Parrot 1.7.0 "African Grey" is out! | Latest modified TT's: icanhaz.com/parrotbugs
Set by moderator on 16 November 2009.
japhb Ah well, guess it will have to be a generic tweet then. 00:03
Whiteknight how do I switch which branch I am on in git?
japhb 'git checkout <branch>'
If you want to create a new local branch at the same time, it's 'git checkout -b <new_branch>' 00:04
Austin git branch
purl git branch is a user command; if that format changes it'll break my one-liner
japhb To see branches and current branch, 'git branch' with no arguments
dukeleto 'ello 00:05
japhb o/
Austin WhiteKnight: I think 7 git branch uses a '*' to indicate the current branch.
treed It does. 00:06
purl if you say so...
dukeleto japhb: i will try to write more tests to make sure that win32 is failing tests again :)
japhb dukeleto, actually, making any platform fail will work. ;-)
We're woefully undertested.
dukeleto git branch -a shows all branches,including remotes 00:07
japhb It would be nice to get complete coverage of Glue.pir and Util.nqp, and a complete set of invalid & missing arg tests for plumage.
with the most recent tests reorg, it should be easier to see what's missing there. 00:08
dukeleto japhb: ok
japhb Doing a complete set of functional tests for plumage is going to be harder, so that's lower priority. "low hanging fruit" and all that.
dukeleto japhb: also, i started nqptap: github.com/leto/nqptap (you and pmichaud have commit access) 00:10
japhb: it is just a skeleton right now, but i am going to properly test the test harness
Coke anyone here familiar with setting up a google group that tracks an existing mailing list?
seen allison?
purl allison was last seen on #parrot 2 days, 23 hours, 11 minutes and 41 seconds ago, saying: japhb: there is definitely also room for introductory documentation separate from src/ [Nov 14 00:58:11 2009]
japhb dukeleto, ah, very cool.
dukeleto++
Coke, no ... though I've been seeing some serious hate directed at Google Groups lately because of an apparent total lack of the anti-spam technology that gmail has .... 00:12
Coke if we're mirroring a mailing list, that doesn't matter so much.
dukeleto dcolish: take a look at tools/dev/mk_language_shell.pl in the parrot repo for how to start a new language
pmichaud Coke: basically I think it's (1) create the group, (2) subscribe the group to the mailing list 00:13
at least, that's what I've done for the groups I've used
japhb dukeleto, nqptap's CREDITS still refers to Kea. Also, you may want to s/nqp/parrot-nqp/g. Or not. Or mention both. 00:14
Coke, assuming you can limit posts to being only from the mirrored list, yes. (I don't happen to know, I haven't tried to do that.) 00:15
dcolish dukeleto: ran that, so now i have a languages/haskell dir in my parrot tree
he_ Hmm, I'm looking at the t/pmc/threads.t test failing, at smolder.plusthree.com/app/public_pr...st_failure 00:16
japhb dukeleto, And of course I hope to see an nqptap.json in Plumage's metadata soon. ;-)
he_ The "Invalid shared object handle" error message comes from ld.elf_so.
specifically from _rtld_dlcheck(), which is a static function, called from dlclose() and in some cases from dlsym(). 00:17
So what magic does parrot try to do which NetBSD's ld.elf_so balks at?
dcolish so what is the simplest grammar tool to use for starting a new language
japhb Grrr. ping.fm fail. 00:18
00:19 Whiteknight joined
Whiteknight okay, I deleted a branch locally. How do I make that branch disappear from github? 00:21
dukeleto japhb:here is your ping.fm gold watch
dcolish: congrats!
japhb dukeleto, OK where did the gold watch meme begin?
Austin squaak?
purl i think squaak is languages/squaak or the language in kj's tutorial at parrotblog.org or svn.perl.org/parrot/trunk/languages/squaak/
dcolish :) well now i need to actually do work
dukeleto japhb: thanks for the heads up. will fix those soon and give plumage a .json metadata file as soon as nqptap works :)
japhb :-) 00:22
dukeleto japhb: the movie "Speed" maybe? it probably existed before that
Austin dcolish: en.wikibooks.org/wiki/Parrot_Virtua...troduction
dcolish excellent, thank you Austin
dukeleto dcolish: squaak is a good example, but some of the tools have improved since then
Austin tutorial?
purl tutorial is search.cpan.org/dist/Catalyst-Manua...torial.pod
japhb Wow, Speed was 15 years ago. 00:23
Austin no, tutorial is the Parrot interpreter tutorial at en.wikibooks.org/wiki/Parrot_Virtua...troduction
purl okay, Austin.
Coke kid51? 00:24
purl kid51 is probably now going to live dangerously.
Coke jkeenan?
purl well, jkeenan is kid51
Coke kid51 email?
dukeleto dcolish: other questions? 00:25
dcolish none for now
i'm going to go over the tutorial tonight
dukeleto dcolish: recently we got a new NQP, called NQP-rx. it makes some things in the PCT a lot easier, but most languages currently built on parrot use PCT, not NQP-rx
it just landed in parrot a few days ago, hasn't been in a release yet 00:26
but it is definitely the best way to write a HLL
dcolish i built from svn today so i might have it then
japhb dcolish, you definitely do.
dukeleto but getting to know the PCT is definitely worth it as well
dcolish: yes, you do. it is in ext/
treed dukeleto: Are there any guides for converting yet?
dukeleto treed: converting languges on PCT to NQP-rx? 00:27
treed Yeah.
dukeleto treed: that is a great question for pmichaud
treed Heh.
dukeleto treed: at least a guide would be very helpful
dcolish: NQP-rx makes defining the grammar/actions simpler
japhb Also, a lot of stuff that was possible but painful in old NQP is now much easier. 00:28
dcolish i'm used to yacc grammars so anything like that and i'll feel right at home
treed PCT used Perl6 Regex 00:29
I'd be surprised if NQP-rx didn't.
It's pretty nice.
japhb dcolish, you're going to wonder how you ever used yacc once you get used to Perl 6 Grammars.
dcolish heh cool
pmichaud I still need to write a tutorial, too
ETOOMANYCOOLTOYS
japhb pmichaud, MAKE MORE TOYS! ;-) 00:30
pmichaud the new toolkit available through nqp is way cooler than what pge/pct provided
especially with the protoregexes
dukeleto dcolish: github.com/leto/kea/blob/3aedd24bd1...a-compiler for an example of NQP-rx
pmichaud: tutorials would be greatly appreciated :) i can help, since i was gonna write one if you didn't get around to it soon :) 00:31
japhb pmichaud, on a serious note -- I'm beginning to be blocked on not having one of the "globals" solutions from wiki.github.com/perl6/nqp-rx/plumage-requests .... In fact, I could actually make some progress "coding in the dark" in a few places if I even know which one(s) you *plan* to make work.
dcolish dukeleto: is factor related to forth at all? 00:34
00:35 plobsing joined
japhb dcolish, yes, via Joy, IIRC. 00:35
darbelo dcolish: It's a stack language. How could it not be? 00:36
japhb dcolish, see factorcode.org/ and concatenative.org/wiki/view/Factor
dcolish realizes the sillieness of the question only moments after asking
chromatic It's not necessarily a silly question.
pmichaud japhb: I'll see if I can remove the block.... one moment
Austin japhb: I've got solutions for a bunch of those requests in Kakapo. (If only I could get strings to work.)
chromatic I thought PostScript had a Forth heritage, but it doesn't. 00:37
dukeleto dcolish: Factor borrows heavily from stack language culture
pmichaud japhb: I'm not likely to do lexical import anytime soon
japhb: I can probably get @GLOBALS::ARGS to work
and contextuals looking in namespace, at least for read. Not sure about binding, though.
dcolish chromatic: fwiw, wikipedia says postscript has a lisp heritage
japhb pmichaud, \\o/ # thank you! 00:38
dukeleto dcolish: Factor devs describe it as "forth with batteries"
japhb Chuck Moore would cringe. 00:39
Austin japhb: Specifically: (1) I use a dependency queue for load/init ordering, so with some explicit specs I can get startup dependencies done. (2) I've got import/export, including a "global" namespace. (3) I've got class/subclass declarations, using attributes.
dcolish dukeleto: heh, i'll have to look at it more, i've always been fascinated with the stack langs
dukeleto dcolish: ok, i am adding you to the kea github repo now :)
dcolish lol
i better stop before i'm involved with every project under the sun!
japhb Austin, I would love to be able to use a common set of such libraries. Unfortunately, I can only do so if you can get shipped with parrot. :-/ 00:40
*can get *kakapo*
Austin Actually, you need to ship a kakapo release with plumage.
