www.parrot.org | Parrot 1.6.0 "half-pie" released: The JIT is dead! | Testing priorities: Exception and MultiSub
Set by moderator on 20 September 2009.
00:02 Whiteknight joined 00:04 dalek joined
darbelo Whiteknight: ping 00:12
Whiteknight darbelo 00:14
pong
darbelo Do you still think the next_for_GC pointer should die? 00:15
From what I can see in the GC code it's used to keep a sort of 'priority list' for destruction, but that's basically invalidated by the way the list is handled. 00:18
Everything is just blindly pushed to the front, without really caring about 'priority'. 00:20
darbelo feels the warnock. 00:23
00:23 dalek joined 00:28 TiMBuS joined
Whiteknight darbelo: no, the comments are lying. next_for_GC *is* the GC marking mechanism right now 00:30
purl okay, Whiteknight.
Whiteknight it's not a priority thing
(sorry about the warnock, watching the season premier of House)
japhb It works! RAWK. Now for cleanup and tweak .... 00:31
darbelo Damned misleading comments! 00:32
00:33 Zak joined
Coke RT: 181 00:35
darbelo Coke++
Whiteknight that damned GC 00:39
purl rumour has it GC is the boehm conservative garbage collector at reality.sgi.com/boehm/cg.html or a really really bad perl "programmer" or GrandCentral.com or branches/gsoc_pdd09 or a travesty against god
00:40 mokurai joined
japhb Any Win32 people about? 00:41
What is the result of parrot_config osname on Windows? 00:42
darbelo mswin32 from what I can see in the Configure code 00:43
japhb darbelo, thank you 00:44
darbelo but I wouldn't rely on case, since I see a /mswin32/i there. 00:45
japhb ew 00:46
Well, one battle at a time
darbelo maybe it's just an overly paranoid check. 00:47
Whiteknight you can never be too paranoid 00:49
well, maybe not "never" 00:50
darbelo "It is as bad as you think and they are out to get you" 00:52
Whiteknight it's worse 00:53
darbelo goes on the hunt for foods. 00:55
See y'all later.
00:55 darbelo left
Whiteknight me too 00:56
later
japhb Tene, darbelo, pmichaud, Austin: Configure.nqp etc. pushed to gitorious.org/parrot-plumage/parrot...ees/master 01:11
I think I'm getting a lock on "Most NQP code that doesn't implement grammar actions" 01:13
And on that note, time to take a short break.
dalek ose: r117 | Austin++ | trunk/include (4 files):
Added some include files
ose: r118 | Austin++ | trunk/ (33 files):
Updated configurability. Added some utility classes.
ose: r119 | Austin++ | trunk/ (50 files):
.pm -> .nqp renaming
 01:26
ose: r120 | Austin++ | trunk/src/parser/ (2 files):
Renamed actions -> grammar_actions
 01:35
tracwiki: v22 | jimmy++ | GitObjections 01:43
tracwiki: add TortoiseGit
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
01:45 JimmyZ joined
JimmyZ TortoiseGit is the coolest Interface to (Git) Version Control 01:46
TortoiseGit?
purl TortoiseGit is code.google.com/p/tortoisegit/
Tene i've heard the name several times, but never any real mention of what it actually is. 01:47
Tene follows link.
I really can't see myself using an interface like that. I guess it's not for me. :) 01:48
JimmyZ purl: no, TortoiseGit is the coolest Interface to (Git) Version Control and porting TortoiseSVN to TortoiseGit, code.google.com/p/tortoisegit/
purl okay, JimmyZ.
JimmyZ Tene: just like TortoiseSVN, but for Git
TortoiseGit? 01:49
purl rumour has it TortoiseGit is the coolest Interface to (Git) Version Control and porting TortoiseSVN to TortoiseGit, code.google.com/p/tortoisegit/
Tene Yeah, I've never used TortoiseSVN.
cotto TortoiseSVN isn't bad if you're stuck on windows.
JimmyZ purl: no, TortoiseGit is the coolest Interface to (Git) Version Control and porting TortoiseSVN to TortoiseGit, see code.google.com/p/tortoisegit/ for more. 01:50
purl okay, JimmyZ.
JimmyZ TortoiseGit?
purl hmmm... TortoiseGit is the coolest Interface to (Git) Version Control and porting TortoiseSVN to TortoiseGit, see code.google.com/p/tortoisegit/ for more.
dalek kudo: 5d3d3a3 | pmichaud++ | Configure.pl:
Add --optimize to default --gen-parrot build.
 01:53
02:12 payload joined 02:15 shockwave joined 02:19 cognominal joined 02:20 Austin joined
kid51 Is it just me ... or does 'make test' seem to be running faster than last week at this time? 02:30
msg darbelo Refactored code within auto::frames::runstep so as to increase testability in t/steps/auto/frames-01.t; cf r41402. 02:31
purl Message for darbelo stored.
dalek rrot: r41402 | jkeenan++ | trunk (2 files):
Refactor some code within runstep() into internal sub

import statements not needed (yet) in this package.		 02:33
02:34 chromatic joined 02:41 janus joined 02:46 dukeleto_ joined
Austin sings ... so here we are, in the Tijuana jail ... 02:46
dukeleto_ Austin: damn, you must be having a crazier night than mee 02:50
me, even
Austin Duke: "refactoring" is a clever term for "I bet you wish you were smarter." 02:51
I've been refactoring for some time now.
02:53 rg joined
dukeleto Austin: throw in some beer and it sounds like a party 02:53
japhb msg darbelo Configure.nqp etc. pushed to gitorious.org/parrot-plumage/parrot...ees/master
purl Message for darbelo stored.
02:53 whoppix joined
Austin Hmm.. isn't there a "Ballmer number" in there somewhere? 02:54
japhb Ballmer number?
Austin Ahh, the "Ballmer peak"
xkcd.com/323/
Woo-hoo! It works. 03:07
03:17 zerhash joined
Austin Stupid programmer. 03:34
dukeleto PEBCAK 03:35
Austin Yep.
Grammar said "HLL". Grammar should have said, "<.HLL>"
03:38 tokuhirom____ joined 03:57 ysth joined
ysth threw up some parrot packages for ubuntu hardy, intrepid, jaunty, and karmic: launchpad.net/~sthoenna/+archive/parrot-ysth 03:58
was going to do rakudo, but that needs parrot_config to be included in one of the parrot packages to build. I'm assuming that would be parrot-devel? 03:59
dukeleto ysth: are you asking if parrot_config is only installed with parrot-devel?
ysth parrot_config is not installed at all, at least in any of Allison's divisions of packages. I am asking if parrot-devel is the one it belongs in. 04:00
thanks to Allison for doing the hard part of debianizing, by the way. 04:01
I even went to file a bug to add Artistic 2.0 to base-files common licenses and saw she'd beat me to it.
though it is mired in the bureaucracy of debian policy :( 04:02
dukeleto ysth: parrot_config seems to get installed for me 04:05
ysth not in the parrot packages 04:19
err, debian packages
04:19 tokuhirom____ joined
dukeleto ysth: hmm, i don't know squat about the debian packages. sorry. 04:20
ysth no, wait. it is there in Allison's packages; it must be something I did wrong?
purl okay, ysth.
ysth thanks, purl
purl sure thing ysth
dukeleto purl, go play in traffic
purl wanders off to dent some cars.
ysth maybe my vision is just failing. oh well, sorry for the noise 04:22
dukeleto ysth: no worries. I have a debian box at home that I can test stuff on if you need 04:23
ysth: but my main dev boxes for parrot are os x and freebsd
ysth is it normal for parrot_config to include a build dir path in libparrot_linkflags? 04:25
04:25 zak_ joined 04:44 ysth joined 04:51 dukeleto joined 05:07 ysth joined 05:15 masak joined
cotto anyone know if allison's going to be as #ps tomorrow? 05:21
dukeleto cotto: i think she said she would be traveling
but that was at the last #ps. things could have changed 05:22
chromatic She's at Linux Con, so she may or may not be present. 05:23
I'm at Linux Con, so I may or may not be present. 05:24
dalek ose: r121 | Austin++ | trunk/ (14 files):
Added Config.nqp to library for getting config data.
