6.2.10 released! xrl.us/hxnb | geoffb's column (/. ed): xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: feather.perl6.nl:8888/ | www.geeksunite.net
Set by Juerd on 20 October 2005.
Ovid HI all. 00:23
Anyone here? 00:24
clkao *yawn* 00:25
Ovid I was beginning to feel lonely.
I assume that pugs does not yet automatically support automatic method creation for class variables? 00:26
svnbot6 r7708 | Ovid++ | Stub recipes for object creation and destruction. 00:29
r7708 | Ovid++ | I was blocked fairly quickly on recipe 13-03 (not committed) because PUGs
r7708 | Ovid++ | apparently does not yet support autogenerating accessors for class instance
r7708 | Ovid++ | variables (or Im doing it wrong somehow)
clkao commit tests and go to sleep, you shall see them implemented the next day 00:40
svnbot6 r7709 | Ovid++ | A couple of array recipes 00:53
Ovid Looks like a lot of those tests are already there and they're failing. 00:59
Night all
eric256__ anyone remember a while back, autrijus mentioned needing to add stub subs to prelude.pm in some cases? yea thats vague i know. ;) 02:14
hey the logs have a search button!! never mind found it 02:17
02:17 eric256__ is now known as eric256
xinming hmm, what are the differences between `class A { has $.a }; class B { has $a }; ` 02:32
hmm, or ` $.a =:= $a ` ? 02:33
though, don't know if it would be valid to use class A { has $a; } 02:34
eric256 the difference is still in flux 02:38
but last i read $.a meant "make public accessors for this" while has $a; meant "leave this a private attribute"
ping Juerd 02:51
juerd, in case you backlog....building pugs seems to hang on feather...i think. maybe...not realy sure. looks like i might have three concurrent builds at the moment 02:53
xinming eric256: hm, In fact, $:a is a private attribute.... 02:55
eric256 or i'm an idiot and forgot my account is used for the auto build. ;) 02:56
from recent rumblings on the list that might change.
so that $.x is public and just plain $x is private...which makes more sense in my head but who knows what will be the next finaly decision
Khisanth it seems nothing is ever final :) 02:57
eric256 nope. ;) thats part of the fun
Khisanth makes it unappealing to use though
eric256 to use? why would you think its ready for use? 02:58
its not. and if that makes it un appealing then good
autrijus: a while back (while trying to fix shift/pop/min/max to only work on lists) you mentioned using a stub in prelude.pm 03:04
i tried that today, but now it hands on 42.shift; any thoughts? 03:05
eric256 whistles and hums while waiting for pugs to compile ;) 03:19
ingy cpan-upload -mailto [email@hidden.address] -user INGY -password ******** Perl6-Bible-0.21.tar.gz 03:29
Finished!
eric256 cool
ingy why does `use No::Such::Module` crash the pugs shell? 04:16
yet syntax errors do not
eric256 dunno. oddly i just did that a minute ago too ;) 04:17
ingy wonders how to use yaml in pugs
ingy greps a little and gives up 04:22
eric256 lol.
what error do you get?
ingy sleep&
eric256 alright. later
gaal : FEAR: The new features won't be used by people 05:12
FEAR: The new features *will* be used by people!
dduncan FEAR: both of those will happen at the same time ... in different universes ... which one will be ours? 05:16
spinclad both of those will happen at the same time... by different people, or to different features 05:58
'use every feature' is not a recipe for excellent programming 05:59
(especially junctions!)
wolverian junctions might be nice for automatically generating arguments in tests 06:00
spinclad .oO ( if junctions are meant to be a step up from Quantum::Superposition, why don't they have complex weights? )
06:17 brother_ is now known as brother
buu Oh dear god. 06:39
wolverian hm? 06:40
buu Weighted junctions
wolverian heh. 06:41
10:28 r0nny_ is now known as r0nny
svnbot6 r7710 | renormalist++ | - The "documentation will never be rewritten for Perl6" fear 11:51
brother What would be the natural score in p6 golf? number of unicode glyphs or number of utf-8 chars? 13:16
Limbic_Region I think it still has to be keystrokes 13:28
so using unicode glyphs may be a penalty rather than a gain
how many keystrokes does it take to generate the zip glyph?
brother depends on how you keyboard is set up 13:29
theorbtwo Depends how you generate it.
For me, three. For a Japanese user, probably 1.
I tend to think the opcount is a better measure for both perl5 and perl6, though it's not as visually obvious.
brother theorbtwo: No. one of the wonderfull thing in golf is to (ab)use precedence 13:32
theorbtwo Then again, I never much cared for golf, and it'd make eval "..." count as the best way to do everything. 13:36
svnbot6 r7711 | juerd++ | +=1 13:52
theorbtwo wonders what that log entry is supposed to mean. 13:53
Juerd One more. 14:04
rafl_ karma points for Juerd, right. ;-) 14:12
jabbot rafl_: points for Juerd, right. ;-) has neutral karma
rafl_ karma jabbot
jabbot rafl_: jabbot has neutral karma
rafl_ karma Juerd
jabbot rafl_: Juerd has neutral karma
theorbtwo karma juerd
jabbot theorbtwo: juerd has karma of 1
rafl_ Well, maybe not. :-)
karma rafl 14:13
jabbot rafl_: rafl has neutral karma
theorbtwo Did jabbot's karma counting get reset awful recently?
rafl_ Hm.. long time no commits..
theorbtwo jabbot, uptime?
jabbot theorbtwo: That is interesting. Please continue.
eric256 morning 14:18
theorbtwo G'morning, eric256. 14:19
eric256 is there something super special about prelude.pm? because i have a fix for an issue that works in a seperate file, but when i try to compile it in prelude.pm it doesn't work anymore. 14:24
autrijus yeah, see the comments on top of Prelude.pm
eric256 ahh... now to wait for another compile and see if that works ;) 14:26
?eval sub x (@array) { shift @array}; x(1); 14:27
evalbot_7711 1
eric256 why does it call x? shouldn't that mismatch the signature? 14:28
autrijus it should. that is a bug. 14:29
theorbtwo Probably the same bug as...
?eval 1.shift 14:30
evalbot_7711 Error: No compatible subroutine found: "&shift"
theorbtwo Oh.
eric256 i fixed that one ;)
theorbtwo Oh, probably not.
OK then.
eric256 i was working on the other part....
?eval my @x = (1..5); print shift @x, " hello";
evalbot_7711 Error: No compatible subroutine found: "&shift"
eric256 i can make that work by adding sub shift (@array) { shift @array }; but that makes 42.shift start working wrong agian ;)
another oddity i found. 14:31
?eval shift (1);
evalbot_7711 Error: No compatible subroutine found: "&shift"
eric256 ?eval shift (1,);
evalbot_7711 1
eric256 ?eval shift (1,2);
evalbot_7711 Error: Can't modify constant item: VUndef
eric256 middle one shouldn't work either, but it does
autrijus why not? 14:35
eric256 why would (1,) work when (1,2) does not? they should be identical shouldn't they? 14:36
wolverian autrijus, they're lists! they can't be shifted! 14:37
or mutated in any other way
autrijus ah. right. 14:38
?eval shift(1,)
evalbot_7711 Error: No compatible subroutine found: "&shift"
autrijus ?eval shift (1,)
evalbot_7711 1
autrijus autoenref
eric256 huh? what about
?eval shift (1,2) 14:39
evalbot_7711 Error: Can't modify constant item: VUndef
eric256 ?eval shift (1,2,)
evalbot_7711 Error: Can't modify constant item: VUndef
autrijus ?eval 1.shift
evalbot_7711 Error: No compatible subroutine found: "&shift"
autrijus so (1,) evaluates to type List 14:40
autoenref into Array
shifts correctly
wolverian wow, that is headache
autrijus it is
not sure if it's worth it
eric256 but (1,2,) doesn't?
autrijus eric256: there is a bug on either side.
wolverian please p6l about it if you have the energy :)
autrijus yes :)
eric256 ahh.. i thought you were saying it was right. ;)
is there somewhere easy i can look for the sub (@a) accepting a scalar? 14:41
autrijus nb: a scalar containing an array ref 14:43
eric256 what? 14:44
i want to look in the source for a way to fix sub shift (@array) accepting a scalar.
robkinyon why? 14:45
wolverian eric256, it must accept arrayrefs, which are scalars
eric256 but it shouldn't accept "hello" or 1
Juerd wolverian: That's not entirely true.
wolverian: It can expect an Array, and thus provide Array context
This is a form of scalar context, but you can look at it in two different ways:
1. arrays are auto-reffed
2. arrayrefs are auto-dereffed
wolverian right 14:46
Juerd Which one occurs internally is irrelevant
wolverian that's what I meant
it has to accept a subcontext of scalar
is that true?
eric256 either way. sub x (@array); x(1); should be an error right?
Juerd Yes
1 can't coerce to an array.
wolverian argh. this container/value context dichotomy makes me cry
(as in, it hurts my head to talk about two kinds of contexts.) 14:47
Juerd wolverian: Yes, that is true, but not all too relevant for this discussion :)
wolverian true. sorry. :)
Juerd There aren't two kinds of context.
