Camelia, the Perl 6 bug

IRC log for #perl6, 2012-08-31

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs

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

All times shown according to UTC.

Time Nick Message
00:17 whiteknight joined #perl6
00:17 tokuhiro_ joined #perl6
00:25 gongyiliao joined #perl6
00:33 s1n joined #perl6
00:35 cggoebel joined #perl6
00:43 scottdware joined #perl6
00:48 ingyfoo_ joined #perl6
00:57 benabik joined #perl6
00:57 krunen joined #perl6
00:57 sftp joined #perl6
00:57 colomon joined #perl6
00:57 larks joined #perl6
00:57 tipdbmp joined #perl6
00:57 Timbus joined #perl6
01:02 dayangkun joined #perl6
01:14 scott__ joined #perl6
01:31 scottdware left #perl6
01:46 cggoebel joined #perl6
01:54 FROGGS_ joined #perl6
01:55 replore joined #perl6
02:24 thou joined #perl6
02:25 orafu joined #perl6
04:21 xinming joined #perl6
04:22 Khisanth joined #perl6
04:37 erkan joined #perl6
04:37 erkan joined #perl6
04:38 birdwindupbird joined #perl6
04:42 stepnem joined #perl6
04:53 kurahaupo joined #perl6
05:20 dayangkun_ joined #perl6
05:23 benabik joined #perl6
05:35 pmichaud good morning, #perl6
05:36 pmichaud I made it home okay, and I should be able to do the Star release later today (08/31).
05:44 pmichaud afk, sleep
05:46 sorear good morning, #pmicaud
05:46 sorear pmichaud
05:55 moritz \o
05:55 phenny moritz: 30 Aug 23:08Z <japhb> tell moritz Browsing over the KidsRuby GitHub info, it looks like we could build a Perl 6 clone using essentially the same technologies.  Looks more or less straightforward, but some of the libraries involved will have quite large bindings, so definitely a good test of Zavolaj and a fair amount of work.  But at least the path forward is well lit.  :-)
05:57 dalek roast: 81d488a | moritz++ | packages/Test/Util.pm:
05:57 dalek roast: remove outdated #OK comment, japhb++
05:57 dalek roast: review: https://github.com/perl6/roast/commit/81d488ae0e
05:58 wamba joined #perl6
05:59 cggoebel joined #perl6
06:02 replore joined #perl6
06:13 birdwindupbird joined #perl6
06:15 Woodi joined #perl6
06:16 Woodi morning everyone :)
06:17 sorear o/
06:19 Woodi sorear: o/
06:24 alvis` joined #perl6
06:25 Woodi I yesterday learned something about SSI programming http://www.akkadia.org/drepper/summit09-stream.pdf  and examples are for data in arrays/vectors... do v6 specs was thinked for such uses ?
06:25 kaleem joined #perl6
06:39 wk joined #perl6
06:43 FROGGS joined #perl6
07:07 Woodi ehmm, s/SSI/SSE/
07:09 Circlepuller__ joined #perl6
07:11 leont joined #perl6
07:24 kresike joined #perl6
07:24 kresike good morning all you happy perl6 people
07:25 tadzik hello there kresike :)
07:25 kresike tadzik, o/
07:25 erkan joined #perl6
07:25 erkan joined #perl6
07:26 tadzik how are you doing?
07:28 cggoebel joined #perl6
07:29 GlitchMr joined #perl6
07:33 daniel-s joined #perl6
07:37 kurahaupo joined #perl6
07:42 dalek doc: 018e939 | (Felix Herrmann)++ | lib/variables.pod:
07:42 dalek doc: [variables] fix example of suberseding $~Regex (samlh++)
07:42 dalek doc: review: https://github.com/perl6/doc/commit/018e939dae
07:43 felher suberseding ...
07:44 * felher sure needs caffeine
07:54 fglock joined #perl6
07:55 fhelmberger joined #perl6
07:58 kresike tadzik, working ... tired ... sleepy, You ?
08:01 tadzik just working, maybe a bit underslept, but I find it hard to remember the last time I properlyslept :)
08:03 kresike I know exactly the last time I properly slept :) 7 weeks and 1 day ago ...
08:03 tadzik oh
08:03 tadzik hold on, you got a child that long ago? :)
08:04 SamuraiJack joined #perl6
08:04 kresike Yep :)
08:04 tadzik haha, congratulations :)
08:04 kresike not much sleep since
08:04 kresike thanks
08:08 SamuraiJack_ joined #perl6
08:11 jnthn morning o/
08:12 kresike morning jnthn o/
08:19 Psyche^ joined #perl6
08:19 daxim joined #perl6
08:21 wooden joined #perl6
08:21 wooden joined #perl6
08:22 felher morning jnthn o/
08:24 spider-mario joined #perl6
08:29 cggoebel joined #perl6
08:32 dakkar joined #perl6
08:57 am0c joined #perl6
09:16 scott___ joined #perl6
09:35 moritz kresike++ # congratulations
09:35 kresike thanks moritz
09:44 gongyiliao joined #perl6
09:44 FROGGS nr: say time.WHAT
09:44 p6eval niecza v21-1-ga8aa70b: OUTPUT«Num()␤»
09:44 p6eval ..rakudo 231137: OUTPUT«Int()␤»
09:48 tadzik :)
09:50 tadzik pugs: say time.WHAT
09:50 p6eval pugs: OUTPUT«Rat()␤»
09:50 tadzik yapsi: say 1
09:51 tadzik :(
09:58 sergot joined #perl6
09:58 sergot hi o/
10:00 tadzik hello sergot
10:10 tadzik moritz: on which server(s) does p6eval run?
10:10 tadzik ISTR it's built on one of them and then copied into another
10:11 ObseLeTe joined #perl6
10:12 moritz tadzik: it builds on feather1 and runs on feather3
10:16 ObseLeTe joined #perl6
10:25 tadzik so feather3 is the lightweight one, and feather1 is the one we're all ircing from, si?
10:25 tadzik sergot and me want to run the Perl 6 powered modules.perl6.org, + possibly emmentaler in a similar way
10:25 moritz correct
10:26 tadzik time to contact Juerd I guess
10:26 tadzik or just put all the things on feather1 for now
10:27 tadzik sergot: do you have an account here?
10:29 MayDaniel joined #perl6
10:29 moritz tadzik: I can give you an account on feather3 if necessary
10:29 sergot tadzik: no, I don't :)
10:32 tadzik moritz: no rush, seems that we can pretty much put it on feather1 for now
10:32 tadzik like the modules API sits
10:32 tadzik sergot: time to get one, I think :)
10:33 moritz tadzik: when you have something running, I can also set up a proxy that internally rewrites something.perl6.org to a feather1 address
10:35 tadzik moritz: that'll prove useful, thanks
10:42 sergot tadzik: how? :)
10:42 tadzik sergot: contact Juerd. I think there're instructions on http://feather.perl6.nl
10:43 tadzik or maybe not
10:50 cggoebel joined #perl6
10:52 sergot tadzik: ok, thanks. :)
10:56 wooden joined #perl6
10:56 wooden joined #perl6
10:57 moritz http://www.mail-archive.com/perl6​-language@perl.org/msg20060.html
10:57 moritz those are the instructions for getting an account
11:05 * Juerd admits that most requests end up in my mailbox, unread
11:06 Juerd It often takes a month or even longer :|
11:06 Juerd If there's anyone who would like to create accounts on feather, please let me know.
11:07 moritz Juerd: I would
11:09 fhelmberger joined #perl6
11:17 marmay joined #perl6
11:17 cggoebel joined #perl6
11:35 fhelmberger_ joined #perl6
11:40 Juerd moritz: Please go ahead if anyone asks. I'll need username, email address and real name per new account.
11:46 crab2313 joined #perl6
11:48 moritz Juerd: per email?
11:48 Juerd Preferred
11:48 Juerd Can I forward some requests to you?
11:52 moritz sure
11:52 JimmyZ joined #perl6
11:54 cognominal joined #perl6
11:54 Juerd moritz: At which email address?
11:54 moritz Juerd: moritz@cpan.org
11:55 Juerd Okay :)
11:57 fernandocorrea joined #perl6
12:02 skids joined #perl6
12:12 Kharec joined #perl6
12:12 Kharec hi guys
12:16 skids o/
12:21 cggoebel joined #perl6
12:29 cognominal joined #perl6
12:33 sergot joined #perl6
12:48 moritz \o
12:48 Kharec o/
12:48 fernandocorrea joined #perl6
13:08 benabik joined #perl6
13:09 BlueT_ joined #perl6
13:13 takadonet morning all
13:14 fernandocorrea joined #perl6
13:16 diakopter hi o/
13:17 colomon \o
13:18 BlueT_ joined #perl6
13:18 PacoAir joined #perl6
13:22 BlueT_ joined #perl6
13:36 grondilu joined #perl6
13:36 grondilu Guys, check out this video http://youtu.be/-Djj6pfR9KU and tell me how you'd write a short program to find out brown numbers.
13:38 GlitchMr Just wondering, does Perl 6 have negative zero?
13:40 tadzik that's easy to check
13:40 GlitchMr perl6: print atan2 0.Num, -0e0
13:40 p6eval rakudo 231137: OUTPUT«3.14159265358979»
13:40 p6eval ..niecza v21-1-ga8aa70b: OUTPUT«3.1415926535897931»
13:40 GlitchMr It seems it exists
13:41 GlitchMr perl6: printf '<%g>', -0e0
13:41 p6eval niecza v21-1-ga8aa70b: OUTPUT«<0>»
13:41 p6eval ..rakudo 231137: OUTPUT«<-0>»
13:41 GlitchMr Uhmmm... what should be result of this?
13:41 Kharec The first result seems to be Pi ?
13:41 gongyiliao joined #perl6
13:42 grondilu p6: say 0**0
13:42 p6eval rakudo 231137, niecza v21-1-ga8aa70b: OUTPUT«1␤»
13:42 GlitchMr But negative zeros exist only for Num type, right?
13:43 Kharec left #perl6
13:44 FROGGS joined #perl6
13:46 japhb Congratulations, kresike!  Is this your first?
13:46 kresike yes
13:46 kresike thanks
13:46 japhb Welcome to the club.  :-)
13:46 kresike :)
13:46 kresike a pretty large club I might say :)
13:47 stopbit joined #perl6
13:48 japhb Makes me wonder what percentage of currently alive humanity is in that club ...
