»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.
colomon anyone out there? 00:26
geekosaur nobody but us camels 00:27
colomon no butterflies?
TimToady flutters by
[Coke] pong. 00:28
colomon not a bad crowd at all!
colomon is watching basketball and trying to debug Panda under Niecza 00:29
ggoebel colomon++ 00:30
LlamaRider one llama is still lingering around too :p 00:32
sorear *pop* 00:33
colomon: what do you think about the Rat.perl issues that came up earlier? 00:34
colomon +<2/3> versus <2/3> ? I wasn't paying too close attention
sorear versus (2/3)
by current spec, <2/3> makes a RatStr, not a Rat 00:35
TimToady also offered a solution where <2/3> (without spaces) makes a pure Rat, but I'm not fond of that
my preferred choice is (2/3)
colomon why (2/3) and not just 2/3? 00:36
sorear precedence
colomon ah, makes sense 00:37
(2/3) is easier on my eye
sorear I'm not sure if we want to go down the Haskell showsParen rabbit hole
colomon and (2/3) doesn't seem to require extra logic in the grammar. 00:38
<2/3> would go completely away, other than as the same as <1> or <1.4>?
TimToady I would still like to see that for a literal syntax 00:40
it's a nice pill
and it's kind of odd to use < ... > for a single string value in any case 00:43
colomon sorear: did you have thoughts on fixing up my patch from this morning? proper implementation of $*CWD and proper error handling on chdir? 00:46
sorear colomon: I didn't understand what you meant 00:54
colomon $*CWD is supposed to always be the current directory. I suspect it should query the current directory each time. (It's also supposed to be writable to do chdir, but I'm not looking to make that work today.) 00:55
If the C# system call in chdir throws, chdir's CATCH doesn't catch it.
you can try S32-io/chdir.t to see a demonstration of the latter. 00:56
sorear any particular reason not to use a C#-level CATCH?
colomon err.... not really, I suppose. 00:57
mostly that I thought the p6-level CATCH would get it
panda/fetcher.t now passes! \o/
sorear oh, your problem is that there's nothing in the CATCH to actually catch the exception 00:59
sorear if you just fall off the end of a CATCH block, it's assumed to be unhandled 00:59
CATCH blocks are supposed to contain when and default clauses 01:00
colomon oh!
ah, that's why the "goto fail" in the other CATCH in CORE.setting works? 01:01
so something like CATCH { default { $ok_code = False; } } should work? 01:02
colomon spectest seems very slow, or possibly looping? 01:22
colomon is glaring in moritz's direction... 01:23
colomon someone added a bunch of Range.roll tests which will be deathly slow unless an optimized version of Range.roll has been added. :\ 01:24
sorear perhaps that's the point :) 01:25
yes, default version will/should work 01:26
colomon I'm sure it's the point of the test, but simply adding it just wasted about ten minutes plus a near complete spectest run for me
dalek ast: 3480b41 | (Solomon Foster)++ | S32-list/ (2 files):
Fudge tests that work but are deathly slow until we can optimize the functions.
 01:28
colomon wishes tadzik were around... 01:34
dalek ecza: e32f383 | (Solomon Foster)++ | lib/ (2 files):
Quick implementation of IO.copy.
 01:40
ecza: a19bc37 | (Solomon Foster)++ | lib/CORE.setting:
Fix CATCH statement in chdir.
ecza: dd8e178 | (Solomon Foster)++ | t/spectest.data:
Turn on S32-io/chdir.t.
colomon woo-hoo! # basketball results, exciting finish 01:52
[Coke] what's the perl6 word for "coerced" ? 02:18
[Coke] also, I never got an answer to this: if you have an optional parameter, and someone passes in an undefined value, can you tell the difference between that and an invocation with no value provided? 02:24
colomon I don't think so 02:27
sorear yes, if your parameter has a defined default value
[Coke] any feedback on feather.perl6.nl/~coke/splice.diff ? 02:28
sorear: nope, just a straight $foo?
sorear [Coke]: I think you should write $size = * - 0 02:30
or $size = Inf maybe
[Coke] sure, that's better. 02:31
[Coke] sorear: slightly updated version. 02:37
sorear no comment 02:39
dalek d: 2300df1 | larry++ | STD.pm6:
check for common (?) mistakes from p5
 02:58
d: 6006414 | larry++ | STD.pm6:
make declarators parse the following initializer
d: 40d10f7 | larry++ | STD.pm6:
make initializer use item vs list assignment
TimToady sorear, jnthn: okay, STD now parses initializers as part of the declaration
sorear reviers 03:03
colomon wonders if that's a language he doesn't know, or a typo for reviews 03:06
sorear: presuming chmod requires the PosixWrapper. Given docs.go-mono.com/?link=T%3aMono.Uni...rmissions, how do I specify the types of the Func< > ? 03:08
dalek ecs: 744f7ca | larry++ | S02-bits.pod:
treat angle literals as slightly special

  <1/2> and <+1-2i> now produce Rat and Complex as (slightly) special
forms. Use whitespace or french angles to get the other behavior.		 03:08
sorear colomon: the latter 03:13
colomon latter? 03:14
TimToady typo
sorear typo for reviews
colomon: You can't. That would need to use a MethodInfo and wrapper function
colomon ah!
