Camelia, the Perl 6 bug

IRC log for #parrot, 2010-06-04

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 jnthn whiteknight: So may have some bad assumptions in the Rakudo C bits as a result.
00:00 cotto_work I'm pretty sure we trace it.
00:00 bacek_at_work we _can_ trace it
00:02 bacek_at_work But we _don't_ do it.
00:02 bacek_at_work Check "Parrot_gc_trace_root" function
00:03 bacek_at_work when GC call it without specifying GC_TRACE_FULL we don't trace stack.
00:03 bacek_at_work But we do sweep of pool after
00:04 nopaste "bacek" at 192.168.1.3 pasted "Prove that we don't trace stack" (18 lines) at http://nopaste.snit.ch/20897
00:05 bacek_at_work flags == 0.
00:05 bacek_at_work gc_trace_root doesn't trace stack in this case
00:05 dalek parrot: r47347 | tcurtis++ | branches/gsoc_past_optimization (2 files):
00:05 dalek parrot: All PAST::Pattern and subclass attributes test for exact matching correctly(except for "symtable"); now for children.
00:05 bacek_at_work Bang
00:05 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47347/
00:07 bacek_at_work I would _love_ to be wrong. But I'm not (at least in this case).
00:07 bacek_at_work And current GC works by accident...
00:23 bacek_at_work .oO( Did I scare everyone? )
00:25 sorear no
00:25 whiteknight bacek_at_work: well, damnit
00:25 whiteknight bacek_at_work: we used to trace it. I've seen it done
00:25 sorear I've long known that our GC runs on steady supply of miracles
00:26 whiteknight miracles? Don't you mean nightmares of young children?
00:26 bacek_at_work whiteknight, probably. But we don't now.
00:26 sorear whiteknight: elder gods are still gods, and what they deliver are still miracles...
00:27 bacek_at_work It was one of "gotchas" in gc_massacre branch when I did some performance testing.
00:27 sorear anyways, this is obviously why allocating objects in imcc doesn't work
00:29 bacek_at_work And all this "if PMC_data(SELF)" inside .mark VTABLE are just wallpapering main problem.
00:34 sorear I've always hated those lines
00:36 cotto So why aren't we tracing the stack?  Performance?
00:36 bacek_at_work "stupidity"?
00:39 cotto then flip that bit and let's get some sanity into this bird
00:39 cotto or does that make things even more broken?
00:40 bacek_at_work cotto, I dunno. But I want to kill current GC anyway :)
00:52 abqar joined #parrot
00:57 snarkyboojum joined #parrot
00:57 snarkyboojum_ joined #parrot
00:59 Coke dukeleto: ping.
01:01 Coke or, anyone with git-fu: ping.
01:02 sorear pong.
01:04 Coke have a git branch 'revive'. wish to replace master with it.
01:05 Coke I did 'git branch -M revive master'.
01:06 Coke I can't push now.
01:06 Coke "fix me". =-)
01:06 Coke ah. found:
01:07 Coke http://stackoverflow.com/questions/2​862590/how-to-replace-master-branch-​in-git-entirely-from-another-branch
01:08 Coke can't do a git push origin :master (deletion of current branch prohibited)
01:09 Coke seen dukeleto?
01:09 purl dukeleto was last seen on #parrot 2 hours, 27 minutes and 52 seconds ago, saying: now i get this: http://gist.github.com/424624 when loading perl6.pbc from C
01:11 bacek_at_work Coke,git checkout revive; git push origin :master; git push origin master
01:11 cotto pmichaud, ping
01:14 sorear Coke: git push -f origin master
01:17 Coke bacek_at_work: http://stackoverflow.com/questions/2​862590/how-to-replace-master-branch-​in-git-entirely-from-another-branch
01:17 Coke er...
01:17 Coke ! [remote rejected] master (deletion of the current branch prohibited)
01:18 Coke I get that on git push origin :master
01:19 bacek_at_work Coke, try: git push origin revive; git push origin :master; git push origin master;
01:19 sorear this is why you use git push -f origin master instead
01:19 sorear no colon, but :
01:19 sorear no colon, but -f
01:20 Coke yup, the -f worked.
01:20 Coke ... does that look right on http://github.com/partcl/partcl ?