13:51 Circlepuller joined #perl6
13:51 FROGGS joined #perl6
13:54 Circlepuller_ joined #perl6
13:59 cggoebel joined #perl6
13:59 cognominal p6:  say 'aaa' ~~  m:g/a/
13:59 p6eval rakudo 231137: OUTPUT«False␤»
13:59 p6eval ..niecza v21-1-ga8aa70b: OUTPUT«[31m===[0mSORRY![31m===[0mâ�¤â�¤Regex modifier g not yet implemented at /tmp/dA0amvxSzV line 1:â�¤------> [32msay 'aaa' ~~  m:g[33mâ��[31m/a/[0mâ�¤â�¤Unhandled exception: Check failedâ�¤â�¤  at /home/p6eval/niecza/boot/lib/CORE.setting line 1435 (die @ 5) â�¤  at /ho…
14:00 cognominal how m:g is supposed to work in Perl 6?
14:01 colomon kresike++
14:01 cognominal there is not much material in S05 about that
14:02 kresike :)
14:05 moritz cognominal: with magic
14:05 moritz cognominal: our current smart matching system doesn't allow it to work
14:05 cognominal p6: my  @a =  'aaa'.match: /a/, :g ; say @a.perl
14:05 p6eval niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Excess arguments to Cool.match, unused named g␤  at /home/p6eval/niecza/lib/CORE.setting line 0 (Cool.match @ 1) ␤  at /tmp/9MoqYnfVG_ line 1 (mainline @ 3) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4138 (ANON @ 3) ␤  at /home/p6e…
14:05 p6eval ..rakudo 231137: OUTPUT«Array.new(Match.new(orig => "aaa", from => 0, to => 1, ast => Any, list => ().list, hash => EnumMap.new()), Match.new(orig => "aaa", from => 1, to => 2, ast => Any, list => ().list, hash => EnumMap.new()), Match.new(orig => "aaa", from => 2, to => 3, ast => Any, li…
14:06 cognominal that's better
14:07 cognominal sounds like some wiring is missing to make m:g// work but that the bulk of the work has been done
14:08 FROGGS go make that wiring :P
14:09 cognominal :)
14:09 moritz cognominal, FROGGS: http://perlgeek.de/blog-en/perl-6/​three-functions-of-smartmatch.html explains why it can't be done in the current setup of smartmatchin
14:09 cognominal moritz++
14:10 * FROGGS reads
14:17 JimmyZ joined #perl6
14:19 * [Coke] hurls http://boingboing.net/2012/08/29/​t-shirt-dunkin-donuts-vs-du.html for rjbs
14:27 moritz https://github.com/perl6/specs/issues/23 # input welcome
14:30 FROGGS ya, my first thought was the dot too
14:31 tadzik input'd
14:32 FROGGS thats pretty unsafe, isnt it?
14:33 tadzik hm
14:33 moritz my concerns with the dot is that it'll lead to people writing things like require "./$filename"  just to ensure there's a dot in it
14:33 moritz and then it'll fail for absolute paths
14:35 jnthn require 'ThisIsAPath'.IO; # ?
14:35 colomon you can always add /./ in the middle of a path, no?
14:35 tadzik jnthn: +1
14:36 colomon isn't there a quoting operator for paths?
14:36 moritz jnthn: .path, not .IO. Otherwise +1
14:36 * colomon thinks jnthn is definitely on the right trail...
14:37 tadzik bah, I accidentally my comment
14:38 moritz should I delete mine too?
14:38 tadzik I put it below
14:40 bluescreen10 joined #perl6
14:42 nnunley joined #perl6
14:43 erkan joined #perl6
14:43 erkan joined #perl6
14:49 kaur left #perl6
15:00 FROGGS damn
15:01 FROGGS I reject my pull request since now does exactly what I wanted time to do
15:01 FROGGS nr: say now
15:01 p6eval niecza v21-1-ga8aa70b: OUTPUT«Instant.new(...)␤»
15:01 p6eval ..rakudo 231137: OUTPUT«Instant:1346425317.946087␤»
15:02 cggoebel joined #perl6
15:02 tadzik but shouldn't time do the floating point per spec?
15:03 fernandocorrea joined #perl6
15:03 FROGGS see this: https://github.com/perl6/specs/blob/ma​ster/S32-setting-library/Temporal.pod
15:03 FROGGS I read older specs before, these doesnt tell you what it returns
15:03 FROGGS the one I posted does
15:04 FROGGS so niecza does it wrong
15:05 [Coke] nr: say now.WHAT
15:05 p6eval rakudo 231137, niecza v21-1-ga8aa70b: OUTPUT«Instant()␤»
15:05 [Coke] they both return an instant. what's the problem?
15:06 FROGGS niecza gives you time as a Num, but it should be Int
15:06 FROGGS nr: say time.WHAT
15:06 p6eval niecza v21-1-ga8aa70b: OUTPUT«Num()␤»
15:06 p6eval ..rakudo 231137: OUTPUT«Int()␤»
15:06 wamba joined #perl6
15:07 FROGGS n: print now
15:07 p6eval niecza v21-1-ga8aa70b: OUTPUT«Instant()<instance>»
15:07 FROGGS how do I get it to print the seconds?
15:11 [Coke] n: print +now
15:11 p6eval niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Cannot use value like Instant as a number␤  at <unknown> line 0 (ExitRunloop @ 0) ␤  at /home/p6eval/niecza/lib/CORE.setting line 293 (Any.Numeric @ 6) ␤  at <unknown> line 0 (ExitRunloop @ 0) ␤  at /tmp/IDKozpItew line 1 (mainline @ 3)…
15:12 [Coke] n: print now.^methods
15:12 p6eval niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Unable to resolve method methods in type ClassHOW␤  at /tmp/0N57h0OCnx line 1 (mainline @ 5) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4138 (ANON @ 3) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4139 (module-CORE @ 571) ␤  at …
15:12 [Coke] n: print now.HOW.methods
15:12 p6eval niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Unable to resolve method methods in type ClassHOW␤  at /tmp/NWSUGOwRz8 line 1 (mainline @ 5) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4138 (ANON @ 3) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4139 (module-CORE @ 571) ␤  at …
15:14 JimmyZ r: say now.^methods
15:14 p6eval rakudo 231137: OUTPUT«new from-posix to-posix Bridge Num Int Rat abs sign conj sqrt rand sin asin cos acos tan atan atan2 sec asec cosec acosec cotan acotan sinh asinh cosh acosh tanh atanh sech asech cosech acosech cotanh acotanh floor ceiling round unpolar cis Complex log exp truncate…
15:15 FROGGS n: say now.Num
15:15 p6eval niecza v21-1-ga8aa70b: OUTPUT«Unhandled exception: Unable to resolve method Num in type Instant␤  at /tmp/IcT1nSebtS line 1 (mainline @ 4) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4138 (ANON @ 3) ␤  at /home/p6eval/niecza/lib/CORE.setting line 4139 (module-CORE @ 571) ␤  at /home…
15:16 [Coke] n: say now.to-posix
15:16 p6eval niecza v21-1-ga8aa70b: OUTPUT«1346426171.3581679 False␤»
15:16 [Coke] niecza source: boot/lib/CORE.setting -> "class Instant"
15:22 kaare_ joined #perl6
15:26 thou joined #perl6
15:32 kresike bye folks
15:33 pmichaud good morning, #perl6
15:33 FROGGS hi
15:33 JimmyZ joined #perl6
15:33 [Coke] pmichaud: hio
15:39 stopbit joined #perl6
15:47 cggoebel joined #perl6
15:56 moritz \o pmichaud
15:58 japhb moritz, does doc.perl6.org have any facility for CGI?
15:58 jnthn hi, pmichaud o/
15:58 jnthn hope your trip back went fine
16:02 pmichaud it did... was just long
16:05 jnthn *nod*
16:09 [Coke] pmichaud: ah, that's why your schedule has seemed a little off this week!
16:10 pmichaud [Coke]: yes; I just got back from .eu last night
16:13 [Coke] \o/
16:14 pmichaud anyway, I'm afk, lunch
16:17 mucker joined #perl6
16:19 fernandocorrea joined #perl6
16:20 japhb tadzik, ping
16:23 leont joined #perl6
16:27 japhb Open question: How much stylistic similarity do we want between perl6.org and its subdomains?  There's a scale there from "styled independently" through "share a common base style, but tweaked for each subdomain" to "use a single, unified style sheet".  Also, what about related domains, like the sites for each implementation?  Note BTW that I am *NOT* assuming perl6.org's current design will remain unchanged (in fact, it will have to chang
16:27 japhb e somewhat to accomodate changes like snippets on the front page).
16:28 japhb For example, one might say that the main perl6.org remains playful and inviting, while doc.perl6.org goes more sober and efficient, but they share similar fonts, borders, etc.
16:28 tadzik japhb: pong\
16:29 japhb tadzik, Is it expected that C< > doesn't work in Pod table cells?
16:29 japhb I'm wondering if it's being parsed wrong, or being rendered wrong.
16:30 tadzik japhb: no, it's not parsed right
16:30 tadzik known bug
16:30 japhb Ah, OK, thank you.
16:30 * japhb wonders what to do for a workaround ... hmmm.
16:30 japhb Is that going to be true of any inline markup in a table?
16:32 tadzik that should be fixed
16:32 tadzik I just dunno how do it sanely
16:33 wk_ joined #perl6
16:40 jnthn r: class A { has $.a; has @.a; }
16:40 p6eval rakudo 231137: OUTPUT«===SORRY!===␤Two or more attributes declared that both want an accessor method 'a'␤»
16:40 jeffreykegler joined #perl6
16:41 mucker joined #perl6
16:42 [Coke] HIGHLANDER ATTRIBUTES!
16:43 dalek doc: f1eb292 | (Geoffrey Broadwell)++ | htmlify.pl:
16:43 dalek doc: Minor regularization of status messages from htmlify.pl
16:43 dalek doc: review: https://github.com/perl6/doc/commit/f1eb292313
16:43 mucker joined #perl6
16:44 sorear japhb: I want a maximum of stylistic consistency, it will help to make perl6.org feel less like a link farm to the end user
16:48 japhb sorear, you mean because all of the subdomains and related domains will look like parts of the whole, thus making it feel like one big website instead of several smaller ones?
16:48 sorear yes
16:49 japhb sorear, OK, thank you.
16:49 mucker joined #perl6
16:51 mucker joined #perl6
16:52 japhb tadzik, Is there a workaround for Pod verbatim blocks getting split at the first line without non-whitespace?  I thought at first it might be splitting on the first empty line, or the first line with less whitespace than the current indent, but adding whitespace to empty lines did not help.  :-(
16:53 tadzik oh hm
16:53 tadzik you can probably go for =begin code, =end code
16:54 japhb OK, will try
16:57 tadzik I have a feeling that PARROT_ARGS in Makefile not worky
16:58 replore_ joined #perl6
16:59 lichtkind joined #perl6
17:01 moritz japhb: common theme, but tweaked per sub domain
17:01 japhb tadzik, Great, =begin code works
17:01 tadzik jnthn: I CAN HAS Unrecognized nqp:: opcode 'nqp::ishash' at line 41, near " {\n       "
17:01 tadzik when compiling rakudo on toqast
17:01 tadzik japhb: good
17:01 japhb moritz, what about related domains, like Rakudo's?