(IOW, copy the .pirs)
japhb Austin, plumage will be shipping with Parrot, so same difference. ;-) 00:41
00:41 mikehh joined
Austin Right. So copy the pirs into plumage and you're done. 00:41
Actually, it's more than that, but I'll send you a merge eventually. 00:42
japhb Austin, have you done a PaFo CLA? 00:45
Austin Nope.
dukeleto dcolish: ok, you are on the kea gtihub repo now
dcolish dukeleto: excellent 00:46
dukeleto dcolish: looks like you need to have 48 more repos to catch up to my level of "involved with every project under the sun" :)
dcolish heh
dukeleto dcolish: please add yourself to CREDITS 00:47
dcolish k
japhb Austin, the other roadblock is that I'm not allowed to put anything into Plumage unless the author has done a PaFo CLA.
Austin Okay.
japhb So please do. ;-) 00:48
00:49 tetragon_ joined, abqar joined
dukeleto japhb: you can accept a patch, but no one can actually commit to plumage with a PaFo CLA, right ? 00:50
japhb: i don't think there is anything wrong with you accepting a patch into plumage from someone who hasn't signed the CLA
dalek TT #1278 created by coke++: remove MT19937 from stdlib 00:53
00:55 mikehh joined
dalek rrot: r42539 | coke++ | trunk/DEPRECATED.pod:
Add a notice for this DEP ticket.
 00:55
00:56 particle joined 00:59 mikehh joined
dalek trixy: 98e5607 | Whiteknight++ | (4 files):
apparently all my complex changes from branch didn't make it through the merge, or whatever. Re-add some stuff
 01:04
trixy: 82ca237 | Whiteknight++ | src/ (2 files):
fix another issue from the missing complex code
trixy: 1e7df46 | Whiteknight++ | t/300-transpose.t:
fix the plan on transpose.t
trixy: 6591db1 | Whiteknight++ | src/builtins/conj.pir:
separate out conj() into it's own file, as per standards
 01:10
kid51_at_dinner make fulltest: Linux/i386: no special configuration options: PASS at r 42538 01:12
cotto_work That's a good sign given tomorrow's release.
kid51 results soon for darwin/ppc 01:14
cotto_work goto home; 01:17
Coke grumbles at svn. 01:20
I wish to revert a commit that involved a few deletions. 01:21
I don't see a one liner for this. (tortoise svn gets this right...) 01:22
01:25 particle joined
pmichaud I need to update NEWS and DEPRECATED for the release. Will do that in a couple of hours. 01:30
dalek TT #1279 created by pmichaud++: Parrot prefers libraries in build tree to installed tree or current ... 01:32
Ryan52 when's the next parrot release? 01:36
chromatic Tomorrow. 01:37
purl o/~ the sun will come out.. tomorrow.. o/~ or the National Day of Slayer, and the National Emo Kid Beatdown day, by www.nationaldayofslayer.org/ and community.livejournal.com/wtf_inc/2805832.html or tomorrow and tomorrow and tomorrow creeps in this petty pace or mañana or free for all or another day or the name of the pitcher
Ryan52 nice.
allison: mind if I make another upload to debian experimental after the release? do you want to review it again or can I just do it? 01:38
dalek tracwiki: v1 | plobsing++ | Lorito 01:51
tracwiki: initial thoughts
tracwiki: trac.parrot.org/parrot/wiki/Lorito...ction=diff
TT #608 closed by coke++: user mailing list (parrot-users) with an associated google group 02:03
TT #1280 created by jimmy++: [patch]removed unused codes in oo.c 02:06
TT #917 closed by coke++: What mailing lists should parrot.org host?
rrot: r42540 | coke++ | trunk (3 files):
re-add mt19937 to the stdlib (See TT #1278)
 02:07
02:07 tetragon joined
Coke AOOOGA. 02:09
parrot-users now exists.
(has for some time, but just twisted allison++'s arm to setup the google groups stuff.)
if you're a dev, please consider joining to help out the users. 02:10
diakopter rakudo: $_="hihi"; .=="hihi".say 02:20
oop.
dalek TT #1281 created by coke++: Remove graph coloring register allocator from IMCC 02:26
japhb dukeleto, My understanding (IA *SO* NAL) from the last time this came up is that by committing a patch, I am claiming that the copyright stuff is OK. For small patches, that's probably rarely an issue. For a large change, with lots of new files, I don't think I can just do it without myself getting the approximate equivalent of a PaFo CLA -- so no reason to do that, rather just insist that major new stuff be sent from someone who has alr 02:27
eady submitted a PaFo CLA.
dalek p-rx: d7b50d4 | pmichaud++ | (5 files):
Add @XYZ::foo and $GLOBAL::bar variables.
 02:28
p-rx: 2925cff | pmichaud++ | (2 files):
Contextuals also look in the global namespace.
p-rx: 49c429d | pmichaud++ | src/stage0/ (3 files):
Bootstrap update for package vars and contextuals in global scope.
02:30 kiwichris joined
dalek TT #1282 created by coke++: OS.cwd returns platform specific slashes... 02:36
pmichaud japhb: r42541 is your friend. :-)
Coke pmichaud: you said you had a comment to make on rt.perl.org/rt3/Ticket/Display.html?id=59576 that would let us close it, any chance you can poke at that tonight? 02:37
dalek rrot: r42541 | pmichaud++ | failed to fetch changeset:
[nqp]: Update with latest contextual and package var improvements.
 02:38
TT #1283 created by coke++: rethrow should keep the backtrace of the original 'die' 02:44
02:53 cconstantine joined
japhb pmichaud++ # thanks again, I will check it out tomorrow 02:54
dalek TT #1284 created by coke++: Integer PMC missing math methods 03:00
03:01 cconstantine joined
dalek TT #1285 created by coke++: Some files generated by Configure.pl should be generated by Makefile 03:10
TT #1286 created by particle++: method overload in pir subclass of pmc pccmethod leaves object undefined 03:16
kid51 Belated report: make fulltest passed on darwin/ppc at r 42538 03:17
dalek TT #1287 created by coke++: Move thread check into Parrot_exit() 03:20
03:22 mokurai1 joined
dalek TT #1288 created by coke++: Limit the required interface for PMCs 03:23
Coke tries to parse Bob Rogers's email client's "reply-to" format. 03:26
Coke fails.
pmichaud Coke: RT #59576 closed. 03:29
03:35 janus joined
Coke pmichaud++ 03:39
dalek TT #1289 created by coke++: Review/Consolidate test infrastructure in t/codingstd and t/distro 03:59
Coke who owns irclogs.perlgeek.de ? 04:02
PerlJam ITYM irclog.perlgeek.de 04:03
Coke ITYKWIM. 04:04
purl i think itykwim is I Think You Know What I Mean
cotto Coke, moritz
Coke msg moritz IWBNI irclog.perlgeek.de knew that #dddd could also mean trac tickets. 04:05
purl Message for moritz stored.
dalek kudo: d3a573b | (Solomon Foster)++ | src/setting/ (2 files):
Add Rat.abs. When doing abs(Any), just redispatch to self.abs rather than taking +self.
 04:11
TT #1290 created by chromatic++: Produce Single PBC from Multiple PIR Files with -o 04:12
TT #1291 created by coke++: update make headerizer to work with PMCs too. 04:26
Coke NotFound: ping.
msg NotFound can you look at migrating rt.perl.org/rt3/Ticket/Display.html?id=39117 ?
purl Message for notfound stored.
