»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
Mouq rnp: "OH HAI"\ \ .say 00:02
camelia rakudo 38a0f9: OUTPUT«===SORRY!=== Error while compiling /tmp/64g7cmsuNJ␤Confused␤at /tmp/64g7cmsuNJ:1␤------> "OH HAI"\ ⏏\ .say␤ expecting any of:␤ unspace␤ horizontal whitespace␤ postfix␤»
..pugs: OUTPUT«*** ␤ Unexpected "\\"␤ expecting "."␤ at /tmp/O8zRvlBRZI line 1, column 11␤»
..niecza v24-88-g1f87209: OUTPUT«===SORRY!===␤␤Method call found where infix expected (change whitespace?) at /tmp/vBWA6fJ56r line 1:␤------> "OH HAI"\ \ ⏏.say␤␤Parse failed␤␤»
Mouq std: "OH HAI"\ \ .say
camelia std 06636e4: OUTPUT«===SORRY!===␤Method call found where infix expected (change whitespace?) at /tmp/4FxeumuiTx line 1:␤------> "OH HAI"\ \ ⏏.say␤ expecting any of:␤ horizontal whitespace␤ infix or meta-infix␤ infixed function␤ postcircumfix␤
..post…
Mouq Making that ( RT #117465 ) work in rakudo was pretty easy 00:03
Just change <.unsp>? to <.unsp>* and <.unsp> to <.unsp>+
...(in src/Perl6/Grammar.pm). However, it isn't in std... 00:04
00:05 dayangkun joined
Mouq (And of course, I'm sure there's a way to do it, just making <.unsp> be eqv to what is now <.unsp>*) 00:06
(err, <.unsp>+)
00:08 colomon left
Mouq running& 00:09
00:28 dayangkun left 00:29 dayangkun joined, abnorman left, abnorman joined
dalek d/multiunsp: d699c5f | Mouq++ | STD.pm6:
Allow multiple unspaces in sequence
 00:34
00:38 _jaldhar left 00:41 sftp joined 00:47 _jaldhar joined
Mouq You see, #perl6, I have my school starts tomorrow. So clearly the best way to prepare is to play around with Perl6. The first day (highschool) is always a joke anyway 00:48
s/I have// perhaps I miss my English class a bit 00:49
TimToady we'd probably say "my school is starting tomorrow" 00:50
we tend to avoid "present tense" in favor of "present progressive"
Mouq Well, I'm a naïve speaker—*ahem* native 00:51
diakopter *we are tending
TimToady I am standing corrected :) 00:52
actually it'll be interesting to see if the trend toward progressive reverses, what with everyone becoming a writer in their texting 00:54
and "starts" is easier to type than "is starting" :) 00:55
So, am I predicting this? Or do I predict this? :) 00:56
geekosaur you might coould be predicting it... 00:57
Mouq You previously will predict it.
jercos You will have predicted that when it does or does not occur.
Your prediction transitioned from unmade to pre-resolution? 00:58
Mouq TimToady: I made PR #5 on STD because I'm cautious to change something not explicitly specced, and it needs review: github.com/perl6/std/pull/5
01:01 drbean joined 01:02 dayangkun left
TimToady Mouq: yes, still thinking about that 01:08
Mouq Okay~, I just wasn't sure if you had seen it... The :dba is inside the []+. I'm not sure if that works... 01:10
r: say "\c[LATIN CAPITAL LETTER A, COMBINING DOT ABOVE]".ord; 01:16
camelia rakudo 38a0f9: OUTPUT«65␤»
Mouq r: say "\c[LATIN CAPITAL LETTER A WITH DOT ABOVE]".ord;
camelia rakudo 38a0f9: OUTPUT«550␤»
TimToady that will fix itself when we switch to NFG mode
since it will default to grapheme mode 01:17
our current string semantics based on codepoints will probably be renamed to Uni/NFD/NFC and such 01:18
Mouq Cool 01:20
TimToady mostly people don't want to care about how many codepoints are in their graphemes
re <.unsp>* I can see some potential use cases; what I'll have more difficulty seeing is whether it causes surprising results of its own; that'll take playing around with it a bit 01:23
for instance, we currently allow 42\i to be an unspace without space, as it were; what happens with 42\ \i then? 01:24
or if we allow that, what about if you want to start the next term with a \ for some reason (a capture?)
so it's not entirely a simple decision 01:25
01:25 benabik left
Mouq Hm 01:30
Okay. Interesting 01:32
TimToady std: 42\i
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
TimToady std: 42\ i
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
TimToady std: 42\ \i 01:33
camelia std 06636e4: OUTPUT«===SORRY!===␤Two terms in a row at /tmp/2UI2xm_ZRj line 1:␤------> 42\ ⏏\i␤ expecting any of:␤ feed_separator␤ horizontal whitespace␤ infix or meta-infix␤ infixed function␤ postcircumfix␤ postfix␤
..postfix_prefix_meta_operator␤ …
TimToady but that has been proposed before, to have a backslash at the end as well, mostly to put at the beginning of a line to say the previous line ended with a \
up till now, it has just seemed like more clutter
the use case for \ \ is more about allowing \ on a blank line, if for some reason you're writing a series of \ lines 01:34
and want a blank line in the middle for some reason
01:35 colomon joined
TimToady but that's different from the stuff\ \stuff problem 01:35
Mouq Well, that's just a matter of changing the quantifier for <.unsp> at that particular place 01:36
TimToady nod, I just always tend to view one change in relation to all the nearby things that might get nudged out of place :) 01:40
Mouq Probably a good tendency 01:42
r: .++()
camelia rakudo 38a0f9: OUTPUT«===SORRY!===␤No such method 'unshift' for invocant of type 'NQPMu'␤»
TimToady afk for a couple hours & 01:49
Mouq std: .++() 01:54
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
Mouq std: .i # ? 01:57
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
Mouq r: .i
camelia rakudo 38a0f9: OUTPUT«No such method 'i' for invocant of type 'Nil'␤ in block at /tmp/uVnqXORnWE:1␤␤»
Mouq r: .i() 01:59
camelia rakudo 38a0f9: OUTPUT«No such method 'i' for invocant of type 'Nil'␤ in block at /tmp/va6R0KAYlz:1␤␤»
Mouq r: sub postfix:<¡> { $^a }; "thing"¡; "thing".¡; 02:01
camelia rakudo 38a0f9: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&postfix:<.>' called (lines 1, 1)␤»
Mouq :? 02:04
r: my @a = ^10; @a.».++ # strange error
camelia rakudo 38a0f9: OUTPUT«No such method 'postcircumfix:<( )>' for invocant of type 'Int'␤ in sub METAOP_HYPER_CALL at src/gen/CORE.setting:15607␤ in block at /tmp/9T1RPAHm0a:1␤␤»
Mouq r: .:<++> 02:05
camelia rakudo 38a0f9: OUTPUT«===SORRY!=== Error while compiling /tmp/_Dm6r538tu␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/_Dm6r538tu:1␤------> .:⏏<++>␤»
Mouq r: $_.:<++> 02:06
camelia rakudo 38a0f9: OUTPUT«===SORRY!=== Error while compiling /tmp/BuInZymics␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/BuInZymics:1␤------> $_.:⏏<++>␤»
Mouq std: .:<++> 02:11
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
02:18 benabik joined 02:53 BenGoldberg left 02:57 colomon left 03:18 preflex_ joined, ChanServ sets mode: +v preflex_ 03:19 preflex left, preflex_ is now known as preflex
Mouq std: .:<nonpre> 03:42
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
Mouq std: $_.:<nonpre>
camelia std 06636e4: OUTPUT«ok 00:00 42m␤»
Mouq std: $_.-+ 03:50
camelia std 06636e4: OUTPUT«===SORRY!===␤Confused at /tmp/5ZddGzXE_j line 1:␤------> $_.⏏-+␤ expecting any of:␤ dotty method or postfix␤ postcircumfix␤ postfix_prefix_meta_operator␤ postop␤Parse failed␤FAILED 00:00 42m␤»
03:52 iravan joined
Mouq 03:59
Oop. Yup. /me
04:00 iravan left 04:29 xilo left
Mouq Really close to fixing rakudo's «.++»/«.:<++>» problem, but I must to bed! ciao #perl6 04:39
04:39 hummeleBop joined
TimToady night 04:40
04:44 Mouq left 05:01 berekuk joined 05:08 berekuk left 05:09 berekuk joined, berekuk left 05:12 SamuraiJack joined 05:26 dayangkun joined 05:38 fhelmberger joined 05:43 dayangkun left 06:07 daxim joined 06:08 benabik left 06:25 kaleem joined 06:34 FROGGS joined 06:36 kaleem left 06:37 dayangkun joined, dayangkun is now known as dayangkun_715, dayangkun_715 left
FROGGS morning 06:38
06:38 dayangkun joined, dayangkun left, kaleem joined 06:39 kaleem left, kaleem joined 06:49 dayangkun joined 06:51 fhelmberger_ joined, dayangkun left
moritz \o 06:52
06:53 fh joined, fhelmberger_ left 06:55 fhelmberger left 07:19 domidumont joined 07:23 domidumont left 07:24 domidumont joined 07:31 salv0 joined 07:37 sqirrel joined 07:39 skulls left
FROGGS .tell Mouq What if <unsp> unspaces more than one char? It could unspace \s+, wouldn't that be better? 07:46
yoleaux FROGGS: I'll pass your message to Mouq.
FROGGS .tell TimToady What if <unsp> unspaces more than one char? It could unspace \s+, wouldn't that be better?
yoleaux FROGGS: I'll pass your message to TimToady.
07:49 abnorman left 08:05 dalek left 08:06 dalek joined, ChanServ sets mode: +v dalek 08:07 benabik joined
nwc10 yoleaux: .tell [Coke] that FRA is about 15 minutes from central Frankfurt by train, and HNH is 124km away by road. 08:08
on, that wasn't the magic
.tell [Coke] that FRA is about 15 minutes from central Frankfurt by train, and HNH is 124km away by road.
yoleaux nwc10: I'll pass your message to [Coke].
nwc10 oh, sweet 08:09
Hahn is nearer to Perl than to Frankfurt. 08:10
moritz t.co/RqfXT3A2XG # one of the YAPC::EU talks by masak++
masak good o/, #perl6 08:11
FROGGS ohh, awesome! 08:12
hi all
timotimo good day 08:15
08:18 dmol joined, denis_boyun joined
timotimo i think i'm going to merge the nqp-optimizer today. don't forget to run Configure*.pl before re-building 08:21
FROGGS k 08:24
timotimo currently running all the tests 08:25
oh woops, there seems to be a debug output left somewhere 08:28
JimmyZ timotimo: Do you have a benchmark to show how faster with nqp-optimizer? 08:30
FROGGS good that it isnt at the right, you might havn't spotted it
08:31 SamuraiJack_ joined
timotimo JimmyZ: yes, hold on 08:33
t.h8.lv/p6bench/nqp-optimization.html
08:34 dmol left
timotimo 215ac4c is the optimized one 08:34
fortunately it has the same color in all graphs
pro tip: you can click on the checkouts in the annotaion on the right to show/hide the lines
08:35 SamuraiJack left
timotimo the optimization only does something for integer arithmetics, so it doesn't do anything to most benchmarks 08:35
but a rakudo translated with the optimized&optimizing nqp binary will run the spectest suite in about 1.5% less cpu time 08:36
08:37 Kiralexis left
JimmyZ oh 08:40
I care about compiling time, it's a pain :P
timotimo the optimizer is blazing fast :) 08:41
didn't measure it. should i?
JimmyZ I mean the rakudo cpmpiling time :P
timotimo yes, i know :)
JimmyZ not the optimizer itself
timotimo ah! 08:42
of course
last time i checked there was no noticable difference
JimmyZ please measure it, that's a good benchmark
a real life benchmark 08:43
timotimo why are there so many spectest failures? =o
oh, it's rakudo-jvm
GlitchMr std: "OH HAI"\. say 08:53
camelia std 06636e4: OUTPUT«===SORRY!===␤Confused at /tmp/ZNF_x0ZWGC line 1:␤------> "OH HAI"\.⏏ say␤ expecting any of:␤ dotty method or postfix␤ postcircumfix␤ postfix_prefix_meta_operator␤ postop␤Parse failed␤FAILED 00:00 42m␤»… 08:54
GlitchMr std: "OH HAI"\. say
camelia std 06636e4: OUTPUT«===SORRY!===␤Confused at /tmp/fGtCvQWNPr line 1:␤------> "OH HAI"\.⏏ say␤ expecting any of:␤ dotty method or postfix␤ postcircumfix␤ postfix_prefix_meta_operator␤ postop␤Parse failed␤FAILED 00:00 42m␤»
timotimo that's not how you do it :)
GlitchMr oh, right
std: "OH HAI".\ say
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
GlitchMr std: "OH HAI".\ say
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
GlitchMr std: "OH HAI".\ \ say
camelia std 06636e4: OUTPUT«===SORRY!===␤Confused at /tmp/YzVs926gmK line 1:␤------> "OH HAI".\ ⏏\ say␤ expecting any of:␤ dotty method or postfix␤ horizontal whitespace␤ postcircumfix␤ postfix_prefix_meta_operator␤ postop␤
..unspace␤Parse…
GlitchMr rakudo: "OH HAI".\ \ say 08:56
camelia rakudo 38a0f9: OUTPUT«===SORRY!=== Error while compiling /tmp/UrcNYQI06l␤Confused␤at /tmp/UrcNYQI06l:1␤------> "OH HAI".\ ⏏\ say␤ expecting any of:␤ unspace␤ horizontal whitespace␤»
GlitchMr niecza: "OH HAI".\ \ say
camelia niecza v24-88-g1f87209: OUTPUT«===SORRY!===␤␤Confused at /tmp/M6slbo8V0S line 1:␤------> "OH HAI"⏏.\ \ say␤␤Parse failed␤␤»
08:56 donaldh joined 09:00 odoacre left 09:03 amkrankruleuen joined, amkrankruleuen left, amkrankruleuen joined 09:12 denisboyun joined 09:14 denis_boyun left
timotimo measures core setting compile times 09:15
09:17 grondilu_ joined 09:21 grondilu left
hoelzro morning #perl6 09:22
timotimo sup rob :)
the new vim release features a faster regexp engine. i wonder if the perl6 syntax rules can make use of that?
hoelzro timotimo: for the perl5 files, the new engine really messed things up 09:25
but that was a while ago, so it might work better now
timotimo oh?
hoelzro any and all contributions to the perl6 file would be most welcome =)
I have so much I want to do for it 09:26
but such little time =/
timotimo the only mess-up i have with vim at the moment is that sometimes regions of the window get damaged and not re-drawn, so i get large areas of black or duplicated or shifted lines :(
moritz throws a round tuit to hoelzro
timotimo but only in gvim
and only on my desktop
hoelzro catches it, and gives it a small bite to verify its authenticity
09:27 berekuk joined 09:32 dmol joined 09:36 wtw left
timotimo compilation times are pretty noisy, JimmyZ. i'm getting 4 measurements with and without nqp-optimize each, but only for the CORE.setting 09:37
09:39 dmol left, berekuk left 09:40 dmol joined
timotimo what's the shortest way between "a list of floating point numbers" to "average and standard deviation"? 09:41
maybe even a box plot? :)
tadzik hmm, I thought we have a statistics module
must've got lost
moritz timotimo: iterate over all values, keep a total of the values and the squares 09:42
and then you can use those two to calculate mean and standard deviation (first and second momentum, really)
timotimo better answer: wolframalpha ;)
09:42 berekuk joined
moritz oh 09:42
tadzik :P 09:43
moritz I thought you wanted to write Perl 6 code
locally I use octave (open source matlab clone) for this stuff
timotimo hehe
moritz v = [1.3, 1, 0.7] 09:44
hoelzro so...what's the schedule for this upcoming workshop?
moritz mean(v)
std(v)
hoelzro I'm looking into flights and wondering when I should arrive at/leave Frankfurt
09:44 fh left
tadzik hoelzro: gist.github.com/lizmat/6177720 09:44
oops
it's gone :)
hoelzro =(
timotimo r: my @samples = 171.305, 165.521, 165.521, 165.521; my ($sum, $sqsum) = ([+] @samples), ([+] (@samples >>**>> 2)); say $sum; say $sqsum; say "and now?"
camelia rakudo 38a0f9: OUTPUT«667.868␤111537.007348␤and now?␤»
moritz the mail to p6c has the details, but I'm not sure if there's a schedule 09:45
09:45 berekuk left
timotimo i have to sum up the difference between the average and the squares at each point, right? 09:45
r: my @samples = 171.305, 165.521, 165.521, 165.521; my $mean = ([+] @samples) / +@samples; my $sqdiffsum = ([+] ((@samples X- $mean) >>**>> 2)); say $mean; say $sqdiffsum; say "and now?" 09:46
camelia rakudo 38a0f9: OUTPUT«166.967␤25.090992␤and now?␤»
moritz en.wikipedia.org/wiki/Standard_devi...on_methods
uhm, that looks more complicated than necessary 09:47
09:49 Kiralexis joined
moritz timotimo: the standard deviation is just sqrt($sqdiffsum / @samples) 09:49
timotimo ah
09:50 fhelmberger joined
timotimo the results so far have been inconclusive (i had a b0rked data point in my first measurement) 09:51
grondilu_ there is an entry about std dev in RC: rosettacode.org/wiki/Standard_Deviation#Perl_6 09:52
09:52 nyuszika7h left 09:53 nyuszika7h joined 09:54 colomon joined
JimmyZ timotimo: so you got a result? how much faster? 09:55
timotimo hold on just a few more minutes
it won't be much
perhaps 1% 09:56
masak .oO( calculating the arithmetic average of something is pretty mean )
timotimo masak: do you have an answer for "gather-based processes are hard to resume without manual effort"?
resume as in: the ircbot has to restart and keep its state 09:57
09:57 iSlug joined
masak timotimo: no idea. I sense I should backlog to know better what you're talking about. 09:58
timotimo i didn't really say much more about that before 09:59
09:59 grondilu_ is now known as grondilu, grondilu is now known as Guest93334
timotimo JimmyZ: with 4 data points each i get: 10:01
optimized: mean 167, std: 1.676 10:02
JimmyZ oh
timotimo unoptimiz: mean 168.9, std: 0.941
so a little bit more than 1% for compiling the CORE.setting
faster, i mean. 10:03
FROGGS 100 of these patches and we can compile in *no time*! whoohoo! *g*
timotimo :D
i try my best, but it's so hard :(
Guest93334 tries to think of a smart way to write a momentum function
FROGGS timotimo: I know what you mean :o) 10:04
timotimo env.parrot.rakudo is known to fail, right?
FROGGS timotimo: yes
Guest93334 r: sub momentum(@a) { gather take ([\*] 1, $_ xx *).item for @a } 10:05
camelia rakudo 38a0f9: ( no output )
timotimo that works? •_•
10:05 nyuszika7h left 10:06 nyuszika7h joined 10:07 Guest93334 is now known as grondilu, grondilu is now known as Guest88156 10:08 Guest88156 is now known as grondilu, grondilu left, grondilu joined 10:10 cognominal__ is now known as cognominal
grondilu please ignore what I wrote as Guest93334 10:10
I still think there has to be a smart way to write a momentum function, but it's not as easy as I thought. 10:11
10:23 sergot joined, synopsebot joined, fhelmberger left 10:24 fhelmberger joined
timotimo measurements for rakudo.jvm optimized is mean 72.85, std 0.938 10:24
grondilu r: sub momentum(@a) { my @momentum := (1 xx @a).item, @a.item, {(@^x Z* @a).item} ... *; map { [+] @momentum[$_].list }, 0 .. * }; say momentum(^10)[^4] 10:25
camelia rakudo 38a0f9: OUTPUT«10 45 285 2025␤»
grondilu r: say [+] ^10 »**»3 10:27
camelia rakudo 38a0f9: OUTPUT«0␤»
grondilu r: say ^10 »**»3
camelia rakudo 38a0f9: OUTPUT«0..^1␤»
grondilu r: say [^10] »**»3 10:28
camelia rakudo 38a0f9: OUTPUT«0 1 8 27 64 125 216 343 512 729␤»
grondilu r: say [+] [^10] »**»3
camelia rakudo 38a0f9: OUTPUT«2025␤»
colomon n: say ^10 »**»3 10:29
camelia niecza v24-88-g1f87209: OUTPUT«0..^1000␤»
colomon that at least sort of makes sense....
grondilu r: sub momentum(@a) { my @momentum := (1 xx @a).item, {(@^x Z* @a).item} ... *; map { [+] @momentum[$_].list }, 0 .. * }; say momentum(^10)[^5] 10:30
camelia rakudo 38a0f9: OUTPUT«10 45 285 2025 15333␤»
grondilu r: sub momentum(@a) { my @momentum := (1 xx @a).item, {(@^x Z* @a).item} ... *; map { [+] @momentum[$_].list }, 0 .. * }; sub stddev { sqrt( (.[0]*.[2] - .[1]**2)/.[0] ) given momentum(@^a) }; say stddev ^10 10:32
camelia rakudo 38a0f9: OUTPUT«9.08295106229247␤»
10:36 donaldh left, donaldh joined
Ulti timotimo I'd do a en.wikipedia.org/wiki/Kolmogorov–Smirnov_test on that to get a pvalue ;P thats what all the TLDRs want cold hard science >:3 10:37
timotimo i don't even ... 10:40
10:41 locsmif joined
timotimo noooo, the unoptimized version is not slow enough! must start more background processes ... 10:41
Ulti jvm rakudo I think the main thing anyone is going to complain about time wise is the ~10s startup time, at least that's what it was for me quite a while ago a lot has happened since then 10:43
timotimo =time ./perl6 -e 'say "hello world"' 5,57s user 0,15s system 162% cpu 3,515 total
moritz does what he should have done all along: two separate directories for the two rakudos 10:44
timotimo i just want to make sure i don't accidentally make it slower by adding my optimization
grondilu r: my @a = ^10; my $m = @a R/ [+] @a; say sqrt ( @a R/ [+] map (* - $m)**2, @a ); 10:45
camelia rakudo 38a0f9: OUTPUT«2.87228132326901␤»
grondilu r: sub momentum(@a) { my @momentum := (1 xx @a).item, {(@^x Z* @a).item} ... *; map { [+] @momentum[$_].list }, 0 .. * }; sub stddev { sqrt(.[0]*.[2] - .[1]**2)/.[0] given momentum(@^a) }; say stddev ^10
camelia rakudo 38a0f9: OUTPUT«2.87228132326901␤»
grondilu here you go
or shorter: 10:47
r: sub momentum(@a) { my @momentum := [1 xx @a], {[@^x Z* @a]} ... *; map { [+] @momentum[$_].list }, 0 .. * }; sub stddev { sqrt(.[0]*.[2] - .[1]**2)/.[0] given momentum(@^a) }; say stddev ^10
camelia rakudo 38a0f9: OUTPUT«2.87228132326901␤»
10:48 fhelmberger left
Ulti timotimo: 162% cpu ? there is some threading happening? 10:48
10:50 fhelmberger joined
timotimo yes 10:50
10:50 berekuk joined
timotimo the jvm has a concurrent GC 10:50
Ulti ahh right 10:51
moritz has a rakudo-jvm startup time of 3.7s
10:52 berekuk left
Ulti I'll have to get myself a fresh copy then 10:52
timotimo perl6 -e 'say [+] lines', so nice and short to have a little terminal to paste stuff into and get a sum out 10:53
grondilu wonders how could we have intermediates sums in "real time" 10:57
moritz with [\+] of course
grondilu .say for [\+] lines
masak flight &
timotimo hm, lines is eager, isn't it? :(
grondilu but it only shows up at the end
moritz timotimo: it shouldn't be
I think that for-loops are too eager
timotimo yes, but it is ;(
that may be it, yeah
using map instead should help, right?
Ulti perl -E 'say $sum += $_ for <>' isn't too long either, but a lot harder to read 10:58
timotimo it doesn't.
hm. how do i read a single line from stdin? 10:59
r: say $?STDIN.^methods
camelia rakudo 38a0f9: OUTPUT«===SORRY!=== Error while compiling /tmp/du0EIHtJ1y␤Variable '$?STDIN' is not declared␤at /tmp/du0EIHtJ1y:1␤------> say $?STDIN.^methods⏏<EOL>␤ expecting any of:␤ method arguments␤»
timotimo r: say $*STDIN.^methods
camelia rakudo 38a0f9: OUTPUT«new BUILD defined Int Num Numeric sink postcircumfix:<{ }> Bool Str gist <anon>␤»
timotimo ... oh?
moritz timotimo: $*IN.get
timotimo ah
thanks :)
perl6 -e 'my $s; map { when /\d+/ { $s += $_; }; default { say $s; $s = 0 } }, ({ $*IN.get } ... *)' # improvements welcome 11:00
11:02 berekuk joined
grondilu don't we have a -n option in perl6, as in perl5? 11:02
timotimo dunno
grondilu tries
moritz I'm pretty sure we have.
timotimo if all that does is wrap for lines() around it, it won't work the way i want it to 11:03
11:03 locsmif left
timotimo however, i could change that to use ({ $*IN.get } ... *) instead 11:03
grondilu Option 'e' in grouped options '-ne' needs a value, but does not have one
^got this
pelr6 -ne '.say'
*perl6 11:04
timotimo r: my $s; for ({ $*IN.get } ... *) { $s += 1; }; say $s;
camelia rakudo 38a0f9: OUTPUT«(timeout)»
timotimo r: my $s; for ({ $*IN.get } ... $*IN.eof) { $s += 1; }; say $s;
camelia rakudo 38a0f9: OUTPUT«(timeout)»
timotimo r: say $*IN.^methods
camelia rakudo 38a0f9: OUTPUT«open close eof get getc lines read seek tell write opened t print slurp spurt copy chmod IO path flush encoding d e f s l r w x z modified accessed changed say Str gist perl <anon> <anon>␤»
moritz I'm pretty sure $*IN.eof is evaluated only once
timotimo oh
grondilu timotimo: you don't have to use $*IN explicitely
timotimo r: my $s; for ({ $*IN.get } ... { $*IN.eof }) { $s += 1; }; say $s;
camelia rakudo 38a0f9: OUTPUT«24␤» 11:05
timotimo yes that's better
hm, well, lines uses a gather/take that really should be lazy. 11:06
and map should be lazy, too! weird. 11:07
11:08 donaldh left
moritz well 11:08
map is lazy
but the question is how it is made eager in the end 11:09
timotimo oh? could that be why i don't see lines echoed back at me in perl6 -e 'map { .say }, lines'?
moritz maybe 11:10
timotimo all in all, that's very unhelpful.
11:10 dmol left, dmol1 joined
timotimo however, why does it work when i do the { $*IN.get } ... { $*IN.eof } hack? 11:10
grondilu guess it's because of the sequence operator with a closure at the end 11:13
timotimo i don't understand. 11:14
grondilu thinks it forces the laziness of the for
kind of the same reason why for 1, 2 ... * { ... } works
timotimo it can't be buffering in the input buffer, can it? 11:16
using gather while True { take $*IN.get } has the same problem lines() does, it seems
have gather blocks suddenly become eager? >_> 11:17
grondilu you mean gather while True { take get; last if eof } 11:18
timotimo yeah.
grondilu or gather until eof { take get }
timotimo eof doesn't seem to be built
grondilu can try that right now as he's recompiling rakudo
flussence r: gather take get until eof
camelia rakudo 38a0f9: OUTPUT«===SORRY!=== Error while compiling /tmp/FaqOU8ura4␤Undeclared routine:␤ eof used at line 1␤␤»
moritz $*ARGS.eof or so 11:19
grondilu sub eof { $*IN.eof } # just had that in you .perl6rc :) 11:20
grondilu knows there is no such thing as .perl6rc, but you know what I mean 11:21
.-2s/had/add/
.-3s/you/&r/ 11:22
daxim ewww, is that viml? 11:24
dalek p: 459e172 | (Timo Paulssen)++ | / (6 files):
initial draft of optimizer, inactive ATM.
 11:26
p: ba977c2 | (Timo Paulssen)++ | src/NQP/Optimizer.nqp:
working build, lots of debug output, more intelligence.
p: 8bea225 | (Timo Paulssen)++ | src/NQP/Optimizer.nqp:
nqp-jvm doesn't like negative nums in substr
p: e63cc6f | (Timo Paulssen)++ | src/NQP/Optimizer.nqp:
this should have been gone already.
p: 60e9b7e | (Timo Paulssen)++ | src/NQP/Optimizer.nqp:
little improvements. actually finds symbols now.
p: 3e45f29 | (Timo Paulssen)++ | src/NQP/Optimizer.nqp:
removed debug output
timotimo ^ seemed good enough to be merged
p: f1d040d | (Timo Paulssen)++ | / (6 files):
Merge branch 'origin/nqp-optimizer'

This gives us an optimizer for NQP that removes some int to num and num to int coercions if they are superfluous.
rlito: 3158ca6 | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - tweak the perl5 pretty-printer
 11:34
11:35 donaldh joined
donaldh testing binary IO ops on JVM, I guess there are encoding differences I cannot see visually: gist.github.com/donaldh/b66e23a436daed71cdcd 11:36
timotimo you can always pipe it through xxd 11:37
also, maybe on the way "out" it gets mangled?
moritz tests pass here 11:38
donaldh: what's your locale?
donaldh moritz: en_GB.UTF-8
moritz ok, that shouldn't be the problem 11:39
donaldh I've just written the binary ops so I've probably messed up the encoding steps.
:)
moritz try printing out the Buf/Blob objects directly
11:51 woolfy joined
dalek rlito: 5b1c58c | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - tweak the perl5 pretty-printer
 11:55
11:57 kbaker joined 12:00 benabik left 12:02 donaldh left 12:03 JimmyZ_ joined
dalek rlito: 819f0a3 | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - tweak the perl5 pretty-printer
 12:07
12:16 xenoterracide left 12:19 fhelmberger_ joined 12:20 jnap joined 12:22 fhelmberger left 12:28 sqirrel left
dalek rlito: 17e8217 | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - tweak the perl5 pretty-printer
 12:28
12:34 census joined
dalek rlito: 22db30e | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar compiler - tweak the perl5 pretty-printer
 12:37
pmichaud FROGGS: ping 12:42
dalek rlito: d6aeb0d | (Flavio S. Glock)++ | / (3 files):
Perlito5 - grammar compiler - tweak the perl5 pretty-printer
FROGGS pmichaud: pong
pmichaud did you really mean <[\s+\r\n]> in your comment on github.com/perl6/std/pull/5#issuec...t-23252504 ?
12:43 ajr joined
FROGGS pmichaud: I think so, yes 12:43
pmichaud \s already includes \r and \n
and we probably don't want to match +
12:43 ajr is now known as Guest76882
FROGGS ahh 12:43
yes
but I posted " <[\s\r\n]>+ ", see :P 12:44
(I edited my comment right now)
pmichaud which is shorter as \s+ :-)
FROGGS and edited again -.-
yeah, I recognized it after writing here 12:45
I just can blame that and me having a cold
s/and/on/
-.-
pmichaud colds can do that, definitely.
I think a lot of the synopses may be to TimToady++ colds. :-P :-P
FROGGS my brain is a bit slow these days
*g*
pmichaud *be due to 12:46
FROGGS pssssst, this channel is logged
12:46 ggoebel left
FROGGS we tend to rant about TimToady in #rant-a-toady 12:46
12:46 ggoebel joined
FROGGS TimToady++ even 12:46
pmichaud I tend to be more public in my rants. :) 12:49
12:50 colomon left, kaleem left
FROGGS yeah, this might bring up more well-considered rants 12:50
12:51 not_gerd joined, kaleem joined
not_gerd o/ 12:51
FROGGS hi not_gerd
not_gerd quick question - if we pass a negative offset to nqp::substr() that greater than the length, should we wrap around or return an empty string? 12:52
*that's
FROGGS nr: say substr("abc", -5, 1)
camelia rakudo 38a0f9: OUTPUT«Start argument to substr out of range. Is: -5, should be in 0..Inf; use *-5 if you want to index relative to the end␤ in method gist at src/gen/CORE.setting:11494␤ in method gist at src/gen/CORE.setting:990␤ in sub say at src/gen/CORE.setting:12390␤ in block a…
..niecza v24-88-g1f87209: OUTPUT«(Str)␤»
FROGGS not_gerd: you meant offset? 12:53
nr: say substr("abc", *-5, 1)
camelia rakudo 38a0f9: OUTPUT«Start argument to substr out of range. Is: , should be in 0..Inf; use *-2 if you want to index relative to the end␤ in method gist at src/gen/CORE.setting:11494␤ in method gist at src/gen/CORE.setting:990␤ in sub say at src/gen/CORE.setting:12390␤ in block at …
..niecza v24-88-g1f87209: OUTPUT«(Str)␤»
not_gerd I was looking at the JVM ops
it checks `if (offset < 0) offset += val.length();` 12:54
logically, argument could be made that it should be `while (offset < 0) offset += val.length();`
FROGGS hold on
12:54 pmurias joined
dalek rlito: 53222da | (Flavio S. Glock)++ | / (4 files):
Perlito5 - grammar compiler - tweak the perl5 pretty-printer
 12:55
moritz not_gerd: I don't see what that would mean
not_gerd: the current thing means "if it's negative, index from the end"
pmurias how can I detect windows in nqp code? 12:56
not_gerd moritz: yes, but what should happen if the offset is greater in magnitude than the string's length?
moritz not_gerd: an error is reported to the user
clkao win2 4
12:56 dayangkun joined
FROGGS I can't find the spec about this 12:57
not_gerd nqp/jvm dies with StringIndexOutOfBoundsException
also, if we overflow at the end, we just return an empty string 12:59
FROGGS true
12:59 iSlug left
FROGGS that speaks for the current impl 13:00
not_gerd (the current impl being throwing a VM error?!)
FROGGS no, just doing an if(off<0) and not while(... 13:01
not_gerd FROGGS: but that *does* result in the vm error because off will still be negative
(which then gets passed down to Java's substring) 13:02
nwc10 not sure if I'm asking the wrong silly question here, but if the negative index is allowed to wrap multiple times, why isn't the positive index allowed to wrap at all?
FROGGS because substr("abc", *-17) is clearly a thinko, and shouldn't be hidden
not_gerd: sure, we need to raise a X::
not_gerd FROGGS: and substr("abc", 42) isn't?
moritz nr: say substr('abc', 42)
camelia rakudo 38a0f9: OUTPUT«Start of substr out of range. Is: 42, should be in 0..3␤ in method gist at src/gen/CORE.setting:11494␤ in method gist at src/gen/CORE.setting:990␤ in sub say at src/gen/CORE.setting:12390␤ in block at /tmp/_yUdcJIo4_:1␤␤» 13:03
..niecza v24-88-g1f87209: OUTPUT«(Str)␤»
FROGGS sanity++
moritz nwc10: we never wrap implicitly in the Perl 6-level substr
not_gerd r: nqp::substr("abc", 42)
camelia rakudo 38a0f9: OUTPUT«Cannot take substr outside string␤ in block at /tmp/XsZs31YsMj:1␤␤»
not_gerd nqp: nqp::say(nqp::substr('abc', 42))
camelia nqp: OUTPUT«Cannot take substr outside string␤current instr.: '' pc 40 ((file unknown):39) (/tmp/5PTBCxr0sk:5)␤»
moritz nwc10: we wrap explicitly if the argument is a callable (e.g. *-42)
FROGGS nqp-jvm: nqp::substr("abc", 42)
camelia nqp-jvm: ( no output ) 13:04
nwc10 moritz: ah OK. I should probably be clear - I don't think that a "while" loop to wrap negative indexes is sensible
moritz r: substr('abc', *-42)
FROGGS nqp-jvm: say(nqp::substr("abc", 42))
camelia rakudo 38a0f9: OUTPUT«Start argument to substr out of range. Is: , should be in 0..Inf; use *-39 if you want to index relative to the end␤current instr.: 'throw' pc 380147 (src/gen/CORE.setting.pir:161329) (src/gen/CORE.setting:10132)␤called from Sub 'sink' pc 415708 (src/gen/CORE.setti…
nqp-jvm: OUTPUT«␤»
FROGGS LTA
nwc10 and I'm not even sure whether NQP should be stricter, and not take negative indexes
moritz nwc10: then we are in violent agreement
13:05 SamuraiJack_ left
not_gerd keep in mind we're talking about nqp:: here - P6 of course should do what's right[tm] 13:05
moritz could argue both for dying outright, and for wrapping negative numbers once
not_gerd I'd just like to have some consistency - either throw errors at both ends, or return empty strings
I agree that wrapping around is probably not a good idea 13:06
nqp-jvm: nqp::say(nqp::substr('abc', 1, 42))
camelia nqp-jvm: OUTPUT«bc␤»
moritz not_gerd: it's probably a good idea not to bomb more than nqp-parrot does right now, because the regex engine (might) rely on how it's done 13:07
not_gerd nqp: nqp::say(nqp::substr('abc', 1, 42))
camelia nqp: OUTPUT«bc␤»
moritz nqp: nqp::say(nqp::substr('abc', 42)) 13:10
camelia nqp: OUTPUT«Cannot take substr outside string␤current instr.: '' pc 40 ((file unknown):39) (/tmp/WONygss6Oe:8)␤»
moritz nqp: nqp::say(nqp::substr('abc', -1))
camelia nqp: OUTPUT«c␤»
moritz nqp: nqp::say(nqp::substr('abc', -2, -1))
camelia nqp: OUTPUT«␤»
moritz nqp: nqp::say(nqp::substr('abc', -2, 1))
camelia nqp: OUTPUT«b␤»
moritz nqp: nqp::say(nqp::substr('abc', 0, -1))
camelia nqp: OUTPUT«␤»
not_gerd note that Parrot and JVM don't agree in all of these cases
13:10 colomon joined, salv0 left
not_gerd (and MoarVM potentially just returns garbage ;)) 13:10
timotimo maybe an evalbot for "the nqps" would be nice to have
since it automatically points out differences 13:11
moritz returning garbage isn't a good idea
not_gerd that's what I'm currently fixing
the question is, what should I aim for?
timotimo moritz: getting at the allocation data of a string with sufficiently negative substr offsets, what's not to like?
not_gerd proposal: translate offset + length to start-index + end-index 13:12
if either is out of range, adjust to fit string size
JimmyZ_ throw exception :)
moritz nqp-jvm: say(nqp::substr('abc', 0, 42)) 13:13
camelia nqp-jvm: OUTPUT«abc␤»
not_gerd I'd only like to thro if boch indices are out of range
ie if the window doesn't cover a substring at all
dalek p/shell3: 3adbc0a | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Make nqp::shell respect the 3 argument on linux.
 13:14
pmurias FROGGS: could you check if the above commit compiles on windows?
FROGGS pmurias: will do 13:16
13:17 PacoAir joined
colomon pmichaud: nok exists and is used 587 times already in roast. dunno about isnt, though 13:17
errr, isnt is there too.
much more rarely used, though.
13:18 fhelmberger_ left
colomon will one day learn to actually check on things that can be checked with 9 characters in the shell before spending 24 characters telling the world he hasn't bothered to check on them yet. 13:20
FROGGS *g* 13:21
pmurias: is there a test I can run afterwards? 13:24
13:25 btyler joined
pmurias FROGGS: currently things work the same under windows, I just want to check things build correctly after moving code 13:26
[Coke] returns from “vacation” 13:27
yoleaux 08:08Z <nwc10> [Coke]: that FRA is about 15 minutes from central Frankfurt by train, and HNH is 124km away by road.
FROGGS pmurias: k, nqp still compiles
pmurias: do you want me do build rakudo too?
pmurias not yet
FROGGS k
pmurias test 78 passes?
13:28 markov joined, kaare_ joined
FROGGS pmurias: yes, nqp builds and 78 passes 13:29
dalek p/shell3: 6cbe2ef | (Pawel Murias)++ | src/vm/parrot/ops/nqp.ops:
Pass the env variables on windows.
 13:30
pmurias FROGGS: what's the output of pastie.org/8270977? 13:32
nwc10 [Coke]: sorry, I typo'd Hahn's airport code. I hope that that information is useful and not too late.
[Coke]: it's probably actaully easier to get to central Frankfurt from Cologne/Bonn airport than from Hahn. But I have not tested this. 13:33
FROGGS pmurias: gist.github.com/FROGGS/393b2db32957ba103f97 13:34
nwc10 (I think that the only time I've actually travelled between Frankfurt-ish and Cologne was by boat down the Rhine.
www.kdrhine.com/rhineschedule.htm ) 13:35
13:36 notjack joined
pmurias FROGGS: that's after my last commit? 13:36
FROGGS pmurias: yes, revision is 2013.08-22-g3adbc0a 13:37
13:37 fhelmberger joined
pmurias where did you get the revision from? 13:38
13:39 xinming left
FROGGS pmurias: git describe and nqp --version 13:39
13:40 xinming joined
pmurias commit 6cbe2ef83b7257ce958c1f4cc93c204ea6bc0955, and the results were after running make? 13:40
FROGGS brb 13:41
[Coke] nwc10: if FRA is a good option, that's fine. 13:42
nwc10 it's Lufthansa's main hub - it's probably one of the easiest airports in Europe to get to (for a real airline) 13:44
13:45 xinming left, xinming joined, virtualsue joined
pmurias rakudo: say $^OS; 13:47
camelia rakudo 38a0f9: OUTPUT«===SORRY!=== Error while compiling /tmp/l9rGMlmXFA␤Cannot use placeholder parameter $^OS in the mainline␤at /tmp/l9rGMlmXFA:1␤------> say $^OS⏏;␤»
pmurias rakudo: say $*OS;
camelia rakudo 38a0f9: OUTPUT«linux␤»
nwc10 oh boy. I would not want to be the owner of a business 94% dependent on RyanAir 13:48
www.hahn-airport.de/Default.aspx?me...&cc=en
moritz :-) 13:55
13:55 markov left 13:56 rindolf joined 13:59 berekuk left
FROGGS pulls and rebuilds 14:00
14:03 berekuk joined
FROGGS pmurias: gist.github.com/FROGGS/393b2db32957ba103f97 14:07
gist.github.com/FROGGS/393b2db3295.../revisions
pmurias ah, I'm passing the env variable incorrectly on windows 14:14
14:17 daxim left, daxim_ joined 14:18 donaldh joined
dalek rlito: ae790ba | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - tweak the perl5 pretty-printer
 14:27
14:28 kaleem left
masak oh hai from an aeroplane! \o/ 14:32
not_gerd github.com/MoarVM/MoarVM/blob/HEAD...ops.c#L342 14:34
FROGGS masak: \o/
not_gerd ^ if someone want to see how nqp::substr() on MoarVM anded up behaving
*ended
14:35 benabik joined
not_gerd at least 3 of the NQPs now disagree what nqp::substr() should do ;) 14:38
FROGGS end_pos = length == -1 ? agraphs : start_pos + length;
do I read that right that in case I want to read from end it always reads to the end? 14:39
r: say nqp::substr("abcdef", -4, 2)
camelia rakudo 38a0f9: OUTPUT«cd␤»
14:40 Psyche^ joined
not_gerd FROGGS: length is the 3rd argument 14:42
if you eant to read from the end, the 2nd needs to be negative
*want
FROGGS r: say nqp::substr("abcdef", -4, 2) # <--- second is negative
camelia rakudo 38a0f9: OUTPUT«cd␤»
not_gerd but length is 2, which is != -1 14:43
FROGGS right
not_gerd so I'm not seeing what you're getting at
14:43 Psyche^_ left
FROGGS ahh, hold on 14:43
length = -1 does mean Inf 14:44
14:44 ajr joined, Guest76882 left, ajr is now known as ajr_
not_gerd FROGGS: right 14:45
FROGGS end_pos = length == -1 ? agraphs : start_pos + length; # another thing: if length==Inf, you read up to strlen, but other wise you read up to offset+strlen
I'd guess it should read up to strlen-offset
not_gerd length != strlen 14:46
agraphs == strlen
14:46 lizmat joined, berekuk left
FROGGS length is number of graphemes to read... nvm then :o) 14:46
I'm a bit dumb today 14:47
dalek rlito: eb84bd7 | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - tweak the perl5 pretty-printer - fix eval-block
not_gerd basically, the substring is bounded by offset and offset+length
14:47 pecastro left
not_gerd we grab any parts of the string at lies in that interval 14:48
then come the small compications that offset can count from the and and length == -1 signifies Inf
14:48 berekuk joined
not_gerd we only throw an error if the upper bound is < 0 or the lower bound > strlen 14:49
neither Parrot nor JVM do it that way ;)
nwc10 JVM presumably copies Parrot. 14:55
not_gerd nwc10: it doesn't
nwc10 oh?
not_gerd nqp-jvm: nqp::say(nqp::substr('abc', 5)) 14:56
camelia nqp-jvm: OUTPUT«␤»
not_gerd nqp: nqp::say(nqp::substr('abc', 5))
camelia nqp: OUTPUT«Cannot take substr outside string␤current instr.: '' pc 40 ((file unknown):39) (/tmp/NKEQ3mtRQ4:7)␤»
not_gerd nqp: nqp::say(nqp::substr('abc', -1)) 14:57
camelia nqp: OUTPUT«c␤»
not_gerd nqp-jvm: nqp::say(nqp::substr('abc', -1))
camelia nqp-jvm: OUTPUT«java.lang.StringIndexOutOfBoundsException: String index out of range: -1␤ at java.lang.String.substring(String.java:1871)␤ at org.perl6.nqp.runtime.Ops.substr2(Ops.java:2087)␤ at 849716E0592128BADBA0F494078445F51EDED46B.qb_10(<generated>)␤ at
..org.perl6.nqp.runtime.Ops.in…
not_gerd totally unbiased that I am, I like MoarVM's behaviour best ;) 14:59
FROGGS *g*
not_gerd oops 15:01
I pushed more than I wanted to push :( 15:02
there's some boying missing
15:04 bluescreen10 joined
not_gerd *boxing 15:09
15:11 fhelmberger left 15:13 btyler left 15:25 pecastro joined 15:26 celesta joined
celesta ahoi #perl6 15:27
moritz \o celesta
[Coke] .seen tene 15:28
yoleaux I saw Tene 21 Aug 2013 19:55Z in #perl6: <Tene> pmichaud: My condolences for your loss.
lizmat .tell [Coke] FRA and yes
yoleaux 25 Aug 2013 21:54Z <[Coke]> lizmat: Which airport to fly to, FRA or HHN?, and also, any rooms left at the hotel?
lizmat: I'll pass your message to [Coke].
lizmat .tell [Coke] no need to worry about accommodation 15:29
yoleaux lizmat: I'll pass your message to [Coke].
lizmat also: only 2 places left for the Rakudo and NQP Internals Workshop, 14/15 Sep in Frankfurt, Germany
15:30 btyler joined
lizmat If you want to get into supporting Perl 6 / Rakudo development with a flying start 15:30
this is the place to be!
[Coke] lizmat - when are accomodations available? (e.g. when should I book flights around?)
yoleaux 15:28Z <lizmat> [Coke]: FRA and yes
15:29Z <lizmat> [Coke]: no need to worry about accommodation
lizmat </and of promotional message>
[Coke] hugme: hug lizmat
hugme hugs lizmat
lizmat accommodation has been booked with checkin on Fri and checkout on Mon 15:31
[Coke] Roger that.
15:32 daxim_ left
lizmat if you're arriving earlier / leaving later, let me know 15:32
colomon lizmat++
timotimo lizmat: do you know if i can park my car for free at that hotel you booked lots of spots in? or perhaps there's a free spot on the venue's parking lot or something? 15:34
[Coke] I'm pretty sure those dates work well with minimum-price booking of flight.
timotimo is pretty new to traveling by car
[Coke] is squee-excited. 15:35
lizmat the hotel policies of the reservation states: Guest parking:
Free public parking is possible at a location nearby (reservation is not possible).
TimToady FROGGS: unsp already allows \s*, the issue is whether to allow another \ in the middle of those spaces 15:36
yoleaux 07:46Z <FROGGS> TimToady: What if <unsp> unspaces more than one char? It could unspace \s+, wouldn't that be better?
FROGGS hmmmm
dalek kudo/nom: 283f763 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Make @a.VAR.name and %h.VAR.name work
TimToady std: 42\ i
camelia std 06636e4: OUTPUT«ok 00:00 41m␤»
FROGGS TimToady: then it is really just about a blank line 15:37
TimToady std: 42\ \ i
camelia std 06636e4: OUTPUT«===SORRY!===␤Postfix found where infix expected (change whitespace?) at /tmp/MTXcim2yp8 line 1:␤------> 42\ \ ⏏i␤ expecting any of:␤ horizontal whitespace␤ infix or meta-infix␤ infixed function␤ postcircumfix␤ postfix␤
.. postfix_…
TimToady unless we decide 42\ \i should work too
15:37 pecastro_ joined
TimToady but I'm not convinced about that part 15:38
FROGGS 42\
\
\ i
TimToady the blank line part seems harmless-ish
that would also work, though lacks a kind of visual symmetry...
PerlJam almost looks like a C++-ism comment block 15:39
s/ism/ish/
15:40 pecastro left
timotimo how am i supposed to properly turn infix:<~> into a list-associative if i have to take care of str, Str and Buf? :| 15:44
15:45 colomon left, Kiralexis left 15:46 FROGGS[mobile] joined 15:47 FROGGS left
TimToady and Uni, and NFD, and NFC... 15:47
timotimo indeed. 15:48
timotimo is a tiny bit more fond of replacing strings of infix:<~> with a join with "" during optimization :| 15:49
15:56 colomon joined 15:58 JimmyZ_ left 15:59 zakharyas joined
pmurias can the sh be located somewhere else then /bin/sh on a unixy system? 16:00
that is can I change execl(...) to execlp(...)? 16:01
lizmat r: my &a; say &a; my Int &b; say &b # somehow expected Any,Int 16:02
camelia rakudo 38a0f9: OUTPUT«(Any)␤(Any)␤»
dalek ast: a6a46c3 | (Elizabeth Mattijsen)++ | S02-names/name.t:
Tests for .VAR.name
 16:04
kudo/nom: efc4efd | (Elizabeth Mattijsen)++ | t/spectest.data:
Make sure .VAR.name tests are run
 16:05
japhb Is --profile supposed to work in rakudo-jvm? (I ask because when I tried it last night, I got a 50 byte file and a java process that sat for hours using only 1% CPU, apparently doing not much at all.)
geekosaur pretty sure POSIX mandates existence of /bin/sh although how it gets there is open (that is, it could be a symlink somewhere else) 16:06
16:07 dmol1 left, domidumont left 16:08 btyler_ joined, colomon left 16:10 btyler left 16:11 FROGGS joined 16:14 JimmyZ_ joined 16:15 JimmyZ_ left 16:16 pmurias left
TimToady timotimo: join '' still doesn't help with the string types though 16:23
and it's not like we've figured out how to enforce types on slurpies yet either...
and how much coercion we allow between related string types is also underspecced 16:25
16:27 fhelmberger joined
TimToady upgrading any well-typed encoding or normal form to NFG is presumably always safe, but maybe not always what the user wants 16:29
timotimo that's true :|
lizmat but NFG would only be internal, no ? 16:30
TimToady yes
lizmat so opaque to the user, I would think ?
TimToady presumably an IO handle has the ability to specify the downcasting
lizmat indeed, pretty much like P5 IO layers, I would think, but then properly implemented 16:31
TimToady how much we can propagate that knowledge through type inference to avoid unnecessary casting is an open question
we can have explicit downcasts as functions, then we only have to know the immediate context of a join to know the intended type 16:32
we could have a pragma saying that we think we're working in, say, NFD, and then ~ might coerce to that
16:32 domidumont joined
lizmat why not always do everything internally in NFG ? 16:32
and decode on the way out ?
TimToady we could have a named arg to infix:<~> etc. that specifies desired result type, if not NFG
well, that's the default, but some code will want to work intentionally in lower abstraction levels 16:33
another approach is to coerce everthing to the type of the first argument
jnthn We did that with integer arithmetic, and if you mixed in to the type of the first arg stuff got fun :P 16:34
Then we fixed it to always make an Int :)
16:34 fhelmberger left
TimToady nodnod 16:34
jnthn Where we was probably moritz++ :)
TimToady so I'm guessing an explicit function to downcast is okay, and if some type information can someday flow backwards into the optimizer, that's good enough for now 16:35
yes, much as we put all our numeric intermediate results into the more abstract type 16:36
timotimo how about getting that into nqp, though? having nqp's infix:<~> be a list-assoc? (or, you know, i could just port the tilde-to-strjoin transformation from the rakudo optimizer to the nqp optimizer)
TimToady NFG is our Int, while Uni/NFD/NFC/utf8 etc are more like native types
timotimo nqp only deals with str and things that would be str-ified
afaict
jnthn gonna go for nomz 16:37
bbl
16:37 lowpro30 joined
jnthn & 16:37
timotimo gonna go for groceries :)
diakopter NF* is lossy 16:39
b/c it reorders things
so you can't use it when you need to output exactly what was inputted bit for bit
16:41 rindolf left, rindolf joined
lizmat diakopter: why would NFG be lossy ? 16:43
or are you referring to *valid* unicode combinations not coming back uncombined ?
16:44 Pleiades` left
diakopter b/c it's the same as NFC 16:44
they reorder combining codepoints
I'm just pointing out that traditional Perl way of using "strings" for binary data doesn't work if those strings are nfg 16:45
lizmat well, maybe NFG should also create virtual codepoints for *any* combining codepoint
so that it can restore it on output ?
benabik O.o
I thought the point of NFG was that the many ways of expressing a character resulted in one codepoint. 16:46
TimToady round tripping is explicitly not a goal of NFG
diakopter yeah
TimToady if you want round tripping, you must work in a lower abstraction level
diakopter so anyone coming from p5 must learn not to use strings for binary data 16:47
lizmat hmmm… not roundtripping feels a bit like a WAT to me
16:47 ajr_ left
timotimo round trippin' with my two favorite languages / fully loaded we got codes and graphemes 16:47
lizmat diakopter: that is clear
16:47 FROGGS[mobile] left
diakopter lizmat: but that's implied by the N[ormalized] - multiple inputs produce one output 16:48
lizmat I'm more worried about cases where a text file gets processed, without actually doing anything, but writing out something else
diakopter right. can't use strings
16:48 Pleiades` joined
lizmat because of normalization 16:48
otoh, that will only happen once, any subsequent processing of the same file would not change anything 16:49
benabik Does it matter if the bits change if the data doesn't? I'd say that the 90% answer is no. 16:50
diakopter by the same program..
other programs may re-corrupt it
lizmat diakopter: yes, that was the assumption
benabik And Bufs are there for the other 10%.
diakopter benabik: but binary data will have "codepoints" outside unicode 16:51
[if you try to decode it as such]
lizmat is reminded of a system in which *each* codepoint would be remapped to a BER encoded integer
benabik diakopter: Treating binary data as Unicode is a mistake?
diakopter so if you want binary data strings, you have to set your filehandle to decode as if it's ascii or whatever 16:52
16:52 dmol joined
benabik If you want binary data strings, tell the FH to treat it as binary, no? 16:52
diakopter that's at the os layer; at the perl layer it wants to decode to something
if you want a Str
lizmat in practice, that meant bytes for most text files, as they would have fewer than 256 different codepoints in them
diakopter lizmat: but that's exactly what nfg does.. :P 16:53
lizmat hmmm.. BER would mean 127, but yes, you get the idea
I thought it only did that for codepoints that could not combine ?
benabik Why do you want a Str for binary data?
diakopter no, also codepoints >127 16:54
16:54 colomon joined
diakopter benabik: it's just that most of the io "ops" return strings 16:54
lizmat but why not go all the way then? 16:55
TimToady all the way where?
diakopter map every used codepoint?
lizmat well, not treat ASCII specially,
map every codepoint?
benabik S32-IO has write(Buf), but not read( --> Buf) 16:56
lizmat every *used* codepoint, that is
benabik Although slurp can return a Buf apparently.
FROGGS benabik: well, the mmd can see if you passed a Str of Buf, but it can't see if you want a Str or Buf 16:57
TimToady lizmat: most strings will work in NFC most of the time, and then you don't have to do extra work unless there are chars that don't graphemize in NFC
FROGGS so, a read( :bin ) would be needed
benabik FROGGS: slurp returns Str|Buf depending on your desired encoding.
diakopter well you could pass a scalar ref. <hide>
TimToady we will have to be more particular in cases where we want a particular type to be read 16:58
16:58 SamuraiJack_ joined, woolfy1 joined, woolfy left
benabik Really, I would think that more of the IO::Handle funcs should return Str|Buf and return Buf when the encoding is 'binary'. 17:00
dalek rlito: 0cbac5f | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - emit subroutine signature
TimToady Str vs Buf is an oversimplification 17:01
17:01 denisboyun left
TimToady maybe it defaults to NFG, but you can requrest NFD, NFC, or Uni (raw codepoints, unnormalized) 17:02
Buf defaults to Buf8, but there are other buf sizes, as well as blobs of utf8, utf16, etc.
benabik If you ask for binary data, you should end up with Buf/Blob not Str, no? Or are we going to have Strs with :encoding<binary> ? 17:03
TimToady Str will mean NFG, period
we can hid some of these differences from the naive user, but it's imperative that the computer not be confused about what type it's dealing with 17:04
*hide
that was the Perl 5 mistake, and we're making different mistakes this time
"binary" is underspecified, so it's kind of a bad word, like "length" 17:05
"string" is also underspecified, so it's also kind of a bad word 17:06
benabik Perhaps we need to have encodings like 'uint8', 'int32', etc. Perhaps with 'binary' aliased to 'uint8'. Then IO with these encodings return Buf[type] 17:07
17:08 spider-mario joined
TimToady well, unless they return Blob[type] 17:08
17:08 berekuk left
benabik Although IO with anything other than 8 bit types also has to worry about endian-ness, etc. :-/ 17:09
masak waves from Oslo
lizmat waves back from Echt
r: my &a; say &a; my Int &b; say &b # somehow expected Any,Int 17:10
camelia rakudo efc4ef: OUTPUT«(Any)␤(Any)␤»
lizmat is this a rakudobug ?
17:10 berekuk joined
benabik What's `my Int &b` supposed to be? A function that returns Int? 17:10
lizmat yes
that's how I read it
masak lizmat: don't think it's a bug.
but I'm not sure. 17:11
benabik Then &b shouldn't return Int, it should return a null version of a function that returns Int, no? :-D
dalek rlito: 3958e2f | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - emit shift left/right
17:11 xinming left
lizmat which would be indistinguishable from Int, no? 17:11
masak I'm not sure Any qualifies as a function that returns an Int, so maybe that's the bug. 17:12
dinner &
benabik r: say Int()
camelia rakudo efc4ef: OUTPUT«0␤»
benabik r: say Any() 17:13
camelia rakudo efc4ef: OUTPUT«No such method 'Any' for invocant of type 'Parcel'␤ in block at /tmp/_OVjj9v3NO:1␤␤»
17:13 colomon left, xinming joined 17:14 ajr joined, ajr is now known as Guest91190 17:15 Guest91190 is now known as ajr_ 17:17 Pleiades` left 17:22 woolfy joined, woolfy1 left 17:23 Pleiades` joined 17:24 btyler_ left 17:25 virtualsue left 17:26 danaj left
donaldh Debugging the slurp test faiures on JVM. Looks like a JVM bug to me: gist.github.com/donaldh/6344106 17:29
17:30 denis_boyun joined
FROGGS what do you mean? 17:31
timotimo what does the limit mean?
FROGGS the chars/elems I guess
29 chars but 39 bytes
benabik I don't see a bug there. The UTF-8 encoding of that string is 39 bytes.
Although Vim saves it as 40 bytes. :-/ 17:32
moritz benabik: trailing newline? 17:33
TimToady or inserted a CR?
benabik moritz: There's a trailing newline in the string.
17:33 Faxmachinen joined
Faxmachinen Hi! 17:33
Anyone know of a clever way to transform a two-dimensional list? 17:34
TimToady o/
benabik Oh. Yes, there is a newline after the newline. moritz++
moritz Faxmachinen: depends on the transformation
FROGGS 39 bytes on my box (and 29 chars)
Faxmachinen Sorry, I meant transpose 17:35
moritz rosettacode.org/wiki/Matrix_transposition#Perl_6 17:36
donaldh hmmm.
When I look at the written file, it has nulls.
Or something.
TimToady though the RC entry is more brute force than "clever" 17:37
Faxmachinen Thanks moritz. I feel like there should be some way to go @foo.invert though
Or something like that
benabik Matrix transposition is not a useful operation on most arrays. 17:38
timotimo i love how the i in the middle of the matrix is highlighted :D
FROGGS ** 17:39
17:39 danaj joined
FROGGS *g* 17:39
no, should be *i*
TimToady ii, captain
^bold on some printers 17:40
donaldh okay, barking up wrong tree. Asymmetry creeping in elsewhere.
