»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:00 dayangkun joined 00:02 tokuhiro_ left, stocks_ joined 00:03 spider-mario left 00:05 MayDaniel left, stocks left 00:12 Bulgaria is now known as Gliner 00:18 mr007x left 00:21 sudokode left 00:24 gdey left 00:27 stocks_ left 00:34 sudokode joined 00:35 VXZ left 00:44 sudokode left 00:47 bowtie left 00:48 dayangkun left 00:57 anuby joined 01:00 hypolin joined 01:01 dayangkun joined
rjbs > my $m = sub ($x) { ... } 01:05
> Sub.HOW.methods($m)
Method 'gist' not found for invocant of class 'Sub'
What did I need to do instead?
01:16 leont left 01:17 plobsing joined 01:31 HarryS left, fgomez left, HarryS joined 01:36 HarryS left 01:38 fgomez joined 01:39 dayangkun left 01:44 PacoAir left, jaldhar_ left 01:58 jaldhar_ joined, havenn joined
[Coke] what are you trying to do? 02:05
find out what methods you can call on $m ?
geekosaur $m.^methods fails the same way for me, fwiw. I think it's tripping over getting a name from a method?
02:24 orafu left 02:31 HarryS joined 02:34 pmurias left 02:36 HarryS left
havenn Any thoughts on this? Wrong? Incomprehensible?: gist.github.com/3991238 02:37
02:37 dayangkun joined
havenn Oops, wrong channel - sorry! 02:40
rking havenn: I was really puzzled. ☺ 02:41
02:47 sudokode joined 02:50 bloonix left 02:51 avar left, bloonix joined, avar joined, avar left, avar joined 02:52 orafu joined 02:56 havenn left
rjbs [Coke]: Yes. 02:58
moritz r: (sub ($x) { }).^methods 03:03
p6eval rakudo 34febe: ( no output )
moritz r: say (sub ($x) { }).^methods
p6eval rakudo 34febe: OUTPUT«No such method 'gist' for invocant of type 'Sub'␤ in method gist at src/gen/CORE.setting:4875␤ in sub say at src/gen/CORE.setting:7304␤ in block at /tmp/DgavXwbshh:1␤␤»
moritz r: say (sub ($x) { }).can('gist')
p6eval rakudo 34febe: OUTPUT«gist␤»
moritz probably a parrot Sub that is the problem 03:04
geekosaur some of the error sI was getting suggested, yes, that
that I was getting a non-p6 Sub in there
moritz r: say (sub ($x) { }).^methods.join(' ') 03:05
p6eval rakudo 34febe: OUTPUT«No such method 'Stringy' for invocant of type 'Sub'␤ in method join at src/gen/CORE.setting:1223␤ in block at /tmp/mZpuHq8BzN:1␤␤»
moritz r: say (sub ($x) { }).^methods.map(*.name).join(' ')
p6eval rakudo 34febe: OUTPUT«Unmarshallable foreign language value passed for parameter '$x0'␤ in block at /tmp/njlb8bMQua:1␤␤»
geekosaur ^^ thgat was one of the things I kept getting
japhb r: Sub.^methods.keys.say 03:06
p6eval rakudo 34febe: OUTPUT«0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41␤»
japhb Well, at least we know there are 42 of them. :-)
03:08 gootle joined 03:14 orafu left, sudokode left 03:15 orafu joined 03:17 stocks joined 03:20 gugod left 03:28 gugod joined 03:30 benabik joined 03:31 HarryS joined
rjbs japhb: hehe 03:51
n: say (sub ($x) { }).^methods.join(' ')
p6eval niecza v22-15-gc600005: OUTPUT«Potential difficulties:␤ $x is declared but not used at /tmp/mWq0zzYWLJ line 1:␤------> say (sub (⏏$x) { }).^methods.join(' ')␤␤Unhandled exception: Unable to resolve method methods in type ClassHOW␤ at /tmp/mWq0zzYWLJ line 1 (mainline…
04:00 jaldhar_ left 04:01 jaldhar_ joined 04:08 dayangkun left 04:11 erkan left 04:12 colomon left
diakopter sweet; error I've never seen before 04:13
r: Sub.^methods.values>>.name.say
p6eval rakudo 34febe: OUTPUT«Unmarshallable foreign language value passed for parameter 'obj'␤ in method dispatch:<hyper> at src/gen/CORE.setting:984␤ in block at /tmp/WiIs0F72N5:1␤␤»
diakopter phenny: tell masak wat. I'm confused, so LTA, imho r: Sub.^methods.values>>.name.say 04:14
phenny diakopter: I'll pass that on when masak is around.
