pugscode.org/ planetsix.perl.org/ | nopaste: sial.org/pbot/perl6 | evalbot: perl6: say 3; (or rakudo:, pugs:, elf:, etc) | irclog: irc.pugscode.org/ | ~340 days 'til Xmas
Set by mncharity on 16 January 2009.
mberends @tell masak re: ./perl6 versus ../../parrot perl6.pbc, RT#62376 illustrates a similar problem and can be demonstrated with r35708 on Debian x86-32. rt.perl.org/rt3/Public/Bug/Display.html?id=62376 00:10
lambdabot Consider it noted.
diakopter ruoso: ri invokes m0ld with the following command line parameter: ",suffix='.c',src_suffix='.ri'" (I added double quotes b/c they're immaterial on windows) 01:44
HyP hy 03:09
i search a help for code in Perl 03:10
araujo ? 03:11
HyP Problem is after log and connect , i've not view of terminal host
sorry i'm frenche and my english is incorrect 03:12
araujo mm... 03:13
this channel is about perl 6
araujo you probably want help for perl 5? 03:13
gwern no pugs channel? 04:16
gwern well, if anyone can tell why HsJudy was removed from pugs I'd appreciate it 04:23
pmurias diakopter: those parameters are incorrect, how did you tweak your SConstruct you may have broken it 11:13
masak std: my @a = $_ for 1,2,3 11:16
lambdabot masak: You have 1 new message. '/msg lambdabot @messages' to read it.
p6eval std 24936: OUTPUT«00:06 105m␤»
masak @messages
lambdabot mberends said 11h 6m 21s ago: re: ./perl6 versus ../../parrot perl6.pbc, RT#62376 illustrates a similar problem and can be demonstrated with r35708 on Debian x86-32. rt.perl.org/rt3/Public/
Bug/Display.html?id=62376
masak @clear messages
lambdabot Messages cleared.
masak rakudo: my @a = $_ for 1,2,3; say @a.perl 11:18
p6eval rakudo 35710: OUTPUT«Lexical '@a' not found␤current instr.: '_block14' pc 162 (EVAL_17:73)␤»
masak it surprises me somewhat that @a is not declared in the above case. but I don't know what the spec says about such a case. 11:19
ruoso diakopter, ping 12:44
masak szabgab++ # C-style for loop RT ticket 13:40
lambdabot masak: You have 1 new message. '/msg lambdabot @messages' to read it.
masak @messages
lambdabot mberends said 13h 30m 18s ago: re: ./perl6 versus ../../parrot perl6.pbc, RT#62376 illustrates a similar problem and can be demonstrated with r35708 on Debian x86-32. rt.perl.org/rt3/Public/
Bug/Display.html?id=62376
masak lambdabot: right, I got you the first time. :/
@clear
lambdabot Messages cleared.
jnthn masak: You and bots just don't get on, huh? :-)
masak jnthn: I think I have high expectations. 13:41
I want the future, now.
gist.github.com/48659 13:44
jnthn High expectations indeed...at least, purl makes them look high. ;-) 13:46
masak purl wilfully violates them.
that's her main flaw.
jnthn Aye, but somehow, when she's gone, she's missed. 13:47
masak right, right. she has some good traits. 13:48
jnthn tries to get more done on his slides, but sorta wants to be coding too...
Think I figured out how I'm going to do junction auto-threading while in the shower this morning.
masak jnthn: I note that you and pmichaud still haven't really made a dent in the pile of bug tickets I left you over the holidays. ;) 13:49
jnthn Yes. ;-) 13:50
On my part, it's because I've got a good chunk of time at the moment without much distraction from anything else. 13:51
And so I'm taking advantage of it while I have it to get the Big Things in my Hague Grant underway.
masak I understand. 13:52
jnthn Also, waiting to see if my Rakudo day will be renewed.
masak renewed?
jnthn Aye, it was agreed until end of 2008.
masak oh.
jnthn They encouraged me to submit a proposal for continuing it in 2009 though, and I've done so.
So just waiting to hear on that. :-) 13:53
masak crosses fingers
jnthn For now though, I've got my hands plenty full on the Hague Grant tasks.
masak ok, no worries. 13:54
I don't have any showstopping bugs right now.
jnthn OK
I do have some I want to address, plus I need to go through RT and close up any that I have fixed already. 13:55
masak just general grumpiness over arbritary segfaults. you probably saw the discussion with chromatic++ yesterday.
jnthn *nod*
They are just plain hard to track down.
