Parrot 3.2.0 released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Goals: Get more GSoC ideas on wiki; close tickets; stable 3.2 release; assess status of roadmap goals for 3/15 meeting.
Set by moderator on 15 March 2011.
whiteknight plobsing: unping 00:04
plobsing whiteknight: unpong then 00:08
whiteknight sorry, I was trying to find out how to mark a PackFile*. I figured out that I had to write a new function forit 00:09
plobsing there is a function that marks all the GCable junk the current interp's packfiles. I'd factor the marking stuff from there. 00:10
s/factor/factor out/
whiteknight that's what I did
mark_const_subs -> Parrot_pf_mark_packfile(PARROT_INTERP, PackFile*pf)
plobsing it does a lot more than marking const subs though. much better name. 00:11
00:11 dmalcolm left
dalek rrot/whiteknight/imcc_compreg_pmc: d82abdc | Whiteknight++ | / (6 files):
Implement a new Parrot_pf_mark_packfile to mark an arbitrary packfile. Delete mark_const_subs, which only marked the ->initial_pf and was named poorly. IMCC now outputs a PtrObj PMC instead of an UnManagedStruct
 00:14
whiteknight plobsing: can you take a look at that last commit and make sure what I added, especially with the PtrObj is sane?
plobsing whiteknight: looks to be good 00:18
whiteknight awesome
plobsing have you figured out your nci_thunks problem yet? 00:19
whiteknight what nci_thunks problem? 00:20
I have so many problems it's hard to keep track
plobsing the one you got executing that program during the build
00:21 cosimo joined
whiteknight I have no idea, to be honest with you 00:21
if I can't remember, it's probable that I haven't had a problem in a while
plobsing OK. awesome. 00:22
dalek rrot/whiteknight/imcc_compreg_pmc: 713860f | Whiteknight++ | / (24 files):
fix merge, it looks like just some param decorations. Go with whatever master uses
 00:25
whiteknight ...and that's all the effort I am spending on this issue tonight 00:26
plobsing whiteknight: sorry to bug you about this, but I just thought of it. you shouldn't use SETATTR* for PtrObj attributes. use the custom macros PTROBJ_SET_*. 00:31
whiteknight oh, I see that
oay
okay 00:32
thaks
....stupid fingers tonight
00:35 davidfetter left
whiteknight FFFFFUUUUUUU 00:36
after all this, removing the hackjob from TT #1990 causes parrot to segfault
plobsing lolwhut? 00:37
dalek rrot/whiteknight/imcc_compreg_pmc: 7960354 | Whiteknight++ | compilers/imcc/main.c:
use the correct macro for setting the mark function on PtrObj. plobsing++ for pointing this out
plobsing try pinning the PMC? 00:39
whiteknight I don't even know what the problem is. I'll need to dig through it tomorrow 00:40
reading bacek's blog posts is making me remarkably happy 00:57
00:57 woosley joined
bacek_at_work whiteknight, glad to hear it :) 00:57
whiteknight okay, I need to go to bed now 01:03
goodnight
01:04 whiteknight left, woosley left
cotto ~~ 01:20
bacek_at_work ENOMATCH 01:22
cotto is matchless 01:24
01:52 bubaflub left 02:09 eternaleye_ joined, eternaleye left 02:33 ShaneC left
dukeleto ~~ 02:43
dukeleto had the pleasure of buying kid51++ a few beer tonight and talking about parrot stuff
bacek_at_work: i didn't know that magic coding robots could wrote blog posts ;)
s/wrote/write/
bacek_at_work: it is very nice to see an explanation of the opsc stuff you have been hacking on 02:49
dukeleto hears crickets 02:59
03:15 ShaneC joined
dukeleto ShaneC: howdy 03:27
ShaneC good evening 03:39
03:39 dafrito left
dalek rrot-embed: 0e28feb | dukeleto++ | Build.PL:
Make 'perl Build.PL' work outside of the Parrot source tree
 03:53
rrot-embed: 30bbf19 | dukeleto++ | .gitignore:
Add a .gitignore
 04:01
cotto bacek++ for those. I like reading them too. 04:19
04:23 JimmyZ joined
cotto dukeleto, ping 04:26
dukeleto cotto: pong
cotto dukeleto, do you know when oscon will accept or reject talks? 04:27
waiting is lta
dukeleto cotto: no clue. they do what they will 04:29
cotto guess so
dukeleto cotto: i hung out with kid51++ tonight and talked about parrot stuff 04:30
cotto awesome
dukeleto cotto: also, i created the parrot-embed repo
cotto what thunk he?