There are infinite.
eric256 yea i don't kow anything about containers and values :) i'm happy in my bliss. either way 1 isn't an array or list so i'm fine ;)
Juerd item and list context each have infinite subcontexts
wolverian Juerd, I meant two classes of context. erm. that's even worse way of putting it 14:48
Juerd, item/list versus type
Juerd eric256: 1 is a list in list context, much the same way anything is a list in list context
something is a list iff it is in list context.
There are no lists that aren't in list context
And there are no non-lists that are in list context
Well, elements of a list :) 14:49
wolverian: type contexts are subtypes of item/list
wolverian: Array context is item/Ref/Array
This hierarchy is pretty simple
wolverian hm. right.
Juerd There is item, with beneath it: Str, Num, Bool and Ref 14:50
And everything else is directly under Ref
List has the same substructure
Oh, and there is Scalar, which equals item but changes type.
Or you could see it as directly under item 14:51
As there's also Scalar list context
wolverian how does one specify list context?
*@foo?
Juerd Yes
wolverian okay. 14:52
Juerd Hence the alias "slurpy" context
wolverian sorry for being so thick about this.
Khisanth is that the left or the right?
Juerd This is in signatures only
Outside signatures, * is splat, which doesn't exactly equal providing list context.
wolverian hm. I thought assignment was unified somewhat with signatures, at some point
Juerd Also, I believe the "list" operator is weird, wrong, and I consider it non-existent.
wolverian my (*$foo) = @bar; 14:53
Juerd And, of course, containers provide list context to the RHS of the =
collection containers at least
wolverian: No, that's aliasing
wolverian: The := operator has a signature on its LHS, as the signature in a sub, but without parens.
wolverian does the * do anything there?
Khisanth then what are := and ::=? 14:54
wolverian Juerd, ah. okay.
Juerd Khisanth: For binding/aliasing
Khisanth * on the left side of an assignment retains it's p5 meaning?
Juerd No, typeglobs go away.
*$foo = @bar is a syntax error, afaik.
Khisanth sounds pretty darn similar 14:55
Juerd * is valid in signatures and sub calls (symmetrically, also the RHS of binding)
robkinyon eric256: wouldn't it be enough to ask "if ( $scalar does Array )" ?
Juerd robkinyon: The signature should fix this.
robkinyon: Not an explicit check
PerlJam Khisanth: * is always either multiplication or "flatten"
eric256 robkinyon....sure. but the scalar should never be passe din and that sub never called 14:56
robkinyon so, there should be a shift( @a Array ) and a shift( $c Array ) ? 15:00
Juerd robkinyon: What is this syntax?
There should be a single shift(@a)
svnbot6 r7712 | juerd++ | more fears; markov++
eric256 wouldn't shift (Array @a) be redundant?
Juerd This handles: shift(@array), shift($arrayref), @array.shift and $arrayref.shift
Possibly it has modifiers like :n(5) to shift 5 at a time, but that's outside the scope of this 15:01
wolverian OT: why does (foo){1,3} only capture once in perl5? 15:03
rather, does perl6 do that too? :)
robkinyon oops
i flipped the type and the variable. :_)
eric256 ?eval sub x ($x) { say $x}; x(1); x (1,2,3,4); x(1,2,3,4); 15:06
evalbot_7712 1 1234 Error: No compatible subroutine found: "&x"
eric256 ?eval sub x ($x) { say $x}; x(1); 15:07
evalbot_7712 1 bool::true
eric256 ?eval sub x ($x) { say $x}; x(1,2,3);
evalbot_7712 Error: No compatible subroutine found: "&x"
eric256 so it does work the other way around. blah
Juerd wolverian: Because that's the only way to make sure you know beforehand which capture variables there will be.
wolverian: Count the left parens. Those are the $n numbers.
wolverian Juerd, right. 15:08
Juerd I know no way around this.
wolverian since $/ is a tree of matches anyway, it can just hold an arrayref 15:09
kolibrie wolverian: in Perl 6, all three values will be captured and put in an array ref
wolverian kolibrie, ah, like I assumed. thanks!
kolibrie :)
PerlJam wolverian: All quantifiers will capture into an array. 15:10
wolverian yay!
PerlJam (this week :) 15:11
wolverian that is much saner than perl5.
Juerd It's more useful 15:12
But imo, equally sane
wolverian hm. perhaps. 15:13
PerlJam Juerd: I don't know ... the perl5 behavior exposes an implementation detail to the user rather than doing something more useful.
I guess that was the sanest thing to do though
wolverian TMTOWTBS (There's More Than One Way To Be Sane) 15:14
do rules have methods? 15:16
eric256 to their madness?
eric256 couldn't resist
wolverian or: are named rules first class citizens outside rules? 15:17
Grammar.rule($string) or so. 15:19
PerlJam wolverian: perl6 continues to follow the principle of least surprise. 15:22
wolverian PerlJam, was that a yes or no? :) 15:23
PerlJam yes, but it would be spelt more like $string ~~ Grammar.rule I think 15:24
eric256 hmmm. stubing shift doesn't seem to work at all
or my computer just dislikes me greatly ;)
PerlJam I don't know that it's been specced anywhere how the "subroutine form" of rules work. 15:25
Perhaps Grammar::rule($string,@args) ?
eric256 ?eval sub shift (@array) { shift @array }; sub shift ($s) { die }; my @x = (1..5); say shift @x, " hello"; 15:28
evalbot_7712 1 hello bool::true
eric256 ?eval sub shift (@array) { shift @array }; sub shift ($s) { die }; my @x = (1..5); say shift @x, " hello"; shift 42;
evalbot_7712 1 hello Error: Died
theorbtwo ?eval rule r {42}; \&r
evalbot_7712 Error: Undeclared variable: "&r"
eric256 but dropping those two definitions into prelude.pm and marking them as 'is primitive' doesn't change anything at all.
btw in prelude.pm i marked them with * to put them global... even tried adding multi in front, it just doesn't like them 15:30
ehh don't 'mind me...if you have pugs open on windows during compile it doesn't replace the existing pugs...good to know 15:31
?eval my @x = (1..5); my @y = splice(@x,5) 15:37
did i do that? or coincedence?
?eval my @x = (1..5); my @y = splice(@x,5);
evalbot_7713 Error: No compatible subroutine found: "&splice"
eric256 ?eval my @x = (1..5); my @y = splice(@x); 15:38
evalbot_7713 [1, 2, 3, 4, 5]
svnbot6 r7713 | iblech++ | * Usual svn props. 15:39
r7713 | iblech++ | * docs/quickref:
r7713 | iblech++ | * README: Mentioned new "fears".
r7713 | iblech++ | * unicode: Fixed the Unicode variant of ??!!.
r7713 | iblech++ | * fears: Extremely minor cosmetical fix.
r7713 | iblech++ | * data:
r7713 | iblech++ | * (42) is neither a list nor an array, but the Num 42.
r7713 | iblech++ | * (42,) and list 42 are one-element arrays.
r7713 | iblech++ | * ($foo,$bar) does not create new containers, while [$foo,$bar] does
r7713 | iblech++ | (also see t/data_types/lists.t and t/data_types/array_ref.t and several
r7713 | iblech++ | p6l discussions linked in these tests).
r7714 | iblech++ | docs/quickref/data: More explanations WRT the constantness of (...,...) and 15:40
r7714 | iblech++ | [...,...]; also see t/operators/binding/arrays.t and related p6l discussions.
r7714 | iblech++ | (BTW, PIL2JS handles most (all?) of these things ((42) is not an array, the
r7714 | iblech++ | comma operator does not create new containers, but [...] does, etc.) correctly.
theorbtwo ??!! has a unicode variant? 15:41
eric256 dunno. i'm confused on his (42,) list 42 notes. ;) 15:42
Khisanth list 42 creates an array? boy that sure is a misleading name ...
theorbtwo Oh, it doesn't -- he added it to the abusive section. 15:44
Do we want to add ¢ to the top section, or is it too early?
PerlJam theorbtwo: release early, release often.