MethodInfo, hmmm, like in stat, eh? 03:15
let me see if I can reverse engineer it from that
TimToady sorear: does the initializer rule seem workable to you? jnthn has been pining for it for some time :) 03:16
presumably each declarator can decide which of the extra forms make sense or not; 'my' will likely distingish '=' from ':=', while 'constant' probably won't, seems like 03:18
hopefully this will take some of the pressure off of trying to guess what the operator means at reduce time 03:21
sorear TimToady: tentatively looks ok
TimToady: I'll give a shot at implementing it later
TimToady will be interested to see how it works out 03:22
[Coke] sorear, colomon: if I have a method in the setting, how can I check to see if one of the arguments is a whatevercode, and then how do I get its value? (.e.g getting -7 from *-7) 03:23
colomon [Coke]: look at substr 03:24
particularly $len 03:25
[Coke] danke. 03:28
colomon sorear: I see obj_getstr, obj_getnum, obj_getbool 03:29
sorear: is there an obj_getint I'm missing somewhere?
sorear colomon: no 03:37
[Coke] is there a way to auto-expand all the sections in lib/CORE.setting in vim? 03:38
sorear zi 03:39
sorear is looking at :help z
[Coke] finds "set nofen"
sorear looks like zR is the actual "expand all folds" command 03:41
[Coke] I just never want that, will add nofen to my .vimrc.
[Coke] argh. stacktrace in niecza that isn't showing any of the lines I'm editing. 03:43
colomon arrrgh, chmod's mode stuff isn't quite working 03:45
colomon bedtime 03:51
[Coke] I'm trying to fudge a test that has a sub that runs 2 tests. the sub is prefixed with a #?DOES 2 - but if I later put a #?niecza todo 'eek' before an invocation of that sub, only 1 test is skipped. 03:58
dalek ecs: 8d48c31 | coke++ | S32-setting-library/Containers.pod:
minor update to clarify splice
 04:17
dalek ast: 2718f25 | coke++ | S32-array/splice.t:
Simplify tests, follow spec.

  (e.g. error on negative args, take WhateverCode args..)		 04:19
dalek ecza: 280d918 | coke++ | / (2 files):
add @array.splice

run S32-array/splice.t		 04:22
[Coke] there's 37 more tests.
my implementation can surely be simplified. 04:23
dalek ast: 4e1c6fc | coke++ | / (2 files):
unfudge now that we have splice.
 04:28
[Coke] will fudge that for rakudo. 04:30
dalek ast: 78b269a | coke++ | S32-array/splice.t:
rakudo fudge
 04:34
[Coke] phenny, tell masak the splice test has been updated, rakudo needs to be updated to pass it again (it's been fudged0
phenny [Coke]: I'll pass that on when masak is around.
tadzik g'morning 05:49
sorear o/ tadzik 05:51
tadzik oh, nieczapanda working :) 05:55
well, starting to 05:56
moritz \o 06:03
phenny moritz: 17 Jan 21:57Z <jnthn> ask moritz if he could run a spectest of Rakudo on latest NQP master to see if there's any serious regressions from fixing the backtracking into subrules. (I assume you have ICU, which I don't.)
dalek kudo/nom: 42068aa | moritz++ | src/core/Range.pm:
remove debugging statement, moritz-- kboga++
tadzik hello moritz 06:05
moritz hello tadzik 06:06
jnthn morning' 06:45
moritz \o jnthn 06:46
jnthn oops, excessive apostrophe was excessive
jnthn isn't used to being up before the sun :)
moritz jnthn: running that spectest now 06:46
jnthn moritz: OK. Hope it won't be too bad.
moritz jnthn: two hangs, S05-mass/stdrules.rakudo and S05-modifier/sigspace.rakudo 06:52
(at least they don't leak)
jnthn hangs? :/
jnthn oh, seems that they have <.ws> in common 06:58
jnthn heading to training center, bbiab 07:03
jnthn her 07:38
er, her
*here
sorear what about her?
o/ jnthn
jnthn She's cute!