04:30 cosimo left
cotto dukeleto, how'd it go? 04:32
dukeleto cotto: it was fun 04:36
cotto: we talked about planning various parrot hackathons for some confs this year 04:37
cotto I'm looking forward to those.
dukeleto cotto: what are you hacking on nowishly? 04:43
cotto: we need to get serious about making the M0 prototype 04:44
cotto dukeleto, figuring out if/how plobsing and bacek's suggestions about ffi can be used 04:45
I'm glad they were brought up. I'd forgotten to think about a C-emitting interp. 04:46
er, compiler
er, code-generation backend
04:50 JimmyZ left
cotto dukeleto, do you have any thoughts on that? 04:54
it's also easier to generate thunks, since the amount of storage needed for the arguments will be knowable at parse-time 04:57
the idea is that if creating a call (static thunk, generated thunk, function call, etc) looks like gen_call $P0, [INT32, INT32, $I0, INT32, $I1], a backend that generates c will have an easier time generating code for that.
(reverse those last two sends)
dukeleto cotto: i will think about. 04:58
dukeleto needs sleep
cotto dukeleto, 'night 04:59
05:19 JimmyZ joined 05:59 JimmyZ left 06:09 jsut joined 06:14 jsut_ left 06:23 alin joined 06:41 theory left 07:04 rurban_ joined 07:06 rurban left 07:07 rurban_ is now known as rurban 07:12 fperrad joined 07:44 alin left 08:10 mtk left
tadzik hello 08:17
08:18 mtk joined 08:40 alin joined 08:50 mj41_nb joined 09:07 contingencyplan left 10:29 ligne joined 11:09 JimmyZ joined
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#12473) fulltest) at 3_2_0-2-g1868a49 - Ubuntu 10.10 i386 (g++-4.5) 11:22
11:57 bluescreen joined 12:08 lucian joined 12:15 JimmyZ left
Coke bacek_at_work: t'weren't nuthin. 12:30
12:34 whiteknight joined, mj41_nb left
whiteknight good morning, #parrot 12:37
12:47 mj41_nb joined
whiteknight msg NotFound Can winxed have a syntax to add properties (setprop) on class and method PMCs? 12:51
aloha OK. I'll deliver the message.
12:52 mtk left, mtk joined 13:05 particle joined, mtk left 13:12 mtk joined 13:19 particle left 13:24 particle joined 13:25 particle left
dukeleto ~~ 13:30
whiteknight good morning, duke 13:41
dalek rrot: 4e9e576 | mikehh++ | src/gc/api.c:
add default to supress compiler warning and update documentation
 13:42
mikehh was trying to do the same to src/pmc/structview.pmc but having a few problems there 13:43
13:47 contingencyplan joined
dukeleto whiteknight: mornin' ! 13:49
whiteknight: what hacketh you on this morn? 13:51
atrodo dukeleto> I assume your still out east? 13:52
dukeleto atrodo: yep, until tomorrow
13:53 GodFather joined 13:54 lucian left 13:55 particle joined
dukeleto Parrot::Embed needs some love 13:56
I looked at the Build.PL for it and shook my head 13:57
whiteknight I was looking at Parrot::Embed a while ago. It's so bad I was honestly considering learning XS so I could rewrite it from scratch 14:06
14:08 particle left 14:10 Andy joined
dukeleto whiteknight: good lord 14:14
whiteknight: don't learn XS, please
whiteknight: i will shield you from the horrors
whiteknight: i am already a lost cause (I know XS) so I will fix Parrot::Embed
whiteknight: the whole point of Parrot is so that people don't need to learn or maintain XS code 14:15
whiteknight: every time you think "I hate PIR", take a look at some XS code, that should make you feel better :) 14:16
whiteknight: mostly, the Parrot::Embed build system is crappy. I didn't even look at the XS yet
dalek sella: b89bcc3 | Whiteknight++ | src/ (5 files):
Update some error message routines. Update Action library to be more robust about error handling
 14:17
sella: 353da15 | Whiteknight++ | src/co (8 files):
some improved error handling in the container library
sella: cf7a285 | Whiteknight++ | src/ (2 files):
cleanups in Container. fix a bug in Action
sella: 8d9aa84 | Whiteknight++ | README.md:
update README with a section welcoming contributions and a section telling about the public API. If a function is marked as being not for external use, don't use it.
dukeleto whiteknight: is your plan to make PLA and friends use Rosella in their test suites ?