Austin Actually, Coke, IWBNI if one of those url-rewriting bots would put those links directly into chat. Then the logs could just log what the bot said. 04:28
PerlJam Austin: I don't think so. More bot-noise is not a good thing. 04:30
(though if it were logging an internal monologue that contained the URLs but that didn't make it on-channel, that would be neat I think) 04:33
dalek TT #1292 created by coke++: Correct destruction of PackFile objects 04:36
cotto Coke, I was going to reject that Array ticket after #ps tomorrow, assuming that people were ok with its deprecation. 04:39
dalek TT #1293 created by cotto++: Array PMC freeze/thaw/visit broken
Austin PerlJam: I haven't had a botnoise problem since I started /ignore-ing them. But given that #123 gets displayed as a link to another IRC channel, having the bot come along with the right url strikes me as worth seeing. 04:41
Coke #123 is just text here. 04:54
if the recommended course of action is to ignore the bots, why have them spew? 04:55
cotto: certainly there's no point in cleaning up array if we're deleting it.
cotto: I wouldn't just ask at #ps, I'd open an RFC ticket.
and at this point, let's transfer the tickets even if we think they /might/ be rejected in trac soon. 04:56
RT: down to 13 tickets.
msg allison I dealt with 4/5 of your tickets in exchange for the google groups. Hopefully someone else will migrate the last one. (I'm hoping pmichaud peeks at it.) 04:57
purl Message for allison stored.
Austin Coke: Different folks care about different stuff. I have no interest, for example, in seeing all the latest commits. Others may not care about tickets.
Coke Austin: fair enough. 04:58
Coke hopes those 13 tickets are gone by tomorrow. =-) 05:22
05:54 FullMetalHarlot joined
dalek tracwiki: v2 | cotto++ | Lorito 06:28
tracwiki: add some counterpoints and clarifications to the opcode proposition
tracwiki: trac.parrot.org/parrot/wiki/Lorito...ction=diff
06:39 kiwichris joined
NotFound Coke: gone 06:43
with the wind. 06:44
06:44 TiMBuS joined 07:04 uniejo joined
diakopter microsoft released the source to .NET micro framework 4.0 under the Apache 2.0 license.. 07:09
07:10 bacek joined
cotto hi bacek 07:10
bacek aloha cotto 07:11
cotto pictures "bacek of today" with Hawaiian shirt bacek 07:14
bacek cotto I'm in pretty boring style today... 07:15
cotto I have a Hawaiian shirt. I feel loud whenever I put it on.
bacek I actually have couple of them. 07:17
cotto, btw, I tried to update ops_pmc to trunk. 97 conflicts... 07:18
cotto I'm not sure how that's possible. I didn't think there were even 97 changes outside of compilers/opsc. 07:20
bacek probably git-svn borked somehow 07:21
cotto either way, svn--
bacek May be create new branch, copy compilers/opsc across and bring life in this project 07:22
cotto I hate that the best way to deal with branches involves clobbering the history periodically
yeah
dukeleto 'ello
cotto I'd also recommend checking the diff between the most recent merge point and the branch to see what we changes outside of compilers/opsc.
hi dukeleto 07:23
bacek hi dukeleto
cotto, it's in config/ for generating opsc Makefile
cotto I'd expect fewer than 97 conflicts for that change. ;) 07:29
07:31 iblechbot joined, mokurai1 left
cotto bacek, I don't think that message went where you wanted it to go. 07:43
bacek cotto which one???
oh shit.. 07:44
cotto Merge Party!
bacek yak...
sorry...
purl It's okay, bacek.
cotto It's fine. I think it's funny since there don't appear to be any important sekrits in the message. 07:45
chromatic Is there beer?
cotto ponders what kind of beers would be good when merged 07:46
bacek It wasn't... But I still wonder how bloody thunderbird managed to send it to wrong list? 07:47
moritz uhm
isn't today release day?
oh, I see, the branches were merged, but not in trunk, right? 07:48
bacek cotto: "Carlton Draught"
cotto moritz, I think it's from bacek's work
moritz oh, ok 07:49
bacek moritz, it was for my $dayjob... Actually it was sent to both lists...
moritz was a bit puzzled by the unfamiliar names
07:49 nbrown joined 07:51 fperrad joined
cotto goes to sleeeeeeeeeeeep 08:11
dalek rrot: r42542 | fperrad++ | trunk/NEWS:
[NEWS] revert r42467, see TT #1278
 08:14
08:16 barney joined 08:24 mikehh joined 08:32 nbrown joined
japhb posts his #parrotsketch report about 10 hours early ... 08:46
And now, off to bed. g'night, all.
dalek rrot: r42543 | barney++ | trunk/NEWS:
Added news about the deprecated Mersenne twisted pseudorandom number generator.
 08:48
chromatic One quick bugfix for pbc_to_exe coming, barney. 08:50
dalek rrot: r42544 | chromatic++ | trunk/tools/dev/pbc_to_exe.pir:
[pbc_to_exe] Fixed pbc_to_exe to allow the generation of installable
 08:51
barney No hurry 08:53
08:54 nbrown joined, baest joined
dalek rrot: r42545 | barney++ | trunk/NEWS:
Added news about parrot-users.
 08:55
09:01 davidfetter joined 09:16 einstein joined
mikehh barney: when are fixin' to do the release 09:41
barney I plan on 13:00 UTC 09:54
moderator Release of Parrot 1.8.0 starting at 13:00 UTC | Latest modified TT's: icanhaz.com/parrotbugs 10:10
nbrown_ joined
dalek rrot: r42546 | barney++ | trunk/docs/project/release_manager_guide.pod:
Moved an announcement to committers into item 0.
 10:22
10:41 riffraff joined 10:56 slavorgn joined 10:57 theory joined 10:58 payload joined
dalek rrot: r42547 | mikehh++ | trunk/t/op/comp.t:
fix test failure on g++ 4.4.1 (with --optimize) on i386 - TT #1275
 10:58
mikehh I still want to try and sort out TT #1142 before the release 11:00
at least the testing side so we don't get TODO passes 11:05
11:12 payload1 joined
barney mikehh: yes, please go ahead 11:13
11:18 nbrown joined 11:54 bluescreen joined 11:58 payload joined 12:01 jsut_ joined
barney Is META.yml still needed? If not I'll create a TT to remove it. 12:02
12:02 cconstantine joined
dalek rrot: r42548 | barney++ | trunk/README:
Removed reference to Bundle::Parrot.
 12:05
moritz barney: I think the stable releases go to CPAN, so then a META.yml is necessary 12:06
where stable = deprecation checkpoint releases 12:07
dalek rrot: r42549 | mikehh++ | trunk/t/pmc/eval.t:
fix test so it only sets TODO for testr on amd64 with cc-gcc and not g++
 12:11
rrot: r42550 | barney++ | trunk/docs/parrot.pod:
[doc] Minor change in wording.
TT #553 closed by barney++: [CAGE] Remove Bundle::Parrot from README? 12:13
barney moritz: I'll ask allison, CPAN is no longer mentioned in release_manager_guide.pod 12:14
dalek rrot: r42551 | barney++ | trunk/README:
[doc] Point to the correct files with platform info.
 12:18
12:27 kid51 joined 12:39 payload1 joined
mikehh All tests PASS (pre/post-config, smoke (#29980), fulltest) at r42550 - Ubuntu 9.10 amd64 (g++ with --optimize) 12:46
barney mikehh++ 12:52
mikehh All tests PASS (pre/post-config, smoke (#29981), fulltest) at r42551 - Ubuntu 9.10 amd64 (gcc with --optimize) 12:53
have we time to run tests on Ubuntu 9.10 i386?
barney doing that right now 12:55
mikehh ok I'll leave it to you
messages 13:05
13:09 whiteknight joined
whiteknight good morning #parrot 13:12
pmichaud good morning, #parrot 13:15
eta to release? I might have some NEWS updates 13:16
whiteknight barney said he was going to start around 9:00AM european time 13:19
I don't remember which timezone he said specifically 13:20
moritz see /topic
that's, uhm, 20 minutes ago :-)
pmichaud 09:54 <barney> I plan on 13:00 UTC
whiteknight oh. I must be going blind. moritz++
pmichaud I never see the topic in the chans. 13:21
so, too late for NEWS updates?
moritz has a handy topicdiff script for irssi
pmichaud I notice that "make install" isn't listed in NEWS.
oh, wait, there it is.
having most of the bullet items start with "The " makes it a bit more difficult to scan. :-| 13:22
barney: ping 13:23
13:26 plobsing joined
whiteknight I think barney is at $day_job 13:27
dalek rrot: r42552 | whiteknight++ | trunk/src/multidispatch.c:
[mmd] remove RT #45959 note from the source here. That ticket has been rejected and this is not the correct place to be evaluating MMD candidates for fitness.
 13:29
pmichaud so, too late to commit changes to NEWS ?