17:01 jnthn tadzik: Are you compiling the nqpqast branch of Rakudo?
17:02 tadzik jnthn: aye
17:02 jnthn And latest nqp/toqast?
17:02 moritz japhb: re CGI, we have a full apache installation, and can make anything happen that doesn't eat too much memory
17:03 tadzik jnthn: yep, freshly
17:03 moritz japhb: related domains... no idea about rakudo.org, probably depends on what kind of style we can offer for it. In the end pmichaud++ decides about it
17:03 jnthn tadzik: Oddness...it builds fine here with both of those.
17:03 japhb tadzik, C<< > >> Seems to not DTRT.  Known bug, or should I use something else?
17:03 tadzik japhb: yep, NYI
17:03 tadzik quite high on my list tho
17:03 moritz C<< >> NYI is known
17:04 jnthn tadzik: Can you check what revisions you are on in each branch, and then give me the dtailed backtrace?
17:04 * moritz too slow
17:04 japhb tadzik, moritz, OK, thank you.
17:04 jnthn tadzik: I recognize the error but it looks like the one you'd get from building rakudo/nqpqast with nqp/master
17:04 japhb tadzik, when you do that, could you also make French quotes work too?
17:05 erkan joined #perl6
17:05 erkan joined #perl6
17:05 tadzik jnthn: probably
17:05 tadzik jnthn: ...you didn't merge nqp/toqast with nqp/master?
17:06 tadzik that'll explain it :)
17:06 moritz jnthn: iirc I got that error in a test file in nqp on nqp/toqat
17:06 tadzik also the fact that tests passes
17:06 tadzik and that it didn't segfault :)
17:07 jnthn tadzik: no, not yet ;-)
17:07 jnthn moritz: Right, that also :)
17:07 jnthn I de-pir::-ified some of the tests
17:07 jnthn but nqp::ishash was not supported by pre-QAST NQP.
17:08 dalek nqp/toqast: a94c294 | jnthn++ | src/NQP/Actions.pm:
17:08 dalek nqp/toqast: Fix an $/ handling issue, which fixes two more of the Rakudo spectest file regressions.
17:08 dalek nqp/toqast: review: https://github.com/perl6/nqp/commit/a94c2944cb
17:09 dalek rakudo/nqpqast: 44f06cf | jnthn++ | src/Perl6/Grammar.pm:
17:09 dalek rakudo/nqpqast: Update a workaround for a bug (that needs a proper fix).
17:09 dalek rakudo/nqpqast: review: https://github.com/rakudo/rakudo/commit/44f06cf0cf
17:09 dalek rakudo/nqpqast: 6c8ef80 | jnthn++ | src/Perl6/Metamodel/AttributeContainer.pm:
17:09 dalek rakudo/nqpqast: Fix subtly buggy code that used to work for entirely the wrong reason. Deals with another regression.
17:09 dalek rakudo/nqpqast: review: https://github.com/rakudo/rakudo/commit/6c8ef80eb3
17:09 BlueT_ joined #perl6
17:09 jlaire joined #perl6
17:10 japhb moritz, Did I miss your response on my question of CGI on doc.perl6.org?  Also, what about www.perl6.org?
17:11 moritz 19:02 < moritz> japhb: re CGI, we have a full apache installation, and can make  anything happen that doesn't eat too much memory
17:11 japhb moritz, I had a feeling I'd missed it!
17:11 moritz and is reasonable to maintain etc
17:11 japhb Thank you.  OK, is there a cgi-bin already?
17:11 moritz no
17:11 moritz what are you planning?
17:12 Liz_ joined #perl6
17:14 japhb moritz, At the moment I was mostly determining my constraints ... but the first problems I was thinking about were: server-side doc search (for non-JS users), smart(er) guesses about which doc to show for non-exact matches, server-side backing for code snippet display, etc.
17:14 moritz japhb: great, those all sound like worthwhile goals
17:18 felher Can a Pod formatting code block contain other Pod blocks (besides other formatting codes?)
17:18 tadzik er
17:18 tadzik Spec used to say that
17:18 moritz I think it's insane
17:18 tadzik ISTR that we removed that because insanity :)
17:19 moritz instead we should have a way to apply formatter-like modifiers to blocks
17:19 moritz =begin :I ... =end :I
17:19 moritz to make an all-italic block
17:19 tadzik yeah, that was the conclusion I think :)
17:20 felher okay. So formatting codes used to be able to contain anything but it was changed to be more sane? :)
17:20 moritz felher: the more sane option was not to implement that part (during gsoc at least)
17:20 felher moritz: ah, i see. Okay :)
17:28 arnsholt japhb: Re: scrollback, a thorough stress-test of Zavolaj would be awesome
17:29 arnsholt Let, the bugreports begin! =)
17:30 [Coke] I am creeped out that someone is asking about a cgi-bin in 2012. :P
17:31 japhb arnsholt, heh.  Well, I don't know if *I* will be writing that code (dang $day_job ...), but I agree it's quite worthy.  :-)
17:31 frettled [Coke]: why? :)
17:32 * frettled is a bit surprised at the directory separation thingy, but whatever.
17:32 japhb [Coke], what would you prefer, that I convince moritz to host a Mojolicious instance?  :-)
17:32 japhb frettled, ?
17:32 [Coke] japhb: yes, I'd prefer that.
17:33 [Coke] the perl6 one, preferentially. ;)
17:33 japhb [Coke], sadly he did say he wanted low memory usage ... :/
17:34 frettled japhb: separating into a directory named cgi-bin is a bit unnecessary in this century, but it is rather quaint
17:35 moritz japhb: I think mojoclicious might be fine
17:35 japhb Well, my apologies to moritz++ that I misread.  I had definitely gotten the feeling that old school was desired.
17:35 moritz japhb: I just don't want to run 5 rakudo instances in parallel, or something like that
17:36 moritz old scholld is not necessary. Maintainable and not too wasteful on resources is needed
17:36 moritz *school
17:36 japhb Well OK then.  That makes things surprisingly easier.  :-)
17:37 * japhb takes a step back to consider a much wider range of possible architectures.
17:38 arnsholt japhb: I'd be happy to help out with some of the Zavolaj hacking. But yeah, dang $work getting in the way =)
17:38 [Coke] I'd be happy with some perl5 mojo in the mean time, and would be happy to contribute code, and help convert to 6mojo when possible.
17:38 __sri what's the state of non-blocking I/O? :)
17:38 * japhb is >.< that 6mojo is not easy at this point, but waddyagonnado besides keep hacking ...?
17:39 leont __sri: see MuEvent
17:39 __sri MuEvent is unusable
17:39 leont It's very limited by Parrot
17:40 __sri polling every socket individually just doesn't work
17:40 * leont agrees, it's just that some interfacing with a c library is going to be required to do this well
17:40 leont libuv may be a good target
17:40 * __sri nods
17:41 * japhb has been reading the libuv docs over the last few days with that in mind
17:41 __sri if rakudo had really cheap threads we could also get away with blocking I/O
17:42 japhb Anyone know the state of the threads merge in Parrot?  Last I heard it was "soon"
17:42 leont Yeah, but I'm having the impression parrot threads are rather similar in costs to perl 5 threads
17:42 __sri eeep
17:43 leont They're still interpreter threads
17:44 leont japhb: ask nine on #parrot, he's doing that work
17:46 BlueT_ joined #perl6
17:47 tadzik jnthn: not my fault \o/
17:47 tadzik wanna see a patch that fixes those tests?
17:48 tadzik https://gist.github.com/3556432 :)
17:53 jnthn tadzik: Oh, I fixed it in the NQP patch a few moments ago
17:53 jnthn japhb: I got a build of libuv done on my laptop at the hackathon and started reading through the docs also.
17:53 japhb jnthn, excellent
17:53 jnthn And sketched a little code. Didn't get anything interesting running just yet.
17:54 cognominal joined #perl6
17:55 japhb whiteknight in #parrot says that the initial threads implementation should merge on Tuesday. They expect to then work on tuning it, but that (given that it is a hybrid tasks/threads implementation), they should be able to get medium-to-light virtual threads spread across several OS threads.
17:59 colomon jnthn: Any notion how hard it would be to implement "lift"?  It would be very handy for Math::Polynomial...
18:01 jnthn colomon: Maybe not too bad. I guess we have CALLER:: support
18:01 colomon jnthn: who do I have to bribe?  ;)
18:01 pmichaud back from lunch
18:01 jnthn Bribe Pm! ;)
18:02 pmichaud anyone have any comments on the rc3 star candidate before I start publishing it?
18:02 jnthn pmichaud: I could try a build of it on this box if you wish.
18:02 [Coke] it should be blue.
18:02 pmichaud jnthn: that would be helpful.  I'm about to build the .msi here on my box;.
18:02 pmichaud *box.
18:03 jnthn pmichaud: url?
18:03 colomon [Coke]: but should it be nasally-fitted?
18:03 pmichaud http://pmichaud.com/sandbox/r​akudo-star-2012.08-rc3.tar.gz
18:04 [Coke] colomon++∞
18:05 jnthn pmichaud: Kicked off a build
18:06 BlueT_ joined #perl6
18:12 jnthn pmichaud: With --gen-parrot it build an NQP and Parrot fine and also told me to next run "nmake"
18:12 jnthn pmichaud: However, during Configure it did also emit:
18:12 jnthn 'make' is not recognized as an internal or external command,
18:12 jnthn operable program or batch file.
18:12 pmichaud oh.
18:12 jnthn It's harmless almost certainly.
18:12 jnthn Just a tiny bit distracging.
18:12 * pmichaud checks.
18:12 jnthn *distracting
18:15 pmichaud I'll go ahead and fix that now.
18:15 jnthn heh, the QAST NQP seems to have fixed a bug that in turn hid a bug in the Rakudo grammar :)
18:16 jnthn And the fix was the copy-pasting from STD :D
18:16 BlueT_ joined #perl6
18:16 thou joined #perl6
18:18 dalek rakudo/nqpqast: f962cf8 | jnthn++ | src/Perl6/Grammar.pm:
18:18 dalek rakudo/nqpqast: Fix $< parsing (just replace what we had before with what STD has). Fixes the S05 failures.
18:18 dalek rakudo/nqpqast: review: https://github.com/rakudo/rakudo/commit/f962cf8fd9
18:18 jnthn pmichaud: Ah, a real fail:
18:18 jnthn Cannot chmod 755 modules/rakudo-debugger/perl6-debug.exe:No such file or directo
18:18 jnthn ry at C:/Perl64/lib/ExtUtils/Command.pm line 274.
18:18 jnthn NMAKE : fatal error U1077: 'C:\Perl64\bin\perl.exe' : return code '0x2'
18:18 jnthn Stop.
