Camelia, the Perl 6 bug

IRC log for #parrot, 2010-07-09

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:10 dalek parrot: r48058 | coke++ | branches/html_cleanup/docs/index/tools.json:
00:10 dalek parrot: fix malformed JSON spotted by kid51++
00:10 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48058/
00:11 nopaste "kid51" at 192.168.1.3 pasted "Patch to aid in diagnosing JSON formatting problems" (27 lines) at http://nopaste.snit.ch/21898
00:11 kid51 Coke:  please evaluate that patch (which is what I used to spot error)
00:31 kid51 The patch enabled me to figure out *which* .json problem had the problem.
00:32 kid51 Err:  The patch enabled me to figure out *which* .json file had the problem.
00:35 TonyC joined #parrot
00:39 nopaste joined #parrot
00:45 Coke kid51_at_dinner: looks good, but I'd remove the dump of the entire contents.
00:45 Coke otherwise, go ahead and apply that, thanks.
01:01 rurban_ joined #parrot
01:09 nopaste joined #parrot
01:12 s1n joined #parrot
01:18 TonyC joined #parrot
01:23 TonyC_ joined #parrot
01:25 nopaste joined #parrot
01:31 dalek rakudo: b30f2a6 | (Matthew (lue))++ | src/ (2 files):
01:31 dalek rakudo: Begin implementation of ::= binding.
01:31 dalek rakudo: Signed-off-by: pmichaud <pmichaud@pobox.com>
01:31 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/b​30f2a6808b99d6408b28e5b01fa8c8f60d0260e
01:31 dalek rakudo: c59e59a | (Matthew (lue))++ | src/core/operators.pm:
01:31 dalek rakudo: Removal of Q:PIR block in := implementation.
01:31 dalek rakudo: Signed-off-by: pmichaud <pmichaud@pobox.com>
01:31 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c​59e59a5e6b3079d2c8caf85b2a909cf30f73d14
01:31 dalek rakudo: 669c9cc | (Matthew (lue))++ | src/core/operators.pm:
01:31 dalek rakudo: Implementation of := binding.
01:31 dalek rakudo: Signed-off-by: pmichaud <pmichaud@pobox.com>
01:31 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/6​69c9cc132a034d4eb57c83ea7fa9c2009e03edd
01:31 dalek rakudo: a3d4b9e | pmichaud++ | src/core/operators.pm:
01:31 dalek rakudo: Add WHENCE code (originally from lue++, but patch didn't apply for some reason).
01:32 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/a​3d4b9e87a3ced6171c0fa2aad6065ca3189da99
01:32 dalek rakudo: c70469a | pmichaud++ | src/core/operators.pm:
01:32 dalek rakudo: A bound variable needs to adopt the 'rw' state of the thing it's bound to.
01:32 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c​70469aa5e9c4cbd6a4d3eab7f7195fef2ee5fbc
01:34 TonyC joined #parrot
01:42 plobsing joined #parrot
01:51 tcurtis joined #parrot
02:05 kid51 joined #parrot
02:11 jimk joined #parrot
02:13 jimk joined #parrot
02:22 dalek parrot: r48059 | jkeenan++ | branches/html_cleanup/tool​s/docs/make_html_docs.pl:
02:22 dalek parrot: In case of failure to decode JSON, identify which .json file failed.
02:22 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48059/
02:30 kid51 joined #parrot
02:31 kid51 r48059 qualifies as my very first Perl 5.12 code "in production"!
02:31 kid51 In the rest of Parrot, we're confined to 5.8.something.
02:31 kid51 At $job, 5.8.something.
02:31 kid51 In toolchain parsexs project, 5.6!
02:32 tcurtis Why do we get to use 5.12 in make_html_docs?
02:32 sorear 5.8.2 doesn't even build with modern gcc
02:32 kid51 So that patch to tools/docs/make_html_docs.pl, which calls 'use 5.12.0', is my first more-than-5.8 code accepted to run someplace other than my own homedirs!
