»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend! | tinyurl.com/p6contest
Set by moritz_ on 28 December 2010.
sorear time profiling isn't currently available; when optimizing, I rely on NIECZA_TRACE_CALLS and the Mono profiler 00:00
pmurias so it will take too long?
dukeleto sorear: perhaps you can use redo
pmurias what's redo?
dukeleto github.com/apenwarr/redo
sorear the Mono profiler can produce exclusive time information for Perl 6 calls, and then I correlate that with the call graph 00:01
dukeleto djb's invention to replace make, implemented by someone else who is very talented
sorear pmurias: more like "forever" for me; the build already swaps
snarkyboojum dukeleto: redo doesn't work on windows though and requires python on all other platforms atm right?
pmurias sorear: swaps?
sorear: how much RAM do you have? 00:02
00:02 sgarthp left
dukeleto snarkyboojum: i believe there is a simple redo that comes with the python one, that only requires /bin/sh 00:02
snarkyboojum dukeleto: I thought that was more a proof of concept, but still doesn't work on windows I guess
sorear pmurias: 384MiB 00:03
00:03 DarthGandalf left
pmurias sorear: why so little? 00:04
lue afk for a bit. 00:06
dalek odel: bbc7776 | jonathan++ | java/compiler/try.bat:
[java] A barebones try.bat so I can get things running on Windows.
 00:10
pmurias sorear: so how long does recompiling niecza after changing one file take for you? 00:11
00:16 DarthGandalf joined 00:18 DarthGandalf left, AzaToth joined
snarkyboojum lue: "I still think S22's authority/author disambiguation is very illegal." - what do you mean by that? 00:21
TimToady s/author/contributors/ 00:22
snarkyboojum still wonders what's happening with cpan6.org :) 00:29
pmurias the site is surely outdated 00:31
as it referes to April 2010 as a future date
snarkyboojum pmurias: aye 00:32
00:35 redicaps joined 00:39 supernovus joined
lue hello again o/ 00:39
snarkyboojum: S11 states that the :auth option for modules (e.g. module A::B :auth<ME!>) is left ambiguous, meaning either 'authority' or 'author', depending on what you want it to mean. 00:40
supernovus good $timeofday 00:41
sorear pmurias: I haven't done that since I set up the makefile.
I'll check out redo later
lue S22 allowing those to be separate fields (instead of synonyms or replaced with one auth option) violates S11
snarkyboojum lue: ah, "intentionally ambiguous". I see. 00:42
dalek odel: bae7169 | jonathan++ | java/compiler/NQPSetting.pm:
[java] Need the things in the simplified setting to have lexical scope.
odel: 422f867 | jonathan++ | java/runtime/Rakudo/Init.java:
[java] Update outdated name.
supernovus I was finding a strange issue with the JSON::Tiny library. I'm seeing if a complete reinstall of my Rakudo and libraries fixes it. 00:45
supernovus dreams of a day when rakudo is as fast as perl5... 00:51
00:53 molaf_ joined
pmurias sorear: redo is not portable to windows based on the makefile 00:54
00:55 AzaToth left
dalek odel: d944750 | jonathan++ | java/compiler/JST2Java.pm:
[java] Fix code-gen of block setup, so now invocation bottoms out.
 00:56
00:56 pmurias left 00:57 molaf left
dalek p-rx/nom: 7852255 | jonathan++ | src/metamodel/multi_dispatch.c:
Finish up getting type-based multiple dispatch to work.
 01:03
01:05 dku joined
TimToady the point is, auth is part of the key, and contributors is just part of the value 01:07
supernovus Well, a nice fresh build of everything and I can still replicate the bug in JSON::Tiny. When you call from-json() on a JSON string that has an escaped " character in it, the \" part turns into Any(). So if you had <a href="blah"> it would become <a href=Any()blahAny()> which of course fails horridly when you try to parse it.
01:12 ashleydev joined 01:14 candotoo joined 01:18 MayDaniel left 01:21 dku left
supernovus Well, it seems I fixed the bug. It was just missing an action for handing an escaped " character. 01:22
Well, I submitted an issue on the json github page with the solution I used. Hope it gets fixed in the upstream :-) 01:27
01:27 supernovus left 01:28 cibs left
dalek p-rx/nom: b6b076d | jonathan++ | src/metamodel/multi_dispatch.c:
Get multi-dispatch on definedness to work.
 01:36
p-rx/nom: f4fe15c | jonathan++ | t/nqp/55-multi-method.t:
Add multi-method tests from nqpclr, which now pass in nqp-rx/nom.
sorear TimToady: what version of mono do you have? 01:40
EVERYONE: What versions of Mono should I test niecza on? 01:44
01:46 Chillance left
dalek p-rx/nom: 78031bf | jonathan++ | src/stage0/ (3 files):
Update bootstrap with multiple dispatch and signature parsing improvements.
 01:53
p-rx/nom: b0fa8db | jonathan++ | src/metamodel/how/NQPMu.pm:
Make ACCEPTS a multi and correct the signature of it.
jnthn night o/ 01:55
sorear night!