01:22 Coke looks like revive still exists, pushing that...
01:23 dalek partcl: bb6b58e | plobsing++ | runtime/tcllib.pir:
01:23 dalek partcl: avoid using uninitialized HLL-mapped strings in startup
01:23 dalek partcl: Signed-off-by: Coke <will@coleda.com>
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/b​b6b58ed12a5684d345f23005c7e9e22471776c8
01:23 dalek partcl: ea247be | Coke++ | config/PARROT_VERSION:
01:23 dalek partcl: Track the parrot version we've gotten back to on this branch.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/e​a247be03f33185d2e87030af071606f7238832c
01:23 dalek partcl: 2579fb0 | Coke++ | t/cmd_namespace.t:
01:23 dalek partcl: Untodo some passing tests!
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/2​579fb0d636cdaf7dbdd49a0bab755db5880e7fe
01:23 dalek partcl: bdc5273 | Coke++ | runtime/builtin/cd.pir:
01:23 dalek partcl: [cd] should return the empty string;
01:23 dalek partcl: previous versions seem to have relied on a bug in PCC processing which has
01:23 dalek partcl: been fixed.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/b​dc52732e4532ad3dc5260feb25caa929b525030
01:23 dalek partcl: 4b56723 | Coke++ | src/grammar/expr/operators.pir:
01:23 dalek partcl: as ** stopped coercing to int properly, force it to be an int for now.
01:23 dalek partcl: (This will likely have to change as we run more tests.)
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/4​b56723d2122fefb6f8a87b5dec43416871f1142
01:23 dalek partcl: b2d258b | Coke++ | config/PARROT_VERSION:
01:23 dalek partcl: track progress.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/b​2d258ba2ed3fc7847ca01ee08f36f1fe6426945
01:23 dalek partcl: 7d30d88 | Coke++ |  (15 files):
01:23 dalek partcl: Get past the TT #389 "fix"
01:23 dalek partcl: (and workaround the resulting TT #1672)
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/7​d30d883763247967c3b78e17c9ee978632df96b
01:23 dalek partcl: 9c289e0 | Coke++ |  (3 files):
01:23 dalek partcl: Track immutable strings mergeback in parrot.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/9​c289e0064e2e60f26327a1b5f3aa619b698a2a7
01:23 dalek partcl: 8e5cc43 | Coke++ | config/makefiles/root.in:
01:23 dalek partcl: Remove unused config var.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/8​e5cc4334f05fc24e8e254a37d8674d1d598165e
01:23 dalek partcl: 3d716f9 | Coke++ | runtime/builtin/lsort.pir:
01:23 dalek partcl: These opcodes were removed from parrot.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/3​d716f96a37532255a622827cf3545cbebc6caf6
01:23 dalek partcl: 1b67909 | Coke++ |  (14 files):
01:23 dalek partcl: track string opcode changes
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/1​b679094821d2e8851696e05c1a35d07c76358de
01:23 dalek partcl: 937f31f | Coke++ |  (2 files):
01:23 dalek partcl: track dynop migration
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/9​37f31f7c409e87767a419781ddb2826a009e5bd
01:23 dalek partcl: 41b6b13 | Coke++ | src/pmc/tcl (3 files):
01:23 dalek partcl: *attrs now required.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/4​1b6b13c50dad90f2e21853c2a9cedf651904845
01:23 dalek partcl: 70fbc44 | Coke++ | config/PARROT_VERSION:
01:23 dalek partcl: works on a recent parrot.
01:23 dalek partcl: review: http://github.com/partcl/partcl/commit/7​0fbc4491a0f49a92a3e0edb920df4fb528887fe
01:26 Coke sorear++
01:26 mikehh Coke: I get conflicts if I try git pull
01:26 Coke bacek++
01:26 Coke dukeleto++
01:26 Coke mikehh: yup. I just killed master and snuck something in its place.
01:26 Coke I think you might need to do a fresh clone.
01:28 tetragon joined #parrot
01:30 cotto pmichaud, unping
01:32 Coke purl, msg whiteknight partcl is once again functional.
01:32 purl Message for whiteknight stored.