diakopter phenny: tell masak I don't know. just.. I don't know. 04:15
phenny diakopter: I'll pass that on when masak is around.
diakopter googled the error and found several other occurrences 04:16
04:18 azawawi joined
azawawi hi 04:18
diakopter hi 04:23
04:30 woosley joined 04:42 CharellKing joined 04:47 CharellKing left 04:49 dayangkun joined 04:56 kurahaupo left 04:59 azawawi left 05:01 kurahaupo joined 05:02 hypolin left, grondilu left, hypolin joined 05:09 plobsing left, Guest0863 joined 05:10 Guest0863 left 05:22 kurahaupo left 05:33 fgomez left 05:39 kurahaupo joined, fgomez joined 05:42 Pleiades` left 05:43 arlinius left 05:47 Pleiades` joined 05:50 SamuraiJack joined 05:53 birdwindupbird joined 05:57 erkan joined, erkan left, erkan joined 06:02 wk left 06:10 kurahaupo left 06:23 kurahaupo joined 06:35 stocks left 06:38 stocks joined, stocks left
diakopter jnthn: short backlog is 07:19
nwc10 jnthn: yes, that change in nqp now means that I can build origin/Q without running out of swap. 07:21
07:29 colomon joined
diakopter r: say Sub.^attributes>>.name.say 07:30
p6eval rakudo 34febe: OUTPUT«$!dispatchees $!dispatcher_cache $!dispatcher $!md_thunk $!rw $!inline_info $!yada $!package $!onlystar $!state_vars $!phasers $!do $!signature␤True␤»
diakopter r: say Str.^attributes[0].WHAT 07:33
p6eval rakudo 34febe: OUTPUT«No such method 'gist' for invocant of type 'BOOTSTRAPATTR'␤ in sub say at src/gen/CORE.setting:7304␤ in block at /tmp/hxCCshKllG:1␤␤»
07:33 flightrecorder joined 07:37 wk joined 07:39 colomon left 07:41 colomon joined 07:54 kurahaupo left
moritz \o 07:56
sorear o/
08:10 hoelzro|away is now known as hoelzro 08:12 brrt joined
jnthn nwc10: Ah, good. 08:37
nwc10: Without it, it would try to swallow over 5GB :)
09:01 flightrecorder left
nwc10 that would be less than awesome. 09:22
jnthn It apparently was for those without 16GB of RAM around to hide what was going on, yes. ;-) 09:30
09:31 am0c_ joined 09:34 domidumont joined 09:35 am0c_ is now known as am0c 09:42 GlitchMr joined 09:53 domidumont left 09:54 domidumont joined 10:08 anuby left 10:11 spider-mario joined 10:20 gfldex left 10:21 gfldex joined 10:24 cognominal joined 10:26 flightrecorder joined 10:29 FROGGS joined 10:36 hypolin left 10:47 leont joined 10:48 gootle left 10:50 am0c left 10:54 cognominal left 10:59 cognominal joined 11:02 nyuszika7h left 11:10 flightrecorder left, flightrecorder joined 11:12 nyuszika7h joined 11:24 MayDaniel joined 11:25 rindolf joined 11:28 Timbus left, PacoAir joined 11:29 Timbus joined 11:30 mr007x joined 11:39 brrt left 11:41 cognominal left 11:46 cognominal joined 11:47 sivoais left 11:51 cognominal left 11:56 cognominal joined 12:00 spider-mario left 12:01 spider-mario joined 12:04 sivoais joined 12:09 brrt joined 12:16 am0c joined 12:21 am0c left 12:26 grondilu joined 12:27 SamuraiJack_ joined
grondilu guys, check out problem 19 of the project Euler: projecteuler.net/problem=19 I've found a very, very quick solution :) 12:27
12:28 SamuraiJack left 12:29 Lola_91 joined
Lola_91 heloo 12:31
grondilu hello
12:32 SamuraiJack_ left
shachaf for y in `seq 1901 2000`; do for m in `seq 1 12`; do cal $y -m $m | head -n 3 | tail -n 1; done; done | grep -c '^ 1' # this totally counts, right? 12:32
grondilu shachaf: wrong channel
12:32 am0c joined
jnthn o/ Lola_91 12:33
Lola_91 jnthn: finally someone I know :(
jnthn ;) 12:34
Lola_91 :D
jnthn Hope you're doing fine?
