»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
chazarvosky Hi, does anybody knows why the new logo is a Butterfly? 00:23
timotimo on perl6.org you can click on the butterfly to get the explanation text if i remember correctly
chazarvosky thank you! 00:24
chansen_ m: say sprintf '%.4d', -12; 00:55
camelia rakudo-moar 2e63ee: OUTPUT«-12␤»
chansen_ ^^ shouldn't the above be -0012? 00:56
timotimo doesn't the number after the . usually refer to after-decimal-point digits? which doesn't make sense for a %d, right? 00:57
m: say sprintf '%04d', -12
camelia rakudo-moar 2e63ee: OUTPUT«-012␤»
timotimo m: say sprintf '%04d', 12
camelia rakudo-moar 2e63ee: OUTPUT«0012␤»
geekosaur timotimo, .n for %d was changed years ago to the above 00:58
because you can't specify the leading 0 in a *-spec (getting it from a variable)
(and I am continually astonished by how many people do not know of a change to printf specs that was made some 15 years ago) 00:59
timotimo oh? 01:16
huh
i don't know much about printf format beyond the simplest things
sjn m: say sprintf '%.4f', -12; 01:17
camelia rakudo-moar 2e63ee: OUTPUT«-12.0000␤»
sjn m: say sprintf '%09.2f', -12;
camelia rakudo-moar 2e63ee: OUTPUT«-00012.00␤»
sjn m: say sprintf '%09.2d', -12;
camelia rakudo-moar 2e63ee: OUTPUT« -12␤»
sjn m: say sprintf '%09d', -12;
camelia rakudo-moar 2e63ee: OUTPUT«-00000012␤»
chansen_ I'll file a bug report tomorrow, (time for bed) 01:23
sjn m: say sprintf '%.09d', -12; 01:25
camelia rakudo-moar 2e63ee: OUTPUT«-12␤»
sjn m: say sprintf '%.9d', -12;
camelia rakudo-moar 2e63ee: OUTPUT«-12␤»
sjn chansen_: '%.4d' really doesn't make much sense, so the bug is probably that a warning is missing 01:26
chansen_ sjn: .n specifies the minimum digits that should be outputted 01:27
The sign doesn't count as a digit 01:28
m: say sprintf '%04d', -12;
camelia rakudo-moar 2e63ee: OUTPUT«-012␤»
geekosaur another person whose printf knowledge is 20 years out of date...
chansen_ m: say sprintf '%.4d', -12;
camelia rakudo-moar 2e63ee: OUTPUT«-12␤»
geekosaur like i said, it's common that people don't know that was added years ago 01:29
everyone remembers what 1970s printf did
and 80s
adu hi all
m: 6.m(3)
camelia rakudo-moar 2e63ee: OUTPUT«Method 'm' not found for invocant of class 'Int'␤ in block <unit> at /tmp/UwbRN1bssY:1␤ in any <unit-outer> at /tmp/UwbRN1bssY:1␤␤»
adu m: 6.n(3)
camelia rakudo-moar 2e63ee: OUTPUT«Method 'n' not found for invocant of class 'Int'␤ in block <unit> at /tmp/phl2S_3XWT:1␤ in any <unit-outer> at /tmp/phl2S_3XWT:1␤␤»
chansen_ m: say sprintf '%.4d', -12; 01:32
camelia rakudo-moar 2e63ee: OUTPUT«-12␤»
chansen_ m: say sprintf '%+.4d', -12;
camelia rakudo-moar 2e63ee: OUTPUT«-12␤»
chansen_ m: say sprintf '%+.4d', 12;
camelia rakudo-moar 2e63ee: OUTPUT«+12␤»
chansen_ m: say sprintf '%+04d', 12; 01:33
camelia rakudo-moar 2e63ee: OUTPUT«+012␤»
chansen_ m: say sprintf '%+04d', -12;
camelia rakudo-moar 2e63ee: OUTPUT«-012␤»
chansen_ anyway, time for bed 01:35
geekosaur pyanfar:12490 Z$ printf '%+.4d\n' -12 01:37
-0012
jdv79 where is anything about testing on perl6.org? i don't see it. 02:04
jdv79 oh, roast has one line under lang design 02:05
jdv79 but where would other testing stuff go like smoking, "cpan" reports, etc... 02:07
ugexe i imagine those things belong on whatever authority 02:54
zostay m: my Int(Any) $x = '4'; 03:58
camelia rakudo-moar 2e63ee: OUTPUT«5===SORRY!5=== Error while compiling /tmp/1FE880fJh1␤Coercion Int(Any) is insufficiently type-like to qualify a variable␤at /tmp/1FE880fJh1:1␤------> 3my Int(Any) $x7⏏5 = '4';␤ expecting any of:␤ constraint␤»