01:41 sorear mikehh: you'll need to use git pull --rebase
01:41 Coke sorear++
01:45 dalek partcl: d931d6e | plobsing++ | config/makefiles/root.in:
01:45 dalek partcl: use newer ops2c
01:45 dalek partcl: Signed-off-by: Coke <will@coleda.com>
01:45 dalek partcl: review: http://github.com/partcl/partcl/commit/d​931d6eedba57534b4ce0e588051dde47c5c76a1
01:47 Coke msg whiteknight forks & pull requests for partcl are cool, btw.
01:47 purl Message for whiteknight stored.
01:50 plobsing joined #parrot
01:54 plobsing hi #parrot
02:08 Coke hio!
02:10 gbacon joined #parrot
02:13 patspam joined #parrot
02:16 plobsing Coke: I discovered a happy accident that might get TT #906 unstuck.
02:38 Coke http://www.boingboing.net/2010/​05/30/maggie-thatcher-perf.html for fun.
02:40 Coke plobsing: what's that?
02:42 plobsing I didn't cover the corner case for PIR syntax sugar using assignments
02:43 plobsing I figure all the ops we want to use assign with that don't fit the assignment mold have sugared forms.
02:43 plobsing which means you can do "$P0 = $P1 * $P2" just fine.
02:49 tcurtis I'd like to clarify that my commit message that will be showing up in a moment is not entirely true. Only exact matching of children works so far. Committed before I realized I omitted the qualifier.
02:50 dalek parrot: r47348 | tcurtis++ | branches/gsoc_past_optimization (2 files):
02:50 dalek parrot: Children matching works!
02:50 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47348/
02:51 tcurtis tcurtis-- for that.
03:00 janus joined #parrot
03:24 snarkyboojum joined #parrot
03:26 dalek rakudo: e92a5f4 | (Solomon Foster)++ | src/core/metaops.pm:
03:26 dalek rakudo: Allow embedded hashes in array hypers.
03:26 dalek rakudo: Needs work (particularly better handling of errors) but works as a first
03:26 dalek rakudo: approximation.
03:26 purl approximation is the trapdoor.
03:26 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​92a5f4a8ff0b0fdc78574fc7b7a0ea9b91b6875
03:54 snarkyboojum_ joined #parrot
04:20 snarkyboojum joined #parrot
04:41 snarkyboojum joined #parrot
04:46 JimmyZ joined #parrot
04:48 jrtayloriv joined #parrot
05:23 TiMBuS joined #parrot
06:02 cotto seen pmichaud
06:02 purl pmichaud was last seen on #parrot 6 hours, 23 minutes and 37 seconds ago, saying: gotta run -- bbl
06:02 cotto pmichaud, ping
06:06 uniejo joined #parrot
06:08 cotto seen khairul
06:08 purl khairul was last seen on #parrot 2 days, 10 hours, 15 minutes and 44 seconds ago, saying: see t/pmc/scheduler.t  [Jun  1 19:52:46 2010]
06:09 cotto msg khairul It turns out I'll be free Friday if you want to chat.
06:09 purl Message for khairul stored.
06:15 dukeleto 'ello
06:15 LoganLK joined #parrot
06:17 cotto hi
06:18 dukeleto cotto: how is gsoc going for you?
06:18 cotto great
06:19 cotto khairul is doing a good job and has been responsive to questions so far
06:20 aukjan joined #parrot
06:21 cotto I do need to ask why I haven't seen much out of him in the last couple days, but thinking would be a valid reason.
06:27 dukeleto cotto: i don't know about him, but many students have finals coming up
06:29 dukeleto cotto: do you both have a regular meeting time ?
06:29 cotto istr him being done with school, but I like remembering things incorrectly
06:29 cotto yup
06:29 cotto monday evenings (my time)
06:29 cotto he prefers privmsgs
06:32 fperrad joined #parrot
06:41 fperrad_ joined #parrot
07:00 mikehh Coke: when I try a git clone on partcl I get the following - fatal: http://github.com/partcl/partcl/info/refs not found: did you run git update-server-info on the server?