And hard to re-produce too.
masak except the ones that occur every time :/ 13:56
jnthn Especially on different platforms.
masak like the double free.
jnthn That one is scary. I don't doubt that if it was an easyish or even a mediumish difficulty fix, chromatic++ wouldn't have done it by now. 13:57
erm, would have done it
When chromatic describes it as hard, I'm worried.
masak :)
jnthn If you get any that are _not_ the double free though, a backtrace is helpful. 13:58
jnthn (We have quite enough backtraces of that one in RT. :-|) 13:59
masak I only have the double free and irregular ones right now. and chromatic tells me that the irregular ones can't exist. :/
jnthn I think the point was more that, the Parrot bug that gives rise to them shouldn't exist in the executable and not exist when running the PBC. 14:00
*but* whether the bug manifests may vary.
Which if it's GC related, is very possible. 14:01
masak I realize that. and he managed to convince me that apart from the double free, ./perl6 is not necessarily worse than parrot perl6.pbc.
just crashes in different places.
jnthn Right.
Because of differing memory allocation patterns. 14:02
masak which means that the consistent lack of crashes in the former since I switched is a coincidence. which I somewhat reluctantly accept.
s/former/latter/
jnthn That analysis is consistent with what I know too. 14:03
The trouble with the segfaults that arise out of GC issues is...
jnthn They're hard to track down without a small program. 14:03
masak is this due to the nature of Parrot or something? I've never managed to make JVM segfault. 14:04
jnthn And a small program in Perl 6 doesn't help so much. Even if you can reproduce it with running the output of --target=pir it's still pulling in perl6.pbc that does a lot of stuff.
It's a stability thing.
And maturity etc.
And I suspect also in part, complexity. 14:05
masak mm
jnthn I've hunted down GC bugs before. They are highly non-trivial. :-( 14:06
masak oh! and I've sort of decided that after the Rakudo leaving-the-nest thing, November will start targetting Rakudo and Parrot versions, not HEADs.
jnthn Last time I patched one it was 5-6 hours of analysis. 14:07
masak I got a bit fed up with three testsuite marathons in a row.
not anyone-in-particular's fault, of course, just a fallout of our current development model.
jnthn That's probably a sensible stratergy for November.
If you actually want to get development done on it rather than just bug hunting. 14:08
masak aye. I'm at the point where I'd like to guarantee stability.
jnthn I think the rvar branch shook a lot of things up.
It was needed, for sure.
masak I'm still not sure what November will become. but if it's to become a good wiki, it needs stability.
jnthn: the thing I'm the least pleased about is class-def-on-line-1. 14:09
jnthn But I don't doubt some things that accidentally and untestedly worked before it, were broken.
masak I realise that's what the spec says, but it _feels_ a bit unintuitive.
jnthn I've always been tending to write classes in blocks.
masak at least when we have like five different modules which worked before with a couple of harmless use statements before the class decl. 14:10
masak I buy the match-against-grammar deprecation thing. that was our fault in not heeding the many warnings about deprecation. :) 14:10
but I don't really see the harm in having use statements above 'class Foo;' 14:11
@tell moritz_ s/hoover/hover/ in perlgeek.de/blog-en/perl-6/parse-tree.writeback -- otherwise, nice post. 14:13
lambdabot Consider it noted.
jnthn Not sure, what the design rationale is.
But it's been Perl 6 spec for a long while.
masak I guess I didn't interpret 'first in file' so... literally myself. 14:14
jnthn Now and then, the spec does actually really mean what it says. ;-) 14:15
masak haha
perl6: my @a = 1,2,3; @a >>+<<= [10,20,30]; say @a 14:17
p6eval pugs: OUTPUT«*** Unsafe function '=' called under safe mode␤ at /tmp/Z6j8KKzUTU line 1, column 16-36␤»
..rakudo 35717: OUTPUT«Method 'list' not found for invocant of class 'Iterator'␤current instr.: '!HYPEROP' pc 14949 (src/builtins/assign.pir:273)␤»
..elf 24936: OUTPUT«Undefined subroutine &GLOBAL::prefix__61 called at (eval 131) line 4.␤ at ./elf_f line 3861␤»
masak both Pugs and Rakudo seem to parse that as prefix:<=>
jnthn >>+=<<
But I don't think we implemnet it yet. 14:18
masak oh. so putting the = outside is wrong? where does it say that?
jnthn I think the implementations are correct to parse that as prefix:=
jnthn S03 14:19
jnthn @array 􏿽xBB+=􏿽xBB 42; # add 42 to each element 14:19
lambdabot I'll crush ye barnacles!
jnthn ...well, that was a purlesque comment from lambdabot... 14:20
masak :)
jnthn: that tells me that >>+=<< is right, which I can accept. 14:21
but where does it say that >>+<<= is wrong?
or, at least, NWIM?
jnthn It'd never occurred to me you'd write it that way. Hmm. 14:22
std: @a >>+<<= [10,20,30];
p6eval std 24936: OUTPUT«00:08 144m␤»
masak but that doesn't tell us if STD also sees it as prefix:<=> :) 14:23
jnthn Yeah, that's what I was thinking after I wrote it... :(
masak oh well. back to more worldly things. 14:24
jnthn I see what you're expecting it to parse as now though.