18:19 pmichaud okay.
18:19 pmichaud did it not build?
18:19 jnthn It build
18:20 jnthn er, hm
18:20 jnthn It ran pbc_to_exe
18:21 pmichaud I'm building on my box too so maybe I'll find the error here as well.
18:22 mucker ok, i have rakudo compiled and running. now, how do i run the test suite(roast) using rakudo ?
18:22 jnthn mucker: make spectest
18:22 pmichaud or make "rakudo-spectest" if using a star distribution.
18:22 mucker is the process same for niecenza ?
18:23 colomon mucker: yes, make spectest
18:23 mucker thnkx
18:23 jnthn pmichaud: I can't find the .pir and .pbc, let alone the .exe
18:23 pmichaud jnthn: for rakudo-debugger?
18:24 jnthn yeah
18:24 pmichaud could be that my make rules aren't windows-aware
18:24 pmichaud looking.
18:24 jnthn pmichaud: Trying "nmake" again does nothing
18:24 pmichaud that's commonly true
18:25 jnthn And nmake clean tries to use "make"
18:25 pmichaud repeating "nmake" often doesn't do anything, because we don't have every file tracked in the makefile (and doing so would be a huge pain)
18:25 jnthn (maybe related to the issue earlier)
18:25 pmichaud I fixed the nmake clean bug already.  :-)
18:25 jnthn ah, makes sense.
18:25 jnthn ok, nice
18:26 dalek star: c4268e0 | pmichaud++ | tools/ (2 files):
18:26 dalek star: Replace some 'make' with '$(MAKE)'  (jnthn++)
18:26 dalek star: review: https://github.com/rakudo/star/commit/c4268e0178
18:27 pmichaud jnthn: you could try "nmake modules-install" and see if it attempts the rakudo-debugger build again.
18:27 pmichaud (I ended up putting the debugger as a module, for a variety of reasons.)
18:27 lichtkind joined #perl6
18:29 jnthn it's running
18:31 jnthn pmichaud: hmm, I don't see it being built again.
18:31 pmichaud okay, I'll try here and see what I can come up with.
18:32 jnthn pmichaud: Oh,hm
18:32 jnthn pmichaud: And it does try to build modules/rakudo-debugger/lib​/Debugger/UI/CommandLine.pm
18:32 jnthn Before ANSI::TermColor is built
18:32 pmichaud yes, it'll build the libs, but the rakudo debugger binary is a separate step
18:33 jnthn but it builds the lib too early
18:33 jnthn Oh
18:33 jnthn Watching how fast it claims to have built it...I can't believe it really did so
18:33 jnthn My suspicion is that it fails silently
18:34 pmichaud make 'nmake perl6-debug-install'
18:34 pmichaud s/make/maybe/
18:34 pmichaud now that my brain is waking up, I remember that modules-install only installs the libs, but doesn't compile the executable
18:34 jnthn Also failed silently.
18:34 jnthn OK
18:34 tadzik jnthn: huh, hm
18:35 jnthn It still tries to build the lib too early though.
18:35 pmichaud what is "too early" in this case?
18:36 jnthn Before Term::ANSIColor, which it depends on.
18:36 pmichaud Term::ANSIColor is supposed to be built long before rakudo-debugger
18:36 pmichaud it's #6 in the modules/MODULES.txt file
18:36 pmichaud whereas rakudo-debugger is last
18:37 jnthn It's odd
18:37 jnthn It build the rakudo-debugger executable last
18:37 pmichaud that's normal.
18:37 jnthn As in, the NQP bit
18:37 pmichaud that's normal.
18:37 jnthn But it builds the thing in lib/ first
18:37 jnthn LIke, as part of module-install.
18:37 pmichaud that's correct.
18:38 jnthn Yes, by first I mean, before any other module.
18:38 pmichaud okay, that's odd.
18:38 pmichaud you mean, it does rakudo-debugger/lib before any other module?
18:38 jnthn Correct
18:38 jnthn After the docs.
18:38 pmichaud using what make target?
18:38 jnthn But before NativeCall
18:38 jnthn It did it with nmake install when I tried that originally
18:39 jnthn And it does it when I nmake modules-install
18:39 pmichaud is nmake perhaps trying to parallelize the build?
18:39 jnthn Shouldn't be.
18:40 jnthn Seems not.
18:40 pmichaud it all seems to be working fine for me here (using gmake)
18:40 jnthn What does the output look like for you when you make modules-install?
18:41 pmichaud doing that now
18:41 jnthn hm, I also get:
18:41 jnthn Can't stat modules/#/lib: No such file or directory at tools/build/module-install.pl line 21
18:41 jnthn AT some point during modules-install :)
18:43 pmichaud okay, so make modules-install for me
18:43 pmichaud 1.  cd rakudo && make install   (installs perl6.exe and modules from rakudo compiler)
18:44 pmichaud 2. == Installing 'ufo'
18:44 pmichaud 3.  Errors about installing '#'
18:44 pmichaud == Instaling 'zavolaj'
18:44 jnthn https://gist.github.com/3557195
18:44 jnthn That's what I get
18:44 jnthn Notice something really odd happens
18:45 pmichaud note that Term::ANSIColor is installed on line 62 of that output.
18:46 jnthn huh
18:46 jnthn then later on it tries to do everything again
18:46 pmichaud ...?
18:46 jnthn that top bit of output scross by really fast
18:46 pmichaud tries to do everything again?
18:46 pmichaud what line?
18:46 jnthn See from line 282 onwards
18:47 pmichaud that's not installing everything again -- that's precompiling the installed modules.
18:47 cggoebel joined #perl6
18:47 jnthn oh!
18:47 BlueT_ joined #perl6
18:47 jnthn jnthn--
18:48 jnthn OK, then it seems it is doing the right thing with modules
18:48 pmichaud there could be a message there saying "precompiling installed modules" though -- that would help.  :)
18:49 jnthn Apparently. ;-)
18:49 pmichaud can you do the equivalent of   make modules/rakudo-debugger/perl6-debug.exe ?
18:50 pmichaud (I've since added a 'perl6-debug' target, but it's not in the rc3 tarball.)
18:50 pmichaud or you can just do 'make perl6-debug-install'
18:52 skids r: sub a ($b --> Int) { fail("oops") unless $b == 2; $b; }; a(2).say; a(1);
18:52 p6eval rakudo 231137: OUTPUT«2␤Type check failed for return value; expected 'Int' but got 'Failure'␤  in sub a at /tmp/PknV_q94S4:1␤  in block  at /tmp/PknV_q94S4:1␤␤»
18:52 skids That seems to be new with latest tarball.
18:52 skids (since 07 star)
18:52 moritz skids: yes, it's new that type constraints with --> are honored
18:52 pmichaud but fail should bypass a type constraint
18:52 moritz previously only 'returns' was honored
18:54 jnthn pmichaud: trying it
18:54 skids Also I have a few files where I've got a bunch of messy and probably wrong export directives, which ran under 07 star.  Now they give: Method 'scope' not found for invocant of class 'QAST::Op'
18:54 pmichaud so the   --> Int   listed above is a rakudobug
18:55 moritz pmichaud: it is, and known
18:55 moritz (RT'ed)
18:55 pmichaud moritz++
18:55 skids Crunched on time to golf it tho.
18:55 pmichaud the -->  Int bug is not a good regression to have, though :-(
18:56 BlueT_ joined #perl6
18:56 skids would not be as bad if --> Int|Failure worked.
18:56 moritz pmichaud: note that before --> AnyType was just ignored
18:56 pmichaud moritz: I note that, yes.
18:57 pmichaud but having correct code that previously "worked"  (didn't crash)   now crash is a regression.
19:00 jnthn Yeah, one of those "bug hiding a bug" things.
19:01 pmichaud well, it's more like an incomplete implementation of --> Type
19:02 moritz more like an incomplete implementation of fail(), if you ask me
19:02 jnthn It's probably an incompleteness in the return type check handling.
19:02 jnthn --> used to be totally ignored
19:02 jnthn Now it's applied, it has the exact same bug that "returns" has.
19:03 skids r: our $foo::bar is export; # I don't know if this is nonsensical but it used to not bail out.
19:03 p6eval rakudo 231137: OUTPUT«===SORRY!===␤Method 'scope' not found for invocant of class 'QAST::Op'␤»
19:03 jnthn I'm...not sure that's valid, in the same way my $foo::bar is dubious.
19:03 jnthn But yeah, it should not fail that way
19:03 moritz r: our $foo::bar
19:03 p6eval rakudo 231137: OUTPUT«===SORRY!===␤Method 'scope' not found for invocant of class 'QAST::Op'␤»
19:03 jnthn I dunno what it used to do... :)
19:04 sorear n: our $foo::bar is export # curious
19:04 p6eval niecza v21-1-ga8aa70b: OUTPUT«[31m===[0mSORRY![31m===[0m��Trait export not available on variables at /tmp/abdyJzy2Kt line 1 (EOF):�------> [32mour $foo::bar is export # curious[33m�[31m<EOL>[0m��:: syntax is only valid when referencing variables, not when defining them. at /tmp/…
19:04 sorear heh, NYI masks any possible bug
19:04 jnthn No, it emits the :: syntax note afterwards.
19:04 jnthn I suspect Rakudo should do the same.
19:05 moritz elsif $twigil ne '*' { $past.scope('lexcial') } # that's the code path that dies
19:05 pmichaud anyone know the RT bug for the return type check fail fail?
19:05 jnthn pmichaud: OK, I'm confused
19:05 moritz pmichaud: I'll try to find it
19:05 jnthn pmichaud: It tries to build the nqp-debug.nqp into PIR
19:06 jnthn And seems to silenlty fail to do so from within make
19:06 jnthn *nmake
19:06 jnthn But if I copy-paste the command it says it runs...it compiles it to PIR.
19:06 pmichaud jnthn: can you nopaste some output?
19:06 jnthn https://gist.github.com/3557511 is the relevant build log
19:07 tadzik jnthn: yep, the tests pass now
19:07 jnthn tadzik: Did you do a spectest run?
19:08 tadzik jnthn: it's still ongoing, but it fixed this Pod issues
19:08 jnthn tadzik: OK
19:08 jnthn pmichaud: Does the build "cd" at some point inside of the makefile?
19:08 pmichaud jnthn: it's not creating modules/rakudo-debugger/perl6-debug.pir ?
19:08 jnthn No
19:08 skids .oO(ah, "is export" is extraneous.)
19:08 jnthn pmichaud: But if I run the command manually it does create one.
