»ö« | perl6-projects.org/ | nopaste: sial.org/pbot/perl6 | evalbot: 'perl6: say 3;' | irclog: irc.pugscode.org/ | UTF-8 is your friend!
Set by Tene on 14 May 2009.
pmichaud $tree.planted = True 00:02
00:02 vxnick left
jnthn :-) 00:03
pmichaud: Which way up is it facing? The computer science tree way, or the natural tree way? ;-)
pmichaud The root of the tree is at the bottom. 00:04
jnthn Judging from the diagrams, I wouldn't have trusted a lot of the folks who lectured me in computer science to know it should be that way.
pmichaud well, "bottom" and "top" are totally relative. 00:05
jnthn :-)
pmichaud To someone living in Australia, I suspect they think the tree I just planted is in the computer-science orientation.
And the tree is a Chinese Pistache, so who knows what the "correct" orientation is. :-) 00:06
Anyway, the parts that were surrounded with dirt when we bought the tree are now surrounded by even more dirt, which I think is the important part. 00:07
jnthn has to google to get an idea of what kinda tree that is
Hmm, looks like they can get quite pretty. :-) 00:08
pmichaud Yes.
We have one already in our side courtyard, and it's beautiful (even though we've been totally mistreating it for about 3 years)
jnthn Nice. :-) 00:09
pmichaud we originally thought it was a Grey Ash with a heavy mistletoe infestation. Turns out it's a Chinese Pistache and the "mistletoe" we've been vigorously cutting out is in fact its fruit.
jnthn Oh!
pmichaud Oops.
jnthn Eddible fruit?
I'm suspecting not. 00:10
pmichaud Anyway, it survived our mistreatment (and a couple of heavy droughts), so when it was time for a new tree in our front yard to replace the (largely dead) oak it was an easy pick.
I don't think its berries are all that edible, no.
However, the plum that sits next to the Chinese Pistache is *very* edible and makes wonderful plum jam. :-)
jnthn Oooh, nice.
pmichaud And our apple tree is covered with almost-ripe apples this year. 00:11
jnthn My family grew various fruits in the garden. The blackberries and rubarb were the best. :-)
pmichaud I've never been able to get our blackberries to really grow.
jnthn These days I live city center though. Which is nice in many ways, but leaves me without a garden.
pmichaud We have a (I suspect somewhat wild) blackberry vine in the courtyard as well, but I think it's not getting enough sun to really produce lots of berries. 00:12
If I try to train it to a trellis, it's probably too much shade.
Oh well.
It always manages to produce about a dozen blackberries, which the birds and rabbits get to long before I get to try one. :-|
jnthn Yeah, ours got a decent bit of sun.
pmichaud We lucked out with this house -- most of the plants/trees were here when we bought it -- we had no idea we had so many fruit-producing trees. 00:13
jnthn My mum tended to make various puddings out of the fruits. Which was nice. :-) 00:14
pmichaud Indeed.
okay, well, time for a shower, then dinner. Catch you tomorrow, likely.
jnthn Yes, I'm going to take some sleep. 00:15
Spending about half tomorrow on Rakudo, half on $other_job.
00:15 bacek left
pmichaud okay. I'll be travelling much of tomorrow. 00:15
I should be able to catch up with things in the evenings and over the weekends.
jnthn I'll try not to break too much. ;-)
pmichaud I'm not too worried about that. :-)
anyway, see ya soon. 00:16
Trey so... how *is* .prime defined in that example then? does it assume Int being reopened and a prime method added?
jnthn Have a safe journey.
pmichaud Trey: I don't know. I think it may be from a much earlier draft.
There was a time when methods did fall back to subs.
jnthn Trey: I suspect you'd have to augment to get that now.
pmichaud More likely one should just write prime($_)
jnthn Yes. 00:17
00:17 Tene left
Trey yeah, i remember that... but one could make C<@primes = (do (do $_ if .prime) for 1..100);> work if you defined an Int::prime... what about if you created a multi? would that work? 00:17
can single-arg multis be called with dot notation? 00:18
jnthn No.
Only methods.
Trey didn't think so
jnthn There's no fallback now.
Trey didn't think so.
thanks
jnthn You can augment the class, but, well, just write a sub. ;-)
OK, sleeps...night all 00:20
00:22 dextius joined, szabgab left
dextius can anyone give an idea why I am getting this error compiling rakudo? (first comment on the site) www.rakudo.org/node/43#comment-6 00:23
00:23 DemoFreak left 00:28 c9s joined 00:29 frew|work joined 00:31 Tene joined
skids dextius: looks like a string is getting created without a specified encoding. 00:34
dextius I've tried to rakudo for the last 3 builds, keep getting the same error.. 00:35
skids If you do a backtrace, there may be a value in the string to give more clues.
(that is, run under gdb)
dextius skids: ok, I'll try and give that a try.. 00:36
skids: I ran "gdb /code/rakudo/parrot/parrot", at the prompt, I typed "run -o perl6_s1.pbc perl6.pir" at the prompt. It dumped the same thing I posted to the web page. Do I need to give it some sort of "really verbose" option to get what you are looking for? 00:40
skids yes you can walk up the stack and look at the code. 00:41
First thing is, it should have also shown you a line of code after the stack 00:43
and "list" should get you more surrounding code.
dextius skids: src/string/api.c:762: failed assertion 'encoding'
skids Right but it should show you source code. 00:44
dextius skids: Yeah, I typed list, I got an "int main" looking routine..
00:44 nsh joined
dextius (sorry, my GDB skills are non-existent, just a very happy perl programmer) 00:44
skids so look at the line 762 in the listing.
and there is probably a string object being tested by an assert 00:46
00:46 Casan left
dextius skids: yeah, Parrot_str_new_init 00:47
skids what's the argument to that function?
dextius ARROT_INTERP, ARGIN_NULLOK(const char *buffer), UINTVAL len, ARGIN(const ENCODING *encoding), ARGIN(const CHARSET *charset), UINTVAL flags 00:48
skids so use "print" to see what are in those values
dextius cool, ok, one sec 00:49
00:49 clkao_ is now known as clkao
skids bets encoding is null 00:50
dextius am I an idiot? "(gdb) print encoding No symbol "encoding" in current context."
skids digs up that part of the source/ 00:51
might be some kind of macro magic in the way 00:53
dextius skids: screwing around, I ran "bt" and it gave me this.. #5 0x00901b4b in Parrot_str_new_init (interp=0x801790, buffer=0x734acc "MutableVAR", len=10, encoding=0x0, charset=0x0, flags=12288) at src/string/api.c:762 00:54
I'm going to guess that 0x0 is "null" 00:55
I'll dump this onto the rakudo page..
skids OK then do "up" and another list to look at where that was called 00:56
dextius skids: I dumped the entire list to www.rakudo.org/node/43#comment-6
skids Ah, what does "print PARROT_DEFAULT_ENCODING" give? 00:59
dextius No symbol "PARROT_DEFAULT_ENCODING" in current context. 01:01
<sigh>
skids agh, it's a macro.
How about Parrot_fixed_8_encoding_ptr
dextius $1 = (ENCODING *) 0x801d60 01:02
Where did you find this (been greping a storm over here)
skids this is in the parrot source code, so under parrot/src/ and parrot/include/parrot/
dextius Ok. I simply did the instructions on the rakudo page to compile (git, then perl Configure.pl --gen-parrot) 01:03
Is it possible parrot built incorrectly on my system? 01:04
skids Oh, you mean you were trying to run against and installed parrot?
s/and/an/
dextius No, I don't have parrot installed. I think Configure.pl svn's out parrot and compiled it for me. 01:05
skids damn, that would have been an easy explanation :-)
So it's weird then, PARROT_DEFAULT_ENCODING should have been Parrot_fixed_8_encoding_ptr, not 0x0. 01:06
dextius heh, sorry, thanks for your help.. i bought both O'Reilly books on Parrot a long time ago, I keep trying to build this thing every few weeks, was trying to avoid coming here and begging for help :-(
skids I don't do anything else useful around here so I don't mind helping :-) 01:07
dextius Is it possible that Rakudo isn't seeing Parrot's encoding?
01:09 araujo joined
skids I don't think so, it's happening inside an API function where the value passed to "encoding" is not dependent on inputs 01:10
s = Parrot_str_new_init(interp, buffer, strlen(buffer), 01:12
678 PARROT_DEFAULT_ENCODING, PARROT_DEFAULT_CHARSET,
679 PObj_external_FLAG|PObj_constant_FLAG);
01:12 eternaleye joined
skids so unless something is screwed with the headers, PARROT_DEFAULT_ENCODING should be defined in parrot/include/parrot/encoding.h 01:13
dextius Is there any way to interrogate the "default" that is being passed as a null? 01:14
(for encoding)?
pmichaud I have a couple of questions -- what platform?
dextius Mac OS X
pmichaud 64 bit, 32 bit?
Intel? 01:15
dextius Darwin victory 9.6.0 Darwin Kernel Version 9.6.0: Mon Nov 24 17:37:00 PST 2008; root:xnu-1228.9.59~1/RELEASE_I386 i386
pmichaud okay.
That's not one of our common platforms but it seems like it should still work.
out of curiosity, try running the parrot command with the -G flag 01:16
dextius smiles.. I'll be your tester then :-)
sure.. one sec
pmichaud parrot/parrot -G perl6_s1.pbc perl6.pir # I think
oh, missing a -o there
parrot/parrot -G -o perl6_s1.pbc perl6.pir # I think
dextius inside of GDB ?
pmichaud yes, or from the command line 01:17
either one
wayland76 Nopaste URL: sial.org/pbot/perl6
lambdabot wayland76: You have 3 new messages. '/msg lambdabot @messages' to read them.
dextius pmichaud: done, same error
pmichaud hmmm.
what are the contents of build/PARROT_REVISION ?
wayland76 Why do my lambdabot messages keep returning from the dead? 01:18
skids waylang76: try @clear
skids: trying typing slower
wayland76 I've done @clear on at least 3 separate days
dextius 39019
pmichaud hmmm.
At NPW we also someone with problems building Rakudo under Darwin. 01:19
*also had someone
dextius I watched a bunch of the videos from NPW on iTunes, looks like you guys had a nice time.. 01:20
pmichaud It was very cool, yes.
dextius: I don't have any quick answers, I'm afraid. You may have stumbled upon a Parrot bug that is going to take a bit of time to locate and resolve. 01:21
dextius hey, no problem, I dumped everything I've found to www.rakudo.org/node/43#comment-6 ... Thank you VERY much for your time/help. I really want to learn and hopefully contribute some day.. Thanks again.
pmichaud But I'll pass your bug report along to the Parrot developers -- there are several people there who are pretty quick about finding and fixing such things.
skids intuits some kind of compiler argument list construction foobar
dextius has to go take care of the dog. buh-bye! 01:22
pmichaud skids: perhaps, but it looks as though it's having trouble loading the .so
01:22 dextius left
pmichaud The "MutableVar" stuff is a dynpmc 01:22
01:22 bacek joined
pmichaud so it's failing during "Parrot_lib_perl6_group_load", which is the part that results from the .loadlib directive in perl6.pir 01:23
skids Ah. probably it then. (but I've never been able to get something to even attempt to run without a dllib found)
Though that's probably due to always doing proper error handling on dl import :-) 01:24
pmichaud well, it's finding the .so okay -- it's just having trouble when initializing
skids is there a way to build static? 01:25
pmichaud not really 01:26
since they're dynpmcs, they're not part of parrot.
skids notes PARROT_NO_EXTERN_ENCODING_PTRS (and the violation of Parrot "no double negatives" coding guide) 01:28
pmichaud I wonder what that is for.... 01:29
oh, I see. 01:30
we should've found out from dextius what compiler he's using 01:31
01:33 viirya joined, wknight8111 left
pmichaud my guess is that somehow Parrot_default_encoding_ptr isn't being initialized. 01:33
There's also a Parrot_default_encoding() function, but apparently it's never used. 01:34
skids Parrot_fixed_8_encoding_ptr had a value, and that's what the macro is set to. 01:35
pmichaud Hmmm. 01:38
I wonder what PARROT_DATA ended up being defined as.
maybe getting a copy of his include/parrot/config.h would yield some clues. 01:41
anyway.
gotta pack for a trip tomorrow.
skids bon voyage! 01:42
Infinoid gah, dextius triggered the encoding/charset race condition 01:43
01:44 IRSeekBot left
skids
.oO(reliably, it seems)
 01:44
Infinoid wait, no he didn't, that isn't the init race. That might be an export/linker issue 01:45
pmichaud Yes, I think it's more likely an export/linker issue.