Lola_91 kinda :P i am working on a new newsletter with varius news and such ^__^
grondilu shachaf: oh you were talking about problem 19, my bad. 12:35
jnthn :)
Lola_91 jnthn: i need your help in something, i have an email list named unsub.txt, and a list called Haifa.txt I need to remove all those in unsub.txt from Haifa.txt. Is there some one liner I can save to do it in perl? 12:36
rindolf shachaf: hi. 12:37
12:37 SamuraiJack joined
grondilu shachaf: neat idea. I rewrote it though: 12:38
for y in {1901..2000}; do for m in {1..12}; do cal $y -m $m |grep '^ 1'; done; done | wc -l
jnthn cat haifa.txt | perl6 -ne ".say if $_ ne BEGIN { any(lines('unsub.txt'.IO)) }" # or so :) 12:39
rindolf grondilu: {1901..2000} is Bash-specific I think.
jnthn: haifa.txt ?
grondilu rindolf: it is indeed.
shachaf hi rindolf 12:40
rindolf shachaf: what's up?
jnthn rindolf: Answering a question like 6 lines back :)
grondilu r: say 100 * 12 div 7; # this was my solution :)
p6eval rakudo 34febe: OUTPUT«171␤»
rindolf jnthn: ah.
Lola_91 jnthn: perl6 wiz :P someday I will free myself and become a developer 8| 12:41
jnthn :) 12:42
ooh, there will be a Saint Perl this year :)
Lola_91 jnthn: i can not afford a plane ticket -_-
jnthn Lola_91: Aww. Yeah, not sure if I'll go, it was just announced though :) 12:43
jnthn is still a Perl conf addict :)
Would need to get visa to go to this one though :/ 12:44
Lola_91 jnthn: hmmmm.. i think i would look like a stranger there :s i barely remember any perl now :(
jnthn Languages. They slip away without use... Both human and computer ones. 12:46
Tend to come back fairly quick as well, though :) 12:47
jnthn figures he'd best have some lunch...
12:47 snearch joined
grondilu does open('file').lines return a lazy list? 12:47
jnthn Pretty sure lines is lazy 12:48
grondilu: Yes, it does gather/take
masak g'day, #perl6 12:49
phenny masak: 04:14Z <diakopter> tell masak wat. I'm confused, so LTA, imho r: Sub.^methods.values>>.name.say
masak: 04:15Z <diakopter> tell masak I don't know. just.. I don't know.
masak seems I should backlog :) 12:50
r: Sub.^methods.values>>.name.say
p6eval rakudo 34febe: OUTPUT«Unmarshallable foreign language value passed for parameter 'obj'␤ in method dispatch:<hyper> at src/gen/CORE.setting:984␤ in block at /tmp/WRBqqrZSrB:1␤␤»
jnthn Just people rediscovering unresolved RT issues... :)
masak :)
.oO( There's no new bugs under the sun... )
arnsholt Heh. Taking the train from Copenhagen to Malmö takes only slightly longer than the train from Oslo to Oslo airport =D
[Coke] gets a package from Hong Kong from sender "Lion" 12:51
jnthn arnsholt: And costs, like, 3-4 times less! 12:52
arnsholt I don't wanna talk about it >.<
shachaf is falling asleep. 12:53
arnsholt I can get the University to cover most of my travelling costs these days though, which is nice =)
[Coke] except for the part where I have to do work when I get there, having work cover travel expenses is nice,ja. 12:54
12:56 leont left
arnsholt Troo, troo 12:57
In my case it's mostly conferences and stuff, so it's slightly different from my normal work
12:58 Pleiades` left 13:04 Psyche^ joined
rindolf shachaf: why are you sleepy? 13:05
shachaf Because it's ~06:00.
