»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.
sorear TimToady: I like the way LALR parsers can always tell you the exact location of the error and the set of expected symbols. 00:00
hello and welcome, Araq
Araq thanks
00:00 att joined
Araq unfortunately I have to sleep soon 00:00
sorear dives into backlog for more on Araq 00:01
Araq maybe I'll come up with another good question before that though :-)
TimToady sometimes good questions have good answers...sometimes the best answer is "huh...hmm..." :) 00:02
00:03 havenn joined 00:04 jferrero left, jferrero joined
sorear Araq: I'm back - that didn't take long 00:05
`patch` has anyone started work on Str.normalize and the normalization forms? 00:08
Araq so if I have operators * and + and then introduce *+ as an operator the longest match rule disambiguates *+ right? 00:11
TimToady correct
Araq what if I don't have a *+ operator? is *+ parsed as two tokens or is it an error? 00:12
TimToady well, it would either be a place where those two tokens make sense in that order, or it'd be an error
perl6: say 2 *+ 5 00:13
p6eval pugs b927740, rakudo 3ca6e5, niecza v14-97-gacd7da0: OUTPUT«10␤»
TimToady perl6: sub infix:<*+> ($a,$b) { $a * $b + $b }; say 2 *+ 5 00:14
p6eval niecza v14-97-gacd7da0: OUTPUT«15␤»
..pugs b927740, rakudo 3ca6e5: OUTPUT«10␤»
TimToady pugs is a bit behind the times
Araq how do sigils and prefix operators get disambiguated?
TimToady sigils are consider part of the term, for the most part 00:15
EXPR never sees them as prefixes
note that the LTM knows it's looking at both terms and prefixes in the same alternation 00:16
so it's not really terribly important which way things are classified
00:17 RandomMonkey joined
TimToady the only difference between a term and a prefix is what they expect after them 00:17
Araq I can't follow. Do you mean they are basically only differentiated in the semantic checking phase? 00:18
00:18 RandomMonkey left
TimToady the 'termish' rule will look for a sequence of prefixes followed by a single term followed by a sequence of postfixes 00:19
so a prefix keeps looking for prefixes and terms, while a term will start looking for postfixes
EXPR mostly only cares about "termish" vs "infixish" 00:20
Araq ok so !ident and $ident are the same
(where ! is a prefix operator and $ is a sigil)
TimToady don't understand the questoion
don't understand 'the same'
Araq I mean syntactically there are the same
I mean the parser treats ! like $ 00:21
TimToady the EXPR method could expect to see either of those at 'termish' time
Araq yeah that's what I mean
TimToady but ! is prefix:<!>, while $ is the <sigil> part of <variable>
but they all get |'d together into a single alternation, as far as LTM is concerned 00:22
Araq hm ok, but what if I have an operator $ ?
or is that not allowed?
TimToady if it's ambiguous, one or the other will go first; the PEG rules are pretty detailed 00:23
Araq perl6: sub infix:<$>($a) { $a; }
p6eval niecza v14-97-gacd7da0: OUTPUT«Potential difficulties:␤ &infix:<$> is declared but not used at /tmp/Eq4gTqbDHL line 1:␤------> sub infix:<$>⏏($a) { $a; }␤␤»
..pugs b927740, rakudo 3ca6e5: ( no output )
Araq perl6: $4 00:24
p6eval pugs b927740, rakudo 3ca6e5, niecza v14-97-gacd7da0: ( no output )
TimToady no state
Araq should have used 'prefix'? 00:25
perl6: sub prefix:<$>($a) { $a; }
TimToady definitions don't carry over
p6eval rakudo 3ca6e5: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&a' called (line 1)␤»
..niecza v14-97-gacd7da0: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'a' used at line 1␤␤Potential difficulties:␤ $a is declared but not used at /tmp/XzoGhOX3Le line 1:␤------> sub prefix:<$>(⏏$a) { $a; }␤␤Unhandled exception: Check failed␤␤ at /home/…
..pugs b927740: ( no output )
Araq perl6: sub prefix:<$>($a) { return $a; } $5;
p6eval niecza v14-97-gacd7da0: OUTPUT«===SORRY!===␤␤Strange text after block (missing comma, semicolon, comment marker?) at /tmp/6aEmsQTN4E line 1:␤------> sub prefix:<$>($a) { return $a; }⏏ $5;␤Other potential difficulties:␤ $a is declared but not used a…
..pugs b927740: OUTPUT«*** No such subroutine: "&a"␤ at /tmp/6iVFbMDWFF line 1, column 29-31␤»
..rakudo 3ca6e5: OUTPUT«===SORRY!===␤Confused␤at /tmp/RF7x2Vb3AI:1␤»
TimToady obviously it's preferring your operator there :)
perl6: sub prefix:<$>($a) { ::('$a') }; say $5 00:27
p6eval pugs b927740: OUTPUT«pugs: Internal error:␤ Invalid sigil ":$"␤Please file a bug report.␤»
..niecza v14-97-gacd7da0: OUTPUT«Potential difficulties:␤ $a is declared but not used at /tmp/YRhvM5xNvj line 1:␤------> sub prefix:<$>(⏏$a) { ::('$a') }; say $5␤␤5␤»
..rakudo 3ca6e5: OUTPUT«5␤»
TimToady there you go
Araq lots of fun ahead of me :-) 00:28
but ugh, it's too late already
thanks for the answers
good night
TimToady rest well
00:29 Chillance left, Araq left
TimToady somehow I don't think Araq lives in the Cocos Islands :) 00:29
00:36 alester left 00:41 noam_ left 01:01 havenn left 01:07 jaldhar joined 01:12 scott_ joined 01:13 scott_ is now known as Guest43686, tokuhirom joined 01:20 aloha left, aloha joined
sorear so, what should I put in the announcement about pmurias' P5 stuff? 01:26
it seems he has missed the submission deadline for documentation
and I don't understand it well enough to write up anything
colomon P5 made progress? 01:28
the art of the vague 01:29
sorear sounds good
01:39 NamelessTee left 01:40 wooden left
dalek ecza: f52f360 | sorear++ | docs/announce.v15:
Add an announce.v15 draft
 01:41
01:42 wolfman2000 joined 01:43 rafl joined 01:46 havenn joined
TimToady looks good 01:49
colomon +1 01:54
lichtkind good night 01:56
01:56 lichtkind left 01:58 kmwallio joined
dalek ecza: 39faf4e | sorear++ | FETCH_URL:
Update bootstrap to v15
 02:14
02:21 PacoAir left 02:29 baest joined 02:33 tokuhirom left 02:40 ascent_ joined
sorear who is Miroslav Silovic? 02:41
02:46 havenn left 02:49 whiteknight left 02:54 ribayr left
dalek ast: 28ef8dc | coke++ | / (43 files):
pugs fudges
 02:57
gs.hs: 725cddd | coke++ | t/spectest.data:
run fudged tests
02:57 kmwallio left 03:00 woosley joined
[Coke] "pugs" , 5497, 6, 2174, 583, 8259, 23561 03:01
pugs (27.07%); if nothing else changes by the run tomorrow.
colomon :) 03:02
dalek gs.hs: 8d195b3 | coke++ | t/spectest.data:
run tests
 03:12
03:12 xinming left 03:14 xinming joined 03:19 orafu left, orafu joined 03:20 xinming left 03:21 lue left 03:22 am0c left 03:31 baest left 03:32 xinming joined 03:38 sftp left 03:41 mucker left, baest joined 03:59 havenn joined 04:04 Patterner left, Guest43686 left 04:08 am0c joined 04:11 Psyche^ joined, Psyche^ is now known as Patterner 04:31 havenn left 04:33 havenn joined 04:42 kaleem joined 04:54 Jerkwad left 04:55 kaleem left 05:10 birdwindupbird joined 05:11 mucker joined 05:15 skids left 05:22 am0c_ joined 05:24 am0c left 05:28 lestrrat left 05:29 lestrrat joined 05:33 thou left 05:35 am0c_ left, noam joined 05:36 kaleem joined 05:45 am0c_ joined 05:48 retupmoca left 05:56 havenn left, havenn joined 05:58 am0c_ left 06:04 birdwind1pbird joined
TimToady rosettacode.org/wiki/Arithmetic-geo...ean#Perl_6 06:10
TimToady wonders if we should have a fixed-point target for ... to aim at 06:13
06:18 birdwind1pbird left
sorear just so long as nobody confuses it with Haskell's fix, which takes fixed points in an entirely different way 06:21
geekosaur has to admit that was how he read it initially 06:27
(and tried to apply it to the ... operator.)
