . Guess I need *eval*_dies_ok then. 20:21
20:21 rv2733 left
pmurias BrowserUk: i think i found it by typing 'rapidshare Erlang Programming' into google 20:23
isBEKaml moritz_: I knew about ack, faster search, better regex support. ## err...
lue Is it alright to fudge part of an advent test? [something doesn't work in rakudo, and I don't want to fix it right now] 20:28
[alternatively, I can wait to upload the test]
moritz_ sure, fudge it
20:31 Charlie joined, Charlie is now known as Guest12194 20:32 Guest12194 is now known as Ross, Ross^ left, Ross left, Ross joined 20:34 masonkramer joined, meteorjay joined
pugssvn r30758 | lue++ | [t/spec] Updated Day 9 Advent test to actually test Day 9 bits of code 20:36
isBEKaml rakudo: my $temp = any(3,2,5); my $another = $temp * 2; say $another.eigenstates; 20:38
p6eval rakudo 9a15b8: OUTPUT«6410␤»
lue Dankon, kaj FDK [later I might feel like undfudging Day 9 Advent :)] 20:42
.btw(FDK = For de Klavaro)
20:42 meteorjay left 20:46 cognomore joined
isBEKaml rakudo: (3&5&8).WHAT.perl.say; 20:50
p6eval rakudo 9a15b8: OUTPUT«Junction␤»
moritz_ EWRONGFILENAMEINBACKTRACE 20:54
isBEKaml huh?
moritz_ in some local back traces
current instr.: 'perl6;Seq;!fill' pc 14956 (src/builtins/Code.pir:161) 20:55
ajs That should probably have its own error code ;-) EEWRONGFILENAMEINBACKTRACEINBACKTRACE
jnthn That's...very wrong. :-| 20:56
isBEKaml :)
moritz_ there's not a single occurence of !fill in Code.pir
ajs Is there a way to turn on some kind of debugging tracing when a regex is executing? 20:58
moritz_ there is... and i forgot how
rakudo: grammar A { token TOP { <?DEBUG(1)> <foo> }; token foo { .+ } }; A.parse('blah') 20:59
p6eval rakudo 9a15b8: OUTPUT«0/1: START foo␤0/1: PASS foo at pos=4␤0/1: PASS TOP at pos=4␤»
moritz_ ah
20:59 zorgnax joined
moritz_ that's how you do it :-) 20:59
ajs oh nice
moritz_ (it's not specced
jnthn But it is useful. :-) 21:00
ajs I'll take it ;)
jnthn has found and fixed bugs with that
zorgnax how is STD.pm6 used? How do you make something to parse Perl6 in Perl6? 21:01
moritz_ 1) you write a grammar 2) you write a script that translates it to another language 3) you run said script 4) ... 5) PROFIT!
jnthn :-) 21:02
zorgnax so its just a way to specify what the language is, I cant run it yet
moritz_ sure you can
ajs * step 4 may involve time travel, rendering steps 1 and 2 hypothetical
moritz_ std: <a b c>
p6eval std 30758: OUTPUT«ok 00:01 114m␤»
moritz_ std: <a b c 21:03
isBEKaml as for time travel, have a look at tardis!
p6eval std 30758: OUTPUT«===SORRY!===␤Unable to parse quote words at /tmp/8nJmbjYIKC line 1:␤------> <⏏a b c␤Couldn't find final '>'; gave up at /tmp/8nJmbjYIKC line 1 (EOF):␤------> <a b c⏏<EOL>␤Parse failed␤FAILED 00:01 110m␤»
jnthn STD.pm6 gets translated to Perl 5. Rakudo also has a grammar that is a lot like STD an converging with it.
Rakudo's grammar engine is bootstrapped - it parses itself.
moritz_ has been chasing bugs in the wrong method for the last two hours
jnthn moritz_: Oh, ouch. :-( 21:04
zorgnax my heads going to explode
moritz_ my $x = 3; Q:PIR { $P0 = find_lex '$x'␤$P2 = box 3␤ $P0[0] = $P2 } 21:05
rakudo: my $x = 3; Q:PIR { $P0 = find_lex '$x'␤$P2 = box 3␤ $P0[0] = $P2 }
p6eval rakudo 9a15b8: OUTPUT«set_pmc_keyed() not implemented in class 'Integer'␤current instr.: '_block14' pc 29 (EVAL_1:0)␤»
isBEKaml what does it mean when I get "unknown file" in a stacktrace? I mean, where do I look for that concerned method implementation? ## I dont' want to grep...