 05:25
bacek_at_work And I'm in Australia. So I will not be present on #ps or Linux Con.
sigh...
cotto bacek_at_work, I finally have some time to try out that pprof2cg optimization you suggested. 05:27
bacek_at_work cotto: that's good. 05:28
dukeleto chromatic: is there someone designated to precede over #ps in your and allison's absence? 05:30
Austin * precede -> preside ? 05:31
dukeleto Austin: yes please and thank you
dukeleto blames beer for any grammatical shenanigans
cotto I feel silly. I didn't think to use Devel::NYTProf on pprof2cg until just now. 05:33
chromatic dukeleto, not formally but if someone wants to, feel free. 05:34
cotto by default, it's the first person who asks who's leading 05:35
05:37 dukeleto joined
dukeleto cotto: good to know 05:47
dukeleto is diving down a wiki hole 05:48
05:52 kurahaupo joined, theory joined
cotto dukeleto, you can conveniently get around it by asking who's going to ask who's leading. 06:01
dukeleto cotto: that was the plan ;) 06:02
cotto what's the typical overhead for Devel::NYTProf?
06:02 uniejo joined 06:04 flh joined
dalek ose: r122 | Austin++ | trunk/library/close/Compiler/Messages.nqp:
Fixed up error-message formatting
 06:08
ose: r123 | Austin++ | trunk/src/parser/ (4 files):
Cleaned up include-directive parsing
ose: r124 | Austin++ | trunk/library:
Updated svn:ignore
TiMBuS cotto, from what i recall from one of the presentations it was like 2-4x slower. 06:12
cotto looks more like 13x, but that wasn't carefully controlled. 06:15
TiMBuS ouch
cotto meh 06:17
TiMBuS i guess it depends on the cpu, since the overhead is probably just a delay per op
are you worried about the speed? 06:18
cotto not very much. I just care about eventually getting the profile, which I now have.
TiMBuS ah
06:20 uniejo joined
cotto I'm surprised how poor the documentation on the C and C++ parts of GCSpy are. 06:21
oic. It's got some examples of how to use the api. 06:25
and now that I've got the profile, it's time for bed. 06:30
zzzzzzzzz
Austin What does the "DEVEL" config option signify? 06:36
dalek ose: r125 | Austin++ | trunk/ (3 files):
Moved String.nqp up in library tree
 06:40
ose: r126 | Austin++ | trunk/library/Config.nqp:
Normalized library/Config.nqp
 06:50
06:59 JimmyZ joined
japhb use.perl.org/~geoffrey/journal/39663 07:04
dalek kudo: e52ed31 | moritz++ | src/setting/Rat.pm:
.pred and .succ for Rat
 07:08
07:14 cotto_w0rk joined 07:22 donaldh joined 07:27 zak_ joined 09:45 bacek joined
Austin Bravo, japhb. Had I known you needed helper functions, I'd have pointed you at the close codebase. But I don't have anything like subst, and yours is definitely nicer than anything I'd have written. 09:46
dalek rrot: r41403 | mikehh++ | trunk/config/gen/makefiles/root.in:
fix some build and clean problems with make cover
 10:11
10:14 toyam joined
dalek rrot: r41404 | bacek++ | trunk (3 files):
[core] Proper fix for TT#1009 made after NotFound++ version
 10:23
10:24 Ron joined
mikehh All tests PASS (pre/post-config, smoke, nqp_test, fulltest) at r41403 - Ubuntu 9.04 amd64 10:45
10:55 DrForr joined
dalek ose: r127 | Austin++ | trunk/ (13 files):
Got cross-namespace symbols resolving
 10:57
11:05 nopaste joined
dalek rrot: r41405 | bacek++ | trunk/src/pmc/continuation.pmc:
[cage] Fix linelength codestd test
 11:07
11:11 bacek joined 11:12 szabgab joined
mikehh bacek: beat me to it :-} 11:14
bacek mikehh: :)
mikehh bacek: I did rerun codetest and test 11:15
bacek stealing karma from NotFound++ and mikehh++ :)
mikehh bacek: I am messing around with make cover at the moment - I think we need to run a basic make test as well as the rest as too many tests are not run 11:16
bacek: I also think that we need to apply that to fulltest as well 11:18
bacek "make cover" is just veeery slow. And not applicable to optimised builds. 11:20
s/And.*$// 11:21
mikehh bacek: :-} - it takes me about an hour to run - but it is missing library and compiler tests and debugger
bacek: in the non-optimised build the core-tests are the same as testb are they not? 11:23
bacek mikehh: no idea...
purl i heard no idea was jays.net/images/noidea.jpg
bacek mikehh: you can add explicit "cover-test" into Makefile 11:24
nopaste "bacek" at 114.73.160.179 pasted "Makefile patch for mikehh++" (23 lines) at nopaste.snit.ch/18021 11:25
bacek mikehh: with patch it should be same 11:26
Don't forget to run Configure.pl :
:)
mikehh bacek: yeah - that is what I am testing now 11:27
I think I may have to modify the COVER_DIRs 11:28
11:33 jsut joined 11:34 Whiteknight joined
Whiteknight good morning Parrot 11:37
Austin good morning, andrew 11:41
Whiteknight hello Austin 11:42
you've been on here bright and early the last few days
Austin close-parrot.blogspot.com/
I've been on overnight. I'll tootle off to bed in a few hours.
bacek Good evening Whiteknight 11:47
Whiteknight good morning bacek
bacek clock?
purl bacek: LAX: Tue 4:47am PDT / CHI: Tue 6:47am CDT / NYC: Tue 7:47am EDT / LON: Tue 12:47pm BST / BER: Tue 1:47pm CEST / IND: Tue 5:17pm IST / TOK: Tue 8:47pm JST / SYD: Tue 9:47pm EST /
bacek Evening! :)
Whiteknight Austin: very cool 11:48
mikehh well afternoon for me :-}
Austin And now I'm stuck in "no simple way to do that"-sville. 11:50
I need to automatically create a "void () :init :load {...}" sub for my namespaces. It'd almost be easier to create a string and recursively parse it. 11:52
Whiteknight be careful, that path is frought with evil and fail 11:57
Austin I'm not sure it is. Right now, it seems pretty easy - just lie about being inside an include file, and away I go. 11:58
Whiteknight ok
how is Close doing now? Get any of your bugs fixed? is it ready for testing? 12:10
12:11 bluescreen joined
Austin Close is doing pretty good. The blog post I just made reflects a *lot* of progress in the innards. Right now I've got a couple of the original test cases working, and I believe that most of the old test cases will fall out pretty quickly. I'm working on the generation-ordering issue today, and it's a straight-line problem, so it'll be done today. Close will probably be back to 0.1 functionality in 12:13
2-3 days.