02:33 kid51 tcurtis:  Umm, because Coke says so?
02:33 * kid51 ducks
02:34 tcurtis kid51: good enough for me.
02:34 kid51 But, in reality, it's mainly the release manager who has to run 'make html'.  It's mainly for docs.parrot.org, where we can do what we want without worrying about backwards compatibility.
02:35 janus joined #parrot
02:36 kid51 Not being able to use anything >5.8 'in production' means that I really have only a vague idea of the new syntax in 5.10 and 5.12 ...
02:37 kid51 ... other than 'say' (and that only because I used to maintain Perl6::Say ;-) )
02:40 Coke joined #parrot
02:44 somebody_ joined #parrot
02:47 somebody__ joined #parrot
02:51 TiMBuS joined #parrot
03:17 Mokurai left #parrot
05:00 hercynium joined #parrot
05:30 cotto ~~
06:11 uniejo joined #parrot
06:21 uniejo joined #parrot
06:34 eternaleye joined #parrot
07:43 snarkyboojum joined #parrot
07:48 dalek rakudo: 1086ff8 | moritz++ | t/spectest.data:
07:48 dalek rakudo: we pass 6 more test files
07:48 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​086ff8a04a3e9a1a0c7ba5a78e9107eb32d61a5
07:50 dalek parrot: r48060 | mikehh++ | branches/html_cleanup/tool​s/docs/make_html_docs.pl:
07:50 dalek parrot: perlcritic does not like code before use warnings
07:50 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48060/
07:56 fperrad joined #parrot
08:12 clinton joined #parrot
08:23 dalek parrot: r48061 | mikehh++ | branches/html_cleanup/tool​s/docs/make_html_docs.pl:
08:23 dalek parrot: fix copyright and remove unnecessary use strict (use 5.12.0 incorporates this)
08:23 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48061/
08:25 hanekomu_9 joined #parrot
08:51 cotto bacek, ping
08:52 cotto fun fact: one of the failing tests in the linscan branch doesn't actually use any registers
08:53 cotto yet somehow the generated pbc is different
08:53 cotto fun times
09:02 rurban_ joined #parrot
09:14 sorear pbc is horribly, horribly underconstrained
09:21 cotto I think it's actually my fault.
09:21 cotto the vanilla allocator allocates one register where mine does 0
09:31 bacek_at_work cotto, (barely here) pong
09:32 cotto What's !SUB ?
09:33 * cotto is also barely here
09:33 bacek_at_work cotto, PMC register for non-constant sub calls
09:33 bacek_at_work e.g. 'foo()' when 'foo' isn't in same PIR file
09:34 bacek_at_work 'foo()' translated to '$!SUB = find_sub_not_null "foo"; $!SUB()'
09:34 cotto ok.  My allocator isn't giving it a register because it doesn't appear to have a range.
09:34 cotto sounds like it'll need a special case
09:35 cotto or something else
09:36 bacek_at_work yeah...
09:36 cotto that gives me my next target once I've slept
09:36 bacek_at_work We do need two-pass Sub processing.
09:36 cotto yes
09:36 cotto or were you referring to something apart from the allocator?
09:37 bacek_at_work Not yet.
09:37 bacek_at_work But we can probably can use result of first pass for some optimizations
09:37 bacek_at_work ETOOMANYCAN
09:38 cotto making it work is a good first target
09:38 cotto sleep is a good zeroeth step
09:38 cotto 'night
09:38 bacek_at_work night
10:34 jsut joined #parrot
10:59 TonyC joined #parrot
11:05 TiMBuS|Away joined #parrot
11:10 nopaste joined #parrot
11:35 TonyC joined #parrot
11:39 mikehh Coke: you around?
