Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2013-10-13

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 kurahaupo_ joined #perl6
00:06 kurahaupo left #perl6
00:12 Psyche^ joined #perl6
00:14 lue .ask jnthn is there a technical reason the NQP Configure[VM].pl files aren't folded into one Configure.pl file, or is just a case of priorities?
00:14 yoleaux lue: I'll pass your message to jnthn.
00:26 fuad left #perl6
00:29 xinming joined #perl6
00:56 rpr223 joined #perl6
01:04 ivanshmakov joined #perl6
01:05 diakopter lue: Yes we would like to merge them..
01:22 benabik joined #perl6
01:25 FROGGS_ joined #perl6
01:31 BenGoldberg joined #perl6
01:45 SmokeMachine joined #perl6
01:47 ivanshmakov joined #perl6
02:00 retupmoca joined #perl6
02:00 ivanshmakov joined #perl6
02:03 btyler joined #perl6
02:06 colomon joined #perl6
02:18 ivanshmakov joined #perl6
02:35 ivanshmakov joined #perl6
02:42 sizz_ joined #perl6
02:53 retupmoca joined #perl6
02:54 xenoterracide__ joined #perl6
02:56 ivanshmakov joined #perl6
03:16 colomon joined #perl6
03:22 ivanshmakov joined #perl6
03:24 zakharyas joined #perl6
03:49 preflex_ joined #perl6
03:52 retupmoca I can't get panda to bootstrap with rakudo 2013.09: https://gist.github.com/retupmoca/6957964
03:52 retupmoca I'm just running ./bootstrap.pl - should I be doing something different?
03:56 dalek perl6-roast-data: 711a2a8 | coke++ | / (2 files):
03:56 dalek perl6-roast-data: today (automated commit)
03:56 dalek perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/711a2a875f
04:14 ivanshmakov joined #perl6
04:16 [Coke] nqp: say(nqp::index("this", "h+", 1));
04:16 camelia nqp: OUTPUT«-1␤»
04:17 [Coke] nqp: say(nqp::index("this", "h", 1));
04:17 camelia nqp: OUTPUT«1␤»
04:17 [Coke] nqp: say(nqp::index("this", "h", 1,2,3,4,5,6));
04:17 camelia nqp: OUTPUT«Error while compiling block : Error while compiling op call (source text: "nqp::index(\"this\", \"h\", 1,2,3,4,5,6)"): Error while compiling op index (source text: "nqp::index(\"this\", \"h\", 1,2,3,4,5,6)"): Error while compiling op indexfrom: Operation 'indexfrom' requ…»
04:17 [Coke] nqp-jvm: say(nqp::index("this", "h", 1,2,3,4,5,6));
04:17 camelia nqp-jvm: OUTPUT«Operation 'indexfrom' requires 3 operands␤  in  (src/stage2/gen/QAST.nqp:319)␤  in  (src/stage2/gen/QAST.nqp:227)␤  in compile_op (src/stage2/gen/QAST.nqp:227)␤  in as_jast (src/stage2/gen/QAST.nqp:3683)␤  in as_jast (src/stage2/gen/QAST.nqp:2958)␤  in  (…»
04:19 BenGoldberg nqp-jvm: say(nqp::index("this", "h", (1,2,3,4,5,6)));
04:19 camelia nqp-jvm: OUTPUT«-1␤»
04:19 BenGoldberg Parens help :)
04:20 BenGoldberg nqp-m: say(nqp::index("this", "h", 1,2,3,4,5,6));
04:20 camelia nqp-moarvm: OUTPUT«Bytecode validation error at offset 22, instruction 4:␤register operand index 50 out of range 0..9␤»
04:20 BenGoldberg :(
04:22 [Coke] that seems LTA, but I think in previous cases, the answer has been "don't do that, then."
04:23 * [Coke] wonders why the asymmetry between index and rindex opcodes in nqp.
04:30 ivanshmakov joined #perl6
04:41 dalek nqp: 50407c1 | coke++ | docs/ops.markdown:
04:41 dalek nqp: document more opcodes.
04:41 dalek nqp: review: https://github.com/perl6/nqp/commit/50407c10d9
04:43 [Coke] updated https://gist.github.com/coke/6799633
04:49 ivanshmakov joined #perl6
04:52 * [Coke] asks folks to pick an op from that gist and write up a sentence on what it does so I can throw that into the docs. Danke.