jdv79 "testing" is a whatever item - idk about that? 04:28
ugexe smoke testing and cpan reports? i dont imagine those things belong in the core of a language 04:54
jdv79 well, maybe from some version of a user perspective its not a top level category but in a cultural sense i'd say its deserving of a higher spot than "misc"/whatever. 04:59
jdv79 maybe the "modules" section could be changed to "dists" which could then include compiler dists and then it would all fall under there 05:02
ugexe are you going to defined what a failed or passed test is? or the proper number of test reports to send to a test server? That is where im confused. What exactly would you be defining 05:06
dalek kudo-star-daily: 3786bdc | coke++ | log/MoarVM- (5 files):
today (automated commit)
 06:13
kudo-star-daily: f1e86ba | coke++ | log/MoarVM- (5 files):
today (automated commit)
[Coke] waves briefly. 06:23
lizmat: rjbs.manxome.org/rubric/entry/2082 06:25
lizmat good *, #perl6! 07:42
[Coke]: thanks for the link 07:43
dalek Heuristic branch merge: pushed 71 commits to rakudo/newio by lizmat 07:48
masak 'antenoon, #perl6 08:59
RabidGravy erp 09:00
lizmat masak RabidGravy o/ 09:01
eli-se hi 09:11
masak \o 09:12
dalek ake: 7b1aa76 | arnsholt++ | src/Snake/Actions.nqp:
Add a TODO note.
 09:13
snake: 157c5e9 | arnsholt++ | src/Snake/Actions.nqp:
snake: Temporarily create tuples as nqp::list.
arnsholt Whoops!
I'm sorry, dalek! 09:14
lizmat arnsholt++ 09:17
# fighting the dalek
jnthn o/, #perl6 09:19
lizmat jnthn o/ 09:21
Q: is there an op that works like "say" but prints on STDERR ?
:op<note>, # doesn't cut it :-( 09:22
masak \o, jnthn 09:24
jnthn lizmat: No, you'll have to compose it out of 2 09:31
arnsholt lizmat: nqp::sayfh(nqp::getstderr(), ...) probably 09:31
jnthn :op<sayfh> with an :op<getstderr> or so as its first arg
lizmat ok, let me play with that 09:32
I'm about to reinstate Devel::Trace functionality as a full lexical "use trace"
so you can switch it on / off
but I would like to see the output of "use trace" on STDERR rather than STDOUT
so you can redirect 09:33
hmmm "sayfh" doesn't show up in any src/Perl6 file
arnsholt I found it in nqp/docs/ops.markdown 09:34
lizmat ah, ok
:-)
lizmat ===SORRY!=== 09:36
Arg count 1 doesn't equal required operand count 3 for op 'say_fhs'
hmmm...
found it, I think 09:37
jnthn you probably put the string to emit in the wrong place
lizmat yeah, but now it does't print anything :-( 09:38
QAST::Op.new( 09:39
:op<sayfh>,
:op<getstderr>,
QAST::SVal.new(:value( $id ~ ": " ~ $/))
),
seems legit, no ? 09:40
jnthn oops
Second line needs to be QAST::Op.new( :op<sayfh> ),
lizmat ahhhh 09:41
eh, but this works: 09:43
QAST::Op.new(
:op<say>,
QAST::SVal.new(:value( $id ~ ": " ~ $/))
),
dalek kudo/nom: d630193 | lizmat++ | src/Perl6/ (2 files):
Revive Devel::Trace as "trace"

While cleaning up, I found some leftover Devel::Trace functionality that seemed useful to reinstate now that we have lexical pragma's.		 09:46
lizmat jnthn: work so far ^^ 09:47
$ 6 'use trace; my $a; $a++' 09:48
1: use trace
2: my $a
3: $a++
$ 6 'use trace; my $a; no trace; $a++'
1: use trace
2: my $a
pmurias snake is back :) 09:49
jnthn lizmat: Dunno if line number is better to emit than the statement ID thing, which is an internal compiler ID
lizmat well, if line number, we would also need file name, no? 09:50
jnthn Dunno :) 09:51
vendethiel o/; #perl6
lizmat in any case, the emit would be per statement, so statement ID seems in place
vendethiel pmurias: snake :o)?