Because at the time this occurs, Parrot is already running and has started processing perl6.pir
(and is at the stage of performing the .loadlib)
Infinoid and has a valid Parrot_fixed_8_encoding_ptr (within the core, at least)
pmichaud although since the encoding isn't coming from the library being loaded.... well, that's just weird. 01:46
Infinoid yeah. something's off there 01:47
01:57 kst` is now known as kst 02:04 dukeleto left, dukeleto joined 02:36 lichtkind left 02:40 sri_kraih joined, ElectricHeavyLan left 02:56 sri_kraih_ left 02:57 Chillance left
meppl good night 03:03
03:07 meppl left 03:08 frew|wor2 joined 03:20 donaldh left, donaldh joined 03:25 alester_ joined
s1n any idea what "Null PMC access in can()" might imply? 03:30
skids That there's a bug in can()?
03:31 nsh left
dukeleto howdy 03:31
skids OH HAI 03:32
s1n i think i found it, lemme see if i can reproduce 03:34
here we go 03:35
rakudo: $test.say; my $test = 5; 03:36
p6eval rakudo 5ed6da: OUTPUT«Null PMC access in can()␤in Main (/tmp/4ovG5cGhvE:1)␤»
s1n that's not quite the error i was expecting
rakudobug?
03:41 orafu left 03:47 orafu joined 03:56 frew|wor2 left 03:57 frew|wor2 joined 03:59 frew|wor1 left 04:02 frew|wor1 joined 04:03 kst left, kst joined 04:27 alester__ joined, alester_ left 04:47 cspencer joined
eternaleye rakudo: multi infix:<union>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item} = 1; }; }; return %result.keys; }; say (1, 2, 3, 4) union (1, 7, 3, 8); 04:52
p6eval rakudo 5ed6da: OUTPUT«123478␤»
eternaleye mmm, tasty set theory 04:53
rakudo: multi infix:<intersection>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item}++; }; }; my @answer; for %result.kv -> $key, $value { @answer.push( $key ) if $value == @sets.elems; }; return @answer; }; say (1, 2, 3, 4) intersection (1, 7, 3, 8); 04:55
p6eval rakudo 5ed6da: OUTPUT«␤»
eternaleye hm
rakudo: multi infix:<intersection>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item}++; }; }; my @answer; for %result.kv -> $key, $value { @answer.push( $key ) if $value == @sets.elems + 1; }; return @answer; }; say (1, 2, 3, 4) intersection (1, 7, 3, 8);
p6eval rakudo 5ed6da: OUTPUT«␤»
eternaleye rakudo: multi infix:<intersection>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item}++; }; }; my @answer; for %result.kv -> $key, $value { @answer.push( $key ) if $value == @sets.elems - 1; }; return @answer; }; say (1, 2, 3, 4) intersection (1, 7, 3, 8);
p6eval rakudo 5ed6da: OUTPUT«␤» 04:56
eternaleye rakudo: multi infix:<intersection>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item}++; }; }; my @answer; for %result.kv -> $key, $value { say "$key: $value -- {@sets.elems}" }; return @answer; }; say (1, 2, 3, 4) intersection (1, 7, 3, 8); 04:57
p6eval rakudo 5ed6da: OUTPUT«1: 2 -- 8␤2: 1 -- 8␤3: 2 -- 8␤4: 1 -- 8␤7: 1 -- 8␤8: 1 -- 8␤␤»
eternaleye Ah, it's flattening 04:58
rakudo: multi infix:<intersection>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item}++; }; }; my @answer; for %result.kv -> $key, $value { @answer.push( $key ) if $value == @sets.elems; }; return @answer; }; say [1, 2, 3, 4] intersection [1, 7, 3, 8];
p6eval rakudo 5ed6da: OUTPUT«13␤»
04:58 frew|wor2 left
eternaleye rakudo: multi infix:<intersection>( Array *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item}++; }; }; my @answer; for %result.kv -> $key, $value { @answer.push( $key ) if $value == @sets.elems; }; return @answer; }; say (1, 2, 3, 4) intersection (1, 7, 3, 8); 04:58
p6eval rakudo 5ed6da: OUTPUT«Parameter type check failed; expected something matching Array() but got something of type List() for @sets in call to infix:intersection␤in sub infix:intersection (/tmp/2lSB7QALap:1)␤called from Main (/tmp/2lSB7QALap:1)␤» 04:59
04:59 wayland_ joined
eternaleye By the way, in my example above using List *@sets and passing in ()'d args, were those lists really supposed to flatten? 04:59
(the reason I used a slurpy was because in my head is was assoc<chain>) 05:00
*it
05:01 wayland76 left 05:03 frew|work left 05:06 wayland_ is now known as wayland76
wayland76 rakudo: multi infix:<∪>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item} = 1; }; }; return %result.keys; }; say (1, 2, 3, 4) ∪ (1, 7, 3, 8); 05:07
p6eval rakudo 5ed6da: OUTPUT«123478␤»
wayland76 I think I like that :) 05:08
So, are you volunteering to write Sets.pm ? 05:11
05:11 cspencer left
eternaleye Sure, just give me a bit to learn set theory first ;D 05:15
Although, subset and superset are easy. all() ~~ any() and vice versa. 05:16
With the benefit of autothreading, unlike a nested loops subset check. 05:17
05:17 skids left 05:18 masak joined
eternaleye Ooh, interesting thought: my Set integers = Set.new( seed => 0, up-generator => sub { $^a + 1 }, down-generator => sub { $^a - 1 } ); 05:21
Double-ended lazy list :D
Implementing subset of _that_ would probably be difficult, but rewarding 05:22
Of course, some mathematical wise guy'll probably try 'ok( $integers subset $real-numbers )' 05:23
"My thesis: Proving cardinality through Perl 6 one-liners" 05:24
05:24 szabgab joined
masak Peano would have been proud. 05:32
eternaleye rakudo: multi infix:<exclusion>( List *@sets ) { my %result; @sets.map: ->$list { @($list).map: -> $item { %result{$item}++; }; }; my @answer; for %result.kv -> $key, $value { @answer.push( $key ) if $value == 1; }; return @answer; }; say [1, 2, 3, 4] exclusion [1, 7, 3, 8]; 05:36
p6eval rakudo 5ed6da: OUTPUT«2478␤» 05:37
eternaleye I love how most of these only require changing the name and a single term (or maybe one expression)
rakudo: multi infix:<subset> ( @sub, @super ) { all(@sub) ~~ any(@super) ?? return( Bool::True ) !! return( Bool::False ); }; multi infix:<superset>( @super, @sub ) { @sub Rsubset @super }; say [1, 2, 3] subset [0, 1, 2, 3, 4]; say [1, 2, 3] superset [0, 1, 2, 3, 4] 05:43
p6eval rakudo 5ed6da: OUTPUT«1␤1␤»
eternaleye Hm.
Oh wait... 05:44
rakudo: multi infix:<subset> ( @sub, @super ) { all(@sub) ~~ any(@super) ?? return( Bool::True ) !! return( Bool::False ); }; multi infix:<superset>( @super, @sub ) { @super Rsubset @sub }; say [1, 2, 3] subset [0, 1, 2, 3, 4]; say [1, 2, 3] superset [0, 1, 2, 3, 4]
p6eval rakudo 5ed6da: OUTPUT«1␤0␤»
wayland76 Maybe you could write some perl code that generates Set.pm :) 05:45
eternaleye Metaops on custom ops, enabling sugar everywhere ;D
wayland76: That's way cooler than I could possibly accomplish!
wayland76 Did you read Dluglosz article on APL? 05:46
eternaleye Not yet, where?
wayland76 www.dlugosz.com/Perl6/web/ and click the "APL" one 05:47
s/Dluglosz/Dlugosz/
05:47 JDlugosz joined
wayland76 JDlugosz: We were just talking about you :) 05:48
JDlugosz Hey, everyone...
lambdabot JDlugosz: You have 2 new messages. '/msg lambdabot @messages' to read them.
JDlugosz lambdabot: really?
wayland76 Did you know that, or did you just turn up now by coincidence?
JDlugosz I was wondering... why was failover of method calls to subs removed (in August)?
No, just coincedence. I needed to ask a question. Noticed client wasn't running. 05:49
wayland76 Ok :)
masak JDlugosz: oh hai. reading your latest article right now.
I like it. thanks for writing it.
JDlugosz masak: updating it from remarks as we speak.
Found a few problems with the Syn along the way. Like failover change not updating ANY other docs. 05:50
masak JDlugosz: I have a slight nit, then: we tend to write S04<77> as S04:77. if you go to the logs, you'll even find that this latter syntax turns into links into the respective spec.
wayland76 Yeah, JDlugosz, I can write stuff, but you've done a really good job there. I'd be keen to see stuff like yours as part of any official tutorials we have :)
JDlugosz I'd be glad for it. You know, I used to write a lot for magazines, back when they were pleantiful. 05:51
Have you looked at any of the others around that directory?
wayland76 I wondered; masak, am I correct in thinking that you're thinking 77 refers to a line number -- and is JDlugosz referring to that, or to a version number?
I read over the APL one, but I haven't read the others. 05:52
masak wayland76: oh, I assumed line number.
JDlugosz That is the synax for version, mimicing the module version for use.
masak aha. I see.
no need to change, then.
JDlugosz I put the pop-up there for those who don't know that convention. Love that hypertext.
wayland76 S04<77>:31 :)
eternaleye Oh, that's an interesting thought. 05:53
JDlugosz BTW, what's the lambdabot message thing?
wayland76 lambdabot is a bot that will store messages for later
masak JDlugosz: missed the pop-up, thought it was a link to the line. :P
JDlugosz It's telling me I have messaages, but it's not giving them up.
wayland76 if you type the thing it says to type, then it will give you messages
eternaleye rakudo: my @array = ( [1,2,3], [4,5,6] ); say [[+]] @array
p6eval rakudo 5ed6da: OUTPUT«say requires an argument at line 1, near " [[+]] @ar"␤in Main (src/gen_setting.pm:2400)␤»
eternaleye Drat.
wayland76 lambdabot: @messages
lambdabot You don't have any new messages.
JDlugosz I copy/paste'd, and it didn't work. Just showed @messages as a message. 05:54
masak p6eval: that's not ever going to work. :)
eternaleye: ^
eternaleye Ah.
So the syntax for that is @array.map: { [+] $_ } ?
wayland76 JDlugosz: try just typing "lambdabot @messages", which will put them in the public channel (hopefully they're not private), and we can hopefully tell you what's going wrong
eternaleye rakudo: my @array = ( [1,2,3], [4,5,6] ); say @array.map: { [+] $_ } 05:55
05:55 dduncan joined
p6eval rakudo 5ed6da: OUTPUT«1 2 34 5 6␤» 05:55
JDlugosz The APL stuff was half joking, but serious in showing how you can warp things to your desire.
OK...
lambdabot @messages
eternaleye rakudo: my @array = ( [1,2,3], [4,5,6] ); say @array.map: { [+] @( $_ ) }
p6eval rakudo 5ed6da: OUTPUT«615␤»
wayland76 Put a colon after "lambdabot"
(sorry)
JDlugosz lambdabot: @messages 05:56
lambdabot You don't have any new messages.
wayland76 I guess they must've gotten cleared. I don't know what happened there
JDlugosz OK, then why tell me I had 2? Or did you think you told me already so they are not new?
eternaleye rakudo: my @array = ( [1,2,3], [4,5,6] ); say [+] @array.map: { [+] @( $_ ) }
masak I should add that lambdabot has been acting unreliably lately.
p6eval rakudo 5ed6da: OUTPUT«21␤»
JDlugosz Well, hopefully whoever it was will get back.
wayland76 Yes, my messages keep coming back after I @clear them
masak wayland76: yes, that's one of the symptoms. 05:57
wayland76: my guess is that someone will write a better bot in Perl 6 fairly soon.
eternaleye Well, there's phenny++ IIRC
masak aye.
phenny: I keep forgetting about you. :)
JDlugosz So, what do you think of all the wrapping that ()'s are doing, now that they are active object-creating elements?
wayland76 I can tell you the magic answer JDlugosz 05:58
JDlugosz And does anybody know why the failover was removed?
wayland76 lambdabot was supposed to give you a message from TimToady saying "the lines() function already defaults to $*ARGFILES; that's its only advantage over the method form"
masak JDlugosz: "Because of that, having whatever if .prime for 1..100 would be a suffix for a statement that already has a suffix, and that is not allowed." are you sure this isn't also old news? 05:59
rakudo: .say if $_ % 2 for 1..10
JDlugosz OK, thanks.
p6eval rakudo 5ed6da: OUTPUT«1␤3␤5␤7␤9␤»
wayland76 And another one from TimToady that was supposed to say "yes, S11:573 specs the separate compilation principle that implies each modules starts in its own SETTING"
JDlugosz masak: I already changed it. Lots of old clutter in the Synopses. 06:00
wayland76 Well, I hope we get to keep our karma with the new bot :)
JDlugosz Thanks again, Wayland.
masak JDlugosz: please make sure to keep a list of the cruft in the synopses too. :)
JDlugosz: or just change them directly yourself. 06:01
JDlugosz Now that I have commit bit, I don't need to keep a dusty list that nobody ever applies anyway. Very frustrating, it was.
