"Please mentor for SoC | parrotcode.org/
Set by moderator on 3 March 2008.
cotto_ Is there any reason not to set PARROT_GC_DEBUG while developing? 00:08
it seems to be good for exposing bugs quickly
chromatic The environment variable, the build flag, or the runcore? 00:09
cotto_ env variable
chromatic I've never used it; what does it do? 00:11
cotto_ makes garbage collection happen as frequently as possible, same as passing --gc-debug to parrot 00:14
s/happen/run/
svnbotl r26209 | chromatic++ | pdd17pmc: 00:29
: [lib] Made PMC parser handle non-register pointers for PMC attributes.
: This has some implications for generated accessors, but if we're cautious we
: can handle things.
r26210 | chromatic++ | pdd17pmc:
r26211 | chromatic++ | pdd17pmc: 00:35
: [PMC] Removed dodgy typecasting in NCI PMC now that attributes can be any type
: of pointer.
diff: parrotvm.org/svn/parrot/revision/?rev=26211
r26212 | chromatic++ | pdd17pmc: 00:42
: [PMC] Fixed an incorrect return value from the readline() method.
diff: parrotvm.org/svn/parrot/revision/?rev=26212
01:07 skv_____ joined
svnbotl r26213 | chromatic++ | pdd17pmc: 01:07
: [lib] Added types matching the form Parrot_\\w+ to the allowed list of PMC
: attribute types. This is necessary for some of the event system PMCs.
r26214 | chromatic++ | pdd17pmc:
: [PMC] Migrated the Scheduler PMC to the new attribute declaration syntax.
01:10 d4l3k_ joined
diakopter svnbotl will be flapping whilst I reattempt to throttle its messages 01:11
01:13 GeJ joined 01:16 LimbicRegion joined 01:17 svnbotl joined
diakopter let's try this. somebody commit something 01:18
pugs can't handle 'sleep', but it can handle system calls to perl -e 'sleep 1'. We'll see if that works. 01:20
chromatic Hold on.
Ah, nevermind. I already checked it in. 01:21
cotto_ should clone() create a shallow clone? 01:30
chromatic Probably not. 01:31
cotto_ ok
chromatic I don't know if we've ever made a ruling on that though.
cotto_ I just misread the hash code. it makes a deep clone
except keys for some reason 01:33
01:53 katie_d joined 02:06 kid51 joined
diakopter waiting for someone to commit something.. 02:18
:D
02:18 GeJ_ joined
kid51 diakopter: Why? Have you got new tricks for svnbotl? 02:19
02:20 dalek joined, particle joined
diakopter kid51: yes; hopefully it will throttle its irc posts 02:23
kid51 Does that mean that it won't leave channel when a commit with many files happens? 02:32
diakopter could be. 02:34
though likely not. likely that means the parrotvm.org svn diff timed out
02:42 Ademan joined 02:44 RichiH joined 02:46 dalek joined 02:51 AndyA joined 02:52 c9s joined 02:53 Maddingue joined 02:55 svnbotl joined
diakopter svnbotl got netsplit. luckily it didn't miss any commits 02:58
diakopter waits more
chromatic Here's one. 02:59
kid51 must sleep 03:00
purl $kid51->sleep(8 * 3600);
svnbotl r26217 | chromatic++ | pdd17pmc: 03:01
: [Tcl] Updated Tcl PMCs to match PDD 17. The previous use of C<does> is now
: C<provides>.
: Some tests still fail, but at least most tests run now.
diff: parrotvm.org/svn/parrot/revision/?rev=26217
diakopter arghhh
must be an off-by-one error, where the one in this case is which file I'm editing. 03:02
Tene pleasedieinafire.net/~tene/gil.tar.bz2 03:03
Generic Imperative Language
Thrown together in just over two hours at a coffee shop tonight.
Going to use it for my presentation.