06:29 samlh joined 06:30 wolfman2000 left 06:33 Alias left 06:34 Alias joined
moritz o/ 06:45
sorear /o moritz 06:51
07:03 noam left 07:04 noam joined
samlh Yay: gist.github.com/1930305 07:09
still need to figure out a way to hook up seeking and pausing 07:10
moritz cool 07:12
samlh++
samlh moritz: been a long time lurker, probably time to get out of the woodwork 07:13
07:14 plobsing left
sorear o/ samlh 07:14
moritz samlh: welcome in the unlurk-world :-)
s/in/to/ 07:15
samlh sorear: \o 07:17
sorear: thanks for a nice implementation :)
moritz: my final goal is a (lossless) mp3 splitter gui, I'll see how it goes 07:18
sorear: btw, run and qx apparently don't run on windows (no big deal) 07:19
07:20 noam_ joined
sorear samlh: for run that's known, but qx *should* work 07:20
if you're having qx issues on Windows bring it up with colomon++ 07:21
samlh sorear: ok, thanks
moritz sorear++ # niecza v15
sorear samlh: btw, you should be able to use .CLR::System::IDisposable::Dispose now instead of .dispose-hack 07:22
07:22 noam left
sorear (not that it's any shorter or faster *sigh* but at least it's less special-casey) 07:22
samlh ah, cool 07:25
qx error gist.github.com/1930380
moritz phenny: tell colomon about gist.github.com/1930380 (qx on windows errors out, samlh++ discovered it) 07:28
phenny moritz: I'll pass that on when colomon is around.
07:56 havenn left 07:57 kaleem left, kaleem joined 08:07 noam joined 08:08 noam_ left 08:16 plobsing joined 08:34 mj41 joined 08:37 havenn joined 08:39 sisar joined
sisar moritz: there ? 08:39
phenny sisar: 27 Feb 17:50Z <moritz> tell sisar I've turned our session into two commits, github.com/rakudo/rakudo/commit/d696855f41 and github.com/perl6/roast/commit/52f6db64fc
moritz sisar: yes 08:40
sisar so i took up line 47 in Hash.pm 08:41
moritz ok 08:42
sisar the method is STORE but 'is hidden_from_backtrace' will that be a problem somehow?
moritz no, not a problem 08:43
sisar (to introduce an exception there)
ok
moritz it just means that the backtrace doesn't list it by default, but the calling location
nom: sub f() { my %h = 1 }; f();
p6eval nom 3ca6e5: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at src/gen/CORE.setting:5546␤ in sub f at /tmp/qjBIod6o9T:1␤ in block <anon> at /tmp/qjBIod6o9T:1␤␤»
moritz uhm, and it doesn't even work :(
sisar you were expecting 'in method STORE' to be missing from the error? 08:44
moritz right
sisar anything else i need to lookout for or is this a straight forward exception with only 'message' as the only method ? 08:46
moritz it's straight forward
08:47 havenn left
sisar i named my clas as X::Hash::Store , or should i use caps: STORE ? 08:47
moritz just remember to add the stub for the exception
sisar will remember 08:48
moritz sisar: I'd suggest X::Hash::OddNumber
or X::Hash::Store::OddNumber
sisar ok, will use X::Hash::Store::OddNumber 08:49
moritz looks forward to seeing a first patch from sisar++
sisar .oO ( i went by the name of fasad earlier, ergo, I already have a rakudo commit :p) 08:51
fasad++
moritz: you missed the anouncement of my nick change 08:52
moritz sisar: or simply forgot. ETOOMANYNICKSAROUNDHERE :-) 08:53
08:54 fhelmberger_ joined, fhelmberger_ left 08:58 havenn joined
dalek ecza: d19c478 | (Solomon Foster)++ | lib/Builtins.cs:
Try to be a little more careful about array boundaries, in hopes of curing the Windows qx issue.
 09:07
colomon samlh: I just pushed a patch which may fix the qx issue. My Windows box is at the shop for repairs, so I can't test it. BTW, are you using mono or .NET? 09:09
phenny colomon: 07:28Z <moritz> tell colomon about gist.github.com/1930380 (qx on windows errors out, samlh++ discovered it)
moritz colomon++
colomon samlh: btw, run still won't work -- I will port the qx code to run once we've got qx working properly. 09:10
samlh colomon: .net 09:13
colomon: thanks for the fix! I just need to get a compilation env up now :)
btw, I've been testing github.com/sorear/niecza/issues/59 - the original bug is gone, just have to comment out a dispose call 09:15
colomon: have any ideas?
09:15 havenn left
sorear samlh: interesting - probably some subtle difference in reflection behavior 09:17
samlh: could you gist a run of sierpinski after 'set NIECZA_CLR_TRACE=1' or however that's spelled now? 09:19
(produces 4278 lines of output here; I only *really* care about the ones with Dispose in them)
although it looks like more context is needed than just what "grep" gives 09:20
colomon heads back to bed... 09:26
09:26 woosley left
jnthn morning 09:30
samlh sorear: gist updated gist.github.com/1931425
09:39 thou joined, thou left
sisar moritz: i have access to the rakudo/nqp dir only, can you give me full access to rakudo 09:40
(commit access)
moritz sisar: I'm sorry, we don't hand out commit bits for rakudo so freely :/ 09:41
sisar: you'll have to make a pull request
sisar moritz: ok, will do
moritz sisar: after a few good patches, you can submit a CLA ("Contributor's License Agreement"), and then the rakudo devs will decide over giving out a commit bit 09:42
sisar ok
can i make a pull request from my terminal ?
moritz sisar: not with the git built-ins, but some clever folks have written extensions that allow that: www.splitbrain.org/blog/2011-06/19-...l_requests 09:43
09:43 dakkar joined
sisar moritz: if i understand it correctly, that extension is for branch oweners to manage pull requests, not for contributors to make a pull request 09:47
moritz oh :/
then I should read more careful in future
sisar moritz: np. i'll fork on github :) 09:48
sorear -> sleep
09:48 proller_ joined, proller joined, proller left
samlh -> sleep too 09:49
moritz oh, defunkt.io/hub/ can do command line pull requests 09:50
09:54 noam left, proller_ left 09:55 noam joined
sisar moritz: github.com/rakudo/rakudo/pull/59 10:01
moritz \o/ 10:02
dalek kudo/nom: fc96b6d | (Siddhant Saraf)++ | src/core/Exceptions.pm:
add class X::Hash::Store::OddNumber
kudo/nom: 584be63 | (Siddhant Saraf)++ | src/core/Hash.pm:
passing an odd number of elements while hash creation now throws X::Hash::Store::Oddnumber
kudo/nom: 83b15cd | moritz++ | src/core/ (2 files):
Merge pull request #59 from Siddhant/nom

typed exception X::Hash::Store::OddNumber
sisar i'll write a test soon 10:03
for now, afk
moritz++
moritz sisar++
10:03 grondilu joined
dalek ar: 419a68c | jnthn++ | Makefile:
Bump version numbers.