Coke I haven't seen any messages either way. 13:30
whiteknight just commit it
Coke I would sneak them in.
whiteknight worst case scenario: it doesn't make it into the tarball
Coke (at which point it's prewritten for next time.)
barney pmichaud: go ahead with NEWS
Coke EEK it's barney. 13:31
barney?
purl barney is a big, purple piece of shit, or gettin' jiggy wid it or see 'grimace'. or purple dupa or o/` I love you / You love me / We're a happy family / with a great big hug /and a kiss from me to you / Won't you say you love me TOO! /`o
Coke botsnack
purl :)
13:33 tetragon joined
dalek rrot: r42553 | whiteknight++ | trunk/src/embed.c:
[embed] Remove old comment for RT #46155. This code isn't used and causes confusion. If we find actual memory leaks later we can create a new ticket
 13:35
pmichaud news updates committed 13:38
(r42554) 13:39
dalek rrot: r42554 | pmichaud++ | trunk/NEWS:
[NEWS]: NEWS updates
13:41 payload joined 13:42 bluescreen joined 13:44 masak joined
dalek rrot: r42555 | barney++ | trunk/NEWS:
Added details about merged branches.
 13:45
rrot: r42556 | whiteknight++ | trunk/src/exit.c:
[exit] Remove note for RT #46403
13:56 lucian joined
barney Any more outstanding commits? (Running 'make fulltest' now) 13:56
dalek TT #1294 created by doughera++: Build failure on OpenBSD 4.6 -- missing dependency? 13:58
Coke barney: assuming that I'm one of the few people that can update docs.parrot.org, lemme know when you get to that step.
barney Coke: will do 14:00
Coke barney: #1294 looks like something we should try to fix if we can. doing a checkdepend run...
whiteknight Coke: how does one get the ability to update docs.parrot.org?
dalek TT #1295 created by whiteknight++: Should FixedPMCArray autovivify nested arrays on set_*keyed()? 14:02
Coke whiteknight: at the moment, you need an account on parrotvm.osuosl.org
though certainly we could setup something to do it automatically pretty easily. 14:03
"check out the tagged release; run make html; copy over the appropriate directories" would pretty much do it.
whiteknight ...and now I have no RT tickets. 7 left
barney whiteknight++ 14:04
Coke that leaves particle (5), infinoid (1), and allison (1). 14:05
whiteknight++
barney Coke: Looks loke foo2 depends on foo 14:12
Coke barney: looks like all those deps are handrolled and don't include the pmc header files anyway. 14:13
Coke added some comments to the ticket.
someone can do a patch job for the release, but the dynpmc deps should be handled just like the PMC deps.
Coke ponders resurrecting his "one makefile to rule them all" strategy. 14:15
14:16 einstein joined 14:18 Zak joined
Coke ah. both gmake and nmake support the 'include' directive; is anyone using a make variant that doesn't? 14:18
mikehh All tests PASS (pre/post-config, smoke (#29986), fulltest) at r42556 - Ubuntu 9.10 amd64 (g++ with --optimize)
Coke :q 14:19
ww
if we had a bunch of generated ".mak" files, where would we put them? 14:20
mikehh I get with smoke #29986 - 11,611 ok, 0 failed, 271 todo, 574 skipped and 0 unexpectedly succeeded
14:23 bluescreen joined
mikehh looking at various recent smoke reports we get 1 failure on MSWin32/i386 - t/op/time.t - not ok 9 - sleep_n slept for at least the integer amount of time specified 14:26
1 failure on solaris/sparc32 but none on sparc64 14:28
dalek rrot: r42557 | barney++ | trunk/config/gen/makefiles/dynpmc.in:
Added explicit dependence foo2$(O): foo.c foo2.c (TT #1294).
 14:29
barney I've seen Test OS.readdir in t/pmc/os.t smolder.plusthree.com/app/public_pr...ails/29987 14:34
14:40 payload joined
Coke is that the same mswin32 failure we always get? 14:43
someone needs to go through and create tickets for those and TODO them.
mikehh no - new one
Coke (smoke needs to be green.0
huh. 14:44
I'd have known that if it was normally green. =-)
what is src/string_private_cstring.h in MANIFEST.generated? 14:47
mikehh we had a failure before with t/op/time.t that seems to have been fixed and now a new failure with t/pmc/os.t - mon MSWin32/i386 at r49556 - thats from smoke reports - I haven't got that platform available
r42556 14:48
14:48 patspam joined
Coke solicits feedback on a make suggestion. 14:54
s/make/build/
14:57 bubaflub joined
barney if the ReadDir - test is skipped on ( $MSWin32 && $MSVC ), why should it work with gcc ? 14:57
14:57 patspam1 joined
whiteknight GeJ fixed the failure in time.t yesterday 14:59
or, I committed his fix
(it was my typo that broke the test)
mikehh got to go out for an hour or so - bbl 15:04
15:05 iblechbot joined 15:09 lucian joined 15:10 Andy_ joined 15:12 Psyche^ joined
barney Is it just me, or it smolder.plusthree.com/ down? 15:33
15:36 plobsing_ joined
allison barney: (from earlier question) only the Jan and July releases go up on CPAN 15:36
dalek website: coke++ | parrot-users list created! 15:38
website: www.parrot.org/news/parrot-users-list-created
barney allison: ok
Coke en.wikipedia.org/wiki/Rakudo_Perl - someone should mention why "way of the camel" is appropriate. =-) 15:40
only 7 tickets left to migrate to trac. How long until #ps? 15:41
15:42 plobsing joined
Coke pmichaud, allison: RT #44471 seems to still be an issue, and pmichaud's suggested workaround doesn't. 15:46
(so if someone can steal RT#44471 and transfer it to trac, +1) 15:47
Coke does it. :|
15:48 plobsing_ joined
allison Coke: rejected it 15:50
Coke: (#44471) the patch is wrong, and it's really just one additional test for TT#389, so I added it as a note there 15:51
15:53 plobsing joined
Coke that leaves one ticket that infinoid opened, 4 from jerry, one from leo. 15:53
allison Coke: looking at Leo's... 15:54
dalek TT #1296 closed by coke++: :vtable is ignored when :anon 15:55
allison Coke: Leo's RT #35391 is one of those really vague ones, that hasn't been closed because no one's quite sure what it's about 15:56
barney is going ahead with the release
Coke allison: implement file:spec:catdir for non-windows. 15:57
barney++ 15:58
allison Coke: aye, it gets into that later in the ticket comments, but the initial task is unusuable
15:59 plobsing_ joined
allison Coke: I'll take a stab at translating it to a sensible TT 15:59
whiteknight barney++
I have developed a particular dislike of vague tickets 16:00
Coke allison: (initial unusable) yes. 16:01
dalek TT #1297 created by Infinoid++: PackFile_Constant_unpack_pmc should call Parrot_thaw_constants(), not ... 16:02
rrot: r42558 | barney++ | trunk (9 files):
[release] Prepare for Parrot 1.8.0
Coke RT: 5
.. 4 16:05
dalek TT #1298 created by particle++: opcodes, warnings, and exceptions
allison Coke: any idea where particle's work on File::Spec::catdir for windows lives?
particle iirc File/Spec.pir has been removed. it was never used, anyway, and the file::spec api on which it was based is *ugly* 16:07
Coke allison: no, that's why i assigned it particle years ago. =-)
RT: 3
allison Coke: this is looking to me like an external module kind of thing, does that get a ticket? 16:08
dalek TT #1299 created by coke++: create tests for Config/JSON.pbc
Coke allison: particle just said it had been removed. if so, reject the ticket. 16:09
allison Coke: will do
Coke (and please cc the list.)
allison Coke: rejected, but before I saw the request to cc the list 16:10
16:11 plobsing joined
allison 2 tickets left! 16:11
Coke allison: just reply to the last reply and cc the list with that. 16:12
whiteknight I'm really starting to believe that we need to completely rethink constant PMCs. It's really not a good implementation that we have now 16:13
allison Coke: okay, done
whiteknight TT #1297 really drives that point home 16:14
TT #956 is a big related issue too
I think we need to be looking much more aggressively at a generational system 16:15
dalek TT #1300 created by coke++: sysinfo_s opcode incomplete
Coke I do not want to migrate 40817, but will do so tonight if I am not beaten to it. 16:16
allison whiteknight: there are lots of definitions of "constant" other than "excluded from GC", we might need to move toward one of them
Coke finding if any of the patches are salvagable would be a good first step.