bkeeler Hallo folks!
moritz_ isBEKaml: it means that parrot screwed up. Or rakudo, or boh :( 21:06
jnthn hi bkeeler!
moritz_ \o/
21:07 gbacon left
jnthn moritz_: Are you attached to what's in the "=head1 Unpacking" section of subs-n-sigs.pod? 21:07
moritz_: I find it kinda confusing-ish.
moritz_ jnthn: feel free to improve it any way you like 21:08
(including ripping it out :)
jnthn moritz_: OK, will try and do so :-)
moritz_: After that, we're much of the way through having subs-n-sigs.pod in decent shape.
dalek ok: e4f473a | jonathan++ | src/subs-n-sigs.pod:
A little explanation about why signature introspection might be useful, to
isBEKaml moritz_: I don't understand it at all. I see core.pir, run.pir - unknown file and HLLCompiler lower level gobbledygook... :(
ajs Is $/.keys supposed to work? Am I missing something? I'm getting no such method for invocant of class 'Regex;Match'... shouldn't that be the names of the named matches?
jnthn ajs: It's a bug in present Rakudo. 21:09
ajs Oh, OK. Cool. As long as it's not me misunderstanding.
jnthn No, it should work.
bkeeler jnthn: do you have any idea why the following doesn't work? 21:10
rakudo: class C does Associative does Positional { }
p6eval rakudo 9a15b8: OUTPUT«get_attr_str() not implemented in class 'Sub'␤current instr.: 'perl6;Code;multi' pc 13975 (src/builtins/Num.pir:234)␤»
moritz_ ajs: it's something I'm working on as we speak
jnthn bkeeler: erm
bkeeler: wtf :-/
moritz_ but it's not trivial
jnthn bkeeler: I don't know why.
bkeeler: But I'm a tad surprised. 21:11
rakudo: class C does Associative { }
p6eval rakudo 9a15b8: ( no output )
jnthn rakudo: class C does Positional { }
ajs oh sweet. That would really help. For some reason, that debug pragma didn't work for me, so I'm walking the match results.
p6eval rakudo 9a15b8: ( no output )
jnthn rakudo: role A { }; role P { }; class C does A does P { }
p6eval rakudo 9a15b8: ( no output )
bkeeler rakudo: class C does Associative ; use MONKEY_TYPING; augment class C does Positional { }
p6eval rakudo 9a15b8: OUTPUT«get_attr_str() not implemented in class 'Sub'␤current instr.: 'perl6;Code;multi' pc 13975 (src/builtins/Num.pir:234)␤»
moritz_ ajs: that's a pain as-is, I know
21:11 whiteknight left
jnthn oh 21:11
bkeeler: I know
bkeeler: Bet it's an error reporting FAIL. 21:12
bkeeler This is the roadblock for making Capture do Positinal and Associative
jnthn bkeeler: Or related fail.
ajs no worries. I'm thick-skinned when it comes to debugging
jnthn rakudo: class C does Associative does Positional { method of() { Mu } }; say "lived";
p6eval rakudo 9a15b8: OUTPUT«get_attr_str() not implemented in class 'Sub'␤current instr.: 'perl6;Code;multi' pc 13975 (src/builtins/Num.pir:234)␤»
jnthn ah, it's deeper
Damm. 21:13
bkeeler confused jnthn! o/
jnthn :-P
moritz_ $ ./perl6 -e '"ab0" ~~ /(.)/; say $/{0}'
a
jnthn ergh
moritz_ parrot emits a set_pmc_keyed for boths $P1[$I0] = $P0 and $P1[$S0] = $P0 21:14
ajs I really like vim, but I wish it would stop underlining code as a means of highlighting... my_sub_match_path() is kind of hard to read
moritz_ ajs: it depends on the color scheme, iirc
arnsholt Yeah, my vim doesn't do that either
moritz_ ajs: "my" vim doesn't do that
bkeeler I don't have that problem with vim
arnsholt Of course, I'm boring and use almost entirely default settings 21:15
ajs Yeah, and for some reason I'm trying to do this from Windows over putty instead of in my Ubuntu VM. I think I'm just dazed by too much sun today.
isBEKaml ajs: give :nohlsearch
that should fix it..
jnthn moritz_: Oh? That's...unexpected.
ajs ah thanks isBEKaml
jnthn bkeeler: ah 21:16
I think I know what's up with it
21:16 cognomore left
moritz_ jnthn: irclog.perlgeek.de/perl6/2010-05-22#i_2358731 yes, set_pmc_keyed. sigh. 21:19
jnthn moritz_: I'm...a tad perplexed by that.
moritz_: oh 21:20
moritz_: Actually, that's not the case, I don't think
(more)
moritz_ or am I doing something execptionally stupid there?
jnthn Parrot has fallbacks for set_pmc_keyed_int and set_pmc_keyed_str. If they don't exist, it tries to box the arguemnt and delegate to set_pmc_keyed. And that then throws the error.