(IOW, add it!)
eric256 abusive section? 15:45
?eval (4,2).ref; 15:47
evalbot_7714 ::Array
eric256 ?eval list (4,2).ref;
evalbot_7714 ::Array
eric256 ?eval (list (4,2)).ref;
evalbot_7714 ::Array
eric256 agrees with Khisanth....thats weird. ;)
?eval (l4,).ref; 15:48
evalbot_7714 Error: No compatible subroutine found: "&l4"
eric256 ?eval (4,).ref;
evalbot_7714 ::Array
theorbtwo eric256: There's a section at the top, unicode operators, then unicode ops in Set.pm, then things that we could use as unicode operators if we weren't evil.
eric256 ahh
mrborisguy so is there also a ::List type?
eric256 i' 15:50
Limbic_Region adds his 6 cents to the "Perl 6 fears" thread
theorbtwo So, I wonderses... 15:51
svnbot6 r7715 | eric256++ | 'Fixed' shift and pop so that all tests pass. Sort of a kludge in prelude.pm that i think needs fixed in the MMD at some point. (@a) signature allows scalars to be passed in (even if they arn't refs).
theorbtwo Where does one find documentation on C-x 8 ... things for emacs? 15:54
C-x 8 gives me a no-such error on my xemacs... 15:55
Jooon theorbtwo: you could try C-h b for keybindings list and search for C-x 8 in that list 15:56
but it doesn't seem that helpful to me 15:57
eric256 ?eval shift 42; 15:58
evalbot_7715 Error: Cannot 'pop' scalar
eric256 aint pretty but it works. ;)
eric256 finds that many of the trans tests were actualy broken, not the implementation ;) 16:01
svnbot6 r7716 | iblech++ | docs/quickref/data: More explanations WRT list context and slurpy signatures. 16:03
r7717 | eric256++ | trans.t - fixed tests.
eric256 ?eval int Inf; 16:13
evalbot_7717 179769313486231590772930519078902473361797697894230657273430081157732675805500963132708477322407536021120113879871393357658789768814416622492847430639474124377767893424865485276302219601246094119453082952085005768838150682342462881473913110540827237163350510684586298239947245938479716304835356329624224137216
Khisanth oO
eric256 ?eval int Inf == Inf; 16:14
evalbot_7717 bool::true
xerox o_O
eric256 ?eval use Test; is (int Inf, Inf);
evalbot_7717 pugs: *** No compatible subroutine found: "&use" at Prelude.pm line 61, column 30-59
eric256 figures. lol
xerox ?eval Inf 16:15
evalbot_7717 Inf
eric256 is does an eq tests instead of ==
?eval int Inf eq Inf;
evalbot_7717 bool::false
eric256 of course that fails
integral ?eval log(int Inf) / log(2); 16:18
evalbot_7717 Inf
integral ?eval my $a = int Inf; $a
evalbot_7717 \179769313486231590772930519078902473361797697894230657273430081157732675805500963132708477322407536021120113879871393357658789768814416622492847430639474124377767893424865485276302219601246094119453082952085005768838150682342462881473913110540827237163350510684586298239947245938479716304835356329624224137216
integral ?eval my $a = int Inf; log($a) / log(2)
evalbot_7717 Inf
integral ?eval my $a = int Inf; log(+$a) / log(2) 16:19
evalbot_7717 Inf
integral ?eval log(179769313486231590772930519078902473361797697894230657273430081157732675805500963132708477322407536021120113879871393357658789768814416622492847430639474124377767893424865485276302219601246094119453082952085005768838150682342462881473913110540827237163350510684586298239947245938479716304835356329624224137216) / log(2)
evalbot_7717 Inf
integral *sigh*
eric256 what?
integral I was hoping log wouldn't overflow 16:20
eric256 ahh
?eval int Inf == 'Inf'
evalbot_7717 bool::false
eric256 ?eval int Inf eq 'Inf'
evalbot_7717 bool::false
eric256 thats just a dumb test anyway...i can't see anyway that would be usefull
?eval '' ~ Inf 16:21
evalbot_7717 "Inf"
theorbtwo ?eval my $a=int Inf; my $b=0; $b++ while ($a/=2); $b;
eric256 i mean that would make the string "Inf" magicaly....seems ludicrous
evalbot_7717 (no output) 16:22
theorbtwo ?eval my $a=int Inf; my $b=0; $b++ while ($a>>=1); $b;
evalbot_7717 Error: unexpected ">" expecting word character, "::", term postfix, operator or ")"
theorbtwo ?eval my $a=int Inf; my $b=0; $b++ while ($a >>= 1); $b;
evalbot_7717 Error: unexpected ">" expecting operator or ")"
theorbtwo Did shift-right change?
Khisanth maybe unimplemented :)
theorbtwo ?eval my $a=int Inf; my $b=0; $b++ while ($a +>= 1); $b; 16:24
evalbot_7717 Error: Can't use readline() or getc() in safemode.
theorbtwo (Yes, it did change... but WTF?)
?eval $x = 2; $x +>= 1;
evalbot_7717 Can't exec "./pugs": Permission denied at examples/network/evalbot//evalhelper.p5 line 46.
theorbtwo WTF?
xerox What is +>= ? 16:25
svnbot6 r7718 | eric256++ | Fixed infinitiy tests to use cmp_ok and == instead of eq. This seemed more logical than the current string comparisons.
theorbtwo Shift-right in-place.
?eval $x = 2; $x +>= 1; 16:27
evalbot_7718 Error: Undeclared variable: "$x"
theorbtwo ?eval my $x = 2; $x +>= 1;
evalbot_7718 Error: Can't use readline() or getc() in safemode.
theorbtwo Hm.
Borken.
xerox pugs> $x +>= 1; 16:28
*** cannot cast from VInt 1 to Handle (VHandle)
at <interactive> line 1, column 1-9
(r7636)
theorbtwo If it was fixed in r7636, why am I seeing it now? 16:29
eric256 ?eval 1**Inf 16:44
evalbot_7718 1.0
eric256 ?eval .9**Inf 16:46
evalbot_7718 0.0
eric256 ?eval 1**Inf
evalbot_7718 1.0
eric256 ?eval 1.1 **Inf 16:47
evalbot_7718 Inf
eric256 ?eval 1.1**Inf
evalbot_7718 Inf
eric256 if a test is just plain wrong? wadda do? lol
theorbtwo Patch the test so that it's right?
eric256 this test claims 1**Inf is NaN everywhere except OSX where it is 1... but that doesn't seem to be true 16:48
theorbtwo More to the point, it doesn't make sense.
eric256 exactly
so just go ahead and fix it?
theorbtwo Yep.
eric256 is finding that lots of tests are failing because the tests are just plain wrong. lol 16:49
?eval 1**Inf == 1;
evalbot_7718 bool::true
eric256 ?eval 1**Inf eq NaN; 16:50
evalbot_7718 bool::false
eric256 ?eval 1**Inf eq 'NaN';
evalbot_7718 bool::false
eric256 ?eval 1**Inf ~ " Test"
evalbot_7718 "1 Test"
eric256 hmm. the is test failes.
okay..on my home computer 1**Inf does equal NaN...but thats a bug 16:53
gotta be. right?
brother You saying prel6 woldn't have transfinite ordinals 16:56
Khisanth where is a mathematician when you need one? :)
PerlJam eric256: 1**Inf == 1 17:00
eric256 hmmm according to a math sie 1**Inf is indeterminate. seems dumb though
lol
PerlJam well the math site iw WRONG! :) 17:01
s/iw/is/
brother my $e = Inf; $e **= $e for (1 .. Inf); say $e
PerlJam brother: that should say Inf forever
xerox ?eval 0/0
evalbot_7718 Error: Illegal division by zero
brother It should only say one thing 17:02
eric256 mathworld.wolfram.com/Indeterminate.html
xerox ?eval Inf / Inf
evalbot_7718 NaN
Limbic_Region . o O ( what is a legal division by zero? )
xerox Limbic_Region: economics :P
eric256 ?eval 0 * Inf
evalbot_7718 NaN
eric256 ?eval 0 / 0
evalbot_7718 Error: Illegal division by zero
eric256 ?eval Inf - Inf
evalbot_7718 NaN
eric256 ?eval 0**0;
evalbot_7718 1/1
xerox ! 17:03
eric256 thast suppose to be NaN too. lol
xerox Yep.
eric256 ?eval Inf**0
evalbot_7718 1.0
eric256 and that too
xerox err.
eric256 ?eval 1**Inf;
evalbot_7718 1.0
eric256 wonders where these test should go. ;) 17:04
brother blame it on ghc 17:05
t/builtin/math/exp.t
or infinity.t
Juerd PerlJam: Given the time Perl 5 was made in, I do not think "exposing implementation" was a bad thing. It meant performance and learnability, and Perl 5 was already hiding more than most other languages. 17:06
svnbot6 r7719 | eric256++ | arth.t Uncommeted two tests that work, and a third that is still debatable. 1**Inf == ??
Juerd eric256: If you don't know where to put a test, dunnowhere.t ;)
eric256: "# I don't know where to put these. Please move them to the appropriate spots."
eric256 juerd works for me ;) 17:07
Juerd eric256: If this allows you to continue where otherwise you need to wait, don't worry about delegating some work.
eric256 just in the base \t directory?
Juerd eric256: Someone will eventually step up and put them in the right place. Just keep the tests pure enough (or in blocks) so they can be moved easily.
Sure.
Limbic_Region there is the t/builtins/math directory though 17:08
as brother pointed out
eric256 yea i like his place
Juerd Whose place? :)
Limbic_Region brother's 17:09
mrborisguy ?eval Inf > Inf
evalbot_7719 bool::false
Juerd Oh, I had overlooked that
xerox eric256: 1^Inf is 1 in Mathematica
Limbic_Region Juerd - I didn't modify fears file as it seemed like you were doing consolidation 17:10
Juerd ?eval 1**Inf
evalbot_7719 1.0
Juerd Limbic_Region: What's consolidation? :)
mrborisguy xerox: that makes sense. 1^Inf should be 1
Limbic_Region has a hard time knowing when Juerd is serious or not
Juerd Limbic_Region: I'm serious 17:11
I almost always am
svnbot6 r7720 | iblech++ | t/builtins/math/infinity.t: Made clearer whether numification (==) or
r7720 | iblech++ | stringification (eq) is tested, using (Perl 6)++'s
r7720 | iblech++ | &infix:<operator_references>. (XXX: should &Test::is be changed to test with
r7720 | iblech++ | eqv?)
mrborisguy ?eval 2*Inf > Inf
evalbot_7719 bool::false
mrborisguy ?eval Inf + 1 > Inf
evalbot_7719 bool::false
PerlJam Limbic_Region: re your complaint. I don't think that can ever be truly "fixed" as most of the time questions on design have no easy answer.