05:10 ivanshmakov joined #perl6
05:33 ivanshmakov joined #perl6
05:41 varna joined #perl6
05:47 dalek nqp: 67d8988 | moritz++ | docs/ops.markdown:
05:47 dalek nqp: [opcode docs] harmonize header levels
05:47 dalek nqp: review: https://github.com/perl6/nqp/commit/67d8988626
05:47 dalek nqp: 68d0dbd | moritz++ | docs/ops.markdown:
05:47 dalek nqp: [doc] document nqp::getattr
05:47 dalek nqp: review: https://github.com/perl6/nqp/commit/68d0dbda57
05:47 darutoko joined #perl6
05:53 dalek nqp: a6a00f1 | moritz++ | docs/ops.markdown:
05:53 dalek nqp: [docs] document nqp::bindattr
05:53 dalek nqp: review: https://github.com/perl6/nqp/commit/a6a00f15d1
05:55 ivanshmakov joined #perl6
05:55 moritz nqp: nqp::call(say, 42, 'foo')
05:55 camelia nqp: OUTPUT«invoke() not implemented in class 'NQPMu'␤current instr.: '' pc 99 ((file unknown):41511373) (/tmp/YRFI0So5pv:1)␤»
05:55 moritz nqp: nqp::call(&say, 42, 'foo')
05:55 camelia nqp: OUTPUT«42foo␤»
05:57 moritz nqp: say(nqp::can(42, 'foo'))
05:57 camelia nqp: OUTPUT«0␤»
05:58 moritz nqp: say(nqp::can(NQPMu, 'bless'))
05:58 camelia nqp: OUTPUT«1␤»
05:59 dalek nqp: a7d7169 | moritz++ | docs/ops.markdown:
05:59 dalek nqp: [docs] document nqp::create, correct nqp::can docs
05:59 dalek nqp: review: https://github.com/perl6/nqp/commit/a7d7169f6d
06:02 moritz nqp: my $x; nqp::bind($x, '42'); say $x
06:02 camelia nqp: OUTPUT«Confused at line 2, near "say $x"␤current instr.: 'panic' pc 14748 (src/stage2/gen/NQPHLL.pir:5229) (src/stage2/gen/NQPHLL.nqp:279)␤»
06:02 moritz nqp: my $x; nqp::bind($x, '42'); say($x)
06:02 camelia nqp: OUTPUT«42␤»
06:02 moritz nqp: nqp::bind(42, 23)
06:02 camelia nqp: OUTPUT«Error while compiling block : Error while compiling op bind (source text: "nqp::bind(42, 23)"): First child of a 'bind' op must be a QAST::Var␤current instr.: '' pc 55082 (src/stage2/QAST.pir:20123) (src/stage2/QAST.nqp:3316)␤»
06:10 varna left #perl6
06:11 dalek nqp: 0cde6db | moritz++ | docs/ops.markdown:
06:11 dalek nqp: [docs] bind, callmethod
06:11 dalek nqp: review: https://github.com/perl6/nqp/commit/0cde6dbf05
06:14 kaare_ joined #perl6
06:15 ivanshmakov joined #perl6
06:18 dwarring joined #perl6
06:35 moritz oh my. Now I was about to write a small script which turns the ops.markdown into HTML, adds some meta info (version, where the file comes from) and run it periodically
06:36 moritz and then I realized that github does that all for us already, and linking to https://github.com/perl6/nqp/blob/master/docs/ops.markdown is all it takes :-)
06:38 [Coke] moritz++: I've been documenting return value/exceptions based on java. If parrot is different, please mark the opcode as working differently across vms. (in which case we need to fix it.)
06:38 [Coke] moritz++ for documenting things. Thanks!
06:39 [Coke] would create be better written as create(Type $type) ?
06:39 [Coke] (if it's not a type it throws an exception, I'm assuming.
06:39 [Coke] ... and bed. moritz++ again.
06:40 japhb__ g'night, [Coke]
07:09 BenGoldberg joined #perl6
07:11 woolfy oO(over 30 minutes of quiet)
07:16 moritz [Coke]: there is no type named Type
07:17 moritz nqp-jvm: say(nqp::can(NQPMu, 'bless'))
07:17 camelia nqp-jvm: OUTPUT«1␤»
07:17 moritz nqp-jvm: say(nqp::can(NQPMu, 'blerg'))
07:17 camelia nqp-jvm: OUTPUT«0␤»
07:23 moritz nqp: say(nqp::sha1('foo'))
07:23 camelia nqp: OUTPUT«0BEEC7B5EA3F0FDBC95D0DD47F3C5BC275DA8A33␤»
07:53 aindilis joined #perl6
07:56 Ben_Goldberg joined #perl6
07:59 itz joined #perl6
08:09 kurahaupo joined #perl6
08:09 kurahaupo left #perl6
08:10 masak aloha on this fine Sunday, com-p6ers.