14:21 mj41_nb left 14:22 particle joined
whiteknight dukeleto: PLA already uses Rosella for the test suite 14:23
dukeleto whiteknight: i figured :) 14:24
whiteknight dukeleto: I recommend everybody use it :)
dukeleto whiteknight: i may want to use it in the PL/Parrot test suite
dukeleto takes PLA for a spin 14:25
whiteknight the PLA test suite may be a little broken right now, I was working on the upgrade
I don't remember where I left off 14:26
dukeleto whiteknight: github.com/Whiteknight/parrot-line.../issues/30 14:27
whiteknight: the rosella harness can't be found
whiteknight is that an issue you are seeing?
dukeleto whiteknight: yep 14:28
whiteknight and you have Rosella installed?
dukeleto whiteknight: nope :)
dukeleto installs Rosella
and it looks like I need winxed to compile Rosella... 14:29
dukeleto falls down a yak hole
whiteknight heh
when I put out the first release, I will include the .pir in the tarball
dukeleto whiteknight: works for me 14:31
whiteknight awesome
14:32 hercynium joined 14:33 clunker3 joined
dalek sella/gh-pages: 1dbcc31 | Whiteknight++ | index. (2 files):
index.html -> index.md
 14:33
sella/gh-pages: 8c02305 | Whiteknight++ | / (2 files):
add supporting files
sella/gh-pages: 1b968f4 | Whiteknight++ | / (3 files):
list libraries in the page header
 14:45
14:52 JimmyZ joined 15:00 theory joined 15:04 rurban_ joined 15:06 rurban left 15:07 rurban_ is now known as rurban
cotto_work ~~ 15:26
15:28 mj41_nb joined 15:31 ligne left
dukeleto cotto_work: top of the mornin' 15:35
NotFound ~~ 15:49
whiteknight Hello Notfound
dukeleto NotFound: i am trying to compile winxed 15:50
NotFound whiteknight: syntax for setprop? Never thought about that,
dukeleto NotFound: error:imcc:The opcode 'concat_s_s' (concat<2>) was not found. Check the type and number of the arguments
NotFound: does winxed need Parrot 3.0.0 ?
NotFound dukeleto: Don't try. Do it, young Jedi.
dukeleto NotFound: yes, I am doing but it is failing :)
whiteknight NotFound: I'm thinking about doing attribute-based programming stuff. So I want to be able to attach metadata to methods and classes at compile time 15:51
NotFound: it's not a big deal, but would be nice
dukeleto NotFound: i am trying to compile it against parrot master
NotFound dukeleto: What winxed version?
dukeleto NotFound: winxed master
NotFound Last time I see such error should be a year ago. 15:53
Coke should be a simple update to concat_s_s_s
dalek sella/gh-pages: c6181d5 | Whiteknight++ | stylesheets/rosella.css:
fix header a little bit
sella/gh-pages: bab66ef | Whiteknight++ | / (2 files):
more fixes
sella/gh-pages: 2e5f780 | Whiteknight++ | libraries/ (7 files):
start adding in documentation webpages for the various libraries. Only the core documentation is filled in at all
NotFound dukeleto: make from svn, or plumage? 15:54
15:55 mj41_nb left
NotFound whiteknight: you can use a pirop for a now, and then let me see use cases. 15:56
dukeleto NotFound: i just cloned the repo on github
dukeleto goes to lunch
NotFound dukeleto: uh, no, that repo was just a test of migration, is not updated.
Coke winxed? 15:57
NotFound I must delete it to avoid mor mistakes.
Coke aloha--
aloha Coke: Pbbbbtt!
Coke karma aloha?
aloha aloha? has karma of 0.
whiteknight aloha, winxed?
aloha whiteknight: winxed is code.google.com/p/winxed/ or winxed.org/
whiteknight karma aloha
aloha aloha has karma of -41.
whiteknight aloha++
aloha whiteknight: Thanks!
