Camelia, the Perl 6 bug

IRC log for #parrot, 2011-03-18

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:03 dalek parrot/opsc_llvm: f28b928 | bacek++ | t/library/llvm/07-basicblock.t:
00:03 dalek parrot/opsc_llvm: Add simple skeleton test for BasicBlock.
00:03 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/f28b928df6
00:04 whiteknight joined #parrot
00:14 dalek parrot/opsc_llvm: dc4aeef | bacek++ | runtime/parrot/library/LLVM/Opaque.pm:
00:14 dalek parrot/opsc_llvm: Add LLVM::Opaque.ACCEPTS implementation.
00:14 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/dc4aeefedd
00:14 dalek parrot/opsc_llvm: 3d942a0 | bacek++ | t/library/llvm/07-basicblock.t:
00:14 dalek parrot/opsc_llvm: Add test for BasicBlock.next
00:14 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/3d942a00a4
00:27 whiteknight I just realized that I didn't have /usr/local/bin added to my PATH, and I had a copy of Parrot 2.6.0 copied into /usr/bin
00:27 whiteknight so I've been running Parrot 2.6.0 on this machine for months
00:27 whiteknight fail
00:27 davidfetter left #parrot
00:30 plobsing upgrade tax time?
00:31 whiteknight yeah, I'm doing the rounds ow
00:31 whiteknight now
00:31 whiteknight build+install parrot, install plumage, install winxed, install rosella, and finally I can test PLA again
00:32 whiteknight plobsing: PLA was relying on that old behavior to allocate 0-sized blocks
00:32 whiteknight I've already fixed it, but it was being used
00:33 plobsing don't look at me. I just handled the situation properly. Andy decided the situation was silly (it is) and broke it.
00:33 cotto ~~
00:34 NotFound Winxed trunk works with Parrot 2.6? Amazing.
00:34 cotto whiteknight, amazing
00:34 whiteknight NotFound: apparently it's been working for months, and I'm too dense to notice
00:34 whiteknight what's really weird about it is that I was still installing libraries to the 3.2.0 directory, so I had some really funky simlinks working
00:36 NotFound Not bad, it shows that winxed is isolates well from parrot subtle changes.
00:37 whiteknight either that, or I'm much better at symlinking than I thought
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: 2560bf4 | plobsing++ | src/pmc/nativepccmethod.pmc:
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: add expected get_multisig() method to NativePCCMethod
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: review: https://github.com/parrot/parrot/commit/2560bf41a1
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: d8203d2 | plobsing++ | lib/Parrot/Pmc2c/PMC.pm:
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: create multi-dispatcher PCC signature based on return type, not NCI signature mangling
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: review: https://github.com/parrot/parrot/commit/d8203d2f8b
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: 4659654 | plobsing++ | lib/Parrot/Pmc2c/Parser.pm:
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: set method type correctly at creation
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: review: https://github.com/parrot/parrot/commit/46596549f9
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: db3ba55 | plobsing++ | lib/Parrot/Pmc2c/ (2 files):
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: eliminate "short sigs" (really NCI signatures) for multis
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: review: https://github.com/parrot/parrot/commit/db3ba55fd5
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: 71c5503 | plobsing++ | lib/Parrot/Pmc2c/ (3 files):
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: put method/multi name mangling in one place
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: review: https://github.com/parrot/parrot/commit/71c5503d87
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: a1d45c4 | plobsing++ | lib/Parrot/Pmc2c/ (2 files):
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: generate PCC variants for multis
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: review: https://github.com/parrot/parrot/commit/a1d45c4585
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: eccfd4e | plobsing++ | lib/Parrot/Pmc2c/PMC.pm:
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: use Parrot_mmd_add_multi_from_long_sig() to register NativePCCMethod multis
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2:
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: also minor fixups to pcc call and return sections
00:38 dalek parrot/pmc2c-multi-nativepccmethod-2: review: https://github.com/parrot/parrot/commit/eccfd4ec16
00:42 whiteknight where do I need to add GC write barriers in my PMC? Pmc2c automatically adds it to several vtables, right?
00:43 Andy plobsing: I assume you mean another Andy?
00:43 plobsing sorry, I can't keep you to separate in my head. I'm name-deficient.
00:43 plobsing s/to/two/
00:44 plobsing whiteknight: only vtables under a [.*] :write category in src/vtable.tbl
00:45 plobsing and in methods that aren't multi.
00:45 plobsing but multis are sometimes vtables.
00:46 whiteknight awesome pawsome
00:54 kid51 joined #parrot
00:54 ShaneC1 joined #parrot
00:55 bacek_at_work whiteknight, if you change PMC "guts" outside of ":write" VTABLEs.