 10:05
ar: 0a79ced | jnthn++ | skel/docs/announce/2012.02:
Add first draft of 2012.02 announce.
ar: f41027b | jnthn++ | skel/README:
README version number bump.
10:05 lestrrat left
grondilu I removed the constraints in my sha256 implementation and it's *way* faster. About 15 seconds against 5 minutes! I guess constraints are much too slow. s0.barwen.ch/~grondilu/Crypto.pm6 10:05
phenny grondilu: 27 Feb 18:49Z <TimToady> tell grondilu if you remove the unnecessary 'where' constraints, your sha256 runs 20 times faster
grondilu: 27 Feb 18:51Z <TimToady> tell grondilu if you really want them, change them to 'where ^33' and it still runs 15 times faster
grondilu TimToady: indeed I had noticed that yesterday as well. 10:06
jnthn Review and tweaks welcome, as are suggestions of any extra modules that should go into the Star release.
moritz reviews
10:06 lestrrat joined
jnthn Away for ~15 mins 10:06
grondilu I also had a few perl6 idioms :)
moritz grondilu: if you want to optimize for rakudo, you can use native ints for the loop variables 10:07
but then it won't run in neicza anymore 10:08
*niecza
and you ahve to write $i = $i + 16 instead of $i += 16 for now
jnthn: star release announcement looks good 10:09
grondilu I noticed I can't use the module with the 'use' mechanism for some reason. It's not implemented or something?
moritz grondilu: what's the error you get? 10:10
grondilu: you'll need to export some subs for it to be useful
grondilu by export you mean 'our sub ....', right? 10:11
moritz no, I mean 'sub foo($x) is export { } 10:12
grondilu oh, my bad
but shouldn't I be able to use the function with fully qualified name even if it is not exported? i.e.: perl6 -e 'use MyPackage; MyPackage::foo();' 10:13
moritz well
in your case, everything is inside the module Crypto 10:14
so it would be
use Crypto; Crypto::util::bytesToWords(@bytes)
grondilu anyway I tried stuffs like that several times and it never worked. 10:15
10:15 plobsing left
moritz works here 10:16
./perl6 -e 'use Crypto; say Crypto::sha256::hex "blurb";'
(wait, wait, wait)
2cf24dba5fb0a30e26e83b2ac5b9e29e1b161e5c1fa7425e73043362938b9824
f0ac34ce3cbd715bafcae72f96bd914151c12e8d478bbb645d94d55cf54a1a14
the first one is from the say sha256::hex "hello"; within the module 10:17
moritz hopes grondilu++ can reproduce his success 10:18
grondilu oh it does work indeed. I don't understand what was going wrong. 10:20
moritz maybe a typo 10:21
dalek ar: bc1338d | tadzik++ | skel/docs/announce/2012.02:
Fix a typo in the release announcement
jnthn tadzik: Nice catch :) 10:24
tadzik it's one of those moments, "there's something in that sentence..." when you stare and stare subconsciously :) 10:25
jnthn I'll do the release late afternoon or early evening to allow time for any shiny extra modules to get added to the list. :) 10:29
10:33 daxim joined
tadzik I'll add benchmark there 10:39
and maybe IoC 10:40
dalek p/bs: aba92d4 | jnthn++ | src/6model/sixmodelobject.h:
Stub in SC barriers.
 10:43
p/bs: d770a5a | jnthn++ | src/ops/nqp.ops:
First crack at scattering SC-related barriers in appropriate places in nqp.ops.
jnthn tadzik: OK. Please mention in the announce too :)
tadzik I'll poke it in a while 10:46
10:48 grondilu left
moritz blug: perlgeek.de/blog-en/perl-6/2011-02-...tions.html 10:57
11:04 huf joined
jnthn moritz++ # nice post 11:05
moritz: "a special please" => "a special pleasure"
moritz++ # for the awesome work too :) 11:06
moritz jnthn: fixed, thanks
au "a dedicated; type" => "a dedicated type" or => "a dedicated X::Hash::Store::OddNumber type"? 11:08
moritz++ # awesome new backtraces 11:09
jnthn oops, missed that stray semi
11:10 ab5tract joined
moritz stray colon fixed, au++ 11:11
11:16 plobsing joined 11:25 Trashlord left 11:30 NamelessTee joined
dalek ast: 8c63c26 | moritz++ | S03-operators/eqv.t:
fix rakudo fudge in eqv.t

Rakudo does not like argumentless todo() calls		 11:34
11:34 benabik left 11:36 frettled left 11:39 PacoAir joined 11:41 xinming left 11:42 xinming joined
dalek p/bs: f9eb249 | jnthn++ | src/pmc/serializationcontext.pmc:
Add slots in the SC for holding object repossession info.
 12:00
p/bs: c11251c | jnthn++ | src/ops/nqp.ops:
Implement support for maintaining a stack of currently compiling SCs.
p/bs: 10d5a9c | jnthn++ | src/HLL/World.pm:
Push/pop SC to the compiling stack.
p/bs: e4937be | jnthn++ | src/ (2 files):
First cut at the barrier applicability logic, which just emits a debug message for now.
p/bs: 93cbec9 | jnthn++ | src/Regex.pir:
Regex library needs to push the SC it builds onto the compiling stack while doing so, to prevent spurious repossession; clears up all but two instances in the build of the barrier triggering (which it never should in the NQP build).
12:08 sftp joined 12:21 pernatiy joined 12:25 bluescreen10 joined 12:28 frettled joined 12:30 agentzh joined 12:31 bluescreen10 left 12:36 benabik joined 12:41 noam left 12:42 noam joined 12:45 bluescreen10 joined 12:49 pernatiy left 12:53 pernatiy joined 12:57 dakkar left, dakkar_ joined 12:58 dakkar_ is now known as dakkar
dalek ast: b92c4a8 | moritz++ | S12-introspection/parents.t:
rewrite parents.t to not use infix:<eqv> with Mu as argument
 13:00
13:04 noam left 13:05 Tedd1 left, noam joined 13:06 dakkar left 13:09 nwc10 joined
nwc10 moritz: is the IRC logger written in Perl 6 yet? :-) 13:09
moritz nwc10: no
nwc10: though rakudo supports sockets and enough NCI for database access now, so it would be entirely possible 13:10
nwc10 it would seem like interesting dogfood. Can't say that I'm volunteering, though.
so, currently "Ain't broken, and a bunch of other stuff is more important" seems like a a valid state of affairs. 13:11
jnthn o/ nwc10
nwc10 hi jnthn
nwc10 isn't dead yet, but does now have a cough 13:12
jnthn GPW next week?
nwc10 assuming that I'm not dead, or faking death convincingly
moritz nwc10: it is a tempting idea though
nwc10 bother. I've distracted you from fixing other things :-/
moritz also I'm now logging perl 5's #onionsketch 13:13
13:13 dakkar joined
nwc10 I know. That's why I went to have a look 13:13
jnthn
.oO( was moritz tempted to re-write the irclogger or fake death convincingly? )
moritz so it would be a double win to provide some infrastructure for Perl 5 with a tool written in p6 :-)
nwc10 wanted to know if Perl 6 was doing the logging for something Perl 5
(snap-ish)
13:16 plobsing left 13:18 dakkar left 13:19 nwc10 left 13:20 mucker left, Tedd1 joined, dakkar joined, abercrombie left 13:21 ifaria joined 13:28 dakkar left 13:30 dakkar joined 13:33 simcop2387 left 13:35 simcop2387 joined 13:39 kaleem left 13:48 overrosy left 13:50 overrosy joined 13:53 kaleem joined 14:16 plobsing joined
fsergot hi o/ 14:28
14:28 mucker joined
jnthn o/ fsergot 14:29
14:30 tokuhirom joined, Trashlord joined
gfldex build under cygwin is failing 14:31
it's not copying cygparrot4_1_0.dll into install
moritz :(
benabik ... Wasn't that fixed not that long ago? 14:32
moritz so I thought too
14:32 xinming left
dalek kudo/nom: 137b16b | moritz++ | src/core/Mu.pm:
make infix:<eqv> autothread as per current spec
 14:34
14:34 JimmyZ_ joined
benabik Yeah 25967be seems like it should install the parrot DLL on Cygwin 14:34
dalek ast: 6e10cfa | moritz++ | S03-operators/eqv.t:
unfudge now passing eqv autothreading tests for rakudo
JimmyZ_ which is failing? parrot or nqp ? 14:35
gfldex parrot 14:36
gfldex is still investigating
14:37 plobsing left 14:39 kaleem left, plobsing joined, agentzh left 14:42 ifaria left
gfldex cygparrot4_1_0.dll is not in MANIFEST* and i think it should 14:43
or should it not?
benabik Hm. My MANIFEST.generated has cygparrot4.1.0.dll instead of 4_1_0... 14:44
gfldex i got that too 14:45
benabik Looks like that may have been a mistake in the release. It used to have 4_0_0.