01:57 coldhead left 02:00 noganex_ joined 02:03 noganex left 02:05 amkrankruleuen left 02:07 amkrankruleuen joined
sorear S03-operators/flip-flop.t is extremely suspect 02:10
it appears to assert that ~[ Nil, Nil, Nil, Nil, Nil ] eq " "
02:12 cafesofie joined
colomon Does the flip-flop operator even work in Rakudo? I can't recall ever running into it in the source.... 02:13
rakudo: say 4 ff 5 02:14
p6eval rakudo 101491: OUTPUT«===SORRY!===␤Confused at line 22, near "say 4 ff 5"␤»
colomon std: say 4 ff 5 02:15
p6eval std 625303c: OUTPUT«ok 00:01 120m␤»
colomon pugs: say 4 ff 5
p6eval pugs: OUTPUT«*** No such subroutine: "&infix:ff"␤ at /tmp/lfZFiZegEU line 1, column 5 - line 2, column 1␤»
colomon niecza: say 4 ff 5 02:16
p6eval niecza v1-149-g56c80b0: OUTPUT«Can't open perl script "niecza_eval": No such file or directory␤»
02:16 dduncan joined, dduncan left 02:17 Su-Shee left
snarkyboojum alpha: say "I still work" 02:21
p6eval alpha : OUTPUT«I still work␤»
colomon alpha: say 4 ff 5 02:22
p6eval alpha : OUTPUT«Confused at line 10, near "ff 5"␤in Main (file <unknown>, line <unknown>)␤»
02:26 coldhead joined
diakopte1 phenny: tell sorear bash on latest cygwin is GNU bash, version 3.2.51(24)-release (i686-pc-cygwin) 02:26
phenny diakopte1: I'll pass that on when sorear is around.
02:26 diakopte1 is now known as diakopter, [Coke] joined 02:35 dukeleto left 02:36 dukeleto joined
PerlJam I just had a minor "moment". I've got a window up with some Perl code in it that I just switched to. For a few seconds I couldn't tell if I was looking at Perl 5 or Perl 6. 02:41
Somehow I think this is a good thing :)
02:43 risou joined
sorear diakopter: what versions of mono should I test on? 02:45
phenny sorear: 02:27Z <diakopte1> tell sorear bash on latest cygwin is GNU bash, version 3.2.51(24)-release (i686-pc-cygwin)
02:46 redicaps left
TimToady sorear: I just upgraded to 2.8.2, but it still fails the say 1 02:49
02:53 ashleydev left
sorear TimToady: is that failure from a clean build after the upgrade? 03:03
TimToady checking 03:06
well, after git clean it doesn't offer to do anything 03:07
git pull forces some stuff to rebuild though 03:08
still can't resolve keys in class Array 03:09
03:10 candotoo_ joined 03:13 candotoo left 03:15 dku joined
sorear TimToady: after 'mono boot/run/Niecza.exe test.pl', do any not oks come up? 03:21
TimToady undeclared routines: chr and ord 03:23
03:32 envi joined 03:42 colomon left
TimToady attempting fresh clone 03:43
03:45 colomon joined 03:48 ashleydev joined
TimToady same error, no chr or ord 03:49
sorear tries to install 2.8 03:54
.2
03:57 helv12 joined 04:19 candotoo_ left 04:25 Su-Shee joined 04:31 KatrinaTheLamia left
diakopter TimToady: weird. 04:35
sorear not weird at all 04:40
I just screwed up the directions
the bootstrap is a little bit older, and doesn't have ord or chr to test
I could ask TimToady to check out an older version of test.pl, or manually delete the last stanza, but I've exceeded my willing-to-ask-favors limit already 04:41
04:48 dku left 04:54 KatrinaTheLamia joined 05:04 ashleydev left 06:14 ashleydev joined 06:15 jdv79 left 06:16 DarthGandalf joined 06:28 ashleydev left, typepeter joined
typepeter I have some questions about grammer. Is the grammer one kind of regex matching for lang parsing and behavior definitions? 06:33
I think it's a very great invention that can implement all the lang in perl6,because it abstract all things in language. 06:37
I do love it :D Though my friends which don't know the existence of perl6(even per5!!),i'll intro it to them 06:42
sorear Could you rephrase that? 06:48
It is not clear what you mean by "one kind of regex matching" 06:51
typepeter oh,i see. i meant,is it one form to do regex matching when lang parsing and behavior definition? I just guess about it 06:55
sorear What do you mean by a form? 06:57
Standard practice is more like grammar Numbers { rule TOP { <number>+ }; token number { \d+ } }; class NumActions { method TOP($/) { ... }; method number($/) {... } } 06:58
So I guess no
huh, my shiny new mono self-identifies as 2.11. I guess I missed the release 06:59
typepeter Mm,i guess it's not equal to regex,but using regex matching actually:P 07:06
07:07 Cyrus_ joined 07:08 Cyrus_ left 07:10 clkao_ is now known as clkao
typepeter if i would like to implement a subset of perl6,regex have to be implemented first,and grammer is required then.I guess they are the essential element. 07:13
07:26 Cyrus left 07:31 risou left, Cyrus joined 07:36 Bzek joined 07:38 Bzek_ left
diakopter typepeter: yes, generally 07:53
typepeter Thank you :D thanks sorear,too 07:55
diakopter gee, I thub 08:05
sorear so I've installed mono 2.8.2 08:09
good : the bootstrap niecza passes all tests
bad: it hangs in the unload phase
diakopter (not for me) 08:24
wfm with 2GB ram
sorear: try it on the p6eval chroot, using the mono (fresh 2.8.2) installed to /home/p6eval/sprixel/clr 08:25
(don't use the system mono) 08:26
08:49 Patterner left 08:57 Psyche^ joined, Psyche^ is now known as Patterner 09:19 typepeter left
dalek odel: 017f52a | (Martin Berends)++ | java/runtime/ (3 files):
[java/runtime] improve Makefile and diagnostics in catch blocks
 09:20
sorear diakopter: why not? does it work? :) 09:40
diakopter: Which machine does p6eval run on again? 09:42
09:44 kaare_ joined 09:47 Bzek left 09:50 Bzek joined 09:51 kjeldahl joined 09:58 cogno joined, coldhead left 10:01 tadzik joined
tadzik bah, feather downtime 10:01
10:08 justatheory left 10:14 cogno left 10:15 am0c joined
sorear hello tadzik 10:20
10:25 cogno joined 10:33 rindolf joined
rindolf Hi all. 10:33
tadzik hi 10:39
rindolf Hi tadzik 10:40
tadzik: what's up?