00:56 whiteknight okay, I think I have it figured out
00:56 whiteknight thanks
01:00 ShaneC left #parrot
01:02 dalek parrot-linear-algebra: 475ba38 | Whiteknight++ | / (3 files):
01:02 dalek parrot-linear-algebra: fix conflicts
01:02 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/475ba3876b
01:02 dalek parrot-linear-algebra: cdad98a | Whiteknight++ | src/pmc/pmcmatrix2d.pmc:
01:02 dalek parrot-linear-algebra: add gc write barriers to pmcmatrix2d
01:02 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/cdad98a7de
01:02 dalek parrot-linear-algebra: 5c2e9ff | Whiteknight++ | t/ (3 files):
01:02 dalek parrot-linear-algebra: fix a few more tests. the order of operands in Assert::throws and Assert::throws_nothing was reversed
01:02 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/5c2e9fff0b
01:02 dalek parrot-linear-algebra: 7486f3c | Whiteknight++ | t/testlib/matrixtest.nqp:
01:02 dalek parrot-linear-algebra: fix another few tests
01:02 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/7486f3cc03
01:02 dalek parrot-linear-algebra: e309317 | Whiteknight++ | t/testlib/ (2 files):
01:02 dalek parrot-linear-algebra: fix most tests
01:02 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/e309317638
01:02 dalek parrot-linear-algebra: 6cb4159 | Whiteknight++ | t/pmc/complexmatrix2d.t:
01:02 dalek parrot-linear-algebra: fix remaining complexmatrix tests
01:02 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/6cb415948d
01:02 dalek parrot-linear-algebra: 0a5c755 | Whiteknight++ | t/harness:
01:02 dalek parrot-linear-algebra: add extra error reporting to the harness
01:02 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/0a5c755ded
01:02 whiteknight dukeleto: PLA is passing all tests for me after the most recent push. Give it a spin!
01:06 Andy Some time when I have copious free time, I need to merge back all my splint mess.
01:06 Andy Maybe I'll publish some pared-down splint error logs and people can pick at them.
01:06 kid51 Why would a Parrot test that segfaults when run by itself run perfectly when run thru gdb?
01:06 whiteknight kid51: I've been asking those kinds of questions myself for years
01:06 Andy kid51: Because running under gdb means that memory that used to be in place X is now in place Y.
01:08 bacek_at_work whiteknight, you did write barriers slightly wrong. Better to put it right after you change PMC. Otherwise GC can be triggered in the middle of method. Which can cause few bad things.
01:08 kid51 Andy: Okay, so what's the implication of that?
01:09 Andy You've got a pointer that is pointing to something bad.
01:09 whiteknight bacek_at_work: which method did I do wrong?
01:09 kid51 Andy:  And when I run under gdb, it's pointing to something good?
01:09 Andy or something that doesn't cause a segfault when you dump crap in it.
01:10 bacek_at_work pmcmatrix2d.pmc, line 87. You should move WB after line 84.
01:10 Andy Pointer problems are very susceptible to memory contents randomness.
01:10 bacek_at_work WB is cheap if it's already set
01:10 Andy Have you run the test under valgrind?
01:10 bacek_at_work afk # lunch
01:11 kid51 No, I doubt I could get valgrind installed on this old machine.
01:11 jjore joined #parrot
01:11 Andy what are you running?
01:11 kid51 (Not that I would know how to use it if I did.)
01:11 kid51 Mac OS X 10.4.11
01:11 Andy Oh, no, valgrind is OS X only.
01:11 kid51 darwin/ppc
01:11 Andy I mean Linux only
01:12 kid51 Yeah, this error in plobsing's branch is not showing up on linux
01:12 Andy lemme try it
01:12 Andy What's the branch?
01:13 kid51 t/pmc/timer.t
01:13 Andy what branch
01:13 kid51 specifically, t/pmc/timer_8.pir
01:14 kid51 pmc2c-multi-nativepccmethod
01:15 Andy ok, gimme a minute
01:15 dalek parrot-linear-algebra: 5fe6aae | Whiteknight++ | src/pmc/pmcmatrix2d.pmc:
01:15 dalek parrot-linear-algebra: fix positioning of the gc write barriers to prevent any weird corner cases. bacek++
01:15 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/5fe6aae205
01:16 plobsing kid51: could you please test pmc2c-multi-nativepccmethod-2 in stead? I suspect it won't have the same problem.
01:16 kid51 ok.  I wasn't aware there was a new branch.
01:16 plobsing I just fired out an email to the list and pushed it.
01:16 whiteknight msg dukeleto: PLA is passing all tests again on my machine. Have some fun with it and let me know if you need anything
01:16 aloha OK. I'll deliver the message.
01:18 whiteknight left #parrot
01:19 kid51 plobsing: k. but this will take a while to configure/build
01:21 plobsing ok. thank you for testing.
01:25 benabik joined #parrot
01:27 bacek left #parrot
01:30 Andy hahaha, wow, the harness is already set to recognize valgrind
01:30 AzureStone left #parrot
01:31 AzureStone joined #parrot
01:31 Andy plobsing: I get plenty of valgrind errors on t/pmc/timer.t
01:31 Andy Interested?
01:31 ShaneC1 left #parrot
01:32 plobsing Andy: sure. old or new branch?
