www.parrotcode.org | Last Release: 0.7.0 "Severe Macaw"
Set by moderator on 3 September 2008.
00:02 tetragon joined
wayland76 Hmm. It appears that parrot-config, once removed from MANIFEST.SKIP, wasn't listed in the other MANIFEST files. 00:03
wayland76 slaps forehead
I'll try that :)
The other alternative would be to remove parrot-config from the RPM .spec file, but since I've no idea what parrot-config does, I don't know whether this is a good idea or not 00:04
00:09 AndyA joined 00:18 TiMBuS joined 00:29 TonyC joined 00:37 Theory joined 01:08 Theory joined
cotto_work uh-oh. 666 open+new tickets 01:12
01:13 rurban_ joined 01:15 kid51 joined 01:25 particle1 joined, petdance joined 01:26 Theory joined 02:02 Zaba joined 02:25 tjh joined
tjh i'm curious if anyone has noticed the test failures on win32 in t/src/compiler.t 02:31
it seems that the /L option passed to the linker isn't recognized. The correct option should be /LIBPATH, but that doesn't work either. Not sure if this is just broken in the Microsoft linker. 02:32
kid51 tjh: Can you write up a bug ticket about that: parrotbug@parrotcode.org [BUG] in subject line. Attach output of failed test (e.g., prove -v t/src/compiler.t). Thanks. 02:56
tjh: You're not alone. It showed up in this Smolder report from early this morning: smolder.plusthree.com/app/public_pr...tails/5249 02:59
Here's the TAP output: smolder.plusthree.com/app/public_pr...m/5249/372
tjh yes, i'm the one doing those smolder reports
kid51 Oh! And do the results in report 5251 reflect a solution to the failures reported in 5249? 03:00
tjh No, that's under gcc on windows, which appears to be fine. This is something Microsoft specific it looks like. 03:01
kid51 I don't have any win32 boxes, so I can't comment specifically. IIRC, we've seen failures on t/stm/runtime.t on Win32 now and then over the past year ... but with which compiler I don't know. 03:02
tjh I was curious if anyone here had more experience with the VC++ toolchain and could point out incorrect usage. The document option doesn't seem to work (/LIBPATH) in this case.
*documented
kid51 I suspect that particle might, because he principally develops on Win32.
tjh There's a workaround I could do, but it means a change in Parrot\\ 03:03
in Parrot\\Test.pm
that does something different for the vc++ linker
kid51 One of the problems we've had is getting a steady stream of smoke/smolder reports on Win32 with specific compilers. So whatever happens with this problem, I encourage you to submit smolder reports steadily for this OS/compiler combination.
Or combinations, if the Win32/gcc reports are also yours. 03:04
tjh I've been running smolder reports for both gcc on win32 and vc++ compilers daily
kid51 Excellent! 03:05
purl EGG-see-lent!
kid51 Franļæ½ois Perrad also mainly develops on win32. So I recommend an RT so that they spot it in the morning. 03:07
tjh Ok. Will do. 03:09
03:13 particle joined 04:17 Whiteknight joined 04:32 Whiteknight joined 04:59 petdance joined 05:28 Ademan joined
dalek r30916 | cotto++ | trunk: 05:56
: [pmc] fix a copy/paste error in function docs
diff: www.parrotvm.org/svn/parrot/revision?rev=30916
r30917 | cotto++ | trunk: 06:02
: [dynpmc] typo fix
diff: www.parrotvm.org/svn/parrot/revision?rev=30917
06:16 uniejo joined 06:56 ian joined 07:18 barney joined 07:21 tetragon joined 07:25 Zaba joined 07:26 XiaoA joined 07:35 Ademan joined 07:36 tetragon joined 07:41 apannu__ joined 07:47 Zaba_ joined 07:51 iblechbot joined 07:57 tetragon joined 08:11 Ademan joined 08:18 mberends joined 08:25 Zaba joined 08:26 megamic joined 08:57 masak joined 09:12 rurban_ joined
dalek r30918 | julianalbo++ | trunk: 09:14
: add paco to CREDITS, paco++
diff: www.parrotvm.org/svn/parrot/revision?rev=30918
09:19 cosimo joined 09:55 bacek joined 10:32 megamic joined 10:39 jan joined
dalek r30919 | kjs++ | trunk: 10:43
: [pirc/new] updates and refactoring.