21:21 dual joined
jnthn So it almost certainly is calling set_pmc_keyed_int first, it's just that Parrot "helpfully" tires to delegate to set_pmc_keyed 21:21
See src/pmc/default.pmc - I get it's in there.
moritz_ thanks
jnthn np
Sorry, I missed the detail the first time I glanced at the Rakudo output.
moritz_ no problem 21:22
jnthn bkeeler:
> class C does Positional does Associative { }
Method 'of' collides and a resolution must be provided by the class
> class C does Positional does Associative { method of() { Mu } }
>
\o/
21:26 pmurias left
isBEKaml rakudo: my $fname; regex namedreg { $fname = (.*)\.lname }; "firstname.lname" ~~ /<namedreg>/; say $/<namedreg><fname>; 21:28
p6eval rakudo 9a15b8: OUTPUT«Malformed regex at line 11, near "namedreg {"␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
moritz_ bkeeler: btw I'm now in the 'mob5' branch, where I don't need ugly workarounds for named captures 21:29
isBEKaml rakudo: my $fname; regex namedreg { $fname = (.*).lname }; "firstname.lname" ~~ /<namedreg>/; say $/<namedreg><fname>;
p6eval rakudo 9a15b8: OUTPUT«Malformed regex at line 11, near "namedreg {"␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
isBEKaml what's wrong? named captures don't work?
moritz_ isBEKaml: looking up named regexes from lexical scopes doesn't work 21:30
jnthn bkeeler: Patch is spectesting now, so will push it soon.
21:30 gbacon joined
isBEKaml moritz_: oh.. :( 21:32
moritz_ wow, adding {set,get}_pmc_keyed_int to Mu.pir causes new dimensions of fails
ajs rakudo: my %x; say %x<apple> :exists;
p6eval rakudo 9a15b8: OUTPUT«Confused at line 11, near "say %x<app"␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
ajs is that not the exists syntax?
jnthn ajs: No adverbs on postcircumfixes yet; in Rakudo use %x.exists('apple') 21:33
ajs oh, cool
thanks
jnthn (In fact, no operator adverbs in general.)
21:34 zorgnax left
jnthn rakudo: // 21:36
p6eval rakudo 9a15b8: OUTPUT«Null regex not allowed at line 11, near ""␤current instr.: 'perl6;HLL;Grammar;panic' pc 501 (ext/nqp-rx/src/stage0/HLL-s0.pir:327)␤»
dalek ok: 82782ca | moritz++ | src/subs-n-sigs.pod:
[subs] remove duplicated words
BrowserUk phenny: seen ruoso?
21:36 moritz_ sets mode: +v dalek 21:40 jakk left
isBEKaml is there any way I can test out my changes without having to build? 21:40
bkeeler jnthn: \o/
moritz_: also \o/
isBEKaml for instance, any changes i make to Seq.pm goes into core.pm. So I have to build.. :(
bkeeler Breaking up the core.pm build would be nice. People have tried before though 21:41
moritz_ isBEKaml: that's how software development works: you change code, you compile it, you test it
isBEKaml moritz_: compile just one, not all. dev switches could be made available? 21:42
moritz_ isBEKaml: compiling the setting separately is non-trivial
bkeeler I think I'm going to go ahead and push my nqp changes. pmichaud's had a chance to review if he wanted to...
isBEKaml moritz_: will have to learn about that some other time. It's over 3 am and I'm a bit dazed.. :) 21:44
moritz_ isBEKaml: I haven't learned about it myself :/
isBEKaml: I've tried to implement it once, and failed 21:45
diakopter bkeeler: somehow I doubt you're correct. 21:46
bkeeler diakopter: about what?
diakopter the last thing you said.