13:08 Patterner left, Psyche^ is now known as Patterner, tokuhiro_ joined
[Coke] aha! now we know which 24th of the world you live in! muahahaha. 13:12
[Coke] sees Gábor just admitted to knowing tcl! muahaha. 13:15
13:16 Pleiades` joined 13:22 plobsing joined 13:33 arlinius joined 13:34 mr007x left, sudokode joined 13:49 bowtie joined 13:52 plobsing left 13:55 sizz left 13:56 sizz joined 13:59 Lola_91 left 14:01 dayangkun left, stopbit joined 14:05 gdey joined 14:06 tokuhiro_ left 14:09 FROGGS left, flightrecorder left 14:10 flightrecorder joined 14:13 am0c_ joined, am0c left 14:22 FROGGS joined 14:28 kiyan joined 14:33 snearch left 14:34 am0c_ is now known as am0c 14:45 am0c_ joined 14:46 am0c left 14:47 gdey left 14:49 am0c_ is now known as am0c 14:50 wk left 14:51 skids left 15:07 thou joined 15:09 mucker joined 15:15 birdwindupbird left
japhb You're doing a lot of evil laughing this morning, [Coke] ... now we know you're really a supervillain! 15:18
[Coke] mua..... ha... ha? 15:23
nothing to see here.
brrt [Coke]: ever figured out the nqp bug thing?
japhb imagines [Coke] in a MegaMind style outfit ... but what would his minion be like? 15:26
dalek p/Qcompat: 4a1f1e2 | jnthn++ | src/ (2 files):
Bring regex nibbler closer to STD.

Means we'll be able to support user-picked stoppers in Rakudo. Should also deal with some over-eager backtracking bugs leading to bad errors. No NQP test regressions with this, but may well bust non-Q Rakudo, so in a branch for now.
15:28 sisar joined 15:29 am0c_ joined
dalek kudo/Q: 51666f9 | jnthn++ | src/Perl6/ (2 files):
Parse single quotes as we do doubles in regexes.

Should make sure we parse them as Perl 6 would.
kudo/Q: 21eb9c7 | jnthn++ | src/Perl6/Grammar.pm:
Generalize nibbler actions handling.

This will mean we can nibble regexes also.
kudo/Q: afa0a2b | jnthn++ | src/Perl6/ (2 files):
Switch /.../ over to using nibble.

Just one test regression, which is because < a aaa > in regexes is not updated for new quote handling yet.