03:10 petdance joined
Tene I think it'll make for a more fun presentation than abc would 03:12
Oh, anyone around to talk about PGE? 03:13
I'm a bit too lazy to put together a bug report for RT without discussing it with someone right now.
Tene tries pushing out patches for lolcode via git-svn 03:14
03:19 Coke joined
svnbotl r26218 | coke++ | trunk: 03:20
: [tcl]
: use less .Undef
diff: parrotvm.org/svn/parrot/revision/?rev=26218
03:22 chromatic joined
Coke chromatic: with reversion 26218, I'm getting the same build errors that kid51 reported; can't get to tcl to verify anything there. (on osx/intel . trying now on feather...) 03:24
.. reversion. There's an interesting typo. 03:26
svnbotl r26219 | tene++ | trunk: 03:45
: Actually let lolcode parse multiple arguments to function definitions.
: There's a big problem here either with PGE or my understanding.
: Add isdecl(1) to variable declarations that don't provide an inital value.
diff: parrotvm.org/svn/parrot/revision/?rev=26219
Tene working on building a test case for the PGE issue. 03:48
katie_d I'm trying to build Parrot, and "make" is failing. I've pastebin'ed the error and some of my system info here --> www.pastebin.ca/928563. Please let me know if there is any other information I should provide. 03:58
Coke katie_d: if no one can help you out while on IRC, please open a ticket. (docs/subm*). good luck. 03:59
katie_d Coke: OK thank you. Did I provide enough information, or should I add other stuff along with what I posted? 04:00
Coke what arguments did you pass to Configure.pl 04:01
(only other thing I see.)
-> zzz 04:02
katie_d I didn't pass any args -- just "perl Configure.pl"
But I'll add that in there. Thanks for your help.
05:00 allison joined 05:01 chromatic joined
svnbotl r26220 | chromatic++ | pdd17pmc: 05:34
: [Lua] Ported PMCs to PDD17. In particular, all METHODs now use Parrot's
: calling conventions. This means that they have to use PCCRETURN appropriately
: and that you can only call them (for now) with some form of PCCINVOKE.
: Not all tests pass, but at least the PMCs all compile and pass some tests.
diff: parrotvm.org/svn/parrot/revision/?rev=26220
05:37 lathos joined 06:27 chromatic joined, allison joined
svnbotl r26221 | allison++ | pdd17pmc: 06:49
: [pdd17pmc] Automatically generate accessors for attributes declared with ATTR.
r26222 | allison++ | pdd17pmc:
: [pdd17pmc] Update Exporter as an example of using new generated accesors.
diff: parrotvm.org/svn/parrot/revision/?rev=26222
r26223 | chromatic++ | pdd17pmc: 06:55
: [PMC] Made classes which inherit from PMCs respect the roles that the parents
: provide. PMCProxy did not check what the proxied PMC returned for does(). Now
: it, well, does.
: I added a test for this just in case. (Amusingly the test fails on the trunk.)
r26225 | allison++ | pdd17pmc: 07:02
: [pdd17pmc] Adding additional notes on attribute generation and inheritance to
: the PMC PDD.
diff: parrotvm.org/svn/parrot/revision/?rev=26225
07:05 contingencyplan joined 07:14 uniejo joined 07:19 slightlyoff joined 07:52 slightlyoff joined 07:56 cosimo joined
svnbotl r26226 | allison++ | pdd17pmc: 08:10
: [pdd17pmc] Missed the continuing slashes on the end of some macro lines.