07:01 moritz mikehh: try the git:// URL, that's usually more reliable
07:20 dalek rakudo: 8b116ff | moritz++ | build/PARROT_REVISION:
07:20 dalek rakudo: bump PARROT_REVISION to get latest nqp-rx fixes for ** quantifier and :sigspace
07:20 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/8​b116ffe60fe44d709042c1db725fcc3c258aa9a
07:37 mikehh moritz: yes - I think I seriously need some sleep
07:43 dalek rakudo: a1695f0 | (Timothy Totten)++ | src/core/ (2 files):
07:43 dalek rakudo: Implemented Date.new(::DateTime) and DateTime.Date()
07:43 dalek rakudo: methods.
07:43 purl methods are slower than constants.
07:43 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
07:43 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/a​1695f01a6584a63423880565ac18abbcbe8a19d
07:57 mikehh joined #parrot
08:16 snarkyboojum joined #parrot
08:51 snarkyboojum joined #parrot
09:13 dalek parrot: r47349 | gerd++ | trunk/tools/util/inc_ver.pir:
09:13 dalek parrot: rewritten with I/O in method style
09:13 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47349/
09:14 clinton joined #parrot
09:18 dalek nqp-rx: 42209fb | pmichaud++ | src/ (3 files):
09:18 dalek nqp-rx: ":my" in regexes is now properly declarative for LTM purposes.
09:18 dalek nqp-rx: Fix a bug with prefix computation on uncaptured subrules.
09:18 dalek nqp-rx: review: http://github.com/perl6/nqp-rx/commit/4​2209fba432a01f1918a14e35d5a2a52cb2878ed
09:40 snarkyboojum joined #parrot
09:47 Hunger joined #parrot
09:51 dalek rakudo: 738fad4 | (Solomon Foster)++ | src/core/metaops.pm:
09:51 dalek rakudo: Extend prefix and postfix array hypers to handle nested hashes.
09:51 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​38fad4ca24951c879cd0542a39fb370b9885547
10:34 jsut joined #parrot
11:04 abqar_ joined #parrot
11:19 tetragon joined #parrot
11:57 dalek parrot: r47350 | fperrad++ | trunk/tools/install/smoke.pl:
11:57 dalek parrot: [install] fix some tests
11:57 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47350/
12:15 bluescreen joined #parrot
12:31 bluescreen joined #parrot
12:41 khairul joined #parrot
12:46 dalek parrot: r47351 | gerd++ | trunk/tools/util/inc_ver.pir:
12:46 dalek parrot: remove whitespace
12:46 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47351/
12:46 dalek parrot: r47352 | khairul++ | branches/gsoc_instrument (3 files):
12:46 dalek parrot: Reworked the data structures for the hooks, use a linked list now instead.
12:46 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47352/
12:47 khairul msg cotto great, i'll come find you then.
12:47 purl Message for cotto stored.
13:10 snarkyboojum joined #parrot
13:11 atrodo joined #parrot
13:17 Coke perldoc checking is broken on ubuntu's moblin. (perldoc is installed, visible in /usr/bin/perldoc, Configure.pl doesn't notice)
13:40 whiteknight joined #parrot
13:42 whiteknight good morning, #parrot
13:55 davidfetter joined #parrot
14:01 JimmyZ joined #parrot
14:06 spinclad joined #parrot
14:12 ash_ joined #parrot
14:44 ash_ does parrot have policies for 3rd party stuff in the core?
14:46 estrabd joined #parrot
14:47 gbacon joined #parrot
14:52 PerlJam ash_: I think the policy is "no"   :)  But that's only a first approximation.
14:54 ash_ just wondering, it just might be convenient to bundle libffi in the source, kinda like python does, or at least i can make a script that will download it if its not installed
14:54 davidfetter hrm. does that include ICU?
14:55 jnthn Parrot used to have ICU in core in the good old days, and we always built it, so you always had unicode support.
14:55 jnthn s/in core/in the repo/
14:55 ash_ i could do a Configure.pl --install-libffi maybe...
14:56 PerlJam ash_: I'd ask allison for something more definitive.  don't take my word for it.
14:57 bubaflub joined #parrot
15:01 Coke second approximation is also no.
15:01 Coke ls
15:01 Coke ww
15:04 NotFound ash_: is not our task to install other things in user's systems.
15:05 ash_ true... i just wanted to make it easier
15:06 NotFound ash_: that's not easy at all. Is a source of problems.