19:08 moritz pmichaud: https://rt.perl.org/rt3//Ti​cket/Display.html?id=64990
19:09 pmichaud moritz++  # thanks
19:09 moritz oops, seems it's marked as 'resolved'
19:09 jnthn pmichaud: But I noticed the paths are relative, thus why I'm pondering if it's a cwd fail
19:09 pmichaud jnthn: yes, the makefile does 'cd' at a few points
19:10 pmichaud but if it wasn't working, I'd expect 'cannot find ...'  errors as opposed to silent fails
19:10 skids Anyway, time to start the long weekend, where's my bottle o rum?  Good luck on the release all!
19:10 jnthn pmichaud: same
19:10 jnthn pmichaud: Another data point: the ufo and panda copying appears to have worked out
19:11 pmichaud jnthn: yeah, that leads me to believe that we're in the correct cwd
19:11 jnthn same
19:11 pmichaud think the / vs \ is an issue?
19:12 jnthn Maybe, but when I copy-paste the exact command there without changing the slashes, it does execute and creates the .pir file
19:12 pmichaud are the other files above being precompiled to .pir?  Those commands also use / instead of \
19:13 BlueT_ joined #perl6
19:14 pochi joined #perl6
19:14 jnthn Yeah; looking in C:/consulting/rstar-rc/rakudo-star-2012.08/i​nstall/lib/parrot/4.6.0/languages/perl6/lib/ I see plenty of pre-compiled .pir files
19:16 birdwindupbird joined #perl6
19:16 pmichaud hmmphs.
19:16 jnthn Where's the thing that builds rakudo-debugger?
19:16 jnthn or tries to...
19:16 pmichaud it's the Makefile that gets generated by Configure.pl
19:16 pmichaud source is tools/build/Makefile.in
19:17 flussence joined #perl6
19:18 pmichaud follow perl6-debug-install in the Makefile
19:20 * jnthn tries getting it to spit out the cwd, just in case
19:20 jnthn but it seems a stretch
19:20 pmichaud anyway, it's building fine for me.
19:20 jnthn yeah, cwd looks fine
19:21 pmichaud maybe some odd characters in the makefile?
19:21 pmichaud I'm kinda grasping at straws :)
19:22 jnthn yeah, me too
19:22 pmichaud perhaps try eliminating the continuation line ?
19:22 jnthn It's really odd.
19:22 pmichaud on the nqp.exe line
19:22 [Coke] do a diff on the 2 generated makefiles?
19:24 pmichaud r:  sub abc() returns Int { 3.5 };  my $x = abc();  say 'alive';
19:24 p6eval rakudo 231137: OUTPUT«Type check failed for return value; expected 'Int' but got 'Rat'␤  in block  at /tmp/1TE1XyShlH:1␤␤»
19:24 * pmichaud files rakudobug
19:25 jnthn pmichaud: ?
19:25 pmichaud S02:1363:
19:25 pmichaud (In fact, the attempt to
19:25 pmichaud return a value that doesn't match the return constraint will internally
19:25 pmichaud call C<fail>, and that failure will be returned according to standard
19:25 pmichaud C<fail> policy, which is dependent on the preferences of the caller.)
19:26 jnthn wtf :/
19:26 moritz wtf.
19:26 pmichaud that sounds about right to me, though.
19:26 [Coke] so that shouldn't die unless you say $x ?
19:26 pmichaud or otherwise try to use $x in a way that causes the exception to be thrown.
19:26 jnthn That's inconsistent with parameter passing's type checking...
19:26 moritz pmichaud: it might be, but it's very bad to switch the behavior until we have better support for failure propagation and blowup
19:26 [Coke] That could really be a compile time failure there.
19:27 jnthn What if we're "returns"-ing a native?
19:28 pmichaud in the case of a native, I suspect it's okay to die immediately.
19:28 jnthn Nowhere else does a type constraint not being met cause a Failure rather than an immediate exception.
19:29 jnthn I don't immediately see why this case deserves to be the exception to the rule.
19:29 jnthn Unless there's some other rule that it not giving back a Failure would be an exception to...
19:31 jnthn pmichaud: Eliminating the continuation line didn't help. Quoting the arg to vmlibs didnt' either.
19:31 pmichaud jnthn: I'm at a loss then :-/
19:32 jnthn pmichaud: me too
19:32 jnthn I'd not hold the relesae for it
19:32 jnthn Especially given we produce MSIs and most Windows users will go for those.
19:32 pmichaud okay.  I have some other changes, so a new rc will be forthcoming soon.
19:34 pmichaud r:  sub abc() returns Int { my $r = 3.5  };  my $x = abc();  say 'alive';
19:34 p6eval rakudo 231137: OUTPUT«Type check failed for return value; expected 'Int' but got 'Rat'␤  in sub abc at /tmp/ndFsrYzcB_:1␤  in block  at /tmp/ndFsrYzcB_:1␤␤»
19:34 pmichaud r:  sub abc() returns Int { my $r = 3.5; $r  };  my $x = abc();  say 'alive';
19:34 p6eval rakudo 231137: OUTPUT«Type check failed for return value; expected 'Int' but got 'Rat'␤  in sub abc at /tmp/P8uacuh4gB:1␤  in block  at /tmp/P8uacuh4gB:1␤␤»
19:35 pmichaud afk, kid pickup
19:37 BlueT_ joined #perl6
19:44 * TimToady is finally back home, after an extra night's stay at MSP courtesy of Delta
19:44 phenny TimToady: 29 Aug 01:03Z <diakopter> ask TimToady should .codes return the original number of codes as inputted, or the number of codes it will output when it's outputted as NFC+marks?
19:44 phenny TimToady: 29 Aug 01:06Z <diakopter> ask TimToady or some other formula I can't imagine at the moment? :)
19:44 phenny TimToady: 29 Aug 01:22Z <diakopter> tell TimToady sorear answered it; see the log..
19:44 diakopter Delta+-
19:44 moritz welcome back, TimToady
19:44 [Coke] welcome back, TimToady++
19:44 TimToady well, it wasn't really Delta's fault, but the security in Amsterdam...
19:45 jnthn Ouch :/
19:45 TimToady after we we all packed in a ready to go, they decided they couldn't decide whether a certain door had been left unlocked, and made us all get off the plane to get rescanned.
19:46 TimToady hmm, my power supply is making funny crackling sounds
19:46 TimToady maybe it didn't like 220 volts...
19:46 [Coke] Oy.
19:46 TimToady so anyway, the resulting 2 1/2 hour delay caused us to miss our connection yesterday
19:46 jnthn :(
19:46 TimToady and of course the motel's WiFi didn't work...
19:46 moritz :(
19:47 moritz TimToady: it's even 230 V these days
19:47 TimToady but we're back home \o/
19:47 colomon \o/
19:48 jnthn nqp: my @x := nqp::split('::', 'A::B'); say(@x);
19:48 p6eval nqp: OUTPUT«[ "A", "B" ]␤»
19:49 jnthn nqp: my @x := nqp::split('::', 'A::B'); say(pir::typeof__SP(@x));
19:49 p6eval nqp: OUTPUT«ResizableStringArray␤»
19:49 jnthn nqp: my @x := nqp::split('::', 'A::B'); say(+@x);
19:49 p6eval nqp: OUTPUT«2␤»
19:49 diakopter parrot leak?
19:50 jnthn diakopter: it's NQP :)
19:50 jnthn I didn't expect it to stringifythat way
19:50 jnthn *that
19:51 [Coke] jnthn: yah.
19:51 sorear huh.  apparently FRA is actually busier than TXL, it seemed quite the other way around when I was there
19:52 wk_ joined #perl6
19:54 jnthn sorear: FRA is rather large, so even if there are more people there, it may feel less crowded.
19:56 jnthn The kast times I've been at FRA, the check-in area has felt almost deserted...
19:56 jnthn *last
19:56 moritz FRA is the biggest German airport
19:56 jnthn Security felt busier though ;)
19:58 sorear moritz: yeah, I was just looking at wikipedia's list of busiest passenger airports, after first looking up "where in the world is MSP"
19:58 moritz and, where is it? :-)
19:58 sorear so, LHR, ORD, LAX, FRA, I've been at four of them in the last two week
19:58 sorear moritz: minneapolis-st.paul, north central USA
19:58 TimToady and you survived!
19:59 TimToady we saved about $500 each flying into LUX instead of FRA
19:59 sorear well I flew into LUX because that's where woolfy wanted to pick me up ...
20:00 dalek doc: c386ae4 | moritz++ | lib/Pod/ (5 files):
20:00 dalek doc: start to document Pod classes
20:00 dalek doc: review: https://github.com/perl6/doc/commit/c386ae4351
20:00 jnthn Eww, LHR... :)
20:01 * moritz isn't too fond of that either
20:01 sorear jnthn: agreed
20:01 scottdware joined #perl6
20:02 * jnthn wasn't especially amused with OSL yesterday, mind.
20:02 sorear ORD wasn't too bad.  land, 5 minute dash on foot to my destination plane .5km away or so, board just before it left
20:03 jnthn I guess it's a nice airport overall, but making people doing a tight transfer hurry through a packed duty free shop is annoying.
20:03 diakopter sorear: nice :)
20:03 diakopter (that you made it)
20:03 sorear not the best transfer, but I couldn't have done it at all in LHR or TXL (they both insist on rescreening connecting passengers)
20:03 ivan`` joined #perl6
20:03 cggoebel joined #perl6
20:04 scottdware hey perl6'ers! just a quick ? any of you guys using perl6 regularly for "production"
20:04 scottdware such as in your day-to-day scripts/programs, etc.
20:04 pmichaud back again
20:04 moritz scottdware: I use it almost daily for small tasks
20:05 moritz (and so do many other of the "core" folks)
20:05 moritz and sirrobert is using it in his business
20:05 dalek nqp/toqast: c07af7a | jnthn++ | src/ops/nqp.ops:
20:05 dalek nqp/toqast: RSA should also report true in nqp::islist, otherwise the result of split does not count as a list, which caused some breakage.
20:05 dalek nqp/toqast: review: https://github.com/perl6/nqp/commit/c07af7a4a5
20:05 pmichaud TimToady: welcome back
20:05 colomon scottdware: I use it routinely for generating sheet music.  often for small tasks for $work.  and I'm currently considering doing a medium-sized $work project in it.
20:05 scottdware moritz: cool, thanks!
20:05 pmichaud back from kid pickup
20:05 jnthn scottdware: I've used it at a $client for some small scripts.
20:06 scottdware what is sirrobert's business, can you say?
20:06 scottdware i really love what is being done with the language, and think that its awesome!
20:06 moritz scottdware: no, either he didn't say, or I forgot, sorry :(
20:06 jnthn OK, I now put out latest nqp/toqast and rakudo/nqpqast for general spectesting and also a module test run.
20:06 jnthn spectests look good here.
20:07 moritz jnthn: btw last I looked, line numbers in NQP error messages were one off (one too large)
20:07 jnthn moritz: Is that a new fail?