08:13 FROGGS morning
08:16 Rotwang joined #perl6
08:20 dansamo joined #perl6
08:36 aindilis joined #perl6
08:38 spider-mario joined #perl6
08:39 kurahaupo joined #perl6
08:56 xenoterracide__ joined #perl6
09:02 spider-mario_ joined #perl6
09:02 spider-mario_ joined #perl6
09:10 berekuk joined #perl6
09:16 dmol joined #perl6
09:28 wsri joined #perl6
09:29 SamuraiJack_ joined #perl6
09:33 grondilu joined #perl6
09:36 dalek panda: ae8936b | moritz++ | lib/Panda/Common.pm:
09:36 dalek panda: remove deprecated * from bless
09:36 dalek panda: review: https://github.com/tadzik/panda/commit/ae8936bcc4
09:41 moritz tadzik: seems like panda bootstrap is broken :( "build stage failed for Shell::Command: Failed building lib/Shell/Command.pm" without any visible error message before
09:45 arnsholt jnthn: No objections to merging nativecall-jvm into master on my part
09:54 dalek nqp: 72427c7 | (Tobias Leich)++ | / (11 files):
09:54 dalek nqp: update stage0, map bool_I and isbig_I, add bool_I's tests
09:54 dalek nqp:
09:54 dalek nqp: moritz++
09:54 dalek nqp: review: https://github.com/perl6/nqp/commit/72427c7f1f
10:02 dalek nqp: d3790e8 | moritz++ | t/nqp/60-bigint.t:
10:02 dalek nqp: import bigint bool_I tests from moarvm/nqp-cc
10:02 dalek nqp: review: https://github.com/perl6/nqp/commit/d3790e8608
10:02 dalek nqp: fad34d4 | moritz++ | src/vm/moar/ (10 files):
10:02 dalek nqp: Merge branch 'master' of git://github.com/perl6/nqp
10:02 dalek nqp: review: https://github.com/perl6/nqp/commit/fad34d4f91
10:31 woolfy joined #perl6
10:32 cognominal joined #perl6
10:35 diakopter why can I never find the man-or-boy test written in nqp
10:36 denis_boyun joined #perl6
10:36 moritz because whoever wrote it didn't check it into nqp as a test, and put a 'This is the famous man-or-boy test' comment in front of it
10:36 moritz then a simple 'git grep boy' would find it withing 2s
10:39 diakopter so.. where is it
10:39 nwc10 um, I used google and the top hitis https://github.com/japhb/perl6-bench/blob/master/nqp/rc-man-or-boy-test
10:40 diakopter sure enough :)
10:41 diakopter except that's the slow one that creates $B needlessly most of the time
10:41 diakopter oh I remember where I put it
10:41 diakopter in the irclog
10:44 diakopter nqp-m: sub A($k, $x1, $x2, $x3, $x4, $x5) { return $x4() + $x5() if $k <= 0; my $B := sub () { $k := $k - 1; return A($k, $B, $x1, $x2, $x3, $x4) }; return $B() }; sub K($n) { sub () { $n } }; my $s := nqp::time_n; say(A(18, K(1), K(-1), K(-1), K(1), K(0) )); say(nqp::time_n - $s);
10:44 camelia nqp-moarvm: OUTPUT«Bytecode validation error at offset 22, instruction 4:␤register operand index 50 out of range 0..9␤»
10:44 diakopter nqp-m: say(1)
10:44 camelia nqp-moarvm: OUTPUT«Bytecode validation error at offset 22, instruction 4:␤register operand index 50 out of range 0..9␤»
10:44 diakopter goo
10:44 diakopter boo
10:45 FROGGS I guess it doesn't know that it needs to rebuild nqp
10:45 diakopter ops
10:45 diakopter oops
10:46 * diakopter hopes "someone else" fixes it before I get back from breakfast
10:49 diakopter nqp: sub A($k, $x1, $x2, $x3, $x4, $x5) { return $x4() + $x5() if $k <= 0; my $B := sub () { $k := $k - 1; return A($k, $B, $x1, $x2, $x3, $x4) }; return $B() }; sub K($n) { sub () { $n } }; my $s := nqp::time_n; say(A(18, K(1), K(-1), K(-1), K(1), K(0) )); say(nqp::time_n - $s);
10:49 camelia nqp: OUTPUT«maximum recursion depth exceeded␤current instr.: '' pc 296 ((file unknown):74239404) (/tmp/aJmEIZJ8f7:1)␤»
10:50 diakopter nqp: sub A($k, $x1, $x2, $x3, $x4, $x5) { return $x4() + $x5() if $k <= 0; my $B := sub () { $k := $k - 1; return A($k, $B, $x1, $x2, $x3, $x4) }; return $B() }; sub K($n) { sub () { $n } }; my $s := nqp::time_n; say(A(15, K(1), K(-1), K(-1), K(1), K(0) )); say(nqp::time_n - $s);
10:50 camelia nqp: OUTPUT«maximum recursion depth exceeded␤current instr.: '' pc 296 ((file unknown):52624812) (/tmp/BLAU0XeTal:1)␤»
10:51 pecastro joined #perl6
11:03 SamuraiJack_ joined #perl6
11:06 pecastro joined #perl6
11:08 diakopter WOW man-or-boy spends >51% of its time in the GC at input 18
11:08 diakopter *awesome* GC optimization opportunity
11:21 masak sounds like.
11:21 masak who knew man-or-boy would be useful in so many ways? :)
11:31 masak the "Heisenberg's uncertainty joke" needs its own joke category: https://twitter.com/patrickdijusto/status/388313018175483905
11:33 jnthn afternoon, #perl6
11:33 yoleaux 00:14Z <lue> jnthn: is there a technical reason the NQP Configure[VM].pl files aren't folded into one Configure.pl file, or is just a case of priorities?
11:33 yoleaux 02:23Z <JimmyZ> jnthn: I got another backstrace: https://gist.github.com/zhuomingliang/6957379
11:34 jnthn Sadly, those backtraces are largely useless for finding the actual problem...
11:34 diakopter erm
11:34 jnthn (Since it tends to have gone wrong a while back)
11:35 diakopter why does one view show 51% in the gc
11:35 diakopter but another view shows only 8%
11:37 jnthn diakopter: lies, dammed lies, statistics? :)
11:39 dalek nqp: 6aba3c2 | (Tobias Leich)++ | src/vm/moar/QAST/QASTRegexCompilerMAST.nqp:
11:39 dalek nqp: remove useless condition, BenGoldberg++
11:39 dalek nqp: review: https://github.com/perl6/nqp/commit/6aba3c2263
11:41 jnthn arnsholt: merging
11:43 jnthn BenGoldberg++ # nice catch
11:44 rafl joined #perl6
11:44 dalek Heuristic branch merge: pushed 23 commits to nqp by jnthn
11:45 xinming joined #perl6
11:47 jnthn Gee, we've been keeping busy. 218 commits to the NQP repo sice the last monthly release...