diff: parrotvm.org/svn/parrot/revision/?rev=26226
08:11 Ademan joined 08:49 DarkWolf84 joined 09:04 IllvilJa joined
katie_d Can someone briefly explain to me what it is that headerizer is doing with __attribute__nonnull__($) ... or at least point me to some documentation (just a search term for Google would be quite helpful) about it. 09:05
lathos points at Andy. He's the headerizer man. 09:06
katie_d Found it :) If I'm understanding correctly, it just makes GCC throw a warning if you pass NULL for that arg. 09:17
09:24 ejs joined 09:59 AndyA joined 11:24 mdiep joined 11:25 ruoso joined
spinclad katie_d: that sounds like it exactly. (varying by compiler, not GCC only) 11:32
11:47 ejs joined 12:34 kid51 joined 12:51 pjcj joined 13:44 Coke joined 13:52 ask_ joined, petdance joined 14:02 gryphon joined
Coke fixes the build on pdd17pmc. whee. 14:04
Coke wonders why it doesn't seem to be broken for chromatic or allison.
davidfetter anybody here looked @ SMOP? 14:08
svnbotl r26227 | coke++ | pdd17pmc: 14:15
: [build]
: add missing headers...
diff: parrotvm.org/svn/parrot/revision/?rev=26227
Coke seen barney? 14:42
purl barney was last seen on #parrot 1 day and 18 hours ago, saying: particle: mailed report [Mar 3 12:56:44 2008]
14:47 jdporter joined
particle davidfetter: i'm familiar with smop 14:50
jdporter where would I find #p5p?
i don't see it on irc.perl.org
davidfetter particle, i'm curious as to your thoughts on this. i spoke recently with daniel ruoso, but of course, his PoV is biased in its favor 14:51
ruoso davidfetter, hi david... 14:52
davidfetter heh
ruoso have you seen www.perlfoundation.org/perl6/index.cgi?smop
? 14:53
davidfetter bookmarks
purl bookmarks is a convenience term for that or unneeded in Congress... they just bend over the pages
lathos Where's the code? 14:54
purl i guess the code is actually compiled - but it's added incrementally to the environment
ruoso lathos, smop's?
lathos Mmm.
particle jdporter: you want irc.freenode.net
ruoso www.perlfoundation.org/perl6/index....op_hacking
particle iirc
ruoso particle, jdporter, is it? I'm in #p5p at this server...
lathos Interesting, reading. 14:55
particle lathos: the code is in pugs repo, under v6/smop/
lathos I know, I'm reading it. :)
particle aha
it's a nice read :)
re #p5p, i don't hang out there, so maybe i'm not the best to ask 14:56
jdporter thanks; see ya there 14:57
lathos I know what this reminds me of. 15:00
Objective C runtime source. 15:01
15:01 rdice joined
lathos Which is a good sign, because they're in a very similar problem space. 15:01
ruoso SMOP is being compared to a lot of runtimes lately... 15:03
erlang, smalltalk, io
lathos Well, it's all a very similar problem space. :) 15:04
davidfetter ruoso, is there someplace we're supposed to post our slides? 15:07
particle allison: ping
15:07 mncharity joined 15:13 ask_ joined
ruoso davidfetter, I think you must send it to cog 15:14
davidfetter email? 15:15
purl i heard email was a view.
ruoso davidfetter, yes... 15:19
15:32 iblechbot joined 15:35 sjansen joined
Coke in pdd17, ../../parrot tcl.pbc -e "puts [concat {expand}[lindex {bar baz} 1]]" 16:01
generates an error that seems to come from PGE. 16:02
16:17 parrot-poke joined
Coke ah. no, it's from inter_call.c 16:20
davidfetter
.oO(intercal.c?)
 16:21
16:27 Patterner joined 16:43 barney joined
particle coke: ping 17:05
Coke pong 17:21
... but not for long. 17:22
17:22 skv_____ joined 17:24 Psyche^ joined
Coke no really.. 17:28
particle coke: 17:29
s/does/provides/ after pmclass before {
Coke particle: Saw your email. that's what chromatic fixed in the revision number in that thread. 17:30
... and I already replied to the email. (you'll get two copies, sorry) 17:31
particle ah. saw your email, now :)
barney Has DYNSELF changed to SELF in pdd17 ? 17:38
allison barney: yes 17:39
17:44 slightlyoff joined
svnbotl r26228 | bernhard++ | pdd17pmc: 17:45
: [Plumhead]
: DYNSELF -> SELF for PDD17 compliance.