tadzik rindolf: not bad. Hard week coming, but I try to keep up 10:41
"EVERYONE: What versions of Mono should I test niecza on?"
sorear: 2.6.7 if you may
rindolf tadzik: ah.
tadzik: why would the week be hard>
?
tadzik rindolf: 4 exams, 2 labolatories, one programming project to finish 11:01
dalek odel: ef7ab8c | jonathan++ | java/compiler/NQPSetting.pm:
[java] Add missing NQPCode to the fake/temp setting.
 11:04
odel: c8ae83c | jonathan++ | java/runtime/Rakudo/Metamodel/SharedTable.java:
[java] Correct capture construction for find_method call. And with this...hello world works on the JVM. \o/ mberends++
11:04 cogno left
rindolf tadzik: oh. :-S 11:04
11:07 kaare_ left 11:12 cogno joined 11:16 cafesofie left 11:32 cogno left 11:44 risou joined
moritz_ tadzik: good luck 11:54
tadzik thanks moritz_
one of these is physics :)
rindolf tadzik: yes, good luck. 11:55
tadzik moritz_: how is ronja? 11:56
* Ronja
moritz_ tadzik: she's very well - no signs of infection left 12:04
in fact once she starts drinking milk on her own, she can leave the hospital
rindolf moritz_: who is Ronja? 12:06
moritz_ rindolf: my little daughter, about 2 days old
rindolf moritz_: ah.
moritz_: mazal tov! 12:07
moritz_ thanks
rindolf Receiving objects: 76% (267766/352071), 64.64 MiB | 28 KiB/s
12:09 mberends left 12:11 MayDaniel joined 12:22 pmurias joined 12:32 risou left 12:37 dsp_ left 12:39 simcop[hireme] left, cogno joined 12:47 risou joined 12:57 cogno left 13:00 simcop2387 joined
dalek p-rx: e657b28 | bacek++ | / (2 files):
Add default value for into RPA.join to follow Perl6 setting closely
 13:05
p-rx: 2613f07 | bacek++ | t/setting/01-resizablepmcarray.t:
Add test for RPA.map
 13:10
13:30 risou left, risou joined 13:32 shi joined 13:38 cogno joined 13:39 Chillance joined 13:47 Fredrek_j joined
Fredrek_j Helo, I was reading about perl5 TK and how the death of "the father of perl/Tk" Nick Ing-Simmons caused the development of perl/Tk to almost stop completely.. is that true that the death of one person can cause a very important modules to stop evolving? 13:52
13:55 VXZ left 13:57 VXZ joined 13:58 Fredrek_j left
colomon It's certainly true that losing a key developer can be a huge setback. 14:00
The perfect Perl 6 example is pugs. It was by far the most successful early Perl 6 compiler, but when its key developer moved on elsewhere, no one else was able to pick up the reins, and it has languished ever since. 14:02
14:06 rindolf left
colomon At the same time, modules which are actually very important will normally have someone take over, or another project replace them, just as Rakudo and Niecza have replaced pugs as the Perl 6 compilers of choice. 14:06
14:08 rindolf joined 14:20 cogno left 14:22 Patterner left 14:26 carlos joined
tadzik oh, someone just updated Rakudo in Gentoo 14:27
jnthn Nice 14:29
14:34 Psyche^ joined, Psyche^ is now known as Patterner 14:41 kaare_ joined 14:51 satyavvd joined, PerlJam left 14:52 cogno joined
colomon \o/ 14:52
14:53 PerlJam joined, PerlJam left, satyavvd_ joined 14:54 PerlJam joined 14:55 mberends joined, satyavvd left, satyavvd_ is now known as satyavvd
dalek p-rx/nom: c759804 | jonathan++ | src/NQP/Actions.pm:
Get basic case of multi-subs in place. Just handles lexical case and multis in nested scopes are probably busted still.
 15:03
p-rx/nom: 46a10cd | jonathan++ | t/nqp/51-multi.t:
Replace old multi-dispatch tests. Multi-method dispatch is now in a separate test file, added yesterday. 51-multi.t will now just test multi subs. Needs more tests, but does cover arity based dispatch and that the proto functions as it should - being called first then delegating to the multi-dispatcher.
15:05 cogno left
PerlJam colomon: re: losing key developers. Right now I think Perl 6 could survive the loss of any single developer except perhaps TimToady. (I don't see who would be his logical "second") 15:06
I guess Perl 6 would survive, its evolution would just be stunted. 15:07
arnsholt A duumvirate of jnthn and pmichaud perhaps. With masak as court jester =) 15:08
15:11 carlos left 15:16 dorlamm joined 15:21 payload joined
pmurias PerlJam: Perl 6 as a whole would survive, but individual implementations are more vulnerable 15:21
PerlJam: niecza withough sorear would have a very bleak future 15:22
PerlJam indeed.
payload hi. where can i look in the spec for the place of overwriting the behaviour of the + in +$a for a type foo? 15:23
ha, never mind... found it 15:24
jnthn S13
payload thanks
jnthn Overloading in Perl 6 is just multis. :)
Though actually overloading prefix:<+> is not really what you want to do in this case, I suspect. 15:25
If you want to control numification, just implement a method Numeric() { ... } in your type.