:P
Realized (I think) on the walk here what's going on...of course, <.ws> will match again and again every time you backtrack into it :) 07:39
Need to see if that's something systemic or specific to the way ws has been written in QRegex.
alas, gotta sort out for my class right now :) 07:40
also, where on earth is the coffee machine... 07:41
moritz is that the saem as quantified zero-width match?
sorear jnthn: what is the symptoms?
moritz a hang
jnthn sorear: I'm pretty sure that it just goes back into .ws, which doesn't do anything about the restart flag. 07:42
And so it matches again...and again...and again... :) 07:43
jnthn oh...I think !cursor_pass sets the backtrack flag by default now, whereas it was done explicitly before. 07:48
Will check later...
sorear So you aren't just using a lazy list of cursors? 07:49
jnthn sorear: Well, backtracking involves an invocation that creates a cursor, so I suspect there's some level of equivalence there 07:55
sorear: You could see this problem as a never-ending iterator if you prefer to look at it from that angle :)
masak mornin' 08:27
phenny masak: 04:34Z <[Coke]> tell masak the splice test has been updated, rakudo needs to be updated to pass it again (it's been fudged0
moritz jnthn: fwiw 'make qregex-test' in nqp also hangs 08:29
masak oh dear -- some people judge overall Perl 6 activity by measuring the activity on p6l :( perlmonks.org/?node_id=948499 08:30
sorear tell them that nobody cares about p6l anymore except fanboys and people who are developing their own vaporware languages 08:37
moritz that's basically what I'm writing now 08:38
... except that perlmonks doesn't accept my post, because of too high load 08:39
masak discussion at perlmonks.org/?node_id=948107 gets quite ugly at times. 08:40
masak I try to steer clear of such discussions nowadays. they lead nowhere and cause blood to boil. 08:41
masak the *only* outcome that matters is putting Perl 6 to good use in various ways. 08:44
moritz should follow that line of thought more closely too 08:45
moritz nom: sleep Duration.new(1/2); say 'alive' 08:45
p6eval nom 42068a: OUTPUT«alive␤»
cognominal_ perlmonks serves a purpose by being a magnet for pointless discussions so they do not pollute developper and pedagogical forums. 08:49
moritz colomon: sorry for the slow tests; on rakudo without the optimizations they were slow, but managable; I forgot to test them on nieza 08:51
mj41 rakudo: class myDate is Date { }; my $a=myDate.new("2011-01-01"); my myDate $b = $a; say $b.WHAT; $b += 1; 08:52
p6eval rakudo 42068a: OUTPUT«myDate()␤Type check failed in assignment to '$b'; expected 'myDate' but got 'Date'␤ in block <anon> at /tmp/lNj705hWlS:1␤ in <anon> at /tmp/lNj705hWlS:1␤»
moritz mj41: oh, that seems Less Than Awesome
but why? 08:53
nom: class A { method new() { self.bless(*) } }; class B is A { }; say B.new.WHAT 08:54
p6eval nom 42068a: OUTPUT«B()␤»
moritz ah, it's the + that produces a Date
I'm afraid there's no good solution to that
mj41 moritz: sorry :-) ... workaround is easy github.com/mj41/FinMag/blob/master...ag.p6#L102 but not nice 08:55
moritz mj41: rt.perl.org/rt3/Ticket/Display.html?id=108052 is what happens when operators return objects of the type of their arguments, and not always a fixed type 08:56
I think the workaround is more desirable than a magic that can go so wrong 08:57
mj41 back at real $work :-( ... will backlog 8 hours later 09:03
sorear -> sleep 09:05
jnthn moritz: ah, good to know...thanks 09:29
moritz I think I know why m/.../ and <prior> don't mix 09:31
m/.../ compiles to $/.match
which in turn calls Str.ll-match 09:32
jnthn aha
moritz where the call to Cursor::set_last_match should be
jnthn yeah, that's prbably it
right
may have to re-order things a bit though - Cusrsor currently comes after Str.
Note it's a private method on Cursor 09:33
You'll have to add a trusts Str in Cursor also
dalek kudo/nom: ed943e6 | moritz++ | src/core/ (2 files):
stringify stuff before passing it to the regex engine

Otherwise lots of tests break; though not sure if this is the right solution		 09:34
jnthn moritz: Another way woulda been to stringify orig in Match.Str
moritz jnthn: no, it died in the regex engine 09:36
jnthn moritz: oh.
moritz so that would have been too late
jnthn Then yeah, we'll have to do it beforehand.
Yeah, I musta misread the stacktrace 09:37
moritz ./perl6 -e 'try { die "foo" }; say $! ~~ /foo/'
now lives again
jnthn moritz++ 09:37
moritz but there are still some test failures related to matching against non-strings
jnthn moritz: btw, were the hangs the only spectest failures the backtracking into subrules patches caused? 09:39
moritz jnthn: no, see the link to the summary diff I pasted earlier 09:40
jnthn huh, I musta missed that
moritz moritz.faui2k3.org/tmp/spectest-summary.diff.txt
jnthn oh, think we were on opposite sides of a netsplit when y ou posted it! 09:42
moritz quite possible :/
jnthn eek, lots of fail 09:43
moritz indeed
jnthn hopes they have a small number of root causes in mind 09:44
Like that there also some passing TODO in there too :)
s/in mind//
colomon tadzik: ping? 12:18
tadzik colomon: pong? 12:26
colomon \o/ 12:27
donri pang!
jnthn tadzik: traceroute?