01:32 Andy new
01:32 particle1 joined #parrot
01:32 dngor_ joined #parrot
01:32 eternaleye_ left #parrot
01:32 ShaneC joined #parrot
01:33 eternaleye_ joined #parrot
01:33 NotFound left #parrot
01:33 Themeruta joined #parrot
01:33 Kulag left #parrot
01:34 hudnix left #parrot
01:34 Kulag joined #parrot
01:34 dukeleto left #parrot
01:34 jjore left #parrot
01:34 jjore joined #parrot
01:34 Hunger left #parrot
01:34 hudnix joined #parrot
01:34 perlite_ joined #parrot
01:34 particle left #parrot
01:35 Hunger joined #parrot
01:35 dngor left #parrot
01:36 perlite left #parrot
01:36 perlite_ is now known as perlite
01:37 Andy plobsing: https://gist.github.com/875483
01:40 Andy That help you at all plobsing ?
01:40 plobsing Andy: thanks, but not really.
01:41 plobsing they're all in GC stack scanning
01:42 plobsing GC stack scanner reading from uninitialized values is expected, and it is written to handle that
01:46 kid51 plobsing: good news ; the new branch passes make pmc_tests
01:46 kid51 will try to run full smolder
01:47 kid51 Also, 'make test' PASS in that branch on linux/i386 (as I would have expected)
01:49 dukeleto joined #parrot
01:49 marcel_r joined #parrot
01:49 dukeleto ~~
01:50 dukeleto marcel_r: welcome to #parrot, what brings you to this fine corner of the interwebs?
01:50 plobsing my theory as to why the old branch failed is that we hit infinite recursion in the scheduler which called a vtable, which called a multi (pcc), PCC dispatch noticed there were active events and called the scheduler
01:56 kid51 plobsing:  I was reading this page on wikipedia yesterday; http://en.wikipedia.org/wiki/Thunk
01:57 kid51 ... and I was wondering which definition of 'thunk' is (or is closest to) the one we use in 'nci_thunk'
01:59 plobsing http://en.wikipedia.org/wiki/Thunk_(compatibility​_mapping)#Thunk_applied_generically_to_software_d​evelopment_to_describe_a_specific_type_of_adapter
01:59 kid51 perhaps a link to that in our docs would be good
02:03 plobsing if you can find an appropriate place to put that, I'm all for it.
02:03 kid51 pmc2c-multi-nativepccmethod-2: make fulltest PASS linux/i386
02:11 dukeleto Check this out: http://www.infoq.com/news/2011/03/apache-harmony
02:11 dukeleto people are considering switching from Java to Parrot because of Oracle and licensing issues
02:12 dukeleto kid51: we need to talk to those people
02:13 plobsing we need to be ready with an implementation of Java-the-language for these people
02:13 kid51 as if we don't already have enough to chew on ;-)
02:14 atrodo dukeleto: wait, what?  That seems awfully out of the plue
02:14 plobsing I proposed it as a GSoC project for exactly that reason.
02:14 * kid51 wondering what chromatic has been up to lately, reads his blog
02:16 dukeleto kid51: he has been coding Java for some reason
02:16 dukeleto atrodo: what seems out of the blue?
02:21 kid51 pmc2c-multi-nativepccmethod-2: make test PASS darwin/ppc: http://smolder.parrot.org/app/​projects/report_details/12522
02:22 atrodo dukeleto: harmony thinking about merging with parrot.  just seems like a strange suggestion to make
02:23 kid51 dukeleto:  Just read that post.   Yeah, yeah last summer my friend Steve Lembark was saying that we should explicitly pitch Parrot to Google as an alternative to a product owned by Oracle.
02:24 kid51 Harmony has PMCs too! ;-)
02:25 plobsing damn Javascript and Java people. both of them have a Harmony project. why can't they pick names that won't confuse half the world?
02:25 cotto I'm going to name my MUMPS on Parrot implementation Harmony.
02:26 * plobsing starts a project to "fully implement JS Harmony". title's it "Harmfully"
02:26 kid51 I had MUMPS when I was very young.  Then kids started getting vaccinated against it.
02:27 cotto there is no vaccination
02:28 * kid51 must sleep
02:28 kid51 left #parrot
02:28 cotto you may be thinking of the more pleasant Mumps
02:28 dukeleto atrodo: read the link
02:28 atrodo dukeleto: i'm working my way through the thread
02:29 dngor_ is now known as dngor
02:29 marcel_r_ joined #parrot
02:33 atrodo looks like my understanding of harmony is wrong.  it makes more sense now
02:34 marcel_r left #parrot
02:34 marcel_r_ is now known as marcel_r
02:37 benabik left #parrot
02:37 benabik joined #parrot
02:47 cotto I'm thinking about the representation of C functions that M0 exposes (I'm calling them function handles for lack of a better term).  I see a couple options, but I don't especially like either.
02:47 cotto The first is to set a function handle's size before setting its arguments, but I don't know if there's a sensible way to set the size that'll make sense for all the ways that M0 code will be used.
02:47 cotto The other option is to just set the signature elements and rely on the function handle to dtrt, but that sounds like the kind of magic we're trying to avoid in M0.
02:48 plobsing function handle? I thought M0 would just pass around function pointers.
02:49 cotto it needs to know about signatures somehow
02:51 plobsing signatures are a property of call sites
02:51 plobsing which is how you go thunkless
02:52 cotto you mean metadata?
02:52 sorear cotto:
02:52 sorear cotto: "size"?