(Or a bug in the tools. Investigating.) 14:46
Bug in the tools. Our update version script does s/4.0.0/4.1.0/. I'll fix manually and then open a bug for the script. 14:47
gfldex++
14:49 skids joined
gfldex there is more while building nqp: del .\\libdyncall_s.a 14:51
make[2]: del: Command not found
benabik I expected dalek to comment, but I pushed the fix for Parrot in 33a524f. 14:52
JimmyZ_ gfldex: sorry, that's my fault 14:53
gfldex: vi 3rdparty\dyncall\buildsys\gmake\os\windows.gmake 14:55
gfldex: rm the "RM= del"
gfldex: I had emailed to dyncall author to fix these windows issues 14:56
gfldex: here is the patch 15:10
gfldex: gist.github.com/1933005 15:11
gfldex i fixed it by hand, it's building now
JimmyZ_ moritz: this should fix gfldex's problem, hopes it goes into nqp
[Coke] perl6: sub matchcheck(*@) { 1 }; say "alive"; 15:13
p6eval pugs b927740: OUTPUT«*** ␤ Unexpected "*@)"␤ expecting parameter name or ")"␤ at /tmp/7gspqhVBcN line 1, column 16␤»
..niecza v15-2-gd19c478: OUTPUT«Potential difficulties:␤ &matchcheck is declared but not used at /tmp/r3EGKDIJTD line 1:␤------> sub matchcheck⏏(*@) { 1 }; say "alive";␤␤alive␤»
..rakudo 137b16: OUTPUT«alive␤»
[Coke] perl6: sub matchcheck(*@a) { 1 }; say "alive";
p6eval pugs b927740, rakudo 137b16: OUTPUT«alive␤»
..niecza v15-2-gd19c478: OUTPUT«Potential difficulties:␤ @a is declared but not used at /tmp/vX1xivqWAW line 1:␤------> sub matchcheck(*⏏@a) { 1 }; say "alive";␤ &matchcheck is declared but not used at /tmp/vX1xivqWAW line 1:␤------> sub matchcheck⏏(*@a…
[Coke] hey, that's worth about 200 tests right there. 15:15
dalek p/bs: 608ee61 | jnthn++ | src/ (3 files):
Initial implementation of object/STable repossession; not really possible to test until the serializer is updated.
 15:16
p/bs: d3c7cd7 | jnthn++ | src/6model/serialization. (2 files):
Initial preparations for repossessions table in serialization.
p/bs: d83bff9 | jnthn++ | src/6model/serialization_context.c:
Use LSB to signify object vs. STable repossession.
p/bs: 3de6217 | jnthn++ | src/6model/serialization.c:
Implement serialization of repossessions info.
p/bs: cff103f | jnthn++ | src/6model/serialization.c:
Implement deserialization time handling of the repossessions. Seems to vaguely work out.
p/bs: f6c39dc | jnthn++ | src/6model/serialization.c:
Clear REPR_data pointer after freeing the data, just in case.
p/bs: 06304cb | jnthn++ | src/ops/nqp.ops:
Temporarily disable triggering of SC barrier on setting boolification protocol.
kudo/bs: 5d20c5a | jnthn++ | / (2 files):
Add missing Stash class in CORE.setting (we stubbed it in BOOTSTRAP, but never finished it up).
15:16 birdwind1pbird joined 15:18 daxim left
felher moritz++ # for blog-post and hacking :) 15:19
15:20 JimmyZ_ left 15:22 tokuhirom left 15:23 daemon left, cognominal left, wolverian left, drbean left, pjcj_ left, BooK left, dju left, Maddingue left, cotto left, jasonmay left, BooK joined, pjcj joined, Maddingue joined, cognominal joined, cotto joined, wolverian joined 15:24 daemon joined, dju joined 15:25 drbean joined 15:26 sisar left 15:28 birdwind1pbird left 15:37 att left 15:44 Psyche^ joined, Patterner left, Psyche^ is now known as Patterner 15:49 birdwind1pbird joined 15:50 att joined
dalek p/bs: 1a15a9a | jnthn++ | src/ (2 files):
Have deserialization handle republication of Parrot v-table handlers, so we catch mixin cases also. With this, CORE.setting now seems to load...but then nothing happens afterwards. :/
 15:51
15:53 grondilu joined
grondilu perl6: my %h = foo => 'bar' 15:53
p6eval niecza v15-2-gd19c478: OUTPUT«Potential difficulties:␤ %h is declared but not used at /tmp/kbFei3V9sM line 1:␤------> my ⏏%h = foo => 'bar'␤␤»
..pugs b927740, rakudo 137b16: ( no output )
grondilu nom: say my %h = foo => 'bar' 15:54
p6eval nom 137b16: OUTPUT«("foo" => "bar").hash␤»
grondilu perl6: module Foo { our %h = foo => "bar" }; say %Foo::h; 15:55
p6eval rakudo 137b16: OUTPUT«Method 'STORE' not found for invocant of class 'Any'␤ in <anon> at /tmp/a23QZ65PK_:1␤ in block <anon> at /tmp/a23QZ65PK_:1␤␤» 15:56
..niecza v15-2-gd19c478: OUTPUT«{"foo" => "bar"}␤»
..pugs b927740: OUTPUT«␤»
grondilu rakudo is wrong here, right?
benabik nom: my %h = foo => 'bar'; say %h 15:57
p6eval nom 137b16: OUTPUT«("foo" => "bar").hash␤»
benabik nom: our %h = foo => 'bar'; say %h 15:58
p6eval nom 137b16: OUTPUT«Method 'STORE' not found for invocant of class 'Any'␤ in block <anon> at /tmp/lgBZwGby7l:1␤␤»
benabik Looks like a bug in our.
jnthn Already ticketed. 15:59
grondilu ok 16:00
16:02 grondilu left 16:04 birdwind1pbird left
jnthn rakudo/bs now runs perl6 -e "say('hello world')" 16:07
16:10 havenn joined 16:14 pernatiy left 16:15 pernatiy joined
jnthn Many of t/00-parrot and t/01-sanity pass also 16:15
libreofficer Hello folks! 16:27
jnthn, I still try to understand metaops 16:28
what means @values.gimme(1) ? Is this a list with one element? 16:29
jnthn It means "make sure the list we're currently considering has at least 1 reified element" 16:32
libreofficer oh, at least
jnthn Lists in Perl 6 are lazy. .gimme(n) forces it to evaluate (or ensure it already did evaluate) up to n items
well
potentially it may do more
libreofficer I see 16:33
jnthn It returns the number of available elements. 16:34
eiro gimme is like take in the other langages 16:35
jnthn Which other languages? :)
libreofficer so, 'take' in p6 is something like push in stack? 16:36
TimToady it's more like a yield, but you can see it as a push on some anonymous array attached to the gather 16:37
dalek kudo/bs: f5f15e5 | jnthn++ | src/Perl6/Metamodel/BOOTSTRAP.pm:
Pun configuration needs to be per-run.