Whiteknight awesome 12:15
Austin Actually, it is pretty cool. 12:16
dalek rrot: r41406 | jkeenan++ | trunk/t/steps/auto/frames-01.t:
Test all branches in internal sub _handle_exec_protect().
Austin The new innardz mean that tree-rewriting stuff, which was destroying me before, is a snap.
Whiteknight okay, how do I build it. It's saying it can't find parrot_config
Austin What is? 12:17
Whiteknight Configure.pl
Austin I have no idea.
Whiteknight do I need to install parrot?
Austin Apparently, I have parrot_config and you don't?
12:17 ruoso joined
Austin I have an installed parrot, if that matters. 12:17
Whiteknight okay, let me try thtat
actually, looking at the Configure.pl source, I think I can write --parrot_config=.. 12:18
12:19 payload joined
Whiteknight nope, need to install it 12:20
Austin Did you see jonathan's Configure.nqp post?
Whiteknight no
Austin gitorious.org/parrot-plumage/parrot...figure.nqp 12:22
Maybe not jonathan. 12:23
Right. Geoffrey, aka japhb 12:24
The j's got me.
use.perl.org/~geoffrey/journal/
Whiteknight some of this Close syntax looks like really old C 12:25
like pre-ANSI C
Austin Could be. What are you looking at?
Whiteknight right now I'm looking a test.clo 12:26
Austin If it has lots of externs in it, it's from a very early version of the language, where storage classes were mandatory
Whiteknight mostly looking at how the parameters are defined after the ()
Austin Well, that's defunct.
Whiteknight oh really? Is there something more recent I can look at? 12:27
Austin t/language/*
Whiteknight nice 12:31
it really is just a better way to write PIR
if the runtime libraries stay small enough and optional to load, it will easily replace all my PIR work 12:32
12:32 tetragon joined
Austin Actually, the runtime libraries will probably grow a lot. For some value of runtime libraries. 12:33
But like C, they'll be selectable.
-lregex, or whatever
Whiteknight gotcha 12:34
and that's fine too
Austin You have any ideas on how to do function inlining? 12:35
12:37 kj joined
Whiteknight Austin: a few 12:38
Austin I'd love to hear 'em.
Whiteknight I don't have time today, but we should definitely chat about it soon
I actually have to start getting ready now, have a funeral to go to at noon 12:39
Austin Yow!
purl Will the third world war keep "Bosom Buddies" off the air?
Austin I'm sorry to hear that.
I hope it's not yours.
Whiteknight no, I'm not dead yet
Austin That's what they all say.
Whiteknight seems a little redundant, if they're saying anything at all 12:40
okay, I'm out of here. I'll talk with you later.
12:52 bluescreen joined 13:44 JimmyZ joined
NotFound bacek++ good fix to my fix 13:53
dalek TT #1057 created by NotFound++: Rename non static list functions 14:04
rrot: r41407 | mikehh++ | trunk/config/gen/makefiles/root.in:
add test to make cover and clean cleanups
 14:05
14:36 Psyche^ joined 14:46 dalek joined 14:49 AndyA joined
Coke ponders writing a patch for rakudo and sees that it was just a stale ticket. 15:03
15:07 Andy joined 15:09 shockwave joined
Coke Andy: ping. 15:22
Andy tessir 15:23
yessir
Coke question about ASSERT_ARGS
e.g. on Parrot_on_exit, we assert_args there; but the assertion is for the interp OR the function; shouldn't that be an AND?
Andy uh 15:24
looking
Coke I can nopaste the bits I'm looking at.
Andy holy shit 15:25
purl only in the Vatican, my friend.
Andy you are right.
Coke I think that explains the few --optimize failures we've been seeing.
Andy ok
And probably many undef checks 15:26
rather
many warnings that something's not going to get executed
Coke I'm happy to let you patch it. =-)
Andy I didn't write ASSERT_ARGS et al
but I'm horrified I didn't notice
Coke ah.
I think I see it in headerizer.pl ... testing... 15:27
Andy ok
Coke that's a few tickets I don't have to port over. :) 15:28
(running make test now. will commit when it finishes) 15:30
Andy++
Andy I didn't do anything.
I wonder what the splint will look like now 15:31
15:33 chromatic joined 15:34 JimmyZ joined, elmex_ joined
Andy make test passed for me with the one change in that one line the headerizer: s/\\Q||/&&/ 15:34
Coke: I'm out of this, then, unless you need something lese. 15:38
Coke Andy: yup. Thanks for the sanity check! 15:44
commit landing shortly.
(big commit)
Andy++ #sanity 15:45
Andy I try to provide what sanity I can. :-)
chromatic I wonder what that will do to optimized builds.... 15:46
Coke I just built optimized.
but I've been doing that for some time.
that had to have been the original intent of assert_args, so I'll hope, "the right thing". =-) 15:47
dalek rrot: r41408 | coke++ | trunk/tools/build/headerizer.pl:
When ASSERTing args, ASSERT *ALL* of them, not *ANY* of them.

to optimization, and perhaps some of Andy++'s splint warnings.
Coke dalek?
purl i guess dalek is #parrot's spammy little rss bot or (see: dalek plugins)
rrot: r41409 | coke++ | trunk (118 files):
rerun make headerizer
15:48 mokurai joined 15:50 tokuhirom____ joined 15:57 fperrad joined
Coke waits for the bug reports to flow in. 15:58
16:03 quek joined
cotto_work cil? 16:25
purl cil is probably the .NET assembly language.
cotto_work cil is also C Intermediate Language or hal.cs.berkeley.edu/cil/
purl okay, cotto_work.
cotto_work anyone seen that?
16:27 flh joined
dalek kudo: 84981fd | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 436 files, 15498 (71.4% of 21695) pass, 0 fail
 16:32
16:33 bluescreen joined 16:36 quek left 16:41 kj joined
mikehh All tests PASS (pre/post-config, smoke, nqp_test, fulltest) at r41409 - Ubuntu 9.04 amd64 16:51
16:53 theory joined 16:55 darbelo joined 16:58 theory joined 17:00 smash joined
smash hello everyone 17:00
everyone hello smash 17:01
17:01 theory joined
Coke smash++ 17:02
hulk?
purl Hulk SMASH!!!
Coke mmm.
17:04 slavorg joined 17:07 MoC joined
smash Coke: every ticket i own on RT is either resolved or rejected 17:10
do u want me to do anything with them ? 17:12
Coke smash: no. only open tickets matter. 17:18
but, if you want to clean up a dozen more, I'd appreciate it! =-) 17:19
RT: 179
17:28 theory joined, AndyA joined 17:30 ysth joined
dalek TT #1009 closed by darbelo++: examples/shootout/spectralnorm.pir shows a segfault in the continuation ... 17:37
17:38 chromatic joined 17:46 joeri joined
dalek tracwiki: v23 | fperrad++ | GitObjections 17:49
tracwiki: trac.parrot.org/parrot/wiki/GitObj...ction=diff
17:52 kthakore joined 18:05 darbelo left, kthakore left 18:06 kjeldahl joined 18:09 kyle joined, darbelo joined
dukelet0 ps? 18:14
purl ps is postscript or process status or see "parrotsketch" or non-vector?! or annoying.
chromatic 17 minutes.
dukelet0 chromatic: danke
chromatic: how is linuxcon?
chromatic Quiet this morning. I'm about to go into another session; we'll see how interesting it is! 18:15
dukelet0 chromatic: will you be at #ps?
chromatic I'll do my best. 18:16
18:20 chromatic joined 18:25 allison joined
Util mikehh: you reported that you intend to "get fulltest to run all tests". Which tests are not running in fulltest? 18:28
dalek rrot: r41410 | NotFound++ | trunk/t/pmc/exception.t:
[t] add 2 Exception tests to cover init_pmc
mikehh Util: nqp_test, library tests and most compiler tests 18:29
Util mikehh: ok, thanks
mikehh time for #ps
darbelo dukeleto++ # Documenting the heck out of the debugger 18:51
18:52 zak_ joined
kyle in what namespace would I find a perl6 global sub? 18:57
pmichaud depends on what you mean by "global" :-) 18:58
by default things start out in the CORE::GLOBAL namespace
kyle ok, yeah, I'm just writing "sub foo{}", not inside a package or anything.