diff: parrotvm.org/svn/parrot/revision/?rev=26228
barney allison: PHP looks now mostly fine with pdd17
allison barney: excellent, thanks! 17:46
Coke tcl, on the other hand... =-)
-> lunch
PerlJam what's pdd17? 17:47
purl pdd17 is new, and we're implementing it in the branch
particle purl++ 17:48
the branch name is pdd17pmc
17:48 stephen joined, stephen left
pmichaud the pdd17pmc branch is for the PMC re-implementation 17:48
17:52 Psyche^ joined
cognominal what is changing? 17:52
purl changing is probably *much* easier than adding
cognominal forget changing 17:53
purl cognominal: I forgot changing
pmichaud iirc, ppd17 changes the base structure and implementation of PMCs
but I don't know the full details
cognominal ok
particle on the surface, the syntax for defining pmcs changes 17:54
below the surface, it opens the door for role-based composition of pmcs (not yet implemented)
cognominal hope this will somehow fix the bugs I filed today.
particle you can svn co the branch and see for yourself :) 17:55
PerlJam yeah, I just did that.
roles certainly are pervasive
cognominal well, I prefer to wait the merge.
particle allison: i'm rebuilding pdd17pmc branch now 17:57
allison particle: great
particle i hope to make the call today, but may miss it 18:00
mncharity particle: hi. could you give me some guidance on when parrot might be able to provide ast yaml dumps of its p6 parsing? I realized the calendar has slid past when I assumed it would exist, but I never asked you for an actual estimate. Do you currently think it more like 1 week, 4, something else? Thoughts? My thanks. 18:01
PerlJam mncharity: what does the yaml buy you? 18:02
mncharity a p6 parser.
PerlJam oh, you've got something that groks yaml and generates code?
mncharity yes.
that part is very easy.
as long as one is only trying to "have something which works, and can be used for simple p5-style p6". 18:03
rather than being full p6.
but enough for people to start being able to write and run significant p6 code.
PerlJam the more p6, the better is what I say.
mncharity currently one simply can't run non-trivial p6 code. having a parser willing to dump yaml (and including oo info, unlike pugs), would allow that to finally, easily, be changed. 18:04
"non-trivial" meaning - uses objects. 18:05
pmichaud rakudo uses objects
particle mncharity: i haven't had tuits to finish, sorry
i hope to hand the task to coke
it's pretty close
PerlJam pm: does it do regex (at least to the extent that PGE allows)? 18:06
particle allison: same failure
pmichaud pj: not quite yet, but we're fairly close on that
particle allison: i can fire up the debugger in 15min or so if you're around to walk through it
pmichaud it does anonymous regex, and named regex
mncharity re tuits, ah, so it's actually stalled out, rather than simply being more difficult than anticipated? ok. I guess that clarifies things. glad I asked, rather than continuing to wait. :)
pmichaud it doesn't yet do grammar, token, or rule
(it will parse token/rule, but they don't yet work)
particle mncharity: yes, just ENOTUITS
allison particle: I'm free now.
particle i *have to* shower. bbi15m 18:07
PerlJam perhaps I'll look at that aspect of rakudo if/when I get some time
pmichaud it requires a change or two in PCT
svnbotl r26229 | bernhard++ | trunk:
: [doc]
: Update imcfaq.pod. Some more replacements IMC -> PIR.