11:39 nopaste joined #parrot
11:53 Mokurai joined #parrot
11:55 Mokurai left #parrot
12:01 kid51 joined #parrot
12:19 dalek rakudo: 92a210d | pmichaud++ | docs/spectest-progress.csv:
12:19 dalek rakudo: spectest-progress.csv update: 513 files, 34010 (84.7% of 40163) pass, 12 fail
12:19 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/9​2a210d6d2eb06699e715447d8625ddc17facaf6
12:19 nopaste joined #parrot
12:20 jsut_ joined #parrot
12:54 ruoso joined #parrot
13:16 gaz_ joined #parrot
13:24 patspam joined #parrot
13:33 tcurtis joined #parrot
13:37 dalek parrot: r48062 | coke++ | branches/html_cleanup/tool​s/docs/make_html_docs.pl:
13:37 dalek parrot: Fix cut-and-paste-o copyright line.
13:37 dalek parrot: review: http://trac.parrot.org/parrot/changeset/48062/
13:41 fperrad joined #parrot
13:46 Coke ~~
13:46 atrodo Good Morning Coke
13:47 mikehh hi Coke
13:51 mikehh Coke: re 48062, check PodToHtml.pm and HTMLPage.pm as well :-}
13:51 lucian joined #parrot
13:55 mikehh Coke: how do you intend to set {RESOURCES_URL}, {NAV_BAR}, title etc
14:00 bubaflub joined #parrot
14:19 patspam joined #parrot
14:22 gbacon joined #parrot
14:35 mikehh joined #parrot
14:41 * mikehh got to go out for a bit - bbl
14:41 Coke mikehh: was going to tackle that this weekend. =-)
14:42 Coke I realize it's just a braindead call to pod2html at this point.
14:42 allison joined #parrot
14:42 mikehh Coke: ok - been playing around with it a bit - work with you then
14:43 * mikehh fresh out of milk and need coffee - bbl
14:44 Coke ok. I will not be offended if commits happen while I'm at DAYJOB, really I won't. a little coordination before is better if possible,but i'll understand if I'm too idle. =-)
14:44 Coke mikehh: AIGH!
14:44 Coke mikehh: aren't you on GMT?
14:45 mikehh Coke: BST = GMT + 1
14:47 Coke close enough. =-)
14:47 Coke I've taken to saying "Eastern" for my time zone because I cannot be bothered to remember if it's EDT or EST.
14:48 dalek rakudo: 1ef278d | moritz++ | t/spectest.data:
14:48 dalek rakudo: run two more test files
14:48 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​ef278dd0c4bd64076448c55e87b40ff207d46bd
14:48 allison joined #parrot
15:13 TiMBuS joined #parrot
15:33 tcurtis How do I tell distutils to add my build/ directory to the library path when running tests?
15:39 mikehh joined #parrot
15:46 mikehh joined #parrot
15:48 cotto ~~
15:52 mikehh joined #parrot
16:04 Coke tcurtis: fperrad is the distutils guy.
16:05 tcurtis fperrad: ping.
16:07 fperrad pong tcurtis
16:11 tcurtis fperrad: I'm trying to use distutils to build my GSoC project now that I'm putting it in a separate repository, and I'm putting my build files in a build/ directory. How can I get distutils to add that directory to the library path when it runs tests?
16:18 fperrad tcurtis, with 'test_exec', you specify the executable + some options (like library path)
16:19 fperrad tcurtis, where is your new repository ?
16:19 cotto tcurtis, what's the advantage of moving your code to a separate repo?
16:20 Coke cotto: near as I can tell, because "bacek is lazy"
16:20 Coke ;)
16:20 tcurtis fperrad: thanks. http://github.com/ekiru/tree-optimization , but I haven't committed the distutils yet.
16:21 cotto The laziest approach seems like it'd be to leave it in svn, apart from having to work with svn.
16:21 Coke cotto: sorry. let me clarify:
16:21 Coke "bacek is the wrong kind of lazy"
16:21 Coke ;)
16:22 mikehh Coke: We should have a few more devs lazy like bacek++
16:22 tcurtis cotto: with it in a branch of parrot's svn, I have to merge trunk every-time there's a change in trunk that someone wants to use it with.
16:23 cotto e.g.?