TimToady note that that declaration will likely default to "my" in Perl 6 18:59
and also that normal sub dispatch never looks in packages anymore
(NYI in rkudo)
kyle alright 19:00
darbelo NotFound: ping me if you need any help setting google code up. 19:06
purl I can't find me in the DNS.
NotFound darbelo: thanks
darbelo karma kill 19:13
purl kill has karma of 6
darbelo karma obliterate
purl obliterate has neutral karma
darbelo karma exterminate
purl exterminate has karma of 1
19:17 ruoso joined
darbelo dukeleto, dukelet0: ping 19:18
kyle if I'm writing PIR, how do I lookup my p6 sub "foo"? something like get_hll_global ["CORE";"GLOBAL"], "foo" ? (that is returning a null pmc) 19:24
Tene kyle: get_hll_global 'foo' 19:25
kyle Tene, thanks 19:34
19:37 theory joined
Tene japhb: You have information for me on what needs to happen to get DB stuff in plumage yet, or do I need to just read the source? 19:43
japhb Tene: Will you be around for a little bit? I'm mentally in the middle of something, but I should be wrapped up in :05-:15
Tene japhb: I should be around on and off for at least four hours. 19:44
japhb Excellent, thanks.
fperrad ping, dukeleto 19:47
dukelet0 fperrad: pong 19:51
fperrad duketel0, in Blizkost, I create a branch load_language 19:52
tests pass, but with a non zero exit status
could you take a look ?
japhb Tene:OK, near as I can manage, you have my full attention. What can I do for you?
nopaste "darbelo" at 200.49.154.173 pasted "t/tools/parrot_debugger.t failure on OpenBSD i386" (29 lines) at nopaste.snit.ch/18030
darbelo dukeleto: ping 19:53
19:53 bacek joined
Tene japhb: I want to put DB libraries in PIR somewhere they can be useful. AFAICT, plumage is supposed to become part of that. 19:53
japhb Tene: Plumage will be used to fetch, build, and install modules into Parrot. I'm very much thinking that 19:54
where it pulls the source from doesn't matter. In other words,
Tene agrees so far. 19:55
japhb As long as the metadata points to a live URI, and Plumage has been taught about the URI scheme in question (git:, svn:, ftp:, what have you), it should Just Work.
darbelo japhb: you are reinventing the BSD ports system, right?
19:56 theory joined
japhb darbelo, I'm reinventing a bunch of things like it, yes. 19:56
dukelet0 fperrad: will check it out
Tene japhb: So I need to work out how the files shoudl be arranged in the repo and what the metadata should contain, yes?
darbelo The basic: Automated "fetch source, build, install" workflow.
japhb darbelo, hopefully I will have something to add to the mix, but at the very least, we will eventually have a ports system that depends in the core only on Parrot-provided functionality. 19:57
darbelo, sure.
There's more to it than that, but yes, that's the basics.
darbelo I wasn't being pejorative, I'm all for what you are doing. Just trying to map it to something I know.
japhb Tene, a week ago I started roughing out the metadata for Blizkost, but since I haven't actually gotten to the point of being able to do the install, I don't know if it's good enough. 19:58
But you can look at it to see where I'm coming from.
Tene japhb: that would be great.
japhb (I already have a couple ideas for improvements, but nothing we can't talk through.)
Tene ah, metadata/blizkost.json 19:59
What can plumage do so far?
japhb Tene: Look in ... yep, that one.
Tene, it builds successfully into a fakecutable, which knows a couple simple help commands, plus 'info', a command which just dumps the metadata for a given project. 20:00
Adding the 'fetch', 'build', and 'install' commands was literally the next day I had planned .... 20:01
... until you changed that plan yesterday.
:-)
Tene :)
Are you planning to just have it install to wherever it wants, or have plumage pass an install prefix?
japhb Part of the deal with doing this all in NQP and PIR is that it's like a rocket taking off. The first few stages expend tremendous effort without seeming to get very far.
Tene Yes, understandably. :) 20:02
darbelo japhb: What are your plans for 'fetch' call an external tool?
Tene So, looks like plumage needs more work before it can be useful here, so now I'm going to ask for you to give me a plumage task to work on tonight. 20:03
japhb Tene, yes, Plumage will try to pass configuration information during Configure/build phases. But that's not something I'd planned on my next hack day.
20:04 chromatic joined
japhb darbelo, for now, yes. The draft spec states that the metadata must list all external tools required. Including the fetch tool. 20:04
That may be something I can relax later, but for now, it seems wise to be specific.
Tene, OK. Give me a couple minutes to think. 20:05
Tene japhb: no pressure. I'm still at $dayjob for another few hours. :)
darbelo Sounds good for the 'grab a snapshot out of my VCS' case, but you might want to leave some slack in the generic http/ftp tool.
Tene darbelo: 'http' could be considered a "fetch tool", in the worst case. :) 20:06
darbelo Yeah, but it means a call to a different executable on just about every platform. 20:07
NotFound japhb: the MySql module right now will just need to put the .pir file in his place and compile it to .pbc
No need to complicated build or configure stages. 20:08
japhb darbelo, that bridge was already crossed with the name of 'make'. You're specifying a virtual package name. 'git' would be answered by 'git-core' on some platforms, and 'make' by 'gmake' on others, for instance. 20:09
(That bit has not been made explicit in the spec, it was still in my head, sigh. ENOTENOUGHHOURS.)
darbelo japhb++ 20:10
japhb NotFound, I thought of that case. Basically empty stages specified in the metadata, and install phase is a manifest-based install 20:11
NotFound japhb: also, it may be the easier module to check basic functionality. 20:12
japhb NotFound: on phone for a minute, brb 20:13
NotFound NP
Coke fperrad: at a guess, use an explicit 'exit' 20:14
20:15 Zak joined
dalek rrot: r41411 | NotFound++ | trunk (3 files):
[core] rename list_delete and list_length functions, TT #1057
 20:15
rrot: r41412 | NotFound++ | trunk/examples/nci/mysqltest.pir:
[examples] unskip the test of MySql prepare
 20:22
japhb Yipes. 20:26
NotFound, OK, bak. Now what were you saying again? 20:27
dalek rrot: r41413 | darbelo++ | failed to fetch changeset:
Remove a duplicated ASSERT_ARGS()
NotFound japhb: I'm thnking about using the MySql module as a simple module to check plumage basic functionality 20:28
Tene NotFound: can you find out about getting a /modules on svn.parrot.org ?
japhb NotFound, oh yes, right. Gotcha.