: + implement .HLL_map
: + implement .HLL
: + implement .loadlib
: + refactoring of files and cleanup.
diff: www.parrotvm.org/svn/parrot/revision?rev=30919
r30920 | kjs++ | trunk: 10:52
: [pirc/new] update README.pod
diff: www.parrotvm.org/svn/parrot/revision?rev=30920
10:57 megamic joined 11:19 Debolaz joined 11:52 iblechbot joined 12:05 pako joined 12:10 pako left 12:21 tetragon joined 13:30 gryphon joined
dalek r30921 | allison++ | pdd27mmd: 13:40
: [pdd27mmd] Implement 'add_multi' and 'find_multi' opcodes for new merged MMD
: system.
diff: www.parrotvm.org/svn/parrot/revision?rev=30921
13:43 rdice joined
dalek r30922 | allison++ | pdd27mmd: 13:48
: [pdd27mmd] Vtable function 'divide' in scalar was masking multiple dispatch,
: should be a multi instead.
diff: www.parrotvm.org/svn/parrot/revision?rev=30922
r30923 | allison++ | pdd27mmd: 13:52
: [pdd27mmd] Remove unneeded .includes from the multiple dispatch tests.
diff: www.parrotvm.org/svn/parrot/revision?rev=30923
13:53 Zaba_ joined 13:59 wolverian joined 14:09 DietCoke_ joined, DietCoke_ left
particle kj++ 14:10
14:13 Zaba joined, davidfetter joined 14:16 bacek joined
bacek g'night 14:16
purl: hi
purl niihau, bacek.
bacek moritz: around?
moritz bacek: yes (partially) 14:17
dalek r30924 | allison++ | pdd27mmd:
: [pdd27mmd] Convert 'cmp', 'cmp_num', and 'cmp_str' opcodes over to new merged
: multiple dispatch.
diff: www.parrotvm.org/svn/parrot/revision?rev=30924
bacek moritz: I've fixed problem for slurpy args. 14:18
unfortunately, I discovered another bug, that causes few spectest failures... 14:19
moritz :(
pmichaud what's the fix/bug ?
NotFound bacek: related to RT#54800 ? 14:20
bacek rt.perl.org/rt3/Ticket/Display.html?id=58278
fix
#58718 is bug.
NotFound: no. It's in signature handling in rakudo. 14:21
pmichaud: fix for 'method signature' in actions.pm... And this method is 300 lines long... It's kinda weird 14:23
pmichaud bacek: I agree that signature handling in actions.pm needs major refactoring
I don't see how #58278 affects slurpy args, though. 14:24
moritz question is, should that wait for jonathan's mmd branch?
s/branch/work/
particle + ' %r."infix:="(%0)',
this looks broken
ah, i guess not 14:25
it just displays funny in html, but looks ok in the patch
NotFound Can you take a look at #54800 and see if the last fix in optional arguments solve it? 14:26
bacek pmichaud: "sub xelems(*@args) { @args.elems }" 14:27
NotFound you = any rakudo guy
bacek going to sleep. very soon. 14:29
pmichaud iiuc, slurpy args shouldn't be using Perl6Scalar at all.
and we definitely shouldn't be calling .item 14:30
bacek pmichaud: oh. it's was in first version of patch.
and it's definetly wrong. 14:31
pmichaud can't keep up with four tickets and patches at once. :-(
bacek rt.perl.org/rt3/Ticket/Attachment/4...ture.patch
it's actually very trivial 14:32
looks like my comment was added twice to ticket...
pmichaud I don't think that the %r = clone %0 can be correct 14:33
of course, I don't know that any of this is really "correct"
bacek why? ('infix:=' doesn't create copy AFAIU)
pmichaud it currently does. 14:34
bacek perl6: my @a=(1..2); my $b=@a; $b[0]=3; say @a
rakudo: my @a=(1..2); my $b=@a; $b[0]=3; say @a
polyglotbot OUTPUT[32ā¤]
bacek pmichaud: no 14:35
or it's another bug
NotFound rakudo: sub foo($x?, :$y = 2){ say "$x"~"|"~"$y"}; foo(:y(3));
polyglotbot OUTPUT[Use of uninitialized valueā¤|3ā¤]
pmichaud perl6: my @a = (1,2); my $b = @a; $b[0]=3; say @a
polyglotbot OUTPUT[32ā¤]
NotFound This code is from RT#54800 . Is this the result expected?
