»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
Tene rakudo: my @a = 1..10; say @a[*-1,*-2]; 00:00
p6eval rakudo 924242: OUTPUT«109␤»
Tene rakudo: my @a = 1..10; say @a[map { *-$_ }, 1,2]; 00:01
masak rakudo: my @a = 1..10; say ~@a[*-1,*-2]
p6eval rakudo 924242: OUTPUT«109␤»
rakudo 924242: OUTPUT«10 9␤»
TimToady masak: re sets and hashes, that's exactly why sets are defined as special hashes with virtual values 00:02
masak well, HashSet is.
TimToady and why sets and bags behave like lists in list context 00:03
the API is consistent for the immutables too
masak ok.
TimToady in the spec, anyway...
masak what does that mean in terms of sigils? can I declare a Set with a % sigil?
it feels pretty natural to do .{} on a Set... 00:04
TimToady my %set of Set; or my %set := set(...)
masak oki, nice.
TimToady well, my %set of KeySet, if mutalbe
*bl
masak *nod*
TimToady er s/of/is/ 00:05
masak oh, good :)
TimToady KeySets are basically just set containers, since sets are immutable
masak std: set 1, 2, 3 00:06
p6eval std 625303c: OUTPUT«ok 00:01 121m␤»
TimToady so tweaking the membership generates a new set
masak std: bag "lady"
p6eval std 625303c: OUTPUT«ok 00:01 119m␤»
masak \o/
TimToady rakudo: say set 1,2,3 00:07
p6eval rakudo 924242: OUTPUT«123␤»
masak rakudo: say set 1, "1"
p6eval rakudo 924242: OUTPUT«1␤»
TimToady oops :)
masak heh.
it stringifies everything, IIRC.
TimToady yes, well, parrot doesn't really support OO keys yet, iiuc 00:08
but they should use === semantics, not eq
masak hm, it doesn't use a hash internally. it uses an array and uniq.
wonder why it only retains one 1... 00:09
TimToady good think it's encapsulated :)
masak rakudo: say (uniq 1, "1").perl
p6eval rakudo 924242: OUTPUT«(1)␤»
TimToady because uniq uses a hash?
masak ah.
is that what we expect of uniq?
masak looks in the spec
TimToady no, I expect === semantics
masak submits rakudobug 00:11
TimToady Cat.new("mitsy") and Dog.new("mitsy") are obviously two different critters
masak yap.
'night, #perl6 00:19
TimToady yap
00:19 masak left 00:21 vmspb left
sorear nice, t0m is a Perl6er 00:46
00:55 stifynsemons left 00:57 jaldhar left 00:58 jaldhar joined 01:00 stifynsemons joined 01:04 guidjos joined 01:08 guidjos left 01:09 guidj0s_ joined 01:11 guidj0s_ left
colomon poop. If I'd spent another couple of hours working on p4, I'd have gotten seen the pattern repeat and fully cracked the problem. 01:12
s/gotten seen/reached/ 01:19
01:21 justatheory left 01:24 risou_ left 01:40 flatwhatson_ left
colomon Well, for definitions of "fully cracked" that don't include proving I was doing the right thing. 01:49
01:57 coldhead left 02:00 benabik_away is now known as benabik
diakopter colomon: ping 02:01
colomon pong
diakopter :D
impromptu non-trivia quiz: 02:02
what is a Perl 6 "Container Type"
(anyone else feel free to respond/answer, of course) 02:03
colomon :)
I'm not sure there is a formal definition.
but, you know, Array or Hash.
I believe there was a thought of having a formal Container role at one point, but no one could work out what it actually did. 02:04
02:05 noganex left
colomon I could be completely wrong about the details, though. :) 02:05
diakopter reads the paragraphs mentioning 'container' in S09 and S12 02:06
02:06 noganex joined
benabik S10 appears to be using Container as a type in the def of CANDO 02:07
sorear the CANDO stuff is completely wrong 02:08
diakopter yay! someone can erase it! :)
sorear it hasn't made any sense since the last overhaul of the package system
benabik The spec being wrong makes things hard to follow. :-D
02:10 coldhead joined
benabik S32/Containers also defines a Container role, which appears to be an marker role. 02:10
But all the usages in S02 and S09 appear to refer to "Array, Hash or similar" 02:11
diakopter it'd be cool if there was a link on each Text::SmartLinks page (the synopses on perlcabal.org) to each's revision history on github 02:13
TimToady anything mutable is essentially a container
diakopter ok 02:14
is the perlcabal.org/syn/S32/Containers.html#Container up-to-date? 02:15
also, is this fossilized as sorear says: perlcabal.org/syn/S10.html#Autoloading 02:16
TimToady a bit distracted at the moment...
diakopter waits
02:18 stifynsemons left
diakopter takes up the mantle of recording sorear's off-hours irc questions for TimToady in a file in the github repo 02:18
and others' 02:19
colomon diakopter: I would argue that the three functions assigned to Container are pretty useless as a role. 02:21
Errr... unless Hash is not supposed to be involved? It looks like the interface is strictly defined in terms of @ variables, lists, and arrays.... 02:24
02:28 icwiener left 02:31 kst left 02:38 kst joined
sorear Array and Hash containers have much more in common with each other than with Scalar 02:39
if Scalar even exists
colomon but what can you do with a variable that is Array or Hash, other than treat one type as the other? 02:48
.elems, maybe. 02:49
sorear colomon: .iterator and .LISTSTORE 02:52
colomon LISTSTORE? 02:53
.iterator makes sense, agreed.
sorear whatever you want to call the function that handles list assignment 02:54
colomon you can do list assignment to a Hash? 02:55
(and I would have thought Container was supposed to read-only?) 02:56
sorear sure
my %hash; %hash = a => 1, b => 2;
colomon ah, assigning a list of Pairs, sure. 02:57
good point.
though that (liststore of Pairs) seems thin stuff for a Role, IMO. 02:58
hmmm... you can call .elems and .iterator on a scalar too, can't you? 02:59
rakudo: say 1.elems
p6eval rakudo 924242: OUTPUT«1␤»
colomon rakudo: say 1.iterators 03:00
p6eval rakudo 924242: OUTPUT«Method 'iterators' not found for invocant of class 'Int'␤ in main program body at line 22:/tmp/0A5wNvgfBA␤»
colomon rakudo: say 1.iterator
p6eval rakudo 924242: OUTPUT«Method 'iterator' not found for invocant of class 'Int'␤ in main program body at line 22:/tmp/Rz95KVLGsL␤»
colomon okay, maybe not that one.
03:04 f00li5h left 03:06 whiteknight left 03:08 rgrau left 03:10 justatheory joined 03:39 jaldhar left 03:40 jaldhar joined 03:52 zy3g0 joined 04:06 zy3g0 left, jimbob joined 04:12 stifynsemons joined
dalek ecza: 64beb12 | sorear++ | / (3 files):
Simplify handling of labelled control exceptions
 04:15
sorear True or false: A when statement must be the immediate child of a topicalizer block.
04:20 benabik is now known as benabik_away 04:21 risou joined 04:26 Su-Shee left 04:28 Su-Shee joined 04:49 wiseguyxp_ joined 04:50 wiseguyxp_ left 04:51 wiseguyxp joined 05:04 felliott left
snarkyboojum o/ 05:07
sorear hi! 05:10
05:13 risou left
TimToady sorear: false 05:16
succeed and proceed are just control exceptions (that can typically be optimized into goto) 05:18
sorear Why aren't succeed and proceed totally dynamic like next and last? 05:19
TimToady next and last are totally dynamic in the absence of a label
with a label they're lexotic
I don't there's as much call for dynamic succeed, but certainly if you succeed in a scope without an outer topicalizer, dynamic is the only thing left to try 05:20
sorear I mean like next and last withou labels
TimToady it doesn't seem to me to be what the typical user would expect 05:21
we could, I suppose, allow LABEL.succeed for pure lexotic, and the implicit succeed would then lexotic as well 05:22
an explicit succeed without a label could conceivably be dynamic 05:23
but certainly it's not specced that way currently
I think consistency with the loop policy has something to be said for it, as long as the implicit ones stay lexotic (even if there's no label) 05:24
I wouldn't expect anyone to ever use LABEL.succeed, but consistency doesn't seem inachievable here 05:25
unachievable even 05:26
I'm not sure what LABEL.proceed would mean though
presumably the label would be on the when that is being left 05:27
05:27 envi joined
TimToady otherwise it's just a goto 05:27
snarkyboojum phenny: tell masak, I particularly enjoyed the release "story" for yapsi this month :D 05:28
phenny snarkyboojum: I'll pass that on when masak is around.
TimToady or maybe it uses the label of the whole structure to figure out which when to exit
sorear Is LABEL required to be able to identify the Block object?
TimToady a thing can be lexotic without a label, but then it's difficult to express intentional lexoticity 05:29
like return from a 'sub', which doesn't care about the subs name, only the keyword and the identity of the sub in question 05:30
so I can also argue that succeed/proceed are always lexotic, and you can use $scope.leave to do anything else dynamically 05:31
so it would be more like return, and less like last 05:32
and, in fact, the proper way to implement succeed is more like return, so that it can return the value of the when block 05:33
sorear LOOP: for 1 { say ?( LOOP ~~ Code ) } # ?? 05:34
LOOP: for 1 { say ?( LOOP === &?BLOCK ) } # ??