+ just calls .Numeric.
payload Numeric? i would have done multi method Int() 15:26
or multi method Num()
or both
or so
didnt know that there is a Numeric ^^ 15:27
jnthn Numeric is the general "give me something that's number-ish" coercion.
.Num and .Int are coercions to a floating point and integer specifically.
payload in Rakudo Star +(1..100000) does the job stupid and slow i think 15:31
i didnt test it with current HEAD
but i want to patch that ^^
jnthn rakudo: +(1..100000) 15:32
payload slow ^^
jnthn ...
p6eval rakudo 101491: ( no output )
jnthn FAIL.
payload yes
15:32 Alias joined
PerlJam That seems like LHF 15:32
payload succ and count
LHF?
jnthn rakudo: (1..100000).Numeric
PerlJam Low Hanging Fruit
payload yes
p6eval rakudo 101491: ( no output )
PerlJam payload++
payload rakudo: (^10).to, (^10).max 15:33
p6eval rakudo 101491: ( no output )
jnthn Go for it, but it shouldn't be overriding prefix:<+>, but rather one of the coercion methods.
payload rakudo: say (^10).to, (^10).max
p6eval rakudo 101491: OUTPUT«1010␤»
payload i also think one of them should return 9
same for .from and .min
15:34 Alias_ left
jnthn rakudo: (^10)>>.say 15:34
p6eval rakudo 101491: OUTPUT«0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤»
jnthn I agree. :)
That looks...odd.
payload but i cant decide which ^^
jnthn Hm
15:35 Kodi joined
payload i would go for .to .from, cause they are methods 15:36
15:36 tri joined
payload and just do $.max $.min 15:36
jnthn Oh, I felt the opposite. 15:37
It felt odd to me because
payload well
jnthn rakudo: (0,1,2,3,4).max.say; (^5).max.say;
p6eval rakudo 101491: OUTPUT«4␤5␤»
jnthn That difference is what bothers me a bit.
payload yes
jnthn Probably best to get more opinions before changing it.
PerlJam setting .min/.max appropriately gets rid of .excludes_min and .excludes_max 15:38
payload but this example makes it obvious
jnthn PerlJam: Possibly that too, yes.
payload: Feel free to propose a patch for this, but I'd like to see what others think before it's applied. 15:39
payload then the information of range definition with excludes is gone... but this isnt relevant i think
sure, will do 15:40
PerlJam except that it's currently used in alot of places.
jnthn I think we may depend on knowing that...I'm not sure off hand though.
PerlJam: Right.
For me the real conflict is that we make .max on a range mean something other than what Any.max means.
payload: I'd do the +(1..1000000) opt as a separate patch - that one is probably quite readily appliable. 15:41
jnthn afk for a little bit 15:42
payload rakudo$ grep -Ri excl * # if this would show all places where the exclusion information is relevant, it is manageable 15:44
colomon The problem you guys are having with Ranges is you're not thinking of the non-Int cases.
payload uh, i have to download the test suite
well, i think of it 15:45
colomon 0..4 is not at all the same thing as 0..^5
payload a fallback to succ and count must be there
15:45 cogno joined, satyavvd left 15:48 Bucciarati_ is now known as Bucciarati
colomon The information of excludes is very important to Range. 15:49
15:49 dorlamm left 15:50 masak joined, cogno left
masak oh hai #perl6! 15:50
rindolf masak: hai.
colomon o/
so, can you tell us how many solutions you got yet? ;)
masak I'm a bit behind on the whole contest thing. been moving house yesterday and today. 15:51
here's the plan for the next hour or so:
I'm going to go through the last batch of solutions and accept/reject.
then I'm going to write a blog post, describing the current status and the plan for the next week or so. 15:52
payload hmm, i want to write something like if #( there is a infix:<-> for $a and $b )# but i dont know how to write this check without going into exceptions 15:53
colomon payload: if $a ~~ Numeric and $b ~~ Numeric 15:54
15:55 sftp left 15:57 am0c left
payload i have to think of the smartmatch more often 15:57
15:58 pmurias left
colomon mind you, that's not exactly saying "infix:<-> is available", but it's probably the closest approximation that doesn't involve invoking the multi dispatcher. 15:58
payload yes 15:59
masak rakudo: multi foo(Str) {}; multi foo(Int) {}; say &foo ~~ :(Str); say &foo ~~ :(Int); say &foo ~~ :(Bool)
p6eval rakudo 101491: OUTPUT«Bool::False␤Bool::False␤Bool::False␤»
masak would like that to say TruenTruenFalse
colomon Actually, if $a ~~ Real && $b ~~ Real is probably closer to what you actually want.
flussence rakudo: multi foo(Str) {}; say :(Str) ~~ &foo; 16:01
p6eval rakudo 101491: OUTPUT«Method 'ACCEPTS' not found for invocant of class 'Perl6MultiSub'␤ in main program body at line 22:/tmp/hbj_maQcVh␤»
flussence :(
colomon apparently is overusing "actually" this morning.
arnsholt awaits masak++'s next blog post with bated breath 16:07
16:08 Bzek left
dalek kudo: f3eb64f | KodiB++ | docs/release_guide.pod:
[release_guide] Added instructions for updating the leap-second tables.
