Parrot 1.1.0 Released | parrot.org/ | 322 RTs left | Weekly Priority: Remove Deprecated Items
Set by moderator on 6 May 2009.
jonathan You might be the first person writing an emitter in NQP. :-) 00:00
bacek At least second after pmichaud :)
jonathan Oh? I didn't realize pm did an emitter too (OK, other than the Perl 6 to LOLCODE translator... ;-)) 00:01
bacek O wait. He emits from PCT.
jonathan Yeah, the emitter for PCT is written in PIR.
And yes, it uses multi-dispatch.
bacek And I using same approach. But I want NQP :) 00:02
I'll kill dalek again with my commit message. 00:05
dalek rrot: r38662 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter/c.pir:
Pass pmclass PAST as first argument during parsing.
 00:07
bacek He survived. That's good 00:08
afk # housekeeping
Whiteknight masak: ping 00:50
or, i guess this question could be answered by any Perl 6 hacker
any of that species around? 00:51
cotto Isn't there a channel for that? ;) 00:52
Infinoid I've been meaning to learn it, but it's a rather daunting task.
cotto ditto
although working with nqp helps, since it's a subset 00:53
Whiteknight yeah, I suppose i can go there, but it's on adifferent server
kid51 Possible 'make' failure: 01:55
make: *** No rule to make target `include/parrot/gc_mark_sweep.h', needed by `src/string/api.o'. Stop.
At: /usr/local/bin/perl tools/build/vtable_extend.pl 01:56
make: *** No rule to make target `include/parrot/gc_mark_sweep.h', needed by `src/string/api.o'. Stop.
doing realclean;svn up;reconfiguring;rebuilding
Got this: ls: include/parrot/gc_mark_sweep.h: No such file or directory 01:57
src/string/api.c was touched in r38654, whiteknight's merging of the gc_api branch into trunk 01:59
Actually, src/string/api.c has been touched several times recently. 02:03
Looks like that was a false alarm. svn up, reconfigure and rebuild appears to have cleared up the problem. 02:10
Infinoid kid51: Do you build with make -j, by any chance? 02:20
02:33 mikehh_ joined
dalek rrot: r38663 | jkeenan++ | trunk/lib/Parrot/Docs/File.pm:
Delete 2 inline comments so that we can close rt.perl.org/rt3/Ticket/Display.html?id=43683.
 02:34
Infinoid hmm, nevermind. I don't see how it could be a missing makefile dependency, I can't find any references to that header at all 02:35
02:35 janus joined 02:38 mikehh__ joined
kid51 Infinoid: No, I did not. But I think the initial report was wrong. I did an 'ls Makefile', inferred that I was ready to 'make' and forgot that earlier tonight I had done 'svn up' and pulled in all those changes from your branch. 02:44
What is 'make -j' anyway? 02:45
I subsequently built and tested successfully: smolder.plusthree.com/app/public_pr...ails/21195
Infinoid ah. make -j runs jobs in parallel, to take advantage of multiple CPUs/cores 02:50
It's also a good way to find missing makefile dependencies, because of its inherent randomness
kid51 Well, I certainly don't have multiple CPUs on this 5-year-old iBook. 02:52
Infinoid Yeah, I just thought your report might be something else, sorry 02:53
03:32 eternaleye joined
dalek rrot: r38664 | bacek++ | branches/pmc_pct/compilers/pmcc/src/parser/actions.pm:
Set pasttype to inline for inline nodes.
 03:43
rrot: r38665 | bacek++ | branches/pmc_pct/compilers/pmcc/src/parser/actions.pm:
Fix handling of macros flags in actions.pm
 03:46
rrot: r38666 | bacek++ | branches/pmc_pct (6 files):
Initial support for MACRO substitution
03:52 cognominal joined 03:57 Theory joined 05:07 Theory joined 05:14 Eevee joined
dalek rrot: r38667 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter (2 files):
Fix passing PMC PAST to C::emit
 05:34
rrot: r38668 | bacek++ | branches/pmc_pct/compilers/pmcc (4 files):
Inprove parsing of MACRO call arguments
 05:37
rrot: r38669 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter/c.pm:
Ooops. I forgot to add emitter/c.pm
rrot: r38670 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter/c.pm:
Remove useless FIXME comment
bacek cotto: Look! I've got something working for MACRO calls! :) 05:42
cotto so bacek, how long before you get L1 specified and implemented
I see what you did there.