 16:38
kudo/bs: 5ab8eea | jnthn++ | tools/build/Makefile.in:
Turn the optimizer down to analysis only - some transform busts things. Will figure out with one later.
TimToady libreofficer: gimme doesn't return anything except the number of elements actually available
munch is more like take
dalek ast: 39b3eb7 | (Siddhant Saraf)++ | S32-exceptions/misc.t:
add test for X::Hash::Store::Oddnumber
gfldex somebody build a jnthn compatible bicycle: thereifixedit.files.wordpress.com/2...iking1.jpg
16:38 plobsing left
TimToady (like take in FP languages, to make jnthn happy :) 16:39
16:40 sisar joined
jnthn gfldex: Only if they fill the basket with *good* beer, though :) 16:40
libreofficer I see. I don't write on FP langs too much, but I almost understand
gfldex jnthn: you can fill it with any beer you want
jnthn \o/
TimToady Haskell take is more like .shift(5), if shift took arguments
jnthn packs it full of Yeti
libreofficer two more questions: :infinite, @values.infinite, and that 'triangle' in metaops funcs - what they does? 16:41
jnthn triangle produces all the intermediate results 16:42
nom: say [+] 1..10
p6eval nom 137b16: OUTPUT«55␤»
jnthn nom: say [\+] 1..10
p6eval nom 137b16: OUTPUT«1 3 6 10 15 21 28 36 45 55␤»
sisar moritz: please review my commit to misc.t
TimToady I have no idea what you're referring to by 'infinite'; that word doesn't occur in P6 that I know of
is this pmichaud++'s list attribute that says it knows something is infinite? 16:43
jnthn iirc, .infinite is a way of asking an array if it is potentially infinite in length. Evidently, that's not the spec name for it. ;-)
TimToady: .infinite is just the accessor for that, yet.
*yes
TimToady afaik there is no spec name for it yet because someone hasn't rewritten S07 yet then :) 16:44
jnthn ah :)
jnthn dreams of pmichaud tuits
TimToady libreofficer: yes .infinite is supposed to tell you if it knows the list is infinite
nom: my @foo = 1..*; say @foo.infinite 16:45
p6eval nom 137b16: OUTPUT«True␤»
TimToady niecza: my @foo = 1..*; say @foo.infinite; # I suspect this fails
p6eval niecza v15-2-gd19c478: OUTPUT«(timeout)»
libreofficer ok, what about ':infinite' ? It looks like this: GATHER({...}, :infinite(@values.infinite)) 16:46
TimToady and @foo ~~ :infinite is just a handy way to write the method call in a smartmatch
jnthn It's just passing on the infiniteness information to the thingy that makes a GATHER
s/makes a GATHER/implements gather/ 16:47
GATHER is what a gather desugars to, iirc.
Ouch, my hands are demanding a keyboard break... Back in a bit.
16:49 alester joined
gfldex i should not have mentioned beer :-> 16:49
TimToady my hands usually demand I break my keyboard, but I ignore 'em
libreofficer nom: say [+] 1..10 16:50
p6eval nom 137b16: OUTPUT«55␤»
TimToady New scientific study indicates Balmer Peak comes from anesthetic effects on sore hands!
nom: say [*] 1..10 16:51
p6eval nom 137b16: OUTPUT«3628800␤»
16:51 retup_work joined
libreofficer jnthn, so, as I understand, METAOP_REDUCE only use this anonymous array to gather values, and then return it. And we need to rewrite it like iteration. 16:52
TimToady perl6: say ([\+] 1..*)[10] 16:54
p6eval rakudo 137b16, niecza v15-2-gd19c478: OUTPUT«66␤»
..pugs b927740: OUTPUT«pugs: out of memory (requested 1048576 bytes)␤»
TimToady perl6: say ([\+] 1..1000000000)[10]
p6eval rakudo 137b16, niecza v15-2-gd19c478: OUTPUT«66␤»
..pugs b927740: OUTPUT«pugs: out of memory (requested 1048576 bytes)␤»
TimToady looks like [\op] is already properly lazy 16:55
(excepting pugs)
16:57 mj41 left
libreofficer I want to try optimize metaops (but for begin I need to fully understand it) 16:57
TimToady have fun! 16:58
16:58 birdwindupbird left 16:59 marmalade joined
TimToady moritz: someone should come up with a whois bot that tells you all the old aliases that came from the same address (which might or might not have helped with the fasad->sisar transition) 17:04
would also help spot known troublemakers (though one should give them the benefit of the doubt, since the reason for changing names *might* be in the repentant category) 17:05
17:06 MayDaniel joined
TimToady 'course it wouldn't help if people change name at the same time they change address 17:06
hmm, you could make the names clickable in the irclog with known/guessed info :) 17:08
libreofficer two technical details. perl6 docs locates only on perlcabal.org? There is no some mans/downloadable files. And second - how to run perl6 code faster? My machine run scripts with some annoying delay
TimToady github.com/perl6/specs 17:12
and which implementation are you using?
niecza compiles slower but runs up to 20 times faster than rakudo
(though rakudo is likely to go much faster in the future)
and it really depends on what you're doing; some things they do at about the same speed 17:13
libreofficer I use rakudo from debian repos 17:14
TimToady that seems likely to be fairly ancient
17:14 thou joined
libreofficer I also cloned rakudo and compile it, but not runs it yet 17:14
TimToady most of us just use the latest from github
libreofficer debian wheezy
oh, its old too, really 17:15
TimToady I just do a git pull --rebase every day and recompile; the test suite keeps it fairly stable 17:16
I also track niecza, and go back and forth to see what works well in which
sigh, my pugs broke because it hardwires libffi.so.5, and my system upgraded itself to 6 17:21
jnthn Rakudo's startup time should be improving soon also. 17:22
benabik jnthn: Any initial results on that, or waiting until you unbreak more stuff?
TimToady he just got to "Hello, world!"
jnthn It certainly *feels* faster. I should do something more scientific, but want to unbust some more bits first. 17:23
benabik TimToady: In other words, he's gotten it running code! It's exciting!
jnthn: Very cool. Didn't mean to put pressure, just curious an excited.
jnthn Well, the most exciting thing about it doing "Hello, world!" is it means that it successfully deserialized CORE.setting and at least some of the stuff in it was usable.:)
moritz sisar: test commit looks good.
jnthn It also passes a bunch of sanity tests. 17:24
sorear good * #perl6
TimToady whoops, gotta go sign a bunch of refi papers &
jnthn Many of the fails look like the same underlying issue, which is just a NYI in the serializer.
o/ sorear
sisar moritz: :-) 17:25
dalek p: 4ff8545 | moritz++ | 3rdparty/dyncall/ (2 files):
windows fix by JimmyZ
 17:26
sisar moritz: src/core/ has both Exception.pm and Exceptions.pm, which is kinda confusing 17:29
17:29 mishin_ joined
moritz sisar: sorry about that 17:35
mostly hysterical raisins 17:36
I should join them into one at some point 17:37
17:38 plobsing joined 17:40 fglock joined
fglock o/ 17:41
[Coke] rant: it should be a TAP error to have two identically described tests.
moritz rant: it should be TAP error to post two identical lines in two different channels :-) 17:45
jnthn rant: it should be a TAP error to rant 17:47
oh, wait...
sisar moritz: also, most (or maybe none) files in src/core/ do not have beginning comments to describe a files's purpose, so i can't just open a file & understand what's it for :-|
moritz sisar: it's usually quite easy. Int.pm implements the 'Int' class and operations that work on Int 17:48
sisar: fairly intuitive, isn't it?
sisar Only the names, give ssome hint... well you said it already :) 17:49
17:50 skipper joined
sisar moritz: can you explain line 220 in control.pm. How do i trigger it ? 17:51
17:53 skipper left, pernatiy left
moritz nom: sub f { gather { take $_ for 1..3; return } }; my @a := f; .say for @f 17:54
p6eval nom 137b16: OUTPUT«===SORRY!===␤Variable @f is not declared␤at /tmp/U8mSjxAla_:1␤»
moritz nom: sub f { gather { take $_ for 1..3; return } }; my @a := f; .say for @a
p6eval nom 137b16: OUTPUT«Attempt to return from exhausted Routine␤ in sub EXHAUST at src/gen/CORE.setting:520␤ in block <anon> at src/gen/CORE.setting:347␤ in block <anon> at /tmp/XJAo6k6kfy:1␤␤»
moritz sisar: this sub returns a lazy list. At the time the return() is run, the subroutine has already returned
sisar hmm 17:57
sisar sleeps zzz...