 16:08
16:10 Bzek joined
jnthn masak: Is that spec'd in the smartmatch table? 16:15
16:16 sftp joined
jnthn ah, seems so 16:16
Callable Signature sig compatibility
16:17 sftp left
colomon ooo, sweet 16:17
jnthn rakudo: multi foo(Str) {}; multi foo(Int) {}; say any(&foo.candidates>>.signature>>.ACCEPTS(:(Str)) 16:18
p6eval rakudo 101491: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 22␤»
jnthn rakudo: multi foo(Str) {}; multi foo(Int) {}; say any(&foo.candidates>>.signature>>.ACCEPTS(:(Str)))
p6eval rakudo 101491: OUTPUT«any(Bool::False)␤»
masak huh.
jnthn rakudo: say :(Str) ~~ :(Str)
p6eval rakudo 101491: OUTPUT«Bool::False␤»
jnthn Oh 16:19
masak o.O
PerlJam heh
masak submits rakudobug
jnthn Just NYI.
masak and probably LHF too :)
jnthn rakudo: say \("foo") ~~ :(Str)
p6eval rakudo 101491: OUTPUT«Bool::True␤»
jnthn rakudo: say \(42) ~~ :(Str)
p6eval rakudo 101491: OUTPUT«Bool::False␤»
jnthn Yeah
Capture ~~ Signautre is.
The Sig ~~ Sig and Callable ~~ Sig just ain't. 16:20
And yes, probably LHF if somebody is feeling hacky. :)
dalek ast: a2cd3d1 | (Kodi Arfer)++ | S02-builtin_data_types/bool.t:
[bool.t] Removed two fudged tests that were contradicted by unfudged tests.
arnsholt masak: Oh, the deadline was friday not saturday. My bad =) 16:21
16:21 sftp joined
masak arnsholt: yeah. sorry about that. 16:21
arnsholt For some reason my mental calendar had "saturday", not the 14th
My fault, not yours =) 16:22
masak but you're right -- it takes all five tasks to win anyway. and if there's any second prize (I don't know yet), it'll also go to a five-task contestant, I guess.
arnsholt Sounds fair 16:23
masak if I find the tuits, I might put people's non-accepted solutions on display as well. might be interesting from the viewpoint of nice code to look at.
colomon: now I can tell you how many solutions I got. 16:25
colomon \o/
16:25 cogno joined, cogno left
masak five per task, except for p4 which got six submissions, two of which are "second submissions". 16:25
16:25 cogno joined
masak five contestants have sent in accepted solutions. *four* contestants have sent in all five solutions. 16:27
16:27 PacoLinux joined
masak those four are colomon, fox, moritz_, and Util. 16:27
colomon fox?
masak that's what his email address line says. 16:28
ok, I'm going to have a good look at all the solutions, taking careful notes. (after I've blogged)
colomon oooo, if there's someone we haven't heard of sending in a full set of solutions, that's very cool! 16:29
masak matthias, who sent in four solutions, visited the channel the other day to fix a NYI in Rakudo :) 16:30
I have a question for everyone: would you prefer I divulged the winner directly after I've looked through the solutions but before I blog my analysis of each problem -- or should I blog the analyses first and then announce a winner? what do y'all think?
arnsholt I say keep us waiting 16:31
masak sure thing. 16:32
colomon I'm inclined to agree with arnsholt.
masak blogging the analyses might take between one and two weeks.
arnsholt If nothing else, having your analyses first gives a bit more context to how you determine the winner as well
colomon I'd rather get to look at the analyses without having my knowledge of who the winner was color things.
also, you might discover that feedback on the blogposts might influence your decisions, too. 16:33
whereas feedback post-decision would just be trouble.
(if you know what I mean.)
Kodi As a student of the study of decision-making, I should point out that just writing the analyses, if you haven't decided already, might affect your decision. This can be a good or bad thing depending on how you look at it. 16:34
masak Kodi: I have a feeling you're right. I haven't looked as closely as I want on all of the submissions yet.
16:40 payload left
masak oh, and now that we're past the stage where players can affect the contest, let me say one thing that I've wanted to say for weeks: 16:40
p4 is where all the action seems to be.
the other problems are relatively simple and straightforward for people. but p4 is where a lot of different solutions, none completely satisfactory, are put forward. 16:41
(oh, and I know the REAL solution!)
masak hides
16:42 robins is now known as robinsmidsrod 16:44 justatheory joined, spetrea joined
spetrea what's up ? 16:44
masak roof. clouds. sky. :)
flussence aaaaaand my .indent pass rate: 8 oks out of 15 16:45
(yes, I haven't forgotten I was doing it)
masak flussence: nice! what seems to be the blocker for the rest?
flussence working on this one at the moment: not ok 5 - If a line contains too few spaces, only those should be removed 16:46
it seems like it unindents the first line and gives up
flussence pastes it somewhere
gist.github.com/781939 16:47
16:48 dju left
PerlJam What's with the eval? 16:49
16:49 envi left
flussence I couldn't get it to figure out what 0..$outdent meant using plain code, that's a workaround :/ 16:49
16:50 dju joined, dju left, dju joined
PerlJam oh 16:50
16:52 Kodi left
masak flussence: it seems to me that you're doing the "check spaces-only or not" check on all lines at once. I read the spec as saying it should be done on each line individually. 16:54
flussence that's the simplest case, if that check fails it's supposed to fall through to the "..." (which will do something smarter with each line :) 16:55
I guess I forgot to put on line 47 a check that none of the lines starts with a \t... 16:57
masak I mean that I see nothing in the spec about checking the lines together. only about checking them individually. 16:58
flussence takes another look 16:59
masak another way of saying that is that it should be possible to comb whole lines, call a helper sub on each line, and then join them back.