TimToady no
it's just a label
presumably you can ask it for the identity of the block in question though
sorear Is that necessary? I just noticed that the model of labels I've been working under doesn't allow it 05:35
TimToady rakudo: LABEL: say LABEL.WHAT 05:36
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "LABEL: say"␤»
sorear rakudo doesn't do labels at all
TimToady pugs: LABEL: say LABEL.WHAT
p6eval pugs: OUTPUT«*** No such subroutine: "&LABEL"␤ at /tmp/pUR53unwQV line 1, column 12 - line 2, column 1␤»
TimToady pugs: LABEL: for 1 { say LABEL.WHAT }
p6eval pugs: OUTPUT«*** No such subroutine: "&LABEL"␤ at /tmp/MEgABLtd6j line 1, column 20-31␤»
TimToady I guess pugs doesn't either
sorear pugs: LABEL: goto LABEL;
p6eval pugs: OUTPUT«*** No such subroutine: "&goto"␤ at /tmp/P_XSWEdhfd line 1, column 8-18␤»
sorear pugs: LABEL: while True { last LABEL; } 05:37
p6eval pugs: OUTPUT«*** No such subroutine: "&LABEL"␤ at /tmp/3Fvi56XQpp line 1, column 21-31␤»
TimToady how can you write a lexotic exception if you can't identify the correct frame to unwind to?
sorear The frame to unwind to is the OUTER:: of the block 05:38
TimToady no
that's like saying you can't return from inside an if 05:39
that's why we invented lexoticness, so we can identify frames lexically
and use that to capture the throw at the right palce 05:40
*place
sorear No, that's not quite what I meant
LOOP: loop { loop { last LOOP } } -> my $LOOP = Label.new(callframe, "LOOP"); "LOOP": loop { loop { last $LOOP } } 05:41
the label doesn't hold a reference to the loop body
if it did, it would prevent loop inlining in a lot of interesting cases 05:42
TimToady labels aren't strings; they're constants 05:43
sorear my $LOOP = Label.new(callframe, 0); 0: loop { loop { last $LOOP } } # if you prefer
I had planned to use names for debugging, but my model works with any unique identifier 05:44
*debuggability
TimToady I don't see how knowing the outer callframe helps you know where to start up again with 'next' 05:45
the code performing the loop needs to know if the block passed in is the one referred to in the control exception it just caught 05:46
so it's more the block needs to know its label
sorear the code performing the loop knows the label name 05:47
the exception is correct to catch if the frame matches the current frame, and the name matches
TimToady it will get the wrong one dynamically if it's scanning for names, and an inner loop has the same label; that's no good 05:48
sorear it's not just scanning for names, it's also scanning for frames
TimToady it has to be a unique identity for the block, that the exception has got from the label used
sorear name,outerframe pairs are a unique identity for blocks 05:49
since the outer block only clones its children once
and names are unique within the outer block
TimToady well, maybe name/outerframe the value of the label constant 05:50
but it ain't just a string
sorear aye
TimToady when you say LABEL.foo, it's got the lexotic identity, and when you say "LABEL".foo, you have to do it dynamically 05:51
(unless we let some things look up caller labels...) 05:52
in which case the latter could behave lexotically, with the cooperation of .foo
anyway, I suspect for optimizability it would be better if succeed defaults to lexotic 05:54
then it has some change of turning into a low-level goto
*chance
which also makes me wonder if we've defaulted next and last the wrong way 05:55
and we should have something explicit to indicate a dynamic throw
"next if" is very heavily used in typical perl code 05:56
sorear actually I find dynamic to be easier to optimize
TimToady that seems counterintuitive to me
PerlJam puts a next inside a nested sub 5 levels deep just for sorear 05:57
sorear there are 2 interesting cases
case 1: next/etc is inside a pile of lexical scopes, with no extra dynamic scopes 05:58
case 2: there are extra dynamic scopes
in case 1, dynamic and lexotic can both be turned into a goto
TimToady I suppose the dynamic scope can make it easier to 'next' things like, say map
sorear in case 2, neither can because the dynamic scopes might have CONTROL blocks
but dynamic is easier because it doesn't require creating label objects 05:59
05:59 arlinius left
sorear I agree it's counterintuitive 05:59
it makes me think I'm attacking the problem entirely wrong
TimToady I think people will tend to write their own loops more than their own switches, but I suppose the argument can carry over there too 06:00
sorear the main argument for lexicality of switchy stuff is that $_ is lexical
TimToady well, but for loops also set $_ 06:01
06:01 arlinius joined
TimToady it's not entirely clear whether we want to say that CONTROL blocks see exceptions that are lexotically aimed elsewhere; if we blinded them to that, then you case 2 becomes easier to optimize for lexotic scopes, since you can ignore dynamic CONTROL blocks 06:03
*your
and really, anything you want to do on stack unwind should really be in something more like a LEAVE block 06:04
so I'm not too scared to say that CONTROL blocks can be auto-skipped, unless there's a good use case for not doing so 06:05
in that sense, it's really the identity of the CONTROL block that is at issue here
06:06 dukeleto left
TimToady well, maybe that's oversimplified 06:06
06:06 dukeleto joined
sorear even a LEAVE block has issues 06:06
it requires a stack unwind
TimToady well, yes, that's the point of 'em 06:07
sorear unwinding to a specific frame doesn't ring to me as much cheaper than unwinding to a frame with the "handles next" bit
TimToady the unwind should not be confused with the throw
the unwind happens after the throw finds the right handler 06:08
PerlJam I sure wish I understood what you guys were talkling about. :-) 06:11
TimToady for example, if a 'return' doesn't find its sub, it throws a new exception from the point of the return
with the original frames still there
sorear PerlJam: last LABEL;
PerlJam Because "auto-skipping CONTROL blocks" makes me wonder how you exit out of nested blocks each with their own LAST semantics. 06:12
06:12 Su-Shee left
TimToady LASTs are processed more like LEAVE, it would seem 06:12
that is, on stack unwind, not on the control exception 06:13
might imply that a stack unwind should be told the reason for unwinding 06:14
dalek ecza: 7742322 | sorear++ | src/ (3 files):
Trap succeed in blocks that signature-bind $_
 06:15
TimToady but again, a 'last' that doesn't find its loop frame as a control exception throws a real exception from the same deep frame instead (which is likely to unwind)
sorear TimToady: Have you seen t0m's TR18 post? 06:16
TimToady exception of either kind never unwind; they only schedule an unwind after they've decide who handles the exception
PerlJam TimToady: and that "find" is lexical?
TimToady no, dynamic, which includes lexotic if there's a lexical identity to scan for
but the scan is of necessity dynamic
sorear: yes, I glanced at it, but haven't had a chance to reply 06:17
it's probably worth someone taking a pass over to see how much the Perl vision and the Unicode vision have diverged over the last year or three
in any case, most of what we've come up with we've done because it's the right thing to do, not because of a consortium report 06:18
sorear TimToady: What do you mean by last throwing a real exception?
TimToady a fatal exception to be caught by CATCH, not a control exception
PerlJam sorear: I parsed that as "non-CONTROL exception"
sorear "No handler for control exception last in dynamic scope" or so? 06:19
TimToady control expections are conceptually similar, but we want a very lightweight implementation
since they aren't, in fact, exceptional
PerlJam maybe they should be "control expectations" rather than "control exceptions" 06:20
TimToady sorear: well, I'd put it into User-ese rather than Language-lawyer-ese, but yes
You used a 'last' that wasn't in the dynamic scope of a loop; shame on you!
I don't expect the casual use to know that 'last' throws a control exception 06:21
this is one of those spots where it's important for budding programmers to do cargo-cult programming
they can learn the exact semantics and technical terms later
PerlJam I've always gotten the impression that the dynamic nature of next/last/redo was surprising to most Perl 5 programmers. I bet that will continue to be the case with Perl 6. 06:24
TimToady same deal with my "Useless use of [] around infix op" in my version of that patch; I don't think the newish user will understand "metaoperator brackets" 06:26
(it's not clear they are intrinsically tied to metaoperators in any case) 06:27
(for instance, there's no metaoperator in the &[+] form, but I think of those as the same brackets)
std: 1 [+] 42 06:28
p6eval std 625303c: OUTPUT«ok 00:01 119m␤»
TimToady I guess it hasn't recompiled lately
sorear niecza: 1 [+] 42 06:29
p6eval niecza v2-10-g7742322: OUTPUT«Potential difficulties:␤ Useless use of [] around infix op at /tmp/ItTwkZ8sAH line 1:␤------> 1 ⏏[+] 42␤␤»
TimToady :)
or I guess it's using the CPAN viv? 06:30
06:30 risou joined
sorear the last build attempt was successful (and last year) 06:31
TimToady niecza: s{foo}{bar} 06:32
p6eval niecza v2-10-g7742322: OUTPUT«===SORRY!===␤␤Unexpected block in infix position (two terms in a row, or previous statement missing semicolon?) at /tmp/zvTneNmG79 line 1:␤------> s{foo}⏏{bar}␤␤Parse failed␤␤»
TimToady aha, you didn't get that one yet :) 06:33
sorear std is suspiciously absent from /chroot/var/spool/cron/crontabs/p6eval 06:34
06:34 arlinius left
sorear phenny: tell moritz_ std isn't in /chroot/var/spool/cron/crontabs/p6eval and the last std: update was Dec 16. Bug? 06:34
phenny sorear: I'll pass that on when moritz_ is around.