11:52 dalek rakudo/nom: e22ac2a | jnthn++ | tools/build/NQP_REVISION:
11:52 dalek rakudo/nom: Get an NQP with basic NativeCall support for JVM.
11:52 dalek rakudo/nom:
11:52 dalek rakudo/nom: Also has a fix that lets us write NativeCall's array stuff in a way
11:52 dalek rakudo/nom: that is more portable between backends.
11:52 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e22ac2aa6d
11:53 dalek zavolaj: eeee9aa | jnthn++ | lib/NativeCall.pm6:
11:53 dalek zavolaj: Use standard nqp::atpos ops.
11:53 dalek zavolaj:
11:53 dalek zavolaj: These are now supported properly on 6model array REPR things on Parrot
11:53 dalek zavolaj: also.
11:53 dalek zavolaj: review: https://github.com/jnthn/zavolaj/commit/eeee9aa1e7
11:53 jnthn eeee!
11:54 jnthn arnsholt: OK, all done. Happy hacking ;-)
12:01 xinming joined #perl6
12:16 FOAD joined #perl6
12:17 woosley1 joined #perl6
12:44 GlitchMr rn: my Int $int = 7; if 0 { $int.Qet }
12:44 camelia rakudo e22ac2, niecza v24-98-g473bd20: ( no output )
12:44 GlitchMr It would be nice if error reporting would exist here.
12:48 jnthn GlitchMr: that's hard to do in general
12:48 jnthn GlitchMr: Because what if you did an augment elsewhere, etc.
12:48 GlitchMr hm, right
12:57 colomon joined #perl6
13:02 tadzik moritz: hmm, I'll take a look at it
13:02 tadzik (hello #perl6)
13:04 jnthn o/ tadzik
13:05 hoelzro ahoy tadzik
13:05 masak GlitchMr: methods are late-bound. very different from subs.
13:06 Rotwang joined #perl6
13:10 masak "Any sufficiently analyzed magic is indistinguishable from SCIENCE!" http://www.girlgeniusonline.com/comic.php?date=20081205
13:12 Rotwang I like all the characters have oversized breasts
13:14 pernatiy joined #perl6
13:14 masak didn't even think of that.
13:22 colomon joined #perl6
13:26 dalek nqp: fb2dbb0 | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
13:26 dalek nqp: Map hllize/hllizefor on MoarVM.
13:26 dalek nqp: review: https://github.com/perl6/nqp/commit/fb2dbb0b57
13:35 spider-mario joined #perl6
13:39 dalek rakudo/moar-support: cc02954 | jnthn++ | .gitignore:
13:39 dalek rakudo/moar-support: Add .moarvm to .gitignore.
13:39 dalek rakudo/moar-support: review: https://github.com/rakudo/rakudo/commit/cc0295428f
13:39 dalek rakudo/moar-support: 2461a84 | jnthn++ | src/vm/moar/Perl6/Ops.nqp:
13:39 dalek rakudo/moar-support: Stub a Perl6::Ops.
13:39 dalek rakudo/moar-support:
13:39 dalek rakudo/moar-support: It's mostly a commented out version of the JVM one. Need to get the
13:39 dalek rakudo/moar-support: extops infrastructure in place to start filling this out.
13:39 dalek rakudo/moar-support: review: https://github.com/rakudo/rakudo/commit/2461a84a27
13:42 iSlug joined #perl6
13:51 kaare_ joined #perl6
13:54 denis_boyun joined #perl6
14:14 PacoAir joined #perl6
14:20 arnsholt jnthn: Cheerio!
14:26 jnthn :)
14:27 arnsholt jnthn: I get 8 failed tests in file-ops.t and one in jvm/02-pipes.t, that expected?
14:29 ajr joined #perl6
14:29 jnthn arnsholt: Think I've seen those too
14:30 jnthn I'd thought they were Windows-specific though...
14:32 arnsholt And it looks like make install on JVM doesn't work properly here
14:32 arnsholt Can't run NQP from outside my nqp directory
14:34 arnsholt Oh, right. I have to run nqp/install/bin/nqp, not nqp/nqp
14:37 hoelzro jnthn: do you have a TODO list for Mokudo?
14:37 timotimo roarkudo
14:38 hoelzro hmm
14:38 hoelzro that's tough.
14:38 arnsholt I guess Mokudo should be the name during Movember, and Roarkudo the rest of the year =)
14:39 jnthn hoelzro: Get the compiler to build, get extops in place, get the MOP to build, get the bootstrap to build... :)
14:39 jnthn hoelzro: Though at the moment Actions seems to get miscompiled
14:40 timotimo i still have no clue why i can't build nqp on moar; there's some misparse going on
14:41 jnthn timotimo: Yeah, that's osme kind of memory corruption bug
14:41 timotimo most probably
14:42 timotimo oh, hold on
14:42 timotimo something changed
14:43 timotimo false alarm, i'm still suffering from the bug
14:43 [Coke] moritz: I know there is no type named type. Please read the top of the docs file. if you want to rework everything to use real perl6 types and signatures, that's fine.