I'll change the things I've found here this (long) weekend.
masak JDlugosz++
JDlugosz Along with a few I recall spotting before. Have to diff my old edits from "offerings" that have grown mold. 06:02
wayland76 anyway, afk &
masak reads "the seat of spirit and courage"
wonderful.
JDlugosz Larry ^^^^ Why was "failover to subs" removed, in August? Problem getting rules to work in general (I recall working on that), or an overwhelimg issues that killed it? 06:03
Thanks, masak. Turns out someone can live without a spleen after all.
masak aye. :)
JDlugosz Notice all the other biology refs? I changed others, like "under the hood" to conform, "under the skin". 06:04
masak oh, nice.
06:05 viirya left
masak JDlugosz: I, for one, am happy that failover is gone. I'm not sure exactly why it did go away, but it feels very much like a p5ism, and not consistent with the fairly strict OO-ish bits of p6. 06:07
06:07 dduncan left
JDlugosz failover is more consistent with generic functions a'la MMD, like in CLOS etc. invocant syntax doesn't work if the invokant is in the middle or multiple. 06:08
masak JDlugosz: TimToady sort of gave a reason in an email from Feb 20: "Anyway, we're trying to maximize flexibility while also giving the compiler as much information as possible to feed the optimizer. So fallbacks are not in vogue these days."
JDlugosz And you can say .prime instead of prime($_).
masak JDlugosz: to me, that looks tantamount to monkey patching. 06:09
JDlugosz That makes sense. Tracking a known static type of a class is one thing, but you lost the ability to give compile-time errors if it must look for more functions at run-time anyway. 06:10
masak JDlugosz: you might also want to check out the speculations by TimToady in an email from Mar 5.
JDlugosz which year? 06:12
masak this one. 2009.
JDlugosz The change was made in August.
masak they were.
the speculation came afterwards.
JDlugosz OK, I'll look back. 06:13
masak and I'm not sure they're directly related. just pointing you to the latest thoughts about subs and methods.
JDlugosz: as to which bracket-like signs show up on my system, I can see all but the second-last pair just fine. I'm on Mac OS X 10.5.7. 06:23
JDlugosz The 567 doesn't show?
masak it doesn't. 06:24
oh, "Sorry, you asked?" shouldn't have the comma.
JDlugosz Sorry, thanks. <g>. 06:25
OK, I updated that essay. Might need polishing where I changed things, but should have fixed everything that was noted. 06:30
If anyone wants to take a fresh look.
06:31 JDlugosz left
masak my tuits ran out for now, but I will come back to it in a few days. 06:31
06:33 finanalyst joined 06:34 alester__ left 06:42 kst left
eternaleye Drat, he left 06:43
06:43 kst joined
eternaleye @tell JDlugosz Your article says: ...just like 2+(3+4) has parentheses that group it the same way it was going to do anyway. 06:43
lambdabot Consider it noted.
masak eternaleye: use phenny! :)
eternaleye I don't know her syntax, and when I PM'd every help invocation I could think of she didn't respond 06:44
masak phenny: msg eternaleye is this the right syntax? 06:45
hm. seems not.
eternaleye I'll look it up in the logs
masak ah -- inamidst.com/phenny/ 06:46
tell eternaleye so this is the syntax?
phenny: tell eternaleye so this is the syntax?
phenny masak: I'll pass that on when eternaleye is around.
masak ah.
eternaleye Aha!
masak phenny++
phenny eternaleye: 06:46Z <masak> tell eternaleye so this is the syntax?
eternaleye Does that Z mean Zulu time? 06:47
masak Zebra time. 06:48
eternaleye Heh, April 28: 13:50 06:49
masak
phenny: pleased to meet you. I will use you for @telling next time.
You liedz! 06:50
masak :)
eternaleye: you mean I've been using lambdabot after that?
that wouldn't surprise me...
I should make a bot to remind me when I'm using the wrong bot.
wayland76 phenny: help me!!! 06:51
eternaleye Well, I seem to have accidentally the whole search function, so...
wayland76 The thing I like about Lambdabot is that it has useful help
That's why I like it better than other bots :) 06:52
masak yes, useful help is a really worthwhile feature in a bot. 06:53
discoverability.
wayland76 Even lambdabot could be improved, but it's better than nothing :) 06:54
06:55 justatheory joined
eternaleye masak: Actually, you seem to have simply been in-channel when you wanted to tell someone something, and they backlogged 06:55
wayland76 s/ lambdabot / lambdabot's help/
eternaleye You haven't used any bots in this channel since you said that to phenny
masak eternaleye: :)
eternaleye Well, except p6eval++
06:56 DemoFreak joined
masak p6eval: you're my favourite bot of them all! 06:56
eternaleye @karma p6eval
lambdabot p6eval has a karma of 1
eternaleye lambdabot: p6eval += 3
@karma p6eval
lambdabot p6eval has a karma of 1
eternaleye :(
masak rakudo: say $*IN.words.grep({$_ eq .ucfirst}).join(' ') 06:57
p6eval rakudo 5ed6da: OUTPUT«IO()<0xb62232f8>␤»
masak hm.
rakudo: say $*IN.slurp.words.grep({$_ eq .ucfirst}).join(' ')
p6eval rakudo 5ed6da: OUTPUT«Land Berge, Land Strome, Land Äcker, Land Dome, Land Hämmer, Heimat Söhne, Volk, Schöne, Österreich, Österreich! Heiß Erdteil Herzen Hast Ahnentagen Sendung Last Österreich, Österreich! Mutig Zeiten, Einig Brüderchören, Vaterland, Treue Österreich, Österreich!␤»
masak Austrian national anthem, all nouns pulled out. :) 06:58
eternaleye rakudo: say $*IN.words.grep( *.comb.pick(*) ).join('') 07:00
p6eval rakudo 5ed6da: OUTPUT«␤»
eternaleye rakudo: say $*IN.words.grep( { .comb.pick(*) } ).join('')
p6eval rakudo 5ed6da: OUTPUT«IO()<0xb62942f8>␤»
eternaleye rakudo: say $*IN.slurp.words.grep( *.comb.pick(*) ).join('')
p6eval rakudo 5ed6da: OUTPUT«␤»
eternaleye rakudo: say $*IN.slurp.words.grep( { .comb.pick(*) } ).join('') 07:01
p6eval rakudo 5ed6da: ( no output )
eternaleye oh wait
rakudo: say $*IN.slurp.words.map( { .comb.pick(*) } ).join('')
p6eval rakudo 5ed6da: ( no output )
masak rakudo: say $*IN.slurp.words.sort.[^50].join 07:02
p6eval rakudo 5ed6da:
..OUTPUT«AhnentagenBerge,Brüderchören,Dome,EinigErdteilHastHeimatHeißHerzenHämmer,LandLandLandLandLandLastMutigSchöne,SendungStrome,Söhne,TreueVaterland,Volk,Zeiten,amarbeitsfrohbegnadetbistdasdemderderderderdiedirdudueinemfreifrühenfürgetragen,gleich.gläubiggroßerhoffnungsreich.hoher␤»
masak oops.
eternaleye Song soup!
masak rakudo: say $*IN.slurp.words.sort({.lc}).[^50].join(' ') 07:03
p6eval rakudo 5ed6da: OUTPUT«Ahnentagen am arbeitsfroh begnadet Berge, bist Brüderchören, das dem der der der der die dir Dome, du du einem Einig Erdteil frei frühen für getragen, gleich. gläubig großer Hast Heimat Heiß Herzen hoffnungsreich. hoher Hämmer, in in inmitten, Land Land Land Land Land Last laß
..lieg…
eternaleye rakudo: say $*IN.slurp.map( { .comb.pick(*) } ).join('')
p6eval rakudo 5ed6da: ( no output )
masak likes the 'das dem der der der der die dir' part
eternaleye rakudo: say $*IN.slurp.map( { .comb.pick(3) } ).join('')
p6eval rakudo 5ed6da: OUTPUT«ear␤» 07:04
eternaleye rakudo: say $*IN.slurp.words.map( { .comb.pick(99) } ).join('')
p6eval rakudo 5ed6da: ( no output )
eternaleye rakudo: say $*IN.slurp.words.map( { .comb.pick(3) } ).join('')
masak 'ear'?
p6eval rakudo 5ed6da: ( no output )
eternaleye rakudo: say $*IN.slurp.map( { .comb.pick(3) } ).join('')
wayland76 Where does the input come from?
p6eval rakudo 5ed6da: OUTPUT«emm␤»
eternaleye rakudo: say $*IN.slurp.words.map( { .comb.pick(3) } ).join('')
masak wayland76: Austria. :)
p6eval rakudo 5ed6da: ( no output )
eternaleye Huh? 07:05
rakudo: say $*IN.slurp.map( { .comb.pick(3) } ).join('')
wayland76 Where abouts in Australia (ie. you know what I mean :) )
p6eval rakudo 5ed6da: OUTPUT«Z f␤»
eternaleye rakudo: say $*IN.slurp.words.map( { .comb.pick(*) } ).join('')
wayland76 (s/Australia/Austria/ -- stupid fingers)
p6eval rakudo 5ed6da: ( no output )
eternaleye rakudo: say $*IN.slurp.words.map( { .comb.pick(20) } ).join('')
p6eval rakudo 5ed6da: ( no output )
masak wayland76: svn.pugscode.org/pugs/misc/evalbot/evalbot.pl
eternaleye Grr, y inputline is not cooperating 07:06
rakudo: say $*IN.slurp.map( { .comb.pick(*) } ).join('')
p6eval rakudo 5ed6da: ( no output )
eternaleye rakudo: say $*IN.slurp.map( { .comb.pick(20) } ).join('')
p6eval rakudo 5ed6da: ( no output )
masak eternaleye: I don't know what you're trying to do... but it's not working.
eternaleye I realize, but have no clue why 07:07
I'm trying to get a jumbled mess of letters, but it's not working :(
masak eternaleye: slurp is one item, so no use running map on that.
eternaleye Well, they why is the words one giving nothing?
*then
rakudo: say $*IN.slurp.comb.pick(*).join('') 07:08
p6eval rakudo 5ed6da: ( no output )
masak rakudo: print .comb.pick(*), " " for $*IN.slurp.words
p6eval rakudo 5ed6da: OUTPUT«Lnda erd eBr,ge dnLa am omtr,Se naLd dre kÄrec, adnL erd mDe,o aLdn dre mm,eräH kszife!thucnru ateHim bsti du oßrger Sn,höe ,oVkl teaenbdg rüf dsa n,hSöce telhisveregüm cri,rhÖseet meeüehtlrgivs !eshrctiÖre ißeH de,fmtehu iwld ntrsetumit sielgt edm tlreidE du nntitie,m einme 07:09
..arntse…
masak there you go.
eternaleye Weird. Is there a limit on chaining method calls?
masak no.
you just have to do it right. :P
07:20 donaldh left 07:21 donaldh joined, justatheory left
eternaleye phenny: tell JDlugosz that the 567 quotes show as unicode-numbered quads on my system in Firefox, but that's a font issue with Serif rather than anything in the OS 07:29
phenny eternaleye: I'll pass that on when JDlugosz is around.
eternaleye phenny: tell JDlugosz (that's amd64 Gentoo, firefox 3.1beta) 07:30
phenny eternaleye: I'll pass that on when JDlugosz is around.
07:36 justatheory joined 07:37 justatheory left
Matt-W Morning 07:55
masak tips hat 07:56
Matt-W looks in the hat for tips 08:02
masak :P
I'm going away for the weekend; will be back on Sunday evening.
08:03 DemoFreak left
Matt-W I'm going away for the weekend as well 08:03
Train tonight, then a day of music tomorrow followed by a barbecue, then some awkward family time 08:04
08:04 bacek left
masak I'll be celebrating the arrival of spring with some fellow Esperanto youngsters. 08:04
Matt-W And probably a lot of pointless questioning from my parents over what's happening with my car, as if I've found out any more information during five minutes when I haven't had any phone calls 08:05
Hmm so no speaking Swedish or English allowed then 08:07
masak Matt-W: we're not very draconian. we like languages, that's all. most people do prefer Esperanto, though. 08:08
especially because it's an international context, where not everyone may feel included if Swedish or English is spoken.
Matt-W Well Esperanto is the common language, really 08:09
And the reason you're all there
masak aye.