02:53 plobsing it could be that way. the 'callc $P0, [ <signature description> ]' proposal is along those lines. $p0 is the function pointer
02:53 cotto sorear, how many arguments are in the signature
02:53 sorear not $I0?
02:53 plobsing the signature description is a constant
02:53 sorear requiring functions to be pmcish is technically problematic
02:54 sorear or is '
02:54 sorear or is 'P' an uninterpreted address in M0-land
02:55 plobsing sorear: I have no return parameter on callc. $P0 is "whatever is equivalent to a pointer"
02:56 sorear plobsing: planning to use "out" arguments instead?
02:57 plobsing sorear: I don't know. it could lie in the signature description, or go as an additional op parameter (we're only using 2/3)
03:06 cotto Let me organize my thinking a bit.  The basic problem is that we need to give the interp enough information to set up an ffi call.
03:06 cotto We can do that by having the information be inline (specified explicitly by ops), in metadata or in a table.
03:07 atrodo cotto> what kind of information again?  ffi parameters?
03:07 cotto atrodo, argument types
03:07 cotto Am I missing any options or making bad assumptions?
03:08 cotto atrodo, and the registers that they map to
03:10 cotto I like having that data be inline because it means no extra machinery is necessary to parse ffi-related code.
03:12 atrodo cotto> related question, is a call to ffi different than a call to an internal parrot function?  Or will the only time the interp leaves m0 land is to call ffi?
03:13 cotto atrodo, in the short term while Parrot is a mix of C and M0, they may be different.
03:14 atrodo long term they would be the same?  What about embedding and extending?
03:15 cotto good question
03:17 atrodo if long term we force extending to use ffi to interact with parrot, that seems like an extra layer that doesn't need to be there, especially since extenders don't necessarily want to be treated like foreign functions
03:18 ShaneC left #parrot
03:18 plobsing that's an extra layer you'll be crossing relatively rarely
03:19 plobsing if you are switching between Parrot and native frequently, you are doing things wrong
03:20 cotto extending doesn't concern me as much because I'd expect that to be mostly in an M0 overlay.
03:21 atrodo If I have parrot embedded in a game, and parrot has to interact with my objects to set states and call often, that seems like it'd happen often, and I don't see a better way to do it
03:21 atrodo *call my functions often
03:23 plobsing lua manages somehow
03:24 atrodo how so?
03:25 plobsing not sure. you might want to look into it if it is of concern to you.
03:27 plobsing although I wouldn't be surprised if the answer was something that Parrot isn't too keen on. I see Lua and Parrot as polar opposites when it comes to how they deal with C.
03:27 cotto We don't give it a big bear hug, that's for sure.
03:31 atrodo I just think we need to be worried not only allowing m0 access to outside libraries, but also make embedding and extending parrot nice to work with
03:32 atrodo I haven't checked out whiteknight's work yet, but as of a year ago, the embed/extend api seemed very hostile
03:34 ShaneC joined #parrot
03:47 marcel_r left #parrot
03:50 Andy left #parrot
03:50 sorear Lua puts the onus of marshalling mainly on the C side
03:51 sorear functions called from Lua need to use the Lua API to access parameters
03:51 sorear sort of like our op bodies
03:52 sorear only "sort of" like though because lua adds an extra isolation layer (which is how lua5 switched out a stack VM for a register VM without breaking client code)
03:53 plobsing my point is that Lua exposes a fat API to C. many facets of Lua are exposed, with C and Lua intermixing. Parrot tries to put up giant walls.
03:53 sorear no comment
04:38 benabik left #parrot
05:09 giwi joined #parrot
05:11 giwi left #parrot
05:33 giwi joined #parrot
06:18 theory left #parrot
06:18 alin joined #parrot
06:47 mtk left #parrot
06:56 contingencyplan left #parrot
07:04 rurban_ joined #parrot
07:06 rurban left #parrot
07:06 rurban_ is now known as rurban
07:48 jimmy joined #parrot
07:49 jimmy_ joined #parrot
07:55 jimmy left #parrot
08:00 jimmy_ left #parrot
08:03 alin left #parrot
08:14 jimmy joined #parrot
08:17 jimmy left #parrot
08:18 bacek joined #parrot
08:27 giwi left #parrot
08:34 fperrad joined #parrot
08:37 jimmy joined #parrot
08:43 jimmy left #parrot
08:47 alin joined #parrot
09:05 bacek ~~
09:12 clunker3 left #parrot
09:15 JimmyZ joined #parrot
09:16 JimmyZ left #parrot
09:16 clunker3 joined #parrot
10:43 alin left #parrot
11:00 lucian joined #parrot
11:19 dalek Rosella: aff81d7 | Whiteknight++ | src/winxed/Distutils.winxed:
11:19 dalek Rosella: +some docs for the distutils wrapper
11:19 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/aff81d77bf
11:19 dalek Rosella: 179fbf6 | Whiteknight++ | / (15 files):
11:19 dalek Rosella: Merge branch 'master' of github.com:Whiteknight/Rosella
11:19 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/179fbf62f1
11:31 alin joined #parrot
12:07 clunker3_ joined #parrot
12:10 clunker3 left #parrot
12:11 woosley joined #parrot
12:23 cognominal joined #parrot
12:26 dukeleto ~~
12:27 whiteknight joined #parrot
12:28 whiteknight left #parrot
12:28 whiteknight joined #parrot
12:30 whiteknight good morning, #parrot
12:30 whiteknight bacek++, again
12:34 dukeleto whiteknight: mornin'
12:34 * dukeleto tries the PLA test suite again
12:43 clunker3_ is now known as clunker3
12:44 dukeleto whiteknight: test suite is actually running!
