»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:07 thou left 00:10 yertalert joined 00:25 pjcj left
dalek osystem: e28fb6a | (Solomon Foster)++ | META.list:
Add Math::ContinuedFractions.
 00:33
00:33 pjcj joined
timotimo what's the shortest/prettiest way for this: multi fib(Int $x where * < 0) { die "negative numbers not allowed"; } ? 00:51
00:51 aindilis joined
sorear Use a subset type and let the dispatcher error for you 00:52
timotimo when i did that, the dispatching failed, because now the normal implementation and the memoized function became "equally special" to the dispatcher
or something
(and the error message wasn't pretty. it just said Cannot call 'fib'; none of these signatures match: :(Int $x) :(Int $x) 00:53
japhb timotimo, try not putting a where clause on the error case, just on the success cases. Then the error case is more general, and would only match when the other ones don't. Er, I would think.
timotimo yeah, now i get "these signatures all match" 00:54
i have subset Positive of Int and subset InMemo of Positive 00:55
japhb timotimo, hmmm. How about making the failure case: multi fib($) { die "Non-negative integers only, please." } 00:56
timotimo i did that now, yes
but now the distinction between "normal positive int" and "int that's also in my memo"
here, let me show the code
japhb timotimo, gist it 00:57
timotimo gist.github.com/1932925
01:00 Vespakoen joined
japhb Ah, I see, as soon as the first fib routine works its way back to the memoized numbers, then boom. 01:01
01:01 FACEFOX left
timotimo yup 01:02
01:02 Timbus|Away joined, Timbus left, Timbus|Away is now known as Timbus
japhb Hmmm, it certainly *intuitively* seems that your program should DWIM, but I'm not sure if it's specced that way. 01:03
01:03 FACEFOX joined
timotimo heh. 01:03
it would need to be in the spec, that a restriction on a subset is less specific than a restriction on a subset of that same subset
01:04 integral left, integral joined, integral left, integral joined
colomon Yeah, I'd be surprised if that ever worked 01:04
timotimo so, how is it fixed? 01:05
one way is to put the error case first, so that it would match immediately
01:06 envi_ left
japhb is still searching the specs 01:07
timotimo i also put a die unless into fib(Int $x), but that's pretty ugly 01:08
japhb Hmmm, S06:2389-2409 seems to indicate subsets get flattened away when building the signature objects 01:09
timotimo how about the Positive subset + an extra where clause to match InMemo 01:10
01:12 scott__ joined 01:14 lumi__ left, lumi__ joined
japhb Bah. 01:15
Yeah, S12:1140 and following confirm it, especially the paragraph at 1163-1169. 01:16
The fact that there *are* constraints matters. *How many* constraints is not considered.
I argue that that breaks POLA when it comes to subsets though: the reality leaks through the sugar a little too strongly. 01:17
colomon I'm inclined to think it's just asking too much of the multi dispatch.
though note that memoization should be built-in with "is cached" (though as far as I know it's not yet) 01:18
01:19 sudokode left
japhb re: 'is cached', sure. But I'm just seeing his code as an example of subset usage, without regard to his implementation of memoization. 01:20
colomon gist.github.com/2026007 (does away with multi dispatch) 01:21
01:21 sudokode joined
japhb Oooh, S12:1214-1222. So perhaps this is just a nom bug? 01:23
colomon gist.github.com/2026007 (now with second script, super-short and efficient)
japhb chuckles 01:24
colomon japhb++
I think you're right that this is a nom bug
which immediately leads to the question... 01:25
niecza: gist.github.com/1932925
p6eval niecza v15-4-g1f35f89: OUTPUT«Potential difficulties:␤ $x is declared but not used at /tmp/RbkGEJymsa line 17:␤------> multi fib(Int ⏏$x) {␤␤Usage:␤ /tmp/RbkGEJymsa <fib>␤»
japhb Need to be able to do args for gist eval
And ... 01:26
benabik If it runs equally constrained versions in order declared, then it would never reach the `fib(InMemo $x)` multi
japhb Given S12:1214-1222, the candidates should appear in the source in the opposite order they are now.
exactly.
And the failing case should be 'multi fib($) {...} 01:27
colomon ------> multi fib(Int ⏏$x) {
Unhandled exception: non-negative arguments required
bah, reversing the mutlis doesn't help. :( 01:28
japhb S12:1214-1222 does have the interesting effect of requiring that information about order of declaration of multis be saved into compiled modules, so that when the module is imported, the multis can be folded into the declaration order list properly. 01:30
01:31 havenn left
japhb Hmmm ... but then how does that interact with the scoping rule of Tiebreaker A (S12:1202-1204) 01:31
What is the relative scope of an imported module for this purpose? 01:32
01:32 havenn_ joined
japhb phenny, ask jnthn Does Rakudo constant-fold bit math on integer constants? Which is to say, can I assume that +|'ed together bit flag constants will turn into a single bitmask in the compiler output? 01:38
phenny japhb: I'll pass that on when jnthn is around.
01:38 leprevost joined 01:42 havenn_ left 01:51 araujo left 01:54 isBEKaml left 02:00 whiteknight left 02:02 araujo joined 02:04 Vespakoen left, Vespakoen joined 02:12 pernatiy left 02:19 skyheights joined 02:24 Vespakoen left, Vespakoen joined 02:28 benabik left 02:29 araujo left, benabik joined 02:31 pernatiy joined 02:35 Vespakoen left, araujo joined, araujo left, araujo joined 02:39 gilm left 02:41 skyheights left 02:42 att joined, att is now known as att_ 03:06 orafu left 03:08 orafu joined 03:10 brxue left 03:30 pochi left 03:31 lestrrat left, lestrrat joined 03:34 leprevost left 03:37 NamelessTee left 03:38 sftp left 04:02 FACEFOX left, FACEFOX joined 04:17 skids left 04:20 japhb left 04:21 jaldhar joined 04:24 lestrrat left 04:26 lestrrat joined, japhb joined 04:32 envi_ joined 04:45 nnunley left 05:05 nnunley joined 05:19 kaleem joined, yertalert left 05:22 tyatpi left 05:33 thou joined, nnunley left 05:35 nnunley joined 05:37 envi_ left 05:47 imarcusthis left 05:48 imarcusthis joined 05:59 mucker joined 06:00 localhost left 06:01 localhost joined
moritz seems it doesn't :( 06:08
well, if you write constant a = 1 +| 4 then it's constant folded 06:09
06:09 scott___ joined 06:10 scott__ left
moritz but a simple my $foobar = 1 +| 4; doesn't constant-fold it 06:10
though it inlines the +| operator
06:14 NamelessTee joined 06:15 imarcusthis left 06:17 imarcusthis joined 06:24 birdwindupbird joined 06:26 rsimoes1 joined 06:30 rsimoes1 left 06:33 aindilis left, rsimoes1 joined 06:34 [particle]1 joined 06:35 buubot_backup left, _ilbot left 06:36 [particle] left, buubot_backup joined, _ilbot joined, rsimoes left, gfldex left, TimToady left, gfldex joined 06:37 NamelessTee left, FACEFOX left, REPLeffect left, colomon left, krunen left, adnauseam left, ranguard_ left, PZt left, ranguard joined, thou left, Trashlord left, apejens left, krunen joined 06:38 mathw left, adnauseam joined, mathw joined 06:39 apejens joined 06:40 mucker left 06:42 FACEFOX joined 06:43 thou joined 06:44 TimToady joined 06:50 NamelessTee joined 06:52 rsimoes joined 06:56 FACEFOX left 06:57 wtw joined, NamelessTee left 07:00 FACEFOX joined 07:05 thou left 07:12 FACEFOX left 07:19 agentzh joined
jnthn good mornin' 07:26
phenny jnthn: 01:38Z <japhb> ask jnthn Does Rakudo constant-fold bit math on integer constants? Which is to say, can I assume that +|'ed together bit flag constants will turn into a single bitmask in the compiler output?
07:26 preflex left, preflex_ joined
jnthn phenny: tell japhb not yet *but* the bs branch landing was the blocker for that and I already have some constant folding code, just not wired up yet. So you can expect it to be in place soon. 07:26
phenny jnthn: I'll pass that on when japhb is around.