16:23 purl e.g. is probably `exempli gratia', latin for `for example'
16:24 tcurtis For example, if there's a rakudo change that needs a later version of Parrot, moritz++ can't merge Rakudo HEAD into his branch until I merge parrot trunk into my branch.
16:24 cotto got it
16:24 cotto that makes sense and does indeed preserve laziness
16:25 fperrad tcurtis, I don't known if parrot-nqp has an option for library path
16:25 fperrad another way, it's to build *.pbc in ./ instead of ./build/
16:26 tcurtis And I expect that just being able to use a normal parrot install with it instead of having to get a branch will make it simpler for other people to try it out.
16:27 moritz in retrospect we could have thought of starting it in a repo of its own from the start :-)
16:27 cotto hindsight?
16:27 purl it has been said that hindsight is 20-20, cotto...
16:27 moritz I mean it's clear that such a framework doesn't really need many parrot changes
16:29 cotto I was just fishing  for that purl factiod
16:45 tcurtis fperrad: my question turns out to have been unnecessary. I accidentally left some of the pbcs out of inst_lib, and the tests run properly once I fix that.
17:02 tcurtis fperrad: Actually, that turns out to be wrong. I had forgotten I'd installed it in the interim.
17:04 rurban_ joined #parrot
17:12 dalek rakudo: 8978801 | moritz++ | build/Makefile.in:
17:12 dalek rakudo: compile Safe.pm, and install the .pir version too
17:12 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/8​97880192ed655b2f6933d180bcb972ed1a3bdc4
17:30 jan joined #parrot
18:09 LoganLK joined #parrot
18:24 tcurtis japhb++ for plumage automatically sudoing for install.
18:35 snarkyboojum joined #parrot
18:36 cotto_work Does plumage support fancy stuff like installing the latest build of a package from a vcs repo?
18:37 cotto_work and updating
18:37 tcurtis cotto_work: yes.
18:37 cotto_work shiny
18:37 cotto_work I'll have to start using it.
18:38 tcurtis On a related note:
18:39 tcurtis msg darbelo Can one of you, dukeleto, or japhb add my GSoC to plumage? The metadata is at http://github.com/ekiru/tree-optimizat​ion/blob/master/tree-optimization.json
18:39 purl Message for darbelo stored.
18:39 theory joined #parrot
18:39 tcurtis msg dukeleto Can one of you, darbelo, or japhb add my GSoC to plumage? The metadata is at http://github.com/ekiru/tree-optimizat​ion/blob/master/tree-optimization.json
18:39 purl Message for dukeleto stored.
18:39 tcurtis msg japhb Can one of you, darbelo, or dukeleto add my GSoC to plumage? The metadata is at http://github.com/ekiru/tree-optimizat​ion/blob/master/tree-optimization.json
18:39 purl Message for japhb stored.
18:39 cotto_work That project is going to be so added.
18:56 somebody_ joined #parrot
19:05 cotto_work anyone here familiar with Cyclone and its attempts to make C pointers safe?
19:08 cotto_work2 joined #parrot
19:08 cotto_work2 http://cyclone.thelanguage.org/wik​i/Cyclone%20for%20C%20Programmers
19:14 LoganLK joined #parrot
19:14 dalek rakudo: e58b166 | (Solomon Foster)++ | src/core/ (2 files):
19:14 dalek rakudo: Tweak Real.atan2 and Num.atan2 so they can handle $x's which are not of the
19:14 dalek rakudo: ideal type.
19:14 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​58b166d508acea739c863e944f05a2d3030bfc8
19:32 japhb tcurtis, I'll do it.
19:32 tcurtis thanks, japhb.
19:33 tcurtis msg darbelo japhb added my GSoC to plumage. Ignore my previous msg.
19:33 purl Message for darbelo stored.
19:33 tcurtis msg dukelete  japhb added my GSoC to plumage. Ignore my previous msg.
19:33 purl Sorry, I've never seen dukelete before.
19:33 tcurtis msg dukeleto japhb added my GSoC to plumage. Ignore my previous msg.