bkeeler diakopter: Oh, about pmichaud's tuit supply
21:46 ShadeHawk joined
ajs I have an object $x and a method name as a string $y. I tried $x.$y with no luck. What's the right way to invoke $y on $x? 21:47
bkeeler diakopter: I suppose you're right :/
ShadeHawk After reading perlgeek.de/en/article/5-to-6 I have one question: how to create empty list, and one-element list in Perl 6?
jnthn $x."$y"()
ajs Ah, ok. thanks jnthn
jnthn $x.$y() always assumes $y is a code object 21:48
moritz_ ShadeHawk: () is an empty list
ShadeHawk: and (1,) a one-item list
rakudo: say (1,).elems 21:49
p6eval rakudo 9a15b8: OUTPUT«1␤»
jnthn Also, for one element, anything that puts a single element in list context will do that too.
dalek kudo: 1499f20 | jonathan++ | src/builtins/Code.pir:
Since methods in Code.pir may also get called on Parrot subs, we should be
ShadeHawk moritz_, jnthn: thanks
jnthn s/single element/single thing)
bkeeler Well, I pushed anway ;)
jnthn my @x = 1; # list context, so like my @x = (1,);
bkeeler: ;-)
bkeeler I'm keen to put the regex interpolation stuff to bed 21:50
jnthn bkeeler: It'd been through one round of feedback before now at least, so hopefully by now it's close enough. If not, we have version control.
bkeeler: Same.
dalek p-rx: f2c2f18 | (Bruce Keeler)++ | (5 files):
Add support for interpolation in regexes
p-rx: 96c2b36 | (Bruce Keeler)++ | (35 files):
Merge remote branch 'upstream/master' into regex-interpolation
p-rx: 3664226 | (Bruce Keeler)++ | src/PAST/Compiler-Regex.pir:
Make it work with Rakduo
p-rx: 0050310 | (Bruce Keeler)++ | src/PAST/Compiler-Regex.pir:
Removed debugging line
p-rx: bbaab49 | (Bruce Keeler)++ | src/HLL/Compiler.pm:
Merge remote branch 'upstream/master' into regex-interpolation
bkeeler jnthn: Yep. It's pretty much exactly what we agreed on. I was just being cautious in requesting the review, perhaps overly so 21:51
jnthn bkeeler: Well, requesting it and waiting a bit was good.
bkeeler Perhaps I should have folded it into one commit. Will have to work on my git-fu
I still don't grok rebase 21:52
moritz_ bkeeler: no need for squashing into one commit
jnthn bkeeler: But if it's what was agreed on from a previous review, then I guess fine to time out. :-)
bacek aloha
bkeeler It will need to make its way into parrot before I can give you the corresponding rakudo patch
jnthn o/ bacek
bacek jnthn, привет!
bkeeler hallo bacek
21:52 ShadeHawk left
bacek bkeeler, hio. 21:53
bkeeler Uh oh, we're back to privet bushes again?
bacek bkeeler, looks like you can help me with nqp :)
bkeeler I can?
jnthn bacek: Привет, как дела? :-)
bacek bkeeler, can you? :)
jnthn, отлично!
jnthn \o/ 21:54
bkeeler bacek: What can I do?
bacek \o/
moritz_ bkeeler: if you 'make boostrap-files' and commit the changes, I can push them into parrot
bacek bkeeler, can you clone github.com/bacek/pir?
moritz_ I can do it too, but first I have to build a newer parrot
which takes some time
bacek bkeeler, I have one reproducable bug (probably "hash ordering" related) 21:55
jnthn moritz_: You going for it?
(doesn't make sense for both of us to :-))
moritz_ jnthn: aye
jnthn moritz_++
bacek bkeeler, nopaste.snit.ch/20618 with little bit of description 21:56
jnthn rakudo: class A {method m { return [~] gather for ^3 {take 'a'}}}; class B {method n { return [~] gather for ^4 {.say;take 'b'}}};my B $x.=new; say $x.n 21:58
p6eval rakudo 9a15b8: OUTPUT«Could not find sub &prefix:<[~]>␤current instr.: 'perl6;B;n' pc 1016 (EVAL_1:0)␤»
jnthn Eww that's disgusting.