bacek++ 05:43
bacek cotto: I need language/pir...
cotto How much more is there to do before pmcc can replace pmc2c?
bacek It's almost straight forward now. 05:44
dalek rrot: r38671 | bacek++ | branches/pmc_pct/compilers/pmcc (2 files):
SELF.method should generate VTABLE_method.
bacek We can start generating get_vtable/update_vtable now.
Last big chunk - rewrite METHODs.
cotto That's nontrivial. 05:45
I knew there was something big.
bacek Agreed... I think it's better wait for allison to finish pcc_rewired branch.
cotto yup 05:46
What about filenames?
bacek erm. filenames? 05:47
cotto are those available to the parts that need them (i.e. the .c, .h and .dump generating code)?
foo/bar/baz.pmc ->foo/bar/baz.c, etc
bacek ah. those filenames... it doesn't look hard to implement 05:48
cotto Also, unserialize_attrs in src/node.pir currently spits a warning to stderr when it can't find an expected dump. I'm not quite sure what dtrt is here, since some PMCs which are extended (mainly default) don't have ATTRs. 05:50
actually, it'll probably need a list of directories to look through for dumps, since dynpmcs can extend normal PMCs. 05:51
bacek current pmc2c silently regenerates .dump
cotto that'd work, as long as the compiler knows where to look for the .pmc files. 05:52
bacek You'll definitely "need a list of directories to look through for dumps" :) 05:53
Actually $cwd and "src/pmc" probably enough.
cotto It looks like dumps and pmcs are currently installed as part of either install or (more likely) install-dev 05:54
bacek O! We can supply them in command prompt
supply list of directories
cotto that would make the installed version easy too
bacek++ 05:55
bacek cotto++ # indeed :)
Ok. Looks like "someone" have to start extending command-line "pmcc" compiler to dump generated files into proper places. 05:59
Any volunteers?
purl Any volunteers are welcome and get beer and pizza.
bacek purl: good girl :)
purl thanks bacek :)
cotto I'll take that, although not tonight.
bacek cotto: ok. Thanks 06:00
I'll take class_init/get_vtable/update_vtable generating.
cotto actually, won't it be enough to pass in the path to the pmc, since the dump will go in the same place? 06:01
bacek dynpmc?
cotto of course there'll still need to be a search path for dumps
bacek :)
cotto dynpmcs are treated the same way as pmcs, as far as dumps go 06:02
bacek dreams about removing ATTRs from headers and reimplement ATTRs handling in pmcs...
Poking inside PMC isn't good idea 06:03
cotto I'm hoping they'll go away eventually, but pmcc won't be usable if it doesn't make them work like pmc2c does. 06:04
bacek agreed.
cotto looks forward to 3.0
bacek And 3.1 06:05
And 3.11 for Workgroups!
cotto It'll be a different bird altogether. 06:06
bacek :) 06:07
06:09 dduncan joined
bacek Ok. Little brake. I hope I'll implement get_vtable/update_vtable today (without MUTLIs of cause.) 06:09
cotto ok. I'll probably be in bed when you get back.
night
bacek night 06:10
06:17 dduncan left 08:50 register joined 08:51 iblechbot joined
dalek rrot: r38672 | bacek++ | branches/pmc_pct (4 files):
Add stub for default.pmc specialised handling.
 09:59
10:03 bacek_ joined
dalek rrot: r38673 | bacek++ | branches/pmc_pct/compilers/pmcc/src/parser/actions.pm:
Add some TODO comments
 10:12
rrot: r38674 | bacek++ | branches/pmc_pct (3 files):
Register PMC::Emitter::PMC::default as ancestor of PMC.
rrot: r38675 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter/pmc.pm:
Split generate_class_init into 3 parts
rrot: r38676 | bacek++ | branches/pmc_pct/compilers/pmcc/t/data/default.pmc:
Add stripped down version of default.pmc into test suite.
rrot: r38677 | bacek++ | branches/pmc_pct/compilers/pmcc/t/07-default.t:
Add (failing) test for default.pmc handling
10:17 gaurav joined
dalek a: 20bf64a | fperrad++ | src/pmc/lua (5 files):
fix build against trac.parrot.org/parrot/changeset/38654 (gc_api merge)

COMPATIBILITY BREAK		 10:45
bacek Guys... 10:51
pobject_lives even wasn't marked as deprecated. 10:52
And marked as PARROT_EXPORT...