17:57 MayDaniel left 17:59 kmwallio joined, att left 18:00 kmwallio left 18:02 phoenix1789 joined 18:04 Chillance joined
dalek kudo/nom: e959358 | moritz++ | / (3 files):
move contents of src/core/Exceptions.pm into Exception.pm

No functional changes		 18:09
18:10 dakkar left 18:12 skipper joined 18:15 skipper is now known as havennn
dalek p/bs: 20ef680 | jnthn++ | src/ops/nqp.ops:
Comment out debug code - don't need it for now.
 18:18
p/bs: cce66db | jnthn++ | src/6model/serialization.c:
Start chasing chains of outer contexts, serializing along the chain rather than just the immediate one. No deserialization of the chaining just yet.
p/bs: 3f742fc | jnthn++ | src/6model/serialization.c:
Serialization/deserialization of lexpads with natively typed lexicals.
p/bs: 3a2f19c | jnthn++ | src/6model/serialization.c:
Fix a comment.
p/bs: 629c0b0 | jnthn++ | src/6model/serialization.c:
Fixup of outer context chains on deserialization. Unbusts attribute accessor generation, which means that Rakudo's error reporting now works again and various sanity tests now pass again.
18:20 havennn left
jnthn dinner & 18:20
18:34 sisar left 18:35 phoenix1789 left
moritz jnthn: fwiw at some point the bs brach in nqp needs to be updated in a similar way as rakudo/master in connection to the take_recont merge 18:42
jnthn: that would make it easier for people to switch between nom and bs, because then they can use the same parrot for both 18:43
18:46 MayDaniel joined, birdwindupbird joined
sorear samlh: still here? 18:48
18:51 MayDaniel left 18:57 mucker left 18:59 Teratogen left 19:00 havenn left 19:02 att joined 19:04 sahadev left 19:09 pernatiy joined 19:10 Teratogen joined
jnthn moritz: Now that rakudo/bs is close to being somewhere interesting, yeah, I should look into that. 19:15
sorear phenny: tell samlh I am curious what version of Mono.Cairo.dll you have 19:16
phenny sorear: I'll pass that on when samlh is around.
19:18 Khisanth joined 19:26 mishin_ left 19:28 MayDaniel joined 19:36 samlh left 19:38 plobsing left 19:46 kaare_ joined
tadzik 'evening 19:49
19:53 pyrimidine left
Tene grue: /join #phasers 19:56
Oops
jnthn hi tadzik
Tene Typing is hard.
jnthn Your command has been eaten by a grue. 19:57
fglock `/leave
19:57 fglock left
tadzik yeah, typins' hard 19:57
* typing's
19:58 MayDaniel left
tadzik ./perl6 -e 'say 1' 0.25s user 0.09s system 97% cpu 0.346 total 20:03
not bad!
jnthn tadzik: That's in the bs branch? 20:04
tadzik aye
jnthn tadzik: What is it for you normally?
tadzik I guess around 1 second
20:05 bluescreen10 left
jnthn OK, so it's about 25% of what it was. 20:05
tadzik more-less-ish
I may rebuild the non-bs nom, I'll probably have to do some hacking on it before GPW anyway 20:06
jnthn Yeah, bs still has a couple of serious issues.
(Not unexpected ones, just things I didn't get to doing yet.)
tadzik jnthn++ 20:08
awesome to see such progress
20:11 drbean left
masak evenin', #perl6 20:13
sorear masak!
jnthn bon soir, masak 20:14
masak sorear++ # release
jnthn++ # I haven't backlogged yet, but I see "25% of what it was" above, and that sounds... promising :)
sorear colomon++ # most of the cool features in v15
20:14 tokuhirom joined
sorear sounds like bs will greatly increase my startup time parity challenges :| :D 20:15
20:15 drbean joined
TimToady niecza: say [1, 0.70710678118654746] eqv [0.85355339059327373, 0.8408964152537145] 20:15
p6eval niecza v15-2-gd19c478: OUTPUT«True␤»
TimToady ^^^
er, sorear ^^^ 20:16
sorear What are the last two numbers?
(is it sad that I can instantly recognize sqrt(1/2) ?)
TimToady the 2nd iteration from rosettacode.org/wiki/Arithmetic-geo...ean#Perl_6
which doesn't work if I change the all >>==<< to an eqv 20:17
masak sorear: no, it's not sad. was gonna say that it's completely normal, but I guess it isn't... :)
TimToady niecza: say [1, 0.70710678118654746] eqv [0.85355339059327373, 0.8408964152537145, 42]
p6eval niecza v15-2-gd19c478: OUTPUT«False␤»
masak sorear: anyway, I recognized it too. I consider it one of the constants that one "should" recognize.
TimToady I suspect niecza is comparing lengths rather than values of the arrays 20:18
masak submits nieczaissue
TimToady niecza: say [1, 0.70710678118654746] ~~ [0.85355339059327373, 0.8408964152537145] 20:19
p6eval niecza v15-2-gd19c478: OUTPUT«False␤»
TimToady well, that works
but eqv is rather more in the spirit of things 20:20
20:20 benabik left
TimToady incidentally, rakudo fails on the line before that 20:20
20:21 benabik joined
TimToady if the masakbot is looking for fuel 20:21
TimToady wonders if Boeing named their first jetliner after that number... 20:22
jnthn my @new := [ 0.5 * [+](@old), sqrt [*](@old) ]; # that line?