(at first I didn't...)
masak aye, and it seems to me that your suggestion is slightly different, although I haven't yet come up with an example that hightlights said difference. 14:25
jnthn They're certainly subtle, afaict. 14:26
masak hm, November p6w doesn't build at all today. investigating... 14:28
jnthn discovers a multi dispatch bug while testing his slide examples :-| 14:33
masak code++ 14:34
jnthn I should write more of it...
masak I know the feeling. 14:36
masak oh! our Test.pm doesn't work, for some reason. 14:41
and neither does the one in languages/perl6. 14:42
masak wonders if that's worthy of a rakudobug 14:43
jnthn :-S
jnthn does svn up do find out what on earth is going on 14:44
masak ./perl6 -e 'use Test' is enough.
...to trigger the problem.
No method named 'pass' to remove in class 'POST;Compiler'.
jnthn Wow.
masak jnthn: sorry to distract you from whatever you were doing.
jnthn sees some patched Parrot core stuff 14:45
Just slide writing, but I was going to move on to hacking on Rakudo soon...
masak that's two Parrot fixes in a row affecting Rakudo in a disastrous way.
looking forward to that 1.0 stable API. :P 14:46
jnthn masak: Reproduced it here.
masak: Did you try a realclean?
masak I _always_ do realclean when compiling.
jnthn OK, so I probably don't need to also do it then. 14:47
masak you could, but the error wouldn't go away.
jnthn Right.
erm. 14:48
How on *earth* does POST::Compiler come into this...
perl6 -e 'multi foo() { }'
Also does it.
masak :/ 14:49
jnthn Well done, somebody...
masak good thing there's not a release coming up or something. :/
jnthn looks through what's been commited of late for suspects
...to revert. :-|
masak jnthn++
masak goes back to learning Chinese words 14:50
jnthn Whiteknight: ping
ruoso hi pmurias 14:59
jnthn And the winner of the "who did the Parrot change that broke Rakudo" is... 15:00
masak (drum roll)
jnthn 35697
(by Whiteknight. I feared so, from the error. :-() 15:01
Even more fun, there are two additional patches on top of the files as of how they were in that patch.
jnthn hates just reverting people's patches
Whiteknight jnthn: pong!
masak sounds like this might take a while. 15:02
Whiteknight: we were just talking about you. :)
Whiteknight yeah, I see that
jnthn Whiteknight: your get_class changes have caused epic breakage :-(
Whiteknight that get_class stuff is insufferable
jnthn Whiteknight: Your later patches succeed in making it fail in a different way though. ;-)
masak I got an angry mob of November hackers outside my window. what should I tell them?
Whiteknight tell them I'll have a fix committed in about 5 minutes 15:03
jnthn Whiteknight++ # thanks
masak they seem pleased.
jnthn (Would much rather have you fix it, than hit the undo button. :-))
Whiteknight what we need right now is all the use cases of get_class that Rakudo is relying on added to the test suite
because I obviously don't know them all, and this is an area that is ripe for changes/updates/improvements
jnthn *nod* 15:04
Whiteknight So if somebody can throw together some tests for that, I'll revert my recent changes of it
Whiteknight where can I go to reproduce this problem? are the failures evident in the Rakudo test suite anywhere? 15:05
I want to make sure my fix actually fixes it 15:06
masak Whiteknight: ./perl6 -e 'use Test'
jnthn Whiteknight: Thing is that your patch touches get_class and stuff that it calls 15:06
It's possible the breakage is because of other things relying on the behavior elsewhere, not specifically get_class. 15:07
Though I'm kinda guessing here.
Whiteknight I'm decently certain that it's only coming through get_class, but I'll do some more research on it all 15:09
I think I can back out most of the changes/fixes and still leave RSA working the way it's supposed to
jnthn OK 15:10
jnthn I'll try and have a look through the Rakudo source and the tests and see what we're doing that is untested in Parrot core. 15:10
Whiteknight thanks 15:11
jnthn (with regard to get_class)
Whiteknight yeah
Whiteknight jnthn, on an unrelated note: what term do you use to refer to the postfix if or unless syntax? 15:12
pmurias ruoso: hi
masak rakudo: class A does B {}; role B {} 15:13
p6eval rakudo 35718: OUTPUT«Null PMC access in find_method()␤current instr.: '!meta_trait' pc 17937 (src/builtins/guts.pir:559)␤»
masak submits rakudobug
Whiteknight masak, hold off on the rakudo bug for a minute 15:14
I've got a fix coming down the parrot-pipeline that might resolve that issue
masak holds off for a minute
Whiteknight thanks
masak meanwhile, what's the semantics of does-ing a role which is declared later in the file? 15:15
jnthn Whiteknight: That's a separate issue to the one you're fixing.