flussence S32/Str:600 says to keep the indent scheme consistent if all lines start the same way, that doesn't appear to apply to the outdenting part... 17:01
I think I might be making this more complicated than it needs to be :) 17:02
17:02 molaf_ left
masak flussence: no, S32/Str:600 only talks about the current line. 17:03
flussence ...I think I might've misunderstood "if these are all the same". whoops. 17:04
masak flussence: if you have any suggestions how to make the spec less ambiguous, I'll happily apply them. 17:07
flussence the "if these are all the same" got me... might be clearer if it explicitly says "for each line" somewhere 17:11
"if the line's indent consists of a single type of space", though that sounds a bit long-winded to me. 17:15
17:16 cogno left
PerlJam flussence: I think your $prefix calculation is a little off. Sometimes tab characters will contribute less than $?TABSTOP spaces to the calculation. (just in case you weren't aware of it) 17:19
flussence I'd thought of that, but the way I interpreted the spec is to just assume they aren't elastic-width 17:20
masak they are. 17:21
flussence hm
masak I guess that only matters when tab characters are exploded during an unindent. 17:24
PerlJam basically.
masak man, this is tricky. I'm starting to think .indent is a really useful function because it abstracts away a lot of details like this :)
flussence ok so if I understand right now, (~@lines).indent($n) should be the same as @lines».indent($n) ? 17:27
i.e. the lines don't affect each other?
(apart from indent(*)) 17:28
masak correct. 17:30
it's as if each line had its own indent character policy.
PerlJam that and you really meant @lines.join.indent($n) since ~@lines will join them with a single space in between :) 17:32
flussence also, is there an easy way to write the first line of the function so it'll make an array of each line's /^^ (\s*) (\N*) (\n?)/ bits?
PerlJam: I'm always getting those two things mixed up :)
PerlJam flussence: that's okay, at least you can focus on something for more than a minute such that you accomplish something. I'm too easily distracted. 17:34
17:37 Su-Shee left 17:45 Su-Shee joined 18:23 rgrau joined 18:35 risou left
masak suddenly everyone is writing Sinatra in Perl 6 :) 18:37
18:39 MayDaniel left
jnthn Well, to be Frank, it is awesome. 18:39
masak :P
jnthn: one of your sharpest in a long time. :)
especially since the sentence has two meanings. 18:40
(hm. that didn't come out quite right. it's not altogether uncommon for puns to have two meanings.) :) 18:46
jnthn :P 18:49
masak "I like this optical illusion -- especially the fact that it can be seen in two ways..." 18:52
18:55 broquaint left, rindolf left 19:03 Vlavv` left 19:05 Cyrus left 19:06 Vlavv joined 19:09 Cyrus joined 19:12 amkrankruleuen left
lue hello zebra PiRats o/ 19:13
masak \o 19:14
what's up with the spelling of the word "aesthetics"? is it "aesthetics", or is it "æsthetics"? 19:18
lue I would go with 'ae' unless you enjoy your ligatures. [I've actually never seen it spelt with æ until just now] 19:19
masak I've seen it spelt with just the 'e' sometimes.
I guess I'm just confused about what preferred and what isn't. 19:20
jnthn masak: It's a weird word. It looks right to me with the "e"
masak .oO( "a wæird word" ) 19:21
jnthn :P
Or at least, at school I was taught to spell it aesthetics, anyways. :)
lue I honestly don't think it matters whether it's ae or æ, it'll be the same word [people don't get in an uproar over wikipædia :)]
jnthn From Wikipedia, it looks like it was probably put into English from German. 19:22
masak ah.
jnthn Yeah, looks like it's the same as when ä in Swedish sometimes gets translitted as ae 19:23
To me it looks mis-spelt without the a at the start, tbh. 19:24
masak good to know.
jnthn But I don't think I could objectively call it wrong either.
lue Maybe you should conſider uſing a long s? 19:25
jnthn I don't þink so. :P 19:26
masak tempting, ðough... 19:28
19:29 shi left
lue There used to be an old rule where you had to use an umlaut to designate another syllable when it could be read as one, such as coördinate. 19:30
masak never really liked that spelling :) 19:31
though I'm fine with it in "naïve" 19:32
jnthn Oh, *that's* what the unlaut in that word means.
19:32 Alias_ joined 19:33 payload joined 19:34 MayDaniel joined
sorear good * #perl6 19:34
lue Oh, English. Your crazy idiosyncracies never fail to amuse us.
hello sorear o/
19:34 Alias left
sorear mm, coördinate 19:35
masak g'whatever, sorear.
sorear always thought Einstein made that one up
masak has blogged: strangelyconsistent.org/blog/the-pe...now-closed
19:36 simcop2387 left
jnthn EWWW! 19:37
colomon masak: so glad to be part of "the people". ;)
masak :D
jnthn decides not to compete with a worse analogy in his forthcoming blog post 19:38
lue
.oO(We're getting five restaurant critiques highlighting the best of all the ad-hoc restaurants built in the commercial sector of Camelia-land)
 19:42
moritz_ back 19:43
masak moritz_! \o/ 19:44
jnthn o/ moritz_
lue hello moritz_ o/
colomon \o
moritz_: everything still going okay?
moritz_ colomon: yep, first steps towards breast feeding showed success 19:46
colomon \o/
masak \o/
19:47 aindilis left 19:49 coldhead joined 19:51 dukeleto left 19:52 dukeleto joined
lue rakudo: my @a = [1,2],[3,4]; say @a[0][1]; say @a[0;1] 19:54
p6eval rakudo f3eb64: OUTPUT«2␤3 4␤» 19:55
lue checks if that bug is reported
masak lue: it's simply a TODO
lue: @a[0;1] is treated as @a[1] because ; in .[] isn't implemented
19:56 [Coke] left 19:57 willco joined
willco codepad.org/CSe2YH1I 19:58
anybody have any idea why this doesn't work? it's killing me.
it just outputs a single ':'
masak willco: hi, this is #perl6. 19:59
when we're in an exceptionally good mood, we give Perl 5 support. but it's not really the channel for that.
try #perlhelp on irc.perl.org 20:00
20:05 cogno joined 20:12 simcop2387 joined 20:14 cogno left 20:19 snikkers joined 20:20 Trashlord left 20:21 shi joined 20:26 y3llow_ joined, y3llow left, payload left, y3llow_ is now known as y3llow, pothos_ joined 20:27 pothos left 20:28 pothos_ is now known as pothos
dalek odel: 74c1fc0 | jonathan++ | overview.pod:
[docs] Add an overview of 6model.