Limbic_Region consolidation in this context means taking input from various sources, finding commonalities and boiling it down to essentials
mrborisguy ?eval (Inf/Inf, Inf-Inf)
evalbot_7719 (NaN, NaN) 17:12
Limbic_Region PerlJam - wasn't asking for it to be fixed.
Juerd Limbic_Region: I'm doing something like that. If two fears are related but different, I keep them as two, though.
Limbic_Region: The plan is to have answers ready in a few months.
r0nny i got a problem with classes
Juerd Limbic_Region: Answers that we are all aware of, so the community can help people overcome their fears by being consistent with eachother. The single-voice principle.
r0nny i want to do class foo { Str $var; };
Limbic_Region but when I can't get a simple answer on will .chars return a count in 1 context and the actual chars in another - what motivation do I have to keep pursuing the answer so that I can write tests for it?
PerlJam Juerd: The fears and responses would make a good article for perl.com :) 17:13
Limbic_Region agrees
Juerd PerlJam: There's already been one like it, but not based on a wide collection like this one.
Limbic_Region though I still think we need to have someplace kept up to date
dev.perl.org/perl6 says to read the latest summary to get caught up on what is new
that IMO, is BS
Juerd Limbic_Region: Edit away, by the way. If something is missing, or there is a sillier/more dramatic way to put something without making the sentences long or complex, please do help out. 17:14
PerlJam r0nny: you're missing a "has" or a "my" or an "our" or a "state"
Limbic_Region Juerd - did you have any problems with the 6 I listed?
Juerd Limbic_Region: I know of no project where something new is created, where documentation is kept up to date during the early stages of development.
PerlJam r0nny: Without more context, I'm not sure which you wanted.
Juerd Limbic_Region: I think this is very annoying, BUT natural.
Limbic_Region: If I didn't change them, then I didn't. 17:15
Limbic_Region Juerd - after 5 years, I don't buy that early stages of development line
Juerd - I didn't change the file, I only replied to your email
I will add the 6 to the file now
r0nny PerlJam: just a simple class wit 2 members 17:16
Juerd Limbic_Region: Ah, then I have just added them
r0nny PerlJam: but it doesnt like the { after class <namr>
PerlJam r0nny: then you probably want to use "has"
svnbot6 r7721 | eric256++ | NaN.t added some test i didn't see anywhere else for Infinity and NaN results.
r7722 | juerd++ | more fear
Limbic_Region will resync
Juerd Limbic_Region: In that case: yes, but the problems I have with them are very minor.
Limbic_Region ok
Juerd Limbic_Region: It involves structure of the sentences. I want everything to be within 80 characters, and explanations to not be there. 17:17
Just the raw fear.
Limbic_Region getting back to the fear/complaint email for a sec
Juerd When the fear is explained or justified, it's not as silly looking anymore.
PerlJam r0nny: class foo { has Str $name; }
Juerd This may be unfair, but it's politics anyway :P
Limbic_Region I find the current process uninviting to contribute
Juerd Which current process?
There are many parallel processes 17:18
Limbic_Region The list
PerlJam Juerd: the discussions that lead no where on p6l
Limbic_Region where is there a publicly accessable to-do list
Juerd Limbic_Region: I haven't been clear, and that is entirely my fault
Limbic_Region where is there a publicly accessible state of the design documents
Juerd Limbic_Region: But so far, people do seem to contribute fears... :)
Limbic_Region where is there a publicly accessisble (and up to date) faq 17:19
PerlJam Limbic_Region: sounds like you've got an email to p6l to write ;)
Limbic_Region Juerd - I sent 2 separate emails for that reasons
the first just listed the fears with a short commentary
Juerd There are indeed too many things that do exist, but are kept privately
Limbic_Region the other was a complaint which was resurfaced in me by your thread
Juerd Which doesn't work with 2005's wiki style development
Limbic_Region PerlJam - I have already written and sent it 17:20
it was lost as a tangent in one of those long threads that never ended up resolving anything
Juerd For example, "that's in my local copy of S\d\d" should not have to pop up so often, or even once.
Local copy? Surely Larry knows how to publish it?
Limbic_Region I tried to frameout the missing pieces of the design docs so everyone would know what needs to be done but even that didn't go anywhere 17:21
the reference document I used I was told was outdated
and that the current work existed in the collective minds of @larry
PerlJam It's been my experience that at least pmichaud and lwall don't really "release early and often". They queue up stuff in their personal workspace and then release when there's something reasonably complete. 17:22
Limbic_Region well - then the most amount of customer feedback is going to happen on only that which has been released 17:23
it is hard to offer insight, ask questions, and implement tests for things we don't know about 17:24
the framework is a means for everyone to know what is left to be done
anyway - I have vented 17:25
eric256 how do you tell that something is marked with an "is *" ? 17:31
ie .. $fh is chomped; ?
Limbic_Region you mean after the fact? 17:32
eric256 yea
Limbic_Region something like my @traits = $foo.traits;
dunno
eric256 except thats not implemented yet? right? lol. can't do $x does chomped?
Limbic_Region good question though
eric256 bastard langauge
Limbic_Region well - I think asking on a case by case basis is just as beneficial as getting a list 17:33
*shrug* - ask the list and see if you get a clear straight forward answer
eric256 lol
i just wants something that works in pugs. ;) 17:34
eric256 is off to investigate in source
brother ?eval int(Inf) 17:38
evalbot_7722 179769313486231590772930519078902473361797697894230657273430081157732675805500963132708477322407536021120113879871393357658789768814416622492847430639474124377767893424865485276302219601246094119453082952085005768838150682342462881473913110540827237163350510684586298239947245938479716304835356329624224137216
brother heh
Limbic_Region ?eval Inf ** ( 1 / Inf ) 17:39
evalbot_7722 1.0
Limbic_Region ?eval Inf ** ( 1 / 2 ** 128 ) 17:40
evalbot_7722 Inf
eric256 ?eval my $x is dumb; print $x.attrs;
evalbot_7722 Error: No compatible subroutine found: "&attrs"
r0nny re
eric256 hmm... from my limited view it looks like traits arn't realy stored even...just parsed and passed along.../me goes looking for traits tests 17:44
Limbic_Region well - if you do end up asking the list, ask for something all encompassing for traits, attributes, do-hickies, and whatcha-ma-callits 17:47
eric256 it appears that traits aren't well defined in tests or in syn's ;) joy 17:48
eric256 can understand how people quickly loose energy and momentum. ;)
Limbic_Region it isn't all doom and gloom 17:49
eric256 no 17:50
just hard to keep momentum with so many things poorly defined ;)
i have a hard enough time that i don't need the extra question of...is this even how its suppose to work? ;)
PerlJam A good strategy is to go with your best guess as if it were fact.
eric256 maybe i could just work on linking the tests to the documentation. i know there was an effort for that..who was leading that effort? 17:51
Limbic_Region my characterization of the situation would be that there isn't much of the design so written in stone that you can implement without fear of having to re-write it
eric256 maybe web based access to test with cross links to documentation could help issues?
Limbic_Region a substantial amount is defined but is still ambigous as you peel back the layers
and there is an unknown amount of stuff that isn't defined at all
eric256 l~r exactly ;)
Limbic_Region when things don't move at the pace people want the mantra "so volunteer" is echo'd without fail 17:52
the problem is that there isn't a list of things (that I know of ) that someone can go to
eric256 true. and i have no fear of my stuff being rewritten, in fact i hope most of my haskell is..i like get it done now, then fix later. ;) 17:53
Limbic_Region they have to learn the design over lots of emails cause the information isn't in sync with the documents
eric256 the list of test..but its hard to tell if the tests are wrong, bad, or truly broken. lol
Limbic_Region and when someone does try to venture out into unknown territory to help frame stuff in they find that there isn't anyway to move forward
except by following the pace of @larry 17:54
robkinyon autrijus: ping 17:58
eric256 if anyone remembers who is cross referencing testing and documentation (Syn) let me know ;) 17:59
kolibrie eric256: I think that's in t/README 18:00
eric256 thanks. found it 18:02
r0nny is there a data type for references/pointrtd 18:04
Limbic_Region ?eval my @foo = 1..4; my $ref = \@foo; say ref $ref; 18:05
evalbot_7722 Array bool::true
Limbic_Region ?eval my @foo = 1..4; my $ref = \@foo; ref $ref; 18:06
evalbot_7722 ::Array
Limbic_Region hmmm
?eval my @foo = 1..4; my $ref = \@foo; $ref.isa;
evalbot_7722 Error: No compatible subroutine found: "&isa"
Limbic_Region r0nny - why do you ask?