lizmat++ 09:56
dalek kudo/nom: 2033bbb | lizmat++ | src/Perl6/Actions.nqp:
Make 'use trace' print on STDERR
 10:03
lizmat figgered it out :-)
jnthn :-) 10:04
jnthn finds another couple of percent off startup
vendethiel jnthn++
lizmat now suppose we want to write to a file, other than by redirecting STDERR
vendethiel jnthn: did you hear about Debugger::UI::CommandLine being broken? Do you have an idea off the top of your head of what might've broke it? 10:05
vendethiel would look at it, but didn't get the error message
vendethiel gets the laptop to triple check
lizmat jnthn: given a file handle, how would I change QAST::Op.new(:op<getstderr>) then ?
pmurias vendethiel: arnsholt++'s python on QAST 10:06
vendethiel pmurias: oh! yes, indeed :D 10:07
jnthn vendethiel: Not sure without looking...
jnthn It looks like the module loading changes may have busted it... 10:10
Or at least, I can't run it in its build directly, 'cus my -Ilib gets ignored... 10:11
lizmat ??
Ven gist.github.com/vendethiel/02b7fc9684c67b786ed8
yoleaux 22 May 2015 16:13Z <brrt> Ven: thanks
jnthn oh, but only at the debugger REPL. Odd. 10:12
lizmat jnthn: could it be that -Ilib adds a "file#lib" to @*INC[0] instead of "lib" ??
jnthn Ven: Hm, what kinda program do you debug to see that? :) 10:14
jnthn guesses he needs a bit more than -e "1" :)
Ven jnthn: a grammar
jnthn: github.com/undo-lang/frontend/blob...00-basic.t and github.com/undo-lang/frontend/blob...rammar.pm6 10:15
jnthn Ven: Reproduced it locally. 10:19
Ven jnthn: to be fair, it hasn't been that much of an issue since I also had your great Grammar::Debugger to help me out :-)
jnthn Ven: Pushed a fix, thanks. 10:27
Ven jnthn: amazing! I'm the thankful one :) 10:28
jnthn Also sprinkled a "unit" in there so we don't get deprecation whines.
Ven wow, yeah, I wasn't about to find how that happened. just for my information, how do you go about tracking this kind of stuff? do you skip some part of the backtrace? 10:29
jnthn --ll-exception was informative in this case
Ven really? it gave me the same stack trace :o 10:30
jnthn Oh? For me I got a way, way longer one. 10:30
Ven amazing, it works for me
jnthn \o/
Ven jnthn++!
Ven Oh, seems like python's PEP0484, "Type Hints", was accepted 10:31
( www.python.org/dev/peps/pep-0484/ )
Ven .tell froogs are you still looking at c++ name (un)mangling? this: github.com/kirbyfan64/unmangler might be interesting (even though it's in C) 10:37
yoleaux Ven: I'll pass your message to froogs.
Ven .tell froggs are you still looking at c++ name (un)mangling? this: github.com/kirbyfan64/unmangler might be interesting (even though it's in C)
yoleaux Ven: I'll pass your message to froggs.
lizmat Ven: telling FROGGS might be better :-) 10:39
Ven lizmat: oh, case matters?
lizmat .tell LIZMAT foo
yoleaux lizmat: Talking to yourself is the first sign of madness.
Ven .tell LIZMAT ++
yoleaux Ven: I'll pass your message to LIZMAT.
lizmat ah, I guess it *is* case insensitive
yoleaux 10:39Z <Ven> lizmat: ++
Ven good. sorry froogs, tho :P 10:40
lizmat jnthn: "finds another couple of percent off startup" where? in Moar or higher up ? 10:44
dalek kudo/nom: ab35a4e | lizmat++ | src/Perl6/World.nqp:
Generalize line number finding into a World method
dalek kudo/nom: 2a81b40 | jnthn++ | src/core/Compiler.pm:
Lazily construct a DateTime in Compiler.