14:44 arnsholt jnthn: Did you try to run the Zavolaj test suite using Rakudo/JVM?
14:44 jnthn arnsholt: No, I just ran the win32 API call example
14:44 arnsholt 'k
14:44 jnthn arnsholt: Trouble with the test suite is that it needs a C compiler, iirc
14:45 arnsholt Yep
14:45 jnthn timotimo: Compiler? GCC?
14:45 arnsholt I just realized =)
14:45 timotimo yes, gcc
14:45 timotimo gcc (Ubuntu/Linaro 4.7.3-1ubuntu1) 4.7.3
14:46 arnsholt Previously we've been able to get cc and friends from $*VM<config>, but that's not set on JVM, of course >.<
14:47 jnthn timotimo: How good are you with gdb?
14:47 timotimo i can sort of kind of handle it
14:47 timotimo i don't have much experience, but i kind of know what commands exist
14:48 jnthn Do you know how to set breakpoints?
14:48 timotimo yes
14:48 timotimo i just need to know where ;)
14:48 diakopter timotimo: have you tried completely separate/fresh clones?
14:49 timotimo i can try that, sure
14:49 jnthn OK, the way I'm looking into these is putting a breakpoint on the first statement in orchestrate.c:run_gc
14:49 jnthn line 262 or so
14:49 diakopter "these"?
14:49 jnthn This will break every time we are doing a GC run
14:49 jnthn diakopter: missing MVMROOT
14:50 diakopter Rotwang: welcome to comics?
14:50 diakopter Rotwang: erm, art in general?
14:50 JimmyZ timotimo: I got the same errot
14:50 JimmyZ *error
14:50 jnthn timotimo: If when it stops there you bt, you'll see where we are at that point in time
14:50 colomon diakopter, Rotwang: certainly Phil Foglio comics, at any rate...
14:50 hoelzro hmm
14:51 hoelzro Moar NQP takes 2GB to build?
14:51 diakopter until yesterday
14:51 jnthn hoelzro: Only if you haven't updated your Moar recently :)
14:51 jnthn hoelzro: If you have it will max out at 160MB or so :)
14:51 jnthn We fixed a, uh, leak. :)
14:51 hoelzro I just updated Moar =/
14:51 colomon \o/
14:51 hoelzro hmm
14:51 hoelzro nvm
14:51 jnthn Did you make install it to the place your NQP is looking?
14:51 hoelzro moar commits!
14:52 timotimo on host07 i can't build the nqp either, but i can debug that later
14:53 diakopter timotimo: what are your exact configure lines
14:54 timotimo perl Configure.pl for moar and perl ConfigureMoar.pl --prefix=../ for nqp
14:54 timotimo actually --prefix=. for moar and ../ for nqp
14:55 diakopter but..
14:55 diakopter you have nqp checked out under moar?
14:55 timotimo yes
14:55 diakopter under "nqp"?
14:56 diakopter oh
14:56 timotimo yes
14:56 diakopter your problem is you need to install somewhere other than the root of moar
14:56 timotimo jnthn: i've set up the break point, but i'm not sure what i should be looking for
14:56 timotimo i ... do?
14:56 hoelzro hmm
14:57 diakopter yes, try ../install for moar and ../../install for nqp
14:57 jnthn diakopter: I'm not sure that's a huge problem, fwiw
14:57 jnthn diakopter: At lesat, I can't think waht it breaks...
14:57 diakopter it's overwriting under lib/ I think
14:57 jnthn oh...
14:57 jnthn yeah, point
14:57 timotimo trying with a separate prefix now
14:57 timotimo i still get that error
14:58 diakopter what error, with what command
14:58 timotimo Starting program: /home/timo/build/MoarVM/nqp/install/bin/moar --libpath=src/vm/moar/stage0 src/vm/moar/stage0/nqp.moarvm --bootstrap --module-path=src/stage1 --setting-path=src/stage1 --setting=NQPCORE --no-regex-lib --target=mbc --output=src/stage1/QAST.moarvm src/stage1/gen/QAST.nqp
14:58 hoelzro yay!
14:58 diakopter paste the configure and make lines
14:58 timotimo Routine declaration requires a signature at line 4107, near "(MAST::Nod"
14:58 hoelzro memory usage is better =)
14:58 skids joined #perl6
14:59 timotimo moar: perl Configure.pl --prefix=nqp/install && make all install
14:59 timotimo nqp: perl ConfigureMoar.pl --prefix=install && make all
14:59 camelia nqp: OUTPUT«Confused at line 2, near "perl Confi"␤current instr.: 'panic' pc 14748 (src/stage2/gen/NQPHLL.pir:5229) (src/stage2/gen/NQPHLL.nqp:279)␤»
14:59 timotimo hehe
14:59 diakopter what OS is this?
14:59 diakopter and what compiler?
15:00 LlamaRider joined #perl6
15:00 timotimo this is linux mint based on ubuntu, the compiler is a gcc (Ubuntu/Linaro 4.7.3-1ubuntu1) 4.7.3
15:01 diakopter 32-bi or 64-bit
15:01 diakopter can you try fresh checkouts that aren't under each other and aren't installing to each other's directories?
15:02 timotimo 64bit
15:02 [Coke] moritz++ # docbugfixing
15:03 timotimo the checkouts were fresh. i'll make new ones completely separately
15:05 diakopter also, is there an nqp in your path?