15:15 dalek parrot: r47353 | NotFound++ | trunk/t/pmc/fixedintegerarray.t:
15:15 dalek parrot: tests for FIA get_repr and is_equal to different type
15:15 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47353/
15:25 whiteknight ash_: What license does libffi use?
15:26 ash_ an MIT liscense
15:26 ash_ but spelled right
15:26 whiteknight ok
15:28 whiteknight better bet is to get a configure probe for the library, and define a macro or something to say whether the library is available
15:28 whiteknight then, conditionally compile your code using that flag
15:29 ash_ i already have the config stuff working
15:29 whiteknight okay, that's most of the battle
15:30 ash_ okay, i need more eyes on this, i am trying to test the parser i wrote to parse signatures, and I cannot for the life of me see where this one bit with the structures is going wrong. https://gist.github.com/5361a2d1bc4e0868eeb4 is a copy of my code currently
15:31 mberends ash_: on debian-ish systems you can install libffi-dev
15:31 ash_ yeah, its already installed on OS X, but windows is the only one that might be tricky i think
15:33 patspam joined #parrot
15:33 ash_ at line 423 i call my function with a struct using the normal way to define a libffi struct, then like 429 i call the function with my dynamically generated one, up to that point everything in the debugger seems correct
15:34 whiteknight weird
15:34 whiteknight so what happens?
15:34 ash_ and the normal libffi definition of the signature works, my signature i made with the parser doesn't
15:34 ash_ i get r3 = 8 for the libffi signature, my parsers signature gives 0;
15:35 ash_ if you put a break point in the add_struct a.x and a.y are both 0
15:35 ash_ :-\
15:35 ash_ normal signatures are working fine though so far
15:35 ash_ and by normal i mean ones without structs
15:43 ash_ I am building it with clang -I/opt/local/lib/libffi-3.0.9/include  -L/opt/local/lib -lffi parse_sig.c -o parse_sig -g if you want to test it out, you'll need to correct the paths for your system of course
15:44 ash_ pkg-config libffi --cflags --libs might work if its installed on your system
15:44 Coke fyi, the hash seed issue is still impacting [incr] in partcl-nqp
15:45 Coke (t/cmd_incr.t fails sporadically )
15:51 ruoso joined #parrot
15:53 davidfetter oi ruoso
16:00 tewk Coke: didn't microsoft offer remote machines for perl/parrot developers?  I'm looking for a contact for another open source project.
16:01 dmalcolm joined #parrot
16:04 estrabd joined #parrot
16:04 dalek parrot: r47354 | NotFound++ | trunk/src/pmc/boolean.pmc:
16:04 dalek parrot: modify Boolean init_pmc to shorten it and allow better coverage
16:04 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47354/
16:07 jan joined #parrot
16:09 particle tewk: alias was working on that, i don't know what ever happened to it
16:09 particle it was supposed to be that anyone with a pause id could get access
16:16 ash_ ... i figured it out, i was just being dumb with a pointer, well now i can parse structs, and signed vs unsigned char/short/int/long/long long's
16:16 whiteknight ash_++ # nice!
16:16 ash_ and pointers work
16:17 pmichaud good morning, #parrot
16:18 whiteknight good morning, pmichaud
16:19 pmichaud whiteknight: I'm really excited about your observations on the return continuations last night... I'm wondering when I'll get a chance to update nqp to use it :)
16:20 whiteknight pmichaud: I'm glad to be of service! My big motivation is to get Matrixy's call/return semantics fixed, and I think these changes will really help
16:20 pmichaud also, I have to figure out if nqp's  "return"  structure needs to be more like parrot's .return()  or rakudo's return(), or if there's even a conflict.
16:20 pmichaud still working out that detail.
16:20 whiteknight does nqp do nested lexical scopes?
16:21 pmichaud yes, at least as far as I understand the term "nested lexical scopes"  :-)
16:21 jnthn pmichaud: At the moment, Rakudo always wants to return just one thing (a Parcel)
16:21 pmichaud jnthn: right.
16:21 jnthn pmichaud: Which is pretty much spec.
16:21 whiteknight I mean, is one function in NQP composed of multiple lexical .subs in Parrot?
16:21 pmichaud whiteknight: yes, unless NQP knows it can optimize one out.