;)
colomon ecosystem.t, the last three tests -- where is that data being sent?
*set
tadzik colomon: it's from t/panda/fakeprojects 12:29
colomon where in the code is it being set?
tadzik #6, especially #8 12:30
it reads the projectsfile as a JSON list of projects 12:31
colomon hmm, so maybe I've got a JSON bug here
danke
tadzik bitte
colomon perl6: class Foo { has @.bar; }; my $a = Foo.new(bar => [1, 2, 3]); say $a.bar.WHAT; say $a.bar[0].WHAT; say $a.perl 12:44
p6eval pugs b927740: OUTPUT«Array␤Int␤\Foo.new(("bar" => [1, 2, 3]),)␤»
..niecza v13-300-g280d918: OUTPUT«Array()␤Array()␤Foo.new(...)␤»
..rakudo ed943e: OUTPUT«Array()␤Int()␤Foo.new(bar => Array.new(1, 2, 3))␤»
colomon perl6: class Foo { has @.bar; }; my $a = Foo.new(bar => [1, 2, 3]); say $a.bar.WHAT; say $a.bar[0].WHAT; say $a.bar.perl
p6eval rakudo ed943e: OUTPUT«Array()␤Int()␤Array.new(1, 2, 3)␤»
..pugs b927740: OUTPUT«Array␤Int␤[1, 2, 3]␤»
..niecza v13-300-g280d918: OUTPUT«Array()␤Array()␤[[1, 2, 3]].list␤»
colomon ... I think that's a niecza-bug 12:45
niecza: class Foo { has @.bar; }; my $a = Foo.new(bar => <1 2 3>); say $a.bar.WHAT; say $a.bar[0].WHAT; say $a.bar.perl 12:46
p6eval niecza v13-300-g280d918: OUTPUT«Array()␤Str()␤["1", "2", "3"].list␤»
masak colomon: looks like. 12:47
colomon niecza: class Foo { has @.bar; }; my $a = Foo.new(bar => <1 2 3>.map(+*)); say $a.bar.WHAT; say $a.bar[0].WHAT; say $a.bar.perl
p6eval niecza v13-300-g280d918: OUTPUT«Array()␤Int()␤[1, 2, 3].list␤»
colomon niecza: say <1 2 3>.map(+*).WHAT
p6eval niecza v13-300-g280d918: OUTPUT«List()␤»
dalek ecza: 188b9c9 | (Solomon Foster)++ | lib/ (2 files):
Implementation of chmod using PosixWrapper.
 13:02
dalek ecza: 0f93f81 | (Solomon Foster)++ | lib/CORE.setting:
Whoops! chmod was supposed to be a method on IO, not a free sub.
 13:19
colomon tadzik: I now have all the panda tests working except the dependency thing. Hmm... and I think I've got a cheesy workaround for that... 13:23
colomon bingo! 13:24
all panda tests now pass 13:25
tadzik nice! 13:26
colomon stubs.t doesn't pass yet, but soon... 13:27
ooo, retract that 13:28
I don't think the introspection it depends on is available in niecza yet
tadzik ah, Test::Mock
colomon yeah
of course, bootstrap still doesn't seem to work. :) 13:29
wishes it conclusively failed rather than just not printing anything other than
==> Installing panda from a local directory '.'
Any()
oh, that's because of the try / CATCH. hmmm 13:30
colomon Unhandled exception: Nominal type check failed in binding self in Pies.resolve; got Str, needed Pies 13:31
I can figure that after breakfast
[Coke] gist.github.com/1476841 - # 01/18/2012 - niecza at 100.86% (from a run at about midnight this morning my time that I left to run while I zz'd) 13:32
b: say 19053-19217
p6eval b 1b7dd1: OUTPUT«-164␤»
[Coke] b: say 186-57 13:33
p6eval b 1b7dd1: OUTPUT«129␤»
[Coke] sweet. so it's not entirely due to the extra failures in rakudo.