12:45 dukeleto whiteknight: and it PASSed!
12:45 whiteknight awesome. happy hacking
12:46 dukeleto whiteknight: do you mostly use github issues for keeping track of stuff to do in PLA ?
12:47 whiteknight yes
12:49 dukeleto whiteknight: have any recommendations for particularly low hanging fruit to get my feet wet with?
12:49 whiteknight i've been using it as both a bug tracker and a TODO list
12:50 whiteknight I don't know. Most of the low-hanging fruit got picked during GCI
12:50 whiteknight damned students, stealing all our tasks!
12:50 dukeleto whiteknight: indeed.
12:52 Coke Andy: peeeng.
12:53 contingencyplan joined #parrot
12:56 * dukeleto can't leave Ithaca soon enough
12:57 whiteknight not happy on the east coast?
12:58 moritz Ithaca sounds Greek to me
12:59 cognominal left #parrot
12:59 whiteknight ba dum-ts
12:59 moritz ah, the greek island is written with s/c/k/
13:04 whiteknight dukeleto: We could probably use a benchmark or two for the PLA types, especially for heavy arithmetic workloads
13:05 whiteknight I've got a few optimizations in mind, but would need benchmarks first to make sure I'm actually optimizing
13:05 whiteknight Or, adding some new algorithms to the section of example code would be fun
13:06 whiteknight I've got one example there which does gaussian elimination already
13:07 whiteknight A routine to convert a matrix to upper- or lower-triangular would be a great starting point for most LAPACK bindings
13:09 plobsing_ joined #parrot
13:09 moritz is that what LU decompositions do?
13:09 whiteknight sort of. LU decompositions will decompose 1 matrix into a set of matrices with particular properies
13:10 whiteknight I think it's three, an upper-triangular, a lower-triangular, and a diagonal matrix
13:10 whiteknight no, just two matrices. no diagonal matrix
13:10 moritz ok, thanks
13:11 whiteknight LU factorization would make calculating determinants extremely easy, which is something PLA can't do yet
13:12 plobsing left #parrot
13:13 * moritz used SuperLU for his diploma thesis, but that was for solving equation systems
13:14 JimmyZ joined #parrot
13:16 JimmyZ left #parrot
13:16 JimmyZ joined #parrot
13:17 plobsing_ left #parrot
13:19 JimmyZ left #parrot
13:19 whiteknight does nqp-rx support heredocs?
13:20 JimmyZ joined #parrot
13:20 JimmyZ left #parrot
13:23 JimmyZ joined #parrot
13:29 JimmyZ left #parrot
13:32 cognominal joined #parrot
13:33 bluescreen joined #parrot
13:40 dalek parrot-linear-algebra: 63262b6 | Whiteknight++ | examples/elimination.nqp:
13:40 dalek parrot-linear-algebra: fix the gauss elimination example. Something has changed in parrot with respect to int/float algebra, and the results were coming out wrong. Now, force the scale factor to be a Float
13:40 dalek parrot-linear-algebra: review: https://github.com/Whiteknight/parr​ot-linear-algebra/commit/63262b6d01
13:47 plobsing joined #parrot
13:59 lucian left #parrot
14:07 Drossel joined #parrot
14:08 Kulag left #parrot
14:25 jsut joined #parrot
14:28 plobsing left #parrot
14:30 jsut_ left #parrot
14:38 mj41 left #parrot
14:45 ambs joined #parrot
14:55 dmalcolm joined #parrot
15:04 rurban_ joined #parrot
15:06 rurban left #parrot
15:06 rurban_ is now known as rurban
15:07 theory joined #parrot
15:10 whiteknight left #parrot
15:19 plobsing joined #parrot
15:23 hudnix left #parrot
15:32 hudnix joined #parrot
15:42 dukeleto msg whiteknight writing benchmarks for PLA sounds interesting, will try to throw some tuits at that
15:42 aloha OK. I'll deliver the message.
15:44 spinclad left #parrot
15:56 dalek Rosella: 3f9fc31 | Whiteknight++ | s (3 files):
15:56 dalek Rosella: winxed_setup now only sets up the stages for winxed, it doesn't call distutils.setup. Update bootstrap files
15:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/3f9fc31e4e
15:56 dalek Rosella: a6da2d6 | Whiteknight++ | s (3 files):
15:56 dalek Rosella: use some proper namespacing for the distutils routines. Update bootstrap
15:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/a6da2d6095
15:56 dalek Rosella: 4bec769 | Whiteknight++ | setup.winxed:
15:56 dalek Rosella: add a small note to the setup sequence when we are using the bootstrapped lib
15:56 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/4bec769a1a
15:56 whiteknight joined #parrot
15:58 preflex left #parrot
16:03 preflex joined #parrot
16:09 * dukeleto got a parrot talk accepted to YAPC::NA 2011
16:09 whiteknight nice. What talk?