pmichaud NotFound: yes, that's the expected result. 14:36
bacek NotFound: yes
NotFound pmichaud: I'll add a comment to the ticket about that part, then.
pmichaud NotFound: thanks.
bacek: okay, I'm not sure what's happening with infix:=
in which case, I'd rather not apply a patch that causes spectest_regression to fail 14:37
especially since I think it all needs refactoring anyway
14:37 Limbic_Region joined
pmichaud anyway, simply having %r = clone %0 is not sufficient, because it needs to be in a container. 14:38
bacek pmichaud: ok. Triggered bug described in #58718. So we can return to this one later
pmichaud however, that should not always be a Perl6Scalar -- e.g., if the parameter is '@a' or '%b' then it doesn't belong in a Perl6Scalar 14:39
and I think the code is not handling that case at all at the moment. 14:40
bacek pmichaud: yes. Side question: how I can create unique labels with PCT (similar to %r)?
pmichaud bacek: there's not a way to do it.
bacek: what do you need a unique label for?
bacek pmichaud: for checking actual type of argument and creating proper container ;)
pmichaud that's to be done in the ast, not as part of the PIR 14:41
in this case the container is determined by the parameter name, not by the argument
bacek ah. ok
pmichaud in general, inline PIR is something we should be avoiding. 14:42
bacek pmichaud: last thing for today according "Wrong creating of Range" - Lists are created wrongly.
pmichaud: (inline PIR) agree. I want to write perl6 in perl6, not PIR
pmichaud how do you mean "Lists are created wrongly?" (more) 14:43
on the other hand, I expect Lists to change substantially when we have lazy lists.
so I'm not too worried about what the current code does. 14:44
bacek $P0 = get_hll_global 'List'; $P1 = $P0."new"(); say $P1
elements() not implemented in class 'Undef' 14:45
dalek r30925 | allison++ | pdd27mmd:
: [pdd27mmd] Install test 'cmp' multi with the correct name.
diff: www.parrotvm.org/svn/parrot/revision?rev=30925
pmichaud bacek: try $P1 = "list"() instead.
bacek pmichaud: inconsistent... 14:46
pmichaud bacek: why?
dalek r30926 | allison++ | pdd27mmd:
: [pdd27mmd] Remove reference to MMD_* constant that doesn't exist anymore.
diff: www.parrotvm.org/svn/parrot/revision?rev=30926
bacek pmichaud: because for copying value I should have to now, what it is.
pmichaud bacek: you lost me there. 14:47
bacek Instead of something like $P0 = arg."WHAT"(); $P1 = get_hll_global $P0; $P2 = $P1."new"()
?
pmichaud oh, if you want to do that, it's just 14:48
$P0 = arg.'WHAT'(); $P2 = $P0.'new'()
but if we're talking about arguments, the correct semantics is indeed "clone"
14:49 Zaba_ joined
bacek ($P0.new) and it will fail for current List 14:49
pmichaud if you're talking about making a copy of an argument that is being passed "is copy", then the clone opcode is the one to use.
if you're talking about making a List in general, then "list"() should work for now. 14:50
If List.new isn't working at the moment, I'm not completely surprised and I'll worry about it when we implement lazy lists.
bacek pmichaud: ok ;) Lets worry later. 14:51
btw, looks like "#49173: [TODO] Allow source written in perl6" can be resurrected now?
pmichaud pretty close.
I still have a couple of pieces to put together -- mainly some examples. 14:52
bacek Need help? I can try to update patch from ticket to current codebase. Or it is not required? 14:53
pmichaud I'm likely to reject that patch. 14:54
in particular, I don't like the references to "BOOTSTRAP"
bacek but how you'll parse perl6 code without (subset of) perl6 compiler? 14:57
pmichaud I prefer STAGE1, STAGE2, etc. 14:58
I'm not saying we won't use Rakudo to parse the library, I just think that "bootstrap" is overloaded and incorrect in this case.
bacek got the point. I can do it if you want 14:59
ok. 1am...
must sleep
purl $bacek->sleep(8 * 3600);
pmichaud I'll keep it in mind -- I'm still deciding how I want to structure Prelude
bacek s/8/6/...