masak: The semantics is, it gives an error.
masak (Илья's question on p6u)
jnthn: ok. I'll reply that to him.
jnthn I explained this to someone else yesterday...
masak I thought it would.
ruoso pmurias, i'm considering implementing named args
jnthn masak: But file, we need a better error.
ruoso pmurias, what do you think of implementing "use"? 15:16
masak no, I'm holding off for a minute.
jnthn In fact, we need a _compile time_ error.
pmurias ruoso: i'm planning to implement use Foo:from<perl5> now 15:17
ruoso cool
masak pmurias++ 15:19
Whiteknight jnthn, okay
masak pmurias: that is the one feature that might make me a smop user. 15:20
ruoso pmurias, I'm thinking about splitting AST::* into separated files as well 15:21
what do you think?
ruoso masak, you realize that only syntax is missing to do that? see mildew/example 15:24
pmurias ruoso: the splitting of VAST seems too have worked nicely 15:25
masak ruoso: yes, I understand. actually, what I'm really after is being able to write whole Perl 6 programs that use Perl 5 CPAN modules.
pmurias ruoso: we have to consider if we want to always create a p5 interpreter or only when it's actually needed 15:34
ruoso pmurias, I was thinking we could pre-declare a variable... 15:35
and kinda autovivify it when someone try to use it 15:36
but while we don't support autovivification
ruoso we could just initialize it always 15:36
tigreton hi 15:37
pmurias tigreton: hi 15:38
ruoso: $P5Interpreter?
ruoso yeah... maybe for now
Whiteknight okay, so maybe my fix will take more then 5 minutes, I have a lot of tests to run 15:39
pmurias $+P5Interpreter maybe
i'll fake $+foo as $foo for now
perl6: $+OUT.print("hi\n"); 15:40
p6eval elf 24936: OUTPUT«Can't call method "Str" on an undefined value at ./elf_f line 649.␤ at ./elf_f line 3861␤» 15:40
..pugs: OUTPUT«*** Can't modify constant item: VUndef␤ at /tmp/KzAU8pkTN6 line 1, column 1-20␤ /tmp/KzAU8pkTN6 line 1, column 1-20␤»
..rakudo 35718: OUTPUT«Scope not found for PAST::Var '$+OUT' in ␤current instr.: 'parrot;PCT;HLLCompiler;panic' pc 146 (src/PCT/HLLCompiler.pir:102)␤»
ruoso pmurias, we don't have a dynamic prelude yet 15:41
nor dynamic lookup
and $+OUT was not yet assumed as the standard 15:42
Whiteknight masak: Can't find ./Test in @INC 15:43
ruoso although TimToady was very much inclined to make it the standard
pmurias perl6: my $foo = 2;say $+foo; 15:44
p6eval pugs: OUTPUT«*** Can't modify constant item: VUndef␤ at /tmp/TPoVWsLEjq line 1, column 13-22␤ /tmp/TPoVWsLEjq line 1, column 13-22␤»
..elf 24936: OUTPUT«Can't call method "Str" on an undefined value at ./elf_f line 649.␤ at ./elf_f line 3861␤»
..rakudo 35718: OUTPUT«Scope not found for PAST::Var '$+foo' in ␤current instr.: 'parrot;PCT;HLLCompiler;panic' pc 146 (src/PCT/HLLCompiler.pir:102)␤»
ruoso hmmm... 15:46
I was surprised to realize that :foo($bar) doesn't expect directly an EXPR inside the parens
but it is a full circumfix with a semilist inside
Whiteknight Okay, I just put through a commit that should fix this issue 15:47
if I could get some testers to double-check me, that would be awesome
jnthn tests
ruoso TimToady, can the compiler simplify :foo($bar) to take directly the EXPR inside the statement inside the semilist inside the postircumfix? 15:48
or is there a different possible use to that?
pmurias, $+foo doesn't look in the current scope... only in the caller scope and onward 15:49
(backward, actually) 15:50
pmurias ruoso: Syntax.pod:1824 15:53
jnthn Whiteknight: Looks encouraging. 15:55
tigreton bye
ruoso pmurias, that looks like an exception 15:55
jnthn Whiteknight: Running spectest to be really sure, but think it's fixed. Thanks! 15:56
Whiteknight++ 15:57
Whiteknight don't thank me, curse whoever wrote this code originally
okay, I'm logging off for the afternoon probably. Talk to you gentlemen later 16:00
pugs_svn r24937 | pmurias++ | [mildew] 16:01
r24937 | pmurias++ | use SDL::App:from<perl5> calls $P5Interpreter.eval('use SDL::App')
pmurias ruoso: we have to figure out what to we want an use Foo::Bar:from<perl5> to export 16:02
pmurias s/to/do 16:03
masak Whiteknight++ 16:04
pmurias we could run the 'use Foo::Bar' using Safe.pm and expose whatever appears in main:: afterwards 16:05
ruoso: seems sane?
szabgab After NPW there is going to be a Perl6 hackathon I am planning to participate, others whom are interested, please put your name on 16:42
www.perlfoundation.org/perl6/index....katon_2009
masak is interested 16:44
szabgab: (that's not the way to use 'whom')
szabgab whome maybe? 16:45
masak szabgab: it's like this: 'who' and 'whom' are like 'they' and 'them'. you can tell that 'them are interested' sounds wrong. so does '...whom are interested'. 16:46
szabgab: 'whome' does not exist as a word in English.