 20:30
odel: 0419a1f | jonathan++ | overview.pod:
[docs] Try and get something that renders nicer with GitHub's POD renderer.
 20:36
odel: c0a00fa | jonathan++ | overview.pod:
[docs] Tssk. Next time I just write plain text. Or HTML.
 20:38
masak :)
jnthn OK, now it looks non-awful. :) 20:39
dalek ast: 2769a7c | moritz++ | S0 (2 files):
fudge relation.t for niecza
moritz_ forgot to push that commit
niecza: say +'24'
p6eval niecza v1-149-g56c80b0: OUTPUT«Can't open perl script "niecza_eval": No such file or directory␤»
moritz_ urks
20:40 tri left
wolverian jnthn: s/stratergies/strategies/ 20:41
lue afk 20:43
dalek albot: 83a6deb | moritz++ | / (2 files):
niecza now uses "make", and the executable is different
 20:44
odel: 7d4f8de | jonathan++ | overview.pod:
[docs] Typo spotted by wolverian++.
 20:45
jnthn
.oO( shudda spellchuck )
 20:46
tadzik o/ 20:49
20:49 aindilis joined
masak "loopy"! \o/ 20:50
20:55 MayDaniel left, araujo left 20:57 countley joined 20:59 countley left 21:00 spetrea left
moritz_ is curious for the new idioms that masak++ has mentioned on his blog 21:03
21:08 araujo joined
masak I'll be sure to highlight them properly when I post. 21:12
21:13 dukeleto left 21:14 dukeleto joined
moritz_ hurry up! 21:17
oh, two new chapters of HPatMoR
masak ooh
arnsholt colomon: BTW, I messed up on the p6cc deadline, so if you want to look at my p4 code, I'll have to put it somewhere =) 21:18
masak is currently watching Evil Dead III: Army of Darkness, indirectly referenced in MoR
the movie's bad in a slightly enjoyable way :) 21:21
21:21 pmurias joined 21:22 mberends left 21:23 pracacc001 joined 21:26 Vlavv left
pracacc001 rakudo: "helo".say 21:27
p6eval rakudo f3eb64: OUTPUT«helo␤»
moritz_ masak: have you found the RSS feed for HPatMoR?
21:27 dual joined, pracacc001 left
moritz_ demented.no-ip.org/~feep/rss_proxy.cgi?5782108 # soemtimes down, but quite reliable when up 21:27
masak moritz_: aye, I have. but I haven't checked my Reader today. 21:29
colomon arnsholt: oh no! please do put it up somewhere... 21:30
Army of Darkness!
arnsholt colomon: heim.ifi.uio.no/arnskj/p6cc2010/p4/code is the frontend code, heim.ifi.uio.no/arnskj/p6cc2010/p4/lib/Kayles.pm is the fun stuff 21:33
From the directory names, the rest of my code should be trivially findable if you're interested =)
moritz_ that's cheating, you used THEORY! 21:34
:-)
justatheory coughs
masak arnsholt++
arnsholt: now I'm even more sorry you didn't make the deadline! :-|
arnsholt moritz_: Why, thank you =) 21:35
moritz_ arnsholt: :-)
arnsholt As I mentioned to colomon the other day, I went kinda bananas on the research on p4
masak arnsholt: seeing the work "kayles" there tells me I want to look closer at your solution.
arnsholt But it was an interesting field to look into 21:36
masak arnsholt: yes, that was the idea. it'll be interesting to write up p4.
arnsholt I vaguely knew of the field, but it was cool to learn a bit more
moritz_ I lacked the initial knowledge, so I had basically no idea what to search for
arnsholt masak: I was happy it was normal play kayles though. Misere play would have been a lot harder
masak misere play doesn't Grundy at all, IIUC. 21:37
arnsholt I didn't know either actually, I was just struck by the idea that "this sounds like perfect play should be possible"
So I just noodled around on Wikipeda after christmas until I found it
masak :)
arnsholt Yeah, from my tidbits of research making it misere instantly makes the game harder, more or less 21:38
masak the ending of Army of Darkness was as random as the rest of the movie :P
arnsholt I really wanted to implement the n log n matrix chain algorithm as well, but since the paper is > 100 pages long I decided DP was enough 21:39
masak "Oh no! I slept too long! Haaahahahahahah!"
moritz_ Misere game sounds like Antichess to me :-)
masak arnsholt: DP was what I was hoping people would go for. 21:40
arnsholt moritz_: Yeah, essentially. The player who has no move left wins
sorear Who currently has trouble with building Niecza?
arnsholt masak: Yeah, it's the obvious solution (at least to my eyes)
I instantly thought "oh, I'll just use CKY parsing" =)
moritz_ suddenly feels so dumb 21:41
masak arnsholt: "CKY parsing"? 21:42
moritz_: why?