Same for VTABLEs stuff... 10:53
jonathan There was a general deprecation notice in 1.0 about C function names not of the form Parrot_subsystem_foo 10:55
moritz what a clever move ;-)
bacek found deprecation notice about "pobject_lives"... 10:56
But can't find deprecation about VTABLEs... Looks like we have to revert my merges of tt631_part* branches. 10:57
jonathan bacek: What changed? 10:58
(that you feel needed a deprecation cycle?) 10:59
bacek We stop exporting generated VTABLEs functions. See #528 and #631
And #651
jonathan OK, but people shoulda been calling those through the VTABLE_foo macros? 11:01
Which still works, if I understand correctly?
bacek They should.
jonathan OK, well, anyone who wasn't IMO gets what they deserve...
bacek But we used to export functions directly.
jonathan Sure. I'd really hope people weren't calling them directly though.
For one it breaks polymorphism. 11:02
bacek Apparently lua called them.
jonathan Oh?
jonathan is a little surprised
moritz lua used pobject_lives
(like rakudo did)
bacek 2. Second realted problem - lua uses direct calls to MULTI functions. I can help with replacing such calls with "Parrot_mmd_multi_dispatch". 11:03
It's from #651.
moritz oh
bacek It was something like "Parrot_Foo_method_TYPE"
dalek kudo: a74a9cf | moritz++ | t/spectest.data:
we pass S12-methods/chaining.t
jonathan Heh, I didn't know that was possible. :-) 11:04
bacek fperrad++ refactored such calls to multi_dispatch. But my point is that we used to export them. So we need deprecation cycle :-/ 11:05
jonathan: it _was_ possible
moritz bacek: maybe ask on next #ps
bacek moritz: I can't attend to #ps... It's too early for me.
Or too late 11:06
4:30AM localtime
jonathan bacek: Suggest sending mail to the list.
So can be discussed there, or suggest it's brought up at next #ps. 11:07
bacek is not very good writer in polite english. 11:08
jonathan bacek: Just try and limit yourself to one swear word per sentnece. :-) 11:09
bacek My dictionary for technical discussions usually includes words like "crap", "bullshit" and "OMG, do you really thinks using your brains?" 11:10
jonathan looks forward to an email containing all three 11:14
szbalint I'm pretty sure linguists know this and there are probably a few of those around here, but swearing in one's native language is harder than doing so in a second or third language. When you swear in your native language, it feels stronger. 11:21
I actually found myself swearing much more in english than I do in hungarian :) 11:22
bacek rewrote mail 3 times to avoid such technical terms.
afk # family
jonathan szbalint: Yeah, I know that feeling. 11:23
You never really know a language until you know how to swear in it too. :-) 11:24
szbalint that's bullsh..I mean you're completely right :)
jonathan ;-) 11:25
11:28 Whiteknight joined 11:31 kj joined 11:57 cognominal joined 12:02 kid51 joined 12:09 kj left 12:23 Whiteknight joined
dalek rrot: r38678 | whiteknight++ | branches/gc_api:
Removing the gc_api branch, it's been merged into trunk already with no major problems. Will do other gc-related cleanup in other branches if I need
 12:24
12:40 rg joined 13:16 he joined 13:37 masak joined 14:05 whoppix joined 14:30 kid51 joined
dalek tracwiki: v68 | jkeenan++ | WikiStart 14:31
tracwiki: Currently there is no information on the original Parrot Development Roadmap page. It's all been translated into the per-release roadmap. We should have a page where we describe our overall development plan, but, until we do once again, we shouldn't have a link to an empty page on the first page of our wiki.