s/any/the latest/
ah. actually, I think I can reject the ticket outright. 16:17
BOOYAH.
allison Coke: yes, that's where I'd lean
Coke (just had to read through the whole thing to kid51++'s assessment at the end.) 16:19
whiteknight allison: agreed. I would prefer constant meant "immutable" as opposed to "uncollectable" 16:20
allison Coke: kid51++ summary is excellent
NotFound For those who asked the meaning of 'winxed', here's a FAQ: winxed.org/
allison Coke: it's one of those bits of knowledge hidden away in a ticket that I wish we could capture somewhere more permanently 16:21
Coke: "Here's why we have MANIFEST.generated" 16:22
whiteknight RT: 0?
moritz karma RT
purl rt has karma of 29
allison whiteknight: yes, they're really orthogonal features (there's no reason we couldn't have a mutable PMC that's kept out of GC) 16:23
Coke rt++
whiteknight allison: we don't want to keep it out of GC entirely, especially if it has child objects that need to be marked
I can see cases where we don't want the PMC to be collected though
allison whiteknight: immutable still would have to have immutable children, though 16:24
whiteknight doesn't currently (I agree that it should)
NotFound Empty? Zero? Hurrah! 16:26
dalek rrot: r42559 | barney++ | tags/RELEASE_1_8_0:
tagged release 1.8.0
 16:28
whiteknight barney++ 16:29
Coke WOOT. beat the release. 16:31
16:31 mokurai joined
whiteknight w00t indeed 16:31
Coke Now someone fix my (*#&@$(#& segfaults. =-) 16:32
dalek TT #1301 created by coke++: Update RT references in repository/wiki
Coke allison: any feedback on my makefile cleanup suggestion?
16:32 darbelo joined
allison Coke: mailing list or ticket? 16:32
Coke allison: mailing list 16:33
I figured I'd get some feedback before making a ticket.
allison looks...
purl looks are free today :)
Coke whiteknight: is TT #1276 closable?
whiteknight looking
Coke TT 1.8.0 tickets - trac.parrot.org/parrot/query?statu...estone=1.8 ... these either need to be closed or have their milestone bumped/cleared. 16:34
whiteknight yeah, I think I will close it. The issue still exists, but is really part of the "taking too much config infomation from Perl5's defaults" problem
I have not been able to get it to build with strawberry perl but MSVC compiler
Coke and I think that's a combination we don't have to bend over backwards to support. 16:35
whiteknight does build with strawberry perl and mingw, or MSVC and activeperl
fair. I'll close it
Coke danke.
allison Coke: it looks like a sensible, gradual refactor, good improvement
Coke k. I'll see about doing some splitting first to make sure 'include' works as expected. 16:36
allison Coke: I'm not sure about platform support for includes, it's worth running a small test
Coke yah. figured I'd do the first one as a patch and ask for feedback.
nmake and gmake look doable.
figured that's 90% if not 100
any suggestions for where to keep .mak files? 16:37
allison Coke: would config/ work?
Coke config/build ?
allison (we're moving toward fewer top-level items and directories, so moving things downward) 16:38
Coke works for me.
dalek TT #1276 closed by whiteknight++: Parrot does't build on Win32 with MSVC
allison there will probably be a top-level build/ directory that contains everything that will be installed
config/build works, or config/make
barney ftp://ftp.parrot.org/pub/parrot/releases/devel/1.8.0/ has landed
whiteknight barney++
allison barney++ 16:39
darbelo barney++
whiteknight barney: change the topic?
purl whiteknight: that doesn't look right
Coke config/make works.
allison Coke: hmmm... looking at the existing directories in config/ I'd go with config/make. they're all confused, and hard to tell what goes where 16:40
config/make nice and clear
whiteknight: what's the status of :call_sig? partially implemented? 16:41
whiteknight: (I just re-encountered your nopaste message about it, but nopaste has expired the code blocks) 16:42
16:43 kj joined
whiteknight allison: I committed that naive first attempt I sent you. Am going to rework it this week hopefully to not suck 16:43
Coke yay, good news from andy. 16:44
moderator Parrot 1.8.0 Zygodactyly released | Latest modified TT's: icanhaz.com/parrotbugs 16:44
whiteknight raises a slight question: What types of objects can I use as a :call_sig on the caller's side? 16:44
allison whiteknight: is there a ticket for that? would be good to have a record of the progress, in case someone else wants to pick it up
whiteknight allison: will create a ticket, yes
dalek tracwiki: v6 | allison++ | AllisonTasklist
tracwiki: trac.parrot.org/parrot/wiki/Alliso...ction=diff
allison whiteknight: it's polymorphic, should accept anything with the necessary interfaces 16:45
whiteknight: (which is currently only CallSignature, but anything could mock it)
whiteknight: (or subclass it)
Coke: ah, make and dmake too, excellent
whiteknight allison: okay, is there a role or something that we can check for? My initial attempts ran into problems trying to be too permissive, which lead to a lot of very ugly special-purpose code 16:46
or do we not check, and jsut assume that a :call_sig has the necessary interfaces (allowing "VTABLE not found" exceptions to handle the problems)?
allison whiteknight: how about we declare CallSignature as "provides callsig", and then check for it? 16:47
whiteknight okay, we can do that
allison whiteknight: then anything else that provides callsig is agreeing to provide the complete interface of CallSignature
whiteknight allison: Okay. Then it would be a very good idea to document exactly what that interface is 16:48
can work on that this week
Coke whiteknight: that's a deficiency of provides in general. 16:49
barney ./parrot tools/util/crow.pir --type=text is broken
Coke well, that sucks.
purl The rock is now off.
whiteknight purl has been hanging around in #perl too often. She has a very dirty mind 16:55
purl whiteknight: huh?
whiteknight I know, purl, it is confusing 16:56
16:57 cotto joined
whiteknight dukeleto: ping 16:58
17:10 kj joined
Coke barney: generating docs from RELEASE_1_8_0 ... 17:18
whiteknight purl msg dukeleto: mind if I join on to the nqptap project?
purl Message for dukeleto stored.
barney Coke: tnx 17:24
Coke barney: PITA to do this from a tag. you have a release lying about? 17:26
dalek website: barney++ | Parrot 1.8.0 "Zygodactyly" Released! 17:32
website: www.parrot.org/news/2009/Parrot-1.8.0
dukeleto whiteknight: pong 17:33
whiteknight: will add you nowishly
whiteknight i msg'd you
w00t. dukeleto++
cotto_work barney++
dukeleto barney++ 17:34
whiteknight dukeleto: I also have a git question for you 17:35
I merged the pla_integration branch of matrixy into head and deleted the branch locally. How do I get it to disappear from github too?
moritz git push origin :pla_integration 17:36
dukeleto what moritz said
whiteknight the colon on front deletes it?
dukeleto whiteknight: git push remote foo:bar means push foo to the remote branch bar. if foo = NULL => delete remote 17:37
cotto_work wishes people were octadactyls
dukeleto delete the remote branch bar, i mean
whiteknight ok 17:38
yay! worked! moritz++ dukeleto++ 17:39
bubaflub i have to google how to delete a remote branch on github about once a week
dukeleto bubaflub: hola! yes, it is slightly non-intuitive until you git-warp your brain 17:40
whiteknight I have stacks of printed git cheatsheets, and still can't ever find the commands I want
Coke is there a way to use sftp to do plain ftp? 17:44
darbelo Coke: nopes.
Different protocol.
Coke is there a way to use http to download parrot?
darbelo has no idea. 17:45
Coke Doesn't appear to be. what is this, 1989?
darbelo Well we *do* restrict ourselves to C 89. 17:46
cotto_work That's surprising.
whiteknight I can't imagine that my brain is ever going to be compatible with git. I use it, but it's command line interface makes no intuitive sense whatsoever
bubaflub well, now what little grasp i had on svn is now gone 17:47
dukeleto whiteknight: you are on nqptap now. feel free to make stuff work
bubaflub git is so much easier esp. branching
whiteknight dukeleto++
dukeleto purl, nqptap is nqpTAP - TAP (Test Anything Protocol) Harness for NQP (NotQuitePerl) github.com/leto/nqptap
purl OK, dukeleto.
dukeleto anybody else want a commit bit to nqptap?
darbelo dukeleto: Hook me up. 17:48
dukeleto darbelo: done
cotto_work nqptap?
purl i think nqptap is nqpTAP - TAP (Test Anything Protocol) Harness for NQP (NotQuitePerl) github.com/leto/nqptap
darbelo Busy untill next week, but I'll dive in later.