I think I know what you are getting at but I am not positive 18:07
in p5, references are of the scalar data type
for most intents and purposes, that doesn't change
Juerd References have to be scalars 18:10
Scalars are Perl's one-item datatype
Limbic_Region Juerd - right but I am not sure that is the question being asked 18:12
it is hard to tell but it might be intepreted as "is there a way to tell the difference between a reference and a regular scalar value"
ref is 1 way to answer that question 18:13
but it is known to be b0rk in a lot of ways
so that's why I was asking the leading question WRT to p6
but as r0nny didn't respond, I guess it doesn't matter
Juerd Limbic_Region: I was only explaining, adding to what you said. 18:20
Limbic_Region oh - okie dokie
forgive me for being a bit anal today - I am in the fog of a head cold
Juerd IIRC, ref() in Perl 6 will be quite boolean.
Limbic_Region well - with real objects I hope it will be possible to tell the difference between a reference to a hash 'HASH' and one blessed into a class called HASH 18:21
;-)
Juerd And there is of course $foo.isa(Ref)
Limbic_Region thought he tried that 18:22
Juerd What are "real objects"?
Limbic_Region Juerd - the smiley at the end was a bit toungue in cheek
Juerd Still, what are real objects?
Perl 6 objects will still be decorated references...
Limbic_Region ?eval my @foo = 1..4; my $ref = \@foo; $ref.isa(Ref)
evalbot_7722 bool::false
Juerd We're just mostly silent about that, because blessed references appear to scare people off. 18:23
Limbic_Region ?eval my @foo = 1..4; my $ref = \@foo; $ref.isa(Array)
evalbot_7722 bool::true
Limbic_Region ?eval my @foo = 1..4; my $ref = \@foo; $ref.isa(Asdf)
evalbot_7722 Error: No compatible subroutine found: "&Asdf"
Limbic_Region hmmm
Juerd <cent>Asdf ;)
Limbic_Region why would $ref.isa(Ref) be false and ref.isa(Array) be true? 18:24
Juerd Ref isn't in the object class hierarchy
(IMO, all these hierachies should somehow be made to be one) 18:25
Aankhen`` is going out of town in a few hours; be back on the 5th.
Be seeing y'all.
Juerd Aankhen``: Havefun
Limbic_Region IMO, "path of least suprise" was violated there ;-)
svnbot6 r7723 | iblech++ | * Usual svn props. 18:28
r7723 | iblech++ | * docs/quickref/unicode: Put the ^K digraph sequences back and added some
r7723 | iblech++ | more (unfortunately, some chars are not ^Kable -- you have to resort to ^V u 1234).
r7724 | iblech++ | docs/quickref/: 18:47
r7724 | iblech++ | * fears: "perl6 will be far too slow", separate compilation
r7724 | iblech++ | * data: Added section about references -- when do they auto-dereference? When
r7724 | iblech++ | is $ref ~~ Ref true? Do arrayrefs flatten in list context?
r7724 | iblech++ | (Also see t/var/autoderef.t, t/var/autoref.t,
r7724 | iblech++ | t/var/refs_point_to_containers.t, t/var/assigning_refs.t, and related p6l
r7724 | iblech++ | discussions (see the tests for links)).
stevan Juerd: Perl 6 objects wont be just decorated Refs anymore actually 18:48
well, they can be
but we have the opaque instance type now
and IIRC we have a Ref type 18:49
which is a reference to ...
Limbic_Region: re: "$ref.isa(Ref) be false and ref.isa(Array) be true" 18:50
I think that is Pugs being weird,.. and not Perl 6
Limbic_Region stevan thanks
stevan Limbic_Region: also remember that my $array_ref = @array; is how we make refs,.. not \ 18:51
Limbic_Region and I didn't answer Juerd's question WRT "real object" because we haven't drawn a black and white line
all we have done is move the gray area around
stevan Limbic_Region: well what a "real object" is, is not always the same from language to language
Limbic_Region stevan - was done explicitly that way for the benefit of the p5 to p6 folks
stevan - I was self-referencing 18:52
errr - rather I was distinguishing between the shoe-horned in objects of p5 in comparison to the native data type in p6
Limbic_Region sneezes, gets confused, and shuts up 18:53
stevan hands Limbic_Region a tissue
autrijus: ping 18:54
Juerd stevan: Then my information is incorrect. 18:55
stevan Juerd: where are you getting it?
Juerd stevan: \ is still available if you want to explicitly reference.
stevan: memory. 18:56
stevan A12 talks about opaque instance types
Juerd That some things evaluate to references in scalar context does not mean the "old" way is incorrect.
In fact, I think that in many cases, it may be very good code documentation.
stevan Juerd: I like ti myself too,.. I just assumed it went away 18:57
I think Ref will be a boxed type as well,.. although that stuff is all fuzzy (I am attempting to start a discussion to clarify that on p6c, but no takers so far) 18:59
robkinyon Ref will have to be boxed, won't it? 19:01
stevan yup 19:02
but ref will be the unboxed version
robkinyon what on earth can be unboxed about a ref?
stevan unboxed stuff is really just for optimization (at least to the best of my understanding0
robkinyon so, it's stuck being a long* ? 19:03
stevan if the runtime is in C :P
robkinyon (or whatever sPASM type is a pointer)
stevan if the runtime is in Parrot :)
it could be a Javascript class called Perl6.Ref too 19:04
the unboxed type will just be the internal runtime type exposed to the p6 user-land
no frills
eric256 where can i check out an uptodate copy of the SAE? "preferrably the repository that the official copies are in" 19:05
or is there such a thing?
stevan svn.perl.org somewhere,.. but Perl6::Bible is keep reasonably up to date IIRC 19:06
eric256 last update to svn.perl.org was 4 months ago...doesn't seem to promissing 19:11
stevan maybe not then... 19:13
I think perl6::Bible is the best bet
it was recently updated too (i think)
eric256 kinda need an repository..i think. maybe not. but i was hoping there was one true source.../me now worries a bit. ;) 19:15
i'll stick with the svn.perl.org one for now. thanks 19:16
stevan eric256: there is a repository, I may be wrong on the location
eric256 i found other docs pointing to that one too
stevan no need to worry
stevan wonders if Juerd's "fears" are maybe not a healthy a thing for the process right now 19:17
eric256 lol 19:18
stevan not that I dont think that getting these feelings out in the open is cathardic
but sometimes negativity can spread like a virus, and even when that was not the original intention 19:19
eric256 feather.perl6.nl/~eric256/test/t/var/temp.html 19:20
aint pretty yet. but i think graphic access to tests + linking to synopsis could be nice. ;) maybe i'll work in some searching etc...who knows
stevan eric256: we had something like this in the early days 19:21
not sure what happened to it
PerlJam stevan: the negativity is already there.
stevan PerlJam: I know
Limbic_Region eric256 - I can find the exact location of the repository if you want
eric256 yea it was a script in util...just going to polish it up and keep a working copy up..maybe after a bit i'll get test.pugscode.org pointed to it ;)
stevan eric256: nice :)
eric256 L~R that would be nice.
Limbic_Region pmichaud mentioned it in my post about needing the rest of the unfinished stuff framed 19:22
eric256 it would be cool to link passing/failing to it...but i'm not sure how possible that is
PerlJam I think Juerd's fears document makes a good rallying point for all of the "neagtive energy" surrounding parrot/perl6 and by answering them maybe we can channel the negativity into something positive
Juerd Thanks. That was exactly my intention 19:23
Limbic_Region svn.perl.org/perl6/doc/trunk
stevan PerlJam: I hope that is how it works
Limbic_Region if you are interested in the thread - it was "State of the Design Documents"
Juerd I've written some responses already. My greatest fear is that those aren't well enough. Please proof read and change where your skills are better than mine!
mrborisguy eric256: are you changing something in the test headers then?
some sort of pod or comment?
Juerd stevan: They may not be healthy right now. It all depends on how they are used. 19:24
svnbot6 r7725 | juerd++ | Some responses and pointers
stevan Juerd: exactly :)
mrborisguy eric256: I can maybe help by starting to find and change some of those
Juerd stevan: But isn't this true for every single commit that is done?
stevan goes to read Juerd's responses
Juerd There is a lot of Perl 6 FUD that is entirely unnecessary
stevan FUD? 19:25
Juerd I think we need to get rid of most of it BEFORE we even get close to the release date.
stevan: fear, uncertainty and doubt.
stevan ah
Juerd stevan: It's a marketing technique.
stevan I think a release might be the best cure for that
Limbic_Region Juerd - a lot of the FUD wasn't FUD at one time
Juerd Limbic_Region: That certainly is true.
Limbic_Region: Still, at this point it is. 19:26
Limbic_Region people say stuff about p6 which at the time is true - it changes, but as there is no single place to get current - it spreads
stevan Juerd: I think the FUD is pretty deeply ingrained,..
I think the best FUD clear-er of late has been Pugs
Juerd Yes, it certainly is
stevan prior to Pugs most people just assumed it was and would always be vaporware
Juerd But that's mostly the "it will never be released" FUD
PerlJam Juerd: I have a fear that I'm not sure quite fits into your document: I won't be able to lookup all these fancy operators in any reference. (books will be poorly indexed and online resources aren't punctuation friendly) 19:27
stevan PerlJam: we will need to re-implement the web in p6 then ;)
Limbic_Region PerlJam - like a wizard type search engine?