...AFAIK.
szabgab ah typo++, 16:47
so how would I say that ? others that are interested ?
masak 'that', or 'who', works. 16:48
szabgab oh crap, just put your name on the wiki :-)
masak I did.
I'm not _just_ a grammar nazi. :P 16:49
I also do things.
szabgab np :-) 16:51
in the end we'll have to ask TimToady about that sentence 16:52
masak not really.
the thing about 'whom' in English is that it's not really needed in current usage. so when it is used wrongly, it sounds grating for at least two reasons. 16:53
jnthn whom is dead. 16:58
In a generations time, it should be mostly forgotten. :-)
masak thing is, I really like some sentences that has'em. :) 16:59
Khisanth replaced by a one letter equivalent? :p
masak 'to who it may concern' doesn't sound nearly as good... 17:00
ruoso pmurias, hmm 17:06
I think we don't need to support export by default
we might support explicit import of symbols 17:07
like...
ruoso use CGI:from<perl5> <header h1>; 17:07
would import the symbols header and h1
pmurias export by default will require use-ing the perl5 module at compile time 17:10
pmurias ruoso: the thing i have most doubts about when should ::CGI be exported 17:11
ruoso what do you mean?
pmurias we need a protoobject for CGI.new to work 17:12
ruoso ah... I see..
yes, we need a P5Package object
that we register as 'CGI::', 'CGI' and '&CGI := { CGI }' 17:13
and CGI.new is the same as CGI::.new in P5
pmurias why do we register &CGI?
ruoso because CGI.new is parsed as CGI().new() 17:14
diakopter ruoso: I'm here now 17:16
pmurias: same
pmurias diakopter: hi
ruoso diakopter, hi
diakopter how should those parameters look
pmurias none at all 17:17
diakopter hm, ok
pmurias m0ld_exe doesn't receive parameters
ruoso: only because is_type is not filled in correctly in STD
diakopter: did you change SConstruct? 17:18
diakopter yeah, I had to, some 17:19
I don't have 'cat'
but anyway, it's stuck at the same spot now.
meppl good night 17:20
pmurias diakopter: what you could do is to commit your changes and i'll see if it works correctly on linux 17:21
diakopter after line 154 of control_exception_return.c is emitted, sm0p.exe is hanging/pausing 17:23
perl6.pastebin.com/f3a71b9b4
buubot diakopter: The paste f3a71b9b4 has been copied to erxz.com/pb/14915
pmurias diakopter: .ri files are emitted by tools/ri not sm0p.pl 17:25
mtnviewmark std: $foo.~^
p6eval std 24937: OUTPUT«############# PARSE FAILED #############␤Syntax error at /tmp/cWFHYolBOG line 1:␤------> $foo.~^␤ expecting postfix_prefix_meta_operator␤00:05 86m␤»
diakopter pmurias: I don't see how that applies to my situation
mtnviewmark std: $foo.:~^ 17:26
p6eval std 24937: OUTPUT«############# PARSE FAILED #############␤Obsolete use of . to concatenate strings; in Perl 6 please use ~ instead at /tmp/jhTO9QQTHC line 1:␤------> $foo.:~^␤ expecting dotty method or postfix␤00:05 86m␤»
mtnviewmark aha!
pmurias diakopter: the should be no such thing as sm0p.exe 17:27
ruoso pmurias, is_type isn't looked for in CGI.new
mtnviewmark so .:<prefix> doesn't work as per Syn03...
diakopter pmurias: what should the sm0p executable be named?
pmurias diakopter: is the paste smop related
ruoso pmurias, it is looked in places like "does Foo" or "is Foo"
masak mtnviewmark: hello. what doesn't work as per Syn03, you say?
pmurias diakopter: what do you mean by the sm0p executable 17:28
diakopter pmurias: yes, it's one of the 3 threads in sm0p.exe while it's "hung", the most active one
mtnviewmark hello masak
Syn03 says .:<postfixop> is a method call op
diakopter pmurias: the one that's built by Setup.exe
mtnviewmark but, on my reading of STD.pm it wasn't so I checked here
and, it isn't
ruoso diakopter, that's m0ld.exe
pmurias m0ld_exe.exe
diakopter oh, ok
masak mtnviewmark: that might have been deprecated.