07:27 preflex_ is now known as preflex
tadzik hello #perl6 07:31
moritz \o tadzik, jnthn 07:32
at least the call is inlined
jnthn moritz: true
But really the answer should be. :) 07:33
Currently that needs explicit use of the constant syntax though
moritz why are all those WHAT ( PIROP perl6_box_int__PI ( ... things on the string heap? 07:34
erm, s/WHAT/WANT/
sorear o/ tadzik , jnthn , moritz
tadzik hi sorear 07:35
jnthn moritz: Inlining info
moritz jnthn: ok
jnthn moritz: the inliner currently encodes some ASTs to a little DLS because we can't serialize PAST
Soon we'll have QAST and that will be 6model based and then we'll be able to serialize it though 07:36
o/ sorear
moritz: That said, the string heap certainly contains some stuff that bothers me
Like, the reference to Perl6::Compiler
Somewhere, somehow, the serializer is finding things it probably should not. 07:37
07:37 FACEFOX joined
moritz commutes 07:38
07:38 FACEFOX left 07:42 FACEFOX joined 08:01 lestrrat left 08:02 drbean left, lestrrat joined 08:03 FACEFOX left, FACEFOX joined 08:08 drbean joined 08:12 alim joined 08:19 icwiener joined
moritz www.perlmonks.org/?node_id=959260 that was a fun p5 question to answer, though not hard 08:21
the p6 solution looks mostly the same
though of course there's a different, cute way in Perl 6 08:24
r: gist.github.com/2027585 08:25
p6eval rakudo bd6742: OUTPUT«6␤»
moritz n: gist.github.com/2027585
p6eval niecza v15-4-g1f35f89: OUTPUT«Unhandled exception: Excess arguments to Cool.match, unused named g␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (Cool.match @ 1) ␤ at /tmp/hF6B9e70pe line 9 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3838 (ANON @ 3) ␤ at /home/p6e…
moritz p: gist.github.com/2027585
p6eval pugs b927740: OUTPUT«*** ␤ Unexpected "0"␤ expecting operator or ")"␤ at /tmp/gnQD2DWDb7 line 14, column 34␤»
moritz ah, pugs doesn't like the series operator 08:26
08:30 mj41 joined 08:32 tyatpi joined 08:37 birdwindupbird left 08:38 Trashlord joined 08:42 icwiener left, icwiener joined, icwiener left, icwiener joined 08:45 birdwindupbird joined 08:49 sisar joined, alim left 08:55 icwiener left 09:08 stepnem left 09:09 stepnem joined
eiro hello there 09:09
jnthn o/ eiro 09:10
moritz \o * 09:11
09:18 FACEFOX left, FACEFOX joined 09:36 alim joined 09:38 Vespakoen joined, Vespakoen left, dakkar joined 09:40 lestrrat left 09:41 lestrrat joined
dalek ast: 1faa73d | moritz++ | S03-operators/value_equivalence.t:
test lexical roles and ===
 09:44
cognominal perl6: say "$_: " ~ (try { say $_; eval $_ }).WHAT.perl for < [] {} () \() :() [] !@@% > 09:45
p6eval niecza v15-4-g1f35f89: OUTPUT«[]␤[]: Array␤{}␤{}: Hash␤()␤(): Parcel␤\()␤\(): Capture␤:()␤:(): Any␤[]␤[]: Array␤!@@%␤!@@%: Any␤»
..pugs b927740: OUTPUT«pugs: Missing required parameters: $_␤»
..rakudo bd6742: OUTPUT«Any()␤[]: X::AdHoc␤Any()␤{}: X::AdHoc␤Any()␤(): X::AdHoc␤Any()␤\(): X::AdHoc␤Any()␤:(): X::AdHoc␤Any()␤[]: X::AdHoc␤Any()␤!@@%: X::AdHoc␤»
moritz rakudo: say WHAT eval '[]' 09:47
p6eval rakudo bd6742: OUTPUT«Array()␤»
moritz rakudo: say WHAT eval $_ for < [] {} >
p6eval rakudo bd6742: OUTPUT«Array()␤Hash()␤»
moritz rakudo: say WHAT eval $_ for < [] {} \() () >
p6eval rakudo bd6742: OUTPUT«Array()␤Hash()␤Capture()␤Parcel()␤»
moritz rakudo: say WHAT eval $_ for < [] {} \() () :()> 09:48
p6eval rakudo bd6742: OUTPUT«Array()␤Hash()␤Capture()␤Parcel()␤Signature()␤»
cognominal moritz: thx for trying to golf it :)
jnthn r: say "$_" for 1..2; 09:49
p6eval rakudo bd6742: OUTPUT«1␤2␤»
moritz rakudo: say WHAT try eval $_ for < [] {} \() () :()>
p6eval rakudo bd6742: OUTPUT«List()␤»
jnthn r: say "{$_}" for 1..2;
p6eval rakudo bd6742: OUTPUT«1␤2␤»
cognominal nom: < [] {} () \() :() [] !@@% >.map: { say "$_: " ~ (try { eval $_ }).WHAT.perl }
p6eval rakudo bd6742: ( no output )
moritz so, the try seems to make the difference
09:49 simcop2387 left, overrosy left, ruoso left, takesako left, sirhc left
moritz rakudo: say try eval $_ for < [] {} \() () :()> 09:49
p6eval rakudo bd6742: OUTPUT« Signature<-1765137524978402461>␤»
moritz rakudo: say try { eval $_ } for < [] {} \() () :()> 09:50
cognominal the .map form works correctly
p6eval rakudo bd6742: OUTPUT«Nominal type check failed for parameter '$code'; expected Str but got Any instead␤ in sub eval at src/gen/CORE.setting:442␤ in block <anon> at /tmp/qrm9UcCSFz:1␤␤Nominal type check failed for parameter '$code'; expected Str but got Any instead␤ in sub eval at sr…
cognominal rakudo: < [] {} () \() :() [] !@@% >.map: { say "$_: " ~ (try { eval $_ }).WHAT.perl }
p6eval rakudo bd6742: ( no output )
moritz does try mess with $_ ?
cognominal hum.
moritz nom: try { say $_ } for <1 b>
p6eval rakudo bd6742: OUTPUT«Any()␤Any()␤»
jnthn Don't think so
cognominal not that one
moritz thinks so
jnthn nom: do { say $_ } for <1 b> 09:51
p6eval rakudo bd6742: OUTPUT«===SORRY!===␤Unsupported use of do...for; in Perl 6 please use repeat...for␤at /tmp/Gdrlr5w4iq:1␤»
jnthn nom: blah { say $_ } for <1 b>; sub blah($c) { $c() }
p6eval rakudo bd6742: OUTPUT«1␤b␤»
jnthn Hmm, does seem hard to find a non-try case.
Still don't see why it should be special though
moritz nom: for <1 b> { try { say $_ } } 09:52
p6eval rakudo bd6742: OUTPUT«1␤b␤»
moritz so, only in the statement modifying case
moritz submits rakudobug
n: try { say $_ } for <1 2> 09:53
p6eval niecza v15-4-g1f35f89: OUTPUT«1␤2␤»
moritz n: try say $_ for <1 2>
p6eval niecza v15-4-g1f35f89: OUTPUT«1␤2␤»
moritz r: try say $_ for <1 2>
p6eval rakudo bd6742: OUTPUT«1␤2␤»
moritz r: try { say $_ } for <1 2>
p6eval rakudo bd6742: OUTPUT«Any()␤Any()␤»
09:53 FACEFOX left
moritz ok, only the block case 09:53
jnthn I'm sure there's another ticket a bit like thsi but not the same 09:54
moritz well, at least we have it golfed quite a bit
jnthn aye 09:55
moritz r: try { say $_ } given 2
p6eval rakudo bd6742: OUTPUT«Any()␤»
09:55 simcop2387 joined, overrosy joined, ruoso joined, takesako joined, sirhc joined
jnthn I think it's to do with nested blocks plus statement modifiers 09:55
moritz sounds about right
cognominal yes, I think so
09:56 FACEFOX joined
jnthn It's a pain to handle. 09:56
The blocks get attached to the outer they are actually written in
moritz but need to get the $_ from the contextualizer, not from the block
jnthn But then the statement modifier forces a block to be later generated that has to go "around" them
And we dont' know about it until too late.