20:07 * moritz starts a build
20:08 moritz jnthn: not new in the nqp/toqast sense
20:08 moritz iirc
20:08 jnthn ok, thanks
20:09 moritz /home/moritz/p6/rakudo/install/bin/parrot --include=src/stage1 -o src/stage1/ModuleLoader.pbc \ src/stage1/gen/module_loader.pir
20:09 moritz make: *** [src/stage1/nqpmo.pbc] Segmentation fault
20:09 moritz oh wait
20:09 moritz it's the nqpmo step that segfaults
20:09 moritz parallel make just confused the order
20:10 [Coke] is rakudo/nqppast smart enough to pull the right nqp with --gen-parrot ?
20:10 moritz no
20:10 * [Coke] really hates it when || make gets borked. :|
20:10 moritz [Coke]: it's not the parallel make that's borked
20:10 moritz [Coke]: rather it's a build step that fails
20:10 moritz and parallel make made be believe it was a different step that failed
20:10 jnthn [Coke]: Just made it so.
20:10 dalek rakudo/nqpqast: 9c1534d | jnthn++ | tools/build/NQP_REVISION:
20:10 dalek rakudo/nqpqast: Update NQP_REVISION to ease testing.
20:10 dalek rakudo/nqpqast: review: https://github.com/rakudo/rakudo/commit/9c1534d3ac
20:12 dalek star: f28d3a8 | pmichaud++ | tools/build/Makefile.in:
20:12 dalek star: Add a perl6-debug target to the Makefile.
20:12 dalek star: review: https://github.com/rakudo/star/commit/f28d3a8b51
20:12 dalek star: 30f8f6b | pmichaud++ | tools/build/ (2 files):
20:12 dalek star: Add better progress messages to module installation.
20:12 dalek star: review: https://github.com/rakudo/star/commit/30f8f6b8f8
20:12 moritz jnthn: https://gist.github.com/3558304
20:12 jnthn moritz: hmmmmm.
20:13 Moukeddar joined #perl6
20:13 moritz somehow I have the suspicion it's the hash in the serialization thing that weeds out duplicates
20:13 jnthn Right
20:13 jnthn Well, that's where it fails at least
20:15 pmichaud rc4 now at http://pmichaud.com/sandbox/r​akudo-star-2012.08-rc4.tar.gz
20:18 sirrobert moose has things like 'after' and 'before' to hook into method execution.  Does p6 have similar?
20:18 moritz sirrobert: there's Routine.wrap
20:18 sirrobert moritz: great, thanks
20:20 [Coke] jnthn: invoke() not implemented in class '$'
20:20 jnthn ah shit...
20:20 jnthn What the hell is that about.
20:20 [Coke] (OS X 10.7.3)
20:20 jnthn And why does it only affect NQP.
20:20 jnthn And not Rakudo.
20:21 moritz [Coke]: is that reproducable?
20:21 jnthn How much RAM do each of you have?
20:21 [Coke] https://gist.github.com/3558426
20:22 moritz jnthn: that was on a machine with 4GB
20:22 [Coke] 4096MB
20:22 [Coke] a lot of which is in use, of course.
20:22 jnthn yeah, I'm pondering how it affects GC
20:22 [Coke] moritz: git clean'ing and trying again.
20:26 * jnthn lies to Parrot about how much RAM he has to see if that'll reproduce it here
20:26 crazed- joined #perl6
20:27 jnthn aha!
20:27 cotto joined #perl6
20:28 jnthn Yes, I can make it explode in interesting ways here now
20:28 [Coke] yes, I can duplicate it.
20:31 pmichaud looks like I'll need a rc5
20:31 sorear sirrobert: Routine.wrap is not a replacement for Moose's before/after, sadly
20:31 sirrobert no?
20:32 sorear no, Routine.wrap modifies subs in place while before/after/around is a principled form of overriding
20:32 sirrobert it looks like Method::Wrap (or whatever) from p5
20:32 sirrobert ok
20:32 sorear if you want to use before/after in p6, you'll need to do it yourself: method foo($x,$y) { callsame() * 2 }
20:32 sorear callsame/nextsame and the saner way we do contexts make this much easier to get right than in p5
20:33 sorear which eliminates much of the need for method modifiers ... on classes
20:33 sorear but there is currently nothing in the p6 spec which can replace method modifiers on roles
20:33 sirrobert hmm ok
20:34 jnthn pmichaud: rc5 for?
20:34 jnthn pmichaud: The segv?
20:35 dalek star: 0cb73f2 | pmichaud++ | / (2 files):
20:35 dalek star: Add 'configclean' target to handle cleanup from Configure.pl.
20:35 dalek star: review: https://github.com/rakudo/star/commit/0cb73f2bab
20:35 pmichaud jnthn: the "make clean" at the end of Configure.pl
20:35 jnthn ah
20:35 jnthn I think I may have found the NQP SEGV.
20:35 pmichaud jnthn: I don't have any fix for the segfault, unless we want to re-issue a nqp release
20:35 * jnthn smacks self
20:35 jnthn pmichaud: If it's what I think it is, then the patch will be pickable into the release.
20:36 jnthn (Visual Studio Debugger)++
20:37 sirrobert ok, gotta head out --wave
20:37 jnthn Also the nature of the problem explains why the issue didn't show up in Rakudo.
20:38 cggoebel joined #perl6
20:38 moritz now I'm really curious.
20:39 rob__ joined #perl6
20:39 dalek nqp/toqast: c0fbe4d | jnthn++ | src/6model/reprs/KnowHOWREPR.c:
20:39 dalek nqp/toqast: Add missing mark in KnowHOWREPR that could lead to segfaults and other corruption.
20:39 dalek nqp/toqast: review: https://github.com/perl6/nqp/commit/c0fbe4d73f
20:39 moritz fwiw the star RC4 looks quite good so far
20:39 jnthn We don't really write knowhow's in Rakudo, we already bootstrapped beyond that point.
20:39 pmichaud (i.e., ..."pickable into the release"?)
20:39 pmichaud jnthn: should I hold off on star release for possible patch?
20:39 jnthn pmichaud: c0fbe4d helps, it seems
20:40 moritz Testing modules/panda...
20:40 moritz t/panda/builder.t .... Missing or wrong version of dependency 'src/stage2/QRegex.nqp'
20:40 moritz in the RC-4
20:40 telex joined #perl6
20:40 jnthn pmichaud: Done multiple rebuilds with that patch where it previously failed and things look good.
20:41 jnthn moritz, [Coke]: Please give it another try.
20:41 moritz jnthn: already rebuilding
20:41 jnthn Thanks.
20:41 jnthn Dang, wish I'd figured out how to reproduce it earlier.
20:42 pmichaud jnthn: c0fbe4d  is only in the toqast branch?
20:42 jnthn pmichaud: yeah
20:42 jnthn it can be cherry-picked
20:42 pmichaud so, cherry-pick c0fbe4d into master, issue a new nqp release?
20:43 moritz cherry-pick the release into 2012.08
20:43 jnthn pmichaud: You can do that, or you can just patch the Star tarball directly. Whichever you prefer.
20:43 pmichaud if I patch the Star tarball, there's really no history of the fact that it was patched.
20:43 leont left #perl6
20:43 pmichaud not sure that's important.
20:44 pmichaud I suppose I can rebuild the 'patch' infrastructure I once had for a previous release.
20:44 * jnthn finds that hard to call
20:44 moritz just make a new nqp release
20:44 pmichaud so, checkout 2012.08, apply patch, make 2012.08.1 release?
20:44 moritz +1
20:45 jnthn Curious fact: if I tell Parrot that I have 1 GB or RAM rather than the 16GB I actually have, the build of Rakudo's CORE.setting totals up with 250MB or so less memory usage.
20:45 moritz you don't even need to make a new compiler release of rakudo
20:45 jnthn However, it takes 10s longer.
20:45 jnthn er, 12s
20:45 pmichaud jnthn: that's not curious.  parrot sets the size of its gc pool based on the amount of memory it thinks you have
20:45 moritz since you can depend on specific nqp versions in star
20:46 jnthn pmichaud: I knew that bit, just surprised that the total usage came out quite a bit less.
20:46 moritz jnthn: nqp build succeeded here, testing rakudo now
20:46 jnthn \o/
20:47 * jnthn crosses his fingers that the issue [Coke]++ saw is also gone
20:47 pmichaud should I make a new compiler release of rakudo that loads nqp 2012.08.1, though?
20:47 pmichaud otherwise it'll grab 2012.08 and we get errors.
20:47 moritz someone can do that after the star release
20:48 jnthn pmichaud: My feeling is that most people either grab * or follow Git.
20:48 jnthn But that may well be off
20:48 pmichaud likely true.
20:48 pmichaud okay.
20:48 moritz I'd just really love to see 2012.08 release in 2012.08 :-)
20:48 pmichaud patching nqp.
20:48 pmichaud oh, I'm sure the 2012.08 release will be in 2012.08
20:48 pmichaud I still have 8 hrs here.
20:48 jnthn ...in some timezone :P
20:48 jnthn :D
20:48 pmichaud it's even possible it'll happen in 2012.08 utc
20:49 pmichaud (likely, in fact, since I have 3 hrs for that)
20:50 pmichaud cherry-picked, now building and testing
20:50 pmichaud I should also build/test on my notebook, since that's wehre I was getting the segfaults earlier
20:51 [Coke] jnthn: testing now
20:51 pmichaud okay, by checking out 2012.08 I'm in detached head.  Will a tag be enough to preserve the commit history so I can recover it?
20:51 pmichaud or do we need to keep a branch?
20:52 moritz tag is enough
20:52 jnthn tag will do it
20:52 pmichaud okay.
20:52 jnthn It's detached because you can't commit to a tag.
20:52 [Coke] aw, you can in subversion!
20:53 jnthn That was NOT a feature. :P
20:53 pmichaud is there a readme or file I should change in nqp where we can explain why the .1 release?
20:53 pmichaud we don't have a ChangeLog or equivalent file, nor announcements.
20:54 jnthn We have neither of those.
20:54 pmichaud (I'm fine if we don't explain it, I think... just curious if anyone knows of a good place for the explanation)
20:54 jnthn Not for now, at least.
20:54 cognominal my $patrx := $pat ~~ Code ?? $pat !! / $pat: /;  # what is the meaning of the colon in this code?
20:54 pmichaud which colon?
20:54 pmichaud (there are two)
20:55 pmichaud the colon in the regex disables backtracking
20:55 cognominal ok, thx
20:56 dalek std: b759d67 | larry++ | / (4 files):
20:56 dalek std: viv+STD_P5 now parses all of viv
20:56 dalek std: review: https://github.com/perl6/std/commit/b759d67357
20:56 pmichaud wow!  TimToady++
20:56 sorear TimToady++ \o/
20:56 moritz \o/
20:58 jnthn wow! :D
20:58 jnthn TimToady++
20:58 lizmat TimToady++
20:58 lizmat :-)
20:59 TimToady though I cheated by fixing a few things in viv, like removing the one indirect object :)
20:59 TimToady and some duplicate package names that STD is pickier about than perl5 is
20:59 moritz but viv does use heredocs
20:59 TimToady abundantly :)
20:59 TimToady including ones without quotes
20:59 TimToady those parse now
21:00 pmichaud created nqp-2012.08.1 tarball, testing it now.