16:22 whiteknight oh, nice
16:22 * whiteknight loves hearing the word "optimize"
16:22 pmichaud (which it can do for blocks that don't have any local lexical declarations)
16:22 jnthn pmichaud: Thing is we may want to do list assignment, or perhaps coerce it to a Capture so we can do a bind and unpack stuff.
16:23 pmichaud jnthn: looking.
16:23 jnthn pmichaud: So it's best to keep it as one parcel until something we return it to supplies a context.
16:23 jnthn pmichaud: Either way, if PCT supports multiple return values, that's fine. Rakudo will just return one value that is a parcel.
16:23 pmichaud jnthn: yes, but the difference between a parcel and a list is really just  :flat/:slurpy, though
16:24 theory joined #parrot
16:24 jnthn pmichaud: We depend on the thingy coming back as being of type Parcel ATM.
16:24 jnthn (If there's multiple values.)
16:25 jnthn I'd really rather not change that, since it's one place where things are easy because we just do it pretty much per spec. ;-)
16:25 pmichaud I'm not proposing that rakudo chagne.
16:25 pmichaud *change
16:25 jnthn OK, then I'm happy. :-)
16:25 pmichaud I'm asking whether nqp absolutely has to do the same, or whether having nqp do a multi-return value sort of thing can work also.
16:26 jnthn pmichaud: One consideration is that VMs having a concept of multiple return values is kinda rare.
16:26 pmichaud in many ways it would be nice (for Parrot)   if    return $x, $y, $z    could result in   .return ($P0, $P1, $P2)
16:27 jnthn Yes, true.
16:27 pmichaud currently   return $x, $y, $z   causes NQP to return a RPA containing $x, $y, and $z  (basically same as rakudo)
16:28 jnthn Thing is that then does NQP let you say e.g. my ($x, $y, $z) := func();
16:28 jnthn (That is, do we have a way to accept multiple return values.)
16:28 pmichaud don't know yet
16:28 pmichaud that's part of what would need to be worked out
16:29 pmichaud (and I'm still thinking about the notion of binding in this manner anyway)
16:29 jnthn pmichaud: Well, in the above case in Rakudo we fall straight into the signature binder.
16:29 sorear multiple return values... in NQP?
16:29 pmichaud right.  and in nqp/parrot, our signature binder is effectively PCC
16:29 jnthn It'd somewhat fit NQP to implement it in terms of Parrot's signature support.
16:29 pmichaud exactly.
16:30 jnthn And for other VMs we'll need the same machinary for Perl 6 anyways, so I guess it's hardly a porting impediment.
16:30 jnthn OK, I think +1.
16:30 pmichaud jnthn: yeah, as I said -- I'm still having to think through the ramifications.  :-)
16:31 pmichaud but the notion that other VMs would need similar machinery is something I hadn't really thought of yet, so +1 for that
16:31 ash_ sweet, now i have a test of it loading a library dynamically, making a signature and calling it with libffi, if anyone wants to see. Now i just have to put that back into parrot...
16:31 jnthn pmichaud: In a sense, PCC just packages all the arguments up into one data structure anyway. ;-)
16:32 jnthn pmichaud: The one place we're likely to hit a nasty difference is named return arguments.
16:32 pmichaud jnthn: yes, that's true also.  It creates a CallSignature or something like that iirc.
16:32 jnthn pmichaud: Because:
16:32 pmichaud I'm not too worried about named return arguments for nqp.  So far we haven't had much of a need for them.
16:32 jnthn sub foo() { return 1, 2, :c<3> }; my ($a, $b, $c) = foo(); # $c is a Pair
16:33 jnthn sub foo() { return 1, 2, :c<3> }; my ($a, $b, $c) := foo(); # also works because of named to positional mapping, but $c is 3
16:33 jnthn sub foo() { return 1, 2, :c<3> }; my ($a, $b, :$c) := foo(); # $c is 3
16:33 jnthn Oh heh
16:33 pmichaud nqp doesn't have assignment, so the first one isn't an issue.  :)
16:33 jnthn The binding case *would* do what Parrot does :-)
16:33 jnthn Gee, good job nqp doesn't do assignment. :-)
16:34 pmichaud well, that's partially why it doesn't do assignment.