17:51 lucian joined
pmichaud Coke: (RT #44471) -- change my workaround to be .sub "" :vtable('get_integer') (i.e., no :anon flag) 17:54
dalek tracwiki: v121 | barney++ | WikiStart
tracwiki: Parrot 1.8.0 released
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
Coke pmichaud: think allison already rejected the ticket.
pmichaud yeah, just noticed that 17:55
anyway, the workaround should just avoid the :anon flag altogether
time for lunch here
#ps in 30 (I will miss it, though) 18:01
Coke barney: docs.parrot.org updated. 18:04
and for 1.9.0, I'll have a script to make it easier. =-)
dalek kudo: 68d5c37 | (Kyle Hasselbacher)++ | tools/update_passing_test_data.pl:
tools/update_passing_test_data.pl will fork according to $ENV{TEST_JOBS}
 18:08
Coke is bemused that our ftp links are www redirects and that our http links are ftp ...
barney Is any of trac.parrot.org/parrot/query?statu...estone=1.8 resolvable ? 18:09
Otherwise I promte it to 1.9.0
Coke we can review it during PS, I suppose. 18:10
barney Coke++
dalek TT #583 closed by coke++: http download links should be exposed on download page 18:11
18:17 chromatic joined
allison Coke: ftp.parrot.org/releases/ 18:22
Coke: (was backscrolling, see you found it)
Coke: our http redirects probably should go to the http:// way of downloading the files (for consistency) 18:27
Coke: the http download didn't exist when we first created the http redirects
18:29 bacek joined
mikehh I make it #ps time 18:30
dalek TT #1302 created by Util++: PIR todo() is frequently misused 18:44
18:47 joeri joined 18:51 nbrown joined 18:59 Zak joined
dalek tracwiki: v1 | chromatic++ | DevelopmentPriorities 19:01
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
tracwiki: v2 | coke++ | DevelopmentPriorities
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
whiteknight had to choose between GC and JIT. Not sure if he chose wisely 19:02
chromatic This is for 2.0, remember. 19:03
Coke chromatic: ... I thought it was for /this week/ 19:04
whiteknight I think big
Coke signs.
Coke sighs, even.
whiteknight maybe we shoudl break up that page "This month", "next 3 months", "next year", etc
Coke I miss bugzilla.
dalek tracwiki: v3 | whiteknight++ | DevelopmentPriorities 19:05
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
tracwiki: v4 | coke++ | DevelopmentPriorities
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
19:07 gaz joined
dalek tracwiki: v5 | chromatic++ | DevelopmentPriorities 19:08
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
bubaflub Util: i might be guilty of misusing todo() ( TT #1302) 19:09
Coke (away) 19:10
Util bubaflub: you are not alone :) Watch the ticket for my update showing how the current interface encourages misuse. 19:12
bubaflub Util: ok. if ya point out where i'll put up some patches for my tests. 19:13
Coke (back) 19:16
dalek tracwiki: v6 | cotto++ | DevelopmentPriorities 19:18
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
Coke someone clobbered my edit. 19:19
trac--
bubaflub karma trac 19:21
purl trac has karma of -15
bubaflub sounds about right
dalek tracwiki: v7 | coke++ | DevelopmentPriorities
tracwiki: That was not a 2.0 priority but a RIGHT NOW priority.
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
Coke cotto_work: so, right now, you can replace Array.pmc with a copy of RPA.pmc 19:22
dalek TT #1303 created by cotto++: [DEPRECATION] Array PMC 19:24
cotto_work Coke, I'd much rather leave it as-is and clobber it after the 2.0 release.
I hate code duplication like that, even for doomed code. 19:25
19:33 lucian joined 19:35 mikehh joined
pmichaud back 19:35
Coke cotto_work: you could duplicate it programmatically... but leaving it there to rot is fine too. 19:39
dalek rrot: r42560 | coke++ | trunk/docs/parrothist.pod:
coke > Will ; mark supported releases; ws to highlite the last supported rel.
 19:40
19:44 hercynium joined 19:45 mikehh joined 19:47 zak_ joined
dalek TT #1304 created by dukeleto++: Parrotcode.org docs show up before parrot.org on web searches 19:48
19:56 mikehh joined
chromatic msg bacek_at_work <chromatic> Who has the responsibility of creating the merged Context/CallSig PMC and where does it happen? 20:00
purl Message for bacek_at_work stored.
chromatic <allison> ultimately, the responsibility is the caller's. unlike currently, where the context is created inside invoke, if they're merged the context will be created long before reaching the invoke vtable
allison chromatic: I think the second half of that message wasn't stored 20:04
chromatic msg bacek <allison> ultimately, the responsibility is the caller's. unlike currently, where the context is created inside invoke, if they're merged the context will be created long before reaching the invoke vtable
purl Message for bacek stored.
chromatic msg bacek_at_work <allison> ultimately, the responsibility is the caller's. unlike currently, where the context is created inside invoke, if they're merged the context will be created long before reaching the invoke vtable
purl Message for bacek_at_work stored.
chromatic That ought to do it.
dalek tracwiki: v7 | allison++ | AllisonTasklist 20:06
tracwiki: trac.parrot.org/parrot/wiki/Alliso...ction=diff
whiteknight I think that's a very good idea. It gives us the ability to access the callee's context before we even invoke it
chromatic: do you have a general estimate for how much work it will take to merge CallSignature/Context? 20:07
chromatic I haven't looked at the branch, but it seemed like bacek had it mostly working. 20:08
I think it's mergeable for 1.9.
... in the amount of work it should take.
dalek tracwiki: v8 | dukeleto++ | DevelopmentPriorities 20:09
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
20:12 lucian joined
dalek tracwiki: v9 | whiteknight++ | DevelopmentPriorities 20:12
tracwiki: Remove my suggestion for now.
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
whiteknight I need to pick somehing better 20:13
dalek rrot: r42561 | coke++ | trunk/docs/intro.pod:
docs cleanup

  - use say, not print "\\n"
  - just point to download page, don't duplicate it
pmichaud reminds folks that trac.parrot.org/parrot/wiki/RakudoTasklist might be a good pool of ideas 20:15
darbelo wants IMCC dead. 20:16
dukeleto pmichaud: i didn't even know about that
KILL KILL KILL IMCC
purl, imcc?
purl i guess imcc is the c of parrot or the problem or the intermediate code compiler
whiteknight kiling IMCC might be the best idea, yes
so that will have my vote then
dukeleto purl, imcc is also will be murdered soon
purl okay, dukeleto.
chromatic pmichaud, can you think of a couple of representative NQP-rx examples which would be good to profile? 20:17
20:17 mikehh joined
dukeleto chromatic: plumage's test harness would be a good one 20:17
chromatic: it is a "real world" example 20:18
chromatic Can you add those to your Valgrind benchmark suite?
darbelo dukeleto: doesn't that spend too much time running other programs?
dukeleto chromatic: which exact versions of everything do you want benchmarked? 20:19
chromatic I don't know; I haven't run any tests yet.
darbelo A test harness spends most of it's time waiting for the tests to run.
chromatic Oh.
dukeleto darbelo: you can just give it a no-op test that prints out some TAP. the time spent in test can also be subtracted
chromatic Let's start from the current HEAD and progress.
dukeleto darbelo: a test harness spends time parsing TAP streams
pmichaud chromatic: A good representative NQP-rx example would be the NQP source itself :)
dukeleto chromatic: which version of NQP-rx, the one in parrot core? 20:20
NotFound That's the beauty of bootstraping.
pmichaud i.e., run nqp on the src/NQP/Grammar.pm and src/NQP/Actions.pm files from the nqp-rx repository
just run them with --target=pir
i.e.
./parrot nqp-rx.pbc --target=pir Grammar.pm
dukeleto pmichaud++ # what he said 20:21
pmichaud: does nqp-rx have any benchmark scripts in the repo? i would be interested in writing some
pmichaud: can i get a commit bit on nqp-rx to make branches, or do you want me to fork on github and do pull requests? 20:22
chromatic I'm adding STRINGNULL right now, but I'll look at the benchmark profile shortly.
pmichaud dukeleto: not as yet. I've been thinking it would be useful to rewrite some of the parrot benchmarks into nqp and see how much we lose (and where/why)
dukeleto: what are you wanting to branch/add ? 20:23
darbelo chromatic++; STRINGNULL++
Coke dukeleto is also unaccountably violent.
purl okay, Coke.
chromatic I'd like to see NQP ports of PIR benchmarks. 20:24
Coke first step in killing imcc is allowing pirc to be swapped in via a config option.