This takes its BUILD from taking 6% of setting loading down to 2%.		 10:46
lizmat also seems we should speed up DateTime.new :-) 10:47
jnthn: I did some work in that area yesterday 10:48
jnthn ok :) 10:49
[ptc] vendethiel: did you by any chance have the GitHub username 'Nami-Doc' once? 11:00
Ven [ptc]: yes! 11:01
then FROGGS++ told me it was ugly, so I changed it :P.
[ptc] Ven: aha, that explains that issue then :-)
Ven issue?
[ptc] right, I'll update the META.info in the Sprockets repo :-)
Ven oh!
[ptc] well, I'm going through the entire ecosystem atm, and found a repo that I couldn't fork automatically
and it turned out to be yours due to the name change :-) 11:02
Ven alright. I havn't updated my name everywhere indeed... 11:04
[ptc] it's not a problem. I was just confused and thought I'd ask :-) 11:05
[ptc] Ven: sweet! Thanks for the quick merge! 11:09
Ven thanks for the PR
[ptc] no probs. Just sent you another ;-) 11:12
Ven it's merged already 11:15
dalek kudo/nom: 7261d4b | lizmat++ | src/Perl6/Actions.nqp:
Add file/line + cleanup to "use trace" output
 11:20
kudo/nom: 948ce64 | lizmat++ | docs/ChangeLog:
Mention 'use trace'
[ptc] Ven: cool! I like the cake emoji, that adds a nice touch to saying thanks to someone after merging a PR :-) 11:21
lizmat afk& 11:23
[ptc] m: sub finished( Code &callback(int) ) returns Int is native('libSDL_mixer') 11:37
camelia rakudo-moar 2a81b4: OUTPUT«5===SORRY!5===␤Shape declaration with () is reserved;␤ please use whitespace if you meant a subsignature for unpacking,␤ or use the :() form if you meant to add signature info to the function's type␤at /tmp/Vg2jUXWrzA:1␤------> 3sub finish…»
[ptc] m: sub finished( Code &callback (int) ) returns Int is native('libSDL_mixer') 11:38
camelia rakudo-moar 2a81b4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/63xfWWPIII␤Missing block␤at /tmp/63xfWWPIII:1␤------> 3) returns Int is native('libSDL_mixer')7⏏5<EOL>␤ expecting any of:␤ new name to be defined␤»
[ptc] m: sub finished( Code &callback (int) ) returns Int
camelia rakudo-moar 2a81b4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/HsAGSwFHVM␤Missing block␤at /tmp/HsAGSwFHVM:1␤------> 3shed( Code &callback (int) ) returns Int7⏏5<EOL>␤ expecting any of:␤ new name to be defined␤»
jnthn Code &callback is a bit dubious in so far as the & already means Callable, so that'd be a callable thing that returns a Code 11:43
[ptc] jnthn: yeah, I'm not 100% sure what the code is trying to do 11:45
jnthn: it's in the SDL module, which hasn't been touched in a few years and I was trying to get the test suite to pass
jnthn: ah, it seems that if I leave off the `Code` and add a space between `&callback` and `(int)` the test suite runs much better 11:48
jnthn: I was getting other errors that the NativeCall interface didn't know what to do with a `Code`, so I think your tip put me on the right track
jnthn: is putting a space between the sub name and its argument in such a situation correct? 11:49
jnthn No, I think you want the colon 11:53
&callback:(int)
At least, I think that works nowadays 11:54
[ptc] I'll give it a go 11:55
nope. I got: Type check failed in binding $sig; expected 'Signature' but got 'Any'
with the space, all of the tests run but one
jnthn OK, stick with the space then 11:56
[ptc] what's the difference between the two? What are they trying to say? 11:57
jnthn &foo:(...sig here...) is meant to mean "a code object whose signature matches that specified" 11:58
&foo (...) is meant to be an unpack
[ptc] ok, thanks! 11:59
jnthn However, 'cus NativeCall processes signatures differently, and 'cus at the time it was implemented we didn't parse the colon case, the unpack syntax was interpreted by NativeCall instead as a specification of the callback's signature
[ptc] which is probably why it works with the space, and not the colon
jnthn Right
I thought we now supported both in NativeCall, but apparently not
[ptc] maybe I just need to update my local stuff
jnthn I'd really rather we were using the one with the colon. It's more consistent with Perl 6 semantics generally.