15:31 am0c left, am0c_ is now known as am0c 15:32 benabik left 15:33 benabik joined
[Coke] brrt: no I just reported it and then stuck with a working version. 15:35
brrt its probably something of a parrot bug 15:37
but i'm having trouble finding where
jnthn brrt: github.com/parrot/parrot/blob/mast...t.pmc#L985 15:38
brrt: I suspect C NULLs are escaping from there
brrt that would make sense 15:39
brrt should figure out the shortest one-liner that triggers it 15:40
15:41 cognominal left 15:44 sizz left
brrt nqp: ~nqp::ctx() 15:45
p6eval nqp: ( no output )
brrt nqp: say ~nqp::ctx()
p6eval nqp: ( no output )
jnthn brrt: I think it was filed in an NQP ticket 15:46
brrt nqp: say ~ nqp::callerctx(nqp::ctx))
p6eval nqp: OUTPUT«Confused at line 2, near "say ~ nqp:"␤current instr.: 'panic' pc 13089 (src/stage2/gen/NQPHLL.pir:4619) (src/stage2/gen/NQPHLL.pm:324)␤»
brrt oh, yes, masak++ has posted it
15:47 cognominal joined, xinming joined
jnthn nqp: my $x := nqp::ctx(); while $x { $x := nqp::ctxcaller($x) } 15:47
p6eval nqp: OUTPUT«get_bool() not implemented in class 'CallContext'␤current instr.: '' pc 59 ((file unknown):48) (/tmp/8oL_CszLJz:1)␤»
jnthn nqp: my $x := nqp::ctx(); while nqp::defined($x) { $x := nqp::ctxcaller($x) }
p6eval nqp: ( no output )
jnthn hm, well, something like that :)
brrt try stringification
jnthn nqp: my $x := nqp::ctx(); while nqp::defined($x) { $x := nqp::ctxcaller($x); ~$x }
p6eval nqp: OUTPUT«(signal SEGV)»
jnthn nqp: my $x := nqp::ctx(); while nqp::defined($x) { $x := nqp::ctxcaller($x); say(pir::typeof__SP($x)); } 15:48
p6eval nqp: OUTPUT«(signal SEGV)CallContext␤CallContext␤CallContext␤CallContext␤CallContext␤CallContext␤CallContext␤CallContext␤CallContext␤»
jnthn Or that
brrt basically
jnthn I guess it hits the NULL and BOOM
brrt it returns a null
which you stringify
which dies somewhere in parrot_pcc_*
15:48 twinshadow joined
brrt so 15:48
hmm 15:49
we can put null pointer checks everywhere
callcontext is a Really Hot Path
jnthn getattribute isn't
brrt fair enough
jnthn Or at least, it'd best not be :)
15:50 xinming_ left 15:51 PacoAir left 15:53 PacoAir joined 15:54 benabik left
brrt lets check it out 15:58
(i have placed the obvious fix in that method) 16:00
well, nqp builds happily now 16:04
and… it exits with a null pmc access 16:05
jnthn++ for the right pointer 16:06
16:06 bowtie left
[Coke] Is this something where parrot is handling us a null or we're making one? 16:06
brrt handing 16:07
apparantly, parrot doesn't initialize with pmcnull
but with a real null
jnthn brrt: yay 16:09
parrot-- for the wrong pointer :P
brrt poor parrot
jnthn :)
brrt whats simpler, pull-requests or posting the diff?
jnthn Well, not initializing with PMCNULL is fine...it's cheaper
Pull request then mention on #parrot I guess 16:10
brrt yes, but that means that you Must Check Null everywhere
jnthn I meant, C-land things
The trick is ensuring the real NULLs don't leak :)
brrt yes, and that proves to be tricky indeed :-) 16:11
jnthn Though if you're the CLR, you use real NULLs and install a segfault handler...
brrt which brings us to the question
why don't we have that in parrot?
hoelzro hehe, segfault handler >:)
brrt my goes it would kill use in an embedding context
japhb hoelzro: Those are actually really useful for a lot of important algorithms.
hoelzro oh, I believe it 16:12
I used one a while back
although all it did was trigger a crash reporter
jnthn brrt: Well, they mostly compared about run on Windows I think. And it's not the only way to implement it.
brrt: Just fairly efficient.
dalek kudo/Q: a466b61 | jnthn++ | src/Perl6/ (2 files):
Switch quote words in regex to new quote parser.
kudo/Q: bf0e21f | jnthn++ | src/Perl6/Grammar.pm:
Remove unused method.
jnthn And yes, you need an OS with the hook :)
hoelzro heh
[Coke] brrt: where in get_attr_str is the NULL coming from? 16:14
brrt [Coke]: my guess is the caller_ctx attribute
the cheapest and universal fix is to replace return value with if (value) return value; else return PMCNULL 16:15
which is with i did, plus some formatting
16:15 benabik joined
[Coke] a better fix is probably to actually initialize all the PMC attributes to PMCNULL. (or fix GET_ATTR to DTRT.) 16:16
brrt DTRT?