: Mention PAST and the PIR tutorial.
diff: parrotvm.org/svn/parrot/revision/?rev=26229
allison particle: I'll be free in 15 minutes too
avar mncharity: does parrot dump something already which you'd want from the yaml but not in that format? 18:08
pmichaud we already have --target=parse
it dumps the parse, albeit not in yaml 18:09
PerlJam Write past2yaml and you're there! :)
mncharity pmichaud: re "rakudo uses objects", yes. But if the question is "so why would one need any other backend?", the answer is "because building on a high level langauge (p5 or ruby), one gets lots of things for free". So which achieves rakudo's goal of "p5 programmer can hack happily" is more like 3 days away, rather than... somewhat further.
pmichaud I wasn't taking a position against other backends, I was simply pointing out that there is something that "uses objects" (which was your definition of "runs non-trivial p6 code") 18:10
mncharity hmm, particle? so the only issue is format, you weren't also having to clean up the tree? /me looks at --target=parse 18:11
re definition, ahhh, ok.
It was intended as necessary, not sufficient. :)
a delta against pugs.
avar tt/w 17 18:13
gah
mm, `make' never works after svn up 18:14
18:15 Psyche^ joined
mncharity hmm. is there anything existing which can parse the --target=parse output format? 18:18
to pull it into p5, say
pmichaud I'm not aware of anything 18:19
mncharity ok.
avar what handles the --target emitting?
pmichaud Parrot's Dumper.pbc
and PGE::Dumper
runtime/parrot/library/PGE/Dumper.pir 18:20
avar sial.org/pbot/30425 18:22
mncharity gets a current version of parrot... 18:24
Tene So, if I want to implement IO stuff in Rakudo, am I going to need to create a PMC, or just add class methods in src/classes/IO.pir
?
particle ...and particle committed runtime/parrot/library/yaml_dumper.pir for yaml emitting 18:25
mncharity looking... tnx... 18:26
pmichaud the IO stuff for Rakudo would likely need to be mapped to Parrot's IO objects
particle allison: i've got about 20m to debug
nopaste "particle" at 24.19.3.148 pasted "patch to enable yaml output with pct" (31 lines) at nopaste.snit.ch/12444 18:27
particle that patch allows you to emit yaml instead of dumper format 18:28
18:28 slightlyoff joined
particle however, the pmc types need a helper function to emit yaml appropriately 18:28
allison particle: okay, first thing that would be useful is to find out where it's segfaulting 18:29
particle likely that just means that the base past object type needs that helper function
pmichaud there isn't a base past object type 18:30
(well, it's Capture_PIR)
but the other pmc types are just the Parrot PMC types 18:31
particle right now, if i type >perl 6 --target=past
> say 1;
---
{
"past" : pmcPAST::Block
!PAST::Block { ... },
}
it's calling the pmcDefault sub in runtime/parrot/library/YAML/Dumper/Default.pir 18:32
Tene pmichaud: can you tell me more about what that mapping would mean, exactly?
pmichaud phone (bbiab)
particle however, it should recognize that past nodes are hashes, and call that func
allison: Parrot_unmake_COW 18:33
there's trouble with mem_sys_memcopy
Tene purl: nopaste?
purl nopaste is at nopaste.snit.ch/ (ask TonyC for new channels) or rafb.net/paste or paste.husk.org/ or poundperl.pastebin.com/ or paste.scsys.co.uk/ or don't bother me while I'm eating
pmichaud Tene: well, I'm not very familiar with Parrot's IO subsystem. (And I think it's changing, anyway.) 18:34
But I know that there's a class called "ParrotIO", and it has a variety of opcodes/methods for doing I/O types of things
so, a Rakudo IO object would need to map its methods and functions into an equivalent ParrotIO object, or a wrapper object that can forward to a ParrotIO object, or something like that 18:35
Tene So just a bunch of wrapper methods, right?
pmichaud it could be done that way, yes.
PAST::Node is more than just Hash
sorry
particle: PAST::Node is more than just Hash
PAST::Block isa PAST::Node isa Capture_PIR isa Hash+ResizablePMCArray 18:36
actually
particle pmichaud: okay, then it needs a helper function called 'yaml'
pmichaud PAST::Block isa PAST::Node isa Capture_PIR
Capture_PIR has Hash has ResizablePMCArray has Scalar
feel free to create a 'yaml' function for Capture_PIR :-) 18:37
(function or method?)