NotFound aloha: take a cookie
Coke builds winxed from svn, no problems.
NotFound++
NotFound Deleted 15:59
whiteknight NotFound: do you have any plans to migrate winxed to git in the future?
NotFound whiteknight: haven't decided yet. 16:00
whiteknight it's not a big deal
NotFound I must figure some politic about release numbers, tags, and such befor deciding anything. 16:03
And a t-shirt. These days any serious project must have t-shirts. 16:07
In the beginning was the web. Then came web 2.0. Now is web-shirt. 16:08
16:08 dmalcolm joined 16:41 ambs joined 17:01 alin_ joined 17:02 alin_ left, hudnix left 17:04 hudnix joined 17:06 alin left, ShaneC left, bluescreen left 17:14 JimmyZ left, ambs left 17:27 ShaneC joined 17:28 bluescreen joined 17:32 lucian joined 17:36 ShaneC left, davidfetter joined 17:40 ShaneC joined 17:51 lucian left 18:11 GodFather left
dalek sella/gh-pages: ea2f50b | Whiteknight++ | libraries/core.md:
fix formatting of lists
 18:11
sella/gh-pages: fb03be6 | Whiteknight++ | libraries/ (6 files):
flesh out several of the pages
18:11 davidfetter_ joined 18:16 davidfetter left, davidfetter_ is now known as davidfetter 18:22 plobsing left 18:35 [hercynium] joined 18:36 [hercynium] left 18:39 Kulag left, Kulag joined 18:41 hercynium left 18:46 Patterner left, Psyche^ joined, Psyche^ is now known as Patterner
dalek sella/gh-pages: c99a95c | Whiteknight++ | libraries/ (7 files):
Add a 'Users' section to each webpage to tell where the libraries are being put to use
 18:49
sella/gh-pages: 471b18a | Whiteknight++ | libraries/ (2 files):
Add in some examples for Action and Container, borrowed from the test suite
18:54 bluescreen left, giwi joined 19:04 bluescreen joined 19:11 preflex left 19:14 kid51 joined 19:15 dmalcolm left, preflex joined
kid51 ~~ 19:16
tadzik hello
kid51 is on vacation
kid51 met up with dukeleto in Ithaca last night; hope to meet Benabik tomorrow in Rochester 19:17
dukeleto kid51: wow, awesome! 19:18
19:18 benabik joined
dukeleto benabik: speak of the devil 19:18
Coke can't imagine going to Rocheter on vaca. ;)
benabik Ahh!
dukeleto: What I do? 19:19
dukeleto Coke: don't insult kid51++'s hometown, it isn't polite...
kid51 Coke: Well, when you were *born* here ...
dukeleto benabik: kid51++ was mentioning that y'all are going to have a mind meld
benabik I can't imagine going to ROC of vay-kay either, but that's probably because I'd rather *go* somewhere.
Coke kid51: seriously though, have a good time. I think my next vacation is probably going to be yapc time. 19:20
dukeleto Coke: you will be at YAPC::NA this year? 19:21
Coke I hope to. 19:22
cotto_work Coke: how likely does it look? 19:23
Coke The only stumbling block is the usual having to pay my own way.
. o O (and burn 3 days off)
actually getting the 3 days off is trivial. 19:24
kid51 Asheville is in foothills of Great Smoky Mtns and is quite pleasant; been there twice in recent years
Coke BBC headline: Google buys Parrot to aid YouTube
sadly it's not about us. ;)
cotto_work not yet 19:25
19:26 alin joined
dukeleto alin: welcome to #parrot 19:29
19:29 giwi left, cotto_work left, cotto_work joined 19:31 giwi joined
dukeleto builds winxed for the first time 19:31
yay, passing tests!