(then we can see where all the tests fall down.)
dukeleto pmichaud: i want the ability to create topic branches in nqp-rx, for when I want to add tests and/or benchmark scripts 20:25
Coke: :)
Coke++ # config option to choose pirc
NotFound Coke: First step will be rewritting main options and arguments parsing, in order to be able to do that.
pmichaud dukeleto: what's your github id?
dukeleto pmichaud: leto
pmichaud dukeleto: you now have a commitbit. 20:26
You can commit tests and examples directly to the master branch, if you wish.
dukeleto pmichaud: thanks, much appreciated! should I add myself to CREDITS or something?
pmichaud dukeleto: sure 20:27
darbelo NotFound: Somebody had posted code to turn command line options into a FPA of Strings a while back. Is that available anywhere?
dukeleto pmichaud: sounds good, I will do some nqp-rx hacking later today! thanks again!
darbelo Or maybe it was a RPA
NotFound darbelo: I don't remember seeing that.
darbelo It happened around the time cotto started the profiling runcore. 20:28
pmichaud command line arguments already come in as an RSA, I think
cotto_work pmichaud, not all of them.
NotFound darbelo: but I remember some comments about the need of processing options before creating an interpreter.
cotto_work only the ones after the script to be run (foo.pir, x.nqp, etc) 20:29
er, hll.pb, etc
*pbc
pmichaud cotto_work: right.
darbelo NotFound: Thinking about it, we might need a two pass scheme for options. 20:30
cotto_work The trick is that to use an RPA for CLI args you need Parrot to be somewhat initialized.
20:31 payload joined
darbelo cotto_work: Exactly. If the GC is selectable on the command line, we can't really allocate PMCs. 20:31
NotFound And if we want to do it right, we need some information about locale and charsets.
Assuming they are always utf8 is wrong.
More general, any assumption is wrong. 20:32
whiteknight cotto had nopasted a routine to convert argc/argv into an FSA at one point 20:33
darbelo NotFound: We could do a 'first pass' that initilaizes parrot based on good old getopt-ish options.
whiteknight I don't think that code ever got put in
NotFound darbelo: getopt is not ansi std 20:34
cotto_work istr doing something like that
whiteknight it's trivial code to rewrite if needed
darbelo NotFound: I said "getopt-ish" 20:35
NotFound darbelo: we already have that.
darbelo NotFound: Yes, in IMCC. We need it out of IMCC. 20:36
Coke NotFound: we can link in /both/ imcc and pirc and just make pirc trump imcc as the compiler.
NotFound darbelo: include/parrot/longopt.h
Coke (as a first pass. but doing it right++) 20:37
NotFound Is not dependant on imcc, just used by it.
bacek_at_work good morning 20:38
cotto_work hi bacek_at_work
darbelo Ok, we can start by moving that into the main parrot executable.
Coke leto: pinged the perl.org admins to see how they can help us with that TT. 20:41
fperrad seen barney 20:45
purl barney was last seen on #parrot 2 hours, 34 minutes and 59 seconds ago, saying: Coke++
Coke pmichaud: message for you on parrot-users. =-) 20:46
purl Message for you stored.
Coke purl, you idiot! 20:47
purl Coke: what?
Coke purl, owner?
purl owner is hachi, see also #purl
Coke purl: msg hachi "pmichaud: message for you on parrot-users." causes a message for "you" to be saved. ^_O?
purl Message for hachi stored.
hachi hah 20:48
for the record, you could just see if I'm awake too
Coke oh, didn't even figure you'd be in here. =-) 20:54
hio. =-)
dukeleto hola 20:56
purl bonjour, dukeleto.
dalek tracwiki: v10 | whiteknight++ | DevelopmentPriorities
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
nopaste "chromatic" at 72.87.39.97 pasted "Symbol visibility problem with STRINGNULL; any thoughts?" (109 lines) at nopaste.snit.ch/18737 20:57
chromatic My C fu is low today.
dalek TT #1305 created by darbelo++: [DEPRECATION] The IMAGE_IO structure. 20:58
chromatic I understand what the error means, but I don't understand why: blib/lib/libparrot.so: undefined reference to `STRINGNULL' 20:59
... nor from where.
dalek rrot: r42562 | darbelo++ | trunk/DEPRECATED.pod:
Add the IMAGE_IO structure and associated VTABLEs to DEPRECATED.pod
 21:00
NotFound chromatic: Do you define it?
Coke (in a .h that is included by a C file?)
(or a linked in C?) 21:01
NotFound The .h just declares it, it also needs to be defined.
whiteknight darbelo++
fixing freeze/thaw and making them sane would be awesome
NotFound Don't remember now where is PMCNULL defined... 21:02
whiteknight cd include/parrot && ack PMCNULL
chromatic A few lines before STRINGNULL in parrot/interpreter.h
Andy_ or ack PMCNULL include/parrot
darbelo whiteknight: I can, at best, disentagle them from the string and io guts.
Fixing is a bigger endeavor than I can tackle right now. 21:03
chromatic As far as I can tell, PARROT_DATA PMC *PMCNULL; and PARROT_DATA STRING *STRINGNULL; should do the same thing.
whiteknight darbelo: that would be enough
dalek p-rx: 19c51c8 | duff++ | CREDITS:
Add myself to CREDITS
 21:05
NotFound chromatic: src/pmc.c:68 21:08
chromatic Ah. 21:09
Because PARROT_DATA does extern "C" and... yes. 21:10
NotFound In C++ jargon, is declared but not defined. 21:11
chromatic lambda-the-ultimate.org/node/3674
I don't speak crazy.
dalek rrot: r42563 | chromatic++ | trunk (3 files):
[string] Added STRINGNULL. Now we can use it appropriately instead of assuming
 21:16
chromatic Now comes the fun part. 21:17
ttbot Parrot trunk/ r42563 i386-linux-thread-multi make error tt.ro.vutbr.cz/file/cmdout/146922.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ )
darbelo fixing the build? 21:18
NotFound I'm going to check if we have char * NULL nci tests...
21:19 cognominal joined
chromatic Ugh, that's a non-optimized build, isn't it? 21:21
darbelo Yep.
cotto_work very broken atm
chromatic No problems here. 21:23
Coke chromatic: broken for me here also non-optimized.
chromatic Did anyone realclean first?
Coke always
nopaste "darbelo" at 190.192.220.13 pasted "bt of build failure on OpenBSD i386" (17 lines) at nopaste.snit.ch/18739
darbelo I realcleaned.
chromatic Oh, I see. 21:24
Coke ORLY? 21:25
purl YA RLY.
darbelo Coke: It's a violated assertion. A encodingless string got created.
NotFound chromatic: I assume that for a 't' NCI type we'll want to convert a STRINGNULL to a NULL char * for argument, and the recyprocal for return type, isn'it?
Coke NotFound: sounds reasonable. 21:26
chromatic Yes, I think so.
dalek rrot: r42564 | chromatic++ | trunk/src/string/api.c:
[STRING] Ensured that charsets and encodings get initialized before
NotFound I'll try to add a test for that. 21:27
Coke chromatic: that's better. 21:29
darbelo chromatic: WORKSFORME
Coke chromatic: "not in the constant table", but PObj_constant_FLAG ?
Util I'll be back in 90 minutes. I still have more to say on TT#1302. 21:31
Coke I wonder why we can't just tack on an optional isTODO flag to is and have it modify the description of the test file. 21:32
ttbot Parrot trunk/ r42564 MSWin32-x86-multi-thread make error tt.ro.vutbr.cz/file/cmdout/146983.txt ( tt.ro.vutbr.cz//buildstatus/pr-Parrot/rp-trunk/ ) 21:35
21:36 theory joined
darbelo _STRINGNULL 21:36
??
Where did that freaking underscore come from?
MSVC--
GeJ Good morning everyone. 21:37
NotFound darbelo: C decorations
Coke NotFound: we'd expect that on any c++, yes?
guessing it needs a PARROT_API
NotFound Coke: C++ ones are a lot worse. 21:38
darbelo Oh, it's a *linker* error. I had totally misread that. 21:39
Still. MSVC--
NotFound I think parrot_debugger must use the function, not the macro.
chromatic Coke, the constant table is a cache. It *is* in the constant pool. Subtle.