NotFound Tene: I'll try to set up an svn repo at googlecode 20:29
Tene Okay.
japhb commit bits for the parrot-plumage repo freely given
Tene japhb: please add 'tene'
NotFound japhb: What can I do to make that easy? Some directory structure? 20:30
japhb
.oO( "You see? You ask for help, and what does it get you? Help, that's what." )
Tene NotFound: I assume that basic PIR modules will be pretty easy... just manifest-based install.
so once we figure out where to install to, and the manifest format, we'll just adjust the repo accordingly 20:31
darbelo manifest-based?
japhb NotFound, yeah, having a MANIFEST and a json file of metadata would be a good start
darbelo Who maintains the manifest? plumage or the module?
Tene ideally, either. 20:32
either give a path in the fetched repository or inline it in the plumage metadata, IMO
japhb darbelo, I've been of the opinion (and some have disagreed) that Plumage ought to be able to make the easy case easy by providing a builtin "build/install" for something that just needs files copied into place (optionally with compiling PIR down to PBC or suchlike) 20:33
And what Tene just said.
NotFound Some idea for a name? parrotmysql? mysqlparrot?
japhb parrot-mysql
Tene japhb: people disagree with that? o.O
darbelo decnum-dynpmcs has nothing like a manifest, it simply shuffles files about. Is that a supported operating mode?
japhb parrot-foo seems a good convention for things that don't have an overriding project name (like 'partcl') 20:34
20:34 bluescreen joined
japhb Tene: Among others, ISTR the #toolchain folks thought it was ... less than wise. 20:34
I didn't agree.
Tene Huh. Any idea what their reasons were?
japhb Honestly, I'm afraid I've forgotten. It was in the middle of a large discussion. 20:35
Tene s'okay
I can dig.
if I care enough. ;)
NotFound Just created code.google.com/p/parrot-mysql/
japhb darbelo, probably just put the proper shuffles in the metadata, or create a MANIFEST and do it that way. 20:36
Oh! Speaking of this general topic, I had a question for the interested parties:
Tene Lime green.
japhb Many package/ports-style systems ...
Tene Oh, I'll wait for the question first...
japhb ... *chuckle* ...
dukelet0 msg fperrad "make test" passes on feather in your blizkost "load_language" branch
purl Message for fperrad stored.
japhb ... have a standard way to apply a patch against the upstream source between the fetch and configure/build stages. 20:37
bacek Good morning
purl And good moroning to you, bacek.
Tene I think that's great.
japhb I think that's useful, but I'm not sure it's useful *yet*. Might help with portability. (System-specific patches, that sort of thing.)
darbelo Ah, yes, the good ol' patches/
Tene Also nice for generating srpms with patches separate from original source. 20:38
japhb Anyway, was looking for a little feedback on whether to have that in the design.
Sounds like a +1 from Tene
Tene +1, yes
darbelo +1
purl 1
darbelo kicks purl
purl bites darbelo!
cotto_work rebootening tiem 20:39
20:41 kjeldahl joined
darbelo japhb: You don't need to have it *working* right away, but it's very helpful. 20:41
Tene Yes, it's very much not urgent at all, but should come in eventually. 20:42
japhb Just added to TODO, committing ... 20:44
... done
pmichaud "rough consensus and working code"
japhb Task for anyone: Get dalek to speak Gitorious.
I will do it eventually if no one else does, but it would be happy making and non-distracting if I didn't have to. 20:45
:-)
Tene japhb: I'm going to go do some stuff for $realjob... if I could have comments on IRC or email on what I should work on tonight, that would be great.
AFK
japhb Tene, will do. later. 20:48
$life just got in the way, AFK for a while.
20:49 cotto_work joined
bacek NotFound: ping 20:53
NotFound pong 20:54
bacek NotFound: you renamed list_delete to Parrot_array_delete, not Parrot_pmc_array_delete. Is it intended?
NotFound Uh... no.
bacek No worries. I can fix it :)
NotFound Doing too much things at the same time
Thanks :) 20:55
20:55 payload joined, chromatic joined
NotFound A copy of current mysql files in examples is now in code.google.com/p/parrot-mysql/ 20:57
bacek NotFound: done 21:03
NotFound bacek++ 21:04
bacek waiting for svn... sigh...
darbelo goes deface the parrot-mysql wiki 21:06
dalek rrot: r41414 | bacek++ | trunk (4 files):
[core] Rename list_foo functions into Parrot_pmc_array_foo. Closing TT#1057
 21:08
TT #1057 closed by bacek++: Rename non static list functions 21:10
NotFound bacek++ verified that mysqltest.pir keeps working 21:13
bacek NotFound: of course! :)
NotFound Verifying does not harm ;) 21:14
dalek tracwiki: v1 | NotFound++ | Modules 21:15
tracwiki: trac.parrot.org/parrot/wiki/Module...ction=diff
tracwiki: v101 | NotFound++ | WikiStart
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
21:18 iblechbot joined
dalek rrot: r41415 | bacek++ | trunk/compilers/imcc/imcc.y:
[cage] Shut one more compiler warning in IMCC
 21:24
21:27 joeri left
cotto_work bacek, make sure those changes get propagated to the generated files. 21:28
bacek cotto_work: oh shi... I forgot --maintainer in Configure.pl args... 21:29
japhb bak 21:33
Tene: do you have a particular short-term goal you want to achieve, or just "help move things along"? 21:34
Tene japhb: My nearest goal is "Put DB and other simple PIR libraries online and be able to give someone else reasonable instructions to use plumage to install them" 21:38
between here and there... eh, whatever. 21:39
japhb OK, gotcha.
Just trying out this project management aikido thing ... :-)
NotFound I share that same goal 21:40
dalek rrot: r41416 | bacek++ | trunk/compilers/imcc/imcc.y:
Revert previous commit. Bison 2.4 (not 2.3) dislike it. cotto++
mikehh hey bacek - you blew me up there - I tried to build at r41415
bacek mikehh: bison 2.3 didn't complain about first changeset :-/ 21:41
NotFound Being able to get postgres and SQLite out of the repo will be nice. 21:42
pmichaud +1
purl 1
davidfetter :/
what happens to them then?
pmichaud they get their own repos 21:43
davidfetter ah :)
NotFound davidfetter: the goal is to simplify parrot config, each module taking care of his own needs.
bacek pmichaud: btw, did you try "git pull --rebase"? It will avoid annoying "merged branch master" commits.
pmichaud bacek: I'm not annoyed by them. :)
davidfetter NotFound, cool :)
bacek pmichaud: fair enough :) 21:44
pmichaud I've been idly thinking of moving nqp into its own repo as well.
bacek pmichaud: no!!! I'm going to use it more heavily for parrot-related stuff. 21:45
davidfetter the down side of that is that there's no automatic way to know whether changes in parrot broke stuff elsewhere
pmichaud bacek: yes, but parrot's deprecation policy is going to be a major break/impediment to nqp devel
s/break/brake/
bacek sigh... 21:46
pmichaud so if I do nqp devel elsewhere, I don't have to worry about backwards compat, and parrot can pull whichever nqp sources it wants
NotFound davidfetter: I suppose that we take care of testing modules owned by PAFO
pmichaud afk # errand 21:47
davidfetter how would that work?
dalek kudo: c584dc6 | moritz++ | src/setting/Operators.pm:
remove some duplicate code in infix:<...>
 21:48
kudo: 0220cc2 | moritz++ | src/setting/Operators.pm:
implement infix:<eqv>(Rat, Rat)
NotFound davidfetter: same as other things, servers setting up to periodically update and launch some tests.
davidfetter do they notify about breakage in a way that's not going to cause people to pipe them to /dev/null ? 21:49
NotFound davidfetter: we are just starting, that things needs to be done when things are walking. 21:50
davidfetter groovy
NotFound We need to have a plumage able to download the modules, at least.
japhb NotFound: I'm currently trying to prep the repo for incoming helpers, because it looks like the time to act is now. 21:52
mikehh bacek: compiling changes compilers/imcc/imcparser.h and compilers/imcc/imcparser.c for me 21:53
japhb tries not to stress about impending day-job related loss of sleep.