diakopter I must have mistyped sm0p somewhere then 17:29
ruoso diakopter, there is also sm0p.pl
masak mtnviewmark: ISTR that filetest adverbs :x got desugared to .:x before, but .x now.
mtnviewmark okay - I'll write up my findings on methodcall operators - as STD is a bit far from Syn03...
diakopter sorry, I meant m0ld.exe is hanging
ruoso diakopter, it's probably waiting for input 17:30
are you sure it is receiving the data via stdin?
mtnviewmark but AFAICT there is no dotted from of prefix operators anymore
pmurias diakopter: could you paste your SConstruct
?
diakopter ruoso: it did indeed output 154 lines of control_exception_return.c
masak mtnviewmark: prefix operators? 17:31
mtnviewmark masak - do you know what .^!method and !method are? the later is named privop in STD
masak - right there is a dotted from of postfix operators: .++ and .i for example
masak mtnviewmark: aye, they are calls/references to private things in classes.
mtnviewmark (wonders how .i is parsed.. operator? or method name?) 17:32
ruoso diakopter, right... maybe tools/ri is trying to talk to m0ld_exe but it's not being able to send the input data
mtnviewmark std: $p!seekret(42)
masak mtnviewmark: TimToady++ decided the other day to always treat .i as a method.
p6eval std 24937: OUTPUT«00:05 86m␤»
jnthn mtnviewmark: .i I think with latest tweaks is...what masak said. :-)
masak :)
mtnviewmark what? not as an operator? so 3 + 4i is gone? now 3 + 4.i
jnthn No 17:33
masak mtnviewmark: no 4i is still there.
jnthn 3 + 4i will work
3 + 4.i wouldn't
mtnviewmark well, in that case i is an operator --
masak sure.
mtnviewmark ah - just .i was the adjudicated result... got it 17:34
masak it just doesn't have a .i form.
ruoso diakopter, can you try to debug tools/ri ?
mtnviewmark so - what is .:method if no longer the dotted form of prefix opeerators?
diakopter ruoso: perl6.pastebin.com/f2a0d8da0
buubot diakopter: The paste f2a0d8da0 has been copied to erxz.com/pb/14916
diakopter ruoso: yes, give me a few minutes 17:35
masak mtnviewmark: I have no idea. I hardly understand the question. :/ 17:36
ruoso diakopter, I don't understand windows command expansion...
but the way tools/ri is called seems wrong to me
diakopter how does it seem wrong? all I changed was how the $SOURCE is fed 17:36
mtnviewmark doesn't think that windows command expansion is deterministic! :-) 17:36
okay - well, i've got to go now - I'll send my methodcall findings out later today! 17:37
caio!
diakopter ruoso: screenshot of process explorer for the process tree. imagebin.ca/view/0Ozu8GM.html 17:43
ruoso, believe it or not, cmd.exe does handle the nested double-quotes correctly, afaict/afaik. 17:45
diakopter installs ActiveState's interactive gui debugger 17:46
masak ok, I'm withdrawing for the evening to concentrate on too many Chinese characters for tomorrow. see y'all. 17:47
diakopter ruoso: any thoughts? 17:52
diakopter ruoso: okay, I'm now stepping through tools/ri 18:14
ruoso check if it is calling m0ld correctly 18:15
diakopter how will I be able to tell?
ruoso what are the parameters send to open3 18:16
?
s/send/sent/
does Open3 works as expected in Windows? 18:18
moritz_ re 18:19
lambdabot moritz_: You have 1 new message. '/msg lambdabot @messages' to read it.
diakopter ruoso: @_[3] in open3() is "m0ld.exe"; that's the only parameter that stringifies to anything other than "" 18:21
ruoso diakopter, so after that step, does it write to m0ld? 18:22
where does it hang? 18:23
diakopter I'm not sure I can trace the fh with this debugger...
ruoso ok... go step by step, and tell me where it hang? 18:24
diakopter during the exec @cmd in Open3.pm 18:25
diakopter I'm pretty confident IPC::Open3 has been specifically tweaked to work well in this Strawberry Perl 18:26
diakopter but yeah it's starting to seem as if it's not sending input to m0ld.exe 18:27
ruoso try open2 instead
ruoso (does win32 have a stderr?) 18:28
diakopter heh; the graphical debugger didn't appreciate me killing the whole process tree
yes
ruoso: switching it to use open2() helped 18:34
it seems to have finished writing to control_exception_return.c
ruoso cool
pugs_svn r24938 | ruoso++ | [mildew] starting the support for named parameters. support for colonpair and fatarrow. added nibbler vast type... only the signature handling is missing; 18:35
ruoso diakopter, I think you can commit the version using open2
there shouldn't be much harm
diakopter ruoso: ok
ruoso: but, see this paste 18:36
erxz.com/pb/14921 18:37
ruoso diakopter, is the control_exception_return compleete? 18:38
diakopter ruoso: well, there aren't any incomplete blocks at the end :) I don't know what the last block should be, though
ruoso diakopter, for some reason scons seems to think your compiler is named 'cl'
diakopter, is there something that looks like the output from m0ld? 18:39
diakopter what does the output of m0ld look like?