cotto_work What Parrot really needs is a MUMPS implementation. 19:32
whiteknight horrors 19:33
dukeleto whiteknight: ok, now i get a different error in the PLA test suite, after install winxed and Rosella 19:34
whiteknight: gist.github.com/874968
whiteknight dukeleto: okay, I figured there would be something. I've been refactoring the rosella test classes lately and haven't been keeping up with changes to PLA
I can get those fixed tonight, now that I know other people are looking at it
dukeleto whiteknight: no worries, I just want to finally be able to run the PLA test suite, so I can help hack on it :) 19:35
whiteknight: i could never get kakapo to work right. It is much closer to working now :)
whiteknight yeah, that was my big frustration too. Kakapo had some awesome goals but it was very complicated and interconnected 19:36
19:38 particle1 joined
dukeleto whiteknight: well, i hope to be able to help hack on PLA and friends 19:38
whiteknight dukeleto: I would love to have your hands! Like I said, I'll have all the tests fixed when I get home tonight
dukeleto whiteknight: awesome
19:41 particle left
bacek Good morning, humans 19:43
19:43 hercynium joined
kid51 Good morning, bacek. Read the blog posts! 19:43
whiteknight yes, bacek++ on the blog posts 19:44
bacek kid51, any feedback/suggestions?
kid51 bacek: One: Could you list the specific source code files from which you're drawing examples?
bacek kid51, I did. In part 2 :) 19:45
kid51 Ah, yes. I recall that that was my reaction to part 1
dalek rrot-linear-algebra: e84dd1f | Whiteknight++ | t/ (5 files):
some fixes so the harness runs again
 19:55
rrot-linear-algebra: 082a241 | Whiteknight++ | t/ (54 files):
more fixes so most tests actually run again on modern rosella. Test fails now because of a recent change in parrot where we cannot allocate zero bytes
whiteknight dukeleto: the test suite runs again, but the memory allocation functions changed in parrot which is breaking some tests. I'll have to fix them 19:56
dukeleto whiteknight: hokey dokey
dalek rrot-linear-algebra: 88aa9ad | Whiteknight++ | src/include/pla_matrix_types.h:
fix PLA to account for memory allocation change in Parrot. This allows tests to run again, although there are a few failures
 20:03
20:06 plobsing joined
dalek rrot/opsc_llvm: 1868a49 | nwellnhof++ | src/pmc/socket.pmc:
Return 0 in socket code for compatibility
 20:20
rrot/opsc_llvm: 4e9e576 | mikehh++ | src/gc/api.c:
add default to supress compiler warning and update documentation
rrot/opsc_llvm: 7374f63 | bacek++ | / (40 files):
Merge branch 'master' into opsc_llvm

Conflicts:
  \tconfig/gen/makefiles/root.in
20:20 giwi left 20:22 giwi joined 20:27 plobsing left
dalek rrot-linear-algebra: f58ccb9 | Whiteknight++ | t/ (3 files):
fix a broken test, and fix the harness to actually output the results again
 20:27
20:28 perlite_ joined 20:29 giwi left 20:32 perlite left, perlite_ is now known as perlite 20:33 plobsing joined 20:35 whiteknight left 20:38 kid51 left 20:44 plobsing left
tadzik hmm. Right now, after being configured and built, Parrot can't be moved to other place and still work. How can that be fixed, is that suitable for GSoC or something? 20:50
20:51 plobsing joined
dukeleto tadzik: are you installing parrot into your build directory ? 21:04
21:04 kid51 joined
dukeleto tadzik: i always install parrot into my build dir 21:04
tadzik: github.com/leto/Util/blob/master/bin/new_parrot 21:05
tadzik: then you can move it around, methinks
kid51 IIRC, we have fixed that problem. Has it reappeared?
tadzik when was it fixed? 21:06
kid51 In fact, 'make release_check' is supposed to guarantee that.
tadzik: What happens when you say: "perl Configure.pl --prefix=/some/where/else; make; make install" and then change the name of your builddir ? Does /some/where/else/bin/parrot --version work? 21:07
tadzik kid51: that's what I want to do. I didn't know that was fixed, if it is: great 21:09
21:09 plobsing left
tadzik owait 21:09
hmm, I think "im asking the wrong questions, nevermind. 21:10
what you say works since I remember
21:14 particle1 left, particle joined 21:31 hercynium left
kid51 tadzik: I just ran 'make relcheck' successfully, and I also did an install in a location other than my builddir, changed name of builddir, and successfully called the installed parrot. 21:32
So I think we're okay. 21:33
21:33 kid51 is now known as kid51_at_dinner, kid51_at_dinner left
cotto_work Does anyone know the purpose of our CLA? I feel like that's something I should understand better than I do. 21:33
davidfetter hopes it'll allow him to negotiate with a smaller group when he asks for parrot under a bsd-compatible license 21:34
NotFound cotto_work: it make us feel important.