So it needs special casing.
moritz submitted as [perl #111704] 09:57
jnthn And every way I think of to fix this issue feels like a hack :/
moritz sounds like the codegen of the statement needs to be defered until the statement modifier has done its magic 09:59
09:59 sporous left
moritz or... something :/ 09:59
10:01 FACEFOX left
dalek ast: c070484 | moritz++ | S04-statements/try.t:
test for RT #111704, try-blocks plus statement modifiers gobble up $_
 10:01
10:02 FACEFOX joined, FACEFOX left, Guest80268 joined
cognominal perl6: perl 1 10:03
p6eval pugs b927740: ( no output )
..niecza v15-4-g1f35f89: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'perl' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1362 (die @ 3) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 33) ␤ at /home…
..rakudo bd6742: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&perl' called (line 1)␤»
cognominal perl6: gist 1
p6eval pugs b927740: OUTPUT«*** No such subroutine: "&gist"␤ at /tmp/zAgis9XNb0 line 1, column 1 - line 2, column 1␤»
..rakudo bd6742, niecza v15-4-g1f35f89: ( no output )
10:03 Trashlord left 10:04 FACEFOX joined 10:12 kaleem left 10:15 envi_ joined, ab5tract joined 10:16 scott___ left 10:24 colomon joined 10:29 pernatiy left 10:33 xinming left, sisar left 10:35 xinming joined 10:41 PZt joined
masak www.perlmonks.org/?node_id=959260 -- "I would be so grateful! -- Pics? ;)" -- I'm thinking of acquiring a sexism gag reflex, like the one sorear has. :( 10:41
oh, hai, #perl6
moritz masak: yes, that part annoyed me too 10:42
oh hai masak
10:42 benabik left
huf did that seriously happen?! 10:42
colomon o/
huf sweet jesus
tadzik :) 10:43
masak I'm glad it's not just me. 10:45
10:48 daxim joined 10:50 xinming left
masak replied: www.perlmonks.org/?node_id=959264 10:51
moritz fwiw I've considered that node for editing (you need a minimum XP level to vote on the consideration)
colomon masak++ 10:52
moritz masak++
10:52 Trashlord joined
masak thank you. still made me sad, though. :( 10:53
on a brighter note, japhb++ found a buglet in gist.github.com/2021108 10:54
tadzik I'm wondering if you don't it too seriously though
moritz well, nobody ever writes such a comment in response to a man 10:55
tadzik that's true
moritz and treating men and women differently is the first step to alienating women
10:56 xinming joined
masak tadzik: the comment implies that katarinahm should express her gratitude to the replier by sending him pictures. 10:58
tadzik well, right 10:59
masak tadzik: (1) what moritz said. (2) expressing the idea, even as a joke, means we somehow don't accept women as first-class citizens.
arnsholt It's just in bad taste, IMO
masak yeah, basically.
11:05 xinming left
masak the polarisation between "you're reading too much into it" and "well, consider what this means" seems fundamental to almost any debate on sexism in the programming communities. 11:06
11:06 xinming joined
eiro tadzik, ack. 11:06
masak, jnthn, tadzik: i asked cognominal to create a repo "perl6 course" and add us at contributors 11:07
11:08 odoacre left
moritz shouldn't we just make one in the perl6 organization? 11:08
and give everyone who wants commit access, just like with roast, the websites etc.
eiro moritz, you tell me :) you guys decide something and i just follow you 11:09
cognominal eiro: it is not yet in the repo, but as a gist. gist.github.com/2028115 11:10
eiro cognominal, can a gist become a repo ? can i clone ? branch ?
moritz yes, it can, though with some command line involvement 11:11
11:12 aindilis joined 11:13 rsimoes1 left, lestrrat left
eiro bbl 11:14
11:14 lestrrat joined
moritz now at github.com/perl6/course 11:14
tadzik eiro: course?
eiro moritz, thx 11:15
arnsholt Oooh, French Perl
eiro tadzik, yep! it's a complete training session we're planning.
really have to go! bbl
arnsholt, yup :)
arnsholt I should read it, if nothing else to brush up on my French =D
tadzik oh nice 11:16
cognominal beware, it is still raw. there are a lot of typos. That's why it was a gist and not yet a depot.
moritz commit early, repo often :-) 11:17
arnsholt cognominal: Don't worry. It's been so long since I read any French that I probably won't notice any typos =)
moritz one thing we can do is set up some perl 6 examples by category, and those can be shared between different courses 11:18
like, examples for classes, roles, regexes, ...
cognominal maintenant, je me re_pose_ un peu. I will pause a bit and get a lunch. 11:19
11:23 bruges_ is now known as bruges, bruges left, odoacre joined
lumi__ Is there a useful way for me to run spectests nightly? 11:24
moritz lumi__: you mean, if there's an automatic smoker client? or a good way to submit the results? or if it's useful for us at all? 11:27
11:27 bruges joined
lumi__ More or less any of these 11:28
moritz yes, would be useful
11:28 NamelessTee joined
moritz you can run 'make spectest_smolder' to submit the reports 11:28
and you'll probably need to write your own cron script
lumi__ moritz: Thanks, exactly what I needed :) I'll set it up 11:30
11:33 pmurias joined 11:38 benabik joined 11:43 grondilu joined, grondilu left 11:47 grondilu joined
grondilu nom: macro talk($s) { quasi { say {{{{$s}}}} } }; talk "hello"; 11:48
p6eval rakudo bd6742: OUTPUT«Block.new()␤»
11:48 mdupont joined
grondilu pugs: macro talk($s) { quasi { say {{{{$s}}}} } }; talk "hello"; 11:48
p6eval pugs b927740: OUTPUT«<SubBlock(<anon>)>␤»
grondilu pugs: macro talk($s) { quasi { say {{{$s}}} } }; talk "hello";
p6eval pugs b927740: OUTPUT«<SubBlock(<anon>)>␤»
grondilu is it too early for these stuffs?
jnthn Yes 11:49
The quasi interpolation ain't impl yet 11:50
(the {{{$s
}}} bit)
grondilu ok
mikec wonders what that does
grondilu do you guys know about wikicreole? They have a EBNF grammar that I'd love to see implemented in perl6: www.riehle.org/2008/01/09/an-ebnf-g...creole-10/ 11:51
jnthn mikec: Lets one chunk of AST be embedded within another 11:52
12:04 mtk joined
cognominal moritz, do I have the proper rights to push to github.com/perl6/course? I got a 403. 12:09
I should learn more about github depot adminstration. 12:10
jnthn depot? :) 12:13
cognominal repository
jnthn :)
cognominal or dépôt 12:14
…which should not be a dépotoir (dumping ground). 12:15
12:15 sporous joined
moritz cognominal: well, you need to use the ssh URL 12:24
cognominal: ... and I've added you to the perl6 group, so now you should have commit access 12:25
wow, did you never ever patch roast or the specs or so?
12:26 PacoAir joined
moritz hugme: add eiro to perl6 12:26
hugme moritz: You need to register with freenode first
moritz hugme: add eiro to perl6 12:27
hugme hugs eiro. Welcome to the perl6 github organization
moritz anyone else missing?
cognominal is looking french people with an interest in Perl 6 and a github account : maddingue, book, dams, franckcuny, fperrad 12:34
grondilu is french, has a github account, but is not a good perl6 programmer yet :( 12:37
benabik He did say interest in Perl 6, not proficiency in.
moritz hugme: add grondilu to perl6
hugme hugs grondilu. Welcome to the perl6 github organization
grondilu :)
moritz resistance is futile. You have been assimilated :-) 12:38
12:38 alim left
cognominal sorry grondilu to have forgotten you. 12:38
grondilu check out my offline wikipedia implementation using perl5: www.perlmonks.org/?node_id=958466 It works but Text::Mediawiki's rendering kind of sucks. I'm pretty sure a Perl6 grammar could help parsing the wiki markup. Thus my remark about wikicreole. Any help welcome. 12:42
12:43 sporous left, sporous joined
dalek p/b64-lookup-table: 1859ad1 | moritz++ | src/6model/base64.c:
base64 decoding now uses a lookup table, not repeated branching
 12:44
moritz phenny: tell jnthn at your convenience, please profile the nqp/b64-lookup-table branch to see if it has made any difference. My own experiments with rakudo startup time remain inconclusive
phenny moritz: I'll pass that on when jnthn is around.