japhb Do The Right Thing
brrt i guess
[Coke] but that's definitely more work. 16:18
brrt fixing GET_ATTR_ isn't necessarily more work, its defined somewhere
[Coke] once you build parrot, it's in include/pmc/pmc_callcontext.h
16:19 bowtie joined
[Coke] I'm not sure we know at that point if it's a PMC or not, though. 16:19
You might know when that macro is generated.
rjbs Looking back on my "how do you find the methods on a sub," I see chatter from overnight: 16:22
> Sub.^methods>>.name.say
but with ^attributes, works
Unmarshallable foreign language value passed for parameter 'obj'
tadzik Interview with moritz++ www.i-programmer.info/professional-...-lenz.html 16:24
brrt [Coke]: thats fair
16:26 hoelzro is now known as hoelzro|away 16:30 brrt left 16:32 FROGGS left 16:33 flightrecorder left 16:35 am0c_ joined 16:36 am0c left
[Coke] does that pull request fix the nqp build, or just make it possible to fix the nqp build? 16:36
brrt++ 16:37
jnthn [Coke]: < brrt> well, nqp builds happily now
brrt++ indeed
jnthn feared a GC issue
16:38 havenn joined
jnthn r: my $x = 'abcd'; $x ~~ s[[ \w ]] = 'x'; say $x; 16:38
p6eval rakudo 34febe: OUTPUT«xbcd␤»
jnthn r: my $x = 'abcd'; $x ~~ s{ \w } = 'x'; say $x; 16:40
p6eval rakudo 34febe: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&s' called (line 1)␤Undefined routine '&w' called (line 1)␤»
16:40 havenn left, havenn joined, Khisanth left
jnthn r: my $x = 'abcd'; $x ~~ s!\w!x!; say $x; 16:41
p6eval rakudo 34febe: OUTPUT«===SORRY!===␤Confused␤at /tmp/MoEi2tcHDR:1␤»
jnthn moritz++ # interview 16:43
16:45 havenn left 16:47 am0c_ is now known as am0c 16:50 kaare_ joined 17:01 havenn joined 17:04 fgomez left 17:06 havenn left, benabik left 17:08 benabik joined 17:11 cognominal left 17:13 am0c_ joined 17:14 am0c left
flussence g'morning, GMT-9 17:16
17:16 havenn joined
jnthn o/ flussence 17:17
17:20 cognominal joined 17:23 topo` joined, jaldhar_ left 17:25 benabik left
diakopter o/ 17:28
17:29 am0c joined 17:30 am0c left, am0c_ left
dalek p/Qcompat: 77c0c7d | jnthn++ | src/QRegex/P5Regex/Grammar.nqp:
Update P5Regex to have user-chosen stopper hook.

Also means that a workaround can be eliminated.
jnthn o/ diakopter 17:32
dalek kudo/Q: 0ad1f12 | jnthn++ | src/Perl6/ (2 files):
Refactor substitutions to be STD-like.

Eliminates the last use of quote_EXPR, and enables use of user-chosen quotes. Seems like some better error reporting also.
kudo/Q: 082194b | jnthn++ | src/Perl6/ (2 files):
Allow user-chosen delimeters for m and rx.
kudo/Q: 7ba48cf | jnthn++ | src/Perl6/ (2 files):
For consistency, nibble in regex_def too.
kudo/Q: 59553bf | jnthn++ | src/Perl6/Grammar.pm:
Remove now-unrequired stopper hack.
kudo/Q: 8078736 | jnthn++ | src/Perl6/ (2 files):
Unbust :P5; supports user-chosen delims too.
17:40 sisar left, rindolf left 17:47 mucker left 17:53 havenn left 17:54 havenn joined 17:58 havenn left 18:00 havenn joined 18:01 rindolf joined 18:06 Chillance joined 18:09 am0c joined, benabik joined 18:12 mr007x joined 18:23 havenn left, havenn joined 18:24 havenn left 18:25 havenn joined 18:28 havenn left 18:29 havenn joined 18:30 jaldhar_ joined 18:31 am0c left 18:33 havenn left 18:37 fgomez joined 18:40 havenn joined 18:42 havenn left, havenn joined 18:43 havenn left, havenn joined 18:52 cognominal left
dalek kudo/Q: 662c219 | jnthn++ | src/Perl6/Grammar.pm:
Implement quote adverbs.