Coke too subtle for me. as long as it makes sense to you. =-) 21:40
chromatic parrot_debugger should probably use the function, but it might also need to include parrot/interpreter.h 21:41
... which seems like a bad place for these macros.
darbelo It does #include "parrot/embed.h"
We might want to give embedders access to STRINGNULL. 21:42
chromatic That should pick it up then.
NotFound Maybe we need a parrot/internals.h header instead of using macros IN_CORE style.
darbelo I only see one use of STRING_IS_NULL() at line 255 21:43
dalek tracwiki: v11 | japhb++ | DevelopmentPriorities
tracwiki: Add documentation goal with subgoals
tracwiki: trac.parrot.org/parrot/wiki/Develo...ction=diff
NotFound darbelo: and then we'll have the most unstable vm in the world.
chromatic When every .c file includes parrot/parrot.h, we have to rebuild a lot of code.
darbelo NotFound: Huh? 21:45
NotFound darbelo: if we make it viewable, it gets modifiable. A little mistake and... BLAM!
Coke chromatic: I'm fixing that. 21:46
er, no i'm not.
(I'm fixing a bunch of other things, but not that. =-)
darbelo Oh. I got embed/extend mixed up again, didn't I? 21:47
21:47 jan joined
NotFound Uh... I think the NCI t thing already has a problem. Looks line PCC is converting a returned NULL to an empty string. 21:49
21:51 Dex joined
NotFound I'll look better at that tomorrow, too late for such a subtlety. 21:52
21:53 cognominal_ joined
darbelo dukeleto: ping 21:57
dukeleto darbelo: pong 21:58
darbelo You did some work on the debugger, right? 21:59
In main(), starting at line 247, what's the point of compiling that snippet of code if there's no arguments? 22:03
dukeleto darbelo: let me see 22:04
darbelo: i didn't write that code and I have no idea :) 22:05
darbelo: Infinoid would know
darbelo Nevermind, without it it has no runloop and segfaults before the (pdb prompt) 22:07
NotFound darbelo: I wrote that. I silly, but people asked for a quick way to start without any pbc loaded. 22:09
darbelo Yeah, PDB_run_code(), unsurprisingly needs code to run ;) 22:10
22:19 cognominal_ joined
cotto_work . 22:22
darbelo ..
cotto_work That's what I get for examining the back of my keyboard.
22:36 zak_ joined 22:43 cconstantine joined, cognominal_ joined
Coke dukeleto: you posted to parrot-users without joining! (blah) 22:49
22:49 jan joined
dukeleto Coke: i need to do that, ooops 22:51
Coke: i am on parrot-users now 22:58
23:00 mokurai left
dukeleto languages? 23:04
purl i think languages is trac.parrot.org/parrot/wiki/Languages
23:09 lucian joined 23:10 Whiteknight joined
darbelo plparrot 23:10
purl it has been said that plparrot is the postgres+parrot integration project or github.com/leto/plparrot
23:12 zak_ joined 23:13 cognominal joined 23:16 Zak joined
darbelo dukeleto: ping 23:26
nopaste "darbelo" at 190.192.220.13 pasted "Explodey PL/Parrot Configure.nqp for dukeleto." (12 lines) at nopaste.snit.ch/18743 23:28
23:29 Whiteknight joined
dukeleto darbelo: pong 23:36
darbelo dukeleto: see nopaste.snit.ch/18743 23:37
PL/Parrot Configure.nqp went kaboom.
dukeleto darbelo: that configure.nqp is just for show :)
darbelo: look at HOWTO
darbelo: but it should work, in theory. we need to be able to detect pg_config 23:38
darbelo: also, i think that worked on some nqp, maybe not the one the you have
Whiteknight did the parrot_nqp binary disappear from this newest release?
dukeleto since nqp still has no version number, this is getting annoying
Whiteknight: parrot-nqp now
darbelo Whiteknight: it transformed into parrot-nqp
Whiteknight ah, yes 23:39
darbelo dukeleto: I thought HOWTO was post-Configure. I'm using nqp that whips with parrot.
s/whips/ships/
dukeleto darbelo: i think i may have only tried nqp-rx master, not nqp-rx-in-parrot-core
dukeleto thinks about whipping with parrot 23:40
ETOOMANYNQPSWITHNOVERSIONNUMBERS
darbelo old nqp is nqp.pbc, new nqp is parrot-nqp, and even newer nqp is 'nqp' 23:41
japhb Now that parrot-nqp exists, just use it, and consider Parrot's revision to be the version number of nqp-rx. pmichaud seems to be doing this from time to time himself even. :-)
Barring an ~1 day delay, there's really not much point in riding nqp HEAD 23:42
darbelo japhb: Until 2.0 then the deprecation police will lock you up if you break compatibility. 23:43
Whiteknight pla Configure.nqp is broken now 23:44
Whiteknight is pumped about that
japhb darbelo, only parrot_nqp is right now subject to that deprecation policy, not parrot-nqp
darbelo unless we declare all of ext/ to be excepted from such things.
nopaste "kiwichris" at 58.172.130.100 pasted "RTEMS: examples/json/postalcodes.pir fails" (9 lines) at nopaste.snit.ch/18744
darbelo Whiteknight: you want to run it with parrot $(BUNCH_OF_ARGS) nqp.pbc 23:45
That't the 'old' nqp left there until 2.1 to comply with the deprecation policy.
Whiteknight darbelo: I don't understand
darbelo the parrot-nqp fakecutable is the new not-totally-compatible nqp-rx that we imported from pmichaud's github repo. 23:46
the old nqp that shupped with parrot 1.7 is still available in pbc form, but there is no fakecutable anymore. 23:47
Whiteknight okay, so what do we have to do to make Configure.nqp work with nqp-rx? 23:48
I had been lead to believe that it wasn't a backwards-incompatible change
darbelo Whiteknight: The only incompatibility is the lack of a fakecutable. The nqp that can tun our Configure is still shipped and installed. 23:49
s/tun/run/
japhb Whiteknight, it's very close to compatible, but not quite. Thus having to keep the old one for two more releases.
Whiteknight darbelo: that's all well and good. How do we make Configure.nqp work with the fakecutable?
I'll fix it once I know what the differences are
japhb Whiteknight, is your Configure.nqp forked from Plumage's? If so, just diff against Plumage's master copy. 23:50
darbelo japhb and dukeleto know that better than I do.
japhb Plumage always uses parrot-nqp (the NQP-rx checked into parrot) for everything now.
Whiteknight japhb: is there are particular diff or diff range I can plunder? 23:51
japhb Whiteknight, looking
darbelo Whiteknight: I can do it for you. Gimme a sec.
japhb gitorious.org/parrot-plumage/parrot...figure.nqp 23:52
Looks like just the top 3 commits there
darbelo Whiteknight: Did it. 23:54
Whiteknight w00t
darbelo++
darbelo Whiteknight: pushed it. 23:55
It was an issue with load_bytecode()
Whiteknight ok 23:56
pmichaud the new version of nqp is nqp-rx.pbc
and parrot-nqp
the old parrot_nqp fakecutable is now gone
dalek rrot-linear-algebra: 9b112a2 | darbelo++ | Configure.nqp:
Update Configure.nqp to work with the new nqp-rx in parrot.
darbelo pmichaud: Maybe we should have kept the old fakecutable.
pmichaud sorry, nqprx.pbc
darbelo: the old fakecutable has only been around for < 1 month
it would just cause confusion. 23:57
japhb Oh man, would it. The current state is bad enough.
(I understand the reasons for the current situation, but it's quite a FAQ)
pmichaud eh, it's not that bad.
if you want to use the old nqp, you continue to use nqp.pbc, as before. 23:58
if you want to use the nqp, you use parrot-nqp
if you want to profile the new nqp, use nqprx.pbc
that's it.
s/use the nqp/use the new nqp/
japhb *I* understand it. I'm just saying people who weren't actively involved, but just kindof peripherally watching, for the last couple months ended up confused. Not much we can do about it, just the way these things go. 23:59
I guess I should be happy that so many people had even half an eye on it. ;-)

Powered by Cro and the Raku® Programming Language.

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

Thank you!