sorear std: given 12 -> $x { } # Is this a topicalizer? 06:37
p6eval std 625303c: OUTPUT«Potential difficulties:␤ $x is declared but not used at /tmp/FeqhRld8XY line 1:␤------> given 12 -> ⏏$x { } # Is this a topicalizer?␤ok 00:01 121m␤»
TimToady not in the sense of setting $_
used to be that way, till we deemed double binding insane 06:38
it got particularly messy if there were multiple parameters; is $_ only an alias for the first one? 06:40
took a long time to detangle invocants from topics too 06:41
sorear there's still mention of invocants in S04 06:42
06:42 _ilbot joined
sorear I take it that's a fossil? it seemed odd. 06:42
TimToady those look accurate to me, on first blush. .foo takes $_ outside, but inside it's the invocant 06:43
or was it a different place you're wondering about? 06:44
sorear S04:869
TimToady inside, the topic and invocant are the same only if $_ is one of the formal parameters
that seems consistent with what I just said 06:45
dalek ecza: 03bd80d | sorear++ | / (2 files):
Implement "given"
TimToady method ($_: $a,$b,$c) {...}
sorear S04:869 seems to imply that method foo($x: $_) { } doesn't set the topic 06:46
TimToady $_ is both invocant and topic there
it should probably s/invocant/any of its parameters, including the invocant,/
06:47 molaf joined
TimToady in some sense $_ = 'foo' is a topicalizer too, but not as far as 'succeed' knows 06:47
it can really only know about sig bindings attached to a scope 06:48
sorear I get the feeling that Perl 6 once thought of "topic" as the master concept
TimToady succeed is really based on the topic container's identity, not the topic's contents... 06:49
sorear whereas now $_ is master and "topic" just means "whatever is in the lexically visible $_"
TimToady $_ as an lvalue is the master now, and the actual topic is kind of epiphenomenal 06:50
so yeah
the actual meaning of "it" is pretty sloppy in natural languages, even if you have official topicalization syntaxes 06:52
sorear (have I mentioned yet that I'm currently studying Japanese?) 06:55
TimToady はい、そう分かりましたよ。 06:57
06:58 arlinius joined
sorear does not know what sou is doing there 07:02
also, I seem to have lost my ability to type macrons since I restarted X earlier today
TimToady sou works much the same as in English, oddly 07:03
dalek ecza: b0e8b22 | sorear++ | / (2 files):
Implement postfix when and for
TimToady "so I have understood"
sou often ends up as kinda meaning "that" when applied to an actoin 07:06
*action
そうします for instance
be we also have "do so" in English 07:07
*but
diakopter oddly and evenly 07:10
07:11 meteorjay left
TimToady std: にほんご: goto にほんご; 07:14
p6eval std 625303c: OUTPUT«ok 00:01 118m␤»
07:18 kaare_ joined
diakopter perlesque: にほんご: goto にほんご; 07:18
p6eval perlesque: OUTPUT«(timeout)»
diakopter FTW
sorear TimToady: When you talk about "doing all your C coding in Perl 6", what do you have in mind? Linux loadable kernel modules? :) 07:20
diakopter would love more context there too 07:23
TimToady well, if we could write the filesystem in Perl 6, that might make it easier to support Unicode filenames. :) 07:25
interestingly, Clearcase's translucent filesystem was written in Perl 5
dalek ecza: 4ebb175 | sorear++ | / (2 files):
Implement block "when"
 07:26
TimToady it actually ran ten times faster than their C version of the code--go figure
sorear ok, niecza has given/when now
that wasn't nearly so hard as I expected
TimToady and having failed to compile a Linux driver module earlier today, I'd certainly prefer to write one in Perl 6... 07:27
sorear and this'll allow a few hundred more lines of STD diff noise to be removed
sorear -> sleep 07:28
TimToady おやすみなさい！
07:36 Chillance joined 07:39 wiseguyxp left
szabgab good morning 07:51
we are in the Perl dev room at FOSDEM on screen, just before the Perl 6 taslk
rakudo: say "good morning" 07:53
p6eval rakudo 924242: OUTPUT«good morning␤»
07:54 justatheory left, justatheory joined 07:56 Mowah joined
diakopter hello screen 07:57
07:59 franckc joined, tommie joined
franckc hi fosdem! 07:59
07:59 justatheory left
tommie :-D 07:59
szabgab rakudo: say "good morning, frank"
p6eval rakudo 924242: OUTPUT«good morning, frank␤»
08:00 awoodland joined, AlviN900 joined 08:01 jeeger joined, jonnie joined, sawyer_ joined
jeeger rakudo: say [*] 1..10 08:01
p6eval rakudo 924242: OUTPUT«3628800␤»
TimToady rakudo: say (1, 1, * + * ... *)[^20].Str 08:02
p6eval rakudo 924242: OUTPUT«1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181 6765␤» 08:03
jeeger rakudo: say 1..30».PERL
p6eval rakudo 924242: OUTPUT«Method 'PERL' not found for invocant of class 'Int'␤ in main program body at line 22:/tmp/_iaorJwfyL␤»
jeeger rakudo: say 1..30».perl
p6eval rakudo 924242: OUTPUT«1␤»
TimToady precedence
jeeger rakudo: say 1..30»perl
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "say 1..30\x{bb}"␤» 08:04
jeeger hm
diakopter rakudo: say (0, 1, * + * ... *)[^20].Str
p6eval rakudo 924242: OUTPUT«0 1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181␤»
jeeger say (1..30)».perl
TimToady rakudo: say (1..30)».perl.Str
p6eval rakudo 924242: OUTPUT«1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30␤»
jeeger ah
08:07 AlviN900 left
TimToady zzz & 08:11
08:12 risou left 08:13 Raven joined
jeeger rakudo: say 1 <= 35 08:14
p6eval rakudo 924242: OUTPUT«Bool::True␤»
Raven raduko: for(1,2,3,4,5) -> $x, $y? { say "$x, $y"; } 08:15
08:17 ecocode joined
jeeger rakudo: my @series = 1..Inf; say @series[-1]; 08:21
p6eval rakudo 924242: ( no output )
Raven raduko:for(1,2,3,4,5) -> $x, $y? { say "$x, $y"; } 08:22
jeeger Raven: rakudo, not raduko 08:23
Raven rakudo:for(1,2,3,4,5) -> $x, $y? { say "$x, $y"; } 08:24
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "for(1,2,3,"␤»
Raven ha, that explains why query did work
jeeger rakudo: say (1,2,3,4,5) -> $x, $y { say "$x, $y"; }
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "say (1,2,3"␤»
jeeger rakudo: for (1,2,3,4,5) -> $x, $y { say "$x, $y"; } 08:25
p6eval rakudo 924242: OUTPUT«1, 2␤3, 4␤Not enough positional parameters passed; got 1 but expected 2␤ in <anon> at line 2:/tmp/2BoZpuOn_n␤ in main program body at line 1␤»
jeeger rakudo: for(1,2,3,4,5) -> $x, $y { say "$x, $y"; }
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "for(1,2,3,"␤»
jeeger hm, that looks like a bug
Raven yay for early morning bugfixes!
08:33 RichiH_FOSDEM joined
ecocode approx 50 persons in Perl Devroom @ Fosdem. 08:33
RichiH_FOSDEM shouldn't the topic say "p6eval usage"? 08:34
08:39 tommie left
Raven I love his @experssions 08:41
08:47 jeeger left, awoodland left
RichiH_FOSDEM experss yourself 08:47
and at the end of the talk, he started telling us about ncurses ;) 08:49
08:52 flatwhatson_ joined, sunnavy joined 08:53 awoodland joined
jdhore quick question (since i'm in here) can you not put curly braces around an if statement if it only contains one line? 08:58
09:00 jimbob left 09:02 stifynsemons left 09:13 envi left, sawyer_ left, envi joined 09:20 awoodland left
tadzik jdhore: nope 09:25
you can use the FOO if BAR though
rakudo: say "foo" if 5 < 7
p6eval rakudo 924242: OUTPUT«foo␤»
jdhore ah 09:27
thanks
09:48 ecocode left 09:49 ecocode joined 09:50 pmurias joined 09:51 ecocode left 10:00 cafesofie left 10:16 sunnavy left 10:17 sunnavy joined 10:48 zenog joined
zenog Stupid newbie question that unfortunately has not been answered by Google: What is wrong with the line "my %hash = map { $_ => $_ } @array;"? 10:49
tadzik my @a = 1, 2, 3; my %hash = map { $_ => $_ } @a; %hash.perl.say 10:53
rakudo: my @a = 1, 2, 3; my %hash = map { $_ => $_ } @a; %hash.perl.say
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "my %hash ="␤»
tadzik rakudo: my @a = 1, 2, 3; my %hash = map { $_ => $_ }, @a; %hash.perl.say
p6eval rakudo 924242: OUTPUT«Can't call map() with a Hash argument, Callable required␤You probably wrote a Hash composer accidentally - try to␤disambiguate it with a ; directly after the opening brace␤ in 'Any::map' at line 1511:CORE.setting␤ in 'map' at line 1863:CORE.setting␤ in main program body at
..line …
10:53 agentzh joined
tadzik sigh 10:53
rakudo: my @a = 1, 2, 3; my %hash = @a.map({ $_ => $_ }); %hash.perl.say
p6eval rakudo 924242: OUTPUT«Can't call map() with a Hash argument, Callable required␤You probably wrote a Hash composer accidentally - try to␤disambiguate it with a ; directly after the opening brace␤ in 'Any::map' at line 1511:CORE.setting␤ in main program body at line 22:/tmp/aKiFU9Mkty␤» 10:54
10:55 f00li5h joined
tadzik I don't think hash is a list of pairs actually 10:55
moritz_ map: {; $_ => $_ } 10:56
phenny moritz_: 06:34Z <sorear> tell moritz_ std isn't in /chroot/var/spool/cron/crontabs/p6eval and the last std: update was Dec 16. Bug?
moritz_ sorear: yes, bug
tadzik moritz_: and that's a bug, above?
zenog tadzik: it has nothing to do with the %hash, "map { $_ => $_ } @array;" also fails. 10:57
moritz_ tadzik: specbug
tadzik I see
zenog: see the moritz_'s example 10:58
10:58 masak joined
masak g'day, zebras. 10:58
phenny masak: 05:28Z <snarkyboojum> tell masak I particularly enjoyed the release "story" for yapsi this month :D
masak :)
it's a true story, by the way. my life is full of incidents where I lose my way. 10:59
coldhead is the Perl 6 And Parrot Essentials book still relevant/recommended (if that's not a rude question)? 11:00
masak not a rude question. :)
moritz_ coldhead: it's very out of date
zenog tadzik: If it is a bug I'll just try a not-so-elegant workaround for now. Thanks. 11:01
masak coldhead: I suppose something can be learned from it, but mostly it'll be something having to do with ancient history.