pmichaud: ok. 15:00
bb
dalek r30927 | pmichaud++ | trunk: 15:12
: [rakudo]: spectest-progress.csv update: 158 files, 3288 passing tests
diff: www.parrotvm.org/svn/parrot/revision?rev=30927
15:22 peepsalot joined 15:25 rdice joined
rurban btw: cannot I use perl6 to generate perl6 to generate pir instead of using pir to generate pir to generate pir? or nqp at least? [jvm] 15:26
pmichaud rurban: yes, at some point (perhaps not yet) you might be able to rebuild the tools in perl 6 15:28
actually, I'll rephrase.
rurban I rather want to wait then. nqp?
pmichaud at some point you *will* be able to rebuild those tools in perl 6. I don't know when that will be possible.
NotFound rurban: just remember to add dependencies in makefiles to any too you use.
pmichaud NQP could probably do it, if you can live with nqp's limited runtime
NotFound s/too/tool
pmichaud chances are that you can. 15:29
rurban NQP is still thousand times better than PIR
NotFound You can even draw with nqp, see examples/nci/xlibtest.nqp ;) 15:30
Someone see a README file in compilers/pirc/new/ ? 15:38
moritz sees neither a Makefile nor a README 15:39
jonathan will probably Rakudo hack tomorrow 15:42
Just so it's known. :-)
moritz YaY 15:43
pmichaud jonathan: works for me -- I should be around. 15:45
jonathan Cool. 15:51
Limbic_Region salutations all
moritz heya Limbic_Region ;) 15:52
rurban NotFound: shouldn't you add a Makefile to examples/nci so that Xlib.pbc gets generated? 15:55
and copied to runtime/library/
NotFound rurban: want to hear some comment at #ps first
15:56 iblechbot joined
NotFound And I suppose that examples/ is not the correct place for a thing that is going to be placed in runtime 16:00
rurban indeed.
I would make Mysql and Xlib regular libs, such as OpenGL and Pg 16:01
NotFound runtime/parrot/library/ , maybe 16:02
rurban sure
NotFound rurban: yeah, but both need a lot of work to be really useful. ATM they are just nice demos. 16:03
rurban Without running XServer they still crash on cygwin 16:04
pmichaud jonathan: out of curiosity, should we be looking at 'morph' as a way of doing "rebless" to get subs into another type? 16:05
NotFound Yes, need some more error checking and throwing.
rurban Otherwise all samples work fine on cygwin. p6, pir and nqp
NotFound Good. 16:06
rurban Need to check the mingw xserver dll name though.
16:07 Zaba joined
NotFound Must add redrawing to nqp version to make it comparable with the pir and p6 ones. Some volunteer? 16:10
jonathan pmichaud: I'd not really thought about doing it that way. 16:12
moritz NotFound: I can take a look 16:13
pmichaud jonathan: it's just a thought. We could have a loadinit sub that morphs the Parrot subs into the different classes, perhaps
moritz NotFound: if there are no results by tomorrow, consider my volunteerness dropped ;)
pmichaud i.e., to distinguish Sub, Method, Regex, etc.
jonathan Thing about morph is that the Parrot Sub has to know how to morph itself to the other type.
pmichaud sure, but we only have to define it once.
jonathan Aye.
NotFound moritz: note taken
pmichaud (in Parrot .Sub)
I would think that fixing .Sub to morph wouldn't be all that difficult. 16:14
anyway, I'll let you mull it over a bit :-)
that might be nicer than having to check properties on Parrot subs
moritz NotFound: how do I run it? ;-) 16:15
it complains load_bytecode" couldn't find file 'Xlib.pbc
and ../../parrot Xlib.pir -o Xlib.pbc fails
pmichaud (-o Xlib.pbc might need to go before Xlib.pir) 16:16
moritz d'oh, -o needs to go first
rurban ./parrot -o runtime/parrot/library/Xlib.pbc --output-pbc examples/nci/Xlib.pir
NotFound moritz: please copy the note on .pir version, while you are at it. 16:17
rurban ./perl6 examples/nci/xlibtest.p6
moritz NotFound: will do
rurban ../../parrot ../../compilers/nqp/nqp.pbc xlibtest.nqp
jonathan pmichaud: Oh, I don't want to go the propreties route, that's for sure. 16:18
dalek r30928 | allison++ | pdd27mmd: 16:19
: [pdd27mmd] Change 'bitwise_xor' and 'bitwise_xor_int' vtable functions to
: multis, so they don't mask default multiple dispatch.