Matt-W And a good chance to practise your spoken Esperanto with lots of people
masak exactly.
well, gotta go.
masak waves 08:10
08:10 masak left 08:14 zpmorgan_ left 08:16 jferrero joined 08:22 payload joined 08:39 dakkar joined 08:54 tulcod joined 09:04 spx2 left 09:06 jiing_ joined 09:20 payload left 09:23 bacek joined 09:24 payload joined 09:26 M_o_C joined 09:55 jan___ left 10:14 riffraff joined 10:23 icwiener joined 10:26 jrockway left, jrockway joined, ZuLuuuuuu joined 10:34 bacek left 10:36 bacek joined 10:40 payload left 10:41 payload joined 10:59 payload left 11:17 clintongormley joined 11:20 donaldh left, donaldh joined, bacek left, bacek joined 11:26 bacek_ joined, bacek_ left 11:29 ssm joined 11:37 payload joined 11:43 pmurias joined
dalek kudo: 5eac9bd | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION so we get the smarter "variable not predeclared" messages.
 11:49
11:49 bacek left 11:50 skids joined 11:59 fridim_ joined 12:02 kst left, kst joined
skids gets Nigerian spam saying it wants to compensate me for funds lost due to falling for Nigerian spam scams. 12:12
Does that mean spam has jumped the shark?
jnthn No, it just means we should send sharks to eat Nigerian scammers. 12:14
wayland76 I think we'd be better of setting the sharks on the Somali pirates, and finding a way to send something else to the Nigerian scammers. Maybe grizzly bears or elephants. No, I have it! Klingons! 12:20
skids Or Somali pirates... 12:23
12:24 jlaire joined
wayland76 Now that would be a feat :) 12:27
12:30 mizioumt joined
Matt-W Well all these scammers are in possession of vast sums of money as inheritances 12:32
So you just need to get them to put it all on a boat 12:33
And sail it past Somalia at a tempting speed
skids They are all officials or royalty, so just their own presence as valuable hostages should do :-) 12:34
Matt-W But they're even more valuable if they come with gold and jewels included, rather than having to buy them afterwards 12:35
wayland76 Hmm. So when we receive Nigerian Spam, we should say we've just gotten free cruise tickets, and we'll meet them on the boat... :)
skids Matt-W: or just safe deposit box keys that they swear contain the wealth of ages 12:36
wayland76 I love Pobox.com. I get so little Spam these days that it's not a real problem :) 12:37
Admittedly, I have to do something about the false positives from time to time :) 12:38
skids I just greylist + dnsbl, it lets enough through so I have a random sampling to amuse myself. but not enough to be burdensome. 12:39
wayland76 Yeah, greylist would be nice :)
pugs_svn r26903 | moritz++ | [t] move closure_traits/* to spec/ 12:40
r26904 | moritz++ | [t] merge regex/context.t into spec/S05-modifier/pos.t
r26905 | moritz++ | [t] remove regex test which is outdated and flawed in multiple ways
skids (spammers peform a valuable service after all, they are the keepalive packets of the mail system :-)
pugs_svn r26906 | moritz++ | [t] move spaceship.t to spec/, minor corrections and fudging for rakudo
r26907 | moritz++ | [t] move subscript.t to spec
r26908 | moritz++ | [t] move lt.t to spec/
r26909 | moritz++ | [t] improve parens.t a bit
wayland76 I receive spam, therefore I am?
clintongormley i used to host my own mail, and just bounced all mails with spam assassin score > $x
pugs_svn r26910 | moritz++ | [t] explain why syntax/signature.t is flawed
clintongormley now moved to google, and after a month, checked my junk folder: 12,000 emails!
wayland76 On the ISP I do some admin for, we reject anything that scores above 40 12:41
And filter anything that scores above 5 into a separate folder
(or we did; the rejecting above 40 part might be broken at the moment)
12:49 payload left, payload1 joined 12:52 abra joined 12:55 cspencer joined, payload1 left
Matt-W rakudo: say [1, 2] X [1, 2] 12:59
p6eval rakudo 5eac9b: OUTPUT«1 21 2␤» 13:00
13:00 payload joined
skids rakudo: say (1,2)X(1,2); my @a = (1,2); say @a X @a; 13:01
p6eval rakudo 5eac9b: OUTPUT«11122122␤11122122␤»
13:02 riffraff left
skids rakudo: say [1, 2] Z [1, 2] 13:03
p6eval rakudo 5eac9b: OUTPUT«1122␤»
PerlJam good morning #perl6 13:07
Matt-W good afternoon PerlJam
13:15 presh joined 13:18 Chillance joined 13:20 jferrero left 13:35 justatheory joined 13:36 [particle] joined 13:46 cspencer left
TimToady rakudo: say 1,2 Z 1,2 13:47
p6eval rakudo 5eac9b: OUTPUT«1122␤»
TimToady rakudo: say 1,2 X 1,2
p6eval rakudo 5eac9b: OUTPUT«11122122␤»
TimToady rakudo is correct on []X[] but not []Z[] 13:48
Matt-W that's not correct??
TimToady Z should treat [] as a single value 13:49
just as X does
Matt-W oh
TimToady list infixes expect to infix lists :) 13:50
13:50 M_o_C left 13:53 justatheory left
Matt-W there's a certain sense to that 13:54
:)
13:55 M_o_C joined
Trey What's the current idiom for P5 "while (<>) { say }" ? 13:55
PerlJam Trey: do you mean a literal <> there, or was that <> meant to stand for reading a line from any filehandle? 13:57
Trey PerlJam: Either. Both. 13:58
PerlJam .say for $filehandle.lines # maybe 14:00
TimToady for lines() {.say} is supposed to work
PerlJam (also depends on if you really meant say there too and the status of autochomping, etc.)
Trey i remember when it was for =<> { .say } and realized with = no longer iterating that that couldn't work 14:01
PerlJam TimToady: The lines sub does the equivalent of reading from STDIN or files given on the command line?
TimToady it defaults to $*ARGFILES, which does that 14:02
PerlJam gotcha
same for get() I take it?
Trey = as iterator went away why? i think that happened when my head was in the sand somewhere.... 14:03
TimToady I suppose, though we're trying to encourage list processing
we lost it somewhere in Norway 14:04
Trey makes sense. it's got a big groove there that can get caught in the ice.
TimToady after we discovered that it was really ugly to teach as the first thing new Perl programmers see :/
14:04 payload left
Trey so now any iterator in scalar context iterates. 14:05
and if you want the iterator itself...
TimToady no
$iterator.net
.get
that what $*IN.get is really 14:06
PerlJam get and lines bother me as the universal iterator interface for some reason but I don't have anything specific to propose :)
TimToady it's an on-purpose symmetry breaking
Trey ahh. i see.
Matt-W finds the idea of lines() working without an object slightly disquieting
TimToady get is a verb because it's essentially procedural 14:07
but lines is a noun because it's essentially FP
skids rakudo: sub infix:<X> (List *@slurp) { @slurp.join("#") }; ((1,2) X (3,4) X (5,6)).say # <-- why we can't just move Z and X to setting yet.
p6eval rakudo 5eac9b: OUTPUT«1#2#3#4#5#6␤»
PerlJam TimToady: I think I'd still like to see something like $handle.gimme (grab them all), and $handle.gimme(1) or something
(luckily I have the option to create that myself :) 14:08
TimToady biab
Matt-W I like $handle.lines()
Conceptually it's like $handle has everything inside it already 14:09
it doesn't matter if it's not been read from disc
skids PerlJam: just ==> @a and use the array interface. (not that that works yet) 14:11
14:12 finanalyst left 14:13 payload joined
PerlJam skids: I think I'm more opposed to having get/lines than having some alternate way to access them. It just "feels" wrong. I haven't done enough introspection on why, but it may just be because of my C background. 14:14
14:15 ssm left
skids (gdb) print _temp_multi_func_list.func_ptr 14:15
$14 = @0x40868042: 0x407aca30 <Parrot_Undef_multi_is_equal_Undef>
(gdb) print Parrot_Undef_multi_is_equal_Undef
$15 = {INTVAL (Parrot_Interp, PMC *, PMC *)} 0x407aca30 <Parrot_Undef_multi_is_equal_Undef>
Can anyone explain why there's two pointers? 14:16
PerlJam one for the undef multi and one for the "true" undef ? 14:18
:)
14:18 justatheory joined
skids now how would gdb know that? 14:18
skids can't find reference to this @ notation in gdb docs. 14:20
(but the @ value is the actual content of the member, just wondering by what mechanism 0x407aca30 is being derived. 14:21
14:22 jferrero joined 14:23 elmex left, elmex joined 14:30 alester left
skids tries to figure out WTF __JCR_LIST__ is. 14:33
14:35 aindilis left
pugs_svn r26911 | lwall++ | [pos.t] fix parsefail 14:47
14:48 payload left 14:50 exodist joined 14:53 payload joined 15:02 nihiliad joined 15:07 [particle]1 joined 15:18 pmurias left, Chillance left, pmurias joined 15:20 donaldh left, donaldh joined 15:34 Psyche^ joined, Patterner left, Psyche^ is now known as Patterner 15:39 ElectricHeavyLan joined 15:50 pmurias left 15:51 claytoris joined
claytoris Hey does anyone know if k3151y comes in here often? 15:52
15:56 claytoris left 16:15 dakkar left 16:16 alester joined 16:18 iblechbot joined 16:20 zostay joined 16:25 zostay left 16:26 zostay joined 16:28 cdarroch joined 16:37 zostay left, zostay joined 16:43 jferrero left 16:44 justatheory left 16:51 zostay left 16:53 donaldh left 16:58 zostay joined 17:20 kst left 17:21 kst joined, ssm joined 17:38 justatheory joined 17:50 eternaleye left 17:55 payload left 18:03 mizioumt1 joined 18:05 aindilis joined 18:09 payload joined 18:17 mizioumt left 18:23 meppl joined 18:27 pmurias joined 18:51 justatheory left 18:52 justatheory joined 18:53 justatheory left 18:54 mizioumt joined 18:55 justatheory joined 19:00 mizioumt2 joined
pmurias how do i create a capture in perl6 which is a copy of $capture_orig except without the first positional 19:10
19:10 mizioumt1 left 19:15 mizioumt left 19:16 icwiener_ joined 19:18 icwiener left 19:23 claytoris joined, claytoris left
jnthn pmurias: Perhaps something like \(|$capture_orig[1..*], |$capture_orig{}). 19:24
19:25 DemoFreak joined
pmurias perl6: (\foo => 1){} 19:26
rhr_ how about something like :($, |$new) := |$orig
p6eval elf 26911: OUTPUT«Undefined subroutine &GLOBAL::prefix__92 called at (eval 123) line 3.␤ at ./elf_h line 5881␤»
..pugs: ( no output )
..rakudo 5eac9b: OUTPUT«Method 'postcircumfix:{ }' not found for invocant of class 'Perl6Pair'␤»
pmurias rhr_: i think |$new would capture the whole list then 19:28
rhr_ hmm
19:28 sri_kraih left
jnthn pmurias: Yeah, I'd got the impression too that if you write |$foo in a signature you get the lot too. 19:29
PerlJam rhr_: you're binding, not copying too :) 19:31
but, I think something like :($, $copy) = |$orig; ?
maybe 19:32
rhr_ dunno, I thought the sig/cap magic only happened for binding
icwiener_ I tried to build rakudo today but it just filled my ram and swap and then aborted. rafb.net/p/6TJ30W31.html
PerlJam you can break down an array [$first, *$rest] ... how do you do the same for captures?
icwiener_ Any idea how to solve that?
jnthn icwiener_: What C compiler / platform? 19:34
icwiener_ GCC 4.3.3 on Debian GNU/Linux AMD64 19:35
jnthn Hmm, nothing so unusual then. 19:36
pmurias PerlJam: method addto (|$args ($self: @x)) {...} # quoting the synopsis
jnthn I'm not really sure. I know that it can take quite a lot of ram to compile perl6.c since it contains a massive bytecode array.
icwiener_ I have a 3rd G RAM here but I usually do not use it since it runs on lower frequency. Do you think 2GB vs. 3GB makes the difference? 19:38
Unfortunately my swap is only 500 MB since I do not need it else.
jnthn icwiener_: I've heard of it taking hundreds, but gigs is kinda...ouch.
icwiener_ Yupp.
It's like compiling Eclipse. ;)
jnthn ;-)
icwiener_ Any ide how I can get more helpful output? 19:39
*idea
jnthn You can try giving it more RAM, but I'm surprised it's using quite that much...
There are efforts underway to generate the .c file differently.
So that it won't require so much RAM>
But different compilers get upset over different things, so it's proved non-trivial... 19:40
icwiener_ I understand.
jnthn Did you build with the --gen-parrot option?
icwiener_ Yes. 19:41
jnthn If you just want to turn Perl 6 and play with it a bit, you can still invoke Parrot with the bytecode file.