moritz_ it still gets some of the Perl 6 essentials across, but lots of details have changed
coldhead that was my fear, it seemed foolhardy to try to document it back then
masak coldhead: especially in retrospect...
moritz_ a nice way to write %h = map { $_ => $_ }, @a is %h = @a Z=> @a 11:02
coldhead i might try it anyway. maybe it could be redeemed by an errata sheet 11:03
masak ...a long errata sheet... 11:07
moritz_ coldhead: I hope somebody already pointed you to "Using Perl 6"? 11:08
11:09 jaffa4 joined
coldhead they did but i liked the perl 6 + parrot option 11:09
jaffa4 rakudo: $_ = 'foo'; s:g/ /a/; say 'alive'
p6eval rakudo 924242: OUTPUT«alive␤»
jaffa4 std: $_ = 'foo'; s:g/ /a/; say 'alive' 11:10
p6eval std 625303c: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/XkMjKRPuJf line 1:␤------> $_ = 'foo'; s:g/ ⏏/a/; say 'alive'␤ expecting quantifier␤Parse failed␤FAILED 00:01 122m␤»
jaffa4 Where is null pattern over there? I cannot see it.
std: $_ = 'foo'; m:g/ /a/; say 'alive' 11:11
p6eval std 625303c: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/TdwOVFDp1V line 1:␤------> $_ = 'foo'; m:g/ ⏏/a/; say 'alive'␤ expecting quantifier␤Parse failed␤FAILED 00:01 122m␤»
masak jedai: s:g/⏏/a/
argh, mistab again :/
jaffa4: the null pattern is where the ⏏ is.
jaffa4 there is space there.
masak jaffa4: whitespace is irrelevant in p6 regexes. 11:12
jaffa4 ok
11:14 dju left
jaffa4 Can I modify rakudo? 11:14
masak please be more specific. 11:15
jaffa4 Do I have the permissions to do so?
Do I have the permissions of rakudo github to do so?
11:18 dakkar joined
moritz_ jaffa4: nope. But you can fork and submit pull requests, or submit patches via RT 11:18
good morning btw
masak mornin', moritz_
moritz_ we had a great (late) burns supper yesterday night
and there's a portion of Stickey Toffee Pudding in the fridge :-) 11:19
coldhead wee muckle kilty tehk the glen! 11:20
jaffa4 How do you debug rakudo? 11:25
Do you use some tools? 11:26
tadzik jaffa4: what do you mean "debug rakudo"? 11:29
jaffa4 perl -d?
I mean are there some debugging switches or settings? 11:30
moritz_ there's the --target switch which lets you see intermediate code
and --ll-backtrace
see docs/running.pod
tadzik use Devel::Trace also
11:32 zenog left 11:39 coldhead left
colomon jaffa4: lots and lots of "say" statements. 11:53
jaffa4 I mean the rakudo compiler
colomon jaffa4: so did I. 11:54
tadzik:
rakudo: my @a = 1, 2, 3; my %hash = @a Z=> @a; %hash.perl.say # just in case you actually were looking for how to do that. :) 11:55
p6eval rakudo 924242: OUTPUT«{"1" => 1, "2" => 2, "3" => 3}␤»
masak jaffa4: when developing things for src/core, it's often best to do it in a standalone file first, and only later integrate it. keeps the development cycle short. 11:57
jaffa4 What standalone file?
moritz_ a .pl file 12:00
masak a script file that you write, instead of starting by integrating things into src/core
moritz_ colomon: re Z=>, I proposed that two hours ago too :-)
jaffa4 How could I extract a part from the whole without breaking the whole?
colomon moritz_: sorry, sketchy half-awake backlogging. ;) 12:01
moritz_ jaffa4: en.wikipedia.org/wiki/Jenga
masak jaffa4: the parts of src/core are relatively independent. in the cases where they aren't, it can get tricky. but often it's solvable even then. 12:02
12:07 franckc left 12:12 MayDaniel joined
moritz_ to be more precises, things in src/core are not needed while compiling rakudo, other parts of rakudo only depend on them at run time 12:13
so if you supply the missing parts at run time, it's all good
tadzik sounds like the day to merge metainfo
12:20 Chillance left
tadzik bah. While JSON parses for 40 seconds, storing the .perl string and evaling it later takes 17 seconds 12:28
not much better, but still, 23 seconds less for installing a module 12:29
maybe sqlite is the way then... bleh
masak tadzik: 17 seconds -- that sounds... unreasonable. 12:30
colomon very unreasonable 12:34
can you paste the .perl string to a gist?
tadzik si 12:37
gist.github.com/raw/813345/c328fe1...jects.perl 12:38
maybe the cpanm way is the way: fetching the metainfo every single time from a web service. But it's nothing but a workaround for the Rakudo slowness 12:39
12:40 yarp joined 12:41 whiteknight joined
tadzik I wonder how niecza... ah, no eval 12:43
12:45 dakkar left 12:51 dakkar joined 12:57 agentzh left 13:15 kfo_ joined 13:18 kfo left, jeeger joined
jeeger Greetings! I'm still looking on how to do something like "@a Z @b Z @c", but with extending the lists. 13:19
masak jeeger: extending the lists with what? 13:20
jeeger masak: repeating them, for example, like @a <<,>> @b would 13:21
masak I seem to recall the spec has something to say about this. hold on.
13:23 felliott joined
masak ah, S32/Containters:101: "The C<zip> function defaults to stopping as soon as any of its lists is exhausted. This behavior may be modified by conceptually extending any short list using C<*>, which replicates the final element." 13:23
rakudo: sub repeat(@list) { gather loop { take $_ for @list } }; my @alpha = 'a' .. 'z'; my @digits = 0..9; say @alpha Z repeat(@digits) 13:26
p6eval rakudo 924242: OUTPUT«a0b1c2d3e4f5g6h7i8j9k0l1m2n3o4p5q6r7s8t9u0v1w2x3y4z5␤»
masak \o/
jeeger rakudo: @numbers,* Z @letters 13:27
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Symbol '@numbers' not predeclared in <anonymous> (/tmp/4cjm9bIaPO:22)␤»
jeeger rakudo: say (1,2,3,4,*) Z ("one","two","three","four","five")
p6eval rakudo 924242: OUTPUT«1one2two3three4four4five␤»
jeeger aha!
rakudo: say (1,2,3,4,Any()) Z ("one","two","three","four","five") Z 13:28
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "say (1,2,3"␤»
jeeger rakudo: say (1,2,3,4,undef) Z ("one","two","three","four","five") Z 13:29
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Unsupported use of undef as a value; in Perl 6 please use something more specific:␤ Mu (the "most undefined" type object),␤ an undefined type object such as Int,␤ Nil as an empty list,␤ *.notdef as a matcher or method,␤ Any:U as a type constraint␤ or
..fail() as a failur…
arnsholt You've got a trailing Z
masak Perl 6 doesn't have an undef.
jeeger rakudo: say (1,2,3,4,undef) Z ("one","two","three","four","five")
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Unsupported use of undef as a value; in Perl 6 please use something more specific:␤ Mu (the "most undefined" type object),␤ an undefined type object such as Int,␤ Nil as an empty list,␤ *.notdef as a matcher or method,␤ Any:U as a type constraint␤ or
..fail() as a failur…
jeeger Yeah, that was a quite exhaustive error message
(for once^^)
masak :)
jeeger rakudo: say (1,2,3,4,Mu()) Z ("one","two","three","four","five") 13:30
p6eval rakudo 924242: OUTPUT«Could not find sub &Mu␤ in main program body at line 22:/tmp/hFSgS8xjcY␤»
jeeger Okay, I'll take this to a query
masak it's Mu, not Mu() 13:31
jeeger That's what I'm trying now^^
13:44 REPLeffect left 13:50 jeeger left, jeeger joined, Axius joined 13:59 yarp left, jeeger left, REPLeffect joined
masak std: sink 14:13
p6eval std 625303c: OUTPUT«===SORRY!===␤Bogus statement at /tmp/prQVAhS8Ix line 1 (EOF):␤------> sink⏏<EOL>␤Parse failed␤FAILED 00:01 119m␤»
14:13 orafu left
masak std: sink say "OH HAI" 14:13
p6eval std 625303c: OUTPUT«ok 00:01 120m␤»
masak rakudo: sink say "OH HAI"
p6eval rakudo 924242: OUTPUT«OH HAI␤»
masak \o/
14:15 cotto left 14:16 Axius left
masak TimToady: $obj.?foo returns Nil if there's no .foo method in the WALK chain. what about $obj.*foo? S12 doesn't seem to say. 14:23
masak checks roast 14:24
xinming_ do -> $x { $x.say } <--- How do we provide the arg for $x in this case? 14:26
14:26 drbean left
xinming_ do (-> $x { $x.say })(1) Got it. 14:27
masak rakudo: -> $x { $x.say }(1)
p6eval rakudo 924242: OUTPUT«1␤»
masak no need for 'do' or parens.
xinming_ I was thinking that how to pass arg to do block. 14:28
Since we can have positional arg
there
do { say $^x }; This 14:29
14:30 Hackbinary joined
masak there's not really any such thing as a "do block". 14:30
there's the "do" keyword, and then there's blocks.
rakudo: do say 42 14:31
p6eval rakudo 924242: OUTPUT«42␤»
masak rakudo: do do do do say "doo-diddly-do"
p6eval rakudo 924242: OUTPUT«doo-diddly-do␤»
masak rakudo: do do do do my $ = { say "not executed" } 14:32
p6eval rakudo 924242: ( no output )
masak rakudo: do do { say "immediate block" }
p6eval rakudo 924242: OUTPUT«immediate block␤»
xinming_ hmm, Got it, I think I need to think more about this.
masak xinming_++ 14:33
14:33 xinming_ is now known as xinming
xinming xinming, not xinming_ :-) 14:33
please ++ again. :-P
The synopsis now is quite different from what I read last time.
masak xinming++ 14:35
yes, synopses tend to behave that way.
from the bright side, the changes are either improvements, or the change before the improvement :P 14:36
xinming When I read the synopsis, I found perl 6 maybe abit over complicated. 14:37
Which made me think, How to implement such a language. :-)
moritz_ we've been asking that ourselves for some years now.