[ptc] agreed 12:00
I'll do a full rebuild and see if that changes anything
[ptc] jnthn: just rebuilt Rakudo, rebootstrapped panda and am still getting the 'Type check failed ...' error. 12:24
jnthn: I'll stick with using the space and submit that as a PR 12:25
Ven [ptc]: yeah, I think it's important to say thanks, it all matters 12:36
arnsholt pmurias: (Re: Snake) Yeah, I got some new inspiration after OSDC.no. Currently working on what I'm calling The Great Objectmodel Refactor; basically getting class and object creation closer in line to how it's supposed to work 12:41
dalek k-simple: a0df606 | paultcochrane++ | lib/GTK/NativeLib.pm6:
Add unit declarator to module declarations

As of Rakudo 2015.05, the `unit` declarator is required before using
  `module`, `class` or `grammar` declarations (unless it uses a block). Code
still using the old blockless semicolon form will throw a warning. This commit stops the warning from appearing in the new Rakudo.		 13:12
k-simple: b191c89 | (Zoffix Znet)++ | lib/GTK/NativeLib.pm6:
Merge pull request #10 from paultcochrane/pr/add-unit-declarator

Add unit declarator to module declarations
isBEKaml [ptc]: (long commit messages)++ :-) 13:23
zostay m: class T { multi method f(Str $p, %o) { say "huzzah" }; multi method f(Pair $r) { my (Str $p, $o) = $r.kv; self.f($p, %($o)) } }; T.new.f("Foo" => { :bar(1), :baz(2) }) 13:30
camelia rakudo-moar 948ce6: OUTPUT«huzzah␤»
zostay m: class T { multi method f(Str $p, %o) { say "huzzah" }; multi method f(Pair $r) { my (Str $p, $o) = $r.kv; samewith($p, %($o)) } }; T.new.f("Foo" => { :bar(1), :baz(2) })
camelia rakudo-moar 948ce6: OUTPUT«Cannot call f(T: Str, Pair, Pair); none of these signatures match:␤ (T: Str $p, %o, *%_)␤ (T: Pair $r, *%_)␤ in method f at /tmp/4AdUXrIJO7:1␤ in block <unit> at /tmp/4AdUXrIJO7:1␤ in any <unit-outer> at /tmp/4AdUXrIJO7:1␤␤»
chansen_ Who do you declare that a routine returns two Int's? 13:59
First thought was, sub foo(Int $bar) returns (Int, Int)
moritz m: sub f { return Int, Int }; say f().^name 14:01
camelia rakudo-moar 948ce6: OUTPUT«Parcel␤»
moritz camelia: you declare a sub that returns a Parcel (or Iterable, if you feel a bit more defensive)
sorry, meant chansen_
chansen_ moritz: ah, thanks 14:02
timotimo tadzik: when will file-find and shell-command get their pullrequests merged? 14:08
dalek ast: 9252444 | usev6++ | S03-operators/precedence.t:
Use existing ticket number, narrow down test for '3 !=3'
 14:24
ast: 0f2f315 | usev6++ | S03-operators/precedence.t:
Really use existing ticket number (RT #121108)
 14:26
chansen_ DateTime generates ill-formed ISO 8601 string representations, the zone designator should not be in basic format when the calendar date and time of day is in extended format. 15:05
2015-05-23T15:22:12+0200
^^^^
moritz what does that mean? 15:06
how should it look like?
chansen_ 2015-05-23T15:22:12+02:00
colon between hours and seconds in the zone designator 15:07
s/seconds/minutes/
chansen_ ISO 8601 defines two formats, basic (minimum number of separators necessary) and extended (additional separators) 15:11
chansen_ I have outlined several of the ISO 8601 date representations and formats in the p5 module Time::Monent, <metacpan.org/pod/distribution/Time...TRINGS> 15:13
moritz chansen_: reading the Wikipedia entry on ISO 8601, I agree with you 15:14
chansen_ Good, I read ISO 8601:2004(E) 15:16
ISO 8601:2004(E) 4.3, the expression shall either be completely in basic format, in which case the minimum number of separators necessary for the required expression is used, or completely in extended format, in which case additional separators shall be used in accordance with 4.1 and 4.2. 15:17
Ven jnthn: new bug with rakudo-debugger. typing "s" fails with the same error 15:30
Actually, --ll-exception gives me a very good error: "Cannot look up attributes in a type object" 15:31
dalek href="https://perl6.org:">perl6.org: 0fcd603 | paultcochrane++ | source/community/index.html:
Mention Perl 6 Testers on the community page
href="https://perl6.org:">perl6.org: 2cd2300 | paultcochrane++ | source/community/index.html:
Whitespace fixup
Ven .tell jnthn typing "s" in the debugger gives me (with --ll-exception) this stack trace: gist.github.com/vendethiel/80b612785402162de022 . It also seems, when stepping (via enter) that the "grammar mode" (that shows which part matched already/didn't match yet) is not appearing anymore 15:34
yoleaux Ven: I'll pass your message to jnthn.
chansen_ moritz: should I file a report/pull request or will you take care of that? 15:37
moritz chansen_: I'm testing a patch locally right now
chansen_ ok 15:38
moritz it seems that updating the tests is much more work than updating rakudo :-) 15:42
chansen_ ok, don't feel bad, php did the same mistake ;o) 15:44
dalek ast: 1cd225a | moritz++ | S32-temporal/ (2 files):
Correct formatting of ISO 8601 time zones, chansen_++
 15:51
dalek kudo/nom: b200c70 | moritz++ | src/core/Temporal.pm:
Correct formatting of ISO 8601 time zones, chansen_++
 15:51
chansen_ moritz++
smls m: say [{x=>"foo", y=>1}, {x=>"foo", y=>2}, {x=>"bar", y=>3}].classify(*.<x>).map({; .key => [.value»<y>]}).perl 15:55
camelia rakudo-moar 948ce6: OUTPUT«(:foo([1, 2]), :bar([3]))␤»
smls ^^ is there a way to do the classify and map step in one go? 15:56
seems like a common tthing, to want to classify according to a sub-value of each element AND take said sub-value out of the classified elements themselves
maybe .classify could get a second parameter for mapping the value? 15:57
Ven m: say [{x=>"foo", y=>1}, {x=>"foo", y=>2}, {x=>"bar", y=>3}].classify(*.<x>).map({ .key => [.value»<y>]}).perl
camelia rakudo-moar 948ce6: OUTPUT«Method 'key' not found for invocant of class 'Any'␤ in block <unit> at /tmp/uVlxW0Pfw1:1␤ in any <unit-outer> at /tmp/uVlxW0Pfw1:1␤␤»
Ven heh, I guess.
smls So the above coul be written as [...].classify(*.<x>, *.<y>)
Failing that, could we possibly get an easier way to map a hash to another hash when you only want to modify the values (and not keys)? 15:58
.map({; .key => ...}) always feels a little icky to me 16:00
b2gills m: sub fib ( $a ){ state @fib := 0,1, &[+] ... *; @fib[$a] }; say fib 10000; say fib 10001; 16:32
camelia rakudo-moar b200c7: OUTPUT«(signal SEGV)»
b2gills m: sub fib ( $a ){ state @fib = 0,1, &[+] ... *; @fib[$a] }; say fib 10000; say fib 10001;
camelia rakudo-moar b200c7: OUTPUT«3364476487643178326662161200510754331030214846068006390656476997468008144216666236815559551363373402558206533268083615937373479048386526826304089246305643188735454436955982749160660209988418393386465273130008883026923567361313511757929743785441375213052050…»
timotimo m: say (0, 1, &[+] ... *)[^6]; 17:02
camelia rakudo-moar b200c7: OUTPUT«0 1 1 2 3 5␤»
timotimo b2gills: i'm investigating 17:04
something seems to be working with int rather than Int ... perhaps 17:05
timotimo i can't quite ... at <unknown>:1 (/home/timo/perl6/install/share/nqp/lib/Perl6/BOOTSTRAP.moarvm::4294967295) 17:13
is that -1, by chance? 17:14
timotimo well, iscont_* doesn't check for containers that are null pointers 17:18
but i have no idea why that'd be in there
especillly since we have VMNull
timotimo hum. indeed, captureposargs gives us a c-level null value. how does it get in there? 17:39
timotimo well, i can't really get a good hint as to why something puts a null in there and replacing a null with the VMNull gives "Cannot call postcircumfix:<[ ]>(Any, Int); none of these signatures match:" 18:05
that's kind of obviously wrong ... ish
er ... huh? @fib ends up being Mu 18:07
well, that explains a few things
it might actually be a null underneath and we just have a problem with state vars and binding?
Ven_ doesn't remember any bug of the sort during his (admittedly relatively short) time here 18:08
timotimo m: sub test { state $test := 1; say $test }; test 18:09
camelia rakudo-moar b200c7: OUTPUT«(signal SEGV)»
timotimo m: sub test { state $test = 1; say $test }; test
camelia rakudo-moar b200c7: OUTPUT«1␤»
Ven_ m: state $x := 1;
camelia ( no output )
Ven_ m: state $x := 1; say $x
camelia rakudo-moar b200c7: OUTPUT«(signal SEGV)»
Ven_ guess you can even drop the sub, uh. 18:10
timotimo for a golf, yeah
well, at least i know where things come from
after sniffing at the problem for an hour without much success
i should have tried golfing *first*
Ven_ .u ≅ 18:11
yoleaux U+2245 APPROXIMATELY EQUAL TO [Sm] (≅)
lizmat m: state $x := 1; $x.gist # just before the segv 18:25
camelia rakudo-moar b200c7: OUTPUT«Cannot call method 'gist' on a null object␤ in block <unit> at /tmp/cNfCb3dUjk:1␤ in any <unit-outer> at /tmp/cNfCb3dUjk:1␤␤»
smls m: say [2, 2.0].classify(*.Int).perl 18:46
camelia rakudo-moar b200c7: OUTPUT«Cannot call infix:<===>(Any, Mu); none of these signatures match:␤ (Any $?)␤ (Any $a, Any $b)␤ (Int:D \a, Int:D \b)␤ (int $a, int $b)␤ (Num:D \a, Num:D \b)␤ (Num $ where { ... }, Num $ where { ... })␤ (num $a, num $b --> …»
smls ^^ .perl bug?
dalek kudo/nom: 73d6f72 | lizmat++ | src/Perl6/Actions.nqp:
Make sure we don't trace in precomped files
 18:47
[Tux] $ perl -pi -e's/\bis_deeply\b/is-deeply/g' t/* 18:48
:(
lizmat what is the problem? That you must do that, or that it must be done? 18:49
smls m: say Hash[Any,Any].new.perl; # golfed
camelia rakudo-moar b200c7: OUTPUT«Cannot call infix:<===>(Any, Mu); none of these signatures match:␤ (Any $?)␤ (Any $a, Any $b)␤ (Int:D \a, Int:D \b)␤ (int $a, int $b)␤ (Num:D \a, Num:D \b)␤ (Num $ where { ... }, Num $ where { ... })␤ (num $a, num $b --> …»
[Tux] I don't like it. it's ugly. personal pref
dalek ast: 33d835c | lizmat++ | integration/advent2010-day16.t:
Adjust to time format correction
 18:50
[Tux] That I must do it? No. it is already done committed and pushed
lizmat :-)
[Tux] what was it exactly you suggested me to do with the mail to you? 18:51
translate to English and post where?
lizmat on perl6-users 18:52
lizmat is tired after cycling 80+ km and is taking the night off to see song and dance 18:53
japhb
.oO( Str.translate(:lang<English>)
yoleaux 22 May 2015 21:51Z <tony-o_> japhb: i have this dimming dalek and camelia pretty well - /hilight -regexp -line -color %K ^rakudo-.*?\s[\d\w]{6}:
22 May 2015 21:54Z <tony-o_> japhb: i have this dimming dalek and camelia pretty well - /hilight -regexp -line -color %K ^(m|r|p):
FROGGS_ o/ 19:11
FROGGS were there any complains about the release? 19:14
yoleaux 10:37Z <Ven> FROGGS: are you still looking at c++ name (un)mangling? this: github.com/kirbyfan64/unmangler might be interesting (even though it's in C)
FROGGS Ven__: yes, hoelzro++ and me are still very interested
Ven__: we only need to mangle though 19:15
timotimo: you are doing the p6w on monday? 19:17
FROGGS would like to be entertained by light reading
timotimo yeah, i'd like to
FROGGS k :o)
hmmm 19:18
FROGGS with the $~MAIN role mixin I did in NativeCall we could even fake 'augment slang MAIN' 19:19
ahh, perhaps not easily if that involves faking an EXPORT sub 19:20
dalek kudo/nom: 800f703 | moritz++ | docs/ChangeLog:
Changelog: DateTime timezone formatting change
 19:21
cxreg isn't name mangling an implementation detail of the compiler? or is it standardized? 19:22
FROGGS cxreg: it is compiler dependent, and not standardized, yes 19:23
FROGGS cxreg: but there are not too many compilers per platform and we can guess the correct mangling scheme and continue 19:23
moritz S17-lowlevel/lock.t occasionally aborts early on MoarVM 19:24
FROGGS my stress- and spectest runs were clean for the release
and I did a lot 19:25
$ time perl6-m -e1
real0m0.098s
moritz only happens on concurrent runs
for i in $(seq 1 5); do make t/spec/S17-lowlevel/lock.t& done
FROGGS first time below .1 ^^
cxreg lots of thoughts about it here stackoverflow.com/questions/1012798...andardized
moritz FROGGS: \o/
FROGGS cxreg: I believe I read that already, but I'll put it on my list before continuing, thanks :o) 19:26
Ven FROGGS: ah, okay 19:28
well, I hope you can find something in that..
FROGGS Ven: wikipedia is informative to get the basics, and the fine tunings happens with code samples and real libraries 19:30
cxreg it looks like clang (libc++) does some weird stuff around short string optimization that you'll need to handle 19:33
zostay m: "".split("/").perl.say 19:35
camelia rakudo-moar 73d6f7: OUTPUT«("",)␤»
eli-se hi 19:42
masak \o 19:45
Ven hi eli-se 19:47
dalek ast: 55dbaf4 | usev6++ | S12-methods/parallel-dispatch.t:
Use unique ticket number RT #124513
 19:52
itz eurovision-- # good reason to be American 20:32
moritz it's not like it's a compulsory event for all Europeans :-) 20:35
Juerd moritz: I wish you would have told me that years ago! 20:36
Ven I admit to never have watched a single eurovision
and I very much intend to keep it that way...
moritz kinda likes it, though not every song 20:37
andreoss doc.perl6.org/language/5to6 20:42
what does syntax highlighting there? 20:43
[ptc] andreoss: pygments
andreoss: to be more explicit: pygments.org/
gnight #perl6 20:44
andreoss wouldn't Text::VimColor be more proper?
*colour
moritz andreoss: feel free to port it Perl 6, then I'll use it :-) 20:45
itz I'm off work shortly :D and was planning on trying to highlight examples.perl6.org to p6-Text-Vimcolour 20:46
moritz: there is a port in progress
moritz \o/
andreoss itz: i almost finished it already 20:47
itz andreoss++ 20:48
jdv79 looks like HTML::Parser::XML is forever looping on tests 20:56
dalek ast: e1ee43f | usev6++ | S05-metachars/tilde.t:
Use existing ticket number RT #62086
 21:05
ugexe hmm, if i try to submit a test report with Panda::Reporter it errors 'This type cannot unbox to a native string' inside its 'to-json' method. Specifically anything with $*KERNEL causes it (removing those keys allows it to work as expected). If I run method 'to-json' directory, it does not error... panda is able to submit test reports just fine with it... and if use $*KERNEL.* (for each method of $*KERNEL used 21:37
in to-json) before Panda::Reporter.submit it will work. Its as if $*KERNEL attributes and/or methods are being lazier than they should be? 21:38
To be more clearn, if I delete these lines I can 'use Panda::Reporter' and Panda::Reporter.new a report from a script just fine: github.com/tadzik/panda/blob/maste...pm#L68-L74 21:42
andreoss should rakudobrew break the current setup when rebuilding? 21:44
ugexe andreoss: rakudobrew has been linking to the wrong directory for me since the last rakudobrew update 21:45
ugexe i've been manually coping .rakudobrew/moar-nom/install/bin/perl6 to .rakudobrew/bin 21:45
suppose i better send a pr for that 21:49
masak 'night, #perl6 22:54
Ven \o 22:56
andreoss prove gets confused when finds ok inside quoted string 23:08
andreoss paste.debian.net/182753/ 23:29
how do i investigate this segfault? 23:30
timotimo andreoss: please configure a moarvm with --debug=3 23:38
andreoss with Configure.pl? 23:41

Powered by Cro and the Raku® Programming Language.

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

Thank you!