nom: my @new = 1,2,3; say [+](@new) 20:23
p6eval nom e95935: OUTPUT«6␤»
TimToady yeah, I get No applicable candidates found to dispatch to for 'Numeric'. when I run the program
maybe it's elsewhere 20:24
jnthn Hm. I know we fixed the reduction parsing thingy not so long ago but long ago enough that I suspect you've a new enough nom. 20:25
TimToady fresh built today
jnthn OK
Then my first guess at why it didn't work ain't it :)
TimToady okay, it's the 'last' that's doing it, apparently throwing away the partial sequence 20:26
so the [*-1] fails
if I comment the last and change *-1 to [5] it works 20:27
oh, wait, no
wrong impl
but different error
now it gets: Nominal type check failed for parameter '@old'; expected Positional but got Num instead 20:28
I wonder if it's flattening the @new return value into the sequence
which is arguably correct... 20:29
yes
if I change the return value to [@new] it works
and still works in niecza 20:30
however, if I uncomment the last it still has the No applicable error
works if the subscript is [3] rather than [*-1], so the problem doesn't seem to be a null return 20:31
so last is probably working right
more like *-1 is not getting a valid *
changing that to [10] returns Nil, so the last is correctly terminating 20:32
jnthn nom: say (1, 3 ... 9)[*-1] 20:33
20:33 broquaint left
p6eval nom e95935: OUTPUT«9␤» 20:33
jnthn nom: say (1, 3 ... 9)[*-2]
p6eval nom e95935: OUTPUT«7␤»
TimToady ah, it's returning Inf for the length
20:33 birdwindupbird left
TimToady because it 'knows' it's ... * 20:34
and doesn't know there's a last
yes, works if I change it to ... 100
that's...an interesting failure mode 20:35
actually, works with ... 0 too, since that never matches +[$a,$g] 20:37
20:38 plobsing joined
TimToady though ... False might be a faster test 20:38
20:39 att left, MayDaniel joined
TimToady wonders if [*-1] should be specced to use the reified length if Inf 20:40
since no one in their right mind will want to find the last element of an infinite list, except maybe a mathematician 20:41
nom: say [].^methods 20:43
p6eval nom e95935: OUTPUT«new bind_pos delete flattens REIFY STORE_AT_POS STORE PARAMETERIZE_TYPE at_pos perl new Bool Int end fmt flat list flattens tree Capture Parcel at_pos eager elems exists gimme infinite iterator munch pick pop push roll reverse rotate shift unshift splice sort classify…
TimToady offhand I don't see a method to get the reified length 20:44
nom: say List.^methods
p6eval nom e95935: OUTPUT«new Bool Int end fmt flat list flattens tree Capture Parcel at_pos eager elems exists gimme infinite iterator munch pick pop push roll reverse rotate shift unshift splice sort classify categorize uniq REIFY STORE_AT_POS FLATTENABLE_LIST FLATTENABLE_HASH keys values pa…
jnthn That'd be equivalent to reifying it I guess
tadzik jnthn: ./perl6 -e 'say 1' 1.05s user 0.16s system 94% cpu 1.288 total 20:45
TimToady I don't want to reify it, just find how much of it real already
esp if it's infinite
tadzik that's nom, so yeah, bs is about 1/4 of the startup time
(at the current state)
20:45 broquaint joined
TimToady nom: say [1,2,3..*].gimme(*) 20:46
jnthn nom: my @x = 1, 3 ... *; say @x[3]; say @x.gimme(*)
p6eval nom e95935: OUTPUT«2␤»
nom e95935: OUTPUT«7␤4␤»
jnthn Ah, we both had the same idea :) 20:47
TimToady so I'd argue that the number fed to a subscript for [*-1] should be that number, not Inf
.elems can still return Inf, i guess 20:48
20:48 havenn joined
masak maybe I'm just being needlessly conservative, but I think it sounds like a confusing feature. 20:49
not all lazy lists are infinite. 20:50
what if I *want* it to find the end and give me the last item?
TimToady true; I guess what I really want is the ... to trap the 'last' and unmark the .infinite
20:51 y3llow_ joined, pothos_ joined
TimToady niecza is producing the correct result because it does not yet support .infinite 20:51
well, and it's not flattening the @new into the ... list 20:52
20:52 y3llow left 20:53 pothos left, y3llow_ left 20:54 y3llow joined, pothos_ left 20:55 pothos joined 20:56 y3llow left, y3llow joined
masak +1 on trapping 'last' and unmarking .infinite -- that's only fair/sane/correct in this case. 20:56
TimToady okay, rosettacode.org/wiki/Arithmetic-geo...ean#Perl_6 now works with both niecza and rakudo (working around current issues) 20:57
20:58 y3llow left, pothos left, pothos joined
TimToady will switch to eqv and * when those issues are resolved 20:58
20:58 y3llow joined
TimToady for now uses ~~ and 0 20:58
oh, and scalar variables to avoid flattening 20:59
I think it's a nieczabug that @new doesn't flatten into the sequence's list context 21:00
21:00 y3llow left
TimToady we have use cases that rely on 1, 1, -> $a, $b { ...; $new-a, $new-b } ... * 21:01
21:01 y3llow joined, pothos left, pothos joined 21:03 y3llow left, y3llow joined
TimToady in fact, one could write the agm program in that form, except that niecza would bind the returned capture all to the next $a, I suspect 21:03
21:04 pothos left
dalek kudo/bs: 038e84c | jnthn++ | src/Perl6/World.pm:
Remvoe todo that's no longer todo.
 21:04
kudo/bs: 662c70a | jnthn++ | src/ (3 files):
Update lexical fixup handling for bs; earlier I just commented it out out. This gets us passing all 00-parrot and all of 01-sanity apart from the final test which requires Test.pm, which we can't quite build yet.
21:04 pothos joined
TimToady niecza: my @x := 1,10, -> $a, $b { $a+1, $b+1 } ... *; say @x[5] 21:05
p6eval niecza v15-2-gd19c478: OUTPUT«3 3␤»
TimToady nom: my @x := 1,10, -> $a, $b { $a+1, $b+1 } ... *; say @x[5]
p6eval nom e95935: OUTPUT«5 14␤»
21:05 y3llow left 21:07 pothos left
TimToady sorear: ^^ nieczabug 21:09
masak nom: my @x := 1,10, -> $a, $b { $a+1, $b+1 } ... *; say @x[0..10].perl 21:10
p6eval nom e95935: OUTPUT«(1, 10, (2, 11), (3, 12), (4, 13), (5, 14), (6, 15), (7, 16), (8, 17), (9, 18), (10, 19))␤»
masak niecza: my @x := 1,10, -> $a, $b { $a+1, $b+1 } ... *; say @x[0..10].perl 21:11
p6eval niecza v15-2-gd19c478: OUTPUT«(1, 10, $(2, 11), $(11, 3), $(3, 3), $(3, 3), $(3, 3), $(3, 3), $(3, 3), $(3, 3), $(3, 3))␤»
masak submits nieczabug
21:17 glass left, kaare_ left, orafu left, sorear left, alvis left, araujo left, tadzik left, doy left, jerome_ left, shachaf left, bacek_at_work left, _ilbot left, japhb left, charsbar___ left, PZt left, gfldex left, TimToady left, mathw left, itz_ left, ruoso left, NamelessTee left, Alias left, havenn left, MayDaniel left, retup_work left, cotto left, BooK left, daemon left, benabik left, frettled left, PacoAir left, ab5tract left, lestrrat left, preflex left, mtk left, mikec left, bonsaikitten left, colomon left, jevin left, pmichaud left, jnthn left, Bucciarati left, conntrack left, ocharles left, __sri left, `patch` left, LoRe left, jtpalmer left, ashleydev left, scottp left, pnu left, lumi___ left, katernya left, Tene left, krakan left, risou left, plobsing left, thou left, apejens left, Vlavv_ left, REPLeffect left, kthakore_ left, awwaiid left, [Coke] left, ranguard_ left, athomason left
dalek ecs: 6f34dd5 | larry++ | S03-operators.pod:
last from ...* must change .elems to finite
 21:17
21:18 samlh joined, att joined, pothos joined, y3llow_ joined, havenn joined, MayDaniel joined, plobsing joined, benabik joined, kaare_ joined, thou joined, retup_work joined, daemon joined, cotto joined, BooK joined, frettled joined, PacoAir joined, NamelessTee joined, ab5tract joined, lestrrat joined, Alias joined, orafu joined, preflex joined, mtk joined, mikec joined, bonsaikitten joined, PZt joined, gfldex joined, sorear joined, TimToady joined, colomon joined, apejens joined, alvis joined, araujo joined, mathw joined, Vlavv_ joined, REPLeffect joined, tadzik joined, glass joined, katernya joined, kthakore_ joined, itz_ joined, doy joined, jevin joined, awwaiid joined, pmichaud joined, [Coke] joined, jnthn joined, jerome_ joined, Bucciarati joined, conntrack joined, ruoso joined, shachaf joined, ocharles joined, __sri joined, `patch` joined, bacek_at_work joined, ranguard_ joined, LoRe joined, jtpalmer joined, charsbar___ joined, japhb joined, _ilbot joined, ashleydev joined, scottp joined, pnu joined, lumi___ joined, athomason joined, risou joined, krakan joined, Tene joined, y3llow_ left, y3llow joined
[Coke] wonders if we should be putting rosetta code offerings in roast. 21:20
(preferably in some way that doesn't require manual syncing)
jnthn++ # such bs.