NotFound, in any case, hopefully you'll have that soon.
mikehh should I say building
svn status is now out of sync 21:55
NotFound japhb: for the MySql module, no urgency, is still very incomplete.
mikehh I get an svn diff on those files 21:56
NotFound mikehh: note that different versions of bison or lex give slightly different results. 21:57
Cosmemtic, mostly.
bacek mikehh: git reset HEAD (or however it's called in svn) 21:59
mikehh I am going to do a clean checkout - but later - I need some sllep 22:02
as you may notice (sleep) 22:03
BTW there is a codetest failure (line length (2) in src/list.c - but I can't commit at the moment until I get sorted out 22:04
darbelo mikehh: I'll get it. 22:05
japhb Brain fail: What is the term for a type of test that is exploratory? The little scripts you write when you just want to see if something is feasible, and generally how you would do it, before you even start on the real prototype? 22:06
darbelo proof of concept?
proto-prototype? 22:07
1.0?
purl i heard 1.0 was 1 is true or SOOO 2006 or “Shipping a 1.0 product isn’t going to kill you, but it will try.” www.randsinrepose.com/archives/2006...20/10.html or just a version number
dalek tracwiki: v2 | darbelo++ | Modules 22:08
tracwiki: Added decnum-dynpmcs to the list of modules.
tracwiki: trac.parrot.org/parrot/wiki/Module...ction=diff
japhb darbelo, that's close, but I think of that as something used to prove some entire project is doable. I'm talking about some little piece of it. Like "How do I do async DNS lookups on this platform" when you're really writing a client app for some protocol.
mikehh I think you are supposed to run a line renumber thinggy or something
japhb
.oO( Proof of Shaved Yaks )
 22:09
mikehh anyway nnl
you see my typing has gone to pot - bbl
japhb
.oO( scribbles? doodles? )
 22:10
darbelo throwaway?
japhb not bad 22:11
darbelo I think that is about as far as my command of the english language can carry me. 22:13
japhb darbelo, no worries, I think we're pretty close. I just copied my question to #perl6 to see if the linguists had an suggestions. 22:14
darbelo (Asking the natives)++ 22:15
japhb :-)
Tene, 'tene' has been added as a parrot-plumage committer. 22:23
Was there anyone else asking for same, while I'm in that dashboard?
darbelo me 22:25
japhb same nick?
darbelo But I'm not on gitorious. So you can get that later.
japhb Ah, OK
Austin Speaking of le plumage, I have a vaguely unrelated question. 22:26
japhb heh
Austin Has there been any thought given (yet) to the installation of parrot extras?
japhb blinks
Austin That is, stuff built on parrot but not part of parrot.
(Like rakudo, I guess) 22:27
darbelo Austin: the plumage is all about that.
Austin Yeah.
And?
Where do I sign p?
*p = up
japhb Austin, I'm not sure I'm understanding your actual question. Since what you're talking about is a major goal of Plumage ...
... ah, OK. 22:28
Austin Right. Is there a manual, or what?
darbelo or what.
japhb Hold, fetching URL
darbelo for now.
Austin darbelo :)
japhb Austin, start here: trac.parrot.org/parrot/wiki/ModuleEcosystem
Other than that, read the SHOUTING FILES in the repo.
speaking of which ... 22:29
purl speaking of which ... is anyone working on dbdi?
darbelo Which is fetchable with "git clone git://gitorious.org/parrot-plumage/parrot-plumage.git"
japhb ... fixed 22:30
(Forgot something in the README. :-)
darbelo 'whirpool' development model? 22:31
japhb d'oh! 22:32
Austin I was going to make a "giant ... sucking" joke there, but you beat me to it.
darbelo is now on gitorius! 22:33
japhb darbelo, fixed
Same nick?
darbelo yeah.
darbelo tries to be darbelo everywhere. 22:34
japhb Good on ya. I couldn't do that for hysterical raisons and nick collisions ....
22:34 ruoso joined
Tene is 'tene' just about everywhere. 22:34
japhb I knight thee "committer", darbelo 22:35
darbelo Any particular rules?
japhb Just what you see in the README. Anything not abundantly obvious, just ask. 22:36
Also, since a lot of stuff right now is in my head, you might want to talk with me about your plans before doing the work, at least for the next couple weeks.
japhb is suddenly racing to catch up with the interest.
darbelo Wait. What? I'm expexted to do work? 22:37
japhb Heh.
If you wish to simple doodle, please create a doodles/ directory to do it in. ;-)
Er, simply 22:39
darbelo git doen't handle empty directories :)
Tene traditionally, you put an empty .gitignore in it.
japhb What Tene said
Damn, you guys keep beating me to the punchline. ;-)
OK, now the hard part -- figuring out tasks I can break off. 22:41
dukelet0 git tracks content. empty directories have no content. At first this is odd, but it makes lots of sense once you start to love the DAG
Austin In a few days I guess I'll be a gitorious user.. 22:42
dalek rrot: r41417 | pmichaud++ | trunk (3 files):
[nqp]: Create a 'parrot_nqp' fakecutable for installation.
rrot: r41418 | darbelo++ | trunk/src/list.c:
Break too long lines to me codetest happy.
darbelo dukelet0: Can you give nopaste.snit.ch/18030 a look. 22:43
japhb Austin, "In a few days"?
Austin Whenever the email gets here.
japhb pmichaud++ # r41417
Wow, sucky email system.
darbelo Austin: took just a few seconds for me.
Austin :)
Yeah. @yahoo.com is a tough domain to send mail to. 22:44
dukelet0 darbelo: bad test. it doesn't take into account the possibility of negative numbers 22:45
darbelo: the regex should look like 242\\s*I0=-?\\d+ 22:46
Austin In the meantime, let me see if I have this right:
Plumage is the repo for search tools, etc. And for specs. There will be - but currently is not? - some central repo for metadata, like a maven repository, that tells the plumage tools how to connect to the source(s) for whatever bundle. 22:47
darbelo dukelet0: Yes, but why does it fail only for me?
japhb Austin, that's a decent description of part of it. 22:48
Austin The metadata should be enough to (1) advertise the stuff; (2) mechanically get the stuff; (3) recursively satisfy dependencies.
japhb There's also the toolchain, the standards (Oh, I see you mentioned 'specs'), and so on.
Austin, configure/build/install, and such, yes. 22:49
Basically: Something needs Parrot to run. How do I find it and get it running?
Sorry, antecedent fail. 22:50
bacek pmichaud: ping. Can you take a look at TT#389? I salvaged stuff from nsentry branch.
japhb How do I find the thing I'm looking for, and how do I get it running with Parrot and whatever else it needs?
Austin Okay, say I write a metadata file. Where shall I put it in my project? 22:51
./PLUMAGE or ./METADATA (bad) or ./etc/PLUMAGE or ?
dukelet0 darbelo: the output of the debugger has garbage data in it. The regex usually passes because the garbage data is usually positive. Evidently on the architecture you are testing it, the garbage data is negative and it fails 22:52
japhb Austin, not decided. But as a quick wack, hmm. Perhaps ./META.json ? or ./plumage/META.json ? Especially if we wanted multiple things there.
dukelet0 darbelo: or, it hates you ;)
darbelo dukelet0: Is the garbage data intended? I can't tell from the tests. 22:53
Austin I'm thinking ./plumage/Close.json
japhb Allison also wants to support the use case that a non-committer to a project supplies the metadata to us separately.