bkeeler moritz_: OK, done the bootstrap 21:59
22:00 Exodist joined 22:01 tlb joined
moritz_ bkeeler++ 22:01
bkeeler bacek: How do I build this thing? 22:02
dalek p-rx: b228f8f | (Bruce Keeler)++ | src/stage0/ (4 files):
Make new bootstrap
bacek bkeeler, parrot setup.pir
bkeeler, than something like "parrot --hash-seed=3 /usr/local/lib/parrot/2.4.0-devel/library/nqp-rx.pbc t/01-parse/05-assign.t" 22:03
(With adjustments of nqp-rx.pbc path) 22:04
22:04 kel__ left
bacek Test passed on my box with --hash-seed=2 and failed with "3" 22:04
22:05 Su-Shee left
bkeeler Hmmm 22:07
bacek bkeeler, nopaste.snit.ch/20620 shorter version of test
moritz_ parrot r46901 includes the new nqp-rx version 22:08
bacek Same "2" vs "3" hash seed problem.
bkeeler So what exactly does --hash-seed do? 22:12
bacek bkeeler, set initial seed for parrot hashes. 22:13
bkeeler So we're just make an otherwise hard-to-reproduce problem reproducable?
moritz_ right
bacek something like this
colomon bacek++ bkeeler++ 22:14
bkeeler Hallo there colomon
moritz_ \o/
colomon o/
22:14 moritz_ sets mode: +ooo bacek colomon bkeeler
colomon lurking here while I show my little guy music videos in a big window. ;) 22:18
moritz_ :-) 22:19
jnthn o/ colomon 22:20
colomon \o
jnthn opens up a La Trappe 22:21
bkeeler praises Belgian beer
bacek warming up with morning coffee
bkeeler for the 1000th time briefly considers joining a Belgian monestary 22:22
ajs What horrible brokenness would I have to commit to get "Could not find sub &prefix:<[,]>" from "return [,] @ap, <a b c>"??
bacek bkeeler, Leffe ftw!
colomon let the record show he watched five videos before deciding he really needed to be playing with the mouse. ;)
isBEKaml hi, colomon ! :) 22:23
jnthn ajs: I was just failing to fix that bug. :-/
moritz_ -> bed
colomon isBEKaml: shouldn't you be in bed? :)
ajs oh heh
colomon 'night, moritz_
isBEKaml colomon: way long past time.. :)
ajs Is there a better way to flatten lists?
or at least a working way? ;)
isBEKaml Nite moritz_
colomon jnthn: what's the deal?
jnthn colomon: Turns out we stick the generated meta-op subs in @BLOCK[0].loadinit(...) 22:24
colomon: So if we have
{ [~] ... }; { [~] ... };
We've a problem.
Erm
Actually not that
If it's namespace.
*namespaced
class A { [~] used in here }; class B { [~] used in here }
ajs Ah, so I could write a global list-flattener sub. 22:25
jnthn We only generated the meta-op once
ajs that'll work for now
jnthn ajs: Yeah, as a workaround
colomon jnthn: woah
jnthn colomon: Then we generate it, mark it as globablly generated in METAOPGEN
But of course, it's namespace. :-(
colomon do we have a test for that? :)
jnthn Obviously not.
I just tried to fix it
And failed 22:26
22:26 Exodist left
jnthn I figured putting it in @BLOCK[+@BLOCK -1] would do it 22:26
But we actually get a failure in the setting of all places then!
BrowserUk backlogs; comes up empty but raises a glass of Duval in salute to jnthn. & 22:27
jnthn Duval - also a good one. :)
bkeeler OK, that does it. It's a little early for me, but off to get a glass of Brother David's Trippel 22:28
jnthn BrowserUk: Oh, wait...Duvel or Duval?
colomon is starting to feel like should have gone to the party store earlier....
BrowserUk But trying tp pour the third (carefully to avoid the sludge) requires patience and practice...............still perfecting my technique, but hey. Practice makes perfect :) 22:29
jnthn colomon: oh...maybe the fail is a bootstrapping/ordering issue.
colomon: Maybe metaops.pm needs to come arlier in the bootstrap too.
BrowserUk jnthn: You are right Duvel. 22:30
(my typing is not good pre-Duvel.
Nite all. 22:31
colomon o
o/
22:31 BrowserUk left 22:32 Exodist joined
jnthn
.oO( s/pre/post/ ? ;-) )
 22:33
colomon: Trying out the setting order tweak. 22:34
colomon jnthn+
jnthn
.oO( what could possibly go wrong )
colomon jnthn++
.oO(this time for sure!)
 22:35
jnthn
.oO( Plz I can haz sharper circularity saw for Christmas? )
Grr, that didn't fix it. :-/ 22:36
colomon
.oO(I want one of those fancy new ones that shuts off rather than cutting off your hand if you make a logic error....)