pat_js hey is there a way to take CallFrame objects as keys in a hash (if there was a "column method this would be easy, but there isn't)? 13:37
moritz pat_js: do you want to uniquely identify the caller's location? 13:38
pat_js yup
moritz pat_js: in rakudo, you can use nqp::callerid() for that (but it's non-standard) 13:39
pat_js thought about it, and that way you cloud implement fff without makros
moritz that's what infix:<ff> now uses
see src/core/operators.pm, lines 163 to 215 13:40
pat_js ah thanks 13:42
and nqp::p6box_s is sort of a hash? 13:43
*hashing algorithm
moritz no, nqp::p6box_s turns a parrot-level strinig into a Perl 6 Str object 13:43
moritz *string 13:43
%ffv is an ordinary hash
pat_js but it's not short cirquiting 13:50
moritz correct
masak hashes? short-circuiting? 13:51
moritz no, ff
masak ahah
nom: False ff (say "OH HAI") 13:52
p6eval nom ed943e: OUTPUT«OH HAI␤»
masak right. 13:52
so "there's a way to implement ff without macros" -- "however, it's not short-circuiting" -- "no, you need macros for that" :P 13:53
moritz well, you can make do with custom action methods too :-) 13:56
masak moritz++ # p6u 13:58
dalek ecza: 9856167 | (Paweł Murias)++ | / (3 files):
Turn p5 undef into Nil.
 14:15
moritz down to 94 mentions of trac.parrot.org in that branch 14:17
[Coke] ww ? 14:18
moritz yes, should have gone to #parrot 14:19
colomon hmmm, seems to be a nextwith issue 14:26
[Coke] Should splice be able to handle something like: my @a=<a b c>; splice(@a,1,1, (1..Inf)) ? 14:29
[Coke] perl6: my @a=<a b c>; splice(@a,1,1, (1..Inf)) ? 14:30
tadzik ach
p6eval rakudo ed943e: OUTPUT«===SORRY!===␤Confused at line 1, near "splice(@a,"␤»
..niecza v13-302-g0f93f81: OUTPUT«===SORRY!===␤␤Confused at /tmp/DLn2qTJRVV line 1:␤------> my @a=<a b c>; splice(@a,1,1, (1..Inf)) ⏏?␤␤Parse failed␤␤»
..pugs b927740: OUTPUT«*** ␤ Unexpected "?"␤ expecting operator␤ at /tmp/JWk9PLImkA line 1, column 41␤»
[Coke] perl6: my @a=<a b c>; splice(@a,1,1, (1..Inf))
tadzik ww
p6eval niecza v13-302-g0f93f81: OUTPUT«(timeout)»
..pugs b927740: OUTPUT«pugs: out of memory (requested 1048576 bytes)␤»
..rakudo ed943e: ( no output )
[Coke] well, at least we're consistent. ;) 14:31
masak Pugs -- so fast it runs out of memory before it times out! 14:31
jnthn nom: my @a=<a b c>; splice(@a,1,1, (1..Inf)); say @a[0,1,2,3] 14:50
p6eval nom ed943e: OUTPUT«a c Any() Any()␤»
jnthn o.O
[Coke] niecza: my @a=<a b c>; splice(@a,1,1, (1..Inf)); say @a[0,1,2,3] 14:58
p6eval niecza v13-302-g0f93f81: OUTPUT«(timeout)»
[Coke] nom: my @a=<a b c>; splice(@a,1,1, (1..Inf)); say @a[*-1]; 15:06
p6eval nom ed943e: OUTPUT«c␤»
[Coke] nom++
jnthn I fear nom gets the right answer for the wrong reason :) 15:07
[Coke] nom: my @a=<a b c>; splice(@a,1,1, (1..Inf)); say @a[10000]; 15:08
p6eval nom ed943e: OUTPUT«Any()␤»
dukeleto ~~ 15:29
masak o/ 15:34
I just realized I haven't seen lue in almost a month...
[Coke] ho, duke. 15:36
colomon dukeleto! 15:37
dukeleto colomon: hola! 15:45
cognominal_ should I file bugs using track or rakudo is moving to github for issues like parrot does? 16:01
moritz rakudo has never used trac 16:02
and it continues to use RT for now
and thus [email@hidden.address]
cognominal_ oops, I confused trac with rt.
moritz we just use the github issue trackers for specs and roast 16:03
(you'll also notice that the rakudo project on github has the issue tracker disabled to avoid confusion) 16:04
masak likes RT 16:05
once you get over the quirkiness of search, it's fine.
moritz it has other quirks too 16:06
but I can handle those as well
for example the separate URLs for public and privileged ticket access can be quite annoying 16:07
masak yeah; what's up with that? :)
[Coke] at least it now prompts if you're logged in that you might want to switch. 16:08
jnthn decommute & 16:09
[Coke] jnthn: more failures in rakudo since yesterday. 16:14
feather.perl6.nl/~coke/rakudo_summary.out 16:16
masak haven't been getting p6cc solutions for a few days now. how're things going for people? 16:19
[Coke] masak: ... when's the deadline? ;)
moritz soon!
masak [Coke]: 2012-01-29.