16:09 dukeleto http://www.yapc2011.us/yn2011/talk/3303
16:09 dukeleto whiteknight: they are still accepting talks...
16:10 whiteknight ain't no way in hell I can make it to YAPC::NA this year
16:10 mtk joined #parrot
16:10 * dukeleto tried :)
16:10 dukeleto whiteknight: well, it would be nice to hack on stuff IRL sometime. You considering any other confs?
16:11 dukeleto whiteknight: maybe the GSoC summit in October (pending our acceptance as an org)
16:11 whiteknight I doubt I'll be doing much traveling this year
16:11 whiteknight unfortunately
16:13 dukeleto whiteknight: is all good
16:14 ambs_ joined #parrot
16:15 ambs left #parrot
16:15 ambs_ is now known as ambs
16:15 dalek Rosella/gh-pages: a71ec65 | Whiteknight++ | / (2 files):
16:15 dalek Rosella/gh-pages: add docs about the winxed library
16:15 dalek Rosella/gh-pages: review: https://github.com/Whiteknig​ht/Rosella/commit/a71ec651c0
16:15 dukeleto whiteknight: the output of examples/elimination.nqp looks wrong
16:16 dukeleto whiteknight: it is saying the reduced row echelon form is the zero matrix, which can't be right
16:16 dukeleto whiteknight: is it truncating floats to ints?
16:17 whiteknight dukeleto: I fixed that bug this morning. Parrot changed it's MMD dispatch rules at some point with respect to int/float
16:18 whiteknight do a pull, I think that should fix it
16:19 whiteknight in the elimination example, the expression "(1/$scale)" was returning 0 instead of a fraction because it was following integer division rules
16:19 hercynium joined #parrot
16:26 cotto_work ~~
16:28 davidfetter joined #parrot
16:30 plobsing left #parrot
16:38 woosley left #parrot
16:49 lucian joined #parrot
16:50 dalek Rosella/gh-pages: b538db7 | Whiteknight++ | / (2 files):
16:50 dalek Rosella/gh-pages: a few style fixes. Add the list of experimental libs too
16:50 dalek Rosella/gh-pages: review: https://github.com/Whiteknig​ht/Rosella/commit/b538db703e
17:10 dip left #parrot
17:10 plobsing joined #parrot
17:16 Andy_ joined #parrot
17:22 spinclad joined #parrot
17:23 lucian left #parrot
17:25 dalek Rosella/test_refactor: 7633acb | Whiteknight++ | src/test/ (5 files):
17:25 dalek Rosella/test_refactor: a few TODO notes ahead of what will become a major refactor
17:25 dalek Rosella/test_refactor: review: https://github.com/Whiteknig​ht/Rosella/commit/7633acb678
17:27 alin left #parrot
17:43 ShaneC left #parrot
18:04 lucian joined #parrot
18:12 ShaneC joined #parrot
18:30 alin joined #parrot
18:47 Patterner left #parrot
18:47 Psyche^ joined #parrot
18:47 Psyche^ is now known as Patterner
19:03 lucian dukeleto: hi
19:07 whiteknight PARROT HAS BEEN ACCEPTED TO GSOC2011!!!
19:07 whiteknight Application accepted
19:07 whiteknight dukeleto++
19:10 plobsing we can has slave labour now?
19:10 whiteknight much slave labour
19:10 lucian is now known as slave01
19:10 whiteknight I prefer to think of them as indentured servants
19:10 whiteknight beautiful, delicious indentured servants
19:11 slave01 is now known as indentured_servant0
19:11 indentured_servant0 whiteknight: does the master wish something?
19:11 plobsing nah. slave01 is awesome for the boba fett reference
19:11 whiteknight heh. No, all I wanted today was parrot's acceptance
19:11 indentured_servant0 is now known as slave01
19:12 marcel_r joined #parrot
19:12 slave01 plobsing: as the master wishes :D
19:20 Themeruta I like that master/slave model.
19:20 Themeruta is now known as NotFound
19:30 kid51 joined #parrot
19:30 slave01 is now known as lucian
19:35 piyushmishra joined #parrot
19:37 piyushmishra left #parrot
19:39 samwho joined #parrot
19:39 dmalcolm left #parrot
19:54 mj41 joined #parrot
19:55 cotto_work whiteknight: awesomeful
19:55 Topic for #parrot is now Parrot 3.2.0 released | http://parrot.org | Log: irclog.perlgeek.de/parrot/today | Parrot is accepted for GSoC2011! | Goals: Get more GSoC ideas on wiki; close tickets; stable 3.2 release; assess status of roadmap goals for 3/15 meeting.  
19:56 hercynium \o/
19:57 * hercynium really wants to work on a clojure-on-parrot implementation someday
19:58 Topic for #parrot is now Parrot 3.2.0 released | http://parrot.org | Log: irclog.perlgeek.de/parrot/today | Parrot is accepted for GSoC 2011! Student application deadline is Apr 8 | Goals: Get more GSoC ideas on wiki; close tickets; stable 3.2 release; assess status of roadmap goals for 3/15 meeting.  