Which in that case would look I think like
./parrot/parrot perl6.pbc
Basically the .exe we build is a loader with the bytecode embedded.
erm, s/trun/run/ 19:42
icwiener_ Let's see if I can do that. :) Thanks for your help. 19:43
19:45 icwiener_ left
[particle]1 i wonder if the trouble is only on x64 bits 19:45
due to larger int size doubling the memory requirements 19:46
jnthn [particle]1: Even so, consuming 2 gigs + swap is nuts! 19:49
(OK, sure, it didn't have the whole 2 gigs, but still...)
pmichaud good afternoon, #perl6 19:56
Tene hi pm 19:58
jnthn hi pm
20:00 [particle]1 is now known as [particle]-
pmichaud rakudo: sub prime($x) { $x == 3 }; say (3 when prime); 20:08
p6eval rakudo 5eac9b: OUTPUT«too few arguments passed (0) - 1 params expected␤in sub prime (/tmp/A5vIHVzFMI:1)␤called from Main (/tmp/A5vIHVzFMI:1)␤»
TimToady rakudo: sub prime() { $CALLER::_ == 3 }; say (3 when prime); 20:10
p6eval rakudo 5eac9b: OUTPUT«Use of uninitialized value␤␤»
TimToady say (3 if 1) 20:11
rakudo: say (3 if 1)
p6eval rakudo 5eac9b: OUTPUT«3␤»
pmichaud afaik, CALLER:: not implemented in rakudo yet. 20:13
jnthn OK, S12 question (oh hai TimToady...) 20:15
In Construction and Initialization
It says:
If the candidate is omitted, a candidate object is implicitly created in
the current class by calling C<CREATE>:
$object = $class.bless(k1 => $v1, k2 => $v2, ...)
pmichaud afk for me -- bbl
jnthn And you can also call bless with the candidate you want blessing, having previously called CREATE. 20:16
However, afaict you can also pass to bless positional arguments, like Animal{ :blood<warm> }
I'd a little fearful about potential for confusion of the first positional.
That is, is it the candidate, or is it just one of the auto-vivifying protos? 20:17
20:21 sjohnson joined
sjohnson hello, anyone awake and wants to chat about perl? 20:22
20:23 M_o_C left
jnthn sjohnson: For some definition of awake. :-) 20:29
sjohnson is it possible to submit something to the developers of perl for consideration into the new Perl 6 spec? 20:33
jnthn A lot of Perl 6 developers hang out here (well, mabye not loads at this exactly point in time... ;-)) 20:34
There's also the perl6-language mailing list.
What were you thinking of?
TimToady jnthn: yes, it's probably ambiguous, and the candidate needs to be required, perhaps with * as a you-pick value
jnthn TimToady: Sounds sane, thanks. 20:35
TimToady sjohnson: for most such suggestions, there's a good chance it has already been considered at some point :)
jnthn TimToady: One of us should whack that into the spec. 20:36
20:36 gbacon joined
TimToady I can't right now, but if you don't I'll try to remember 20:36
jnthn TimToady: I'll have a crack at it. 20:37
Just trying to tidy that lot up in Rakudo.
sjohnson ok i will tell my ideas 20:38
do you think Mr. Wall would consider a whitespace trim function like in every single other language, instead of 1) having to depend on CPAN or 2) having to write your own regex function or 3) doing the raw regex each time yourself
TimToady it's already there 20:39
sjohnson in perl 6 or perl 5?
TimToady perl 6 20:40
S32-setting-library/Str.pod:=item trim
jnthn rakudo: my $x = " abc "; say $x.trim;
p6eval rakudo 5eac9b: OUTPUT«abc␤»
jnthn sjohnson: Implemented in Rakudo even. :-)
sjohnson YESSSSSSSSSSSSSSSSSSs 20:41
!!!!
TimToady told you we might have considered it already :)
sjohnson i am very happy to ehar that
[particle]- (Mr. Wall)++
TimToady biab, will backlog 20:43
20:44 alester left
sjohnson will Perl 6 let you do: if (1) return 1; .... or give you a syntax error 20:45
jnthn You must have the curly braces. 20:46
return 1 if cond; # is fine though
sjohnson ahh so perl 6 will still make the curly braces necessary for that one statement eh.. :( 20:48
20:48 tulcod left
jnthn sjohnson: Dangling syntax considered nasty. 20:49
sjohnson: On the up side, the parens around the conditional are optional.
if $x == 42 { say "the answer" } 20:50
20:50 tulcod joined
sjohnson thats kind of nice 20:51
as i hated those too
but i can live with just writing it backwards
do something if 1;
tho in perl 5 you dont need parens when it's written like that either which is nice 20:52
jnthn Yes, true.
sjohnson but that is still good news to me 20:53
can you edit strings by doing things like $string[3] = 'a'; 20:57
in perl 6?
jnthn sjohnson: No, not for a string, but there's a type called Buf (short for buffer) that lets you do that kind of thing. 20:59
So far as I know anyway.
One thing to note is that Perl 6 is very careful in the whole area of units. 21:00
"length" is kinda a dirty word. :-)
sjohnson length($string) will behave a bit differently?
jnthn length is gone.
You gotta say what you want.
chars($string) # characters
japhb rakudo: say %*VM.perl
jnthn bytes($string) # how many bytes
p6eval rakudo 5eac9b: OUTPUT«{"config" => {"-falign-functions=16" => "1", "-funit-at-a-time" => "1", "-fvisibility=hidden" => "1", "-maccumulate-outgoing-args" => "1", "-W" => "1", "-Waggregate-return" => "1", "-Wall" => "1", "-Wbad-function-cast" => "1", "-Wc++-compat" => "1", "-Wcast-align" => "1", 21:01
.."-Wcast-q…
jnthn Note you can call 'em as methods too.
$string.chars
japhb ouch, way too long.
jnthn wow.
japhb Can .perl be told to pretty-print? 21:02
sjohnson thats not too bad 21:03
jnthn japhb: No, but try .fmt
21:05 pmurias left
japhb jnthn: ooh, hash fmt works quite well, thanks. 21:05
jnthn: is there any way to get Rakudo's version and commit string, as you can for Parrot's using %*VM? 21:07
sjohnson will Perl 6 contain a switch / case structure?
jnthn japhb: Not that I'm aware of. 21:08
japhb sjohnson: given/when.
sjohnson: and it's quite powerful.
sjohnson wow i am excited
japhb sjohnson: also exists in a weaker version in 5.10, btw.
jnthn: awww. :-)
21:08 PhatEddy joined
sjohnson japhb: Perl has a sort() for arrays without the need for cpan. Will a shuffle() be available for Perl 6? 21:09
skids sjohnson: it's called given/where 21:10
And the given is in a lot of cases optional.
japhb sjohnson: yup. It's called pick
skids (where works on the topic, i.e. $_)
japhb rakudo: say <1 2 3 4 5>.pick(*)
p6eval rakudo 5eac9b: OUTPUT«53241␤»
japhb rakudo: say <1 2 3 4 5>.pick(2)
p6eval rakudo 5eac9b: OUTPUT«35␤»
jnthn rakudo: say "Perl 6 is {<awesome cool beautiful twisted>.pick}" 21:11
p6eval rakudo 5eac9b: OUTPUT«Perl 6 is twisted␤»
PhatEddy jnthn: I can vouch for the other fellow's experience with requiring a few gig of memory to build a perl6 executable with 64 bit linux.
skids oops. right. when. I must be too "constrained" today.
jnthn payload: Really?
PhatEddy I had the same problem with 64 bit ubuntu.
jnthn oops
japhb sjohnson: skids is referring to the fact that 'given' merely sets the topic ($_). If you're getting the topic set another way, you don't actually need the given. Just do when clauses.
jnthn PhatEddy: Wow.
japhb skids: 'where' is for subtypes. 21:12
sjohnson there's a shufflefunction called pick() to shuffle arrays? that's an odd name 21:13
japhb sjohnson: pick is the more general function.
Tene jnthn: yes, I cna also verify. Is this not already-known?
japhb shuffle is just the special case of pick on the entire array.
21:13 abra left
Tene I thought it was well-known, or something. 21:13
jnthn sjohnson: .pick without an argument gets one item.
skids rakudo: for <a b c> { when <c> { say "OH HAI" }; default { say "OH" } } 21:14
p6eval rakudo 5eac9b: OUTPUT«OH␤OH␤OH HAI␤»
jnthn Tene: That it took gigs? Maybe known to others, that sounded an awful lot to me.
Tene ah
jnthn Tene: I thought in the hundreds.
Tene ah, OK
PhatEddy I had two gig of ram and swap disabled. It wouldn't build until I re-enabled the 4 gig swap. I think the system used 800 meg and the build required 1.5Gig to 2 gig I think. Is this an RT kind of thing?
Tene I sometimes forget that I have more ram than I usually know what to do with.
jnthn But if that's normal, fair enough...I just wasn't aware it was *that* bad. :-|
Tene (although lately firefox has been finding creative uses...) 21:15
(a web browser should never need 1.5G :( )
sjohnson jnthn: does it remove it out of the array? im worried about "picking" twice
jnthn sjohnson: No, it doesn't.
skids Tene: close those windows with flash in them, they waste electricity :-)
Tene skids: this is with no flash at all. 21:16
jnthn sjohnson: You can pass a number to .pick to say how many you want.
Tene flash uses another 600M or so
jnthn sjohnson: If you to .pick(*) it means "all of them"
When you ask for more than one, though, the order can vary.
(as in, is random) 21:17
japhb sjohnson: both pick-without-replace and pick-with-replace are available. It's just an adverb on the pick method. 21:18
jnthn Ooh, I'd forgot about pick with replace. :-)
skids Tene: yeah well, mostly it would just be nice if people used AJAX loops that knew how to sleep. but too OT.
sjohnson Q: is there a uniq(@array); function too?
japhb rakudo: say <1 2 3 2 5 6 4 3>.uniq.perl 21:19
p6eval rakudo 5eac9b: OUTPUT«["1", "2", "3", "5", "6", "4"]␤»
japhb yup
:-)
And before you even ask, note the '.perl' method there. It's Data::Dumper on steroids, and built in. 21:20
Man, being able to say "already done" over and over is really nice. 21:21
sjohnson wow on irc parser eh
rakudo: print "hello sjohnson";
p6eval rakudo 5eac9b: OUTPUT«hello sjohnson»
skids jnthn: WRT memory usage I think there was a misunderestimation on just how many really small hash/array/strings would be created. The out is to make all those guys store small values directly in the PMC. 21:22
sjohnson japhb: is what you wrote a real perl6 syntax?
i dont think you can do that in perl 5... using dots after things like that. am i wrong / right? 21:23
japhb sjohnson: Because Perl 6 is a spec/testsuite with a number of implementations, the evalbot can actually show you the current response from each known implementation of Perl 6.
jnthn sjohnson: When you prefix something with rakudo: here, it takes it and compiles/runs it.
Yes, rakudo is just one implementation of Perl 6.
japhb sjohnson: Yep. $object.foo is roughly $object->foo in Perl 5.
jnthn sjohnson: The . is the Perl 6 method call operator.
ETOOMANYEXPLAINERS :-)
japhb There can never be too many. 21:24
sjohnson fuck i cant wait for perl 6
japhb There Must Not Be Only One
jnthn japhb: Oh yes, I agree. :-)
japhb sjohnson: It's here. Just not 100% implemented yet.
sjohnson print "happy "house".trim;
rakudo: print " happy house \n".trim;
p6eval rakudo 5eac9b: OUTPUT«happy house» 21:25
skids jnthn: That E'1atonce
sjohnson fuck i am in love
sorry for the course language
but all of this is WAY better than perl 5
japhb It's a strong emotion, we understand. ;-)
jnthn sjohnson: There's even more cool stuff. :-)
japhb Like this:
rakudo: print "1 and 2 is {1 + 2}"
skids sjohnson: rumor has it perl6 might swear occasionally too :-)
p6eval rakudo 5eac9b: OUTPUT«1 and 2 is 3»
21:26 Whiteknight joined
sjohnson soi am impressed 21:29
say <1 2 3>.class 21:30
rakudo: say <1 2 3>.class
p6eval rakudo 5eac9b: OUTPUT«Method 'class' not found for invocant of class 'List'␤»
sjohnson rakudo: " moose".class
p6eval rakudo 5eac9b: OUTPUT«Method 'class' not found for invocant of class 'Str'␤»
japhb rakduo: say <1 2 3>.WHAT
rakudo: say <1 2 3>.WHAT
p6eval rakudo 5eac9b: OUTPUT«List()␤»
skids rakudo: say <1 2 3>.WHAT
p6eval rakudo 5eac9b: OUTPUT«List()␤»
skids Most of the "introspection" stuff is allcaps methods like WHAT HOW etc. 21:31
japhb rakudo: say <1 2 3>.^methods
p6eval rakudo 5eac9b: OUTPUT«invoke() not implemented in class 'ResizablePMCArray'␤in Main (src/gen_setting.pm:3166)␤»
sjohnson rakudo: say 3.WHAT
p6eval rakudo 5eac9b: OUTPUT«Int()␤»
sjohnson wow it's like ruby
jnthn japhb: .^methods is kinda screwed up/work in progress atm. :-|
japhb jnthn: I figured I had a 50-50 shot. ;-) 21:32
jnthn japhb: Your luck varies by type. ;-)
japhb heh
jnthn japhb: It's on my todo list.