[Coke] ah, plenty of time. :| 16:20
moritz nom: say Date.new('2012-01-29') - Date.today 16:20
p6eval nom ed943e: OUTPUT«11␤»
masak not really. :)
moritz nom: say (Date.new('2012-01-29') - Date.today) / 5 # days left per problem
p6eval nom ed943e: OUTPUT«2.2␤»
[Coke] masak: what is the maximum number of solutions submitted for a single contestant so far?
masak [Coke]: that's classified.
[Coke] aw, I only want to do the minimum amount of work. ;)
masak not possible. 16:21
[Coke] b: say 11/4
masak the game is set up so you'll be worried enough to submit all five tasks :)
p6eval b 1b7dd1: OUTPUT«2.75␤»
masak nom: .new('2012-01-29') - .today given Date
p6eval nom ed943e: ( no output )
masak nom: say .new('2012-01-29') - .today given Date
p6eval nom ed943e: OUTPUT«11␤»
moritz masak: that's both cute and creepy somehow :-) 16:22
masak :P
moritz commute &
masak nom: sub next-such-day(Str $wd) { my $date = Date.today; my $wdn = <Monday Tuesday Wednesday Thursday Friday Saturday Sunday>.pairs.first(*.value eq $wd).key or die "No such weekday"; $wdn++; $date++ until $date.day-of-week == $wdn; $date }; say next-such-day("Sunday") 16:27
p6eval nom ed943e: OUTPUT«Date.new(2012, 1, 22)␤»
masak a bit long, but it works :)
there's some golf potential there, I guess. 16:28
[Coke] nifty.
masak nom: sub next-such-day(Str $wd) { my $date = Date.today; my $wdn = <Monday Tuesday Wednesday Thursday Friday Saturday Sunday>.pairs.first(*.value eq $wd).key or die "No such weekday"; $wdn++; $date += $wdn - $date.day-of-week; $date }; say next-such-day("Sunday")
p6eval nom ed943e: OUTPUT«Date.new(2012, 1, 22)␤»
masak oh course :)
of*
nom: sub next-such-day(Str $wd) { my $date = Date.today; my $wdn = <Mon Tues Wednes Thurs Fri Satur Sun>.pairs.first(*.value~"day" eq $wd).key or die "No such weekday"; $wdn++; $date += $wdn - $date.day-of-week; $date }; say next-such-day("Sunday") 16:29
p6eval nom ed943e: OUTPUT«Date.new(2012, 1, 22)␤»
skids perl6: class A {}; A.DEFINITE.say; 16:31
p6eval niecza v13-302-g0f93f81: OUTPUT«Unhandled exception: Unable to resolve method DEFINITE in class A␤ at /tmp/vm52QG_ycX line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3108 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3109 (module-CORE @ 64) ␤ at /hom…
..pugs b927740: OUTPUT«*** No such method in class A: "&DEFINITE"␤ at /tmp/hWrtB6tLaq line 1, column 13-27␤»
..rakudo ed943e: OUTPUT«Bool::False␤»
masak skids: .DEFINITE only implemented in Rakudo as yet. 16:34
skids If I'm getting it right, though, rakudo's output is correct, and classes themselves don't count as objects for the purpose of DEFINITE. Which leaves me withthe question, how do you test if a class has been declared or not? 16:39
benabik nom: say A.DEFINITE; 16:40
p6eval nom ed943e: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&A' called (line 1)␤»
benabik nom: say GLOBAL::<A>; class A {}; say GLOBAL::<A>
p6eval nom ed943e: OUTPUT«Method 'at_key' not found for invocant of class 'GLOBAL'␤ in method postcircumfix:<{ }> at src/gen/CORE.setting:1188␤ in block <anon> at /tmp/bQBwArU3sL:1␤ in <anon> at /tmp/bQBwArU3sL:1␤»
benabik Awh.
masak skids: what is it you're trying to do? 16:41
skids: if the class hasn't been declared yet, the program won't parse.
skids work around "class A {}; class A {...}; failing for now, so load order of different files containing interdependent classes does not matter. 16:42
masak just predeclare anything as 'class A { ... }' and be done with it.
skids Except if I have already loaded a fail that declares "class A {}" then the "class A{...}" predeclaration does not work. 16:43
*file
.o0(loading fails does happen though) 16:44
masak so, you predeclare stuff before you load stuff.
that's what I ususally end up doing, anyway.
skids Not good from a module usability perspective to ask users to do that. 16:45
std: class A { }; class A {...}; 16:49
p6eval std 40d10f7: OUTPUT«===SORRY!===␤Illegal redeclaration of symbol 'GLOBAL::<A>' (from line 1) at /tmp/1fesqH4qZ3 line 1:␤------> class A { }; class A ⏏{...};␤Check failed␤FAILED 00:01 107m␤»
masak skids: I don't really see how an end user of modules would end up in a bad situation -- even with the current state of the art -- if all modules do use statements "inwards" to a central module, which predeclares everything. 16:51
skids masak: that might work. No less ugly than "if <something> eval class A {...}" I suppose. 16:55
masak skids: that alone wouldn't work anyway. when you eval, you're already past compilation. 16:56
unless you eval in a BEGIN block, which is just a desperate thing to be doing.
decommute & 17:05
PerlJam still ... seems like it would be useful to have some sort of predeclaration mechanism that didn't care if it was already declared or not.