16:34 sorear all of what you're talking about now has been on my wishlist for weeks
16:34 pmichaud A lot of things become much simpler if we don't have to worry about list assignment versus item assignment versuss...
16:34 pmichaud and if we treat everything as a bind
16:34 jnthn Indeed.
16:35 pmichaud sorear: glad we're hitting your wishlist :)
16:42 lucian joined #parrot
16:43 Coke tewk: theoretically. I heard something from alias about that a long time ago, pinged him recently, and am pretty sure I got no helpful reply. Checking...
16:47 tewk If microsoft is still doing the program I'd like to just get a contact I can send an email to.
16:48 particle i know folks at ms that may be aware of a program... cotto_work for example.
16:48 Coke tewk: ah, I remember. I emailed alias's cpan addr, which bounced, then I tracked him down on IRC. pretty sure I got a "yes, I'm the guy doing that" and not much else.
16:48 Coke but that irc window is long long gone.
16:48 particle Coke:  your client doesn't capture history?
16:49 particle my every conversation is logged
16:49 atrodo_ joined #parrot
16:50 Coke I'll be sure to mention that to the lawyers. ;)
16:50 particle logged "for training purposes" ;)
16:51 cotto ohai
17:01 whiteknight all the smolder reports I've submitted this morning look god
17:01 whiteknight good
17:02 whiteknight There isn't a single linux/gcc report in the list of the last 5 reports
17:07 PacoLinux joined #parrot
17:11 PacoLinux_ joined #parrot
17:22 theory joined #parrot
17:26 tcurtis joined #parrot
17:34 jsut_ joined #parrot
17:40 japhb Anyone here from the PaFo board?
17:43 hercynium joined #parrot
17:43 Coke IWBNI if 'make smoke' didn't wait until the end to complain about not having LWP::UserAgent
17:43 Coke japhb: yes.
17:44 japhb Coke: Have you heard anything about whether my Plumage grant application was shopped around, and if so, whether anything came of it?  Last I heard, it had been board approved, and shopping was soon to commence ... then nothing.
17:45 Coke "board approved" is strong language. =-)
17:45 japhb "Not board rejected"?  :-)
17:45 Coke I believe Allison was going to do the shopping, and she is in the middle of finals, IIRC.
17:45 japhb *sigh*
17:46 japhb Any idea when her school year is over?
17:46 Coke we just don't have a grants/fundraising presence at the moment. it's something I would definitely vote for if someone ran for the board on. =_)
17:46 Coke I think she said she'd be freed up next week. I have no idea if shopping you about is still on her radar.
17:47 japhb Fair enough.
17:47 japhb Thank you for the info.
17:47 Coke I wish I had contacts to raise money.
17:48 japhb Ditto.  ;-)
18:06 clinton joined #parrot
18:19 dukeleto Parrot could get funding via running on RTEMS. Parrot on RTEMS opens up many, many doors.
18:20 LoganLK joined #parrot
18:23 theory joined #parrot
18:23 bubaflub booyah.
18:24 theory joined #parrot
18:24 cotto_work tewk: ping
18:29 joeri joined #parrot
18:30 tewk cotto_work: yes
18:30 cotto_work what was that MS program you wanted to know about?
18:31 tewk MS?
18:31 purl somebody said MS was 'WE HATE USERS AND WANT YOU TO DIE'
18:31 cotto_work microsoft
18:31 purl microsoft is not a software company, it's an abuse company that delivers it's abuse through software or http://moishelettvin.blogspot.com/20​06/11/windows-shutdown-crapfest.html or banking on their secret R&D.. robotic voice controlled helper monkeys or http://www.cracked.com/blog/2008/04/18/​the-first-and-last-time-ill-ever-work-f​or-steve-ballmer-the-friday-nooner-est/
18:32 tewk alias had a contact that supposedly was going to provide remote desktop access to windows machines for developing opensource software.
18:32 tewk a microsoft contact that is
18:33 cotto_work I got that started but I'm not sure who his direct contact at MS is.
18:34 tewk Well I was looking to make the same connections for PLT Scheme
18:35 cotto_work Have you tried talking to alias?