[particle]- 'j' ~ <a-z>.pick(4)
jnthn japhb: As is introspection generally.
sjohnson rakudo: say <1 2 3>.shuffle 21:33
p6eval rakudo 5eac9b: OUTPUT«Method 'shuffle' not found for invocant of class 'List'␤»
sjohnson damn
japhb rakudo: class Foo { method foo {...}; method bar {...} } my Foo $a .= new; say Foo.^methods
p6eval rakudo 5eac9b: OUTPUT«Statement not terminated properly at line 1, near "my Foo $a "␤in Main (src/gen_setting.pm:0)␤»
jnthn japhb: missing semi
[particle]- rakudo: print [~] 'j', <a..z>.pick(4)
p6eval rakudo 5eac9b: OUTPUT«ja..z»
japhb sjohnson: you want '.pick(*)' instead of '.shuffle'
rakudo: class Foo { method foo {...}; method bar {...} }; my Foo $a .= new; say Foo.^methods;
[particle]- rakudo: print [~] 'j', ['a'..'z'].pick(4) 21:34
p6eval rakudo 5eac9b: OUTPUT«foobar␤»
rakudo 5eac9b: OUTPUT«jwytg»
[particle]- finally got it.
japhb rakudo: class Foo { method foo {...}; method bar {...} }; my Foo $a .= new; say Foo.^methods.perl;
p6eval rakudo 5eac9b: OUTPUT«[{ ... }, { ... }]␤»
japhb oooh
sjohnson rakudo: print <a..z>.pick(*)
jnthn japhb: methods doesn't return the names, but the actual methods.
p6eval rakudo 5eac9b: OUTPUT«a..z»
skids [particle]-: it's like simon says ;-)
japhb jnthn: yep, I forgot that.
jnthn rakudo: class Foo { method foo {...}; method bar {...} }; Foo.^methods>>.name>>.say 21:35
p6eval rakudo 5eac9b: OUTPUT«foo␤bar␤»
sjohnson how come my pick(*) didnt work?
jnthn sjohnson: Because <...> is like qw in Perl 5 21:36
a..z is just one string
sjohnson rakudo: say <1 2 3 4 5 6 7 8 9 10>.pick(*)
oh i thought it was a range
p6eval rakudo 5eac9b: OUTPUT«67513910284␤»
jnthn rakudo: ('a'..'z').pick(*)
p6eval rakudo 5eac9b: ( no output )
jnthn rakudo: say ('a'..'z').pick(*) 21:37
japhb sjohnson: Of course, you can do what jnthn just said ...
p6eval rakudo 5eac9b: OUTPUT«gwzltucsriekjdymaxpbnfohvq␤»
jnthn
.oO( if you do that enough times, some of the resulting strings will contains swears! )
japhb clbuttic.
sjohnson will i be made fun of if i say i'm 10x as excited about Perl 6 now than I was 10 minutes ago? 21:38
japhb Nope. We'll merely suggest you get a paper bag to breathe into. :-)
skids rakudo: (1, -2, 3, 4)>>.abs.say # dare we reveal hypers this soon?
p6eval rakudo 5eac9b: OUTPUT«1234␤»
sjohnson all this stuff larry wall decided should go into Perl 6? 21:39
japhb sjohnson: and more. How about this one:
rakudo: (<a b> X <1 2>).say
p6eval rakudo 5eac9b: OUTPUT«a1a2b1b2␤»
[particle]- sjohnson: it's a community rewrite of perl, not (just) larry's rewrite 21:40
sjohnson [particle]-: am i too late to suggest other things?
[particle]- no 21:41
skids Go for it, we want to see how far the "already there" score can get today :-)
[particle]- the perl6-language mailing list is one place, and here is another.
jnthn Gah. I'm waaay to butterfingered tonight to be refactoring object init code... 21:43
skids jnthn: then work on destruction :-)
sjohnson well let me think of another one 21:44
japhb Having just seen HTTP::Daemon leaking all over the floor, I'm all for better destruction. ;-)
Of course, it may be a Parrot issue.
jnthn japhb: It has a memory leak?
japhb: Quite believable, I'm curious if it's notably more leaky than any other long-running Perl 6 program though. 21:45
[particle]- most perl 6 programs are long-running these days 21:47
skids ;-)
jnthn [particle]-: :-P
sjohnson $string = "moose"; $string[2] = "O"; # "moose" => "moOse"; 21:48
how would that be done in perl 6 properly?
skids my Buf $b := $string; $b[2]="0"; 21:49
I think.
Buf isn't quite fleshed out yet though.
sjohnson in Perl 6, can there be a qx/ / without $string interpolation? 21:51
21:51 justatheory left
japhb jnthn: Yes, I ran a thousand hits of the sample httpd root page, and it ate ~ half a gig. 21:52
jnthn: I'll be looking at it in more detail in a bit
sjohnson: yes, there are both interpolating and non-interpolating versions of qx. 21:53
skids sjohnson: there's just about any quoting form you could desire, Q: takes many adverbs that can be mixed together.
jnthn japhb: That is srsly ouch.
sjohnson my brother asks: can you declare variables without dollar signs in Perl 6? 21:55
skids hahah! 21:56
[particle]- rakudo: constant pi = 3; print pi;
p6eval rakudo 5eac9b: OUTPUT«invoke() not implemented in class 'Integer'␤in Main (/tmp/EhsQtdYtpZ:1)␤»
[particle]- feh.
rakudo: constant pi := 3; print pi;
p6eval rakudo 5eac9b: OUTPUT«No applicable methods.␤in Main (/tmp/4gGeTI4F6n:1)␤»
sjohnson skids: i take it no, i dont mind tho 21:57
i dont mind the @ and $'s
skids In perl6, though, you don't change @ to $ when using [], though.
sjohnson what do you mean 21:58
i dont remember doing that in perl 5 even
oh
you mean like $list[5];
what do you do instead?
skids in perl5 @a = (1,2,3); $a[1];
rakudo: my @a = (1,2,3); @a[1].say 21:59
p6eval rakudo 5eac9b: OUTPUT«2␤»
sjohnson skids: will say @a[1]; work too? 22:00
pugs_svn r26912 | jnthn++ | [spec] We now always require the candidate to be passed to bless, otherwise there's potential for confusion with the first auto-vivifying type object.
jnthn sjohnson: ye
s
sjohnson what about hashes and stuff? 22:01
my brother hates having to do all the ${@{%hash{list}}} shit
i dont like it either much
ZuLuuuuuu sigil doesn't change anymore taking an element from array or hash
sjohnson any improvements on that?
skids Same general deal -- the parens determine how you expect it to behave, the $@% are part of the variable name.
ZuLuuuuuu *when taking element from ...
sjohnson but @a[1]; didnt really change its behaviour 22:02
even tho i said list context with the @ snail sign
jnthn sjohnson: The sigils are here to stay, they're just invariant now. That is, part of the name. 22:03
sjohnson: Additionally, the enforce an "interface contract".
*they...
sjohnson can i see some examples? 22:06
that make it easier in perl 6 jnthn?
jnthn Perl 5: my @a = (1,2,3); print $a[1]; 22:07
Perl 6: my @a = 1,2,3; print @a[1];
sjohnson what about... hash things?
jnthn $a and @a are always separate variables.
Same rule.
sjohnson it just knows what you want to do?
jnthn my %h = a => 1, b => 2; say %h<a> 22:08
skids Hrm, I haven't had to use many multi-level data structures due to better control flow, come to think of it.
jnthn You can do [1] and <a> and {'a'} on any variable you want.
But @a promises you that [...] will always work, for example. 22:09
You can put something into a variable with an @ sigil that can't be indexed into positionally.
skids perl5: $a = +{ a => [1,2,3], b => [4,5,6] }; print $a->{b}->[0] 22:11
$a = +{ a => [1,2,3], b => [4,5,6] }; print $a{b}[0]
rakudo: $a = +{ a => [1,2,3], b => [4,5,6] }; print $a{b}[0]
p6eval rakudo 5eac9b: OUTPUT«Symbol '$a' not predeclared in <anonymous> (/tmp/uLlQqlWcox:1)␤in Main (src/gen_setting.pm:3166)␤» 22:12
skids rakudo: my $a = +{ a => [1,2,3], b => [4,5,6] }; print $a{b}[0]
p6eval rakudo 5eac9b: OUTPUT«Could not find non-existent sub b␤»
skids rakudo: my $a = { a => [1,2,3], b => [4,5,6] }; print $a{b}[0]
p6eval rakudo 5eac9b: OUTPUT«Could not find non-existent sub b␤»
jnthn skids: <b>
{...} never auto-qoutes in Perl 6.
skids Yeah I just realized Iwas pasting the perl5 line back in.
:-)
jnthn :-=)
skids rakudo: my $a = { a => [1,2,3], b => [4,5,6]}; $a<b>[0].say 22:13
p6eval rakudo 5eac9b: OUTPUT«4␤»
skids helps if you cut the right text :-)
sjohnson rakudo: say { 1, 2, 3}.WHAT 22:14
p6eval rakudo 5eac9b: OUTPUT«Block()␤»
sjohnson perl6: say <1>.WHAT 22:15
skids rakudo: say { 1 => 2, 3 => 4 }.WHAT
p6eval pugs: OUTPUT«Str␤»
..elf 26912: OUTPUT«Array␤»
..rakudo 5eac9b: OUTPUT«Str()␤»
rakudo 5eac9b: OUTPUT«Hash()␤»
skids (mines the last one)
pugs_svn r26913 | jnthn++ | [t/spec] Spec change means bless always needs a candidate now; update construction.t. 22:16
22:16 smtms left
jnthn y'know, I think my patch is - after tweaking that spectest - actually going to pass make spectest. 22:16
Too bad it's only doing the first easy bit of what I eventually want to do. 22:17
22:17 smtms joined
skids rakudo: my %h = :a(1), :b(2); %h.perl.say 22:18
p6eval rakudo 5eac9b: OUTPUT«{"a" => 1, "b" => 2}␤»
sjohnson how will one learn about all the tricks in Perl 6 that are new, if one has already purchased Programming Perl (v5)
22:18 justatheory joined
skids sjohnson: for the meantime, watch changelogs on pugs cvs. 22:19
jnthn s/cvs/svn/ # thankfully!
sjohnson: There are various places to follow Perl 6 news, and at least one documentation project.
skids embarrassed at showing his age, nervously sweeps hair over bald spot. 22:20
jnthn Plus this place is pretty good for questions. :-)
22:21 nihiliad left
japhb sjohnson: Just in case it wasn't clear earlier, [particle] was trying to demonstrate that according to spec, you can declare *constants* without a sigil, but it isn't working yet everywhere. 22:21
sjohnson without using "use constant ANIMAL => 'goose';" ? 22:22
japhb jnthn: Yeah, that leak is pretty bad. I am going over the HTTP::Daemon source looking for any obvious places the Perl 6 code does the wrong thing, and then I'll probably have to look deeper. :-/ 22:23
skids sjohnson: there are enums also.
22:23 tulcod left
japhb sjohnson: 'constant' and 'enum' are now declarators, there's no module to use. 22:24
dalek kudo: 23718a8 | jnthn++ | src/classes/Object.pir:
First steps in refactoring object building/initialization. We now are a bit more consistent with the spec about new/bless/CREATE and their relationship, plus follow a recent spec change. Plus adds placeholders for the next step of refactoring creation and BUILD.
japhb (They're in the core syntax, I mean)
jnthn rakudo: constant foo = 42; say foo 22:25
p6eval rakudo 5eac9b: OUTPUT«42␤»
japhb Go jnthn!
jnthn rakudo: constant π = 3.14; say π
p6eval rakudo 5eac9b: OUTPUT«3.14␤»
japhb sjohnson: note the unicode ...
jnthn What did particle try earlier that didn't work... :-S
japhb "rakudo: constant pi = 3; print pi;" 22:26
jnthn rakudo: say pi
p6eval rakudo 5eac9b: OUTPUT«3.14159265358979␤»
jnthn Oh!
We already have a pi 22:27
japhb Builtin!
jnthn And it's not detecting the collision.
Bad Rakudo. No cookie.