:-)
xinming Though, Design direction is consistent.
moritz_ but we've made progress 14:38
xinming yes, a great progress, I am trying rakudo these days.
too much to learn, But I believe perl 6 will be the programming language in last 30 yearts. :-P 14:39
A bit like perl 5 released in last century.
masak Perl 6 is a connaisseur language, for people with great taste. 14:41
14:41 dakkar left
masak rakudo: say Nil.elems 14:43
p6eval rakudo 924242: OUTPUT«0␤»
masak where the spec is implicitly undecided on what .*foo returns when there's no .foo, roast is *explicitly* indecisive. interesting. 14:44
it just says "empty list", which might be () or Nil.
it doesn't have a test for the definedness of the value.
masak adds a test
14:46 dakkar joined
masak rakudo: class A {}; A.new.+foo; say 'alive' 14:48
p6eval rakudo 924242: OUTPUT«Could not invoke method 'foo' on invocant of type 'A()'␤ in main program body at line 22:/tmp/E3axhhS400␤»
masak rakudo: class A {}; A.new.*foo; say 'alive'
p6eval rakudo 924242: OUTPUT«alive␤»
masak rakudo: class A {}; say A.new.*foo.perl 14:49
p6eval rakudo 924242: OUTPUT«()␤»
dalek ast: 63157bc | masak++ | S12-methods/calling_sets.t:
[S12-methods/calling_sets.t] made test more precise

  .*foo when there's no .foo is now tested to give not only the empty list
  (), but Nil, just like .?foo		 14:50
14:53 jferrero left 14:57 sji joined
jaffa4 rakudo: m/hello/; 15:05
p6eval rakudo 924242: OUTPUT«Method 'match' not found for invocant of class ''␤ in main program body at line 22:/tmp/sjOqeFwzl5␤»
15:05 cafesofie joined
jaffa4 rakudo: /hello/; 15:07
p6eval rakudo 924242: ( no output )
jaffa4 rakudo: m/hello/;
p6eval rakudo 924242: OUTPUT«Method 'match' not found for invocant of class ''␤ in main program body at line 22:/tmp/sFNaRD0_w5␤»
jaffa4 rakudo: :m/hello/; 15:08
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near ":m/hello/;"␤»
masak jaffa4: the difference is that /hello/ is a (Regex) object, but m/hello/ is an operation.
jaffa4: there's nothing like :m/hello/, and I don't know what made you think there is.
15:09 jferrero joined
jaffa4 Is it possible to change / as a separator? 15:09
masak rakudo: $_ = "heffo"; s!ff!ll!; say $_ 15:10
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "s!ff!ll!; "␤»
masak hm.
masak checks S05
sorear good * #perl6 15:12
masak o/ sorear
std: s!ff!ll! 15:13
p6eval std 625303c: OUTPUT«ok 00:01 119m␤»
masak submits rakudobug
jaffa4: so, the answer seems to be "yes, it should be possible to use a separator other than '/', but Rakudo doesn't allow that yet".
jaffa4 right 15:15
dalek kudo: 684fdf8 | masak++ | src/glue/dispatch.pir:
[src/glue/dispatch.pir] .*foo returns Nil, not ()
15:22 sji left
jaffa4 $*key is like local, right? 15:25
sorear right. 15:27
in terms of usage, anyway
15:27 shi joined 15:28 wamba joined
TimToady I'm not sure that Nil is correct for something that is expected to return a list 15:30
shower &
masak hm. it feels odd to me that there'd be an inconsistency between .?foo and .*foo 15:32
we care enough about the *consistency* between /<foo>?/ and /<foo>*/ to package the former into an array.
(despite the overhead and confusion that creates) 15:33
jaffa4 !paste 15:43
?paste
tadzik phenny: paste? 15:44
jaffa4 What does this mean <quote_EXPR: ':qq'>? 15:45
15:46 mikehh joined
moritz_ it calls the method quote_EXPR on the Curosr with argument ':qq' 15:48
15:48 dakkar left
jaffa4 You mean rule? 15:49
sorear Rules are just one kind of method 15:50
masak ...in the OO inheritance sense, even. 15:56
moritz_ in any kind of sense, really 15:57
15:58 dakkar joined
sorear is doing s!!! for niecza now 15:59
jaffa4 rakudo: $e~~s//g/;
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "$e~~s//g/;"␤»
moritz_ sorear: if you figure out sane semantics for return value and $/ setting for s/// I'd like to steal that for rakudo :-)
16:00 pmurias left
sorear moritz_: What's the problem with the return value ? 16:00
jaffa4 rakudo: $e~~s://g/;
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "$e~~s://g/"␤»
masak ...also, the problem with $/ in .subst blocks...
jaffa4 rakudo: $e~~s:/ /g/;
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "$e~~s:/ /g"␤»
16:00 vmspb joined
jaffa4 rakudo: s:/ /g/; 16:00
p6eval rakudo 924242: OUTPUT«===SORRY!===␤Confused at line 22, near "s:/ /g/;"␤»
sorear jaffa4: might I suggest declaring your variables?
std: $e~~s//g/; 16:01
p6eval std 625303c: OUTPUT«===SORRY!===␤Variable $e is not predeclared at /tmp/vwaCvtG6bh line 1:␤------> $e⏏~~s//g/;␤Null pattern not allowed at /tmp/vwaCvtG6bh line 1:␤------> $e~~s/⏏/g/;␤Parse failed␤FAILED 00:01 120m␤»
masak jaffa4: even if you fix the variable, you shouldn't try to use the null pattern.
jaffa4: the null pattern isn't allowed in Perl 6 regexes. we talked about that yesterday.
jaffa4 yes, that is clear
masak good. 16:02
sorear niecza: s:pos(1)/x/y/ 16:04
p6eval niecza v2-13-g4ebb175: OUTPUT«===SORRY!===␤␤Action method sibble not yet implemented at /tmp/pxxMlS6fwj line 1 (EOF):␤------> s:pos(1)/x/y/⏏<EOL>␤␤Action method quote:s not yet implemented at /tmp/pxxMlS6fwj line 1 (EOF):␤------>
..s:pos(1)/x/y/⏏<EOL>␤␤Unhan…
16:06 cafesofie left
jaffa4 Am I right that recompiling rakudo is slow? 16:08
sorear yes
jaffa4 more than 2 minutes. 16:09
masak oh, probably on most platforms.
sorear I've had it take 600 on me
masak especially core.pir is slow.
sorear the fastest I've seen is 8
jaffa4 8?
8 what?
sorear minutes
600 was also in units of minutes 16:10
xinming Why will $hash = { "a" => 1, $b, $c, %stuff, @nonsense }; be hash, but $code = { "a" => 1, $b, $c, %stuff, @_ }; will be parsed as code?
jaffa4 600 minutes? 16:11
or rather seconds?
sorear 10:10 < sorear> 600 was also in units of minutes
xinming jaffa4: seconds.
sorear xinming: because the latter block has an implicit parameter
xinming: wrong, minutes 16:12
xinming sorear: So, you mean, with $_, @_, %_ will cause the {} be parsed as block?
sorear yes.
xinming Thanks, A bit strange for the first glance. :-) 16:13
jaffa4 what is <?[/]>?
sorear It's always parsed as a block - there's a postprocess that turns parameter-less blocks consisting of a single list of items, the first of which is a pair constructor or %-variable 16:14
jaffa4: Is the next character /? Zero width assertion
Like <?before '/'>, but faster on nqp-rx
16:14 Patterner left
moritz_ sorear: what should s/// return? 16:15
jaffa4 rakudo: "//"~~m/'/' \s* '/'/ 16:16
p6eval rakudo 924242: ( no output )
16:17 Psyche^ joined, Psyche^ is now known as Patterner, envi left
masak moritz_: irclog.perlgeek.de/perl6/2011-01-13#i_3184316 16:22
sorear wonders if smartmathing against * should be more like .so
masak sorear: then what would replace the old * behvaior? 16:23
sorear: ~~ Any ?
um. ~~ Mu ?
sorear True 16:24
moritz_ masak: did you spec that (return value of s///)?
masak sorear: that's currently syntactically disallowd because people misunderstand it.
moritz_: no.