21:00 TimToady it now treats <<EOT as <<"EOT"
21:01 moritz jnthn: failing tests in t/spec/integration/error-reporting.rakudo
21:01 TimToady and I'm sure there are loads of bug-compatibilities that I haven't covered yet...
21:01 * lizmat has a warm fuzzy feeling again ;-)
21:01 jnthn moritz: Just that one?
21:01 jnthn moritz: Argh, it's a #icu test.
21:02 moritz jnthn: yes, just that one
21:02 stopbit joined #perl6
21:02 jtpalmer joined #perl6
21:02 moritz jnthn: but the failures don't look ICU-y
21:03 jnthn moritz: no, but that's the reason I don't see the errors
21:03 moritz http://perlpunks.de/paste/show/5041268e.7402.22b
21:03 jnthn moritz: No ICU build
21:03 TimToady so it might be more accurate to say that viv thinks it parses viv :)
21:03 jnthn moritz: hmmmmm.
21:03 lizmat fwiw, I'm looking at installing ICO on OSX and not liking the experience one bit
21:03 diakopter TimToady: the '->(' you added in STD_P5 - can you put a space inside that?
21:04 lizmat s/ICO/ICU
21:04 jnthn lizmat: Once upon a time Parrot bundled ICU. I spent ages getting it to build on Windows. Then they debundled it. I don't think I ever managed to build with ICU since.
21:04 TimToady that would be one of the bug compatability modes I spoke of :)
21:04 moritz jnthn: try:  ./perl6 -e 'use v6; class A { has $.x is rw }; A.new.x(42)'
21:04 [Coke] jnthn: invoke() not implemented in class '-5119310890791185874'
21:04 diakopter eval: my $a = sub { print $_[0] }; $a-> (3)
21:04 buubot_backup diakopter: 31
21:05 moritz it spits out four warnings
21:05 jnthn [Coke]: Wait, is that with latest NQP?
21:05 moritz and then think it's a runtime error
21:05 TimToady I think we'll probably reimplement most of ICU soon anyway...
21:05 jnthn [Coke]: I didn't bump NQP_REVISION again after the patch...
21:05 jnthn TimToady: Same ;)
21:05 [Coke] jnthn: ARGH.
21:05 [Coke] no, it was with latest rakudo. ;)
21:06 moritz latest rakudo doesn't have that patch yet
21:06 jnthn moritz: s/runtime/compile time/?
21:06 moritz because the patch was only in the nqp/toqast branch
21:06 moritz jnthn: erm, yes
21:06 jnthn moritz: e.g. it includes the SORRY when it shouldn't?
21:06 jnthn hm
21:06 [Coke] moritz: I'm IN that branch.
21:06 moritz oh
21:06 pmichaud nqp tarball passed
21:07 dalek rakudo/nqpqast: b79e158 | jnthn++ | tools/build/NQP_REVISION:
21:07 dalek rakudo/nqpqast: Another revision bump.
21:07 dalek rakudo/nqpqast: review: https://github.com/rakudo/rakudo/commit/b79e1588b3
21:07 jnthn [Coke]: ^^
21:07 jnthn moritz: I'm guessing something odd is going on with backtrace printing.
21:08 moritz jnthn: that's likely
21:08 lizmat wrt "- >", not allowing that in p5 on p6 would not be a big thing in my book
21:08 * pmichaud tests rakudo 2012.08 using new nqp tarball
21:08 moritz my $ x;
21:08 jnthn moritz: oh...I see something right away
21:08 * pmichaud wonders about a "std5:" target for p6eval :)
21:08 jnthn nqp::join(';', $bt[$_]<sub>.get_namespace.get_name),
21:09 jnthn 'nqp;HLL;Compiler'
21:09 jnthn moritz: NQP doesn't emit Parrot namespace directives any more.
21:09 moritz oh
21:09 TimToady std: use v5; / [a-z] /
21:09 p6eval std d93fffe: OUTPUT«ok 00:00 52m␤»
21:09 pmichaud or that.
21:10 sorear std: use v5; @ _
21:10 p6eval std d93fffe: OUTPUT«[31m===[0mSORRY![31m===[0mâ�¤Confused at /tmp/ZBuYlaBoPm line 1:â�¤------> [32muse v5; @ [33mâ��[31m_[0mâ�¤    expecting any of:â�¤   p5infixâ�¤        statement modifierâ�¤Parse failedâ�¤FAILED 00:00 52mâ�¤Â»
21:10 * flussence wonders if anyone's ever made a serious attempt to spec-ify p5
21:10 TimToady you'll note the caveat about bug compatibility
21:10 jnthn std: use v5; / [[] /
21:10 p6eval std d93fffe: OUTPUT«[31m===[0mSORRY![31m===[0m�Couldn't find terminator ]] at /tmp/GZOrbqm6WG line 1 (EOF):�------> [32muse v5; / [[] /[33m�[31m<EOL>[0m�Parse failed�FAILED 00:00 52m�»
21:10 jnthn ;-)
21:10 TimToady jnthn: that's fixed in the latest
21:11 moritz flussence: the perl5docs come as close as anything, I think
21:11 jnthn TimToady: oh...I looked at it and thought it wasn't
21:11 jnthn (as in, read the diff)
21:11 jnthn I musta missed something.
21:12 flussence oh right... I forgot how thorough the docs are :)
21:12 * moritz -> sleep
21:12 diakopter std: use v5.14; / [a-z] /
21:12 TimToady it's quantified with +, so it requires at least one charspec
21:12 p6eval std d93fffe: OUTPUT«ok 00:00 52m␤»
21:12 dalek rakudo/nqpqast: 119716f | jnthn++ | src/core/Exception.pm:
21:12 dalek rakudo/nqpqast: Fix is_runtime to unbust backtrace stuff in some cases.
21:12 dalek rakudo/nqpqast: review: https://github.com/rakudo/rakudo/commit/119716f349
21:12 jnthn moritz: ^^ should do it, guess I find out tomorrow ;-)
21:12 jnthn It helped with the -e line you gave me anyway :)
21:12 jnthn Sleep well :)
21:13 jnthn eval /[]/
21:13 buubot_backup jnthn: ERROR: Unmatched [ in regex; marked by <-- HERE in m/[ <-- HERE ]/ at (eval 20) line 1.
21:13 jnthn ah, and the empty one is invalid.
21:14 TimToady it even accepts / [^[a-z-] /
21:14 sorear what about []] ?
21:14 TimToady that too
21:15 TimToady what it isn't smart about is any backslash longer than .
21:15 * jnthn will have to compare it for what he did for Rakudo's m:P5/.../
21:15 TimToady so \xff will feyl
21:15 sorear (?{ print STDERR "foo" }) hehehe
21:15 jnthn Ah, I did handle that bit :)
21:15 pmichaud jnthn: is the 119716f commit something I need to worry about for the 2012.08.1 nqp release?
21:15 * sorear sneaks an indirect object into a regex to toy with TimToady
21:15 jnthn pmichaud: repo?
21:16 jnthn pmichaud: oh, the one I just did
21:16 jnthn pmichaud: No
21:16 pmichaud rakudo/nqpqast above.
21:16 pmichaud okay, just checking.
21:16 jnthn pmichaud: It's a fix required to nqp/toqast.
21:16 diakopter eval: use v5; / [^a--] /
21:16 buubot_backup diakopter: ERROR: Invalid [] range "a--" in regex; marked by <-- HERE in m/ [^a-- <-- HERE ] / at (eval 20) line 1.
21:16 pmichaud I'm spectesting rakudo 2012.08 now with nqp 2012.08.1
21:16 sorear eval: / (?{ print STDERR "foo" }) /
21:16 buubot_backup sorear: No output.
21:16 pmichaud so far looks reasonable.
21:16 sorear eval: / (?{ print STDERR "foo" @ }) /
21:16 buubot_backup sorear: ERROR: syntax error at (re_eval 21) line 2, near ""foo" @  ;" Compilation failed in regexp at (eval 20) line 1.
21:16 jnthn pmichaud: OK. Glad the segv got patched ahead of this release. :)
21:16 pmichaud yeah, me too.
21:16 jnthn Well, the Star.
21:17 TimToady eval: / [^ --] /
21:17 buubot_backup TimToady: No output.
21:17 jnthn I'd struggled to reproduce it here before, which made it a real pain to figure out...
21:17 TimToady diakopter: I suspect it doesn't like that a comes after -
21:17 diakopter std: use v5; / [^a--] /
21:17 p6eval std d93fffe: OUTPUT«ok 00:00 52m␤»
21:18 TimToady still the old std I suspect, unless someone has rebuilt it
21:18 diakopter std: use v5; / [---] /
21:18 p6eval std d93fffe: OUTPUT«ok 00:00 52m␤»
21:18 sorear new std has a hash startingwith b
21:18 diakopter eval: use v5; / [---] /
21:18 buubot_backup diakopter: No output.
21:18 [Coke] jnthn: now building rakudo, thanks.
21:19 pmichaud spectests pass with nqp-2012.08.1, pushing/publishing
21:20 pmichaud https://gist.github.com/3559196   # huh?!?!
21:21 pmichaud where's that 2012.11 tag coming from?
21:21 sorear check your reflogs?
21:21 sorear find out exactly when it was created
21:21 flussence I have it locally too
21:22 flussence aha
21:22 sorear I don't
21:22 flussence the date on it is 2011-12-22
21:23 flussence guess it's just a typo
21:23 TimToady anyway, to do cclass correctly probably entails copying over the cc role from STD, which postdates the original copy of STD_P5
21:23 pmichaud I don't know how to check the reflogs for tag creation
21:23 pmichaud but, more importantly, I want to get rid of it.
21:23 flussence `git show 2012.11`, `git tag -d 2012.11`, lemme figure out how to push tag deletions...
21:24 FROGGS git push --tags?
21:24 pmichaud I thought I had already deleted it before, though.
21:24 pmichaud that's why I'm surprised to see it come up again.
21:25 flussence FROGGS: not sure if it needs special syntax like for deleting branches remotely
21:25 pmichaud pmichaud@kiwi:~/p6/nqp$ git push origin :refs/tags/2012.11
21:25 flussence well, now I know :)
21:25 FROGGS right
21:25 pmichaud okay, it's now gone from github, at least.
21:25 FROGGS ;o)
21:25 pmichaud hopefully it won't show up again.