Austin Since you've got "blizkost.json", it seems appropriate to have the file "named right"
japhb Austin, how does Plumage know the name of the file if it can't read the json yet?
darbelo Allison++ on that
japhb Austin, AH! Different issue.
Austin japhb: find ./plumage -name '*.json' ? 22:54
dukelet0 darbelo: the debugger prints out the previous value of a register when you trace. if it has not been set, this is garbage data. My regex/test is too strict and requires the garbage data to be positive, since it usually is. But obviously, it can be negative as well.
japhb That was because what is the metadata/ folder in the repo right now will eventually be data in a server somewhere. For now, I didn't want name collisions.
Austin, that assumes only one file in that directory.
or rather, one json file. 22:55
dukelet0 darbelo: i don't explicitly care about that part of the output, but I care about things after it, so a very loose match in that part is fine
22:55 chromatic joined
dukelet0 darbelo: does that make sense? 22:55
dalek rrot: r41419 | bacek++ | branches/nsentry2:
Recreate nsentry branch against current trunk
rrot: r41420 | bacek++ | branches/nsentry2 (17 files):
Salvaged stuff for nsentry branch
darbelo dukelet0: Yeah, I get it now. I'll go add some -?
Austin Next, how often do I update the file? Is it part of the release cycle (seems likely) or should I point at #HEAD or what?
bacek Can anyone with svn delete old nsentry branch? 22:56
japhb The only reason it's called blizkost.json now is because plumage is still in the prototyping phase. :-)
dukelet0 darbelo: yes, other tests might benefit from that as well
darbelo: other debugger tests, that is
japhb Austin, my thought is that we eventually want to support the use case of someone saying:
plumage install blizkost-5.9 22:57
or
plumage install blizkost-blead
or what have you
Austin Okay.
darbelo bacek: done.
Austin What does that mean to me? Will all those revs be in my .json file, or will there be scads of Close.json files in the server someplace?
japhb Austin, that brought up the idea of having includes or other forms of "inheritance" for metadata files, but that is another undecided. 22:58
Austin Meta-meta-data?
japhb Austin, I just don't know yet. You're too far ahead of me to give a definitive answer yet.
bacek darbelo: thanks!
japhb More-meta-than-yo-mama-data
dalek rrot: r41421 | darbelo++ | branches/nsentry:
Kill the uinmerged nsentry branch.
 22:59
Austin What's an authority?
purl an authority is as authority does.
Austin Is that a user, or a repo?
japhb Austin, "owner" in some sense. 'CPAN:RJBS', for instance. 23:00
23:00 rhr joined
japhb It's to handle forking properly. 23:00
Austin The one for blizkost is a url.
japhb Austin, the JSON file for blizkost is only at the sketch stage. Getting my thoughts on paper so that I could go into my next hack day ready to code a simple fetch/build/install sequence. 23:01
Take nothing as gospel. ;-) 23:02
This project takes rather more thinking than you would, er, think.
23:02 tetragon joined
dalek rrot: r41422 | darbelo++ | trunk/t/tools/parrot_debugger.t:
Loosen a few regezes in t/tools/parrot_debugger.t to accept negative numbers.
 23:02
Austin :)
darbelo has stupid fingers. 23:04
Austin Okay, here's some thoughts. 23:05
1. Parrot is talking about switching from svn to git. So you should probably associate versions with repositories either as associations, or with a has-a of some kind (inclusion) 23:06
2. Likewise, hosting providers change, so maybe support for different/alternate ones. 23:07
japhb good so far
Austin 3. Storing a list of versions is horrible. Better to provide some mechanism for querying available ones. 23:09
japhb Oh, I forgot to explain why an authority could be a URI ... my thought was that anything that uniquely identified a particular source for the source could be considered a valid authority identifier. Kind of like how OpenID works -- your persona is just a URI. 23:10
Austin 4. But dependency info is version-dependent, which puts everything inside the archive.
japhb I quibble with #4. *nix distro packaging systems keep version info outside of the source archive ... in their own package system. 23:11
Austin 5. So you sort of want both worlds - here's a project, ask the sources about versions, versus here's a project-version, with some sources that may have other versions, too. 23:12
japhb nodnod
darbelo Hmm. "Copyright (C) 2009, Parrot Foundation." Are you going to require a CLA for plumage contributors? 23:13
Austin Yeah, the thing with #4 is one of balance. For distros, it's easy to store metadata externally. But for things like maven repositories, where versions can be 1.0.1, 1.0.2, 1.0.3, etc., it's going to be a lot easier to ask a repo what is provides.
*what it provides 23:14
japhb darbelo: oh frack, I forgot about that stupid CLA thing. I was just making sure the eventual merge would not have the most obvious copyright and licensing fail. 23:15
"maven repositories"?
pmichaud bacek: (TT #389) I'm guessing that rakudo doesn't need a whole lot of patching to make it work. More to the point, what existed in the branch in december is probably not completely relevant to today 23:16
japhb Someone ping a PAFO person about whether I need to enforce CLA on a project that I am gifting to them?
bacek pmichaud: indeed. But some rakudo patching required.
japhb pmichaud, you're on the board now, yes? Do you know the answer to that question?
chromatic japhb, I think so.
japhb frak with frak sauce.
japhb really wanted to go the audreyt route. 23:17
darbelo So far, I think you are in the clear, your only commiters are also parrot commiters, so no problem there.
pmichaud japhb: iiuc, any code that makes it into the parrot repo has to have CLAs on all of its contributions
bacek anyway, $dayjob time
japhb Hmmm.
pmichaud I suspect the same would be true for anything that we expect parrot to officially distribute
japhb OK, well, I'll note that in the README for now, as darbelo is right about not having actually hit the problem yet. 23:18
Austin OTOH, something that's going to distribute parrot ...?
japhb Austin, not a problem here. Plumage fully depends on Parrot already being there.
pmichaud (I'm not entirely in agreement with that policy, but I understand the reasons behind it and I'm not objectionable to it)
something that decides to distribute parrot only has to comply with parrot's license (artistic 2.0) 23:19
japhb Allison brought up that bootstrapping issue, but I'm not convinced yet.
Austin If you distribute module metadata, what's their coverage?
Say I provide a Close.json file. Do I have to be CLA'd before Plumage can have it? 23:20
pmichaud whoever checks it into Plumage has to be CLAd
dalek rrot: r41423 | bacek++ | branches/nsentry2/compilers/pct/src/PCT/HLLCompiler.pir:
Mark HLLCompiler.parse_name with :nsentry. It heavily used in Rakudo
pmichaud it doesn't have to be the original author (more)
rrot: r41424 | bacek++ | branches/nsentry2/runtime/parrot/library/P6object.pir:
Don't try to deref null list of methods
pmichaud however, the committer is responsible for making sure that all licenses are complied with
Austin That seems wrong. 23:21
darbelo japhb: As I understand it, the CLA is the paper trail that PaFo can use to show that the copyright is accurate.
pmichaud bacek: I think that parse_name should always be used as a method, never as a sub.
(I grant that perhaps Rakudo isn't doing that) 23:22
darbelo This CLA here says that dude agreed to putting a "(C) PaFo" sticker on his work.