22:27 justatheory left
japhb ("It's easy to detect collision with pi -- just check if your face is covered in whipped cream ...") 22:27
jnthn Aww. You just made me want pi(e)... 22:28
skids (disclaimer: does not work in case of french meat pie)
japhb heh 22:30
jnthn Ah well, if I can't have pie I'll have pivo. :-) 22:33
sjohnson what are enums 22:34
japhb watches as Rakudo valiantly tries to overtake Firefox as biggest process on my laptop ...
sjohnson "true", "false" type datatypes?
pugs_svn r26914 | jnthn++ | [spec] Oops, forgot to save before commit.
japhb sjohnson: that is one example, yes.
skids uses firefox to horde cache to check rakudo misses. 22:35
japhb sjohnson: days of the week, months of the year, that sort of thing
skids rakudo: say (2.296,0.5671,"pork") >>~>> " the other " >>~<< (pi,e,"white meat") >>~>> "\n"
p6eval rakudo 23718a: OUTPUT«2.296 the other 3.14159265358979␤0.5671 the other 2.71828182845905␤pork the other white meat␤␤»
jnthn rakudo: enum Day <Mon Tue Wed Thu Fri Sat Sun>; given Day.pick { when Mon { say "yawn" }; when Tue|Wed|Thu { say "work work work" }; when Fri { say "oh phew it's nearly over" }; when Sat { say "off t'pub" }; when Sun { say "good morning, vicar" } } 22:36
p6eval rakudo 23718a: ( no output )
japhb A classic example, that ....
jnthn erm... 22:37
why no output though. :-|
japhb Rakudo wins by a nose!
jnthn oh well, I blogged the working one somewhere...
japhb rakudo: ((436 * 1024 - 69308)/1000).say 22:39
p6eval rakudo 23718a: OUTPUT«377.156␤»
japhb ouch. Leaking a bit under 400 kB per hit.
jnthn ouch.
sjohnson the squiggle operator is new to me 22:40
>> ~ << etc
jnthn sjohnson: stirng concat
japhb sjohnson: since Perl 6 uses . for method calls, we needed something else for concatenation
jnthn rakudo: say "oh" ~ "ai";
p6eval rakudo 23718a: OUTPUT«ohai␤»
japhb sjohnson: in general, there was a massive reshuffle of operators, as A became B and B became C and so on. 22:41
skids and =~ is now ~~
japhb skids: ~~ is how you say 'smart match' in Perl 5.10, too. Perl 5 =~ was regex match only. 22:42
skids Oh I never learned much of .10 new stuff.
22:42 dalek left 22:43 dalek joined
japhb skids: 5.10 gives me heartache. Because it's SO much better than 5.8, but it's so not 6.0. 22:43
skids sjohnson: and the >> stuff are hyperoperators, which are very interesting.
japhb Oh! We forgot reduction! 22:44
22:44 jferrero joined
japhb rakudo: say [+] 1..10; 22:44
p6eval rakudo 23718a: OUTPUT«55␤»
skids Yes reduction before hypers :-)
japhb rakudo: say [\+] 1..10
p6eval rakudo 23718a: OUTPUT«say requires an argument at line 1, near " [\\+] 1..1"␤in Main (src/gen_setting.pm:2400)␤»
jnthn japhb: didn't do tirangles yet
japhb rakudo: ([\+] 1..10).say
p6eval rakudo 23718a: OUTPUT«Syntax error at line 1, near "([\\+] 1..1"␤in Main (src/gen_setting.pm:0)␤»
japhb jnthn: dang
jnthn japhb: Hey, we handle most of the rest. :-P 22:45
:-)
japhb true
skids
.oO(tirangle: n. the half of the town square closed off during dictatorship)
jnthn
.oO( why on earth is this krušovice so insanely fizzy?! )
japhb sees comment '# inefficient workaround - remove when Rakudo gets a qx operator', complies ... 22:48
sjohnson do you guys in general, prefer Perl over Python? 22:49
reason i ask is because the Perl fans seem to live in caves / don't post their thoughts on the debate 22:50
it would be refreshing to hear some opinions that they like Perl better for once
japhb sjohnson: It just doesn't seem a debate worth fighting. Especially since Parrot allows both to run in the same process and commingle nicely. 22:51
jnthn I can't speak for everyone, but on my part getting into "my language is better" debates always feels a tad pointless. For one because part of language preference is just what people personally like too. 22:52
I'm happy to point out factual inaccuracies if I spot them and it looks like a genuine mis-understanding rather than a troll.
wayland76 sjohnson: Yes, I prefer Perl over Python. But I mostly agree with jnthn about these debates being somewhat pointless.
sjohnson not so much a debate i would like to see, but someone sincerely admit that they prefer Perl over Python ... with that person having worked with both
wayland76 Let me give an example of pointfulness, though
22:53 JDlugosz joined
wayland76 There's a book by Raphael Finkel called "Advanced Programming Language Design", and it talks about the criteria that people may or may not use for judging a language 22:53
Perl 5 happily violates almost all the principles mentioned in order to maximise one principle -- expressiveness 22:54
JDlugosz I've read books like that from a more classic era. Things were more interesting then, I think. Now they just study Java and C.
lambdabot JDlugosz: You have 1 new message. '/msg lambdabot @messages' to read it.
phenny JDlugosz: 07:29Z <eternaleye> tell JDlugosz that the 567 quotes show as unicode-numbered quads on my system in Firefox, but that's a font issue with Serif rather than anything in the OS
JDlugosz: 07:30Z <eternaleye> tell JDlugosz (that's amd64 Gentoo, firefox 3.1beta)
sjohnson wayland76: i c
wayland76 That aligns with my preferences, so I like Perl better than Python 22:55
JDlugosz: Well, the Finkel one actually means "Advanced". This is "What to do next after you've done a basic Compilers course"
(Oh, the opinion that Perl maximises expressiveness is mine, naturally, and didn't come from the book) 22:56
Btw, the Finkel book is free online as a PDF, but I liked it so much that I bought it anyway
JDlugosz Algol, PL/1, LISP, SnoBol, REXX, ... 22:58
Got a link for the PDF? Like catnip...
japhb jnthn: Is the time API implemented in Rakudo yet? 22:59
jnthn japhb: time() function will work
japhb jnthn: but none of the methods?
jnthn japhb: The Temporal module work my mberends++ isn't integrated yet. 23:00
japhb OK
jnthn japhb: When he tried, the stage 1 compiler exploded. Then I spent two fruitless hours trying to work out why. :-(
JDlugosz ➸ I'm thinking of a function signature.
japhb nice
jnthn japhb: So that's on my crap-I-have-to-be-bothered-to-debug-but-gah-it's-hard list. :-) 23:01
JDlugosz DId I miss the boat somewhere maybe five years ago? Consider: sub foo ($x, @y, @z) { ... }
japhb jnthn: not exactly a fun list to work through, I bet ....
sjohnson rakudo: $x = { if (0 == 1) { 5; } else { 6; } }; say $x = { if (0 == 1) { 5; } else { 6; } }; say $x;
jnthn japhb: No, there's the fun times in implemetning Perl 6, and the painful ones.
JDlugosz The books of that era always covered stuff like that. I learned much more than any "modern" guy. 23:02
p6eval rakudo 23718a: OUTPUT«Symbol '$x' not predeclared in <anonymous> (/tmp/1nw1yawVgA:1)␤in Main (src/gen_setting.pm:3166)␤»
jnthn sjohnson: You need to declare your variables. :-)
JDlugosz: I'm looking over your article at the moment.
sjohnson rakudo: $x = { if (0 == 1) { 5; } else { 6; } }; say $x;
p6eval rakudo 23718a: OUTPUT«Symbol '$x' not predeclared in <anonymous> (/tmp/ZIlE0YH9jY:1)␤in Main (src/gen_setting.pm:3166)␤»
jnthn JDlugosz: The one you just sent mail to p6l about.
sjohnson rakudo: my $x = { if (0 == 1) { 5; } else { 6; } }; say $x;
japhb sjohnson: 'my $x = ...'
p6eval rakudo 23718a: OUTPUT«_block21␤»
JDlugosz jnthn: The thing I just posted, or the one we've been discssing all week?
jnthn rakudo: my $x = { if (0 == 1) { 5; } else { 6; } }; say $x();
p6eval rakudo 23718a: OUTPUT«6␤» 23:03
jnthn JDlugosz: The one you just posted
JDlugosz jnthn, thanks.
jnthn www.dlugosz.com/Perl6/web/passing_examples.html
JDlugosz: I need to find time to look at more of your work...
sjohnson jnthn: do you have to do $x();?
JDlugosz I need to update most of those short pieces anyway.
jnthn sjohnson: If you want to run the block, yes.
JDlugosz: Your surname looks Polish - is it? 23:04
JDlugosz I think you are missing a "do" there. The immediate block gets executed, not saved as a &. It is a statement. Needs 'do' to get the value back into an expression.
jnthn JDlugosz: It's not an immediate block there, it's a closure. 23:05
JDlugosz Yes, Name is Polish. Actually spelled with an Ł
sjohnson rakudo: $x = { if (0 == 1) { 5; } else { 6; } }() ; say $x;
jnthn Ah, OK. :-)
p6eval rakudo 23718a: OUTPUT«Symbol '$x' not predeclared in <anonymous> (/tmp/iI3mbBC2L1:1)␤in Main (src/gen_setting.pm:3166)␤»
sjohnson rakudo: my $x = { if (0 == 1) { 5; } else { 6; } }() ; say $x;
p6eval rakudo 23718a: OUTPUT«6␤»
JDlugosz What makes it a closure instead?
jnthn JDlugosz: I'm living in Slovakia at the moment, so I run into the odd bit of Polish now and then. :-)
In general { ... } gives a closure, I guess it's the context that makes it "immediate" in some sense. 23:06
skids JDlugosz: it's not where a statement was expected. 23:07
my $x = -> $a, $b { if ($a == $b) { 5; } else { 6; } }; say $x(3,3); say $x(3,4)
rakudo: my $x = -> $a, $b { if ($a == $b) { 5; } else { 6; } }; say $x(3,3); say $x(3,4)
jnthn skids: Ah, that sounds about correct.
p6eval rakudo 23718a: OUTPUT«5␤6␤»
jnthn So I guess it's a syntactic decision.
JDlugosz Yes, the context. I'm reviewing S06 now... guess it's not as simple as I recall it saying.
Probably right, really executes if it becomes a statement itself. So the grammar will give complex behavior in general, but the common cases are designed to DWIM. 23:09
jnthn JDlugosz: You write "Coerce the Positional role out of α,"
It's really "check that α does Positional"
The sigil on the parameter implies that a certain role is done.
wayland76 JDlugosz: www.nondot.org/sabre/Mirrored/AdvProgLangDesign/ 23:10
But it would've been impossible to find if I hadn't looked up the filename on my local computer, and googled that :)
JDlugosz It's my upcoming formal language that I'll use later, trying out now... that's what I mean here. I'm preparing a full treatment of coersions. 23:11
Deals with iheriting from roles and the problems with "Isa" and derived types not being proper subtypes. That kind of thing.
jnthn OK, but to me coercion has always meant something very different from a type check.
JDlugosz That's my pet from last year: my sabbatical research was on F-bounds polymorphism. On my web page. 23:12
term coersion: I'm running with what I saw used already. Maybe needs a different word for the general case?
I've come up with "identity coersion" is one that returns the same object or a conforming proxy to the one it was called on.
where conforming means something a page or two long, too. 23:13
I'll hit the thesarus. But the Synopses mention Conversion and Coersion.
and doesn't distinguish checking for an interface on an object (static type conversion, dynamic type same). 23:14
Suggestions welcome.
jnthn Sure.
JDlugosz Thanks for the PDF.
jnthn I've implemented much of the Perl 6 type stuff in Rakudo.
So I'm kinda coming from the angle of, how that implementation looks. 23:15
JDlugosz At the very least, I should hyperlink the term. But this page wasn't meant to be shown yet. Just want to get past my issue before going on.
jnthn www.dlugosz.com/Perl6/web/typesyste...mmary.html is probably an interesting read for me too... 23:16
JDlugosz Yea, I need to hear from you on "what really happens" since the spec is too vague, but then come up with something that's not implementation-specific to re-describe it.
jnthn The spec on roles has got a tad less vague of late. :-)
JDlugosz typesystem-summary: more lamenting on what's not in the spec. 23:17
good to here.
jnthn Yeah, on the type stuff the spec is a bit hand-wavey for sure.
JDlugosz Also last year I figured out a complete algorithm for role composition.
jnthn A lot of what I've been doing is working out how to reconcile the requirements and get an implementation.
JDlugosz Anyway, is there anything wrong with defining @-varaibles as formal parameters?
jnthn I gather you're trying to do the same but get a formal spec out of the process rather than an implementation. :-) 23:18
skids is the mmd decision process "specced"
sjohnson rakudo: say 1..10.WHAT
jnthn skids: Yes.
p6eval rakudo 23718a: OUTPUT«Use of type object as value␤␤»
skids where at?
sjohnson rakudo: say (1..10).WHAT
jnthn skids: S12.
p6eval rakudo 23718a: OUTPUT«Range()␤»
sjohnson fuck Perl 6 is almost exaclty like Ruby... am i right? i am happy about it 23:19
JDlugosz jnthn: I would appreciate it if you read "specdoc". There are a few things in there that are well worked out. Even if out of date, some important things that still apply.
sjohnson oops gotta watch my mouth
just so excited!!!
jnthn JDlugosz: I don't see any issue with calling them that off hand.