15:05 diakopter (shouldn't be)
15:05 timotimo there is a parrot nqp in my path
15:06 timotimo making the prefix completely separate seems to have helped
15:08 timotimo i get a working nqp from that
15:08 diakopter yay
15:08 timotimo that's an annoying way to break :(
15:08 hoelzro anyone have a ConfigureMoar.pl for Rakudo?
15:09 diakopter jnthn has a moar-support branch
15:10 hoelzro ah, I should've thought to look at branches, thanks!
15:13 hoelzro is building Rakudo with Moar supposed to segfault?
15:13 hoelzro I mean, is that common?
15:14 jnthn hoelzro: We've only had a ConfigureMoar for like, 48 hours, and it's only got further than the 2nd file for me in the last 3 hours :)
15:14 jnthn hoelzro: So we've no idea what's common yet.
15:14 hoelzro ok, just checking
15:14 jnthn You're probably the second person after me to try this...
15:14 jnthn Where does it segv?
15:14 hoelzro /home/rob/.mokudo/bin/nqp --target=mbc --output=blib/Perl6/Actions.moarvm --encoding=utf8 \ src/Perl6/Actions.nqp
15:14 jnthn I don't get a segfault, but I do get bad output.
15:14 jnthn Ah, that file mis-compiles for me (resulting bytecode is wrong)
15:15 hoelzro I'm trying to see if I can figure out what needs to be done
15:15 jnthn Can you get a backtrace?
15:15 hoelzro lemme try
15:15 jnthn thanks
15:18 hoelzro https://gist.github.com/hoelzro/6963437
15:19 jnthn wow, that's a new one...
15:19 moritz (for anybody who likes table tennis: http://www.laola1.tv/de-de/live/ws-final-li-swe-shan-ger/138654.html live stream of European finals)
15:20 moritz Sweden vs. Germany
15:23 [Coke] is Mu:U sufficient to say "object represents a type" ? is "Mu:T" better, or does literally mean Mu's Type?
15:24 moritz :T is actually better
15:24 moritz Int:U       Allow only undefined (abstract or failure) Int values
15:24 moritz Int:T       Allow Int only as a type object
15:24 moritz is what S12 says
15:24 moritz so, I was wrong.
15:28 [Coke] Ok. so I'll push an update to the doc file that replaces Any with Mu, and replaces Type with Mu:T
15:29 tadzik I'm amused how she can throw a ball so high, and yet perfectly upwards
15:30 * masak loves the fact that the hanzi for "ping pong" are 乒乓
15:30 tadzik :
15:30 tadzik :D
15:31 arnsholt =D
15:31 arnsholt What's the reading of those characters?
15:31 masak apparently, Chinese borrowed the term from English, not the other way around.
15:31 masak arnsholt: I'm on it.
15:31 masak they must be relatively new, at least in this meaning.
15:32 masak arnsholt: or, you mean the pronunciation? it's PING1 PANG1.
15:33 masak I'm getting nothing useful about either the character etymology or the constituent parts.
15:34 masak which increases the prior that they're constructed completely custom for the sport.
15:35 arnsholt Indeed
15:35 arnsholt Neat that they made up characters to mirror the sound!
15:36 masak yeah, that's what fascinates me about them. to me, those characters really *look* like "ping pang" :)
15:37 rindolf joined #perl6
15:52 JimmyZ That's onomatopoeia in chinese
15:54 JimmyZ It's most about hieroglyphic or onomatopoeia
15:58 masak if by "it" you mean hanzi, allow me to respectfully disagree. most hanzi are radical+sound pairs.
16:01 hummeleB1 joined #perl6
16:03 Rotwang joined #perl6
16:04 dalek nqp: 391db37 | coke++ | t/docs/opcodes.t:
16:04 dalek nqp: Refactor.
16:04 dalek nqp:
16:04 dalek nqp: Pull code into subs, use keys instead of top level vars.
16:04 dalek nqp: Improve error reporting so that undocumented opcodes appear only
16:04 dalek nqp: once as errors no matter how many vms they are undocumented for.
16:04 dalek nqp: review: https://github.com/perl6/nqp/commit/391db37bdd
16:08 [Coke] https://gist.github.com/coke/6799633 is updated - now more accurately reporting 217 failures in opcode docs. (each opcode can only be undoc'd once, no matter how many vms you implement it on)
16:09 [Coke] it also gives a much better view of which opcodes are currently vm specific.
16:10 [Coke] moritz: do you agree that Mu is a better option for "any type" in the opcode docs than Any?
16:11 masak [Coke]: depends if you're a local variable or a parameter, I think.
16:11 iSlug joined #perl6
16:12 masak r: my $x; say $x.^name
16:12 camelia rakudo e22ac2: OUTPUT«Any␤»
16:13 masak r: sub foo($x?) { say $x.^name }; foo
16:13 camelia rakudo e22ac2: OUTPUT«Any␤»
16:13 masak hm, guess not.