Juerd PerlJam: That's not new for Perl 6 :)
Limbic_Region how do I interleave 2 arrays? and get the zip operator spit out as a possible answer?
Juerd PerlJam: The solution is to have a table of operators in the index.
Limbic_Region disagrees with Juerd's solution 19:28
Juerd PerlJam: We need to (NEED TO) create our own online resources
Limbic_Region how is having a list of possible operators going to help me find the one that does what I want
Juerd PerlJam: Google will not cut it and has been reluctant to add any punctuation to the list of characters that can be found
Limbic_Region If I knew what I wanted I wouldn't be asking
Juerd Limbic_Region: There are several forms of documentation 19:29
stevan searches for "that sqwigly thing with the dots" and gets 0 results back
Juerd Limbic_Region: The most important is not "I want this, how do I do it?", but "What the hell does this thing mean?"
Limbic_Region: A cookbook has an entirely different structure, of course.
PerlJam really really really likes the answer to "FEAR: My Perl knowledge will be useless!"
stevan knowing what a character looks like, and what it is called are 2 different things
Juerd PerlJam: Good to know :) 19:30
stevan: Who cares, really
stevan Juerd: where are the fears?
Juerd stevan: # has 1001 names, and that has only been a source of fun :)
stevan ah
Limbic_Region docs/quickrefs
stevan true
Juerd stevan: pugs/docs/quickref/fears 19:31
stevan Limbic_Region: thanks
Juerd Quickref may not be the right place, but for now, it's as good a place as any
It was a somewhat different document at first (never commited)
stevan Juerd++ # nice responses 19:34
stevan feels better now
Juerd Yay :)
Please help add them
And help more people feel better
stevan oh no,.. I can only scare people with metamodels :) 19:35
Juerd hehe
PerlJam stevan: /me looks at the metamodel ... but ... but ... where are my objects? ;) 19:38
stevan PerlJam: :) 19:40
Juerd /Who dares to object to the metamodel?/
PerlJam Juerd++
robkinyon You know what my biggest P6 fear is? 19:42
It's the lack of definitive leadership
There's no place that says "This is the decision, here's the references, and that's that. Bring it up only if you have a truly compelling reason not answered in those references." 19:43
Limbic_Region robkinyon - I tried to figure out how to say that myself as it is something I have heard repeatedly
robkinyon I can't bring it up, though
stevan elects robkinyon President of Perl 6
Limbic_Region but I see it as a root cause of a lot of other fears
stevan I will bring that up 19:44
Juerd That's brave.
stevan doesn't give shit,.. he is crazy man.. crazy
Limbic_Region wonders why robkinyon can't bring it up?
PerlJam robkinyon: The Synopses are reference-like. Isn't that good enough?
stevan Limbic_Region: robkinyon has a wife and kids to support ;P
eric256 perljam...not if they arn't up to date
robkinyon PerlJam: No, given that the SAE are out-of-date
stevan I think it is a reasonable thing to say
Limbic_Region fwiw - the fears I listed aren't necessarily one's that I have
Juerd Limbic_Region: That's good :) 19:45
Limbic_Region just one's that I have heard in various forums in my pursuit of p6 advocation
stevan its been what,.. 4+ years,.. a freeze on *something* is probably long overdue
Juerd Limbic_Region: I started with those I remembered, and then visited IRC to collect more
PerlJam robkinyon: the A&E aren't meant to be kept up-to-date, just the S.
robkinyon PerlJam: none are up-to-date
Here's the thing - there needs to be a freeze list 19:46
Juerd PerlJam: That's something that should be made very, very clear and currently isn't at all.
stevan accoring to $Larry, A12 is for entertainment purposes only
robkinyon stuff that is going to be in P6, end of story
Limbic_Region PerlJam - I bet you ask two different heads in @larry and you will get 2 different answers on that
stevan volunteers robkinyon for that job ;)
eric256 is there a general way to split a path into its directories? or just use split "/|\" ? which looks cool but seems evil
robkinyon wonders if stevan ever wants him to do any $work
Limbic_Region pmichaud updates the As quite a bit to keep them in sync with the Ss
Juerd eric256: Cough.
eric256: That looks like PHP 19:47
PerlJam Limbic_Region: I bet I won't. I've only talked much to 4 @larry (lwall, dconway, pmichaud, and chromatic) and all give the same answer
Limbic_Region even after the decision was made to expand the scope of S
Juerd eric256: split takes a regex.
robkinyon prefers split "\|/" - it looks cooler
Juerd eric256: Regexes are in //, not ""
eric256 it wasn't code..just the idea... working on this test to pod script and it has catdir....but where do i find the reverse of that?
Juerd Also \" is escaping the "
robkinyon So, is it /\|\// ? 19:48
gaal Juerd: you can give a string to split (in p5); it is interpreted as a regexp.
robkinyon or /\/|\\/ ?
eric256 juerd...chill. wasn't code.. but it got the point accross...you guys need to take a deep breath. lol
robkinyon or /\\|\// ? :-)
Juerd gaal: Unless it is " ", I know this.
gaal good.
eric256 i'm looking at the modules it uses but none seem to even have a catdir function ... 19:49
File::Spec::Functions does have themm...odd name for the module. seems like it should be getting info about the file. anyway
hmmm....now to use an array to travel down a hashref...never done that before 19:53
PerlJam "travel down"?
eric256 my $index = { }; my @paths = splitpath( $path ); now i want the path populated into the hash 19:54
PerlJam like [{}] \%hash, @paths ? 19:55
eric256 yea but in perl5 ;)
PerlJam in perl5 you write a loop
eric256 thank you perljam for that wisdom. lol.
the problem is that its building the hash as it goes..so the code i tried didn't work...lemme paste it 19:56
pasteling "eric256" at 66.102.136.66 pasted "blah" (5 lines, 113B) at sial.org/pbot/13896
eric256 i'm pretty sure it doesn't work because the hash doesn't exist to start with. gonna be a pain...;( 19:57
eric256 wonders if the easiest thing might be to build a string and eval it
robkinyon what about : my $hash; my $str = "\$hash->" . join("", map { "{$_}" } @paths) . " = \$file;"; eval $str; 19:58
eric256 yea thats what i'm doing now. seems to be the easiet way to do it.
robkinyon it's an ad-hoc implementation of like [{}], frankly
like() has to be implemented with eval in P5 19:59
(or a loop, but the loop version sucks rocks and is actually slower than the eval version, which is funny.)
oh, and buggy no matter what you do
eric256 yea the loop would have too many special cases to create the hashref it it doesn't exist yet etc..
eric256 is trying the eval version currently and crossing his fingers ;) 20:00
Juerd Lack of autovivification sucks.
I think it's the most important thing keeping me away from Ruby
wolverian Juerd, you can hack it by giving the constructor your own code block, but I don't know if you can do that for every hash, so that you don't have to recode the block for every hash 20:07
20:15 Lopo_ is now known as Lopo
r0nny yo 20:19
is there a way, to tell there is a array consists of references to a special kind of object ? 20:20
theorbtwo my @foos is Array of Foo; 20:21
r0nny hmm 20:22
and if this array consists of references ?
theorbtwo It consists of Foo objects. 20:23
stevan is Array of Ref of Foo?
r0nny i need it to consit of references to foo objects 20:24
stevan why refs?
r0nny stevan: i have a array of depencyes
stevan why do you need the added level or indirection?
r0nny each of them has depencyes within this array 20:25
so references are usefull for me
stevan why?
r0nny since its the only way to point to the depencyes
stevan is still not understanding though
theorbtwo s/yes/ies/. Neither am I. 20:26
stevan objects are not copied by value or anything like that
so the extra level of indirection is not really needed IMO
r0nny stevan: i have a array of a speical class - and each of the arrays elements needs to maintain a list of the elements of the array wich it depends on
stevan ok 20:27
r0nny tell me a better way if u know
stevan dont use refs
thats all
r0nny then how else should i do ?
eric256 thinks there is a general conflict of understanding 20:28
r0nny dont tell me "dont use the only way u know" without telling another way
eric256 in these cases r0nny it would probably help to show code
theorbtwo I think you're right, eric256.
r0nny atm im creating this code ;P
stevan my @foos = (Foo.new(), Foo.new()); @foo[1].add_dependency(@foo[0]);
theorbtwo You know an object is a reference, right? 20:29
stevan my @foos = (Foo.new(), Foo.new()); @foo[1].add_dependency(\@foo[0]); just adds a level of indirection which is not really needed
r0nny blessed reference
hmm
stevan r0nny: in p5 yes
in p6, it is not 20:30
r0nny if objects are referneces there must be some kind of blessing
either the reference tells, or the object data itself
hmm
knowdegde of internals would be usefull
stevan r0nny: in p6 you dont need to know this, just know it is an object type 20:31
like in Java, C#, etc.
theorbtwo ...and more importantly... 20:32
my $a = new Foo; $b = $a; $a.setBar("x"); $b.bar eq 'x';
r0nny hmm 20:33
so cloning is explicit
stevan r0nny: no, no cloning involved
theorbtwo If you want to clone, be explicit about it.