ruoso a function call with a lot of numbers as arguments
numbers and C symbols, that is 18:40
diakopter in the .c file? no. 18:40
also, m0ld.exe is still running (twice), orphaned 18:41
ruoso hmm... ok... so it looks like it didn't work as expected
the warnings were pointing in that direction
pugs_svn r24939 | moritz++ | [t/spec] 18:41
r24939 | moritz++ | * wrote tests for multis with multiple signatures
r24939 | moritz++ | * corrected redo.t
r24939 | moritz++ | * some unfudges
r24939 | moritz++ | * fixed slurpy-params-2.t to use [+] @a instead of @a.sum
ruoso diakopter, it sure has to do with IPC in windows
diakopter yeah 18:42
ruoso need to go for a while...
ruoso later &
pugs_svn r24940 | moritz++ | [t/spec] unfudge subref.t for rakudo 18:43
diakopter pmurias: what do you think about me trying to make tools/ri read and write to files based on filename parameters and not STDIN,STDOUT 18:45
I mean as an attempted solution, not necessarily a permanent change
ruoso diakopter, feel free to try it 18:49
pugs_svn r24941 | ruoso++ | [mildew] named parameters working!!!! YAY!!! 19:12
pugs_svn r24942 | ruoso++ | [smop/src-s1p] remove uncessary parens 19:13
ruoso and that means Multi.pm can be compiled! 19:14
pugs_svn r24943 | moritz++ | [t/spec] fudge submethods.t to be runnable (but doesn't pass yet) 19:17
ruoso we now need the mildew so loader
jnthn moritz_: Is that, none of it passes? 19:18
I think I looked at that and it was the lexicals outside classes issue getting in the way as much as anything...
moritz_ jnthn: that's right 19:19
jnthn added submethods just the other day
Then looked to add that test file and realized we couldn't...and I think that was about the last thing I did on Rakudo. :-) 19:20
ruoso mildew: sub foo(:$named) { $OUT.print($named.FETCH)}; foo(:named<Hello from Mildew>);
p6eval mildew: OUTPUT«Can't locate object method "emit" via package "AST::Pair" at src/AST.pm line 255.␤»
ruoso hmm... outdated eval
ruoso pmurias, I think I'll create another directory in smop to host prelude that is already compilable, and make smop load that prelude from start 19:22
pugs_svn r24944 | moritz++ | [t/spec] unfudge passing test in basic-types.t for rakudo 19:32
r24945 | moritz++ | [t/spec] more unfudges 19:33
r24946 | moritz++ | [t/spec] another unfudge 19:34
ruoso mildew: sub foo {$OUT.print("gone")}; sub bar(|$capture) { &foo.postcircumfix:<( )>((|$capture),:cc(&?ROUTINE.back)); $OUT.print("won't print") }; bar(); $OUT.print("mad"); 19:37
p6eval mildew: RESULT«Use of uninitialized value in hash element at ../../src/perl6/Cursor.pm line 819.␤Use of uninitialized value in concatenation (.) or string at ../../src/perl6/Cursor.pm line 881.␤Can't call method "emit_m0ld" on an undefined value at src/VAST/noun.pm line 34.␤» 19:38
ruoso hmpf... still outdated 19:40
moritz_ do I need to do anything else than building smop?
(which I just did) 19:41
ruoso moritz_, I think you need to rebuild std, and cleanup lex/ 19:42
moritz_ ruoso: try again please 19:44
ruoso sub foo { $OUT.print("gone\n") }; sub bar(|$capture) { &foo.postcircumfix:<( )>((|$capture),:cc(&?ROUTINE.back)); $OUT.print("won't print\n"); }; bar(); $OUT.print("mad\n");
mildew: sub foo { $OUT.print("gone\n") }; sub bar(|$capture) { &foo.postcircumfix:<( )>((|$capture),:cc(&?ROUTINE.back)); $OUT.print("won't print\n"); }; bar(); $OUT.print("mad\n"); 19:45
p6eval mildew: RESULT«(null pattern for prefix_circumfix_meta_operator:*) at ../../src/perl6/Cursor.pm line 464.␤Can't call method "emit_m0ld" on an undefined value at src/VAST/noun.pm line 34.␤» 19:46
moritz_ sh: line 1: 10544 File size limit exceededperl mildew /tmp/hLS5dzUjs9 >> /tmp/yLLJCp6LkD 2>&1
ruoso: I can increase that limit - how much space do you need?