18:57 mr007x left
jnthn If possible, could somebody (pref with ICU) build rakudoQ (needs nqp/Qcompat) and spectest it? 18:59
jnthn thinks it's mergeable 19:00
[Coke] brrt's fix (and knock-on fix from cotto) pushed.
jnthn brrt++, [Coke]++, cotto++ # getting stuff tracked down/fixed 19:02
cotto got anything else while I'm in there?
19:03 havenn left 19:04 havenn joined
[Coke] bother. I cannot git clone a git:// address with my default firewall settings. 19:04
cotto sshuttle is great for that if you can't control the firewall directly 19:05
On some networks, I just forward everything over port 443 to a Linode vm.
jnthn cotto: I see some patches for the IO stuff we ran in to have landed a bit ago, so I guess we should try using latest Parrot with IO changes again.
[Coke] I really shouldn't be doing that.
cotto might be frowned on 19:06
japhb I can build on 32-bit Linux with ICU, but spectest would take a LONG time. Anyone with a faster machine planning to test RakudoQ, or should I give it a try? 19:07
19:07 havenn left 19:08 havenn joined
japhb Actually, jnthn, don't you have a working Debian VM from when you were building * that you could do it on? ICU should be there for sure ... 19:09
jnthn oh, point...
japhb (well, you might have to install libicu-dev, but still)
flussence hm, I can do a build (as soon as I figure out the manual build stuff; --gen-parrot won't work) 19:10
oh that was easy 19:12
jnthn japhb: Yeah, installed that now 19:13
Let's give this a go...
moritz builds NQP for a rakudo/Q spectest run 19:14
jnthn: t/spec/S02-literals/misc-interpolation.rakudo faisl a test 19:24
not ok 25 - curly brace delimiters don't interfere with closure interpolation
# got: 'a{chr 98}c'
# expected: 'abc'
jnthn moritz: oh, shoulda mentioned that one
Yes, the test is wrong.
19:25 Chillance left
moritz that's what I would have checked enxt :-) 19:25
jnthn TimToady++ mentioned that nested delims are meant to take precedence over escapes
19:25 Chillance joined
jnthn gee, building in this VM is a load slower than I'm used to things building on this machine... 19:26
moritz fwiw setting compilation wasn't slower than in nom 19:27
116s for the parse+past stage, which is normal on that machine
jnthn moritz: Yeah, that matches my measurements
(that it's stayed the same)
19:30 domidumont left 19:33 Targen joined 19:34 havenn left, havenn joined
jnthn oh, my VM is configured with 1 CPU. No wonder parallel spectest is sucking... 19:34
19:38 havenn left, sirrobert joined
moritz jnthn: gist.github.com/3995934 19:38
19:38 havenn joined
flussence I got quite a few more than that... 19:38
moritz $ ./perl6 -Ilib t/spec/S12-enums/thorough.rakudo
Method 'returns' not found for invocant of class 'Parcel'
jnthn huh, passes here 19:40
sirrobert hallo
jnthn oh...
There's been unfudge that I hadn't pulled in t/spec
hm, now I get it 19:41
19:42 sjohnson left, mj41 left
jnthn The Pod tests are still too faily for me here to be likely to easily track down the S26 one 19:43
19:43 mj41 joined
flussence uses that handy tap archive feature 19:44
jnthn um, how did this code ever work, I wonder... 19:45
@values.push($*W.compile_time_evaluate($<term>, $<term>.ast));
Then in the loop over @values 19:46
if istype($_.returns(), $Pair) {
That code is always doomed if we get a real value
moritz doesn't see why
19:46 sjohnson joined
jnthn moritz: Because compile_time_evaluate returns a real Perl 6 object, not a QAST node 19:47
moritz oh, right
jnthn r: Mu.returns
p6eval rakudo 34febe: OUTPUT«No such method 'returns' for invocant of type 'Mu'␤ in block at /tmp/fH9gbT2_Dd:1␤␤»
jnthn Something has changed that makes it take that path, though...