16:13 masak r: class A { has $.x }; say A.x.^name
16:13 camelia rakudo e22ac2: OUTPUT«Cannot look up attributes in a type object␤  in method x at src/gen/CORE.setting:2657␤  in block  at /tmp/LFFbD5kPAd:1␤␤»
16:13 masak r: class A { has $.x }; say A.new.x.^name
16:13 camelia rakudo e22ac2: OUTPUT«Any␤»
16:14 masak r: my %h; say %h<x>.^name
16:14 camelia rakudo e22ac2: OUTPUT«Any␤»
16:15 daniel-s joined #perl6
16:15 masak [Coke]: looks like everything I can think of defaults to Any.
16:17 [Coke] masak: that's r, not nqp.
16:17 [Coke] sorry, this was in the context of nqp's docs/op.markdown
16:17 dalek nqp: b0e8901 | coke++ | docs/ops.markdown:
16:17 dalek nqp: Improve type declarations to follow perl6 rules.
16:17 dalek nqp:
16:17 dalek nqp: Any -> Mu
16:17 dalek nqp: Type (not even a real thing) -> Mu:t
16:17 dalek nqp:
16:17 dalek nqp: Minor rename of some argnames for consistency.
16:18 dalek nqp:
16:18 dalek nqp: moritz++
16:18 dalek nqp: review: https://github.com/perl6/nqp/commit/b0e89010c7
16:18 [Coke] And I just pulled the trigger anyway. <sheepish grin>
16:19 masak [Coke]: oh, ok.
16:19 [Coke] ^^ says :t , but is really :T
16:20 [Coke] Ok. only 176 opcodes left to document. ;)
16:23 ivanshmakov joined #perl6
16:25 xinming joined #perl6
16:26 benabik joined #perl6
16:28 btyler joined #perl6
16:37 TimToady mokudo would be the way of wood/tree, or the way of the eye, or the way of silence
16:38 TimToady not sure any of those are quite right...
16:43 TimToady well MO 模 means "pattern", MO 茂 means "grow thickly", and there are a bunch of KU possibilities
16:46 * masak thinks of https://en.wikipedia.org/wiki/The_Mikado
16:48 TimToady 九 nine; 口 mouth, opening; 工 work; 宮 palace, shrine, prince; 苦bitter; 区 ward, section; 功 merit, service; 久 long time, lasting; 句 phrase, clause; 供offer, attendant; 紅red, crimson, rouge; 駆 gallop, spur on; 貢 tribute
16:53 TimToady but only 区 ward, section; 句 phrase, clause; and 駆 gallop, spur on are unambiguously KU; the rest can be KYUU or KOU or such, so might not read as well to a native speaker
16:53 masak "way of the wood gallop" :P
16:53 masak sounds uncomfortable.
16:53 TimToady no, wood is already 'moku'
16:53 masak oh.
16:53 TimToady was just thinking of KUs that coulld go with MOs
16:54 TimToady pattern ward, or grow-thickly phrase   :)
16:54 masak in Mandarin, KU1 哭 means "cry".
16:54 TimToady either of those is kinda okay
16:54 TimToady pattern gallop is nice
16:54 masak I like "pattern ward" because I don't really know what that would be.
16:55 TimToady it's where the padded walls have designs on them
16:55 cognominal joined #perl6
16:55 masak :P
16:56 masak I typed "mokudou" into Google Translate's Chinese translit, and it gave me 魔窟都 -- "pandemonium all".
16:56 masak I doubt that's very grammatical, though.
16:57 TimToady but most suitable :)
16:57 masak 魔窟 translates to "den", "cave" or (apparently) "pandemonium".
16:57 ssutch joined #perl6
16:58 TimToady 哭 can be used in Japanese, but apparently is pronounced KOKU
16:59 masak looks like "two mouths like a dog" :)
16:59 TimToady a dog howling with the power of two mouths, more likely :)
16:59 masak right.
16:59 masak "bawl"
17:00 TimToady anyway, we don't have a single gloss for anything in English, so no reason it has to be only one in Japanese or Chinese either :)
17:01 * TimToady likes the -kudo suffix in a punny way
17:02 TimToady we have lots of kudos :)
17:02 masak haha
17:02 TimToady ra ra ra
17:02 masak someone should build a VM called Kudo, so that we can talk about porting Rakudo to it and calling it Kudokudo.
17:03 TimToady Three cheers for the Sun God / He sure is a fun god / Ra! Ra! Ra!
17:04 masak hu-ra, hu-ra, hu-ra!
17:05 TimToady there's no kudo.com, but there is a kudo.com
17:05 TimToady er, org
17:06 TimToady :2nd
17:20 TimToady nr: my &foo = -> $x? { say $x.^name }; foo
17:20 camelia niecza v24-98-g473bd20: OUTPUT«Unhandled exception: Nominal type check failed in binding '' in 'ClassHOW.name'; got Mu, needed Any␤  at /home/p6eval/niecza/lib/CORE.setting line 0 (ClassHOW.name @ 1) ␤  at /tmp/wqqvQ0QvgD line 1 (ANON @ 4) ␤  at /tmp/wqqvQ0QvgD line 1 (mainlin…»
17:20 camelia ..rakudo e22ac2: OUTPUT«Mu␤»
17:21 TimToady rakudo has that one right
17:22 masak oh, that's where Mu comes in.
17:23 TimToady S02:1363
17:23 synopsebot Link: http://perlcabal.org/syn/S02.html#line_1363
17:30 xenoterracide__ joined #perl6
17:36 masak TimToady: yes, but what's the reason it ended up taking Mu and not Any? I'm guessing it has something to do with autothreading, but I can't think what.