 22:37
bkeeler jnthn: If you care to merge the 'regex-interpolation' branch in my github fork...
or anyone else who cares to for that matter
jnthn colomon: Can you remember where we set up != as an alias to !== ? 22:40
colomon in actions.pm?
jnthn oh, I found it
nah, it's in operators.pm
...so metaops.pm maybe wants to come before operators.pm in the bootstrap. Hmm. 22:41
Or maybe those aliases can get booted into metaops.pm.
colomon: aha. Fixed. 22:46
spectesting.
ajs rakudo: my @a = <1 2 3>;my @b = @a, <4 5 6>; say @b.perl 22:50
p6eval rakudo 1499f2: OUTPUT«["1", "2", "3", "4", "5", "6"]␤»
ajs That's perfectly sane 22:51
grr... something is off
22:52 meteorjay joined
ajs In my code I "return @a, <host>" and get [[@a], 'host'] 22:52
isBEKaml ajs: looks like you haven't interpolated your variables. 22:53
ajs interpolated in what sense? 22:54
isBEKaml rakudo: my @a=<1 2 3>; my @b = @a, <4 5 7>; say "@b"; #this..
p6eval rakudo 1499f2: OUTPUT«@b␤»
isBEKaml rakudo: my @a=<1 2 3>; my @b = @a, <4 5 7>; say @b[]; #this..
p6eval rakudo 1499f2: OUTPUT«123457␤»
isBEKaml rakudo: my @a=<1 2 3>; my @b = @a, <4 5 7>; say "@b[]"; #this..
p6eval rakudo 1499f2: OUTPUT«1 2 3 4 5 7␤»
isBEKaml see the quotes there.. :)
ajs Oh, sorry. I was shorthanding...
I get [[1,2,3],4] from @a=(1,2,3) when I "return @a, 4" or "my @b = @a, 4" or whatever. 22:56
No, no I figured it out!
silly me 22:57
isBEKaml :)
jnthn ajs: It'll not flatten at the point of return, but it should flatten if you put it in list context. 22:58
ajs: It returns a Parcel, fwiw.
ajs yeah, my @a contained a list of lists
jnthn Parcels themselves maintain structure until something makes 'em not
Ah, OK. :-)
That'd do it too ;-)
ajs yep ;-)
22:59 whiteknight joined, Psyche^ joined 23:00 skangas left 23:02 jaldhar joined 23:03 Patterner left, Psyche^ is now known as Patterner
isBEKaml hits the sack. Night, folks (even as it's early morning here) :) 23:05
lue 'ello!
night
23:06 isBEKaml left, whiteknight left, whiteknight joined
bkeeler Heya lue 23:07
dalek kudo: 5cb546e | jonathan++ | (2 files):
Make sure that since we build meta-ops once globally, we install them at the top

issue, so done a little re-ordering.		 23:13
lue Just curious: If i went insane and tried to convert a .pir file (say in src/builtins) to its equivalent .pm file, would it work/be ok? 23:23
bkeeler bacek: I think the issue is NQP trying proto regexes in a different order. Enabling parse tracing helps a lot
bacek: nopaste.snit.ch/20623
lue [assuming, of course, everything is done right: changing the makefile appropriately, etc.]
bacek bkeeler, yes, I suspected it. Any ideas how to fix it? 23:24
bkeeler bacek: I generated parse traces with --hash-seed=2 and =3 and compared them
ajs rakudo: sub a { return [ <1 2 3>, <4 5 6> ] }; for a() -> $x { say "List: $x" }
p6eval rakudo 1499f2: OUTPUT«List: 1 2 3 4 5 6␤»
ajs did I miss something there? Shouldn't that be List: [1, 2, 3] and List: [4, 5, 6]?
bkeeler bacek: Possibly your grammar is commiting to an alternative that it shouldn't. I don't have more time right now, but I might be able to look at it more tomorrow 23:25
bacek bkeeler, thanks a lot!
bkeeler Anyway, I'm out for the evening. See y'all anon 23:26
ajs ciao bkeeler
jnthn ajs: BTW, 5cb546e fixed that bug you mentioned earlier. 23:28
(the one with the reduction ops)
23:34 lestrrat is now known as lest_away, lest_away is now known as lestrrat

Powered by Cro and the Raku® Programming Language.

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

Thank you!