I could, though.
sorear masak: that's problematic from a p5 consistency viewpoint
moritz_ masak: it's not disallowed, it just warns 16:25
masak ok.
sorear if $string ~~ s/.../.../ { is idiomatic
moritz_ and rakudo treats all compile time warnings as errors
sorear (and used in STD)
moritz_ well, then returning $string on success is unreliable
sorear so I'd rather have s/// return true iff there was a substitution
moritz_ fwm too
sorear right. p5 idioms breaking is fine, having them become slightly unreliable is BAD 16:26
moritz_ erm, wfm
masak I think it works for me as well, but I'll need to ponder it a bit more.
moritz_ so s/// becomes do { $_ = $_.match; so $/ } ?
masak it does simplify the 'while s/.../.../' case.
moritz_ uhm 16:27
masak moritz_: well, .subst
moritz_ *danger alert*
while s:g/.../../ will behave quite differently than in Perl 6
thing is, people sometimes want to access the captures from the substituion 16:28
sorear according to perlre, s/// returns the number of substituions made
always
moritz_ so they want something like for s:g/.../.../ -> $/ { say $0 }
sorear so I think I'll just use that in niecza 16:29
moritz_ somehow I feel we can't make it right for everybody
jaffa4 Who should I sent a patch?
moritz_ jaffa4: [email@hidden.address]
with [PATCH] + description in the subject 16:30
masak I'd rather see we tried to work out something consistent than try to please everybody.
moritz_ does s:g[foo] = do { .... } loop the RHS automagically over all matches?
I think so 16:31
jaffa4 what about the web interface?
moritz_ so that would replace the old while(s///g) { ... } idiom
jaffa4: what about it?
masak jaffa4: there's no way to submit patches through the web interface.
jaffa4 that is weird. 16:32
masak I suggest you try to live with it. :)
moritz_ you are entitled to that opinion.
jaffa4 thanks for understanding. 16:33
masak anytime.
jaffa4 well, mail sent. 16:36
masak jaffa4++ 16:38
16:42 kjeldahl left
dalek p-rx/nom: 57d4969 | moritz++ | build/Makefile.in:
missing dependency on nqp-builtins.pir
 16:47
p-rx/nom: e778745 | moritz++ | / (2 files):
add skip() builtin to avoid test-out-of-sequence for skipped tests
moritz_ this makes the test summary in nom a bit smaller :-) 16:48
16:48 Ultali joined 16:50 Sarten-X left 16:51 araujo joined 16:55 Ultali left 16:56 Sarten-X joined
jnthn evening, #perl6 16:57
masak jnthn! \o/
jnthn masak! \o/
moritz_: Thanks for taking care of the smartmatch.t mess. :) 16:58
sorear hello jnthn! 16:59
jnthn o/ sorear 17:01
moritz_ jnthn: no problem
it was the lowest hanging fruit of all the test failures :-)
jnthn :)
Yes, the others are harder targets. :) 17:02
masak jnthn++ # line numbers on errors in Rakudo
tadzik what, where?
jnthn masak: Huh? I did those ages ago :P
masak nevertheless :)
17:02 risou joined
masak they're *really* convenient. 17:02
jnthn
.oO( Some features are a subscription program for karma )
masak exactly. 17:03
17:18 justatheory joined 17:26 Su-Shee joined
Su-Shee good evening all. 17:26
masak evening, Su-Shee
sorear hello Su-Shee 17:27
jnthn o/ Su-Shee :) 17:28
Su-Shee how's the fosdem going for perl6? :) 17:29
dalek ecza: 6d45faa | sorear++ | / (2 files):
Add &succeed and &proceed
 17:35
ecza: f9417bc | sorear++ | / (3 files):
Implement substitutions
sorear moritz_: check the tests for f9417bc and see if there's anything interesting there 17:36
17:36 benabik_away is now known as benabik
benabik o/ 17:36
masak benabik: \o 17:38
17:43 icwiener joined 17:47 mauror joined 17:48 wamba left 17:51 stifynsemons joined 17:52 mauror left
moritz_ sorear: nice, but nothing really unexpected 17:52
17:52 cafesofie joined
diakopter perl6: say Str 17:55
p6eval pugs: OUTPUT«Str␤»
..rakudo 684fdf: OUTPUT«Str()␤»
..niecza v2-13-g4ebb175: OUTPUT«␤»
TimToady masak: actually, I'm starting to think that ? in regex should return Nil for 0 matches, but * shouldn't 17:56
so I think the reasoning should go the other direction 17:57
masak ok.
diakopter so
TimToady Nil is a useful value for the boundary between scalarness and listness
and ? items are perceived both ways
masak aye. 18:01
dalek ecs: a56ffc3 | larry++ | S (2 files):
make ?/* semantics consistently Nil/()
 18:09
TimToady rakudo: 'x' ~~ /(y)? (z)*/; say $0.defined, $1.defined; 18:11
p6eval rakudo 684fdf: OUTPUT«Bool::TrueBool::True␤»
TimToady that should be False True 18:12
by the new rules
masak submits rakudobug
TimToady and actual, the first was supposed to be False already by the old spec
*ly
masak rakudo: 'x' ~~ /<alpha>?/; say $<alpha>.WHAT 18:14
p6eval rakudo 684fdf: OUTPUT«Array()␤»
masak TimToady: so this should be Match now?
TimToady no, it's still a list 18:15
masak so... always Nil or a list of one item?
jnthn TimToady: (by the old spec) how? An empty list is defined?
TimToady we just mark the 0 case as undefined rather than defined
jnthn Or did that change also?
TimToady rakudo: say ().defined
p6eval rakudo 684fdf: OUTPUT«Bool::True␤»
TimToady rakudo: say Nil.defined 18:16
p6eval rakudo 684fdf: OUTPUT«Bool::False␤»
TimToady that is correct
jnthn *nod*
masak TimToady: when $obj.?foo succeeds, it doesn't return a list of one item.
TimToady *shrug*
jnthn So if (y)? by the old spec gave back and empty list, then .defined on it shoulda been true, as Rakudo has it? 18:17
TimToady old spec said Nil for that
jnthn oh
TimToady it just didn't specify that * returned ()
jnthn ah, OK 18:18
I misunderstood what had changed.
TimToady it's possible we can relax /x?/ to not return a list now that Nil and () are different 18:19
esp if .[0] is a no-op on scalars 18:20
rakudo: say 1.[0] 18:21
p6eval rakudo 684fdf: OUTPUT«1␤»
TimToady so again, where .?meth and x? are inconsistent, it's probably the regex that wants to change 18:22
I'd like to let pmichaud++ comment on that notion first though
18:22 stifynsemons left
masak it's been proposed before. 18:25
TimToady it makes more sense now that we have the Nil/() distinction
18:26 dual left
TimToady the main difference is +$0 would do something different 18:26
masak I'd never do that on <rule>? anyway. I'd do ?$0 18:27
TimToady rakudo: say +Nil
p6eval rakudo 684fdf: OUTPUT«0␤»
TimToady arguably, +Nil should warn, while +() shouldn't
because the idea of Nil is that there's a least one less value than you expected 18:28
which is why listy things return () instead, if 0 items is permissable 18:29
Nil is the absence of an expected value
masak aye. 18:30
TimToady (ignoring the fact that using the ? quantifier means you are expecting 0 values :)
masak was just going to say that :P 18:31
18:38 dual joined 18:43 cspencer joined
dalek kudo: 9898837 | pmichaud++ | src/Perl6/Actions.pm:
Fix RT #83420; statement_control:<if> returns Nil if no else clause given.
 18:44
jnthn yay pmichaud++ patch! 18:45
masak \o/ 18:48
18:50 cotto joined 18:54 Raven left
sorear TimToady: would it be ok for Nil !~~ Any? 18:56
S02 talks about magical behavior for assigning Nil to variables
I don't like magical assignment behavior unless it can be done in the type check slow path
(c.f. Parcels)
masak somehow likes Nil !~~ Any 18:57
"it's not even anything!"
moritz_ installs a cronjob for std rebuilding on p6eval 19:09
colomon moritz_++ 19:16
sorear std: 1 [+] 2
p6eval std 625303c: OUTPUT«ok 00:01 119m␤»
jnthn std: 1 [[[+]]] 2 19:19
p6eval std 625303c: OUTPUT«ok 00:01 119m␤»
jnthn rakudo: say 1 [[[+]]] 2
p6eval rakudo 684fdf: OUTPUT«3␤» 19:20
jnthn :)
19:20 cafesofie left
masak rakudo: say 1 >>[[[[[RRRRR+]]]]]<< 2 19:20
p6eval rakudo 684fdf: OUTPUT«===SORRY!===␤Could not find sub &infix:<[[[[[RRRRR+]]]]]>␤»
masak huh.
moritz_ rakudo: say 1 [[RR+]] 2 19:21
p6eval rakudo 684fdf: OUTPUT«3␤»
19:22 dorlamm joined
moritz_ rakudo: say 1 [RR[RR+]] 2 19:22
p6eval rakudo 684fdf: OUTPUT«3␤»
moritz_ rakudo: say 1 >>[RR[RR+]]>> 2
p6eval rakudo 684fdf: OUTPUT«===SORRY!===␤Could not find sub &infix:<[RR[RR+]]>␤»
moritz_ seems it doesn't like that hyper'ed
jnthn hmm 19:24
masak rakudo: say 1 >>R+<< 2 19:26
p6eval rakudo 684fdf: OUTPUT«3␤»
masak rakudo: say 1 >>[+]<< 2
p6eval rakudo 684fdf: OUTPUT«===SORRY!===␤Could not find sub &infix:<[+]>␤»
masak seems to be hyper and reduction together.
masak submits rakudobug
jnthn oh
that's not a reduction 19:27
that's a bracketed infix
Or should be....
benabik Does >>[+]<< Make sense? Isn't [+] a listop, not a binop? [+] 1, 2 not 1 [+] 2
colomon but rakudo screws it up
If you use [+] as an infix, it's just +
benabik Non-intuitive, but okay.
jnthn The [+] in >>[+]<< is not related to [+] 1,2,3 style things. It's just the same as >>+<<
Yeah, I was like "wtf" at first :) 19:28
colomon benabik: I'm not sure it isn't non-intuitive and not okay. :)
It seems a very very confusing overload.
jnthn colomon: Yeah, not sure why it screws it up
masak jnthn: oh, true.
benabik colomon: I wouldn't make it work that way, but I'm not the BDFL. I'm willing to accept there may be reasons for it, even if I can't think of any. 19:29
colomon and we've already seen a bug where someone accidentally left out an operator and converted a desired [+] reduction into normal infix:<+> with a list on the right-hand side.
jnthn: parser issue? 19:30
rakudo: say 5 R[-] 4
p6eval rakudo 684fdf: OUTPUT«-1␤»
jnthn colomon: I suspect the munging in the actions.
sorear needs to get hypers working in niecza 19:33
19:37 stifynsemons joined, shortcircuit left
masak rakudo: my %info = x => 1, y => 2; my ($x, $y) := %info; say ($x, $y).perl 19:43
p6eval rakudo 684fdf: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in 'infix:<:=>' at line 1␤ in main program body at line 22:/tmp/kudgcUuXVo␤»
masak jnthn: should something like that work?
xinming perl6: Num.new(5).perl.say;
p6eval niecza v2-15-gf9417bc: OUTPUT«(timeout)»
..pugs: OUTPUT«*** Must only use named arguments to new() constructor␤ Be sure to use bareword keys.␤ at /tmp/KPV3d9A7wS line 1, column 1-20␤»
..rakudo 684fdf: OUTPUT«0␤»
jnthn masak: Can't remember if you need to explicitly capture-ize the RHS
19:43 wiseguyxp joined
xinming rakudo: Num.new(5).perl.say; 19:44
p6eval rakudo 684fdf: OUTPUT«0␤»
xinming hmm, Why won't this returns the object please?