21:34 benabik left 21:36 bluescreen left
cotto_work davidfetter: "when"? 21:40
21:45 jjore left 21:46 plobsing joined
dukeleto cotto_work: it is so that the intellectual property of parrot is owned by a legal entity and not individuals 21:50
cotto_work: so if somebody wants to sue for intellectual property/etc they sue Parrot Foundation and not individual parrot hackers
davidfetter: don't hold your breath for parrot under a bsd license 21:51
plobsing I'm skeptical about needing separate BSD licensing for inclusion in a BSD-licensed project. My legal-fu is weak, but section 4 of the AL2 seems pretty straightforward and easy to satisfy. 21:53
dukeleto plobsing: you legal-fu is very weak, grasshopper 21:55
tadzik oh, that reminded me of Usagi Yojimbo :)
dukeleto davidfetter: you seem to be the only person that wants a BSD license for anything, and I honestly haven't heard a good argument for it 21:56
davidfetter dukeleto, that's ok. you haven't made a good argument against it either :)
dukeleto the BSD license helps corporations and usually is not beneficial to the people that actually write the software
davidfetter: i don't have to. The burden is on you
dukeleto doesn't have a soft spot for corporations 21:57
davidfetter there's hocus-pocus and mumbo-jumbo about how somehow the GPL is more friendly to developers. i have not yet seen one scintilla of evidence that this is true
what i *have* seen is corporate Free-washing using the GPL
tadzik I don't see no GPL in Parrot
plobsing how is this even BSD vs. GPL? 21:58
dukeleto davidfetter: can you clearly state how releasing any parrot-related project or parrot itself under a BSD license is beneficial to the parrot community?
davidfetter dukeleto, if you're going to say the BSDL is unfriendly to developers, it's on *you* to provide a specific example. and to show that the example is on point
sure. you mentioned those corporates. they're delighted to contribute if they can hold off for awhile after they publish, or at least they think that's an option 21:59
dukeleto davidfetter: what koolaid are you drinking? 22:00
davidfetter: fosdem.org/2011/interview/david-fetter-2011 <-- this interview caused a lot of confusion and problems for me
davidfetter as a bay area denizen, i'm a little sensitive to that metaphor, but let's go on
dukeleto davidfetter: people were asking me if Parrot Foundation needed to hire lawyers and junk like that
davidfetter: don't speak for the parrot community if you have not consulted us first 22:01
davidfetter can tell he's offended dukeleto somehow, but is a little mystified as to how this happened 22:02
dukeleto davidfetter: i had to do damage control. No one, except you, wants PL/Parrot in PostgreSQL core. I do not want PL/Parrot or PL/Perl6 in PostgreSQL core.
davidfetter: i am not offended, but I needed to get some of that off my chest
davidfetter: no bad feelings, but I just need to speak my peice. 22:03
davidfetter that interview was with me. i was pretty clear on whom i was speaking for: me
re: the licensing thing, this is well-trodden territory on the pg side. the community ships only and strictly code that's under TPL or more permissive 22:04
dukeleto davidfetter: obviously it wasn't, because I had to assure parrot foundation people that we didn't need to hire lawyers
davidfetter and yes, we've checked with IP attorneys, etc.
dukeleto davidfetter: TPL ?
davidfetter well, somebody wrote that CLA. if there weren't lawyers involved, you're gonna have to hire some anyhow, when it comes to court
"the postgresql license," which is somewhat BSD-ish, somewhat MIT-ish. mostly came from editing without checking with an attorney 22:05
fortunately, no significant damage was done, except to create a wacky new license
davidfetter kinda shocked at how many hackers think they could do an IP attorney's job 22:06
bacek_at_work ~~ 22:08
davidfetter kinda curious as to who asked dukeleto about this, and more significantly, why those people didn't reach out to him directly 22:09
davidfetter even more curious as to why dukeleto would answer with anything other than, "davidfetter said it. take it up with him." 22:10
dukeleto davidfetter: i am on the board of PaFo and other board members asked me
davidfetter: people asked me my opinion of the situation and I told them 22:13
bacek_at_work: the llvm stuff you are hacking on is very exciting 22:15
bacek_at_work dukeleto, welcome aboard. I need help with finishing it. 22:16
davidfetter heh
dukeleto bacek_at_work: if you describe some low-hanging-fruit, i am happy to hack on it 22:19
bacek_at_work dukeleto, very-very low-level is finishing wrapping of BasicBlock. 22:21
dukeleto, little bit higher is to wrap BitReader/BitWriter 22:22
which imply wrapping of MemoryBuffer functions. 22:23
dukeleto bacek_at_work: which is easier? 22:24
bacek_at_work dukeleto, you can start with some simple things like wrapping LLVMNextBasicBlock/Prev into BasicBlock.pm
dukeleto bacek_at_work: don't do the easy stuff, let me do it :) You hack on the hard stuff
bacek_at_work it's straight forward.