r0nny theorbtwo: thats what i said
and stevan missunderstood again
stevan r0nny: sorry I read it wrong too :)
theorbtwo You could stand to be a little more explicit too. 20:34
PerlJam stevan: you could've just said you were in violent agreement ;)
stevan just wants everyone to get along,.. no violence man,.. just peace brother
r0nny well - as long as my perl programms dont run planes / controll nukes everything will be fine ;P
PerlJam r0nny: strange ... that's *exactly* what I hope my perl programs will do. 20:35
stevan as long as perl does not run planes, or control nukes,... :)
r0nny PerlJam: maybe if perl6 is complete 20:36
hmm
i got strange errors
PerlJam When there's a problem with the air traffic control software and fixing it will decide whether a plane crashes or not, would you rather have to wade through oodles of C or just a few lines of perl? ;)
stevan PerlJam: if it is a life or death thing, I think fixing software is maybe not the best approach 20:37
PerlJam r0nny: I'm still hoping for a nice christmas present this year.
r0nny hehe
PerlJam stevan: well, one hopes that software that is life-or-death critical has been vetted properly.
stevan was amazed to learn you can debug LISP in outerspace
r0nny PerlJam: afaik they use ada for planes
and i dont think its easy to get them away from it 20:38
stevan PerlJam: unless p6 has a predictable GC, its not even an option anyway
r0nny: that is what Boeing uses, not sure about Airbus
r0nny hmm
PerlJam stevan: it will be as predictable as it needs to be. (you can trigger GC whenever you need to)
stevan PerlJam: can you disable it though? 20:39
that is what Java is not used for hard-realtime stuff
cause you dont want the GC kicking in at a critical moment
r0nny: Ada in airbus too (www.adaic.org/atwork/control.html) 20:40
PerlJam stevan: If someone raises the point in the appropriate forum, I'm sure that our intrepid implementors can see that it will be done.
r0nny i dont see perl as language for realtime systems
PerlJam What are "realtime systems"?
stevan PerlJam: I dont know about you, but I wouldnt pass the FAA review process :P
PerlJam (Everyone has a different definition of "real time")
r0nny PerlJam: everything is allways predictable
stevan PerlJam: and where timing is critical to the operation
PerlJam In that sense, I use perl in "real time" systems right now. 20:41
r0nny perl is only soft realtime
stevan is glad he doesn't live near PerlJam ;)
r0nny hehe
PerlJam timing is critical, but there's usually several minutes between events.
stevan r0nny: that is probably stretching it a bit
r0nny perl isnt suitable for hard-realtime stuff
stevan: compare it to "realtime" linux ;P 20:42
stevan doesnt know realtime linux :)
r0nny a brunch of hacks to make linux more suitable for realtime apps 20:43
stevan ah
r0nny but there is stuff where it cant keep the timings
stevan nice
:P
r0nny so it is no hard-real-time
stevan hard-realtime is still done in assembler and tightly controlled C AFAIK 20:44
and Ada
r0nny hmm
maybe there is a way to make perl6 suitable for kernel-programming
would be cool
theorbtwo Parrot's design has always been for highly user-tunable GC, including extreme tuning like "don't run". 20:45
r0nny it would be nice, to have the gc running in a extra thread
PerlJam terms like "soft" and "hard" real time are meaningless to me.
integral hmm, you don't necessarily need to disable the GC, just make it more bounded in runtime, surely?
although things like regions, or apr's pools seem more ideal 20:46
PerlJam Now, if you talked about response time in terms of minutes, seconds, microseconds, whatever
stevan PerlJam: "hard" realtime is: nuclear reactors, airplanes, etc
r0nny PerlJam: well - using "soft" stuff in envs where "hard" stuff is needes could create whorst-case situatuins
stevan "soft" realtime is: phone switching systems
stevan cant think of another example 20:47
theorbtwo hard realtime has specifications that say things like "this function takes 3 ms +/- 2ns".
r0nny routers
theorbtwo Soft realtime says "this function takes no more then 5 ms".
(Or at least this is how I understand it, having never done either...)
stevan PerlJam: usually picoseconds actually
eric256 feather.perl6.nl/~eric256/t_index/
stevan nanoseconds
stuff like that
eric256 tree of all tests. ;) now to make those link ;)
theorbtwo What's a couple orders of magnitude between friends? 20:48
stevan depending upon the application: life and death,.. or my burrito takes longer to microwave
r0nny hmm 20:49
btw - does normal regex syntax work allready ?
theorbtwo Yes, for at least two definitions of "normal".
r0nny or do i still need to use the prefix? 20:50
theorbtwo (Perl 5 regular expressions, and perl 6 rules, but not grammar... I think.)
r0nny two definitions of "normal"?
stevan PerlJam: see this (www.embedded.com/story/OEG20011016S0120)
r0nny no need for gramar
theorbtwo Ah. If you didn't mean perl5 regular expressions, they're called "rules" now. 20:51
r0nny i just need to scan a string, get some special strings, and replace them with data got from a hash
stevan PerlJam: to quote the article "A hard real-time system is one in which one or more activities must never miss a deadline or timing constraint, otherwise the system fails. Failure includes damage to the equipment, major loss in revenues, or even injury or death to users of the system." 20:52
r0nny baslcally i want to replacy every %foo-bar either with $hash{foo-bar} or $hash{foo}{bar}
PerlJam stevan: yeah, I read it.
stevan ok
:)
PerlJam "hard" and "soft" are still too imprecise for my tastes. 20:53
theorbtwo Oh. I'd use a s:perl5(g)/-/}{/. (IIRC.)
r0nny oh - and every %_foobar by the results of a method call
PerlJam r0nny: p6 rules are available if you have parrot in your path.
theorbtwo ?eval $a = '$hash{foo-bar}'; $a ~~ s:perl5(g)/-/}{/; $a; 20:54
evalbot_7725 Error: Undeclared variable: "$a"
theorbtwo ?eval my $a = '$hash{foo-bar}'; $a ~~ s:perl5(g)/-/}{/; $a;
evalbot_7725 Error: No compatible subroutine found: "&g"
theorbtwo ?eval my $a = '$hash{foo-bar}'; $a ~~ s:perl5/-/}{/; $a;
evalbot_7725 \"$hash\{foo}\{bar}"
theorbtwo Something like that, anyway.
PerlJam theorbtwo: you mesn s:perl5:g/-/}{/;
r0nny PerlJam: any texts on them ?
PerlJam: i want to use perl6 as much as possible 20:55
no perl5 if i could avoid
PerlJam r0nny: other than the Apocalypse and Synposis, you'll need to read the docs on parrot/compilers/pge
eric256 feather.perl6.nl/~eric256/t_index/t...first.html now links to the actual tests ;) 20:58
too few tests are linked to any documentation. ;) 21:01
r0nny is there a good way for recursive replacement ? (ie the substring just got replaces needs other replacements) 21:02
Juerd eric256: If you see a task for yourself, by all means get started ;)
eric256 i am ;) 21:03
svnbot6 r7726 | putter++ | Declare PIL-Run inactive.
r7726 | putter++ | I began an integrated PIL-Run and Perl6-Value/Container rewrite,
r7726 | putter++ | but -Ofun prunes it.
eric256 getting the visualization there..then i can read through those SAE and make sure there are plenty of links to test demonstarting ...but first many of the links to documentation are broken so that needs fixed too. 21:04
this is a task i can manage without waiting for consensu on differnt things...mind if i have the build process on feather build the test structure too? not yet but when it gets a bit prettier
Juerd eric256: No, of course I don't mind. 21:05
eric256 r0nny keep scanning the string for replacements (over and over) until there are none todo
Juerd eric256: Why do you feel you have to ask?
eric256 juerd..just being polite. ;) since you where here..wouldn't have seeked you out, but i do appreciate the enviroment to do this stuff in so i ask our of courtesy 21:06
s/our/out/
Juerd eric256: The motd should speak for itself :)
It's a productivity platform 21:07
eric256 yea. well then i wont ask anymore!! lol ;)
Juerd Asking me isn't very effective in getting things done :P
:)
eric256 i just worry with automated tasks
Juerd And you should
It should probably be ulimited somehow
eric256 so i ask...this way you know it exists too if something goes wrong while i'm away..but i wont start it now. 21:08
Juerd If something goes wrong, I'll find the responsible person anyway. Or not.
Mostly, if something goes wrong, the box will be unavailable for a while. That is a bigger problem than anyone's guilt.
eric256 and i thought maybe i'd have it build once a night or something. since tests don't change all that rapidly..depends if we can get other people interested in fixing doc links...dang i wish i could remember who said they were working on that...thought there was someone dedicated to it a while back
Juerd The motd also says something about this :) 21:09
You can ask who said that in the mailing lists
eric256 yea.../me puts nose back to grind stone for now. ;) 21:10
Juerd "grind stone"? What does that mean?
eric256 just a phrase for going back to work
think it originated from cattle pushing the grinding wheel or something like that
Juerd English is such a wonderful..ly annoying language. 21:11
eric256 hehe. your telling me
Juerd My telling you?
Nice illustration of my point, though.
eric256 makes note not to use any odd phrases aroudn Juerd for fear of needing to explain them. lol
Juerd :) 21:12
Dutch doesn't have that many set phrases 21:13
eric256 on reflection that phrase doesn't strictly make sense, but i've never had anyone who speaks english question it. lol. when i was learning spanish they had stuff like that to, or when i listen to spanish on TV i'm like...what?