Just curious.
jnthn rakudo: my %info = x => 1, y => 2; my ($x, $y) := \%info; say $x, $y;
p6eval rakudo 684fdf: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in 'infix:<:=>' at line 1␤ in main program body at line 22:/tmp/LsMKGXbXkQ␤»
jnthn er, wtf
I'm sure that _did_ work...
masak yeah 19:45
jnthn oh, wait
rakudo: my %info = x => 1, y => 2; my (($x, $y)) := %info; say $x, $y;
p6eval rakudo 684fdf: OUTPUT«12␤»
jnthn ah.
That's telling.
sorear xinming: Num doesn't have a user-accessible construcot
19:46 kaydsoft left
moritz_ it should have 19:46
19:46 rgrau joined
masak jnthn: can I submit as a rakudobug that it doesn't work without the double parens? 19:46
moritz_ just for consistency sake
jnthn masak: Check if the spec says it should first.
masak oki.
jnthn masak: I sorta expected it to. 19:47
19:47 stifynsemons left
xinming sorear: So, Int, Num such kind of types don't have constructor? 19:47
masak jnthn: hm.
(:@even, :@odd) := |classify { $_ % 2 ?? 'odd' !! 'even' }, @list;
19:47 cspencer left
masak seems the | is needed, too. 19:48
rakudo: my %info = x => 1, y => 2; my ($x, $y) := |%info; say ($x, $y).perl
p6eval rakudo 684fdf: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in 'infix:<:=>' at line 1␤ in main program body at line 22:/tmp/1QbM2ij3UZ␤»
masak but that doesn't work either, so... :)
masak submits rakudobug
jnthn I'm not sure | knows what to do there
oh, or maybe it does...but does something wrong, hmm
I bet there's some screw-up somehwere in coercion to capture.
colomon xinming: They don't right now, but that is most likely a bug in Rakudo, not an actual design decision. 19:49
xinming Ok, thanks.
jnthn Note that the way you get from native 2 to an Int object representing 2 isn't a constructor call. But there's no reason not to support a constructor style too.
sorear If something isn't specced, you can't assume it exists
moritz_ well, in the long run the correct way to create a Num object is to box a low-level equivalent 19:50
but what jnthn++ said
jnthn Right, but Perl 6 does auto-boxing. :)
So you don't have to. :)
sorear Rakudo and Niecza both recognize Num.new but fail, since it doesn't actually have any useful meaning
19:50 [Coke] left
jnthn sorear: Right. 19:50
sorear: They just hit the default constructor.
moritz_ niecza: Num.new(3) 19:51
p6eval niecza v2-15-gf9417bc: OUTPUT«Unhandled exception: Excess arguments to SAFE Mu.new, used 1 of 2 positionals␤ at /home/p6eval/niecza/lib/SAFE.setting line 0 (SAFE Mu.new @ 0)␤ at /tmp/DK9rOlq0Aq line 1 (MAIN mainline @ 1)␤ at /home/p6eval/niecza/lib/SAFE.setting line 1044 (SAFE C435_ANON @ 2)␤
..at /home/p6eva…
sorear the default constructor builds Num objects with a P6opaque representation, which the rest of the runtime is sadly not very robust about handling
that's part of what I've roadmapped to fix this month
19:52 sji joined 19:53 shi left
TimToady colomon: 1 [+] 2 now warns, but std hasn't updated itself 19:56
niecza: 1 [+] 2
p6eval niecza v2-15-gf9417bc: OUTPUT«Potential difficulties:␤ Useless use of [] around infix op at /tmp/AOZZgjrWwJ line 1:␤------> 1 ⏏[+] 2␤␤»
colomon TimToady: and no one has fixed Rakudo to have the same warning, so far that I know. :) 19:57
19:58 risou left 19:59 coldhead joined 20:08 Chillance joined 20:09 [Coke] joined 20:10 stifynsemons joined 20:14 Mowah left 20:16 dukeleto left 20:17 dorlamm left, dukeleto joined
moritz_ masak: am I to understand that i need at least the first two volumes of "Winning Ways" to get to the interesting games? :-) 20:20
masak moritz_: I have the first volume checked out from the library. it wouldn't cross my mind to call it devoid of interesting games. :)
but, by all means, get all four volumes. 20:21
the first two are a bit more "theoretical", the latter two a bit more "practical".
20:24 vmspb left
moritz_ well, they aren't cheap :( 20:24
masak if only there was some way to win books...
moritz_ well, I've put it on my wishlist and hope that I didn't mess up p4 too badly :-)
masak :)
in the interests of suspense, I won't more than smile for now. 20:25
arnsholt moritz_: Well, apparently noone who actually submitted on time did the perfect play variant =) 20:26
So I'd assume you're still in the running
20:29 solarion left, solarion joined
colomon I don't know how poor masak is going to decide between different broken implementations. 20:29
inter-version duel to the death?
masak I'm thrilled that people did badly on at least one task. 20:30
otherwise I'd have started to doubt the level of the problems.
colomon masak: don't know if you saw my comment yesterday. Turns out that if I'd just worked a couple more hours on p4, I'd probably have noticed the 34-repeat pattern. :) 20:31
masak colomon: no, I missed that comment.
masak looks
colomon that was the entire sum of the comment, no point looking for it now. :) 20:32
masak found it.
colomon I actually got far enough to hit the repeat in the code, but not far enough that the repeat was obvious to me. 20:33
masak ok.
colomon just doing it via shear brute force, mind you, no smarts involved.
masak counterfactuals are always fascinating. :)
colomon I'm not asking for credit for it, believe me. :)
masak other contestants might have been close to cracking it and not known it, too :)
colomon naturally.
what caught my attention is that I worked on the problem for probably something like 40 hours. 20:34
masak wonders how long arnsholt spent on the problem
arnsholt Good question
colomon I thought I was chasing down completely the wrong path, but I was actually really close to having it worked out. But close doesn't count. 20:35
arnsholt I spent quite a bit of time on the research, and just figuring out how this octal game business worked
colomon arnsholt++ 20:36
arnsholt Once I managed to wrap my head around it, it turned out that the user-facing code took more time than the AI part proper
masak :)
...which shows that the problem was in the same class as the others: "easy if you know how" 20:37
arnsholt Which, TBH, I think is in part because the AI code is very abstract
My gut feeling is that with a differently designed AI code the UI part could probably be simplified
moritz_ actually did two DP attempts, but each time failed to properly combine the sub-solutions 20:40
masak the combination operations are quite non-obvious and specific to this type of games, I would say. 20:41
arnsholt Definitely. I don't think I'd gotten it without Wikipedia
There's a part of me that wants to know how that stuff falls out of the theory 20:42
masak it's the "unknown unknowns" that prevented most people from reaching for Wikipedia.
arnsholt: how do you mean?
arnsholt The other part of me suspects that it's really, really hairy =)
masak (maybe I can help) :) 20:43
arnsholt Actually, your posts were quite helpful =)
masak the only thing I don't understand is where 34 comes from :P 20:44
arnsholt Good question. I guess it's either completely arbitrary, or hsa to be like that because of deep and inscrutable number theory =) 20:45
masak I don't feel comfortable with either answer...
arnsholt A true mathematician's answer 20:46
20:48 ofir joined
masak arnsholt: I'm looking at lib/Kayles.pm for the first time. nice code. 20:48
20:49 shortcircuit joined 20:50 kjeldahl joined 20:53 wooden left
arnsholt Thanks! I'm quite happy with that part of the code, actually 20:53
20:53 wooden joined
arnsholt The theory turned out to be pretty straightforward to put into code 20:53
You don't have to look too much at the UI bits ;) 20:54
masak :)
moritz_ masak: maybe you can review arnsholt's code for p4 too (and just mention it's post deadline) 20:57
20:57 mikehh left
masak moritz_: yes, I'll end up doing something like that. 20:58
note that even if arnsholt had made the deadline, which he deplorably didn't, he'd still have sent in only four solutions.
huh, someone on parrot-dev just released something called 􏿽xE2􏿽x84􏿽xA6􏿽xCE􏿽xB7;)XD
plobsing++ 20:59
moritz_ nods
21:00 Axius joined 21:04 Chillance left 21:15 kaydsoft joined 21:16 kaare_ left 21:17 jasonmay is now known as JasonMay 21:18 kaare_ joined 21:21 kaydsoft left 21:24 Axius left 21:31 JasonMay is now known as jasonmay 21:32 kaydsoft joined
ofir is perl5 expected to be totally deprecated after perl6 is officialy released ? 21:35
tadzik no
jnthn No
masak ofir: Perl 6 has been officially released.
ofir oh nice to know
masak and Perl 5 hasn't been totally deprecated. 21:36
ofir so those are 2 different dev branches ?
masak exactly.
you got it in two guesses :)
ofir heh
masak nicely done.
mathw perl 6 isn't backwards compatible with perl 5, so we can't declare that perl 5 is gone
the perl 5 community would hate that anyway :)
masak yeah, they'd be pretty pissed off, I think.
heck, I'd be a little angry. 21:37
jnthn :)
jdhore Perl 6 seems VERY nearly 100% backward-compatible to Perl 5 to me, but what do i know.