ruoso the question is how mucg std needs 19:47
moritz_ I don't think STD.pm writes so much into one file 19:48
pugs_svn r24947 | moritz++ | [evalbot] increase max file size limit
pmurias ruoso: the mildew so loader works 19:52
moritz_ mildew: sub foo { $OUT.print("gone\n") }; sub bar(|$capture) { &foo.postcircumfix:<( )>((|$capture),:cc(&?ROUTINE.back)); $OUT.print("won't print\n"); }; bar(); $OUT.print("mad\n");
p6eval mildew: OUTPUT«gone␤mad␤»
rakudo_svn r35722 | moritz++ | [rakudo] add four files and a bit more documentation to t/spectest.data 20:30
ruoso moritz_++ 20:38
and now you saw mildew call/cc
mildew: sub foo(:$named) { $OUT.print($named.FETCH)}; foo(:named<Hello from Mildew>); 20:40
p6eval mildew: OUTPUT«Can't locate object method "emit" via package "AST::Pair" at src/AST.pm line 255.␤»
ruoso interesting...
ruoso hmm... that means I didn't implement the compiler hack for named argument in routine call... 20:42
pugs_svn r24948 | moritz++ | [t/spec] correct a test in code.t, and add another one for RT #62432 20:55
pugs_svn r24949 | ruoso++ | [mildew] implement the named args compiler hack in VAST::term; 21:16
rakudo_svn r35724 | moritz++ | [rakudo] correct file name in spectest.data 21:20
pugs_svn r24950 | moritz++ | [t/spec] regression tests for RT #62006 21:38
shinobi-cl hello every1! 21:51
moritz_ good localtime 21:52
shinobi-cl can i have your attention for a little?
i just started a perl6 wiki
on perl6 dot wikia dot com
i just have one article by now :( and it sucks hehe
you are all welcome to contribute ! 21:53
moritz_ shinobi-cl: why another wiki? we already have a few
most prominently www.perlfoundation.org/perl6/index.cgi?perl_6 which links to most others 21:54
shinobi-cl thhat wiki is sooo incomplete 21:55
moritz_ right
jnthn moritz_: Fudge fail? I get
t\spec\S12-attributes\mutators.rakudo 23 2 24-25
moritz_ shinobi-cl: so is the new one
jnthn (failed that is) 21:56
moritz_ jnthn: I'll investiage
shinobi-cl well, but it only has a few hours of life
anyway, there it its:)
seeya, got to make more articles
moritz_ good luck!
jnthn moritz_: Oh, I think it's test count 21:57
1..23
moritz_ jnthn: aye
jnthn later
ok 24 - # SKIP class Proxy
ok 25 - # SKIP class Proxy
OK, cool, not so bad then. :-)
moritz_ I just counted them, there *are* 25 tests in that file 21:58
fix committed
pugs_svn r24951 | moritz++ | [t/spec] fix plan() noticed by jnthn++ 21:58
moritz_ but I did find a fudge bug today 21:58
it forgets half of the file on 'make t/spec/S06-signature/slurpy-params-2.t' 21:59
@tell TimToady t/spec/fudge --keep-exit-code rakudo t/spec/S06-signature/slurpy-params-2.t emits an incomplete file, many tests are missing at the end 22:00
lambdabot Consider it noted.
pugs_svn r24952 | moritz++ | [t/spec] some more unfudges found by autounfudge 22:20
pugs_svn r24953 | moritz++ | [t/spec] get rid of some passing TODO tests by skipping them until we get 22:31
r24953 | moritz++ | clarification from p6l
shinobi-cl just some advertising...New perl6 wiki, on perl6.wikia.com .....Please contribute with guides and how-to s!!!!! thanks! 22:36
moritz_ rakudo now passes 6352 spec tests 22:51
jnthn moritz_++ # nice! 22:58
shinobi-cl hello 23:04
moritz__ the network in which the irclogs live seems very much down 23:10
shinobi-cl hi! a new wiki about perlt is online: perl6.wikia.com 23:22
an article about reading binary files (seek, and all that) would be a nice addition :)
diakopter shinobi-cl: you already mentioned it. 23:24
froele hi guys 23:27
just wondering anyone actually used lua
froele does perl have TST 23:31
ternary search trees? 23:32
it's a fractal of binary trees.
froele anyone familiar with this? 23:34
froele anyone designed MUDs here? 23:47
perhaps a Multi-threaded MUD codebase in perl
diakopter @tell scrottie froele asked about MUDs 23:48
lambdabot Consider it noted.
froele diakopter: who is scrottie? 23:49
diakopter froele: google him? 23:50
froele yeah done 23:54
but anyone in here who will give me some direction on this

Powered by Cro and the Raku® Programming Language.

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

Thank you!