Juerd I'm not used to thinking so many levels deep to figure out what something means.
eric256 that probably makes life easier.
Juerd We have euphemisms, of course, but that's a separate category. 21:14
eric256 my favorite is " i could care less"...which should realy be "i couldn't care less" but the 'not' normaly gets dropped making the phrase mean the exact opposite of how it is intended
r0nny if %info.exists($cmd) <- if doesnt like the i after the % - what did i wrong ? 21:18
eric256 ?eval my %x = (1=>2); say "yea" if exists %info<1>;
evalbot_7726 Error: Undeclared variable: "%info"
eric256 ?eval my %x = (1=>2); say "yea" if exists %x<1>; 21:19
evalbot_7726 Error: No compatible subroutine found: "&exists"
eric256 ?eval my %x = (1=>2); say "yea" if %x.exists<1>;
evalbot_7726 Error: No compatible subroutine found: "&exists"
eric256 ?eval my %x = (1=>2); %x.perl;
evalbot_7726 "\{(\"1\" => 2),}"
eric256 got me. ;) doesn't seem to work at all
r0nny ?
eric256 ?eval my %x = (1=>2); say "yea" if %x.exists("1"); 21:20
evalbot_7726 yea bool::true
eric256 nm. works fine. if you use it right. ;)
xerox What does '%' mean?
eric256 hash
r0nny it probably doesn't like somewhere else..whats the error?
perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
r0nny is there a way to do multi-dimensional access to hashed with a array of keys ? 21:23
Juerd Yes. 21:24
svnbot6 r7727 | putter++ | Update STATUS of 'other implementations of p6 rules' to 'inactive'.
eric256 ?eval my %x = (1 => (2=>(3=>"hello"))); %x<1><2><3>;
r0nny Juerd: how ?
evalbot_7726 \"hello"
xerox lisppaste3: url?
lisppaste3 To use the lisppaste bot, visit paste.lisp.org/new/perl6 and enter your paste.
Juerd r0nny: [.{}] %hash, @keys
eric256 ?eval my %x = (1 => (2=>(3=>"hello"))); [{}] %x (1, 2, 3); 21:25
evalbot_7726 Error: No compatible subroutine found: "&x"
eric256 ?eval my %x = (1 => (2=>(3=>"hello"))); [.{}] %x (1, 2, 3);
r0nny omg cool ;P
evalbot_7726 Error: cannot cast from VList [VStr "1",VList [VInt 2,VRef <Pair>]] to Pugs.AST.Internals.VCode (VCode)
eric256 ?eval my %x = (1 => (2=>(3=>"hello"))); [.{}] %x, (1, 2, 3);
evalbot_7726 \"hello"
eric256 sweet it even works ;)
dduncan fyi, if anyone cares, MySQL 5.0 is now final/stable/GA
r0nny Juerd: does it work with exists, too ?
Juerd r0nny: Yes and no. 21:26
r0nny Juerd: manual ?
or can i do a multi sub for it ? 21:27
Juerd r0nny: All but the -1st autovivify the element left of it, afaik.
Unless this changed. Did it? It'd be very weird.
r0nny how do i tell a multi sub it works ona hash, and gets a array as argument 21:28
Juerd r0nny: multi sub foo (%foo is rw, @array)
eric256 multi sub what (hash $self: Array @array) {}
or the %...actualy probably has to be % 21:29
Juerd eric256: An invocant in a sub?
eric256 if you do %hash.x isn't %hash the invocant?
or not?
Juerd Hm, yes.
eric256 thats what i thought he meant by on a ;) other wise its just a sub with two args 21:30
svnbot6 r7728 | putter++ | Grammars/MatchX.pm deleted - Obsoleted by more recent/correct Match classes in PIL2JS and PIL-Run. 21:36
r7728 | putter++ | Grammars/rx_grammar0.pl deleted - A never-used copy of rx_grammar.pl.
eric256 wonders if we can have a $x.html method. ;) for pretty printing in html in addition to .perl 21:40
r0nny eric256: just use a lib for syntax-highlighting as soon, as its portet 21:42
svnbot6 r7729 | putter++ | INSTALL: Removed "There is currently an undiagnosed bug in eval_yaml().",
r7729 | putter++ | as it's not clear this is true.
r7730 | eric256++ | catalog_test.pl - Updated some, created index.html file, fixed e to be exe. Going to work on having templates, color coded test files, stuff like that.
eric256 r0nny...embedded .html that tagged each element with a div and a class tag would make highlighting as simple as a css file. ;) assuming you could load and parse a file without executing it and then call .html on the result ;) 21:44
as its portet could also equal long from now and is going about the issue in a backwards kind of fashion IMH
pasteling "r0nny" at 84.184.148.187 pasted "recursion in multi sub deepexists broken :/" (56 lines, 1.2K) at sial.org/pbot/13899 21:45
stevan r0nny: it fails because nothing is being returned from $cmd.split() 21:47
no array at least
r0nny eric256: its not the task of perl to give html syntax highlighting
stevan: ?
stevan regarding your paste
eric256 r0nny...why not? wouldn't it be a handy feature? besides if perl has already parsed it, why parse it agian..i assume this could be an extension that walks the AST or something 21:48
stevan r0nny: oh, wait, it failing the recursion
hmm
r0nny stevan: as long as i dont use stuff like fkk-bar it works
stevan: i allready wrote it fails the recursion
stevan is not reading well today ;) 21:49
r0nny any idea how to repair
stevan why are you making it a multi?
r0nny i suck at this
hmm
right - no need of multi
eric256 and realy you could call it deepexists(%h, @a) just to simplify things 21:50
stevan however, if you do multi, then you could just do this:
eric256 why do you say its failing recursion? what is it doing? if you print @a inside deep exists does that shed any light
stevan multi sub deepexists (Hash %h) { return 0 }
eric256: when @a is undef the multi-sub will no longer match
eric256 chat window was sitting over that note in the cood...sorry 21:51
lol
stevan :)
r0nny will it work with a scalar if array has 1 elements
stevan probably not, although I am not sure 21:52
eric256 you want to return 1 when @a is undef not 0
because they all existed all the way to the last element
and i don't think you want @a to be rw....doesn't appear to be a need
r0nny i need to use exists if a has the size of 1 21:53
eric256 hmmm so? 21:54
r0nny hmm 21:57
eric256 lol
r0nny the multi-extension wont work :/
stevan: the sub with the array seems to match allways 21:59
damn - how to get this working :/ 22:00
all i need is a deep exists 22:01
pasteling "eric256" at 66.102.136.66 pasted "deep exists" (12 lines, 300B) at sial.org/pbot/13900 22:06
eric256 not pretty but it will work ;)
r0nny hmm 22:07
eric256 hmmm.....what?
pasteling "eric256" at 66.102.136.66 pasted "deep exists - better ;)" (12 lines, 287B) at sial.org/pbot/13901 22:08
r0nny sweet 22:09
Juerd sub deep_exists (%hash, *@_) { @_ or return 1; %hash.exists{shift} or return 0; return deep_exists(%hash, @_); }
Oh, good point,
+@_ or return 1
eric256 yours checks multiple keys of the same level, mine goes deeper with each check 22:10
Juerd Eh, I forgot something :)
eric256 lol
Juerd sub deep_exists (%hash, *@_) { @_ or return 1; %hash.exists{my $key = shift} or return 0; return deep_exists(%hash{$key}, @_); }
There
eric256 there you go...mines still prettier though . ;)
Juerd It's multi line. 22:11
nice use of slurpy scalar
eric256 sub dexists (%hash, *$first, *@rest is copy) { %hash.exists($first) or return 0; +@rest or return 1; dexists(%hash.{$first}, @rest);} 22:12
now neither are pretty. ;)
thanks btw ;)
i like these new function signatures...when they all work it will be even cooler. ;) 22:13
r0nny now it works :) 22:15
so 22:19
finally i can sleep :)
eric256 lol
i know that feeling
r0nny *afk*
crenz Hi all, I have a quesiton on using parrot subroutines in Pugs. 22:58
leo crenz: while I can't answer the question, parrot backend is currently still stalled 23:01
svnbot6 r7731 | eric256++ | Cleaning up test to documentation cross references. 23:03
crenz Basically, I wanted to be able to "read" a imc, pir or pbc file and have the symbols appear in the Pugs namespaces. Is this possible with any other compiler that targets parrot? 23:05
(In fact, is there somewhere a list of compilers that target parrot?) 23:06
Juerd Perhaps irc.perl.org #parrot is a better place to ask 23:09
leo hmm seems more like a bugs/haskell question to me 23:11
(the former question) 23:12
list of compilers: languages/* in the parrot tree
leo wonders why pugs got misspelled like this ;-) 23:13
crenz thanks for the answers so far! 23:22
So I can safely assume it's not possible to use parrot libraries in Pugs right now.
leo crenz: the pugs lambdacamels seem to be absent currently 23:42
crenz that's okay, I'll go to bed and ask again later :-) Danke! 23:49
leo crenz: bitte & gute nacht
23:52 _SamB_ is now known as SamB

Powered by Cro and the Raku® Programming Language.

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

Thank you!