diff: www.parrotvm.org/svn/parrot/revision?rev=30928
r30929 | allison++ | pdd27mmd: 16:21
: [pdd27mmd] Fix test multi for 'bxor' so it matches the signature of the
: 'bitwise_xor' multi and vtable function.
diff: www.parrotvm.org/svn/parrot/revision?rev=30929
r30930 | rurban++ | cygwin070patches: 16:25
: [jvm] remove SRM docs, add some KAFFE structs
diff: www.parrotvm.org/svn/parrot/revision?rev=30930
r30931 | rurban++ | cygwin070patches: 16:26
: [SRM] Generalize docs, more java friendly
diff: www.parrotvm.org/svn/parrot/revision?rev=30931
16:28 sjansen joined
moritz NotFound: done ;) 16:44
dalek r30932 | moritz++ | trunk:
: [examples/nci] add re-drawing to xlibtest.nqp
diff: www.parrotvm.org/svn/parrot/revision?rev=30932
moritz NotFound: it wastes a bit of memory, but is actually quite fast (subjectively)
NotFound moritz: great! 16:46
Now we can start blaming rakudo guys until they reach the same speed in perl6 }:) 16:47
moritz not amused
NotFound moritz: don't worry, I've already punched with pir 16:48
rurban it is not redrawn while still holding down the resize left-mouse 16:50
moritz just copied the redrawing condition from the rakudo example
NotFound rurban: need more X protocol support to do redrawing while resizing. 16:51
rurban I'm just comparing it to the clisp Xlib samples... 16:52
NotFound X is a complex beast
moritz ... and should be replaced by something better 16:53
rurban How can I test pirc/new? KJS said on #ps that cygwin does not work 16:59
NotFound First you need to buit it... but no Makefile and no README 17:05
moritz I gave kjs shell access to a Debian box so that he can also try on his own 17:06
rurban Iguess I can just replace src with new in the Makefile, but...
When is 18:30 UTC. in 15 min or in 1:15? 17:07
dalek r30933 | rurban++ | cygwin070patches:
: [WMLScript] fix t/runtime.t test 2/2
diff: www.parrotvm.org/svn/parrot/revision?rev=30933
cotto_work rurban: in 1:15
particle current gmt time is 17:07
i use foxclocks addin for firefox 17:08
rurban Ok, so I'll gonna eat something before.
particle i'm going to the airport, i may make it back in time 17:09
NotFound date --utc
particle purl: what time is it?
purl It's nearly ten past five in the afternoon, particle.
particle that's gmt
paco good hint, NotFound 17:10
rurban purl: how late in GMT?
purl rurban: no idea
rurban purl: how late?
purl i don't know, rurban
NotFound purl: utc? 17:12
purl utc is date -u or the timezone you use if you don't have a 1mm penis
NotFound Interesting
purl: zulu time? 17:13
purl i think zulu time is GMT
17:14 rurban_ joined
NotFound paco: forgot to tell you: the skipped tests in your S/390 build are fine, and the -0 errors are not importante ATM 17:20
paco NotFound: yes, I think so
17:37 iblechbot joined 17:47 barney joined 18:03 chromatic joined
dalek r30934 | moritz++ | trunk: 18:10
: [caged] make codetests happier in compilers/pirc/
diff: www.parrotvm.org/svn/parrot/revision?rev=30934
moritz where do I report smolder bugs? 18:16
18:17 Ivatar joined
particle perhaps michael peters left info on his messages to p2 18:17
moritz I'll sent to p2, and cc him 18:19
18:20 jhorwitz joined
pmichaud #ps in 8 18:22
18:25 LimbicRegion joined 18:26 ian_ joined 18:35 rdice joined 18:37 Zaba_ joined 18:43 sjansen joined 18:44 s1n joined 18:53 jan joined 18:55 peepsalot joined 19:02 hachi_ joined 19:11 Infinoid joined 19:25 hachi_ joined 19:29 Infinoid joined
obra yo 19:52
er. apologies. mischan 19:53
particle oy 19:54
19:55 sjansen joined 19:57 sjansen joined
chromatic just made all of Parrot and PGE and TGE build in the MMD branch. 20:00
davidfetter chromatic++
chromatic Tests don't all pass, but there was clearly an error there.
particle how close is the branch to merging? 20:01
chromatic 21 core tests fail. 20:02
Pheme builds, but fails tests. Perl 6 doesn't build.