23:22 kid51 joined
pmichaud if someone can't vouch for the origin/license of some piece of code, they shouldn't be committing it 23:22
that goes along with "responsibility of being a committer" 23:23
Austin PMichaud, sure. But I'm specifically questioning the metadata here.
23:23 jrtayloriv joined
japhb Austin, once the metadata is no longer stored in the parrot-plumage repository, I think it becomes a different case. 23:24
pmichaud Austin: I'm not sure where that boundary lies. Allison would be the person to ask there.
darbelo Austin: metadata does isn't necessarily subject to that restriction. We were talking source code in the plumage repo.
japhb In particular, it becomes "data managed by the Plumage service", not "a part of Plumage".
pmichaud japhb +1
chromatic I'm not sure that metadata is copyrightable, at least in individual pieces.
The *collection* of said metadata may be copyrightable. 23:25
Austin But currently that can only be managed by including it in the repo.
japhb Austin, yes, right now. But as of this instant, the only people who can include it in the repo are people with a signed CLA, so that's a non-issue on a technicality. 23:26
pmichaud agreed, delayed binding is good here :) 23:27
japhb Anyone object to the new wording in the first paragraph of the CONTRIBUTING section of the README? gitorious.org/parrot-plumage/parrot...ter/README
Austin I guess I should have asked for the commit bit before you made that change :-) 23:28
japhb Sorry Austin ... 23:29
On the flip side, as I understand the discussion so far, you *don't* need to be a Parrot committer ... only to have sent in a PaFo CLA.
Austin That's okay. I'll just use "purl message japhb nopaste.org/..." for all my updates.
pmichaud in the "fwiw" category, I've been thinking that although we require CLAs for the Rakudo compiler, a Rakudo distribution is likely to include components that aren't covered directly by the CLA
japhb Austin, ouch .... :-) 23:30
darbelo japhb: +1 here
pmichaud afk # dinner and errands
Austin Maybe I can create a svn commit-hook to automate that..
japhb Austin: BPFH ... 23:31
Austin What's a comment in JSON? 23:32
japhb / I think ...?
er, //
Austin Oka
Wjat 23:33
japhb www.json.org/ has the meat of the spec on the front page
23:33 jrtayloriv joined
japhb Hmmm, come to think of it, I don't know if official JSON even *has* comments. 23:34
Austin purl message japhb nopaste.org/p/awEkRdTF3
purl Message for japhb stored.
japhb OIC 23:35
Austin So here's a specific question: how do you handle two versions?
Create two of those things? 23:36
0.1.2
japhb Right now: Two different files, mostly the same. Yech. Mid-term: probably some is-a or has-a method for cutting out the duplication and only listing differences. 23:37
Austin Actually that's wrong. The installed parrot is required for building
japhb Right now, I'm assuming people pointing to blead.
Yes, I was going to say that, but hadn't had the chance.
Austin No. I'm right. Whiteknight had the problem on configure.
japhb ?
Austin Cause it runs parrot_config 23:38
I don't know if I need perl5 for build..
japhb The *plumage* rule is that you can just assume an installed parrot. If I could make it break on a non-installed parrot, I'd consider it.
Austin Okay. Document that.
japhb I thought I did, somewhere, hmm. 23:39
Austin Maybe you did.
darbelo Austin: Artistic 1.0 ? not Artistic 2.0 ?
japhb: installed os install-deved parrot?
Austin It's what I got from google. The give-a-shit meter doesn't read real high on that.
japhb install-dev. I think it's silly to assume you can build something on top of a Parrot that has not been install-dev'ed. 23:40
To me, install'ed Parrot is only for embedded or other never-changing situations.
Austin japhb: Doesn't that argue that install-dev is the new install?
darbelo Austin: FWIW Perl 5 is (Artistic 1 + GPL) parrot is Artistic 2.0
japhb Austin, I certainly feel it is. 23:41
You can't install any HLL without install-dev, so that pretty much nails it for me.
darbelo It should be. I'll try to remember that for the next parrotsketch.
Austin: www.opensource.org/licenses/artisti...se-2.0.php if you care enough to read it. 23:43
Austin Updated to Art+GPL
japhb Short form: If you don't care, use Artistic 2.0. Only use 1.0 for specific cases.
Austin, Artistic 2.0 is designed to be GPL-happy.
So no need for disjunctive license anymore. 23:44
Austin Yeah, but it's not google-code -happy.
japhb Someone needs to kick someone over at Google, then, because it's not exactly new.
Austin :)
23:44 Limbic_Region joined 23:45 TiMBuS joined
darbelo You can label it as 'artistic' there and clarify on the README 23:45
... or add an explicit LICENSE file. NotFound did that for his mysql libs 23:46
code.google.com/p/parrot-mysql/
code.google.com/p/parrot-mysql/sour...nk/LICENSE
Tene japhb: I'm going home now. 23:48
japhb Tene, very sorry to not get you
the task list you wanted yet.
As you can see, I've been a tad distracted.
Tene japhb: I've got a bit more to do first. :)
s'fine
japhb Tene: will you still be available via IRC later? 23:49
Tene japhb: Yes. 23:51
Austin So, given that Close has a plumage metadata file, how does someone go about installing Close?
Tene Austin: they can't yet.
Austin WHAT THE HELL IS TAKING YOU PEOPLE SO LONG????????????
japhb It's the Next Task
Tene Austin: the next step for plumage is getting it able to download something.
Austin :)
darbelo (0) They get plumage.
(1) They build plubage 23:52
Austin Profit?
purl i think Profit is overrated ... but payroll isn't :)
darbelo (2) #comment out, doesn't work yet...
dalek TT #1058 created by jkeenan++: auto::funcptr: config step not needed if 'jitcapable' is never true
Austin Bummer.
japhb I can see I'm just going to have to code my brains out just to get some peace around here. ;-) 23:53
Austin Perl5 version first?
dalek ose: r129 | Austin++ | trunk/plumage (2 files):
Added ParrotPlumage metadata
23:53 jrtayloriv joined
darbelo plumage info blizkost 23:54
"load_bytecode" couldn't find file 'compilers/json/JSON.pbc'
current instr.: 'parrot;Config;JSON;ReadConfig' pc 0 (runtime/parrot/library/Config/JSON.pir:21)
called from Sub 'get_project_metadata' pc 1156 (plumage.pir:449)
japhb ?
darbelo Ouch.
japhb that's ... odd.
OH DANGIT 23:55
Austin compilers/ ?
purl compilers/ is for core utilities
japhb I know what happened.
Forgot to push a parrot commit to fix Config::JSON.
compilers is the installed version of languages/
Austin It is? I thought that was the other way around... 23:56
darbelo No, compilers is the installed version of compilers/
purl okay, darbelo.
darbelo I hate purl some days.
japhb What Austin said.
darbelo No, compilers/ is for core utilities
purl i already had it that way, darbelo.
Austin Umm, ls /usr/local/lib/parrot/1.6.0-devel has a languages/ but has no compilers/
japhb ETOOMANYTASKS
darbelo No, compilers is for core utilities
purl okay, darbelo.
japhb Austin, you're right, I was trying to say languages/ was the installed version of compilers/ . Config::JSON is broken; give me a minute to find that patch and commit. 23:57
Austin Take all the time you need.
darbelo Somebody needs to go get a time machine, travel back in time and design an install prcedure for parrot. 23:58
japhb darbelo, try r41425 . 23:59

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!