12:44 tokuhiro_ joined 12:45 preflex left, preflex_ joined
masak it's hard to tell if one Anonymous Monk is infix:<===> to another, but www.perlmonks.org/?node_id=959296 was perhaps not entirely unexpected. basically a nonsensical reply. as for me, I'm not digging further down that thread. 12:45
12:46 preflex_ is now known as preflex, NamelessTee left
moritz I've just downvoted and otherwise ignored it 12:47
masak good.
12:56 sisar joined 12:57 Chillance joined 13:13 [particle] joined 13:15 [particle]1 left
au grondilu: I've been working at $job on www.mediawiki.org/wiki/Parsoid#Getting_started -- the core pegTokenizer.pegjs.txt is written in PEG.js pegjs.majda.cz/documentation and perhaps a PEGjs->Perl6Rules translator would be useful 13:19
or PEGjs<->Perl6Rules, even. 13:20
grondilu checks this out 13:22
13:25 drbean left 13:27 drbean joined
masak nice, simple syntax. 13:32
also, an interesting way to inline actions into grammar.
13:33 Vespakoen joined
grondilu au: I installed npm on my sid machine but it doesn't seem to work :( Isn't there a tar ball for the whole thing instead? 13:33
au npm 1.0+? 13:34
also, if "it doesn't seem to work" means "it fails hundreds of tests on 'make test'", then it's quite normal :) 13:35
grondilu no, even a simple "npm -v" throws an exception 13:36
au grondilu: no tarballs atm afaik; you might want to install nodejs.org/dist/v0.6.12/node-v0.6.12.tar.gz from source to get npm working; then it should be possible to checkout only the svn.wikimedia.org/svnroot/mediawiki...ualEditor/ part and "cd tests/parser ; make test".
(feel free to continue this conversation at #wikimedia-dev) 13:37
grondilu ok 13:38
13:39 pernatiy joined 13:40 skids joined
moritz phenny: tell jnthn MiniDBI fails to build on newest rakudo: 'Missing serialize REPR function'. Do the NCI stuff needs to be update to deal with bs? 13:48
phenny moritz: I'll pass that on when jnthn is around.
jnthn moritz: sounds lajk yes :)
phenny jnthn: 12:44Z <moritz> tell jnthn at your convenience, please profile the nqp/b64-lookup-table branch to see if it has made any difference. My own experiments with rakudo startup time remain inconclusive
jnthn: 13:48Z <moritz> tell jnthn MiniDBI fails to build on newest rakudo: 'Missing serialize REPR function'. Do the NCI stuff needs to be update to deal with bs?
jnthn Guess we ain't lucky enough to know which REPR was missing it.... 13:49
Hmm. I didn't think that would happen.
oh, wait...I know.
Though can't fix here... 13:50
13:50 aindilis left 13:52 aindilis joined
moritz no hurry 13:53
flussence seems to be just MiniDBI, my NativeCall-using code seems to work 13:59
(then again I'm not doing anything particularly exciting in it...)
moritz flussence: well, 'make' in MiniDBI precompiles stuff
flussence oh, that'd be why :)
jnthn It'll just show up in precomp
moritz without precomp... 14:01
ok 2 - MiniDBD::mysql version 0.01
Segmentation fault
felher After i had a problem with a program i golfed down the issue for quite some time. But i don't get why the two golfed-down one-liners produce different output: 14:04
@g.push(3);
my @g = @g => 'x';
@g.
...
nom: my $x; my @a = 1,2; $x = Array.new(@a) => 'x'; $x.key[0]; @a.pop(); say $x;
p6eval rakudo bd6742: OUTPUT«[1, 2] => "x"␤»
felher nom: my $x; my @a = 1,2; $x = Array.new(@a) => 'x'; @a.pop(); say $x; 14:05
p6eval rakudo bd6742: OUTPUT«[1] => "x"␤»
felher Might this be a bug?
moritz the difference is just the $x.key[0]; in sink context, right?
felher moritz: yes 14:06
flussence nom: my $x; my @a = 1,2; $x = Array.new(@a) => 'x'; $x.key[0]; say $x;
p6eval rakudo bd6742: OUTPUT«[1, 2] => "x"␤»
moritz probably is a bug, yes
felher Ah, okay. This was quite funny to debug, since everytime i added debugging-output everything worked (because i printed it). And everytime i deleted the print statements for debugging it failed again :) 14:08
colomon is really happy with the next stage of the continued fractions project... 14:10
masak felher++ # nice golf!
felher: should I submit it? 14:11
felher masak: sure. Press your "masak submits rakudobug"-Button.
masak submits rakudobug
felher masak: you do have that button, don't you?
masak smiles mysteriously 14:12
felher :D
moritz smiles too
jnthn moritz: Hm. I only tested the mysql thing on Windows. :) 14:15
moritz: Can I have a stacktrace?
moritz jnthn: gist.github.com/2029048 14:17
jnthn wtf... 14:18
moritz that's a good comment. 14:19
jnthn Innevitably heap corruption of soem kind.
*some
I guess many things allocate out of Parrot pools, but bigint and NCI do with malloc 14:20
masak nom: my @a = 1,2; my $x = [@a] => 'x'; $x.key[0]; @a.pop(); say $x; 14:22
p6eval rakudo bd6742: OUTPUT«[1, 2] => "x"␤»
masak nom: my @a = 1,2; my $x = [@a] => 'x'; @a.pop(); say $x;
p6eval rakudo bd6742: OUTPUT«[1, 2] => "x"␤»
masak oh, 'course.
moritz nom: my @a = 1,2; my $x = @a => 'x'; @a.pop(); say $x; 14:23
p6eval rakudo bd6742: OUTPUT«[1] => "x"␤»
moritz that part seems to make sense
jnthn That bit looks right.
masak aye.
nom: my @a = 1,2; my $x = @a => 'x'; $x.key[0]; @a.pop(); say $x; 14:24
p6eval rakudo bd6742: OUTPUT«[1] => "x"␤»
masak so Array.new is a necessary component of it all.
jnthn ah
I guess Array.new probably ain't exercised a great deal 14:25
moritz aye
14:25 pernatiy left
moritz and it's probably written by me, in a rather naive way 14:25
14:25 pernatiy joined
jnthn nom: say Array.new(1,2,3).WHAT 14:26
p6eval rakudo bd6742: OUTPUT«Array()␤»
14:27 tyatpi left
jnthn wonders if Array.new can't just be written method new(*@a) { @a } 14:27
Well, besides subclassing...
But the current implementation gets that wrong too anyway 14:28
oh
no, it doesn't...
But it looks like it may well be doing soemthing too shallow
14:29 tokuhiro_ left 14:30 kaleem joined 14:48 att_ left 14:51 envi_ left 14:57 tokuhiro_ joined 15:05 plutoid joined 15:07 mj41 left 15:12 JimmyZ joined, JimmyZ left 15:23 mj41 joined 15:27 colomon left 15:32 sftp joined 15:35 grondilu left, tokuhiro_ left 15:37 plutoid left
masak std: sub foo($0) {} 15:38
p6eval std 1ad3292: OUTPUT«===SORRY!===␤Cannot declare a numeric parameter at /tmp/G65mBZ4rAk line 1:␤------> sub foo($0⏏) {}␤Parse failed␤FAILED 00:01 109m␤»
jnthn nom: sub foo($0) {}
p6eval rakudo bd6742: OUTPUT«===SORRY!===␤Cannot declare a numeric variable␤at /tmp/36eFpJi7Yr:1␤»
jnthn Close E. Nuff 15:39
masak :D
moritz nom: try eval 'sub foo($0) { }'; say $!.perl
p6eval rakudo bd6742: OUTPUT«X::Syntax::Variable::Numeric.new(what => "parameter", filename => "eval_0", line => 1, column => Any)␤»
jnthn oh wow, the exception info even knows it's a parameter 15:40
moritz++
moritz now it should also say so :-)
jnthn Probably just tweaking method message to get to what STD has
moritz aye
mikec wow, that's cool
moritz compiles a fix 15:41
one thing I want to add soonish is the context
we might not get it as accurately as STD, but should still be better than nothing 15:42
jnthn "the context"? 15:43
15:43 wtw left
moritz the surrouding text from the parse 15:44
jnthn aha
moritz "context" is a totally overloaded word in here :-)
masak .oO( its meaning depends on a lot of things surrounding its use ) 15:45
15:46 Patterner left
jnthn should start using www.contexteditor.org/ just to make things worse :P 15:47
15:47 colomon joined
dalek kudo/nom: fee8919 | moritz++ | src/core/Exception.pm:
awesomize error message for numeric parameters
 15:47
15:50 agentzh left 15:56 bluescreen10 joined, sftp_ joined 15:57 sftp left 16:00 att joined 16:01 Psyche^ joined, Psyche^ is now known as Patterner
masak I believe I have what is the final word in the search for an even-distribution balanced string generator. 16:05
r: sub c { state @c = 1; return @c[$^n] if @c > $n; 2 * (2 * $n - 1) / ($n + 1) * c($n - 1) }; sub b($l) { return "[]" x $l if $l < 2; my $r = (^c($l)).roll; my $s; for reverse(^$l) Z ^$l -> $l1, $l2 { $s += c($l1) * c($l2); return [~] "[", b($l1), "]", b($l2) if $r < $s } }; say b 5
p6eval rakudo fee891: OUTPUT«[][][][[]]␤»
moritz r: sub c { state @c = 1; return @c[$^n] if @c > $n; 2 * (2 * $n
- 1) / ($n + 1) * c($n - 1) }; sub b($l) { return "[]" x $l if
$l < 2; my $r = (^c($l)).roll; my $s; for reverse(^$l) Z ^$l ->
$l1, $l2 { $s += c($l1) * c($l2); return [~] "[", b($l1), "]",
p6eval rakudo fee891: OUTPUT«===SORRY!===␤Unable to parse blockoid, couldn't find final '}' at line 2␤»
moritz eeks
c&p fail 16:06
masak I'd say.
&c calculates catalan numbers.
&b calculates balanced strings recursively.
the important part is the dice throw $r, which is then compared against a cumulative sum of products of catalan numbers. 16:07
the details of this comparison guarantee an even distribution, because the cumulative sum takes into accound how many strings of the form "[X]Y" there are for X and Y of various lengths. 16:08
account*
it's a bit of a waste to be making a new roll on each recursive call -- the first $r really has enough bits to pin down a string entirely. but it doesn't hurt either. 16:10
moritz suspects it's not quite the final say after all :-) 16:11
masak so in the end, it turns out that the distribution can be mapped down to (a recursive) [\+] of the catalan numbers.
moritz decommutes 16:14
16:15 wolfman2000 left 16:18 sisar left
masak r: sub c { state @c = 1; return @c[$^n] if @c > $n; 2 * (2 * $n - 1) / ($n + 1) * c($n - 1) }; sub b { return "[]" x $^l if $l < 2; my $r = (^c($l)).roll; return "[&b($l - 1 - $_)]&b($_)" if $r < (my $s += c($l1) * c($l2)) for ^$l }; say b 5 16:19
p6eval rakudo fee891: OUTPUT«===SORRY!===␤Variable $l1 is not declared␤at /tmp/9yek541Jg4:1␤»
masak r: sub c { state @c = 1; return @c[$^n] if @c > $n; 2 * (2 * $n - 1) / ($n + 1) * c($n - 1) }; sub b { return "[]" x $^l if $l < 2; my $r = (^c($l)).roll; return "[&b($l - 1 - $_)]&b($_)" if $r < (my $s += c($l - 1 - $_) * c($_)) for ^$l }; say b 5
p6eval rakudo fee891: OUTPUT«[[[]][[]]]␤»
masak r: sub c { state @c = 1; return @c[$^n] if @c > $n; 2 * (2 * $n - 1) / ($n + 1) * c($n - 1) }; sub b { return "[]" x $^l if $l < 2; my $r = (^c($l)).roll; return "[&b($l - 1 - $_)]&b($_)" if $r < (my $s += c($l - 1 - $_) * c($_)) for ^$l }; say b 10
p6eval rakudo fee891: OUTPUT«[[][[[]]][]][][][][]␤»
masak \o/
note: no implicit or explicit computation of a big triangular matric required. 16:20
16:20 kaleem left
masak matrix* 16:23
16:24 thou joined, NamelessTee joined 16:30 Patterner left 16:31 birdwindupbird left 16:39 vytas joined
japhb perl6: ?(all() ~~ /^ \d+ $/) 16:39
phenny japhb: 07:26Z <jnthn> tell japhb not yet *but* the bs branch landing was the blocker for that and I already have some constant folding code, just not wired up yet. So you can expect it to be in place soon.
p6eval rakudo fee891: OUTPUT«This type cannot unbox to a native string␤ in any <anon> at src/gen/BOOTSTRAP.pm:105␤ in any !cursor_init at src/stage2/QRegex.nqp:881␤ in method ACCEPTS at src/gen/CORE.setting:8275␤ in block <anon> at /tmp/aKQYyCuIaz:1␤␤»
..pugs b927740, niecza v15-4-g1f35f89: ( no output )
japhb phenny, tell jnthn Ah, great, thanks! 16:40
phenny japhb: I'll pass that on when jnthn is around.
16:40 vytas left, Psyche^ joined, Psyche^ is now known as Patterner
japhb perl6: ?(all() ~~ /./) 16:40
p6eval rakudo fee891: OUTPUT«This type cannot unbox to a native string␤ in any <anon> at src/gen/BOOTSTRAP.pm:105␤ in any !cursor_init at src/stage2/QRegex.nqp:881␤ in method ACCEPTS at src/gen/CORE.setting:8275␤ in block <anon> at /tmp/wBfthRSBT2:1␤␤» 16:41
..pugs b927740, niecza v15-4-g1f35f89: ( no output )
japhb ^^ Any more golfing that anyone can see for the above bug?
16:46 alester left
japhb r: all() ~~ /./ 16:47
p6eval rakudo fee891: OUTPUT«This type cannot unbox to a native string␤ in any <anon> at src/gen/BOOTSTRAP.pm:105␤ in any !cursor_init at src/stage2/QRegex.nqp:881␤ in method ACCEPTS at src/gen/CORE.setting:8275␤ in block <anon> at /tmp/3G3E2ZywtM:1␤␤»
japhb Another 3 chars off 16:48
:-)
For some reason, I enjoy golfing errors more than working code. :-)
masak there's more feedback in golfing errors :)
japhb: should I submit this?
japhb masak, Yeah, it's probably small enough to be clear at this point. 16:49
masak submits rakudobug
japhb Thanks!
masak bows
japhb BTW, when did r: and n: get added to p6eval? Very nice. 16:50
moritz 9 days ago 16:51
japhb Thank you.
masak p: say "And meeeee!" 16:52
p6eval pugs b927740: OUTPUT«And meeeee!␤»
japhb How could I forget you, Pugs? 16:53
moritz did, too
16:54 ab5tract left
japhb So ... S17-concurrency -- it looks like it was redesigned (more than once?), but it still looks like a collection of ideas rather than something that's had any implementation behind it. Is my assessment correct here? And is anyone attached to the current design? 16:55
moritz yes, and no 16:56
japhb Ah, good.
moritz sorear++ has implemented a tiny bit of concurrency, but I don't know how well that corresponds to S17 16:57
japhb What did he implement?
(sorear, are you around?)