I need to go get lunch for Paula, bbl
I may also miss today's phone -- lots of errands
particle i'm hoping coke can backlog, see what's up, and continue 18:39
heh, or tene
allison particle: since you've gotten to the point that parrot.exe is compiled, could you run some other tests on it? 18:40
particle: make test won't work, but a 'prove' with a specific test file will
particle sure
mncharity pmichaud: thanks for your help 18:41
18:46 Coke left
nopaste "mncharity" at 76.24.29.201 pasted "irc comments re dumping yaml version of --target=parse" (95 lines) at nopaste.snit.ch/12445 18:47
particle crappity, i've gotta run :(
i'll be offline for the next 6h or so. hope to make the design call in ~75m. 18:48
mncharity particle: thanks for your help.
not sure how to proceed. pondering.
18:50 peeps[work] joined
svnbotl r26230 | bernhard++ | trunk: 18:57
: [docs]
: s/Perl/Perl 5/
diff: parrotvm.org/svn/parrot/revision/?rev=26230
19:00 kismet left
Tene I had a weird problem with PGE yesterday, but everything seems to be just fine today... 19:00
svnbotl r26231 | bernhard++ | trunk: 19:04
: [doc]
: Spelling and markup fixes.
r26232 | bernhard++ | trunk:
: [docs]
: Update a couple of copyright years.
diff: parrotvm.org/svn/parrot/revision/?rev=26232
19:16 slightlyoff joined 19:25 Ron joined 19:30 Coke joined
svnbotl r26233 | bernhard++ | trunk: 19:54
: [docs]
: Fix POD.
diff: parrotvm.org/svn/parrot/revision/?rev=26233
TimToady phone 20:00
20:29 Theory joined 20:42 stephen joined, stephen left
svnbotl r26234 | chromatic++ | pdd17pmc: 21:02
: [Plumhead] Fixed mark() so that it marks *all* hash entries, not every other
: bucket.
diff: parrotvm.org/svn/parrot/revision/?rev=26234
cotto_ heh. I found that bug too while doing some other debugging 21:03
seen barney 21:21
purl barney was last seen on #parrot 3 hours and 37 minutes ago, saying: allison: PHP looks now mostly fine with pdd17
21:42 Theory joined
svnbotl r26235 | chromatic++ | pdd17pmc: 21:52
: [PMC] Changed ParrotLibrary PMC creation so that they are now constant, as is
: their metadata.
: When performing full global destruction, occasionally a ParrotLibrary PMC would
: get destroyed before the dynpmc which had vtable entries from the shared
: Now that ParrotLibrary PMCs are constant, they get destroyed only after all 21:53
: non-constant PMCs.
diff: parrotvm.org/svn/parrot/revision/?rev=26235
tewk exit 21:54
21:54 davidfetter joined 22:12 Ademan joined 22:22 Coke joined
mncharity fyi, from #perl6, misc/winter_jig/ast_rakudo/parse_t_spec.t: Created. Runs rakudo --target=parse against pugs t/spec/ .t files. Currently 35 of 128 files are successfully parsed. One non-terminates (S29-conversions/ord_and_chr.t). 22:34
particle and everyone: I suspect addressing the parsefails should be higher priority than finishing the yaml dumper. Unless the two tasks are non-conflicting. 22:41
22:49 japhb joined 23:16 Limbic_Region joined 23:19 Andy joined 23:27 Ademan joined 23:28 slightlyoff joined
allison any windows devs around? 23:47
davidfetter might be able to scare one up. what's shakin'?
23:48 silug joined
allison is this guide at all up-to-date? www.perlfoundation.org/parrot/index...on_windows 23:54

Powered by Cro and the Raku® Programming Language.

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

Thank you!