19:59 cotto_work That might be a bit premature.  They start accepting applications on the 28th.
19:59 * cotto_work applies the pressure anyway
20:00 samwho Congrats on getting accepted for GSoC :)
20:01 samwho I'm a student looking to apply to GSoC this year but I'm struggling to find things I think I'd do well at...
20:01 samwho It's all a bit scary ^_^
20:01 cotto_work samwho: you're in a good place then.
20:01 cotto_work samwho: what gets you excited?
20:01 * sorear is wondering whether to apply, and if so to who
20:01 samwho cotto_work: Learning new things is awesome but I'm scared to apply to a project because I want to learn how it works.
20:02 samwho An open VM that runs a variety of languages sounds like an epicly awesome project but I know next to nothing about how it works.
20:02 cotto_work samwho: the best ways to learn are by trying to do something and by documenting
20:02 moritz samwho: just do something with parrot before applying, then you know if you find it approachable
20:03 bacek Good morning, humans
20:03 moritz sorear: just apply to perl foundation, I'll mentor any niecza tasks (note that mentoring is mostly about organization, not necessarily anwsering technical questions)
20:03 sorear hello baceks.
20:04 bacek aloha, sorear
20:05 sigue left #parrot
20:05 sigue joined #parrot
20:06 cotto_work samwho: is there anything in particular that you find interesting?  e.g. writing a new language with a specific focus, implementing a new gc, dealing with calling conventions, io, threading, etc
20:07 samwho cotto_work: All of it sounds interesting. I love learning how things are implemented at a low level, if that helps? I really don't know how most of what you just mentioned works, though. Which is what's making me a bit nervous.
20:08 atrodo samwho> That's okay.  Doing something in one area doesn't generally mean you have to learn all the other areas first
20:08 cotto_work samwho: fortunately you don't have to know how all of Parrot works to be able to hack on some of it.
20:09 samwho I'm relatively new to working on open source software. The only other one that I have worked on is ThinkUp :)
20:10 samwho I'm having a read of the "Parrot Babysteps" page now.
20:10 samwho Very reassuring introduction :p
20:12 marcel_r left #parrot
20:15 simcop2387_ joined #parrot
20:16 whiteknight samwho: Open source software is just like any other software. nothing scarey about it
20:16 samwho whiteknight: I know. It's just all a bit overwhelming for a new comer :)
20:17 simcop2387 left #parrot
20:17 simcop2387_ is now known as simcop2387
20:18 NotFound whiteknight: except maybe killer zombies games.
20:19 samwho Would Parrot be welcoming to students that are new to the project?
20:19 atrodo NotFound> As opposed to the friendly zombie games?  Those are my favorites
20:19 whiteknight samwho: yes, if you start getting involved now
20:19 whiteknight samwho: most of our students are new to the project each summer.
20:19 samwho whiteknight: unpacking the tar for the latest stable release as we speak :p
20:20 lucian left #parrot
20:21 lucian joined #parrot
20:21 samwho What about if the student is new to the entire idea of virtual machines? (I know what they are but not how they work :) )
20:21 NotFound atrodo: my favourite genre is pacifist non-violent kill-them-all games.
20:21 whiteknight samwho: that's fine. Parrot is a big project. Only a small part of it is actually the VM
20:22 whiteknight we have libraries, and language compilers, and then libraries written in those languages
20:22 atrodo NotFound++ I need to find those games, sounds awesome
20:22 samwho I don't really know how compilers work, either... Another thing I'd be really interested in learning :)
20:23 NotFound The awesomeness is great, but the playability suffers.
20:23 dalek Rosella: 7633acb | Whiteknight++ | src/test/ (5 files):
20:23 dalek Rosella: a few TODO notes ahead of what will become a major refactor
20:23 dalek Rosella: review: https://github.com/Whiteknig​ht/Rosella/commit/7633acb678
20:23 dalek Rosella/test_refactor: e053fd3 | Whiteknight++ | src/test/ (11 files):
20:23 dalek Rosella/test_refactor: First stage of the big refactor. Loader and Testcase are deleted. In their place we have SuiteFactory, TestFactory, TestContext, and TestCase
20:23 dalek Rosella/test_refactor: review: https://github.com/Whiteknig​ht/Rosella/commit/e053fd308d
20:23 dalek Rosella/test_refactor: 16e85fc | Whiteknight++ | s (2 files):
20:23 dalek Rosella/test_refactor: small fix to get the library to build
20:23 dalek Rosella/test_refactor: review: https://github.com/Whiteknig​ht/Rosella/commit/16e85fc793
20:23 dalek Rosella/test_refactor: 2087d8c | Whiteknight++ | src/test/ (4 files):
20:23 dalek Rosella/test_refactor: fixes so that we are running a simply dummy test again
20:23 dalek Rosella/test_refactor: review: https://github.com/Whiteknig​ht/Rosella/commit/2087d8cb6d
20:23 dalek Rosella/test_refactor: fdbc05b | Whiteknight++ | / (42 files):
20:23 dalek Rosella/test_refactor: some fixes, then I ran into a VERY strange problem involving proxies and I've littered a bunch of debugging statements around to try and find it. I'm just committing my work now because I can't finish debugging now
20:23 dalek Rosella/test_refactor: review: https://github.com/Whiteknig​ht/Rosella/commit/fdbc05bfac
20:24 whiteknight samwho: First step is figuring out what kinds of things you like, then we can help match you up with a great project
20:24 whiteknight so...what do you like?