19:33 purl Message for dukeleto stored.
19:35 japhb pushed, enjoy!
19:36 particle where is the plumage metadata parser source?
19:37 japhb I'm wrapping the data_json Parrot language.
19:37 japhb Lemme find the module
19:38 japhb src/lib/Plumage/Metadata.nqp
19:39 japhb Interestingly, data_json is used to parse string inputs, and Config::JSON is used for file inputs.
19:39 particle ah, thanks.
19:39 particle do you need both, or is that only a refactor away from correct?
19:41 japhb I honestly don't remember.  It's been a long time since I touched that code.
19:41 japhb May have been limitations of one or both modules at the time.
19:41 particle no worries
19:43 cotto_work cotto@cotto-desktop:/usr/src/parrot/plumage\ 0 $ ./plumage help
19:43 cotto_work I don't know how to 'help'!
19:43 tcurtis japhb: The link was for the html page to view it on github. The actual text file is at http://github.com/ekiru/tree-optimizat​ion/raw/master/tree-optimization.json
19:44 tcurtis cotto_work: ./plumage usage
19:44 cotto_work sure
19:44 particle 'perhaps you meant usage' would be a nice addition there
19:44 japhb D'oh!
19:44 japhb That's what I get for doing a wget without looking at it.
19:44 particle ha
19:44 japhb Sigh.
19:45 tcurtis That's what I get for not specifying what exactly I'm linking to.
19:46 masak joined #parrot
19:46 dalek parrot-plumage: de66330 | japhb++ | metadata/tree-optimization.json:
19:46 dalek parrot-plumage: [METADATA] New project metadata tree-optimization.json; thanks tcurtis++
19:46 dalek parrot-plumage: review: http://gitorious.org/parrot-plumage/parrot-plumag​e/commit/de6633012150f6f351b56db05aba9d78d24c4fbd
19:47 japhb tcurtis, pushed again.
19:48 tcurtis japhb, thanks again.
19:49 japhb particle, re: 'perhaps you meant usage' ... one of the "after things Just Work" tasks (scheduled for Round 3 of the Grant Series That Didn't Happen) was to make a real help system for it.  So I was "leaving space" for a proper help command.  :-)
19:49 particle ah
19:50 particle well, maybe some one will beat you to a 'MooseX::App::Cmd' module in nqp
19:50 japhb Since it is seemingly unlikely that I will get to that bigger goal in the near future, you are free to make a stub help command as you suggest.
19:50 japhb Have you a commitbit?
19:52 particle not sure
19:53 particle not sure i have a gitorious acct
19:53 * japhb checks ... no
19:53 japhb (No you don't have a commitbit yet, don't know about the acct)
19:53 japhb Get one, and you can have one.  ;-)
19:53 particle i don't see an acct in my pw manager, creating...
19:55 cotto_work I note a distinct lack of PIRATE in plumage.
19:55 cotto_work "someone" may have to remedy that
19:55 japhb FEED ME METADATA ... NOMNOMNOMNOMNOM
19:56 baest joined #parrot
19:56 particle http://gitorious.org/~particle
19:57 eternaleye joined #parrot
19:57 dalek parrot-plumage: bc7f887 | japhb++ | metadata/tree-optimization.json:
19:57 dalek parrot-plumage: [METADATA] OK, let's try that again.  This time minus the HTML wrapper.  japhb--
19:57 dalek parrot-plumage: review: http://gitorious.org/parrot-plumage/parrot-plumag​e/commit/bc7f88716735b0e98bd5e12713440274b27a85ef
19:58 japhb You have the power!
20:00 particle i feel gitorious!
20:02 masak NotFound: ping
20:07 Coke (interestingly) ... please do not get me started. =-)
20:08 Coke (git git.... gitorious, gitorious)
20:12 masak anyone see immediately why the following PIR segfaults at line 25? http://gist.github.com/469980
20:16 masak I get the segfault regardless of whether I use pio.'print'(s) or pio.'puts'(s)
20:16 cotto_work masak, is that the complete code?