23:20 dalek left
wayland76 sjohnson: Almost! But More Better :) 23:20
23:20 dalek joined
jnthn sjohnson: Well, Ruby was a source of inspiration too. 23:20
wayland76 (Actually I haven't used enough Ruby to make a judgement call :) )
jnthn Perl has always been a langauge that borrows from others though.
japhb rakudo: say defined %*ENV<some_env_var_that_does_not_exist>
p6eval rakudo 23718a: OUTPUT«1␤»
JDlugosz That's what I'm hoping. But I got suspicious when no such thing is exampled in the synopses. I distinctly remember being shown that in the very early days.
wayland76 rakudo: say exists %*ENV<some_env_var_that_does_not_exist> 23:22
p6eval rakudo 23718a: OUTPUT«No applicable candidates found to dispatch to for 'exists'␤in Main (/tmp/Nh3DbYC4m2:1)␤»
wayland76 rakudo: say %*ENV.exists('some_env_var_that_does_not_exist');
p6eval rakudo 23718a: OUTPUT«0␤»
wayland76 rakudo: say %*ENV.defined('some_env_var_that_does_not_exist'); 23:23
23:23 M_o_C joined
p6eval rakudo 23718a: OUTPUT«too many arguments passed (3) - 1 params expected␤in Main (src/gen_setting.pm:3166)␤» 23:23
wayland76 rakudo: say %*ENV('some_env_var_that_does_not_exist').defined;
p6eval rakudo 23718a: OUTPUT«invoke() not implemented in class 'Env'␤in Main (/tmp/DbIV7iYqTx:1)␤»
jnthn JDlugosz: The biggest thing that I see missing in your description of "Multi Dispatch" is the separation between what we do per invocation and what we compute ahead of time.
JDlugosz: We sort the candidate list completely independently of any invocation. 23:24
23:24 ZuLuuuuuu left
JDlugosz jnthn: where is that? specdoc? 23:24
jnthn And all we do per invocation is walk that ordering and pick the first thing that matches, being aware of amgiguities.
JDlugosz: I was looking at www.dlugosz.com/Perl6/web/typesyste...mmary.html
JDlugosz Oh, thanks.
jnthn If you've defined it more clearly elsewhere, that's fine - it just stood out to me as missing there.
Or at least, groking that it was one of the things that helped me most when implementing it. 23:25
sjohnson i am so excited i can hardly stand it
JDlugosz I recall trying to get to the bottom of things in the mailing list in June or so. That doc should be updated.
sjohnson im not even joking
any time frame for an official Larry Wall release? hopefully not super long 23:26
wayland76 sjohnson: Perl 6 has that effect on me too sometimes :)
sjohnson: Perl 6 is a specification, there are multiple implementations of that specification
No one implementation is the "Official" one
sjohnson is there one written in C tho / will Larry Wall pick one of them to be official? 23:27
or will he write one himself with his buddies?
jnthn JDlugosz: On the "Relationships Between Types" section of the same doc.
japhb Sorry, was AFK. Shouldn't %*ENV<nonexistant> be undefined?
wayland76 There's one called "Rakudo" that's built on the virtual machine called "Parrot"
And there's one called SMOP that is, IIRC, written in C
jnthn JDlugosz: I found that things work out quite well if you just use smart-matching to define acceptance.
JDlugosz: That's the model I've gone down in Rakudo, and it's worked out well. 23:28
wayland76 The big advantages of Rakudo are a) it seems the most complete at the moment, and b) it will be very interoperable with other Parrot-based languages
JDlugosz Yea, the smart match of sig/capture is intended to tell you whether it can be called. So they'd better agree!
jnthn So far anyways...
wayland76 The big advantage of SMOP is that it will integrate well with Perl 5
japhb sjohnson: Larry is specializing in spec and canonical parser (STD). (Most) implementation is left to others.
s/parser/grammar/
jnthn JDlugosz: Yes but I mean at the "does this value do this type" level.
wayland76 Yes, i forgot to mention STD is an official Larry Wall grammar 23:29
23:29 PhatEddy left
jnthn JDlugosz: I see that as just boiling down to TypeThing.ACCEPTS(value) 23:29
wayland76 I'm expecting it will be used by both Rakudo and SMOP at some point
(I'm not involved in either project, except on the periphery of Rakudo)
JDlugosz Oh, tells you if it can (claims to), but that doesn't return the interface or casted object.
jnthn JDlugosz: And it also works for comparing two types.
skids std: my $a # STD doesn't run code it just says whether the syntax is valid
p6eval std 26914: OUTPUT«ok 00:02 36m␤» 23:30
wayland76 (So my impressions of completeness, etc, are based on eg. what I see people saying)
23:30 ruoso joined
JDlugosz I thought ACCEPTS was a fallback if it can't tell which rule to use at compile time. Normally will use a more specific method. 23:30
jnthn I see that more as an optimization.
skids std: "I am the last word."; "Don't like it?, talk to the Wall.".say
p6eval std 26914: OUTPUT«ok 00:04 36m␤» 23:31
jnthn But yes, if we know we can save a level of indirection.
It turns out knowing isn't always quite so easy though.
wayland76 rakudo: "I am in progress".say
p6eval rakudo 23718a: OUTPUT«I am in progress␤»
JDlugosz That means that ACCEPTS must be inline-able and optimizable.
That's hard to do in the general spec, to make sure crazy things have well-defined meanings and implementations agree. 23:32
jnthn JDlugosz: Yes and no. Rakudo optimizes the bulk of multi-dispatches in such a way that we never actually have to call ACCEPTS on a lot of invocations. 23:33
23:33 mizioumt2 left, revdiablo joined
skids pugs: "I'm cold, and I forgot where I put my teeth.".say 23:33
p6eval pugs: OUTPUT«I'm cold, and I forgot where I put my teeth.␤»
wayland76 (pugs is an old implementation that doesn't seem to have anyone working on it any more) 23:34
JDlugosz I like a simple spec, rather than having the compiler know all the rules directly. Anything more is just the optimizer, or a harrier compiler implementation.
jnthn JDlugosz: In which way is "call .ACCEPTS, and to know what that does go see the smart-matching section of the spec" not simple? 23:35
JDlugosz So, your original point... I think I see what you mean. You don't need different casting functions, you just let ACCEPTS do it all.
Rather than having the type-cast form for ACCEPTS turn around and call the special function for that.
Right?
I was agreeing. I like a simple spec. 23:36
jnthn I'm curious about the notion of "casting" you mention.
JDlugosz The "notion" in general?
jnthn OK, I know what casting is. Well, I know what my definition of casting is. :-) 23:37
I'm just not quite sure how it applies in the Perl 6 world.
JDlugosz Given something of one type, produce something of a (possibly different) type that is conceptually the "same thing".
23:37 sjohnson_ joined
sjohnson_ hello! 23:37
perl6... the FUTURE of scripting languages 23:38
jnthn JDlugosz: I fear that definition confuses casting and coercion.
JDlugosz How it applys: implicit conversions, "isa" relationships via static type information providing differing types for the same object, proxies.
What's the difference between casking and coercion?
jnthn I've always defined coercion as a change of representation. 23:39
Whereas casting is just "we see the type as being A, because A is one of the types we can see the thing as being"
JDlugosz Ah. I see it as no change in representation, just coaxing a different facet out of the same object.
casting: cast int to double, ...
jnthn Yes, but that's not a cast. That'a a coercion because a double has a different representation to an int. 23:40
JDlugosz So our uses are directly opposite to each other.
jnthn At least, for the definitions of these things that I'm used to.
JDlugosz In C++, everything is a cast.
sjohnson_ wayland76: are there Perl 6 things out right now that are fully functional that i dont have to install 3 separate things to get it to inpret, so i can start scripting in it?
and that aren't much slower than perl 5 23:41
wayland76 sjohnson: Not sure what you mean
No. Perl 6 is not complete, and Rakudo at least are not worried about speed very much until it's complete
jnthn JDlugosz: I tend to see it more as, in C++ I see there's one syntax and it may do a cast and it may do a coercion.
revdiablo sjohnson_: Do you mean is there a perl 6 implementation ready for production use? No. Not yet.
23:42 sjohnson1 joined
JDlugosz But grabbing different interfaces from one full object is handled at a different level in C++, and so Perl would need to have a notion of "give me a static type from this... and BE SURE it's the same object". 23:42
skids "First make it work"
23:42 exodist left
wayland76 IRC log: irclog.perlgeek.de/perl6/today 23:42
JDlugosz So what would you call those things?
jnthn JDlugosz: I think what I'm missing is the "give me an X" type of thing.
JDlugosz missing? 23:43
jnthn sub foo(Num $x) { } # Num here means "make sure whatever is passed to go in $x has $x ~~ Num giving a true value"
23:43 sjohnson left
wayland76 sjohnson1: Basically, there's nothing out there that's simple to install; i expect that to change within the next month or two, but feature complete and production ready are definitely further away 23:43
23:43 nsh joined
jnthn It's a check, there's no "getting" as such. 23:43
23:43 sjohnson1 left
JDlugosz Yes, it's a check. But if you already know the static type of $x now, you can optimize calls via a jump table. 23:44
There are some things where knowing the static type matters. But I agree, that doesn't mean you "get" something. 23:45
But, if you have a proxy, then getting is important.
23:45 sjohnson_ is now known as sjohnson
sjohnson there we go 23:45
JDlugosz You might generate the guts to "do" that on the fly, when asked.
sjohnson i used to think ruby would solve all the scripting problems in the universe, but now i think perl 6 will
is the Perl 6 spec done, finished, over? or is larry still cooking up some ideas of his 23:46
JDlugosz Thanks for the insight though. That is definitly something to chew on. And certainly something to explain better if I describe it that way.
jnthn Oh, foc sure the compiler can use the knowledge that "if the check fails we never run this code".
wayland76 The Perl 6 spec is still changing
jnthn And the opposite - if we're running the code $x does Num. 23:47
wayland76 For example, Larry just recently added some more specific stuff about how to override sublanguages (eg. regex) in the grammar
But some sections of the spec are moderately stable
23:48 cdarroch left
JDlugosz If you call $x.foo, the compiler can emit a simple jump table jump to foo, based on a table created for Num. It has that table prepared for the object in $x right now. That's what it "gets". But I agree, that's a much lower level detail that Perl 6 sees. 23:48
jnthn JDlugosz: The thing that worries me about using "coercion" is that the Perl 6 spec does use it.
wayland76 What tends to happen is the implementors get arguing about what *should* happen, and then Larry resolves everything with a brilliant new idea (That doesn't happen with all arguments, but it's what drives some of the spec changes)
JDlugosz Oh, that description also was from when I was thinking that contextualizers can be unified with coersions too. All aspects of the same coin, as it were. 23:49
That's exactly why I did use it.
I was just seeing things differently, as you so reciently enlightned me. 23:50
sjohnson does anyone know of the -devel mailing list for Perl 6, and if Larry himself ever reads it?
jnthn That kinda fits with how things work in Rakudo too.
But it still is a different matter to a "type check".
wayland76 sjohnson: There's a "Language" list and a "Compilers" list
skids He reads perl6-language and also comes here. 23:51
wayland76 Language is for arguing about the Spec. I'm on it, and so is Larry.
(perl6-language is the "Language" list I mentioned)
I'm not on Compilers, so I can't speak for it
sjohnson skids, how often does he come here? what's his nickname when he does? 23:52
wayland76 lambdabot: @seen TimToady 23:53
lambdabot TimToady is in #perl6. I last heard TimToady speak 3h 9m 56s ago.
skids sjohnson: you already talked to him, in fact, when you first logged in.
jnthn JDlugosz: If I can show anythign convincing, then maybe: sub foo($x of Str) { } vs sub foo($x as Str) { } 23:54
23:54 iblechbot left
jnthn The first checks the value passed and bound to the parameter $x does Str. The second takes that value - whatever type it is - and (tries to) coerce to to Str. 23:54
They're two quite different operations, and I fear trying to unify them both under the heading "coercions" is the wrong direction, unless you want to parameterize your coercer to say "check or coerce", in which case you probably might as well have defiend two things anyway. 23:55
sjohnson wow 23:57
i had no idea that was him!!!
skids :-)
revdiablo jnthn: I wouldn't call the first coercion at all... rather, a form of type checking 23:58
jnthn revdiablo: Right, that's my point. :-)
revdiablo jnthn: Oh, I thought you were calling them two types of coercion. But scrolling up now I see the unification was JDlugosz's suggestion...
I guess I should have read the scrollback before, rather than after ;) 23:59
jnthn ;-)
skids well, i suppose if downgrading to a more general type... but that's just exposing stuff the object already had.

Powered by Cro and the Raku® Programming Language.

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

Thank you!