ah, yeah 19:48
The structure of the QAST that shell words returns has probably changed a bit
moritz it was probably me who wrote the offending code 19:50
jnthn *nod*
flussence here's what I get from it: flussence.eu/~ant/perl6/etc/rakudoQ...run.tar.gz
jnthn I think I see what I changed that made it unhappy
19:51 benabik left
flussence there's quite a few "todo passed" in there too 19:51
sirrobert ooc, is there a timeline for implementing heredocs in rakudo at the moment? 19:52
moritz sirrobert: the Q branch is a prerequisite for it
jnthn heredocs are one of the main motivations for doing the Q branch :) 19:53
sirrobert hmmm that's a branch dealing with quoting?
19:53 fgomez left
moritz yes 19:53
sirrobert ok, thanks -- will read =)
jnthn flussence: are there...test results amongst that lot? 19:56
flussence that's just what "make rakudo_test_run.tar.gz" gives me, I dunno if there's an easy way to extract the useful bits 19:57
oh wait, yes there is
19:58 SamuraiJack_ joined
flussence runs make spectest_smolder 19:58
19:58 bluescreen10 joined, SamuraiJack left
jnthn Grr, the obvious change doesn't help 19:59
20:01 wk joined
jnthn moritz: Hm. The sockets one passes for me both under Windows and Linux. 20:03
moritz jnthn: which parrot revision? 20:04
jnthn This is Parrot version 4.4.0-devel built for amd64-win32
[Coke] is reminded to restart work on his smolder replacement.
flussence jnthn: this should be more readable - smolder.parrot.org/app/projects/rep...ails/32197 20:06
20:07 Util joined
flussence don't trust the "only show failures" thing, there's a few that show as "100%" but apparently crashed 20:07
e.g. smolder.parrot.org/app/projects/tap...m/32197/14
20:10 havenn left, SamuraiJack_ left, havenn joined 20:12 birdwindupbird joined
jnthn [Coke]: Please do, this thing doesn't even render results properly in firefox :/ 20:13
bah, I'm not sure I trust anything about this result set
20:13 havenn left
flussence the short version: gist.github.com/3996180 20:14
20:14 havenn joined
jnthn flussence: That looks terribly like something ICU related 20:14
20:15 havenn left 20:16 havenn joined
flussence I've seen these particular files have failures going back a year or more, so I'm not convinced rakudo's causing the failures 20:16
jnthn flussence: Ah, so they're probably not specific to the branch? 20:17
sirrobert r: class A { method b (Str $str, &blk) { }; multi method c (@arr) { }; multi method c (List $list) { }; }; say A.^methods[0].perl; say A.^methods[1].perl;
p6eval rakudo 34febe: OUTPUT«method b(A : Str $str, &blk, Mu *%_) { ... }␤method c(: , Mu *%_) { ... }␤»
jnthn They do look like a really odd bunch of regressions...
flussence I definitely recognise the S05 ones being problematic
jnthn (if they were)
sirrobert How do I get a hold of the individual multi-methods?
jnthn OK
The role one also is odd. 20:18
sirrobert (in particular I want the signatures)
jnthn Though the ones that fail are error reporting.
20:18 havenn_ joined 20:19 havenn left
flussence sirrobert: this might help - github.com/flussence/perl6-XMMS2/b...p6xmms2#L7 20:20
sirrobert fluessence: going
jnthn In Pod like 20:21
=begin code :allow<B>
What' the :allow<B> thing called?
flussence config something-something
tadzik just pod configuration, I think
like =begin table :caption<foo>
we have tests for some of those
jnthn Yeah 20:22
I've busted them...
tadzik (:
jnthn 09-configuration.t 20:23
tadzik good that I wrote them then ;)
jnthn token pod_configuration($spaces = '') {
[ [\n $spaces '=']? \h+ <colonpair> ]*
Does that look like what parses it?
tadzik does that what?
.oO( that should probably be <colonpair>* too )
jnthn Is that colonpair what matches the :caption<foo>
tadzik yes
jnthn OK