21:20 pothos left 21:21 pothos joined, y3llow left, havenn left, y3llow joined
dalek p/bs: a1249b9 | jnthn++ | src/6model/serialization.c:
For now, just treat Coroutine like Sub. Gets us able to compile Test.pm and - after copying it elsewhere because of some PROCESS bug that screws up module location - we can run spectests again. Many pass, quite a bit of fail, but hopefully a few common root causes for most of them.
 21:21
masak nom: sub infix:<*+> ($a,$b) { $a * $b + $b }; say 2 *+ 5 21:22
p6eval nom e95935: OUTPUT«10␤»
masak submits rakudobug
moritz nom: sub infix:<*+> ($a,$b) { $a * $b + $b }; say infix:<*+>(2, 5) 21:23
jnthn wtf
p6eval nom e95935: OUTPUT«15␤»
21:23 y3llow left, pothos left, y3llow joined
felher hm. Does nom think its 2 *(+5)? 21:24
21:24 pothos joined
masak aye. 21:24
niecza gets it right.
jnthn yeah but...even the cheating LTM that Rakudo's parser depends on is meant to get *that* case right.
masak that's why I'm submitting it as a bug :)
jnthn Well, I dobut I'm gonna hunt down a bug in a load of code I plan to eliminate as soon as I get the tuits. :) 21:25
Hopefully the new LTM engine gets it righter.
21:26 y3llow left, y3llow joined, pothos left
masak well, RT is a catalogue of things that are currently wrong, and makes no allowances for hope. ;) 21:26
21:27 pothos joined
jnthn I wasn't disagreeing with the ticket. :) 21:28
Just saying I won't jump on it right away. :)
21:28 y3llow left, y3llow joined 21:29 pothos left
masak nodnod. 21:30
21:30 pothos joined
dalek kudo/bs: 7c6ebd7 | jnthn++ | src/Perl6/Metamodel/BOOTSTRAP.pm:
Fix PROCESS handling. Now all 00-parrot and 01-sanity pass and we can run the spectests with the build produced right off.
 21:30
jnthn Yesterday: 21:34
22:05 <jnthn> Wonder if I can get Rakudo passing most of them spectests again with the bs. :)
22:06 <masak> sounds like a challenge :)
Depends how you pick "most", but: challenge achieved. :)
timotimo "the bs"? :)
cognominal bs is no bs.
jnthn timotimo: "Bounded Serialization" 21:35
timotimo: A Rakudo branch I've been workong on.
21:35 havenn joined
jnthn timotimo: It cuts hundreds of megs off compiling CORE.setting, quarters startup time and will enable various missing features, extra optimizations, etc. 21:36
prammer 'for' lop => 'for' loop # perlgeek.de/blog-en/perl-6/2011-02-...tions.html 21:37
jnthn OK, gonna take a break, and when I come back I'll cut this month's R* release. 21:39
masak: When you did the Rakudo release did you cut an NQP release tarball also?
github.com/downloads/perl6/nqp/nqp....02.tar.gz
gives 404 error
masak huh :/ 21:40
jnthn: I did not. it's not in nqp's docs/release_guide.pod
jnthn :/
Release guide fail. 21:41
masak feel free to do it now. I would, but I'm a bit tied up.
jnthn Sure, I do it after le break. :)
21:46 havenn left
masak rant: it should be a TAP error to read the backlog and then make fun of emergent memes in it. 21:46
21:48 mj41 joined
masak moritz++ # perlgeek.de/blog-en/perl-6/2011-02-...tions.html 21:50
[Coke] rant: all y'all. :P
22:00 jaymibbs joined, preflex_ joined 22:01 preflex left, preflex_ is now known as preflex 22:03 skids left 22:04 jaymibbs left
araujo :) 22:13
22:14 MayDaniel left
jnthn is back 22:17
22:24 colomon left 22:25 havenn joined 22:27 Araq joined
masak Araq! \o/ 22:27
Araq how would templating systems or preprocessors work without source code filters? 22:28
(hi masak :-) )
masak your question invites all manner of cheeky/witty responses, but I'll go this route instead:
what do you mean?
22:29 kaare_ left
Araq well I heard perl6 has no source code filters anymore 22:29
masak Perl 6 rejects the notion of Perl 5 source filters.
since the latter qualify as multi-pass parsing.
and Perl 6 excplicitly parses source in a single pass. 22:30
22:30 snearch joined
masak thus, the kinds of templating systems or preprocessors you're thinking of cannot be Perl 6, since Perl 6 is single-pass. :) 22:31
22:31 fsergot left
Araq ok so something like www.csse.monash.edu.au/~damian/pape...igata.html is simply not supported anymore? 22:32
masak not through Perl 5-style source filters, no. 22:33
basically, we hope that our gift of general, extensible, hookable parsing will more than make up for that. 22:34
timotimo jnthn: sounds great!
the bs branch i mean 22:35
Araq ok, now the cheeky/witty responses please
timotimo quarter of startup time is good. last time i checked, startup time was horrible ;)
benabik bs always _sounds_ great.
Araq I can't image what's funny about my question, but english is not my mother tongue ;-) 22:36
masak Araq: how would they work? "very well." 22:37
Araq oh alright :-) 22:38
jnthn wtf, the Rakudo * makefile freezes while trying to check out Config-INI
masak Araq: I wasn't able to tell, fwiw, that English is not your mother tongue. now that I know, I still think it's more than adequate.
22:39 noam left 22:47 tokuhirom left
dalek ar: 7a3322c | jnthn++ | skel/Configure.pl:
Update versions in Configure.pl (grr, gotta automate this stuff more...)
 22:50
ar: 60ba95e | jnthn++ | skel/tools/build/Makefile.in:
Dammit, forgot to update this too. :-/
 23:04
Araq README: "This is the 2011.07 release of Rakudo Star." 23:05
nope ;-)
masak Araq++
geekosaur someone needs to write a script to catch that, isn't this like the second or third time someone has caught that before release? 23:06
(and at least once nobody caught it)
jnthn Yes, this is waht I bloody mean. there's fo fscking automatio.
*no
23:06 XooX joined
jnthn Which is why I'm still cutting this stupid release rather thatn doing something more fun. 23:06
masak XooX: welcome.
masak hugs jnthn
23:07 XooX left
jnthn Not to mention I have to do it under Linux 'cus the script to cut one doesn't work on Windows. :( 23:07
masak yeah. much as I dislike using Windows, having a Linux-only script sounds too monocultural. 23:10
geekosaur (obviously it should be written in perl 6)
jnthn Indeed. 23:11
masak yeah, because a build script is where you'd want to battle with bootstrapping problems.</sarcasm> 23:12
geekosaur :) 23:14
23:14 havenn left
jnthn fwiw, 2012.02 uses 2.2GB of RAM to build CORE.setting. In the bs branch it peaks around 1.3GB. 23:15
23:15 alvis left, PacoAir left
jnthn nom: say 1.3/2.2 23:15
p6eval nom e95935: OUTPUT«0.590909090909091␤»
Araq good night guys
jnthn 'night, Araq 23:16
23:16 Araq left, mj41 left
jnthn (That's on 64-bit...) 23:17
Which, yes, is still way to high. But it's a move in the right direction. 23:18
At least, I has a release tarball 23:19
23:20 havenn joined
masak 'night, #perl6 23:22
felher o/ 23:23
23:25 alvis joined, att left
jnthn rakudo.org/2012/02/28/rakudo-star-2...-released/ # it's out 23:26
23:35 havenn left 23:37 whiteknight joined 23:42 samlh left
felher jnthn++ :) 23:54
23:57 Chillance left 23:58 lestrrat left 23:59 lestrrat joined, havenn joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!