20:24 samwho whiteknight: Will do :) I'll follow the babysteps guide and then work through from there. Thanks for being patient with me, I really appreciate it ^_^
20:25 whiteknight samwho: no problem! We're happy to have you. We chat in here all the time, so stop in if you have questions or anything
20:25 samwho whiteknight: I really love programming and learning about new languages and things. I think I'd enjoy a project based around learning how languages are designed and how they work.
20:25 cotto_work We think about that a lot.
20:26 samwho I can imagine :D
20:26 sorear The language design dept. is more in #perl6
20:26 whiteknight samwho: what's your favorite programming language?
20:26 sorear #parrot-land focuses on the "how they work" more
20:26 samwho whiteknight: I try and avoid having a favourite... Makes it easy to get in to language wars :D
20:26 whiteknight okay, what languages have you used?
20:27 * whiteknight has to pack up and go. will be back on later
20:27 whiteknight left #parrot
20:28 samwho I had a big reply written for him them ^_^ Bad times.
20:28 cotto_work samwho: post anyway
20:28 cotto_work we have a logging bot (see topic)
20:30 samwho In descending order of my ability, I have used: PHP, Java, C#, ActionScript, Python, C-C++, LISP
20:30 samwho When you get to Python my ability is very basic, small programs.
20:31 sorear The only language from that list used in Parrot proper is C
20:31 samwho I'd love to get better at C.
20:31 Tene aloha: tell whiteknight that ideally we'll get into language design more here, someday.
20:32 cotto_work samwho: that's good.
20:33 sorear #perl6-land also uses C#, Java, and LISP
20:33 sorear we have an outstanding idea to write a PHP compiler for Parrot
20:34 bacek Tene, aloha doesn't understand ':' as nick separator. Use ','.
20:34 Tene bacek: Ah, that explains why aloha always ignores me. :)
20:34 bacek Tene, "aloha, msg someone"
20:34 Tene bacek: Not important enough; maybe he'll see it in the log
20:34 Tene Thanks, though
20:39 fperrad left #parrot
21:02 dalek parrot/opsc_llvm: eeb52d5 | bacek++ | t/library/llvm/06-builder.t:
21:02 dalek parrot/opsc_llvm: Fix test.
21:02 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/eeb52d58ca
21:04 mtk left #parrot
21:05 kid51 left #parrot
21:12 hercynium left #parrot
21:12 alin left #parrot
21:18 marcel_r joined #parrot
21:38 bluescreen left #parrot
21:41 bluescreen joined #parrot
21:42 lucian dukeleto: ping
21:51 bluescreen left #parrot
21:56 mj41 left #parrot
22:15 alin joined #parrot
22:17 marcel_r_ joined #parrot
22:23 marcel_r left #parrot
22:23 marcel_r_ is now known as marcel_r
22:24 noitisnt joined #parrot
22:29 ambs left #parrot
22:33 particle1 congrats! parrot foundation made it into gsoc2011
22:33 particle1 is now known as particle
22:33 particle org admins need to update the org profile. or make me an org admin and i'll do it now
22:34 alin left #parrot
22:35 particle oh, i didn't read to the end of carol's mail... looks like the org admins will be added over time for new-to-gsoc-orgs like pafo
22:36 sorear pafo is new?
22:40 cotto_work looks like PaFo needs to complete its org profile before mentors can be accepted
22:40 plobsing left #parrot
22:40 cotto_work sorear: for the last 3 years we applied under The Perl Foundation.  Now we're our own org.
22:43 noitisnt left #parrot
22:53 soh_cah_toa joined #parrot
22:56 soh_cah_toa left #parrot
23:06 rurban_ joined #parrot
23:08 rurban left #parrot
23:08 rurban_ is now known as rurban
23:32 benabik joined #parrot
23:33 whiteknight joined #parrot
23:34 cotto_work hio whiteknight
23:34 cotto_work happy weekend
23:34 whiteknight good evening, cotto_work
23:34 whiteknight and an especially happy weekend it is! GSOC!
23:36 sorear hello
23:42 cotto_work dukeleto: ping
23:42 benabik left #parrot
23:44 plobsing joined #parrot
23:51 soh_cah_toa joined #parrot
23:54 soh_cah_toa hey guys, just took a look at the gsoc 2011 homepage and it said that the parrot foundation was accepted but hadn't completed its organization profile. what exactly does that mean?
23:55 davidfetter left #parrot
23:56 soh_cah_toa ?
23:57 soh_cah_toa anybody?
23:59 dukeleto we got into GSoC
23:59 * dukeleto has to get on a plane

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

Parrot | source cross referenced