tracwiki: trac.parrot.org/parrot/wiki/WikiSt...ction=diff
tracwiki: v14 | cotto++ | ParrotQuotes
tracwiki: various language-related observations
tracwiki: trac.parrot.org/parrot/wiki/Parrot...ction=diff
15:08 pmichaud joined 15:13 tetragon joined 15:17 Theory joined
dalek tracwiki: v51 | jhorwitz++ | ModParrotArchitecture 15:23
tracwiki: trac.parrot.org/parrot/wiki/ModPar...ction=diff
tracwiki: v52 | jhorwitz++ | ModParrotArchitecture 15:26
tracwiki: trac.parrot.org/parrot/wiki/ModPar...ction=diff
tracwiki: v42 | jhorwitz++ | ModParrotHLLDocs 15:33
tracwiki: trac.parrot.org/parrot/wiki/ModPar...ction=diff
kudo: cbc1bf6 | moritz++ | t/spectest.data:
regress on a few tests tests that now use "augment" instead of "is also"
 15:44
15:56 eternaleye_ joined
Tene purl: tell WhiteKnight the 'issues' link in your blog post just links to the "MAKE A PASTE" page, not to an actual paste. 16:05
purl Tene: what?
Tene purl: msg WhiteKnight the 'issues' link in your blog post just links to the "MAKE A PASTE" page, not to an actual paste. 16:06
purl Message for whiteknight stored.
16:18 cognominal joined 16:29 yuvipanda joined 17:08 mikehh_ joined
dalek kudo: e6363e8 | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 383 files, 11085 passing, 1 failing

   S32-io/IO-Socket-INET.t 2 - echo server and client		 17:14
19:04 jan joined 19:14 yuvipanda left
dalek kudo: 496e1f0 | moritz++ | t/spectest.data:
a few more files for t/spectest.data
 19:25
19:31 particle1 joined 19:46 indy joined
indy Hi there! 19:47
purl i guess Hi there! is there a smart way to use inheritance of a db in dbic?
indy currently trying to get rakudo running and I am wondering whether anybody would be available to help
jonathan indy: hi 19:49
You having trouble building? 19:50
indy yep, sort of
I installed the win binary for Parrot-1.1.0
jonathan OK, the one hosted on SourceForge? 19:51
indy it's location is C:\\Parrot-1.1.0
yes
sourceforge.net/project/downloading...-1.1.0.exe
jonathan OK.
And that works OK? 19:52
indy and rakudo-2009-04.tar.gz
I think it does
when I type parrot I do get 19:53
parrot -[acEGhprtvVwy.] [-d [FLAGS]] [-D [FLAGS]][-O [level]] [-o FILE] <file>
basically I am stuck here:
perl Configure.pl --parrot-config="C:\\Parrot-1.1.0\\bin\\parrot_config.exe"
NMAKE : fatal error U1073: don't know how to make 'C:\\fperrad\\Parrot\\RELEASE_1_1 19:54
jonathan I think there are some issues with using the Parrot Win32 installer, and then doing a source build of Rakudo.
Did you try the Win32 installable version of Rakudo from Sourceforge?
indy NMAKE : fatal error U1073: don't know how to make 'C:\\fperrad\\Parrot\\RELEASE_1_1_0\\parrot.exe'
jonathan sourceforge.net/project/showfiles.p..._id=677853
indy no, I haven't 19:55
thanks for the link
jonathan Yeah. Building Rakudo from source against a Parrot installed this way currently doesn't work. We'd like it to in the future. But for now you really need to either build both from source, or just use the supplied binaries.
indy Ok.. thanks for the clarification 19:56
jonathan I do from source, 'cus I'm hacking on Rakudo a bit. But I know people have had success using the Win32 binaries too.
indy cool stuff 19:57
I'd like to get in it a bit as well 19:58
but first need to get it running and have some more time to play around ;-)
jonathan Yeah, the binaries make for an easy way in. 19:59
Once you get to wanting to work from the source, it builds easily enough under the free versions of MS VC++ compiler. And I think it's OK under MinGW too. Cygwin I fear we're a tad broken on though. 20:00
Infinoid strawberry perl (which includes mingw) has proven to be a nice all-in-one for me 20:02
it gives you perl5 and c compilers, pretty much everything parrot needs
indy Well, I am using the commercial version of VS2008, so that hopefully should do it as well I hope ;-)
Infinoid that should work too
jonathan Yeah, should be fine. 20:03
So long as nmake and cl are in your path, should work out.
indy does the rakudo installer automatically set the environment path? 20:04
jonathan Not sure, sorry. I didn't use it. 20:05
I meant about the path in relation to if you wanted to compile from source.