16:58 overrosy left
japhb I remember back a long time ago, one of the compilers (Pugs?) experimented with STM, but I recall that being at most a mixed success. 16:58
masak decommutes 17:01
17:08 kaare_ joined 17:15 MayDaniel joined 17:18 MayDaniel left 17:22 fglock joined
[Coke] r: say "" ~~ /./ 17:29
p6eval rakudo fee891: OUTPUT«#<failed match>␤»
[Coke] r: say Mu() ~~ /./ 17:30
p6eval rakudo fee891: OUTPUT«Method 'Mu' not found for invocant of class 'Parcel'␤ in <anon> at src/gen/BOOTSTRAP.pm:805␤ in <anon> at src/gen/BOOTSTRAP.pm:802␤ in any <anon> at src/gen/BOOTSTRAP.pm:797␤ in block <anon> at /tmp/aButAF_v15:1␤␤»
[Coke] r: say Mu ~~ /./
p6eval rakudo fee891: OUTPUT«use of uninitialized value of type Mu in string context␤#<failed match>␤»
[Coke] I'm not sure why any() ~~ /./ is worse than that.
er, all() 17:31
(what does all() mean, anyway?)
(that is, what is a junction of everything of nothing?) 17:33
r: say all() ~~ 10
p6eval rakudo fee891: OUTPUT«all()␤»
japhb I'm guessing it's junction autothreading that makes one()/any()/all() worse than Mu in this case.
[Coke] r: say Nil ~~ /./
p6eval rakudo fee891: OUTPUT«#<failed match>␤»
moritz nom: say nqp::unbox_s(Junction) 17:34
p6eval rakudo fee891: OUTPUT«Cannot unbox a type object as a native str␤ in block <anon> at /tmp/PEbdxSyZx0:1␤␤»
japhb And for the purposes of this bug, it doesn't matter whether the junction composers have arguments or not.
moritz nom: say nqp::unbox_s(all())
p6eval rakudo fee891: OUTPUT«This type cannot unbox to a native string␤ in block <anon> at /tmp/C3_koKdTvx:1␤␤»
[Coke] shuts off the part of his brain that's thinking about this.
japhb r: nqp::unbox_s(all())
p6eval rakudo fee891: OUTPUT«This type cannot unbox to a native string␤ in block <anon> at /tmp/eZYLZwBom3:1␤␤»
japhb 'say' not necessary
[Coke] we're not suggesting that unbox_s(junction) should do something, though, are we? 17:35
just that ~~ should autothread better? 17:36
japhb Right.
[Coke] hokay
japhb The tests with unbox_s are just to check if that's what caused the exact error message we're getting, I think.
17:37 MayDaniel joined, mj41 left 17:41 Vespakoen left 17:44 simcop2387 left 17:46 pmurias left 17:47 simcop2387 joined
[Coke] +# 03/13/2012 - rakudo++ ; niecza (96.22%); pugs (38.97%) 17:51
+"niecza", 20309, 1, 757, 1540, 22607, 23763
+"pugs" , 8227, 0, 3022, 1344, 12593, 23572
+"rakudo", 21106, 31, 626, 1888, 23485, 24026
niecza inches up 2 tests while everyone else holds.
17:55 mj41 joined, dakkar left 18:00 mj41 left 18:06 MayDaniel left 18:07 NamelessTee left
cognominal error golfing, we should propose that sport to the olympic committee 18:09
japhb I can just see the events: fewest strokes to demonstrate a given error, fewest strokes to trigger a VM-native error, fewest strokes to a segfault ... 18:11
18:14 awwaiid left
japhb phenny, ask sorear What have you implemented in the way of concurrency support in Niecza? (I mean that in the general sense, including threads, AIO, event loops, STM, actors, ....) 18:20
phenny japhb: I'll pass that on when sorear is around.
TimToady perl6: say (<[ ]> xx 10).pick(*).join 18:21
p6eval niecza v15-4-g1f35f89: OUTPUT«[][[]][[][]]]][[[]][␤»
..pugs b927740: OUTPUT«[]]][[]]][[][][][[[]␤»
..rakudo fee891: OUTPUT«[][][][][][][][][][]␤»
TimToady rakudo feyl
moritz should it flatten all those parcels? 18:23
TimToady that's the default for most listy methods, which is why we need .lol 18:25
admittedly the spec could be clearer here
japhb TimToady, how "deep" a concurrency model are you really looking for? Right now the best feel I've got is "1. In general syntax, give the programmer ways to specify that things *can* be done concurrently, but leave reality to the implementations (e.g. hyper, race). 2. If the user wants to have greater control, give them features that allow them to do "native" threading, events, etc. but with a portable veneer that hides platform dependent 18:26
details." Is that about right?
TimToady perl6: say (<[ ]> xx 10).map(~*).join
p6eval pugs b927740: OUTPUT«*** Cannot cast from VList [VStr "Inf"] to VCode (VCode)␤ at /tmp/LSDUtFmiID line 1, column 5 - line 2, column 1␤»
..rakudo fee891, niecza v15-4-g1f35f89: OUTPUT«[][][][][][][][][][]␤»
TimToady you'll note rakudo flattens there, so it's being inconsistent 18:27
perl6: say (<[ ]> xx 10).lol.map(~*).join
p6eval niecza v15-4-g1f35f89: OUTPUT«Unhandled exception: Unable to resolve method lol in class List␤ at /tmp/dbGphD48ko line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3838 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3839 (module-CORE @ 65) ␤ at /home/p6…
..pugs b927740: OUTPUT«*** No such method in class Array: "&lol"␤ at /tmp/hxg5aZza1N line 1, column 5 - line 2, column 1␤»
..rakudo fee891: OUTPUT«Method 'lol' not found for invocant of class 'List'␤ in block <anon> at /tmp/jZqVfffvUO:1␤␤»
TimToady perl6: say (<[ ]> xx 10).tree.map(~*).join 18:28
p6eval niecza v15-4-g1f35f89: OUTPUT«Unhandled exception: Unable to resolve method tree in class List␤ at /tmp/vp42Bk4zo1 line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3838 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3839 (module-CORE @ 65) ␤ at /home/p…
..rakudo fee891: OUTPUT«[ ][ ][ ][ ][ ][ ][ ][ ][ ][ ]␤»
..pugs b927740: OUTPUT«*** No such method in class Array: "&tree"␤ at /tmp/BSEAjlSVcZ line 1, column 5 - line 2, column 1␤»
18:47 tyatpi joined
TimToady japhb: yes, that's pretty close to the mark, though I'd view events more in category 1, with some way to say "each dispatch to this object/multi spawns a thread" 18:50
18:54 overrosy joined 18:57 birdwindupbird joined 19:02 overrosy left
japhb TimToady, ah, OK, thank you. 19:04
[Coke] so, now that i'm trying to run pugs ona mac, I have the same problem ... benabik? had. all the test files seem to be in a utf-8 or us-ascii charset. so why is pugs complaining about utf-8 (and requiring iso8859-1 on linux?) 19:06
benabik [Coke]: I'm pretty sure GHC is reading the file and pre-processing the UTF8 19:07
I have no idea if this is caused by version of GHC, OS, or what.
[Coke] benabik: what version of ghc are you using on os x? 19:08
(I'm using 6.10.4)
benabik [Coke]: 7.0.4 from haskell-package
*-platform 19:09
19:09 overrosy joined
[Coke] tries to find a test file that complains. 19:11
sorear good driveby #perl6 19:15
phenny sorear: 18:20Z <japhb> ask sorear What have you implemented in the way of concurrency support in Niecza? (I mean that in the general sense, including threads, AIO, event loops, STM, actors, ....)
colomon \o
[Coke] benabik: so I just ran it with no setting, and didn't get that many failures. 19:16
(and more TODO passes than failures)
benabik No setting? 19:17
[Coke] LC_ALL is not set
sorear japhb: that which is in lib/Threads.pm6, no more, no less
japhb: it's just Java/CLR style threads+monitors concurrency primitives 19:18
benabik [Coke]: I do that too but still get lots of decodeUTF8 warnings. Is LANG set?
[Coke] (I didn't save the output, so I don't know if I got a lot of the warnings, but it doesn't seem to impact the tests on this box.)
sorear I want somthing realistic written on top of it
[Coke] benabik: can you tell me a file you get the warnings in?
benabik t/spec/integration/advent2009-day07.t
[Coke] benabik: LANG is set to en_US.UTF-8
nope, seems fine. 19:19
weird.