PerlJam a defensive "I want to make sure this name is available" 17:06
colomon #phasers in 12 minutes? 17:18
benabik Hm. NQP seems to be hanging in the qregex tests. 17:24
sorear good * #perl6
colomon o/ 17:25
benabik bisects.
sorear EOBACKLOG 17:27
jnthn benabik: known 17:40
benabik jnthn: Alrighty then. 17:43
jnthn benabik: I really shoulda done it in a branch. 17:56
benabik jnthn: That would have been nice for this poor bastard that runs "make test" on everything. ;-)
colomon sorear: if you pull the latest colomon-panda (github.com/colomon/panda) and try running bootstrap.sh, you'll find it fails with a weird nextwith issue. I'd appreciate it if you could take a look at it sometime. No rush, it will probably be a good 8+ hours before I get a chance to work on panda again. 18:29
thanks.
masak home 18:52
sorear masak!
sorear wonders if @masakyst is any relation 18:53
dalek ecza: 24441d8 | coke++ | t/spectest.data:
run fudged integration/99problems-21-to-30.t
 18:55
ast: d4bdbd3 | coke++ | integration/99problems-21-to-30.t:
niecza fudge
 18:56
masak sorear: I've never met @masakyst, and doubt there's any relation ;) 19:13
masak installs Strawberry Perl for the first time 19:22
arnsholt Does panda work with nom these days? 19:24
PerlJam arnsholt: it didn't the last time I tried (a few days ago) 19:25
arnsholt Right. I'll just assume it doesn't for the time being then. Not that many modules I need anyways 19:26
not_gerd hello, #perl6 19:40
how do I replay to a ticket (in particular rt.perl.org/rt3/Public/Bug/Display...l?id=64032 )?
^reply
masak not_gerd: need to be logged in to RT. 19:41
not_gerd masak: no way to do this via mail?
masak I think there is, if you have another mail from that ticket. 19:42
but I may be mistaken.
I always make my ticket comments via the web interface.
masak `10: greetings. 19:51
tadzik masak: wow, I'd bet you never open your web browser to submit a bug 19:53
zjmarlow hi all. building on windows ... is mingw32-make not supported? the nqp/Configure.pl only checks for nmake... 19:57
not_gerd anyway, see gist.github.com/1635180 if someone wants to take a shot at unbreaking Cygwin... 19:59
masak zjmarlow: it's quite possible that mingw32-make works but no-one has tried it yet. 20:00
not_gerd zjmarlow: if you use Strawberry Perl, it might work out-of-the box 20:01
MSYS is not supported, though
jnthn zjmarlow: The nmake check in there is a "make the dyncall build happy under the Microsoft toolchain on 64-bit" fix mostly. 20:02
jnthn PerlJam: panda didn't work with nom? tadzik++ and I did some work to fix that a while back... 20:10
PerlJam hold on, let me try again 20:14
PerlJam jnthn: gist.github.com/1635283 20:16
jnthn :( 20:18
tadzik PerlJam: update your Rakudo
jnthn: we've fixed that few days ago
and there's a test in roast too'
PerlJam ah.
jnthn tadzik: oh, phew...it's just that one.
PerlJam makes a new rakudo 20:19
shinobicl_ q 20:32
quit
jnthn fail 20:33
flussence quite 20:43
.oO( grr... netbook ran out of memory twice in a row )
 20:44
arnsholt I'm getting errors running the test suite for Zavolaj on OS X 21:07
Looks like the flags in $*VM<config><ld_share_flags> and $*VM<config><ld_load_flags> conflict
jnthn ugh ;(
arnsholt -bundle and -dynamiclib are mutually exclusive, apparently 21:08
jnthn hmm...maybe I should only be using one of those two
arnsholt Yeah, on OS X at least the only difference between them is that one has dynamiclib and the other has bundle
jnthn dynamiclib sounds like the right one :) 21:09
Try deleting tother
arnsholt Yeah, killing ld_share_flags gets me a step further
Now I just need to install libgmp =) 21:10
arnsholt Or not. I just need to get the linker to find it, it seems 21:11
jnthn libgmp?
jnthn is confused 21:12
The Zavolaj tests don't use that.
moritz arnsholt: on linux I needed to add LD_LIBRARY_PATH=. to the environment
otherwise 'make test' wouldn't find the newly compiled .so files
arnsholt Yeah, I'll probably need that too 21:13
Right now it doesn't even get that far
dalek p: 76377d0 | jonathan++ | src/Q (2 files):
Try to fix the hangs that implementing backtracking into subrules introduced.