17:41 hummeleB1 left #perl6
17:50 xinming joined #perl6
17:54 BenGoldberg joined #perl6
17:59 isBEKaml joined #perl6
18:03 pmurias joined #perl6
18:03 sqirrel joined #perl6
18:30 btyler joined #perl6
18:34 cognominal joined #perl6
18:49 iSlug joined #perl6
19:03 btyler joined #perl6
19:03 LlamaRider Not the most helpful error message out there:
19:03 fgomez joined #perl6
19:03 LlamaRider You cannot create an instance of this type ( in method bless at src/gen/CORE.setting:797 )
19:03 cognominal joined #perl6
19:04 LlamaRider not too sure if this regressed from my working Parrot backend (now trying out the JVM stuff)
19:04 FROGGS hmmm
19:04 FROGGS maybe you try to instanciate a role?
19:05 FROGGS err, wait, this should work... you wasn't able to augment it
19:07 LlamaRider hm, there is something fishy. Instantiating an object cleanly just throws a warning:
19:07 LlamaRider perl6 -Ilib -e ' use XML::LibXML; my $parser = XML::LibXML.new(); '
19:07 LlamaRider Passing an object candidate to Mu.bless is deprecated
19:11 LlamaRider ah!
19:11 LlamaRider seriously... compare the above to the next two broken examples:
19:12 LlamaRider perl6 -Ilib -e ' my $parser = XML::LibXML.new(); '
19:12 LlamaRider Could not find symbol '&LibXML'
19:12 LlamaRider ...
19:12 LlamaRider This error makes perfect sense.
19:12 LlamaRider but then if I do:
19:12 LlamaRider perl6 -Ilib -e 'use XML::LibXML::Common; my $parser = XML::LibXML.new(); '
19:12 LlamaRider You cannot create an instance of this type
19:13 timotimo does it autovivify XML::LibXML or something like that?
19:13 LlamaRider it's almot empty
19:13 LlamaRider it has two empty methods and no mention of any other classes
19:13 LlamaRider almost*
19:13 timotimo try XML::LibXML::.keys
19:14 LlamaRider 'use XML::LibXML::Common; XML::LibXML::.keys.say '
19:14 LlamaRider prints only a single key 'Common'
19:15 LlamaRider ok, then it does create that stash
19:16 LlamaRider I am still unhappy with the error message (though I am happy I figured out what was going on :-))
19:18 LlamaRider btw, does anyone know if zavolaj/NativeCall can now hook into values from XS?
19:18 LlamaRider that was poorly phrased
19:18 LlamaRider I have a commented out TODO that looks like:
19:18 LlamaRider # class LIBXML_RUNTIME_VERSION is OpaquePointer is native('libxml2') is symbol('xmlParserVersion');
19:19 LlamaRider I remember stopping the LibXML port because of it some months back
19:27 spider-mario joined #perl6
19:27 FROGGS timotimo: if you declare a package Foo::Bar::Baz, it must (sort of) stub Foo and Foo::Bar too
19:27 FROGGS s/too//
19:29 LlamaRider Hm, NativeCall doesn't bill under the JVM backend. I suppose that was to be expected.
19:29 LlamaRider build*
19:29 moritz LlamaRider: just today jnthn did some nqü and NativeCall hacking that should make it build
19:30 moritz LlamaRider: but you actually a very fresh copy of rakudo + nqp
19:30 LlamaRider will update and try again, thanks!
19:31 moritz LlamaRider: though not all features work in the JVM version yet
19:50 iSlug joined #perl6
19:58 lue hello world! o/
19:59 itz_ joined #perl6
20:07 LlamaRider joined #perl6
20:13 arnsholt NativeCall HEAD should work with NQP and Rakudo HEAD
20:13 dmol joined #perl6
20:13 arnsholt The test suite won't run however, and it looks like I'll have to fiddle a bit with the nativecall bits in NQP to get it working properly
20:14 arnsholt NativeCall on JVM at this point is only useful for those who want to hack on the nativecall internals in NQP/JVM
20:19 bbkr__ joined #perl6
20:35 bruges joined #perl6
20:43 xinming joined #perl6
20:54 ajr joined #perl6
21:00 telex joined #perl6
21:34 xinming joined #perl6
22:19 dalek nqp: 6fa6d36 | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
22:19 dalek nqp: Map named lexical lookup ops.
22:19 dalek nqp: review: https://github.com/perl6/nqp/commit/6fa6d36596
22:19 dalek nqp: 246aaae | jnthn++ | src/vm/moar/QAST/QASTCompilerMAST.nqp:
22:19 dalek nqp: Remove fossil.
22:19 dalek nqp: review: https://github.com/perl6/nqp/commit/246aaaebda
22:30 grondilu Off topic:  after the GPU, here come the NPU (Neural Processing Unit)  http://www.33rdsquare.com/2013/10/brain-inspired-computing-in-qualcomm.html
22:45 lue .oO(The robot apocalypse has finally begun. Slowly, but surely.)
22:56 benabik joined #perl6
23:15 xinming joined #perl6
23:18 ajr joined #perl6
23:23 dalek perl6-roast-data: f912eeb | coke++ | / (3 files):
23:23 dalek perl6-roast-data: today (automated commit)
23:23 dalek perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/f912eebe1e
23:50 dansamo joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo