»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.
00:00 akakcolin left
japhb profan: perl5's startup is a really difficult bar (it's only 2-3 ms). But we already beat 'perl5 -Moose' 00:00
*perl5's *base* startup 00:01
00:01 anocelot left, diakopter left
profan ahh, well, i'm sure there are strides to be made either way :), even if you dont get to where perl5 was, i'm sure you'll get to where it's most certainly fast enough 00:02
japhb But personally I'd be happy if we got down to 20-30 ms for perl6 startup. That's fast enough to be able to start *and get actual work done* before the user can react or the window can scroll.
profan at that point most people won't register the difference most likely, sounds good
japhb Right. 00:03
Certainly anything less than 17 ms, and you're faster than a single screen refresh, which is freaky, because then you can do things in such a way that the user feels like you time-warped on them. But that's just showing off. :-) 00:04
profan japhb: gaslighting, in perl 6? apparently more likely than you think :) 00:05
japhb Well, we try to only abuse the implementors, not the users. ;-) 00:06
00:07 TEttinger left 00:12 lolisa left 00:13 skids joined 00:20 BenGoldberg joined 00:24 akakcolin joined 00:25 akakcolin left, akakcolin joined 00:26 akakcoli_ joined 00:30 akakcolin left 00:33 akakcoli_ left 00:34 Wolf joined 00:35 Wolf is now known as Guest29580 00:36 akakcolin joined 00:39 akakcoli_ joined 00:41 akakcolin left, Guest29580 left 00:45 laouji joined 00:48 Akagi201 joined 01:24 uncleyear left 01:25 uncleyear joined 01:32 aborazmeh joined, aborazmeh left, aborazmeh joined 01:46 dayangkun joined 02:11 akakcoli_ left 02:12 asdf12z_ joined, akakcolin joined 02:13 akakcolin left, akakcolin joined, akakcolin left 02:14 akakcolin joined 02:19 yqt left 02:20 noganex joined, telex left 02:22 telex joined, noganex_ left
hoelzro anyone around who can help me upload the release tarballs? 02:25
02:26 jepeway left 02:28 jepeway joined 02:34 meLon joined 02:38 rmgk_ joined, rmgk is now known as Guest48893, Guest48893 left, rmgk_ is now known as rmgk 02:46 cognominal left 02:51 aborazmeh left 03:04 AlexDaniel left
dalek p: 2488091 | hoelzro++ | tools/build/MOAR_REVISION:
Bump required MoarVM version
03:13
p: 61b5259 | hoelzro++ | VERSION:
Bump our version
03:15 nys left
hoelzro whenever someone gets a chance, please upload the NQP tarball and signature; they're at hoelz.ro/files/nqp-2015.06.tar.gz and hoelz.ro/files/nqp-2015.06.tar.gz.asc 03:15
03:19 akakcolin left 03:23 dwarring left 03:37 jack_rabbit left 03:49 jack_rabbit joined 03:51 jepeway left 03:52 skids left
TimToady waves from Wilson, WY (near Jackson) 04:04
04:06 prevost joined 04:17 BenGoldberg left 04:20 akakcolin joined 04:32 laouji left, laouji joined 04:37 hobbs left, laouji left 04:44 Sqirrel left, Sqirrel joined 04:46 laouji joined 04:48 prevost left 04:51 kaare_ joined 04:53 hobbs joined
zostay is there a way to bind to a Pair value? 04:58
04:58 laouji left, laouji joined
zostay m: my $x = 'A'; $p = X => \$x; $x = 'B'; $p.perl.say; 05:07
camelia rakudo-moar 02c791: OUTPUT«5===SORRY!5=== Error while compiling /tmp/KnFCyyl9K6␤Variable '$p' is not declared␤at /tmp/KnFCyyl9K6:1␤------> 3my $x = 'A'; 7⏏5$p = X => \$x; $x = 'B'; $p.perl.say;␤»
zostay m: my $x = 'A'; my $p = X => \$x; $x = 'B'; $p.perl.say;
camelia rakudo-moar 02c791: OUTPUT«:X(\("B"))␤»
zostay that's the closest i can come up with
TimToady m: my $p = x => 42; $p.value = 43; say $p 05:22
camelia rakudo-moar 02c791: OUTPUT«x => 43␤»
TimToady why do you need to bind?
TimToady --> zzz & 05:25
zostay i'm not sure i *need* to bind... 05:27
night night 05:28
05:38 ponbiki joined 05:39 ponbiki is now known as Guest33354 05:54 spider-mario left 05:57 xinming_ joined, spider-mario joined 05:59 domidumont joined, xinming left 06:03 domidumont left, domidumont joined 06:05 diana_olhovik joined 06:06 _mg_ joined 06:09 akakcolin left
ugexe colomon: ABC might actually install if you fix your META.info provides section. You put the same file path for nearly every module 06:11
otherwise its impossible to create a proper build order
looks like it might have been autogenerated? 06:13
06:15 Sqirrel left 06:18 akakcolin joined 06:21 Sqirrel joined 06:45 espadrine joined 06:46 kaare__ joined 06:48 RabidGravy joined 06:49 kaare_ left 06:51 FROGGS joined
RabidGravy marning! 07:00
moritz \o 07:02
07:04 zakharyas joined 07:06 rindolf joined 07:11 TEttinger joined 07:12 larion left 07:15 lizmat left 07:17 [Tux] left 07:28 [Tux] joined 07:30 domidumont left 07:31 domidumont joined 07:44 rindolf left 07:45 kaare__ is now known as kaare_ 07:53 amurf left 07:54 amurf joined, amurf left 07:55 brrt joined 08:01 zakharyas left 08:03 CQ joined 08:14 akakcoli_ joined 08:15 akakcolin left 08:18 espadrine left, darutoko joined 08:19 dakkar joined 08:25 larion joined 08:32 uncleyear left 08:33 uncleyear joined 08:34 spider-mario left 08:37 Ven joined
timotimo alart! 08:46
(i think that's spanish for "i'm awake"?) 08:47
08:51 xinming_ left 08:52 xinming joined
DrForr or "Aiee! LART!" 08:58
09:01 espadrine joined 09:02 kurahaupo left 09:06 zakharyas joined
TEttinger hm, that wasn't what I expected trying to build latest NQP 09:09
$ perl Configure.pl --backends=moar,jvm --gen-moar
Bareword found where operator expected at Configure.pl line 220, near "'tools/build"
09:09 abraxxa joined
TEttinger that is either 'tools/build/Makefile-Moar.in', or $MAKEFILE, 09:12
it also says it may continue line 217, which is: // 'make';
FROGGS ups 09:14
TEttinger: can you change the // to ||? 09:15
TEttinger sure, will try
nwc10 I'd guess that Configure.pl is missing a 'use 5.010'
FROGGS your perl binary is probably to old to know what // is
:/
timotimo wow 09:18
isn't 5.010 already extremely old?
DrForr Oldest version not EOLed, maybe?
(haven't checked, just guessing.)
moritz 5.010.0 is 2007-12-18
timotimo that's just "very old", not "extremely old" 09:19
DrForr Though it would be (sigh) handy to be able to go back to 5.8 for hardcore legacy users, which is to say "most of them" :/
RabidGravy AVAST!
moritz DrForr: iirc two or three major releases are supported, so 5.22, 5.20 and maybe 5.18
hardcore legacy users that want Perl 6? :-) 09:20
timotimo people on strange platforms who want to port perl6, perhaps? :)
TEttinger yep, I have 5.8.8 because MSYS overrode the path
trying to figure out how to get it to see a newer version 09:21
itz_ vendors do try to backport at least some fixes to older perls sometimes
timotimo oh, it's windows
RabidGravy yeah but not "features" 09:22
DrForr moritz: There's a difference between 'legacy users' and 'legacy fanboys' :)
dalek c: ec67361 | moritz++ | lib/Type/Str.pod:
Str.encode returns Blob, not Buf. mt1957++
09:23
TEttinger ok, so you can't build perl6 with a pre-built older perl6, is this right?
I don't see any perl binary named 'perl' in rakudo star from march 09:25
moritz you need a perl 5 to configure/build rakudo. 09:26
timotimo rakudo star doesn't ship a "perl" binary, it's just "perl6" 09:27
TEttinger right ok. installing strawberry perl, should be 5.22 if the version is accurate
itz_ the wrong perl5 version is specified in Configure.pl 09:28
moritz TEttinger: that's new enough :-)
09:29 hoelzro_trying_w left
moritz itz_: what do you mean? That the minimal version requirement in Configure.pl is too low? 09:29
itz_ yes
dalek p: 0dd9463 | FROGGS++ | Configure.pl:
use || instead of 'defined or' to work on older perls
FROGGS we don't need // there anyway 09:30
TEttinger this community rocks :)
itz_ has anyone tried it in 5.8 (I will if noone else is about to)
FROGGS :D
itz_: that's most likely a 5.10 feature 09:31
09:31 larion left
itz_ yes it is .. but is it the only >5.8 feature? 09:31
moritz fwiw perl 5.8.8 is from is from 2006-01-31
TEttinger ha, I was in high school then
moritz m: say Date.today - Date.new(2006-01-31)
camelia rakudo-moar 02c791: OUTPUT«Potential difficulties:␤ Leading 0 does not indicate octal in Perl 6; please use 0o1 if you mean that␤ at /tmp/sXAN9nlNT7:1␤ ------> 3say Date.today - Date.new(2006-017⏏5-31)␤Default constructor for 'Date' only takes named arguments…»
09:31 Ven left
bonsaikitten ... and still the current version in CentOS 5 09:32
moritz m: say Date.today - Date.new('2006-01-31')
camelia rakudo-moar 02c791: OUTPUT«3425␤»
itz_ its still in RHEL 5.x until 2017 or something
nwc10 and the RHEL with 5.8.8 is common.
bonsaikitten nwc10: no upgrade path
sigh
DrForr Lots of CentOS users out there stuck on older perls :/ 09:33
bonsaikitten DrForr: the only 'upgrade path' is nuke and reinstall, forget about reusing anything 09:34
so people stay on centos5 until the pain is big enough
DrForr Been there many times over.
bonsaikitten I had to install newer java on a centos5 machine this week 09:35
09:35 bin_005 joined
DrForr I'm simply pointing out that being able to build on older boxes might help adoption. (and having much the same issue here at work, CentOS 5 for the main project and not able to install a newer perl to bootstrap to perl6.) 09:38
Though it heartens me that BiP is on 6.5. 09:40
09:40 andreoss joined
timotimo what is BiP? 09:42
DrForr Built In Perl.
09:42 amurf joined
andreoss [ptc]: have you had a chance to look at hanged build process? 09:43
09:44 brrt left 09:45 lolisa joined
TEttinger sweet, it looks like it's building now with 5.22 09:45
09:47 amurf left
TEttinger aaand no 09:51
anyone built NQP on windows?
probing whether your compiler thinks that it is gcc Can't compile simple gcc probe, so something is badly wrong at build/probe.pm line 92.
FROGGS TEttinger: that message is from MoarVM
TEttinger I actually have gcc with both mingw and cygwin
itz_ grrr 5.8.9 doesn't even install on the three platforms I tried without hacking
FROGGS TEttinger: try this in nqp: perl Configure.pl --backend=moar --moar-option='--os=mingw32' 09:52
TEttinger: though, if the cygwin gcc is in PATH before the mingw gcc, you won't be lucky 09:53
DrForr itz_: Maybe perlbrew to the rescue?
TEttinger I haven't done gen-moar yet
FROGGS TEttinger: but your message is from moar, I'm pretty sure about that 09:54
TEttinger woo FROGGS++
FROGGS ncie
nice*
TEttinger yeah I added --gen-moar to that, it's building now
FROGGS awesome
itz_ DrForr: I doubt that works unless it contains backports to clang etc 09:55
it would be easiest just to spin up Centos 5.x VM
09:55 hoelzro_trying_w joined
DrForr yep, that's pretty much what I ended up having to do. 09:56
TEttinger ok, I got an error building pastie.org/private/1yumfq1kbnv6sg1u6ht9ia 09:57
timotimo /bin/sh: c:Strawberryperlbinperl.exe: command not found
yeah, that looks great!
TEttinger beforethat
itz_ DrForr: actually if you have that .. you could just perl -wc on the various Configure.pl (rakudo, nqp etc) with the 'use 5.008;' lines
TEttinger dyncall_callf.c:90:75: error: invalid conversion from 'DCpointer {aka void*}' to 'DCstring {aka const char*}' [-fpermissive] 09:58
timotimo right
09:58 dayangkun left, zj joined, Ven joined
timotimo that's 3rdparty code we include, i wonder why it doesn't work on your system 09:58
TEttinger steve ballmer? 09:59
timotimo it probably wants to have a cast in there, but it seems that cast isn't necessary with other compilers
though you're using gcc, so you're not on "another compiler"r eally
TEttinger I'll check that it's a recent mingw
mingw. gcc.exe (GCC) 4.8.1
timotimo is -fpermissive the flag to allow this to not be an error?
if you set CFLAGS in your environment when building dyncall, maybe it'll work 10:00
TEttinger this is an interesting experience so far
really not that many windows devs huh
FROGGS there are some, but they're not using cygwin 10:01
TEttinger nor am I
DrForr There's a reason :) Though I should really start.
FROGGS TEttinger: can you rename your cygwin folder, and try again?
TEttinger mingw
I have cygwin, but it's not using that gcc
FROGGS but perhaps stuff from cygwin is in PATH 10:02
TEttinger this is the mingw bash shell, `which gcc` outputs: /mingw/bin/gcc.exe
FROGGS I just want to make sure it does not pick anything up
ohh... I use cmd.exe only
can you try that instead?
TEttinger aha!
`which gmake` : /c/Strawberry/c/bin/gmake.exe 10:03
itz_ --moar-option=--cc=/whatever/cc?
TEttinger strawberry perl's gmake
10:03 kaare_ left
TEttinger I didn't get cmd to work yet 10:03
FROGGS gmake is fine...
I just fixed that last week
TEttinger itz_, I will try that in cmd
FROGGS or this week or so
10:06 abraxxa left, abraxxa joined 10:08 larion joined 10:10 espadrine left
TEttinger ok, this is freaky. 10:11
the first gcc on the PATH is MinGW's GCC. which gcc confirms this
it isn't correctly responding to the probe
10:11 _mg_ left 10:14 larion left
TEttinger what gcc doe you use, FROGGS? 10:14
FROGGS strawberry's 10:15
TEttinger I tried TDM-GCC-64, it didn't work either
FROGGS and its gmake
TEttinger oh!
FROGGS in cmd.exe
bbiab lunch&
10:15 akakcoli_ left
JimmyZ or you can install msvc and use it in CMD too :) 10:16
lolisa Hi, total beginner here, are there anyway to stop list of list from automatic flattening?
10:16 Ven left
TEttinger well strawberry perl's gcc doesn't work either. 10:16
same error, doesn't respond to compiling a probe 10:17
10:17 espadrine_ joined
DrForr lolisa: Sure, use a reference to the list contents. foo( [1,2,3], [4,5,6] ); sub foo($r1,$r2) { $r2.[2]++ } 10:17
lolisa DrForr, thx 10:18
DrForr new to perl5 and 6, or just 6?
lolisa new to dynamically typed language
DrForr Cool. How'd you find perl6? 10:19
TEttinger JimmyZ: I have MSVC, can I use its weird toolchain to compile Moar?
10:19 Akagi201 left
lolisa one Haskeller has been telling me how great perl6 is and I dont have the taste for python anyway, so there isnt really much choice to choose what dynamic language to learn (well, scheme is another choice but that's it 10:20
JimmyZ TEttinger: yeah, open the MSVC tool cmd, and make sure perl is in your PATH
DrForr Also, potential bug-o-the-day - 'Use of Nil in string context in method element...' # The ' ' suggests that there's a missing something in there.
'string context in' # The spaces are there, my irssi chose an awkward spot to linebreak. 10:21
10:22 itz joined
TEttinger ok, seems to be doing... something with VS now. thanks, JimmyZ++ 10:31
woah. 10:35
never ever seen this before.
src\strings\unicode.c(45034): fatal error C1026: parser stack overflow, program too complex
andreoss m: enum Letters <A B C E D F G H>; my @a = A..C; my @b = A,B,C; say @a eq @b;
camelia rakudo-moar 02c791: OUTPUT«False␤»
andreoss why enums are converted to Ints? 10:36
TEttinger enum Letters <A B C E D F G H>; my @a = A..C; my @b = A,B,C; say @a eqv @b;
m: enum Letters <A B C E D F G H>; my @a = A..C; my @b = A,B,C; say @a eqv @b; 10:37
camelia rakudo-moar 02c791: OUTPUT«False␤»
TEttinger interesting
andreoss m: enum Letters <A B C E D F G H>; my @a = A..C; say @a».perl;
camelia rakudo-moar 02c791: OUTPUT«0 1 2␤»
TEttinger andreoss, oh, because they're enumerated types, probably :)
RabidGravy I guess, that the range constructor coerces them to either Ints or Strs before creating the range 10:38
andreoss TEttinger: how can i stick with aliases i gave them here?
TEttinger this is my, just starting 2nd day of being aware of perl 6 10:39
so my answer is I dunno?
psch m: enum Letters <A B C E D F G H>; say A ~~ Str; say A ~~ Int # this is why they get converted 10:40
camelia rakudo-moar 02c791: OUTPUT«False␤True␤»
moritz m: enum Letters <A B C E D F G H>; say A.succ
camelia rakudo-moar 02c791: OUTPUT«1␤»
moritz andreoss: you can't, I think
TEttinger names are names, their values are those numbers
10:40 uncleyear left
psch also hey #perl6 o/ 10:41
TEttinger hey psch
so compiling Moar with VS 2015's cl.exe has an interesting error
10:41 uncleyear joined
TEttinger src\strings\unicode.c(45034): fatal error C1026: parser stack overflow, program too complex 10:41
timotimo m: enum Letters <A B C E D F G H>; say A.WHAT.values.perl 10:42
camelia rakudo-moar 02c791: OUTPUT«()␤»
timotimo m: enum Letters <A B C E D F G H>; say A.WHAT.keys.perl
camelia rakudo-moar 02c791: OUTPUT«()␤»
timotimo m: enum Letters <A B C E D F G H>; say A.WHO.keys.perl
camelia rakudo-moar 02c791: OUTPUT«("F", "B", "A", "H", "G", "D", "E", "C")␤»
timotimo m: enum Letters <A B C E D F G H>; say A.WHO.values.perl
camelia rakudo-moar 02c791: OUTPUT«(Letters::F, Letters::B, Letters::A, Letters::H, Letters::G, Letters::D, Letters::E, Letters::C)␤»
timotimo you could work with this?
RabidGravy andreoss, if you really want to do something like that you will need to make them back into the enums: 10:45
m: enum Letters <A B C E D F G H>; my @a = (A..C).map({Letters($_)}); my @b = A, B, C; say @a ~~ @b
camelia rakudo-moar 02c791: OUTPUT«True␤»
RabidGravy for istance
instance 10:46
10:53 larion joined
timotimo m: enum Letters <A B C E D F G H>; say Letters::[A..C].perl; 10:53
camelia rakudo-moar 02c791: OUTPUT«Method 'parameterize' not found for invocant of class 'Perl6::Metamodel::EnumHOW'␤ in block <unit> at /tmp/8lMQYtMyOk:1␤␤»
timotimo m: enum Letters <A B C E D F G H>; say Letters::(A..C).perl;
camelia rakudo-moar 02c791: OUTPUT«5===SORRY!5=== Error while compiling /tmp/4tFt5FKK1p␤Combination of indirect name lookup and call not supported␤at /tmp/4tFt5FKK1p:1␤------> 3s <A B C E D F G H>; say Letters::(A..C)7⏏5.perl;␤ expecting any of:␤ argument list␤»
timotimo m: enum Letters <A B C E D F G H>; say Letters::.[A..C].perl;
camelia rakudo-moar 02c791: OUTPUT«(:F(Letters::F), :B(Letters::B), :A(Letters::A), :H(Letters::H), :G(Letters::G), :D(Letters::D), :E(Letters::E), :C(Letters::C), Failure.new(exception => X::OutOfRange.new(what => "Index", got => 1, range => "0..0", comment => Any), backtrace => Backtrace.…»
timotimo m: enum Letters <A B C E D F G H>; say Letters::.list.[A..C].perl;
camelia rakudo-moar 02c791: OUTPUT«(:F(Letters::F), :B(Letters::B), :A(Letters::A))␤»
timotimo m: enum Letters <A B C E D F G H>; say Letters::.values.[A..C].perl;
camelia rakudo-moar 02c791: OUTPUT«(Letters::F, Letters::B, Letters::A)␤»
timotimo yikes! :)
never rely on hash ordering, folks 10:54
or package ordering for that matter
10:57 bin_005 left 11:01 abraxxa left 11:02 abraxxa joined, brrt joined 11:06 laouji left 11:07 laouji joined 11:09 regreg joined
FROGGS TEttinger: I know that VS 2012 works 11:09
11:09 itz left, brrt left
timotimo anyway, the working version of that code above would be this: 11:09
TEttinger trying with 2013
timotimo m: enum Letters <A B C E D F G H>; say Letters::{'A'..'C'}.perl
camelia rakudo-moar 02c791: OUTPUT«(Letters::A, Letters::B, Letters::C)␤»
11:10 itz joined 11:14 gagalicious left 11:15 Guest33354 is now known as ponbiki 11:16 brrt joined, gagalicious joined
RabidGravy testers.perl6.org is doing that "Test report submitted as: testers.perl6.org/reports/</html>.html" again if anyone is interested 11:16
11:17 zacts left 11:18 zacts joined 11:19 _mg_ joined
dalek osystem: 1dc6aa5 | RabidGravy++ | META.list:
Add AccessorFacade to the modules list
11:20
RabidGravy there, not big, not clever but if nothing else it may serve as an example that you can do funky things with traits without having to bake them into the language 11:21
11:25 laouji left 11:27 dalek left 11:28 dalek joined, ChanServ sets mode: +v dalek 11:29 Ven joined 11:32 amurf joined 11:36 larion left 11:37 larion joined, amurf left
ab5tract good *, #perl6 11:41
hoelzro o/ ab5tract
ab5tract I was playing with ranges recently, and was a bit bemused as to what op, if any, would take two ranges and return True/False based on whether there is any overlap between the two 11:42
moritz why I start to prefer web apps over desktop applications: find-as-you-type 11:43
ab5tract I didn't see any obvious methods defined on Range, at least in the docs
11:44 abraxxa left, kaare_ joined, abraxxa joined
moritz I don't know about such an op. 11:45
ab5tract moritz: funny, find-as-you-type is something I would love to add to `p6doc` (or through a separate wrapper)
11:45 aborazmeh joined, aborazmeh left, aborazmeh joined 11:46 skids joined
ab5tract moritz: would adding an 'overlaps' method to Range make sense? 11:48
11:49 uncleyear left, uncleyear joined
moritz ab5tract: overlaps-with 11:50
maybe
though since it's a symmetric operation, it would better fit into a sub or method
m: say (1..3) ~~ (0..5) 11:51
camelia rakudo-moar 02c791: OUTPUT«True␤»
moritz m: say (1..3) ~~ (2..5)
camelia rakudo-moar 02c791: OUTPUT«False␤»
moritz ah, it's "contained in", not "overlaps"
m: say (1..3) ~~ (1^..3)
camelia rakudo-moar 02c791: OUTPUT«False␤»
moritz m: say (1^..3) ~~ (1..3) 11:52
camelia rakudo-moar 02c791: OUTPUT«True␤»
ab5tract moritz: Indeed. I thought of it as a moment where smart match has to make a choice where DWIM is a bit fuzzy. 11:55
Have there ever been adverbs on smart match?
I imagine that it would get too contextual and dirty, quickly 11:56
moritz the idiomatic way is to say all(range1) ~~ range2 11:57
but it's neither efficient nor totally correct
11:58 AlexDaniel joined, abraxxa left 11:59 abraxxa joined
Mouq m: say {$^a.min <= $^b.min <= $a.max or $b.min <= $a.min <= $b.max}(1..5,2..6) 12:04
camelia rakudo-moar 02c791: OUTPUT«True␤»
ab5tract moritz: do you mean any(range1) ~~ range2 ?
12:05 skids left
Mouq m: say {$^b.min ~~ $^a or $a.min ~~ $b}(1..5,2..6) 12:05
camelia rakudo-moar 02c791: OUTPUT«True␤»
moritz ab5tract: ah, yes
ab5tract Mouq: that second one is nice and idiomatic as well
moritz m: say so any(^1) ~~ (0.1..0.9)
camelia rakudo-moar 02c791: OUTPUT«False␤»
moritz ... and ignores excluded endpoints
Mouq moritz: Ahh, dang 12:06
ab5tract Mouq: however I still think there is an argument for encapsulating that inside a method that ranges can carry around with them
moritz: true that :)
Mouq m: say {$^b.min ~~ $^a or $a.min ~~ $b}(1..5,5^..6)
camelia rakudo-moar 02c791: OUTPUT«True␤»
Mouq Hm… yeah, trying to account for that is a mess. You can be elegantish but not real succint 12:08
12:08 akakcolin joined
ab5tract Mouq: which also makes me yearn for something like an 'effective-min', which can do nearly-inevitable ternary conditional phrase 12:09
so that you don't have to
moritz so, what would effective-min return? 12:10
RabidGravy 85 unique module authors, that's quite healthy 12:12
ab5tract m: my $exc-min = False; my $min= 5; sub effective-min { say $exc-min ?? $min+1 !! $min }; effective-min; $exc-min = True; effective-min 12:13
camelia rakudo-moar 02c791: OUTPUT«5␤6␤»
Mouq ab5tract: My thought is that the min of 5^..6 is 5.00000000…00000001 12:14
:P
ab5tract Mouq: very fair point :)
moritz 5 + Epsilon 12:18
masak this is why real math makes a distinction between "minimum" and "infimum". 12:19
the range 5^..6 strictly has no minimum.
it has an infimum, though. it's 5.
moritz because for every Epsilon, you can find a smaller Epsilon' so that 5 + Epsilon' is still part of the range 12:20
masak en.wikipedia.org/wiki/Infimum_and_supremum
moritz: right. that end of the interval is open, which amounts to saying the same thing: you can always get closer to the edge.
hoelzro jnthn: regarding changelogs for NQP: should that be rolled into the rakudo changelog?
moritz hoelzro: example? 12:21
DrForr paging Dr. Zeno: Cleanup on aisle 8... no, 4..., no, 2... :)
hoelzro moritz: for example, I fixed that bug yesterday in NQP. NQP doesn't seem to have a changelog, so where should that be recorded?
jnthn hoelzro: I don't think we bother producing an NQP one at the moment.
hoelzro: Well, it's recorded in the commit log... :)
masak DrForr: is the message getting twice as fast with every correction too? :P 12:22
jnthn The chnage log is really just a curated version of that in a sense
hoelzro fair enough =)
DrForr You just can't see the bandwidth restrictions :)
12:23 aborazmeh left
hoelzro so it seems that m-stresstest is passing, but not j-stresstest; how much should I investigate into JVM-specific failures for the release? 12:23
moritz hoelzro: if it was a bug that manifested itself in Perl 6 code, you can mention it in the rakudo changelog (if it seems noteworthy; we don't mention every change or bugfix) 12:25
hoelzro alright
Mouq m: class ε {}; role Greater[$n] { method near { $n }}; multi infix:«<»(Numeric:D $a, Greater $b) { $a == $b.near ?? True !! $a < $b.near }; multi infix:<+>(Numeric:D $n, ε:U $) { Greater[$n] }; say 5 < 5 + ε; say 5.00000001 < 5 + ε 12:26
camelia rakudo-moar 02c791: OUTPUT«True␤False␤»
12:27 uncleyear left, uncleyear joined
dalek pan style="color: #395be5">perl6-examples: 6c2842a | andreoss++ | categories/euler/prob054-andreoss.pl:
[euler] prob054 using objects as hash keys
12:35
andreoss can i import java::util::*: with perl6-j? 12:41
12:43 Ven left
dalek kudo/nom: d71640a | hoelzro++ | docs/ChangeLog:
Update changelog for release
12:56
kudo/nom: 1fd3f0d | hoelzro++ | docs/announce/2015.06.md:
Update 2015.06 announcement with final changes/contributors
12:57 telex left 12:58 telex joined
dalek kudo/nom: 1ee7c94 | hoelzro++ | tools/build/NQP_REVISION:
[release] bump NQP revision
13:00
kudo/nom: 8532c0f | hoelzro++ | VERSION:
[release] bump VERSION
13:05 Akagi201 joined
masak ooh. blog.ionelmc.ro/2014/08/03/the-most...-python-3/ 13:09
how do we handle this one in Perl 6?
(displaying nested exceptions) 13:10
13:11 [Sno] left, [Sno] joined
masak they call it "chained exceptions" in the HN comments. 13:11
13:11 aborazmeh joined, aborazmeh left, aborazmeh joined
masak I'm pretty sure we have support for the nesting/chaining itself... but how do we display that, as a stacktrace? 13:11
13:12 uncleyear left 13:13 uncleyear joined
FROGGS masak: I just had that here in channel the other day 13:14
timotimo we have that for failure objects
when they get thrown
FROGGS m: sub f returns Int { fail "a" }; f() # like this
camelia rakudo-moar 02c791: OUTPUT«Earlier failure:␤ a␤ in block <unit> at /tmp/CfDNyAcd95:1␤␤Final error:␤ Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:639␤ in sub f at /tmp/CfDNyAcd95:1␤ in block…»
FROGGS masak: ^^
13:16 yqt joined
psch fwiw, that's changing, in reference to RT #64990 13:17
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...l?id=64990
13:17 molaf joined
psch i've finished moar and jvm letting Failure slip through p6typecheckrv, but still to adjust the Optimizer to not inline non-native return types anymore 13:17
13:17 uncleyear left, uncleyear joined
psch +need 13:18
m: sub f returns Int { fail "a" }; say f() ~~ Failure # will print "True" then
camelia rakudo-moar 02c791: OUTPUT«Earlier failure:␤ a␤ in block <unit> at /tmp/bmS5b1iraL:1␤␤Final error:␤ Type check failed for return value; expected 'Int' but got 'Failure'␤ in any return_error at src/vm/moar/Perl6/Ops.nqp:639␤ in sub f at /tmp/bmS5b1iraL:1␤ in block…»
psch i suppose that's only relevant to that specific example, though 13:20
i.e. the "Earlier failure: ...\nFinal error:..." things stays 13:21
hoelzro how much should I care about JVM stresstest failures pre-release? 13:22
there's a handful, but they seem to be around GLR-looking things
moritz ususally doesn't care too much about JVM stresstest failures 13:23
hoelzro alright, I'll not worry about them, then 13:25
psch jnthn: i'm not sure how the guarding for native types only is supposed to look. in my understanding i'd just skip all of Perl6::Optimizer.inline_call..? 13:27
masak FROGGS: oh, nice. then it ought to be possible to do something nice for chained exceptions as well.
TimToady: suddenly I'm curious about why it's infix:<.> in Perl 5. do you remember what went into that particular language decision?
(not asked in a critical way at all. pre-OO, it seems to have been a very sensible choice. and I know fivers who would still be ready to defend it.) 13:31
13:35 vimal2012 joined
zostay is the .key of a Pair always a string? 13:38
13:39 zj left, domidumont left
zostay i ask because i see some eq comparisons in Enum for AT-KEY and EXISTS-KEY 13:39
i would have thought those would be eqbv
13:39 domidumont joined
zostay eqv 13:39
timotimo m: say (True => 1).key.perl 13:41
camelia rakudo-moar 8532c0: OUTPUT«"True"␤»
timotimo i believe Pair coerces
13:41 brrt left
timotimo oh, wait 13:41
that's autoquoted
m: say ((True) => 1).key.perl 13:42
camelia rakudo-moar 8532c0: OUTPUT«Bool::True␤»
13:42 skids joined
timotimo so yeah 13:42
but hashes are String-keyed by default
13:44 uncleyear left 13:45 uncleyear joined
zostay it'd be nice if there was a TypedPair to provide parity with TypeHash 13:45
13:47 mrf left
hoelzro would someone mind uploading the rakudo tarball and signature for me? they're at hoelz.ro/files/rakudo-2015.06.tar.gz and hoelz.ro/files/rakudo-2015.06.tar.gz.asc 13:50
itz can someone close rt.perl.org/Ticket/Display.html?id=123728 please?
hoelzro itz: done! 13:51
itz ty
moritz hoelzro: has the nqp tarball been uploaded already?
FROGGS hoelzro: I can
moritz: not that I can see
hoelzro moritz: no, it's on my server too
FROGGS hoelzro: is the nqp tarball ready too? 13:52
hoelzro whenever someone gets a chance, please upload the NQP tarball and signature; they're at hoelz.ro/files/nqp-2015.06.tar.gz and hoelz.ro/files/nqp-2015.06.tar.gz.asc
FROGGS and moarvm also?
hoelzro MoarVM too
FROGGS that's what only jnthn can do
hoelzro if anyone's around with upload permissions, I've uploaded the tarball to hoelz.ro/files/MoarVM-2015.06.tar.gz
moritz hoelzro: nqp downloaded
erm, uploaded 13:53
whatever :-)
13:53 vimal2012 left
FROGGS moritz: and the signature? 13:53
moritz done 13:54
FROGGS moritz++
hoelzro moritz++ indeed
moritz rakudo done too
hoelzro horray
we just need jnthn to upload MoarVM, and I'll send out the release announcement 13:55
jnthn hoelzro: Where's the release now? 13:56
hoelzro hoelz.ro/files/MoarVM-2015.06.tar.gz
jnthn Done. 13:57
hoelzro \o/
alright, out goes the email
psch hoelzro++
skids hoelzro++ 13:58
moritz hoelzro++ # release
nwc10 hammers reload on the website to see when it will change 13:59
er, moarvm.org/ that is
hoelzro++ # shipping stuff 14:00
jnthn nwc10: I didn't change *that*
14:00 mrf joined
jnthn Just wget'd the file into place 14:00
FROGGS hoelzro++ 14:01
hoelzro "You're done! Celebrate with the appropriate amount of fun."
\o/
FROGGS *g*
14:02 lolisa left
CQ is there a roadmap somewhere on what has to happen (major tasks / milestones?) for perl6 to turn RC / release this year? 14:05
14:10 JimmyZ_ joined 14:16 lizmat joined 14:17 _mg_ left 14:18 regreg left 14:19 brrt joined, regreg joined 14:20 lucasb joined
PerlJam hoelzro++ 14:21
sjn hoelzro++ # well volunteered 14:23
lucasb Thanks for the release! My installed rakudo was two months old, will upgrade now. 14:24
14:24 maettu left 14:25 run4flat joined 14:26 maettu joined
run4flat yesterday jnthn mentioned that there is no email forum for discussing PDLish stuff for Perl 6 14:26
I was thinking more about this
PerlJam CQ: Since it's the test suite that defines the language, I think it's mostly just marking up the tests appropriately.
run4flat since the PDL-related stuff in S09 hasn't been touched much, it'll need some work to bring up to date
wouldn't perl6-language be the forum for discussing that? 14:27
I just feel like I would like an extended discussion on some of these things, i.e. not on irc
PerlJam run4flat: Sure, it would be *a* forum for discussing that :)
run4flat: so would #perl6
run4flat Or, should I take my ideas to perl6-compiler? 14:28
brrt what is PDL
PerlJam run4flat: Are you ideas compiler-specific? Like ways to implement some optimized data structure?
run4flat Perl Data Langauge
brrt ah
run4flat like matlab, but for Perl (5)
PerlJam, no 14:29
I'm interested in talking about specs
PerlJam run4flat: then, perl6-language sounds like the right place to me
run4flat ... I guess I could write my ideas in a blog post, or discuss them on one of PDL's mailing lists, too 14:30
I just feel like writing something lengthy, not chit-chatty
run4flat cogitates on the blog-post idea
PerlJam run4flat: IF you do the blog post, make sure to mention the URL here 14:31
run4flat yeah, good call
PerlJam kinda wishes we could take all of the work the PDL folks have already done and somehow integrate that into Perl 6. I haven't looked, but my gut tells me that's probably not possible 14:32
14:33 bjz joined
run4flat shares that wish, but doesn't believe it's possible, or the right way for Perl 6 14:34
14:34 muraiki joined 14:42 abraxxa left, abraxxa joined 14:43 uncleyear left
psch hm, i don't know about changing Optimizer.nqp — it already looks like it works right from here: gist.github.com/peschwa/f399cf10914f1d287a74 14:44
14:44 uncleyear joined 14:51 gcole left 14:55 run4flat left 14:56 FROGGS left 14:58 gagalicious left 14:59 preyalone joined 15:00 aborazmeh left, brrt left, brrt joined 15:03 gagalicious joined 15:05 uncleyear left 15:06 uncleyear joined
JimmyZ_ psch: you forget to release failure_o 15:09
15:09 amurf joined
psch JimmyZ_: thanks, corrected that 15:13
15:14 amurf left 15:19 larion left 15:24 FROGGS joined 15:25 uncleyear left 15:26 gcole joined, uncleyear joined 15:29 Ven joined 15:33 gfldex joined 15:34 zakharyas left
psch jnthn: can i provoke the inlining problems somehow? 'cause as mentioned, it looks fine as-is. 15:37
regreg is there any reason why parrot fell out of favor other than implementation issues? 15:42
like non-inherent performance issues
15:45 slavik left 15:49 slavik joined 15:50 diana_olhovik left
brrt hmm, wish you hadn't asked that question 15:52
ugexe because they want to release in December and maintaing support for 3 VMs would be too much. MoarVM is built for p6, so its the obvious choice for VM to get things done in time. JVM gives them a stable VM to test stuff with
15:52 uncleyear left 15:53 uncleyear joined
brrt JVM is also 'popular' in an IT sense 15:53
ugexe right. its got countless man hours behind it, hardening, etc
15:54 domidumont left
brrt that too, and IT managers trust it 15:54
ugexe i think NQP might be maintaining parrot support
PerlJam ugexe: I don't buy your "3 VMS is too much" argument :) 15:55
ugexe its not 'too much', just too much to get done before december
[ptc] andreoss: you mean in the examples website? 15:56
brrt anyway; i think you can safely say that parrots goals were too ambitious, and that this is inherently a performance issue
[ptc] andreoss: last time I looked, it built ok, just took aaaaaaagggges to run, and no colour came out via Text::VimColour :S
PerlJam brrt: mostly a "people performance" issue :)
15:57 Ven left
brrt i.e. parrot wanted to support all dynamic languages. all dynamic languages are really different. ergo, these differences have to be abstracted. irreducible abstraction means poor performance 15:57
well, yes, parrot had lost the wind from its sails
to make matters worse, parrot had it's own PMC object system that didn't quite map onto perl6's object system 15:58
PerlJam brrt: to be fair .... Parrot had PMCs long before Perl 6 knew what kind of object system it needed/wanted 15:59
brrt oh, yes
in retrospect one of the greater mistakes (if i may call it that?) was starting on parrot before properly knowing what perl6 would be like 16:00
PerlJam but, yeah, momentum kind of impeded parrot's progress towards Perl 6.
16:00 lichtkind joined
PerlJam brrt: I don't think that was a mistake. We had to start somewhere. 16:01
brrt: perhaps the real mistake is building something of unknown suitability for Perl 6, then building infrastructure on top of it making it harder to change.
brrt: also, all the while trying to cater to "all dynamic languages" rather than just Perl 6 16:02
moritz if you think of parrot as the prototype to throw away, as in the old saying "plan to throw one away, because you will anyway"
then parrot was quite a success
PerlJam moritz: indeed
moritz and still is, because we wouldn't have MoarVM without it
brrt yes, that is true 16:03
also as a good cold lesson in the nature of programming languages
JimmyZ_ stepping-stone
fork mkarcvm 16:04
16:04 uncleyear left
JimmyZ_ fork moarvm and s/moarvm/parrot/g and you will have a new parrot if it is allowed ;) 16:05
itz I think the vim colour thing should work now .
with examples.perl6.org .. I added a style sheet 16:06
brrt moritz: my only counterpoint would be that parrot was awfully overengineered to be a prototype
but overengineering is also a form of experimentation
itz [ptc]: when did you last try it?
16:06 muethos joined 16:11 brrt left
moritz stefan-marr.de/papers/pldi-marr-et-...ogramming/ 16:13
16:14 yqt left 16:15 yqt joined 16:18 akakcolin left
itz how did the first stage0 files get generated? :) 16:30
16:31 domidumont joined
moritz it got bootstrapped by nqp-rx 16:31
16:32 diana_olhovik_ joined
moritz which in turn got bootstrapped by the previous nqp 16:32
which in turn was bootstrapped by PGE, which was written in PIR
16:33 spider-mario joined
hoelzro here I was thinking it was sung into existence 16:34
moritz
.oO( sung cost )
16:39
mst sungo cost: one live badger 16:42
[ptc] itz: a couple of days ago, I forget exactly when. Around the time you recommended to me that I should give it another go 16:43
itz: on the 14th of June at just after 1pm, to be more exact 16:44
andreoss [ptc]: i guess Text::VimColour is not installed, also vim 7.4 should be installed too 16:45
with plugins for perl6 16:46
16:47 larion joined 16:49 mohij joined
[ptc] andreoss: hrm, I thought it was. I'll have another look 16:49
andreoss also it seems to use a bit outdated rakudo 2015.04 16:50
[ptc] andreoss: I was getting the line numbers, but not the colour
andreoss: ok, I need to update the rakudo used for the examples 16:51
colomon ugexe: I’m very confused by your patch, as panda installs ABC for me without any issue already. 16:53
andreoss [ptc]: line numbers in html?
[ptc] andreoss: the rakudo used on examples.perl6.org should use the most recent rakudo as installed on hack 16:54
andreoss: yes
16:54 nys joined
[ptc] andreoss: that's in my local build of the examples 16:54
AlexDaniel hoelzro: fixed? What error are you seeing? I've just built the whole thing with rakudobrew 15 minutes ago and the problem is still there (rt.perl.org/Public/Bug/Display.html?id=125339)
andreoss probably some local .vimrc caused that
[ptc] duh, yes, just realised that was me... sigh 16:55
ugexe colomon: well, i use stricter rules that only use whats in the meta. anyways if you look at what is actually changed there is little to be confused about
AlexDaniel m: take 1
camelia rakudo-moar 8532c0: OUTPUT«take without gather␤␤»
[ptc] strange that I don't get the colours though... I use the perl6 plugin which is more up to date than that delivered with vim on debian jessie
andreoss [ptc]: probably css is missing, let me check 16:56
16:57 JimmyZ_ left
[ptc] andreoss: does Text::VimColour require that 'syntax on' is set somewhere in the process calling vim? 16:58
andreoss: it's just that the output on examples.perl6.org is also without colour
andreoss syntax on is default 16:59
[ptc] ok 17:01
itz you need 7.4 which is in jessie I think
[ptc] itz: yup, it's 7.4
it's also 7.4 on hack 17:02
andreoss [ptc]: also are there perl6 plugins on examples.perl6.org?
it does set lang=perl, but there might be no coloring for that
itz I imported html/css/vim.css
[ptc] andreoss: I just opened a perl6 file on hack.perl6.org (where examples.perl6.org is hosted) and was able to correctly syntax highlight using perl5 and perl6 17:03
my guess is that the plugins delivered with debian are the ones used on hack 17:04
should we be using the updated ones from vim-perl?
itz I used vim from hg without upgraded vim plugins 17:05
17:05 espadrine_ left 17:06 preyalone left
itz hmmm I wonder if it's missing a debian package 17:07
[ptc] hrm, let me check that hack actually has Text::VimColour installed...
itz try :toHTML within vim 17:08
oops :TOhtml
andreoss i get no highlighting without vim-perl plugin 17:10
[ptc] itz: should I be seeing anything in particular with :TOhtml (other than the html, of course)
itz html I guess
I have vim-common vim-runtime vim-tiny 17:11
[ptc] just installing Text::VimColour in hack.p6c.org's rakudobrewed perl6 17:13
andreoss itz: your name on github is stmuk? 17:14
itz yes that's me seve
err steve
17:15 Psyche^ joined
[ptc] is building the examples website by hand 17:16
itz I'm wondering if there is a dependency on a particular .vimrc or maybe the filename vim is called with 17:18
17:19 uncleyear joined, abraxxa left
timotimo or maybe vim-perl or vim-python or something like that? 17:21
[ptc] if code can be syntax highlighted with vim on hack, then it should work 17:22
my guess is that vim-perl is installed, however it's a virtual package, so 'aptitude search vim-perl' isn't that helpful...
itz examples.perl6.org/css/vim.css does need to be in place
I'm sure it doesn't need a perl linked into vim 17:23
[ptc] itz: that css file doesn't seem to be available on examples.perl6.org; I'm getting a 404
andreoss git pull? 17:24
itz I checked it into git in the last week or so
[ptc] it's there on hack
17:25 ShimmerFairy left
[ptc] but maybe it's not getting synced to the web site for some reason 17:25
ugexe [ptc]: made a comment about panda-test for travis-build
[ptc] ugexe: ok, haven't seen that yet 17:26
ugexe: we're installing panda as part of the rakudo installation on travis 17:27
ugexe ouch. panda still breaks regularly, so it will end up blocking results 17:29
[ptc] ugexe: hoelzro recommended that we use panda-test, since it is part of the current toolchain (I had used a similar incantation to what you recommend as an alternative)
ugexe: how should users then install dependent modules?
ugexe install panda, its 1 extra command. and then they can choose how its installed as well 17:30
[ptc] ugexe: my guess was that we should provide something similar to cpanm so that they can install dependencies required for them to test their modules
ugexe make it easy to install panda. but if you have no dependencies then you have no need for it. cpanm doesnt break regularly, so thats different
[ptc] it's all still somewhat up in the air if everything works on travis the way we want it from the word go; testing the travis scripts on my dev machine were... difficult 17:31
ugexe: ok, that's a good argument
strange that BanzaiMan pushed to get the updates in but then hasn't said boo since... 17:32
17:32 meisl joined 17:34 mrf left
[ptc] itz, andreoss: btw: if things on examples.perl6.org don't seem to be working the way we expect, then examples.perl6.org/build-log/ is a good place to look 17:34
17:34 mrf joined
itz ok thanks 17:34
[ptc] I just found that the html wasn't being built since Text::VimColour wasn't installed :-) 17:35
meisl hello #perl6
17:35 akakcolin joined
itz I've just discovered Text::VimColour doesn't pass the tests if I delete my .vimrc anyway :) 17:36
17:37 ShimmerFairy joined
ugexe [ptc]: looks like he regularly does blog posts, but hasnt in the last 18 days. so he is probably on vacation or something 17:38
[ptc] good point
hoelzro AlexDaniel: oh, my bad; I did a 'take;'
andreoss i wonder if it would be possible to highlight syntax based on AST in Perl6, like Eclipse does for Java 17:40
timotimo i'm pretty sure i saw a simple proof-of-concept recently 17:41
andreoss probably parse code with STD, and it would be a graph coloring problem
[ptc] hrm, vim.css is now there, however the perl6 code still isn't being highlighted 17:42
andreoss though the really interesting thighs to have would be a scope coloring or context coloring. 17:43
17:43 diana_olhovik_ left
itz is this v0.2? do the tests pass? 17:43
andreoss [ptc]: is the code inside <pre>?
[ptc] itz: well, it installed via panda, so the tests should have passed 17:44
itz actually i noticed the brew panda I just used seemed to skip tests
[ptc] itz: yup, it's v0.2 17:45
andreoss: yes, it's inside a <pre> 17:46
itz I think there is a further problem .. there is a hidden dependency on some lines in my .vimrc (which is quite long)
andreoss [ptc]: it tries to call vim, but if it fails for soe reason the CATCH block wraps code in <pre> 17:47
itz I think its the issue where it just hangs .. I see that with the tests and no .vimrc present on jessie
andreoss paste.debian.net/238648/ tests pass with my .vimrc 17:49
itz there is something in our vimrc which isn't present in the vim shell line
andreoss probably vim in debian has some smart stuff in it 17:51
itz thats quite possible .. I had problems with zsh configuration on debian recently caused by overzealous debian customisation 17:52
my vimrc is 96 lines : 17:53
:/
[ptc] and when there's no vimrc? What happens then? 17:59
itz hangs
18:02 dakkar left
tony-o i always end up building vim from source in debian 18:04
it behaves strangely 18:05
[ptc] itz: I just tried `vim -T builtin_dumb -c 'set noswapfile|set bg=light|set ft=perl6|TOhtml|wq! P01-scottp.html|quit' P01-scottp.pl` in categores/99-problems on hack 18:08
itz oh?
[ptc] itz: it ran through fine (doesn't hang), however the html is wrapped in a <pre>
more specifically, <pre id='vimCodeElement'>
itz thats correct behaavior 18:09
[ptc] should that be highlighted, or is there something missing there?
htm
*hrm
itz that should be the html
I think there is an embedded style sheet by default 18:10
[ptc] itz: yes, it looks that way 18:11
itz does the test pass?
[ptc] gist.github.com/paultcochrane/e6e5...e59852741a 18:13
itz: the tests pass afaik (well, they pass when installing via panda)
itz: btw, what is vim-let? 18:15
itz ask andreoss :)
andreoss you could pass aditional arguments to vim shell with it
like set numbers 18:16
[ptc] ah, ok
must dash for a while # dindins
PerlJam m: use Test; is-approx(1,1); # why does it choose the 3rd variation of is-approx?
camelia rakudo-moar 8532c0: OUTPUT«Absolute tolerance must be a positive number greater than zero␤ in sub is-approx at lib/Test.pm:254␤ in block <unit> at /tmp/LiirJJH2vd:1␤␤»
[ptc] PerlJam: several links are mentioned in github.com/rakudo/rakudo/pull/366 concerning the implementation. Hopefully one will help you answer that question! 18:20
18:20 kjs_ joined
[ptc] PerlJam: improvements to the algorithm are most certainly welcome! And more discussion as to its correctness etc 18:21
PerlJam: I wish more discussion had happened before the patch got merged...
[ptc] really afk now
PerlJam [ptc]: Well, that questions was really about multi dispatch. But it was triggered by a problem I have with the current implementation that I hope to soon fix (just as soon as I figure out what the right thing is :) 18:22
but [ptc]++ for is-approx
tony-o http::server::async has a branch converted to conform to htt::server but async inet is even more broken with the cross thread read stuff 18:25
18:27 [Sno] left, ][Sno][ joined 18:28 ][Sno][ is now known as [Sno], andreoss left 18:32 _mg_ joined 18:36 kjs_ left 18:38 kurahaupo joined, kurahaupo left 18:39 bin_005 joined
dalek osystem: 3a5ea41 | PerlJam++ | META.list:
Add Math::Trig
18:40
18:40 kjs_ joined
dalek rl6-most-wanted: 53e10cb | PerlJam++ | most-wanted/modules.md:
Add WIP for Git::Wrapper
18:44
18:44 hudo joined 18:47 lucasb left, amurf joined 18:49 kjs_ left 18:51 molaf_ joined, amurf left 18:54 molaf left 19:02 brrt joined 19:05 vendethiel joined 19:06 vendethiel- left 19:08 vendethiel- joined
[Coke] (pdl discussion) I would recommend against perl6-language unless you're ready to deal with a million bikeshedders. 19:09
19:09 vendethiel left
nwc10 we have a million interested people? \o/ 19:10
Or does it just feel like that, sometimes? :-/
[Coke] FSVO interested. 19:12
masak the bikeshedding in p6l is less bad than it used to be. 19:13
and it tends to be OK if you set your expectations accordingly.
sometimes I have the feeling that p6l is filled predominantly with people who are unaware that Perl 6 has an implementation... :) 19:14
moritz aye
also, it's mostly inverse bikeshedding
PerlJam
.oO( p6l tends to be better if you have your mail filter set to drop mail from known trolls )
19:15
masak what's inverse bikeshedding?
moritz questions about how to paint the bike shed are answered with nuclear power plant construction musings
masak hehe
PerlJam: I never thought of anyone on that list as a troll, to be honest. but there sure are people who are a little too helpful. 19:16
PerlJam honestly I don't have an automated mailfilter, but every once in a while I have a mental filter that tends to identify certain people as unproductive and I just don't read anything they have to say after the first couple of messages. 19:17
masak *nod*
PerlJam (if my mentail filter recognizes your name right away, that's either really good (you're masak, moritz, pmichaud, etc.) or really bad :-) 19:18
itz I think things are worse in the perl5 community tbh .. I was sent a Reini Urban blog post I wish I never saw :(
moritz is honored to be on PerlJam's "really good" list 19:20
jnthn wonders if he's on the "really bad" one for his puns... :D 19:21
nwc10 itz: it's hard to see how to avoid the more general problem - you want to have an open access project. You *need* to have an open access project, else you can't recruit new people
masak jnthn: no, it's just that you haven't made any significant contributions to Perl 6 :P:P:P 19:22
PerlJam heh
jnthn :P
nwc10 itz: and you have a lot more people who care about *something*, but don't have the ability to see the big picture. Or don't have the time to learn to see it. Or don't have the skills to implement what they want doing. Or a whole bunch of things like that.
19:22 bin_005 left
nwc10 and there's a lot of ways to break something 19:22
and very few paths through that don't break much 19:23
19:23 brrt left
itz nwc10: yes people believe because they have personal investment of time in things that those things are Good 19:23
PerlJam ponders responding to nwc10 in a worst-of-p5p-parody sort of way, but totally lacks the energy. 19:24
masak is currently fixing 007, which was broken in a few places by Z semantics shifting 19:26
19:29 kjs_ joined
nwc10 also, I hope this doesn't upset too many folks, but I think it's an unavoidable symptom of success 19:29
if people aren't upset about changes/things not changing in the future
then that means that Perl 6 hasn't made "it"
(Although having the grammar lexically scoped is going to avoid a whole chunk of problems of legacy languages) 19:30
RabidGravy aye, I warrant this channel will be unusable a year hence
PerlJam RabidGravy: I'm not sure if you're an optimist or a pessimist 19:31
itz RabidGravy is a glass full sort of guy ;) 19:32
19:32 domidumont left
masak this is not a purely theoretical concern. I remember what happened to this channel around the first R* release. 19:32
a foretaste of a bigger community, with more chaos in it.
19:33 hudo left, bin_005 joined
RabidGravy the touchstone is when you can't go for a piss before the scroll buffer runs out ;-) 19:33
PerlJam RabidGravy: that's why we have irclogs! :) 19:34
19:34 akakcolin left
masak I will miss the thoughtfulness and slowness of the #perl6, but it's a thing I'm willing to sacrifice... 19:35
I do hope we manage to scale kindness pretty far, though.
jnthn grmbl 19:36
At some point, "make spectest" has gotten busted on Windows.
nwc10 Star is now an asterisk in a 4 letter word? :-( 19:38
jnthn It's not that it fails any tests. It that it doesn't try to run any tests. 19:40
Use of uninitialized value in split at t/harness line 155.
Files=0, Tests=0, 0 wallclock secs ( 0.00 usr + 0.00 sys = 0.00 CPU)
Result: NOTESTS
RabidGravy boo!
jnthn I'll grant it's faster this way, but I don't think it's going to enhance the quality of my commits...
itz is it something clever to do with lists? 19:41
19:41 molaf_ left 19:44 _mg_ left, lolisa joined
jnthn omfg 19:45
19:46 kjs_ left
masak guesses GLR somehow 19:46
FROGGS O.o
no
that's about P5 only I guess
masak oh, because t/harness is p5.
jnthn No, this is a Perl 5 script
bahaha 19:47
Epic fail.
FROGGS jnthn: do tell
jnthn my $cmd = join ' ', $^X, 't/spec/fudgeall',
@pass_through_options, $impl, @_;
return split ' ', `$cmd`;
FROGGS ohh, space in path?
nwc10 space in path?
jnthn `$cmd` comes back undef
Nope
Hint: if I comment out any spectest in spectest.data, it's OK
itz line endings?
FROGGS ETOOMANYTHINGS? 19:48
jnthn Bigger hint: print length($cmd) . "\n";
FROGGS ups
nwc10 oh. it's recently got a bit too big?
jnthn With all the spectests it's 32776. When I delete one, it's 32734...and works fine :P
Apparently, something doesn't like more than 32KB of command line args :P
nwc10 we have to many tests. obviously. delete some!
itz cmd.exe? 19:49
FROGGS hmmm
RLE to the rescue! /o/
jnthn oh you're kidding, it's a Windows API level limitation 19:50
[Coke] (embiggening #perl6) I suspect we will have to split into dev/help channels.
jnthn It uses the UNICODE_STRING struct 19:51
typedef struct _UNICODE_STRING {
USHORT Length;
U r short indeed...
19:51 brrt joined
FROGGS :D 19:51
hehe
[Coke] we prefudge everything at once, that doesn't need to happen.
jnthn "32KB should be enough for everyone"
itz [ptc]: it does work! see pl6anet.org/examples//cookbook/09di...-file.html
19:51 telex left 19:52 telex joined
FROGGS jnthn: we could batch it at around line 78 19:52
jnthn FROGGS: Yeah, or in "sub fudge" itself 19:53
FROGGS aye
RabidGravy I don't suppose windows has am xargs? 20:00
;-)
PerlJam RabidGravy: perlpowertools.com/ :) 20:02
dalek kudo/nom: df0c3a3 | jnthn++ | t/harness:
Batch files before sending to fudgeall.

We reached enough spectests to exceed the maximum command line length possible on Windows when calling fudgeall.
20:04
jnthn apologizes for his probably non-idiomatic Perl 5 :)
nwc10 jnthn: the solution to *that* is for someone (else) to re-write the test script in Perl 6. 20:05
jnthn :)
yeah, then I'd have just used .rotor :)
FROGGS isnt illegal to use undeclared vars in Java? github.com/perl6/nqp/blob/master/s...e.java#L34 20:07
vendethiel- FROGGS: github.com/perl6/nqp/blob/master/s...e.java#L20 :)
FROGGS ohh, eeks 20:08
(twiggels ftw)
20:11 darutoko left
[ptc] PerlJam: really I just copied the algorithm from the Python people :-) 20:16
itz: wow, that looks good! 20:17
masak if you have an array of arrays of primitives, and you want to clone it all, how would you write the clone logic?
this is what I came up with:
[ptc] itz: now to work out why it doesn't work on examples.p6.org ...
masak m: my @a = [1], [2]; my @b = @a.map({ [.list] }); say @a[0] === @b[0]
camelia rakudo-moar 8532c0: OUTPUT«False␤»
itz [ptc]: I have made some minor tweaks as a result of testing on debian vim 20:18
raydiak m: my @a = [1], [2]; my @b = @a».clone; say @a[0] === @b[0] 20:21
camelia rakudo-moar df0c3a: OUTPUT«False␤»
raydiak m: my @a = 1; @a.clone # hrm 20:24
camelia rakudo-moar df0c3a: OUTPUT«Unhandled exception: This type does not support elems␤ at src/gen/m-CORE.setting:10953 (/home/camelia/rakudo-inst-2/share/perl6/runtime/CORE.setting.moarvm:REIFY:34)␤ from src/gen/m-CORE.setting:11471 (/home/camelia/rakudo-inst-2/share/perl6/runtim…»
RabidGravy oops 20:29
20:33 xaduha joined 20:36 amurf joined
[ptc] itz: is the new version already available via panda? 20:38
itz [ptc]: yes I bumped the version to 0.3 20:39
also the test actually tests for the presence of the colour syntax where it didn't really before
20:41 amurf left
[ptc] hrm, now the test fails on hack... 20:43
20:43 xaduha left
[ptc] failed test 4 20:43
unfortunately no further output than that (prove is sometimes not verbose enough....) 20:44
dalek ast: 3192739 | usev6++ | S06-advanced/lexical-subs.t:
Change two 'dies-ok' to 'throws-like'
ast: 0867f11 | usev6++ | S06-advanced/lexical-subs.t:
Add tests for RT #109322
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=109322
20:45 nige joined
nige masak++ like your idea of scaling kindness 20:47
20:48 [TuxCM] joined 20:49 [Tux] left, ribasushi left 20:50 uncleyear left
masak nige: I think that's the only thing one can affect as a community grows. seeding with good examples and good influence. it won't always help, but it will make a mark. 20:50
20:50 uncleyear joined 20:51 ribasushi joined
nige yes - indeed - as a long time lurker from O(fun) days I hope the vibe continues - although I can see more success will bring challenges 20:52
20:52 [Tux] joined 20:53 meLon left 20:54 kjs_ joined
nige the artistic licence and camelia will also play a big part 20:54
brrt join us now, share the software :-) 20:55
20:59 kjs_ left
itz \o/ join us now and freeee the software # </rms> 20:59
21:03 bin_005 left 21:04 dolmen joined 21:05 pyrimidi_ is now known as pyrimidine, muraiki left 21:06 skids left
nige camelia is doing a good job of troll detector and defender 21:06
she may be called upon more - post christmas
itz tries to unhear the stallman singing voice 21:07
nige yes - me too - never drunk that KoolAid
I like the theme in the artistic licence - about code as expression 21:09
and perl6 is hyper expressive - so I see this as part of the brand 21:10
vendethiel- had perl6 not believe perl6.org was the real website at first. most probably the colors :D 21:11
21:13 Sqirrel left 21:14 bin_005 joined
dalek pan style="color: #395be5">perl6-examples: 9dc9344 | paultcochrane++ | t/000-check-dependencies.t:
Check that dependencies for tests are installed
21:15
21:18 Sqirrel joined
[Coke] needs to work on a pull request to push down non-rakudo compilers at this point. 21:18
(on the perl6.org site)
21:19 muethos left, Vlavv left 21:20 grondilu left
dalek ast: 046ad7e | usev6++ | S32-array/pop.t:
Add test for RT #111720
21:20
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=111720
21:20 grondilu joined 21:23 Vlavv joined
masak 'night, #perl6 21:33
bartolin o/ 21:34
21:38 Sqirrel left 21:41 itz left, Sqirrel joined 21:45 Sqirrel left 21:50 Sqirrel joined 21:53 dolmen left 21:55 Sqirrel left 21:59 Sqirrel joined 22:04 brrt left 22:06 dolmen joined, bonsaikitten left 22:08 mohij left 22:09 xiaomiao joined, kjs_ joined 22:13 nige left 22:15 larion left, dolmen left 22:22 lichtkind left 22:23 bin_005 left 22:32 kaare_ left 22:40 skids joined 22:46 meisl left 22:49 sivoais left 22:50 ruoso left, sivoais joined 22:52 kjs_ left 22:55 grondilu left, ribasushi left, spider-mario left, Akagi201 left, ponbiki left, pecastro left, moritz left, Possum left, petercommand left, lestrrat left, agentzh left, bartolin left, cibs left 22:56 kjs_ joined, lolisa left
timotimo hoelzro: you can remove yourself from the "future planned releases" list 22:56
22:57 ruoso joined 23:05 kjs_ left 23:06 grondilu joined, ribasushi joined, Akagi201 joined, ponbiki joined, pecastro joined, moritz joined, Possum joined, petercommand joined, lestrrat joined, agentzh joined, cibs joined, bartolin joined 23:07 RabidGravy left, Akagi201 left 23:08 Akagi201 joined 23:10 amurf joined
[ptc] timotimo: are you going to the Swiss Perl Workshop? 23:12
23:15 grondilu left 23:16 grondilu joined
[ptc] timotimo: anyway, if you are going, I was wondering if you'd be interested (or have the time) to work on the issues Coverity shows up for MoarVM 23:22
23:26 Akagi201 left 23:44 bjz left 23:45 kst] joined 23:47 xinming left, kst joined