 21:14
moritz t/qregex/01-qregex.t (Wstat: 0 Tests: 611 Failed: 10) Failed tests: 23-25, 42, 44-45, 216, 545, 559, 611 21:15
arnsholt Calling ld with -L/path/to/lib fixes it, but neither LD_LIBRARY_PATH nor DYLD_LIBRARY_PATH does
arnsholt ponders
jnthn moritz: I'm not sure how many of those I've regressed, and how many failed before I started. Do you know if they all passed before by backtracking patches? 21:16
moritz jnthn: I think so, yes
jnthn: but I can recheck 21:17
yes, all qregex tests pass in 97760d4540bb4cc06baea3b5a3b1194ed4f656ac 21:18
jnthn :( 21:19
shinobicl__ hi!: i used to use this alias for testing my p6 scripts: alias prove6='prove -e perl6 ' 21:26
moritz ufo && make test 21:27
shinobicl__ how i'm receiving an 'open3' error....: open3: exec of perl6 t/Schedule.t failed at /usr/share/perl/5.10/TAP/Parser/Iterator/Process.pm line .....
oh, i see... thanks! i'll read about this 'ufo' then 21:28
moritz shinobicl__: is perl6 in $PATH?
arnsholt shinobicl__: It's in masak's github 21:29
dalek p: baf0b9b | jonathan++ | src/QRegex/Cursor.nqp:
Fix for !BACKREF, which fixes test 42.
 21:37
jnthn Need to fix more, but a bit tired for it now :( 21:43
jnthn sleep & 21:54
masak 'night, #perl6 22:07
bacek_at_work ~~ 22:59
moritz, ping
perl6: $*IO.WHAT.say 23:00
p6eval rakudo ed943e: OUTPUT«Failure()␤»
..niecza v13-302-g0f93f81: OUTPUT«Unhandled exception: Unable to resolve method say in class Any␤ at /tmp/kjmN40LWrr line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3108 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3109 (module-CORE @ 64) ␤ at /home/p…
..pugs b927740: OUTPUT«Scalar␤»
bacek_at_work perl6: say +$*IO.lines 23:01
p6eval rakudo ed943e: OUTPUT«Method 'lines' not found for invocant of class 'Failure'␤ in block <anon> at /tmp/csO_RUU5fL:1␤ in <anon> at /tmp/csO_RUU5fL:1␤»
..niecza v13-302-g0f93f81: OUTPUT«Unhandled exception: Unable to resolve method lines in class Any␤ at /tmp/zN9VM2DZVQ line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3108 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3109 (module-CORE @ 64) ␤ at /home…
..pugs b927740: OUTPUT«*** No such method in class Scalar: "&lines"␤ at /tmp/CV7OJg3iJI line 1, column 5 - line 2, column 1␤»
sorear bacek_at_work: There is no $*IO
bacek_at_work: perhaps you meant $*IN?
bacek_at_work meh... 23:01
sorear, thanks
perl6: say +$*IN.lines
p6eval rakudo ed943e, niecza v13-302-g0f93f81: OUTPUT«23␤»
..pugs b927740: OUTPUT«*** No such method in class Scalar: "&lines"␤ at /tmp/AuUGHPn_Wk line 1, column 5 - line 2, column 1␤»
bacek_at_work perl6: "oof".reverse.say 23:02
p6eval rakudo ed943e: OUTPUT«oof␤»
..pugs b927740: OUTPUT«foo␤»
..niecza v13-302-g0f93f81: OUTPUT«Unhandled exception: Unable to resolve method reverse in class Str␤ at /tmp/RND2cuhT_T line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3108 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3109 (module-CORE @ 64) ␤ at /ho…
bacek_at_work perl6: "oof".flip.say
p6eval pugs b927740: OUTPUT«*** No such method in class Str: "&flip"␤ at /tmp/50E647pvne line 1, column 1 - line 2, column 1␤»
..rakudo ed943e, niecza v13-302-g0f93f81: OUTPUT«foo␤»
bacek_at_work sorear, how is "reverse" spelled for strings? 23:03
( apart from comb.reverse.join )
bacek_at_work need more coffee... 23:05
sorear bacek_at_work: flip 23:12
bacek_at_work sorear, thanks. 23:37
btw, looks like $*IO.get isn't properly specced. 23:38
At least not in S32-io
ah
my bad
found it

Powered by Cro and the Raku® Programming Language.

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

Thank you!