21:37 pmurias joined
masak jdhore: well, they're not totally dissimilar. but saying that they're close isn't exactly it either. 21:37
ofir like anything in life, it is a matter of time until Perl 5 will no longer be maintained
masak jdhore: it was more true around 2002-2003. 21:38
diakopter ofir: most would disagree with that
ofir diakopter, why, what's the point in the duplicites of two different languages in one langauge ?
sorear perl 5 will die
very slowly
mathw jdhore: there's really not much perl 5 you can feed to a perl 6 compiler and have it work 21:39
sorear it will probably still be in use 40 years from now
ofir diakopter, it's as if I'll tell you that you can still burn CD's, and floppy disks, but you wouldn't
diakopter it's installed just too many places, and too many places rely on it for building/bootstrapping other software that must live on forever
benabik Perl 6 will replace Perl 5 the same way C++ replaced C.
sorear I assign 80% odds that perl 5 will reach "dead language" status before Perl 6
masak benabik: that is to say, "not really"? :)
diakopter ofir: not even close.
masak ofir: what diakopter said.
ofir: I'm an ardent Perl 6 user, and I use Perl 5 quite a bit. 21:40
diakopter there are tens or hundreds of perl5 installations, and I fully agree with sorear that it'll still be in use 40 years from now.
masak ofir: some of the reasons is of the "at this point in time" kind, but not all of it.
ofir masak, until one day it'll give you the creeps and you're mind will be split to 5 & 6 chunks, and you'll decide that there's a room for one perl only
masak diakopter: s/hundreds/hundreds of thousands/? 21:41
diakopter in an individual's mind, maybe, but that has nothing to do with the market as a whole
mathw hundreds of thousands, more like
jdhore Also, it'll prolly take a very long time till most of the CPAN is P6
diakopter masak: bwahaha; I meant to say tens or hundreds of millions :D thanks
masak diakopter: :)
mathw and many, many of those aren't even a very recent Perl 5
masak ofir: actually, I feel I'm moving in the other direction.
21:41 wallberg joined
mathw (which is a big shame) 21:41
masak ofir: I'm learning more and more programming languages, and Perl 5 and Perl 6 keep coming out as my favorite two. 21:42
Su-Shee ah well that old perl 5 feeds 55 people here for example. ;)
21:42 wallberg left
jdhore ascent_, for me, that's Perl 5, Perl 6 and C 9in no particular order) 21:42
*(
21:43 wallberg joined
diakopter .( 21:43
mathw Perl 5 skills pay part of my salary...
a very small part, but part :)
jdhore er, i highlighted the wrong person, i suck.
masak, ^^
mathw mostly, I get paid for being very patient
:P
Su-Shee well it boils down to environment.. every time I try something else but perl I'd either have to use java or python to get something even resembling CPAN's power. 21:44
mathw Mmm I recently got the freedom to do a script with as many CPAN modules as I wanted/needed
found some very helpful ones indeed 21:45
Su-Shee last weekend I got stuck with something as stupid as javascript's \w not working properly with umlauts.
mathw although there wasn't one which did precisely what I wanted, which was disappointing because I thought there might be (or at least, I couldn't find one)
Su-Shee: eeww
Su-Shee yeah that's what I said but with more words ;)
pmurias mathw: what did you intend to do?
21:47 pmurias left
diakopter ofir: last I looked, even git depends on perl... (someone correct me where I'm wrong) 21:49
ofir diakopter, correct 21:52
diakopter (my point is that all the (tens/hundreds? of) thousands of free/open-source software packages/projects that depends on Perl 5 (NOT TO MENTION the 100x that amount (probably, by lines of code) of unseen/corporate/private Perl 5 code) would all need to recursively remove/port all those Perl 5 dependencies, or go dormant/obsolete themselves, in order to satisfy your claim)
Su-Shee yeah. if it's out there, it's out there. ;) 21:53
diakopter (by a wide definition of "maintained") 21:54
for example.
Su-Shee I don't want to know how many perl scripts are silently running for a decade noone ever touching them until they get forgotten..
ofir diakopter, well I guess you are right, it doesn't count by number of weeks nor months, but It must happen sometime
diakopter lifetimes 21:55
Su-Shee ofir: there is no "must happen" - a perl renaissance could happen as well.
diakopter (the other half of my point is that if there are so many softwares dependent on it, it will be "maintained" *enough* for them, b/c they need it) 21:56
dalek p-rx/nom: ad21c53 | jonathan++ | / (5 files):
Role composition for when a class just does a single role (we don't do the role summation case just yet). Seems method and attribute composition works, methods in the class beat ones from the role and mentions of $?CLASS are correctly generic.
ofir diakopter, and who guarantees that support & maintenance ?
masak liked the phrase "recursively remove/port all those Perl 5 dependencies" 21:57
that's not really true, though. Perl 5 could stop being maintained, and all that code could still exist out there. 21:58
just as some people still run Perl 4 code, even though Perl 4 isn't being maintained.
diakopter eh, I was using a wide definition of maintained
masak does it matter? unmaintained Perl versions don't just disappear... 21:59
BinGOs Perl5::ButMaintained
diakopter like, "still available to download", and "folks available/willing to be paid to consult on it privately to maintain/install/support it if it's not publicly maintained"
masak oh, ok. 22:00
22:00 wallberg left
diakopter i guess another portion of my argument would need to be that those private supporters would help keep it public too. 22:00
(seems reasonable to me) 22:01
22:03 dakkar left 22:06 icwiener left 22:10 kjeldahl left
dalek rixel: 86dcfe6 | diakopter++ | / (9 files):
[perlesque] (compiler) yet another iteration of improvements to generics on mono
 22:12
diakopter sorear: re "80% odds"... that's interesting 22:13
22:19 cosimo joined
sorear jnthn: Any progress on subset/native types in 6model? 22:19
22:19 cafesofie joined
jnthn sorear: 'fraid not. I had a lack of tuits last week to do much Perl 6 stuff at all. Much of the thinking time I have had has gone into how to factor roles and trying to work out a roadmap for the rakudo/nom branch. 22:21
sorear: Though I think roles may soon make me hit the same issue that subset types bring up in terms of the type cache.
sorear jnthn: any type 'does' a finite number of roles
but an infinity of subset types 22:22
jnthn sorear: That's not exactly true.
sorear ?
jnthn In the non-parametric case you're right.
but PR[S] ~~ PR[T] is true where S ~~ T
It's still finite in a sense 22:23
But not automatically in a useful way.
Well, the totally non-parametric case of roles doesn't actully exist. 22:24
Since any mention of $?CLASS is generic.
But you get a certain amount of simplification if you only handle that case.
xinming What is twigil mean please? 22:30
jnthn $!foo $*bar $.baz # in these, the twigils are !, * and .
xinming jnthn: thanks
jnthn It's a "secondary sigil" essentialy
benabik Name is derived from sigil number two, or something similar. 22:31
xinming Ok, got it. thanks
masak 第二个sigil :) 22:33
diakopter chimchimney 22:35
masak I don't even see the characters anymore. to me, they're just blonde, brunette... 22:36
jnthn That's kinda shallow. They all have personalities and stuff. 22:37
masak jnthn: www.imdb.com/title/tt0133093/quotes?qt0324317 22:41
jnthn Oh!
masak you might have seen it... :) 22:42
jnthn missed that reference completely, and was trying to work out why masak had randomly context-switched... :)
masak (too bad they never made any sequels)
diakopter or prequels 22:44
22:44 garo joined
tadzik yeah 22:44
sorear I'm thinking parametrized roles need to have a little bit of the subset-nature 22:47
jnthn sorear: In so far as "a simple type cache won't get us all the way", certainly. 22:48
sorear: The main issue at the moment is that the type cache is "all or nothing". 22:49
Which isn't really ideal.
I just didn't settle on exactly how to change it yet.
22:49 garo left 22:50 am0c joined, kaare_ left
jnthn We'd really like "is this some kind of Positional/Callabe/Associative" to be fast. 22:50
At the moment it's painfully not.
jnthn wonders if NQP roles should supporting punning or just demand they be composed into something 22:52
23:05 drbean joined 23:06 leprevost joined 23:07 leprevost left 23:08 leprevost joined 23:11 leprevost left, jql joined 23:12 benabik is now known as benabik_away
masak '♞, #perl6. 23:16
23:16 masak left 23:17 jaffa4 left 23:18 guidj0s joined
colomon "black knight, #perl6"?? ;) 23:19
arnsholt White knight on my end =p 23:21
23:22 guidj0s left
jnthn sleep & 23:22
23:27 orafu joined 23:29 am0c left
diakopter TimToady: knap01 works now on mono :) 23:32
3 seconds on the p6eval machine from the shell, but oddly 30 seconds from p6eval
I'm trying to diagnose why
hm, maybe p6eval needs another execution mode ;) 23:33
...at least for perlesque
23:36 MayDaniel left 23:39 jasonmay left 23:49 rgrau left, hudnix joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!