The installer won't need those.
Or you maybe meant if the installer sets the path of where it installs the perl6 binary, in which case I don't know. 20:06
indy no worries
just realized what I did wrong 20:07
working now :-)
jonathan yay 20:09
indy thanks a lot for your help! 20:11
20:27 knewt- joined, szbalint joined
indy hmm 20:31
just tried the simple "hello world" example and I don't get any output at all... 20:32
say "Hello" ~ " World" ~ "!";
jonathan indy: Odd. You ran it as perl6 and then did it at the interactive prompt? 20:34
indy yes, that is working 20:35
but also perl6 -c hello.pl doesn't return any output
jonathan -c ?
purl well, -c is real 0m3.827s
jonathan Oh, -c is "check syntax" - should have said the syntax was OK though.
does perl6 hello.pl work? 20:36
indy sorry, not explaining myself well enough 20:37
What I meant was that just executing perl6 is taking me to the interactive move
trying to run something like ' 20:38
print "Hello";
however returns
Unhandled exception at 0x011065a8 in perl6.exe: 0xC0000005: Access violation.
when I run
perl6 hello.pl
I do not get any output
when I try to check the syntax with
perl -c hello.pl
I also do not get any output 20:39
not even "OK"
jonathan Ouch. :-( 20:41
That doesn't sound good.
What does after it fails to give output this give:
echo %ERRORLEVEL%
indy -1073741819 20:42
purl -1073741819
jonathan ouch 20:45
What version of Windows are you running on?
20:46 donaldh joined
indy Server 2003 R2 SP2 20:46
jonathan OK, odd. 20:47
I've not heard of issues with the binaries on that platform. On the other hand, I don't know how widely used they are... 20:48
It is probably worth reporting you ran into issues with them though.
purl, rakudobug?
purl rakudobug is, like, mailto:rakudobug@perl.org
indy ok, will send a mail 20:51
jonathan Thanks. 20:52
indy Thank you
cotto bacek, ping 21:01
21:03 szbalint joined
cotto (and happy Monday) 21:06
jonathan no no no Monday means WORK.
cotto not for me :( 21:07
jonathan Ouch. :-( 21:08
cotto well, not paid
jonathan Sorry to hear that. 21:14
cotto more time for parrot 21:18
21:31 eternaleye joined 21:32 indy left
bacek good morning 21:37
cotto: pong 21:38
21:51 __darbelo joined
dalek rrot: r38679 | chromatic++ | trunk (6 files):
[IMCC] Allowed Unicode identifiers in named parameters, arguments, and return

The resulting IMCC code could use some refactoring: create a new token which represents a constant SymReg that is already a properly constructed STRING or Unicode STRING. As well, this new code needs memory profiling, as the expected lifespan of this memory isn't always clear. Still, it fixes the immediate bug.		 21:56
rrot: r38680 | bacek++ | branches/pmc_pct/compilers/pmcc/src/emitter (2 files):
Move few methods from pmc.pir to pmc.pm
22:19 gryphon joined
dalek kudo: a27bbb6 | moritz++ | build/PARROT_REVISION:
bump parrot revision to get Unicode identifiers in named arguments, chromatic++
 22:50
cotto bacek, sorry about that. I figured out my question. 22:52
22:57 Ademan joined
dalek tracwiki: v14 | coke++ | BranchDescriptions 23:11
tracwiki: trac.parrot.org/parrot/wiki/Branch...ction=diff
23:12 eternaleye joined
dalek rrot: r38681 | cotto++ | branches/pmc_pct/compilers/pmcc/pmcc.pir:
[pmcc] don't load a couple unneeded libraries
 23:15
23:18 kid51 joined 23:50 Whiteknight joined 23:54 japhb joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!