19:19 overrosy left 19:21 ascrazy joined 19:24 havenn joined
japhb sorear, Just to be clear, you mean "monitors" in the sense of "object with all methods implicitly run under mutex", not as in "registered event listeners", yes? 19:26
japhb has discovered in the last week that the concurrency literature is anything but consistent in the use of certain common words ....
19:27 pernatiy left
arnsholt japhb: Welcome to academic literature. Everyone uses their preferred terminology =) 19:30
TimToady japhb: I might argue that event dispatch to an object should behave as a monitor because objects are implicitly stateful, but dispatch to a multi should not, on the assumption that the multi code is likelier to be FP oriented 19:33
so should probably handle its own atomic {} bits 19:34
japhb arnsholt, The last time I went deep on fundamental concepts (as opposed to tight specializations, e.g. GPU programming), I was an undergrad, and my professors seemed to do a decent job of agreeing on common terminology. With this kind of stuff? Not so much.
19:34 pochi joined
TimToady in this case I'm scoping the "monitor" to the object, not the class, so I guess it's not a classical monitor 19:34
npi 19:35
heavy revision to all but first entry in rosettacode.org/wiki/Balanced_brackets#Perl_6 19:36
japhb arnsholt, though given that it's been a couple decades since college, perhaps the academics have just had more time to come up with diversity of definition. :-) 19:38
TimToady but certainly per-object locking/queueing will scale better than per-class, which doesn't even make a lot of sense under inheritance 19:41
japhb TimToady, I'm not entirely sure I buy that dispatch to a multi should require the programmer to know when to atomic {}. There may be a use case I haven't thought of yet, but it seems at present that's mental load that needn't be there, if you're already fine with single method dispatch for events using monitor semantics
TimToady, agreed with that last statement (and honestly, I can't see a strong need for per-class locking that isn't better served with a different WTDI)
TimToady I'd also like to be able to easily write code with multiple transactions running through it simultaneously, though it doesn't necessarily have to be default 19:42
but I'd like some way of marking the proto, so you don't have to throw async into every body 19:43
proto foo () is atomic vs proto foo () is async, I suppose
we also need a way to dispatch that doesn't expect a response, or that manages any response via some callback mechanism 19:45
sink context isn't good enough, since you need to distinguish a sequence point at the ; boundary for imperative programma\ing 19:46
19:47 awwaiid joined
japhb TimToady, can you expand what you mean by that last statement? 19:47
TimToady foo; bar has to wait for foo to finish
send foo; bar doesn't
fsdo send
felher really likes the FP-oriented solution of rosettacode.org/wiki/Balanced_brackets#Perl_6 :) 19:48
TimToady++
japhb Oh I see what you mean, you're saying that sink context shouldn't imply asynchrony automatically.
TimToady yes, we need to mark the 'go' routine calls :)
19:49 thou left
TimToady we need a shorter verb for fire-and-forget 19:49
faf foo :)
japhb BTW, I noted an attempt in the existing S17-concurrency to try to do some of that "async return" via CONTROL exceptions. What are your thoughts on that? 19:50
TimToady maybe it's another contextualizer...maybe I already added it, and it's 'race foo' 19:51
19:51 thou joined
TimToady well, race in sink context, to thorw away the return 19:52
japhb: I think that's probably insane
unless we teach exceptions to work like that anyway 19:53
we do need a story for where thread exceptions go
japhb I think I had roughly the same sequence of thoughts. 19:54
TimToady but merely blowing up the spawner when the spanee goofs up seems a bit untoward
19:54 araujo left
japhb Eh, what's a little blowing up between friends? 19:54
TimToady druther have a bit more error encapsulation there somehow
much like a waitpid() can tell you how the child process finished 19:55
19:56 daxim left
TimToady but possible propagation to the parent might be a default, in the absense of other handling 19:56
hafta think about that
japhb Maybe "child thread needs to report an error" is just another event you can listen to 19:58
(going with the SIGCHLD path)
TimToady or if you are emulating call/return, could callback with a failure value 20:02
more the "you made a promise and then broke it" model 20:03
20:04 araujo joined
TimToady if we allow async routines to return lazily via promises 20:04
another way of breaking the sequence pointness of ;
but we've avoided scalar laziness for the most part till now 20:05
but you can view a promise as a scalar feed, a pipe with one value coming on it 20:06
so certainly promises can be emulated, if not native
in any case, feeds are very much *not* sequence points 20:07
hmm, nobody ever fed me lunch... 20:08
no wonder my head feels strange(r)
20:16 Guest80268 left 20:17 overrosy joined 20:19 retup_work left
masak submits (<[ ]> xx 10).pick(*).join rakuobug 20:22
TimToady: '\\' «leg« $s.comb # that's sick/cool :P 20:23
TimToady darn, missed the chance for a Xleg pun though
and X can be lazy, so we could reject a string in the middle that hits gambler's ruin without combing it all out 20:24
masak I think merlyn would go on a rampage if he saw the results of infix:<leg> being summed by [\+] -- or is it just summing boolean values that he doesn't like? 20:25
TimToady assuming .none is also sufficiently lazy
yes, well, he would not like the interchangability of map/for either, I suspect 20:26
masak I remember one p6l email where he said something like "Bool::True should be 42 or something to dissuade people from doing this kind of stuff!"
TimToady but we all change our views over time, so I don't want to prevent anyone from repenting earlier views :)
masak nodnod 20:27
it just stayed with me for some reason.
nom: constant True = 42; say True
p6eval rakudo fee891: OUTPUT«True␤»
masak hm.
benabik r: constant True = 42; say +True
masak submits rakudobug
p6eval rakudo fee891: OUTPUT«1␤»
benabik Huh
masak good rakudobug day today ;)
benabik Oh, I was thinking it was an enum not getting numified. blah. 20:28
jnthn Great, not being able to change the truth is a bug :P 20:30
phenny jnthn: 16:40Z <japhb> tell jnthn Ah, great, thanks!
20:31 retup_work joined
TimToady the opposite of OUTER truth is INNER truth... :P 20:31
ranguard Hi, is there someone who looks after perl6.org? - basically I'd like to redirect dev.perl.org/perl6/* somewhere and didn't know if you want to create an archive and host it? 20:32
20:32 pmurias joined
ranguard isn't sure what is useful / what's not etc 20:33
masak ranguard: hi, we all look after perl6.org -- some more than others.
are you saying dev.perl.org/perl/* is going away?
TimToady if you have a github account, you could put it there yourself
jnthn masak: perl6/* 20:34
masak uhm. meant that. 20:35
ranguard masak: hthe stuff in /perl6/ is not current, and just a bit confusing, I'd like to just link to it on the perl6.org site then you've got everything in one place
20:35 birdwindupbird left
masak sounds like a generally good idea to collect things under perl6.org, yes. 20:35
ranguard TimToady: ok, will do :)
TimToady hugme: add ranguard to perl6 20:36
hugme TimToady: You need to register with freenode first
masak I've always loved this page: dev.perl.org/perl6/architecture.html
[Coke] hugme: add ranguard to perl6
hugme [Coke]: sorry, you don't have permissions to change 'perl6'
[Coke] hugme: biteme
TimToady hugme: add ranguard to perl6
hugme hugs ranguard. Welcome to the perl6 github organization
ranguard thanks :) 20:37
TimToady uh, assuming that's your github nick
ranguard it is
TimToady have the appropriate amount of fun
hopefully not negative
ranguard I'll call it /archive/ and put everything in there - can be pulled out by someone who knows what stuff actually is :) 20:38
TimToady thanks 20:39
ranguard github.com/perl6/perl6.org is still readonly for me :( 20:41
TimToady checking 20:43
jnthn iirc, you need [email@hidden.address] in order to push
arnsholt You do. HTTP isn't really a write-y protocol 20:44
But if you've cloned from HTTP, you can switch over to the SSH one by editing .git/config 20:45
ranguard Sorry, I mean if I goto that URL github only shows me the read-only URLs 20:46
20:46 overrosy left
TimToady they had some security issues recently, so maybe passwords got reset of something 20:46
japhb You *can* do write access via HTTPS to git in general, but perhaps just not github.