21:26 TimToady nap &
21:27 pmichaud 2012.08.1 now released.
21:27 jnthn \o/
21:27 jnthn pmichaud++
21:27 pmichaud now to build a rc5 candidate
21:28 dalek star: c53748c | pmichaud++ | tools/star/Makefile:
21:28 dalek star: Bump bundled nqp version to 2012.08.1 .
21:28 dalek star: review: https://github.com/rakudo/star/commit/c53748cf69
21:29 [Coke] jnthn: built.
21:29 masak morning, #perl6
21:30 masak ooh, a star is born. pmichaud++
21:31 masak oh wait. we're still on rc5.
21:32 pmichaud I'm about to do rc5, then test, then release.
21:32 masak nice.
21:33 pmichaud http://pmichaud.com/sandbox/r​akudo-star-2012.08-rc5.tar.gz
21:33 jnthn [Coke]: Great :)
21:44 pmichaud nqp just built on all of my platforms w/o segfaults
21:44 jnthn yay
21:44 pmichaud so, hopefully things are all good now :)
21:54 * pmichaud has 2 hrs until 2359 utc
21:55 diakopter TimToady: NFC specifies that a series of "orphaned" combining marks leading a string be left alone.. do you want those to combine to a synthetic, or leave them alone in case the string is appended to another, possibly forming a combinable sequence?
21:57 dalek star: c7cd324 | pmichaud++ | tools/build/module-install.pl:
21:57 dalek star: Skip blank lines and comment lines in input to module-install.pl.
21:57 dalek star: review: https://github.com/rakudo/star/commit/c7cd32438b
21:58 Guest9874 joined #perl6
21:59 Guest9874 hello
21:59 diakopter Guest9874: howdy
22:00 Guest9874 m super and u?
22:00 diakopter 3 o'clock and all's well
22:01 nuba joined #perl6
22:02 masak no pains or aches anywhere. mood: happy.
22:12 pjcj any qr/.*/ => sub { template "404" };  seems to do for 404 what error_template does for 500, but it seems that there should be a not_found_template or something
22:13 pjcj or am I barking up the wrong tree?
22:13 sorear qr/.*/ is Perl 5 syntax
22:13 pjcj ha, sorry, wrong channel
22:13 sorear masak: happy masak is best :D
22:17 masak \o/
22:24 * lizmat wishes pmichaud strength in getting rakudo * out in time
22:25 lizmat sleep &
22:25 masak oh right. new month.
22:26 pmichaud jnthn: ping
22:27 jnthn pongmichaud
22:27 pmichaud when I run perl6-debug from a dos cmd prompt window, I get the escape sequences instead of colors.  Should I set something differently?
22:28 jnthn pmichaud: For the demos I was running it in another shell.
22:28 pmichaud okay.  I suspect that will be a faq for this release then.
22:28 diakopter what shell?
22:28 jnthn ansicon
22:28 pmichaud (just so we're prepared for it)
22:28 jnthn yeah
22:29 jnthn https://github.com/adoxa/ansicon/downloads
22:29 pmichaud okay, I think I have the star tarball and .msi ready to publish -- any last blockers before I do that?
22:29 jnthn I wanna work out how Git makes it work on the normal console
22:29 diakopter windows 8 smartscreen blocked that app
22:29 jnthn diakopter: How...unsmart
22:30 jnthn what is smartscreen? :)
22:30 pmichaud it's just windows being "smarter than you" again.  :-P
22:30 jnthn Every release, there's more crap to disable... :-)
22:30 diakopter http://www.ghacks.net/2012/08/24/windows-8-sm​artscreen-filter-reporting-back-to-microsoft/
22:31 jnthn heh
22:31 jnthn Anyway, I should figure out a better way
22:31 jnthn And make a module that does the Right Thing.
22:32 diakopter EBADSHELL
22:38 pmichaud star tarball and .msi published to http://github.com/rakudo/star/downloads
22:39 jnthn pmichaud++
22:39 diakopter oops, smartscreen blocked that .msi too
22:39 jnthn smartscreen--
22:39 diakopter because it's not signed?
22:39 jnthn Guess I'll sticked with Win7 :)
22:39 jnthn diakopter: What does "blocked" mean?
22:40 jnthn Asks if you really want to install it?
22:40 diakopter full-screen darkened error, nope, no option to run it anyway
22:40 jnthn wat. :/
22:40 diakopter "Windows protected your PC"
22:41 jnthn Is it an IE thing or a Windows thing?
22:41 diakopter Windows SmartScreen prevented an unrecognized app from starting. Running this app might put your PC at risk. Then a link for More info, which says: Publisher: Unknown Publisher
22:41 diakopter oh wait. there is a Run anyway button
22:42 diakopter still, that error could be very scary
22:45 pmichaud oops
22:45 pmichaud need to create a new one; the announcement is wrong about Bailador.
22:45 * pmichaud does that.
22:46 diakopter pmichaud: the .msi doesn't appear to be signed. is there a way for you to do that?
22:46 diakopter I bet that would make smartscreen like it more
22:47 pmichaud diakopter: I have no clue about windows signing.
22:47 jnthn Not automatically
22:47 jnthn And signing costs.
22:47 jnthn http://www.hanselman.com/blog/UsingCodeS​igningCertificatesToSignDownloadedMSIsAn​dBuildReputationWithIE9SmartScreen.aspx
22:48 pmichaud and there's only so far that I'm willing to go in the name of bringing freedom to those who select products that limit freedom.
22:48 jnthn Indeed.
22:48 pmichaud the whole idea that everyone needs to check with M$ before installing a package... well, I disagree with that.
22:48 jnthn Same.
22:49 jnthn If somebody else wants to go through the cost/trouble of building a distribution that deals with these issues, fine, but it's not somewhere I feel like spending resources.
22:49 jnthn Well, that's why we have separate compiler and distribution releases.
22:50 jnthn Really though, fuck Microsoft for this one. :/
22:50 diakopter Adam signs his Strawberry Perl .msi
22:51 jnthn diakopter: Then there's hope somebody will make a signed MSI for Rakudo some day. :)
22:51 diakopter "strawberryperl.com - Open Source Developer"
22:51 diakopter kmx@cpan.org
22:51 diakopter a Certum Level III certificate
22:52 diakopter .. I thought it was adam..
22:52 pmichaud diakopter: I'm not saying I won't do signing.  I'm saying I don't know how, and that I'm not inclined to do a lot of research on it.  I already spent days figuring out how to create the .msi in the first place.
22:53 pmichaud if someone else will point me to a quick-and-easy signing resource, I'm all for it.
22:53 jnthn pmichaud: The resource I just linked to shows it's anything but quick and easy :(
22:53 diakopter I'll get a cert
22:53 pmichaud jnthn: yes, I know.  the same was true for putting together the .msi in the first place.
22:54 diakopter "The actual signing, once the cert shows up is not too hard."
22:54 diakopter (from that page)
22:54 diakopter a one liner
22:54 jnthn Getting the cert seems to involve more effort than for me to get a Swedish ID card. Something's weird. :)
22:54 dalek star: 9727c86 | pmichaud++ | docs/announce/2012.08:
22:54 dalek star: Remove "Bailador not working" entry from announcement.
22:54 dalek star: review: https://github.com/rakudo/star/commit/9727c8655c
22:54 jnthn Anyways, sleep time for me.
22:56 pmichaud I'm not sure I want to sign things using someone else's cert, fwiw; I want to make sure I know the details behind the cert.
22:56 masak 'night, jnthn, #perl6
22:57 diakopter pmichaud: I meant a new cert.. is that good enough?
22:57 pmichaud depends on how the cert is registered, I suspect.
23:01 sorear if we want to go through the trouble of getting code signing set up, it seems like it would be a lot better to have 'The Perl Foundation' rather than 'Matthew Wilson'
23:01 diakopter :) yeah
23:02 diakopter I bet we can do it
23:04 sorear fwiw, Apple has set up their own semimandatory code signing system for mac os 10.8
23:04 diakopter besides, it appears most of the code signing cert providers won't register one to an individual
23:05 * sorear wonders how happy code signining people are with bundling of 3rd party code like is done in Star
23:05 sorear hopefully they won't insist on certs for PaFo and all our module authors
23:07 dalek star: ba342ac | pmichaud++ | docs/announce/2012.08:
23:07 dalek star: More release announcement updates; fix some obsolete text.
23:07 dalek star: review: https://github.com/rakudo/star/commit/ba342ac4dc
23:07 dalek star: f6ef640 | pmichaud++ | tools/star/release-guide.pod:
23:07 dalek star: Typo fixes in release-guide.pod.
23:07 dalek star: review: https://github.com/rakudo/star/commit/f6ef640fc1
23:11 pmichaud sorear: yes, that was my thought -- something like "Rakudo" or "TPF" is better than "person who bought a certificate"  :-)
23:12 pmichaud TPF might have an issue with signing for third-party code, though.  So I'm not exactly sure what umbrella it should go under.
23:12 sorear afaik Rakudo is not a legal entity
23:13 sorear well, you're already signing for third-party code, the CLA is not an ownership assignment
23:13 diakopter I bet we could get one with perl6.org in the title
23:13 pmichaud Rakudo Star includes code that doesn't come from a CLA.
23:13 pmichaud and TPF would likely feel more comfortable about third-party code that comes via CLA.  I meant code that isn't via CLA.
23:13 sorear does stuff get bundled into star without first being signed off on by PaFo or you?
23:14 sorear (does it include an unreviewed ecosystem dump?)
23:14 pmichaud sure, most of the modules aren't via CLA
23:14 pmichaud heck, even nqp itself isn't via CLA
23:15 sorear I think then I am not quite understanding the constraints
23:15 sorear I was thinking this is a matter of "what if star contains bona fide malware, who do we fire?"
23:16 pmichaud that's not an issue I was thinking of, no.
23:16 pmichaud besides that, there's no "fire" involved except in a community sense -- there aren't employees involved here
23:18 diakopter if an incident occurred, more stringent controls would be set up; that's all
23:19 sorear this whole discussion is moot without a TPF representative.
23:19 sorear [Coke]: Who should we ping?
23:22 pmichaud I'm not sure I _want_ to go through TPF, so it's not something I want to be pushing them about yet.
23:23 yarp joined #perl6
23:23 diakopter chatting with DigiCert.. they will sell a cert only to orgs registered with the IRS
23:24 pmichaud registering with the IRS is no big deal; I have two already.
23:24 sorear diakopter: will they do DBAs or does it have to be "The Yet Another Foundation"?
23:25 sorear I beleive that's the primary IRS name of TPF
23:28 diakopter Yet Another Society
23:29 benabik joined #perl6
23:33 fgomez joined #perl6
23:47 lichtkind masak:
23:50 scottdware joined #perl6

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

Perl 6 | Reference Documentation | Rakudo | Niecza | Specs