18:42 tewk cotto_work: no coke just told me to talk to alias.  I just found a local windows box I can use, but it would be nice to just rdesktop to one a microsoft, instead of hunting around and installing one locally every time i need it.
18:43 cotto_work ok.  There's #msopensource on irc.perl.org, though it's not very active.
18:44 tewk cotto_work: ok thanks
18:49 cotto_work If you'd like, I can also take a shot at getting you a downloadable copy of VS 2010.
18:51 cotto_work2 joined #parrot
18:52 tewk cotto_work: thanks, I have access to msdnaa, I just need different architectures and large multi-core boxes to run tests on.
18:52 cotto_work2 ok
18:56 aukjan joined #parrot
19:08 dalek parrot: r47355 | tcurtis++ | branches/gsoc_past_optimization (3 files):
19:08 dalek parrot: Fix my pir:: ops to use the appropriate signature syntax. Previously, they has been working purely by accident, seemingly.
19:08 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47355/
19:12 jjore joined #parrot
19:17 cotto_work2 who adds feeds to planet parrot?
19:18 cotto_work2 planet parrot?
19:18 purl planet parrot is an aggregate of any parrot-related blogs. or editable by coke or changable with a trac ticket. or http://planet.parrotcode.org/
19:19 cotto_work msg coke Khairul's blog http://parrot.mangkok.com/ should be added to planet parrot.
19:19 purl Message for coke stored.
19:25 jjore_ joined #parrot
19:41 dalek parrot: r47356 | tcurtis++ | branches/gsoc_past_optimization (2 files):
19:41 dalek parrot: Begin implementing smart matching PAST::Pattern children & tests for smart matching of children that are also PAST::Patterns.
19:41 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47356/
20:03 Psyche^ joined #parrot
20:06 davidfetter joined #parrot
20:07 theory joined #parrot
20:08 theory joined #parrot
20:08 dalek rakudo: 7c83e65 | (Solomon Foster)++ | src/core/Any-list.pm:
20:08 dalek rakudo: Switch sub version of map to take slurpy values.
20:08 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​c83e65fb2cda0ddd85e6abe98b18ee7da4a465f
20:17 dmalcolm joined #parrot
20:43 theory joined #parrot
20:45 bluescreen joined #parrot
20:47 dalek parrot: r47357 | tcurtis++ | branches/gsoc_past_optimization (2 files):
20:47 dalek parrot: Smart matching on PAST::Pattern, regex, and sub children tested and working.
20:47 dalek parrot: review: http://trac.parrot.org/parrot/changeset/47357/
21:00 jsut joined #parrot
21:27 dalek partcl-nqp: 25f4b03 | Coke++ | src/Partcl.pir:
21:27 dalek partcl-nqp: printerr is now a parrot dynop.
21:27 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/25f4b0369af4025134969718b582cc05938ea10c
21:31 sorear Coke: What's the status of partcl vs. partcl-nqp?
21:31 Coke sorear: both now work against parrot HEAD!
21:32 Coke basically the same state as master vs. ng in rakudo, before ng became master.
21:33 Coke need more?
21:33 sorear no
21:39 Coke cotto_work: done.
21:55 hercynium joined #parrot
22:15 cotto_work Coke: thanks!
22:19 GeJ Good morning everyone.
22:20 davidfetter g'day, GeJ
22:21 GeJ Bonjour davidfetter
22:22 bacek Good morning, humans
22:22 sorear Greetings.
22:22 purl somebody said Greetings. was anyone awake?
22:22 cotto_work hi bacek
22:22 bacek aloha, cotto_work
22:23 GeJ G'Day bacek.
22:24 bacek G'Day GeJ
22:41 ash_ joined #parrot
22:50 LoganLK joined #parrot
22:58 ash_ what is the nci attribute 'long_signature' used for? i don't see it actually used anywhere in the source code
23:08 tetragon joined #parrot
23:11 KatrinaTheLamia joined #parrot
23:13 bubaflub joined #parrot
23:18 dalek tracwiki: v15 | bacek++ | GCMassacre
23:18 dalek tracwiki: http://trac.parrot.org/parrot/wiki/G​CMassacre?version=15&amp;action=diff
23:32 joeri joined #parrot
23:41 ash_ it seems to be used in multi-dispatch

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

Parrot | source cross referenced