Loose guess: there are a couple of other bugs like the one I diagnosed.
particle that's encouraging 20:06
NotFound Nice and simple bug: load_bytecode dies with a null string argument.
moritz shouldn't it magically load the right bytecode? ;-) 20:07
NotFound moritz: ETOOMANYMAGIC
I think change the signature of Parrot_load_bytecode to allow null, and check for it, is the way to go. 20:08
rurban I think dying with NULL is good 20:09
To catch programming errors 20:10
NotFound rurban: is not good to let a VM machine die. 20:13
rurban ok, die == exception 20:14
NotFound The program can be killed, not the machine.
20:29 Andy joined 20:40 Whiteknight joined
paco interesting reading : gcc.fyxm.net/summit/2003/Porting%20...%20bit.pdf 20:41
dalek r30935 | julianalbo++ | trunk: 20:42
: throws on load_bytecode with NULL string
diff: www.parrotvm.org/svn/parrot/revision?rev=30935
rurban mk_language_shell.pl is full featured now 20:56
dalek r30936 | rurban++ | cygwin070patches:
: extend the language template by pdd30_install suggestions, and a sample pmc and ops
diff: www.parrotvm.org/svn/parrot/revision?rev=30936
rurban even pmc's and ops are in
21:07 Ademan joined 21:12 LimbicRegion joined
dalek r30937 | allison++ | pdd27mmd: 21:39
: [pdd27mmd] Deprecation note for future reference.
diff: www.parrotvm.org/svn/parrot/revision?rev=30937
r30938 | allison++ | pdd27mmd: 21:41
: [pdd27mmd] Update several multiple dispatch tests to use current semantics and
: storage location for multis.
diff: www.parrotvm.org/svn/parrot/revision?rev=30938
22:00 particle joined
dalek r30939 | chromatic++ | pdd27mmd: 22:02
: [lib] Made PMC emitter trust the declared return value of a MULTI rather than
: applying incomplete heuristics; this allows PGE and TGE to build and avoids the
: error where Parrot tries to convert a returned INTVAL into a PMC pointer (see
: "cmp" MULTIs). Some tests still fail, but all of Parrot builds.
diff: www.parrotvm.org/svn/parrot/revision?rev=30939
r30940 | allison++ | pdd27mmd: 22:29
: [pdd27mmd] Fix coding standards test for cuddled elses.
diff: www.parrotvm.org/svn/parrot/revision?rev=30940
r30941 | allison++ | pdd27mmd: 22:35
: [pdd27mmd] Fix failing coding standards test for line length.
diff: www.parrotvm.org/svn/parrot/revision?rev=30941
r30942 | allison++ | pdd27mmd: 22:39
: [pdd27mmd] Fix failing coding standards tests for trailing spaces.
diff: www.parrotvm.org/svn/parrot/revision?rev=30942
r30943 | allison++ | pdd27mmd: 22:41
: [pdd27mmd] Fix failing file metadata tests.
diff: www.parrotvm.org/svn/parrot/revision?rev=30943
22:51 Ademan joined 23:05 davidfetter joined 23:10 Zaba joined 23:19 nopaste joined 23:20 pjcj joined 23:31 chromatic joined 23:32 cognominal joined 23:33 tetragon joined 23:44 cognominal joined
dalek r30944 | Whiteknight++ | pdd27mmd: 23:50
: [pdd27mmd] change over two more old-style function calls to use the new MMD functions
diff: www.parrotvm.org/svn/parrot/revision?rev=30944
23:52 wknight8111 joined
wknight8111 I've missed #ps two weeks in a row now 23:54
chromatic We're taking away your birthday. Also Flag Day.
wknight8111 work is really starting to encroach on the important things in life
And I spent all night last night installing a new water heater, so I couldn't do any work I wanted
how are you chromatic? I feel like we never talk anymore :) 23:58

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!