TimToady it's making me say I forgot my passowrd...
doy fairly sure you can write to github over https 20:47
japhb The passwords didn't get reset, but all of the SSH keys got de-trusted. You have to manually re-trust them.
arnsholt TimToady might be onto something. They had hole that let people upload keys, so you have to audit your keys
japhb arnsholt, exactly.
20:47 mj41 joined
TimToady I'm guessing hugme is not considered trusted 20:49
and I seem to have misplaced the button to add new users on the website o_O 20:53
okay, you're added 20:55
(I think)
ranguard TimToady: wonderful, thanks :)
20:56 vlkv joined
dalek href="https://perl6.org:">perl6.org: c210507 | (Leo Lapworth)++ | README:
Fix spelling
 20:57
20:57 vlkv is now known as gv
TimToady obviously it worked \o/ 20:57
ranguard guess it worked, I'll sort it in the next couple of days :)
masak ++ranguard
20:59 ksi joined 21:00 fglock left
japhb (backlogging) TimToady, going back and reading S06's section on the feed operator, it looks like you had already specced feed to a scalar to be relatively smart: spawning a thread on a the blunt end, making the scalar contain a lazy, fully structured output from that thread. As long as we say that exceptions from that subthread are not thrown immediately but instead show up in the feed result, that gets us halfway there. 21:08
21:11 pernatiy joined 21:16 bluescreen100 joined 21:18 skids left 21:26 bluescreen100 left, bluescreen10 left 21:40 sudokode left 21:42 havenn left
masak does `my @nums of Int` mean the same as `my Int @nums`? 21:42
ah; S09 seems to indicate it does. 21:43
21:44 kaare_ left
jnthn Yes 21:45
Though not in Rakudo yet.l
Also, why are you calling it @nums if it's full of Ints :P 21:46
21:47 thou left
japhb @lies 21:48
pyrimidine @stars
masak .oO( because I'm a p6cc contestant, and my code is being ruthlessly deconstructed by strangers )
benabik my @it of Star
masak nom: my Int @a = 1, 2, "HAHAHA", 4; say "alive" 21:49
p6eval rakudo fee891: OUTPUT«alive␤»
masak nom: my Int @a = 1, 2, "HAHAHA", 4; say @a
p6eval rakudo fee891: OUTPUT«1 2 HAHAHA 4␤»
21:49 donaldh joined
masak nom: my @a of Int = 1, 2, "HAHAHA", 4; say @a 21:49
p6eval rakudo fee891: OUTPUT«1 2 HAHAHA 4␤»
masak neither works in enforcing type, it seems.
jnthn masak: Only for indexing.
masak so I don't see what jnthn means by "not in Rakudo yet" :)
oh, indexing.
nom: my Int @a; @a[0] = "HAHAHA"; say "alive"
p6eval rakudo fee891: OUTPUT«Type check failed in assignment to '$v'; expected 'Int' but got 'Str'␤ in block <anon> at /tmp/RuJsGGlCTw:1␤␤» 21:50
masak nom: my @a of Int; @a[0] = "HAHAHA"; say "alive"
p6eval rakudo fee891: OUTPUT«alive␤»
masak ah.
jnthn I think I didn't do it for the assignment because I expected pmichaud++ to be fixing that up at some point "soon". :/
21:55 thou joined
jnthn (iirc there is still some buglet in assignment, but it's a corner case) 21:55
It's the one that very occasionally means you see whinings about splicing Mu.
masak ah.
21:59 ksi left 22:08 att left 22:09 FACEFOX left 22:10 FACEFOX joined, att joined
masak nom: my $s = "ha ha ha haaa ha haa"; say ($s ~~ m:g/ha+/).WHAT 22:17
p6eval rakudo fee891: OUTPUT«Bool()␤»
masak "if you want the list of matches in Perl 6, use .comb", right?
22:17 retup_work left
masak it's not reasonable of me to expect m:g// to return a list of things, right? 22:18
benabik r: (m:g/ha+/).WHAT.say
p6eval rakudo fee891: OUTPUT«Method 'match' not found for invocant of class 'Any'␤ in block <anon> at /tmp/zHslu2LtBQ:1␤␤»
22:19 retup_work joined
masak nom: my $s = "ha ha ha haaa ha haa"; say $s.comb(/h(a+)/) 22:19
p6eval rakudo fee891: OUTPUT«ha ha ha haaa ha haa␤»
22:20 pmurias left
masak nom: my $s = "ha ha ha haaa ha haa"; say $s.comb(/h(a+)/, :match) 22:21
p6eval rakudo fee891: OUTPUT«ha ha ha haaa ha haa␤»
masak nom: my $s = "ha ha ha haaa ha haa"; say $s.comb(/h(a+)/, :match)>>[0]
p6eval rakudo fee891: OUTPUT«=> <a>␤ => <a>␤ => <a>␤ => <aaa>␤ => <a>␤ => <aa>␤␤»
masak nom: my $s = "ha ha ha haaa ha haa"; say $s.comb(/h(a+)/, :match)>>[0]>>.Str
p6eval rakudo fee891: OUTPUT«a a a aaa a aa␤»
jnthn -> sleep, night
felher o/ 22:23
22:23 mdupont left
masak 'night, jnthn-san. 22:25
22:26 tyatpi left 22:28 FACEFOX left 22:38 jlaire left 22:39 jlaire joined, Jerkwad joined 22:40 Jerkwad left
masak 'night, #perl6 22:41
tadzik 'evening #perl6 22:44
felher night masak, welcome tadzik :)
tadzik huh, am I a night owl or what :) 22:45
22:45 tr-808_ joined 22:46 tr-808_ left
felher sure ;) 22:46
tadzik I was looking for a new phone today; I have a feeling that I only like things that angular (edgy?) 22:48
22:49 att left
felher and, were you lucky? 22:50
tadzik well, I think I'll settle for Nokia N900, which is not available in Poland anymore it'd seem :)
22:51 overrosy joined
felher tadzik: now that i see you and you are into all the POD-stuff: how do the specs get converted to html? 22:51
ah, i phone with a real keyboard, not just touch :) 22:52
tadzik felher: using github.com/perl6/Pod-To-HTML/blob/...To/HTML.pm
felher: essentialy, it looks like when it'd collide with a car, the car will turn out more damaged ;) 22:53
felher tadzik: that is _only_ real kind of phone :D 22:54
tadzik++ # pod2html
tadzik yeah, a true descendant of good, ol' 3310
flussence++ for all the real work
felher flussence++ :)
22:55 mj41 left, overrosy left 23:01 donaldh left 23:02 skids joined, Vespakoen joined 23:05 wentam joined 23:06 sudokode joined 23:07 benabik left 23:10 PacoAir left
wentam how would I match anything except for a particular character in a regex? 23:12
tadzik <-[c]> 23:13
23:14 Vespakoen left
tadzik that is "every character but not »c«" 23:14
wentam tadzik, ty
tadzik yw
welcome to #perl6, we have cookies 23:15
moritz returns from a lost table tennis game, and wants a cookie
tadzik poor moritz. Have a cookie!
moritz -> sleep 23:18
moritz waves to the cookie monster
tadzik waves back
have a nice rest
23:30 benabik joined
colomon cookies? 23:35
23:35 whiteknight joined
tadzik yeah, go ahead 23:35
colomon what kind?
tadzik whatever you want, I'm paying today :) 23:36
colomon \o/
benabik solid gold. :-D
colomon uck, too chewy
tadzik cookie gold at least
flussence
.oO( hey now wait a minute, I only wrote part of a thing that does stuff with a Pod AST, tadzik++ wrote the parser - that's the hard part! )
tadzik flussence: from us two, you're the only who knows HTML :P 23:37
colomon: have another, today we party :)
colomon mmmm, peanut butter 23:38
tadzik . o O ( much butter )
23:44 thou left 23:45 wolfman2000 joined 23:52 gv left

Powered by Cro and the Raku® Programming Language.

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

Thank you!