20:17 masak cotto_work: along with ./perl6 -e 'my $binary_out_fh = open("foo", :w, :bin); $binary_out_fh.write("föö".encode("ISO-8859-1"))' and the rest of Rakudo, yes.
20:17 Tene masak: change it to pio.'print'('OMG HAI')
20:17 purl Tene: that doesn't look right
20:17 masak but this is the relevant part.
20:17 masak Tene: ok.
20:17 Tene and check if it still fails
20:17 Tene I could imagine it segfaulting if there was something funny in s
20:18 Tene also try: say s
20:22 Tene masak: How do those fail (or not)?
20:22 masak still building.
20:22 masak still building your first suggestion :)
20:22 Tene bah
20:22 Tene You need a faster build.
20:22 masak ok, segfault on pio.'print'('OMG HAI')
20:23 Tene I keep forgetting that my work laptop is a bit atypical.
20:23 masak maybe there's something wrong with pio, then.
20:23 Tene Then there's probably nothing funny in s.
20:23 Tene Yeah.
20:34 Coke ah, that's not really a PIO object.
20:34 Coke from #perl6 - it's a container that has a different pmc data pointer. they need to unbox it before calling it.
20:34 masak pio = deref_unless_object pio
20:34 masak jnthn++
20:35 Coke (that being a custom p6 op)
20:35 ash_ left #parrot
20:42 somebody_ joined #parrot
20:49 cotto_work clock?
20:49 purl cotto_work: LAX: Fri 1:49pm PDT / CHI: Fri 3:49pm CDT / NYC: Fri 4:49pm EDT / LON: Fri 9:49pm BST / BER: Fri 10:49pm CEST / IND: Sat 2:19am IST / TOK: Sat 5:49am JST / SYD: Sat 6:49am EST /
20:50 ash_ joined #parrot
21:06 theory joined #parrot
21:11 Mokurai joined #parrot
21:15 somebody__ joined #parrot
21:16 LoganLK joined #parrot
21:19 Mokurai joined #parrot
21:25 dolmen joined #parrot
21:57 patspam joined #parrot
22:02 tcurtis How can I set something to PMCNULL in NQP? Do I have to use Q:PIR?
22:07 cotto_work my $p := pir::null__P
22:07 cotto_work ;
22:12 jnthn probably needs ()
22:13 tcurtis Turned out that nulling it wasn't what I needed anyway.
22:15 cotto_work () isn't needed in that case
22:23 patspam joined #parrot
22:43 dalek rakudo: 692aa15 | masak++ | src/core/IO.pm:
22:43 dalek rakudo: [IO] implemented IO.read and IO.write
22:43 dalek rakudo: There's still some mis-encoding somewhere, because a few tests still fail in
22:43 dalek rakudo: t/spec/S16-filehandles/io.t.
22:43 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/6​92aa15f2538858028934b8e26910199cc5fdc53
22:43 dalek rakudo: 2d0a573 | masak++ | src/core/IO.pm:
22:43 dalek rakudo: [IO] added :$bin parameter to &open and .open
22:43 dalek rakudo: When called, the file is opened without expecting utf-8. The argument
22:43 dalek rakudo: 'binary' to the .encoding method on the filehandle could actually have
22:43 dalek rakudo: been any non-'utf-8' string, since Parrot only has two modes right now.
22:43 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/2​d0a57359c94fdd97476ae69cf831b77ea2af62d
22:46 tcurtis Ah, the joys of accidentally using PIR dereferencing syntax in NQP.
23:12 TonyC joined #parrot
23:19 tcurtis pmichaud: ping
23:19 nopaste joined #parrot
23:21 tcurtis pmichaud: Does the pirop<tailcall> PAST::Op work?
23:43 Chandon joined #parrot
23:45 TonyC joined #parrot
23:46 nopaste joined #parrot
23:49 kid51 joined #parrot
23:55 Psyche^ joined #parrot

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

Parrot | source cross referenced