arnsholt CKY is a DP CFG parsing algorithm
21:42 Vlavv_ joined
jnthn moritz_: fwiw, I haven't understood almost any of the last 4 screenfulls. :) 21:42
arnsholt In this context it means my matrix chain solver was a bottom up DP algorithm
moritz_ masak: because I missed some readily available theory on a subject, and handed in a quite imperfect solution 21:43
arnsholt Starting by finding the costs for adjacent pairs of matrices, and working its way upwards through bracketings
masak moritz_: don't worry, everyone did. you all blew p4, except arnsholt who blew the deadline... :)
moritz_ masak: :-)
masak arnsholt: you do realize that you're up for at least an honourable mention... :)
arnsholt Thank you =) 21:44
masak arnsholt: yes, exactly. I didn't know there was a name for that. gotta look up CKY...
arnsholt Also, you should use the "everyone blew p4 ..." line. It has a nice ring to it
masak :)
oh, I'm going to have such a field day with p4.
moritz_ masak: actually I tried a DP solution for p4, and failed
masak I can imagine. 21:45
arnsholt CKY is just one application of the technique, I think. But it's an interesting algorithm
A bit limited in that it only does binary branching grammars, but oh well
masak anyone who likes can consider the secret agenda of this whole contest being to spring Sprague-Grundy theory on people :)
it's an "only works once" trick. :P
arnsholt In that case, consider me educated =)
masak arnsholt: binary branching grammars fit quite well with binary operators :) 21:46
moritz_ arnsholt: well, you can always use Chomsky normalization to get the grammar in binary branching form
arnsholt Yeah, Chomsky normalisation is usually mentioned right before or right after SKY is introduced 21:47
masak huh? SKY or CKY?
arnsholt But it's not always desireable, especially because IIRC the grammars are only weakly equivalent
Er, CKY. Typo
pmurias sorear: i had problems compiling the mono from github, but niecza works on 2.8.1
moritz_ arnsholt: actually CKY is the only reason that Chomsky normalization is taught at all :-) 21:48
arnsholt And if you're parsing a natural language grammar, converting to CNF may lose some of the structures you're interested in
Yeah. The whole binary-branching thing makes the linguist part in me cry
I had to read up on it for a course I took. It's so ridiculous! >.<
moritz_ just consider it isomorphic to...
that's like "it all makes perfect sense in n-dimensional space" :-) 21:49
arnsholt Oh, and about feeling stupid, I tried to read up about the Ising model earlier today. Didn't exactly feel very clever either then =)
jnthn I blug! 6guts.wordpress.com/2011/01/16/nqp-...t-and-jvm/
sorear pmurias: I thought you said earlier that it was non-functional on 2.8.1
something about Array.keys
moritz_ arnsholt: I never understood the analytical solution of 2D Ising model either :-)
arnsholt I didn't understand any of it, more or less =) 21:50
I think I'mstarting to get graphical models, though
jnthn: Excellent, the English language needs more strong verbs! 21:52
21:54 IllvilJa left
moritz_ it dus 21:54
masak it dudn't! 21:55
pmurias sorear: sorry it works on 2.8.2 21:56
it's non functional on 2.8.1 21:57
masak I'm on mono 2.6.7, and niecza doesn't build here. 22:03
22:03 MayDaniel joined
tadzik oh, 6model LHF! 22:04
jnthn You know you want to. 22:05
tadzik riiight 22:06
jnthn cheers on tadzik :) 22:07
masak moritz_: which mono version did you have again?
moritz_ masak: 2.6.7-4 22:08
(where the -4 is Debian specific, iirc)
masak moritz_: can you build niecza?
tadzik I have 2.6.7 and I can build, but not run
moritz_ masak: yes 22:09
tadzik well, I can run, but I can't Helloworld :)
masak I'm getting "MSBUILD: error MSBUILD0003: Please specify the project or solution file to build, as none was found in the current directory."
moritz_ masak: you need to use 'make' now 22:10
not xbuild anymore
masak oh.
right, it's even in the README.pod
22:10 orafu left
masak moritz_++ 22:10
22:10 orafu joined 22:12 am0c joined 22:15 jedai left 22:16 kjeldahl left 22:17 jedai joined 22:19 sftp left 22:27 shi left 22:35 shi joined 22:39 Trashlord joined
snarkyboojum moritz_: congrats on the healthy bub and healthy mum :) 22:39
dukeleto submitted a talk to PGCon 2011 about PL/Perl6 today 22:40
moritz_: congrats! 22:41
22:41 MayDaniel left
dukeleto jnthn: i am reuding your blug 22:41
22:44 sftp joined
jnthn dukeleto: :) 22:47
dukeleto jnthn: thanks for writing a 6model overview! 22:49
jnthn dukeleto: Hope it makes sense. :)
dukeleto jnthn: i will let you know if it doesn't :)
jnthn :)
OK, sleep time for me tomorrow...gotta get up earlyish.
22:50 kaare_ left
jnthn wtf 22:50
...how did the tomorrow end up 4 words too early? :S
sleep &
22:52 vmspb joined, cafesofie joined
masak that's simply how early tomorrow comes :) 22:52
22:59 vmspb left
masak sorear: I think tonight's attempt to build niecza has taught me that giving Virtualbox 1024 MB of RAM isn't enough. 23:00
the build hasn't failed yet, but it seems to have reached some kind of swapping plateau.
masak decides to try again tomorrow with more RAM
'night, #perl6.
23:00 masak left 23:07 vmspb joined 23:15 [Coke] joined 23:17 am0c left
snarkyboojum how does one run the test harness for Niecza now? 23:19
t/run_spectests was working.. but I guess needs to be updated now there's not niecza_eval around anymore? 23:20
23:38 ashleydev joined 23:42 PacoLinux left, flatwhatson left 23:45 KatrinaTheLamia left 23:52 shi left 23:55 KatrinaTheLamia 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!