dukeleto bacek_at_work: ok, should I work in your branch, or branch from your branch?
bacek_at_work dukeleto, just jump on my branch 22:25
next easy step is finishing Builder.pm. Which is wrapping rest of LLVMBuildFoo functions.
dukeleto hops onto opsc_llvm 22:26
bacek_at_work hmm... I've got idea for next blog post - "Implementation details of LLVM bindings" 22:27
dukeleto bacek_at_work: so I am looking at runtime/parrot/library/LLVM/BasicBlock.pm
bacek_at_work dukeleto, yes.
it's pretty much empty :)
dukeleto bacek_at_work: if you can make your next blog post a "howto" for other parrot hackers to get involved, that would be hella useful
bacek_at_work: i see that :)
bacek_at_work: which other file should I look at to get an idea for what I need to do? 22:28
bacek_at_work: i have a vague idea from your blog posts
bacek_at_work LLVM.pm
it holds all bound functions from LLVM
you need LLVMGetNextBasicBlock/GetPreviousBasicBlock/etc
look at Value.pm for idea 22:29
it's kind of 1 line of code for each of them :)
dukeleto bacek_at_work: ok. which llvm docs are you reading to figure out all the method names?
bacek_at_work Core.h :) 22:30
cotto_work bacek++
use the source, duke
bacek_at_work dukeleto, reading of Kaleidoscope Tutorial was really helpful to understand LLVM basics.
22:31 Kulag left
bacek_at_work dukeleto, it should be something like "method next () { LLVM::BasicBlock.create(LLVM::call("GetNextBasicBlock), self)) } 22:31
22:31 Kulag joined
bacek_at_work and Constant.pm can be filled with more stuff :) 22:32
dukeleto bacek_at_work: you have a link to the tutorial? 22:33
cotto_work: ;P
dukeleto looks at llvm.org/docs/doxygen/html/Core_8h_source.html 22:36
dukeleto gets lost
tewk llvm.org/docs/tutorial/ 22:37
dukeleto tewk++ 22:38
bacek_at_work: i am going to break the opsc_llvm branch, and when you fix it, I will learn :) 22:40
bacek_at_work: is this expected? gist.github.com/875280
bacek_at_work: which version of llvm are you working aginst? 22:41
bacek_at_work: i think i was on an old version of the branch 22:42
dalek rrot/opsc_llvm: ecbda7c | dukeleto++ | runtime/parrot/library/LLVM/BasicBlock.pm:
Add the next method to LLVM::BasicBlock
 22:44
dukeleto bacek_at_work: let me know if I broke stuff. The branch doesn't compile for me
dukeleto goes to nom some $food 22:46
davidfetter bon appetit! 22:47
22:53 fperrad left 23:03 rurban_ joined 23:06 rurban left, rurban_ is now known as rurban 23:14 alin left 23:15 Hunger left 23:24 Hunger joined 23:36 mtk left 23:39 cognominal left 23:42 mtk joined 23:46 jsut_ joined 23:51 jsut left
dalek rrot/opsc_llvm: 99ec214 | bacek++ | / (55 files):
Merge branch 'opsc_llvm' of github.com:parrot/parrot into opsc_llvm
 23:59
rrot/opsc_llvm: 16cb003 | bacek++ | runtime/parrot/library/LLVM/BasicBlock.pm:
Fix BasicBlock. dukeleto-- for committing non-compilable stuff.

Powered by Cro and the Raku® Programming Language.

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

Thank you!