FROGGS .oO( a 2nd captain? )
geekosaur just zaphod
flussence ha, I couldn't remember whether there was a way to round() to a specific decimal precision so I just tried the first thing that came to mind... and it worked :D
TimToady captains are supposed to be bold
except for your airplane captain
Faxmachinen Hmm. Well, the zip operator kinda does it
FROGGS Picard was never bold, though Kirk was around the middle 17:41
lizmat some captains are even bald
[Coke] hurls mathwithbaddrawings.com/2013/08/13/...-decimals/ and shivers.
Faxmachinen rakudo: say [Z] ([1..3], [4..6], [7..9]) 17:43
camelia rakudo efc4ef: OUTPUT«1 4 7 2 5 8 3 6 9␤»
TimToady r: say [Z]([1,2,3],[4,5,6],[7,8,9]).tree.perl
camelia rakudo efc4ef: OUTPUT«([1, 4, 7], [2, 5, 8], [3, 6, 9]).list␤»
Faxmachinen Then I just gotta figure out an elegant way to split a flat array into n parts of equal size
timotimo it's not flat :) 17:44
TimToady this one should work with .tree
well, the Xxx should work too, but doesn't...
Faxmachinen Or that, yes :)
You're right, it's not flat
Genius! 17:45
TimToady r: say [Z]([1,2,3],[4,5,6],[7,8,9]).lol.perl 17:46
camelia rakudo efc4ef: OUTPUT«LoL.new(ListIter.new())␤»
TimToady .perl should not give up on a list like that
r: say [Z]([1,2,3],[4,5,6],[7,8,9]).lol.[0] 17:47
camelia rakudo efc4ef: OUTPUT«ListIter.new()␤»
TimToady bleah
r: say [Z]([1,2,3],[4,5,6],[7,8,9]).lol.[0][0] 17:48
camelia rakudo efc4ef: OUTPUT«ListIter.new()␤»
TimToady !lol
[Coke] pugs: say (1,2,3).lol 17:49
camelia pugs: OUTPUT«*** No such method in class Array: "&lol"␤ at /tmp/FfNKGX8ijR line 1, column 5 - line 2, column 1␤»
TimToady lol is post-pugs
[Coke] Anyone seen .au lately? I hate to give up on pugs, but am having issues getting it running on the new roast box. 17:50
dalek rlito: a1b4639 | (Flavio S. Glock)++ | / (2 files):
Perlito5 - perl5 - emit function deref
 17:52
notjack r: say.WHAT 17:56
camelia rakudo efc4ef: OUTPUT«␤»
timotimo r: say.WHAT.say
camelia rakudo efc4ef: OUTPUT«␤(Bool)␤»
notjack r: say.WHAT? 'no':'he di-int';
camelia rakudo efc4ef: OUTPUT«===SORRY!=== Error while compiling /tmp/SdemB7xier␤Unsupported use of ?: for the conditional operator; in Perl 6 please use ??!!␤at /tmp/SdemB7xier:1␤------> say.WHAT?⏏ 'no':'he di-int';␤»
[Coke] r: $^u.WHAT.say 17:57
camelia rakudo efc4ef: OUTPUT«===SORRY!=== Error while compiling /tmp/Tsrslnjce1␤Cannot use placeholder parameter $^u in the mainline␤at /tmp/Tsrslnjce1:1␤------> $^u⏏.WHAT.say␤»
FROGGS r: say.WHAT?? 'no'!!'he di-int';
camelia rakudo efc4ef: OUTPUT«␤»
notjack r: say.WHAT?? 'no' !! 'he di-int';
camelia rakudo efc4ef: OUTPUT«␤»
TimToady std: r: say.WHAT?? 'no' !! 'he di-int';
camelia std 06636e4: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/laASFVzzuk line 1:␤------> r: say⏏.WHAT?? 'no' !! 'he di-int';␤Check failed␤FAILED 00:00 42m␤»…
donaldh <-fail. aaaaaargh. 17:58
17:59 sftp_ joined
donaldh There was a bug in the file writing code that I did fix. In the process of investigating, I uncommented the clean up of the test files, so that I could read them. 17:59
They contained extra garbage. After the fix, they still contained extra garbage. They don't get truncated.
So the tests kept failing. 18:00
timotimo ooooh
18:01 anocelot_ joined
donaldh I'm guessing that the behaviour we want from nqp file open for "w" is WRITE, CREATE, TRUNCATE_EXISTING 18:01
not just WRITE, CREATE 18:02
18:02 BooK_ joined
FROGGS true, wa does not truncate of course 18:02
donaldh wa does WRITE, CREATE, APPEND 18:03
18:03 zby_home_ joined
FROGGS right 18:03
18:04 dagurval_ joined, sftp left, anocelot left, aindilis left, dylanwh_ joined 18:05 dylanwh_ is now known as dylanwh, gdey joined 18:06 japhb__ joined
donaldh yay, unfledged the binary slurp tests 18:07
*unfudged
18:09 btyler joined 18:10 rindolf left
timotimo very cool! :D 18:10
donaldh++ 18:12
18:14 lizamae joined
lizamae Hi 18:15
Helo
18:15 lizamae left
timotimo that was a short visit 18:15
sjohnson he's a busy guy. 18:16
18:19 felher joined
dalek kudo/nom: fe289bf | (Elizabeth Mattijsen)++ | src/core/Str.pm:
Naive implementation of Str.tc and tc(Str)

Waiting for RaNIW to implement nqp::tc as an exercise		 18:23
tadzik :) 18:24
18:25 not_gerd left
grondilu nqp: say("hello".tc) 18:25
camelia nqp: OUTPUT«Method 'tc' not found for invocant of class 'String'␤current instr.: '' pc 49 ((file unknown):43) (/tmp/L1bpK3j6LW:1)␤»
grondilu has no idea what Str.tc is
PerlJam grondilu: tc == titlecase
lizmat method "tc" on a Str
grondilu oh yeah. 18:26
I thought this was implemented already.
PerlJam me too
lizmat was going through spectest to see whether any rakudo skip's could be removed, or changed into todo's
and found that .tc wasn't implemented
PerlJam lizmat++
grondilu r: say "bonJour tout le monde :)".tc 18:27
camelia rakudo efc4ef: OUTPUT«No such method 'tc' for invocant of type 'Str'␤ in block at /tmp/aHrwx5t_IE:1␤␤»
lizmat you will have to wait for a bit before camelia has caught up
grondilu would it turn "heLLo" into "Hello" or "HeLLo"?
lizmat HeLLo 18:28
you want tclc for Hello
grondilu ok
18:31 ssutch left, abnorman joined 18:32 cognominal left, ssutch joined
lizmat r: my $?OS = BEGIN $*OS # wouldn't this be the way to implement $?OS in settings ? 18:37
camelia rakudo efc4ef: ( no output )
diakopter r: (role WAT { }).WAT.say 18:38
camelia rakudo efc4ef: OUTPUT«No such method 'WAT' for invocant of type 'WAT'␤ in block at /tmp/xnr3XMapqk:1␤␤»
lizmat hmmm… seems $*OS is unspecced ? 18:43
tadzik I think so 18:44
lizmat r: say $*OS # should this be $*DISTRO ?
camelia rakudo efc4ef: OUTPUT«linux␤»
lizmat r: say $*OSVER 18:45
camelia rakudo efc4ef: OUTPUT«2.6.32-5-amd64␤»
lizmat also not specced 18:46
TimToady: is this an implementation problem, or a spec problem?
18:46 spider-mario left 18:47 spider-mario joined
diakopter camelia needs auto-complete 18:49
*giggle* 18:50
lizmat is there a way to freeze a compile time value into serialized code ?
diakopter if you put it into a begin block and compile it to a .pbc then it's there 18:51
18:52 bluescreen10 left
lizmat but wouldn't that compile the *code* to get at the value ? 18:53
in other words: would "my $?OS = BEGIN $*OS" serialize the value of $*OS, or the opcode to get the value of $*OS ?
you're saying the former ? 18:54
diakopter no, it compiles the resulting value; it runs the begin block at compile time
lizmat ok, cool, so it *would* be the way to implement *?OS
diakopter then during its "fake compile time" (loading the module) it loads it
lizmat *and* it would be a reason for an "is readonly" variable attribute after all as well 18:55
no, of course not
my $?OS ::= BEGIN $*OS
if I remember the spec right :-) 18:56
diakopter I thought readonly was implemented with a scalar with an overridden no-op STORE
18:56 colomon joined
diakopter or exception-throwing STORE 18:56
TimToady you probably want: constant $?foo = stuff 18:59
lizmat indeed
TimToady since $?foo is supposed to be compile-time constants, generally 19:00
lizmat just finding the tests for that :-)
TimToady and ::= is only readonly, and can change meaning on the next invocation
lizmat ok, clear
TimToady: wrt $*OS and $*OSVER, are the not specced 19:01
or are they prematurely implemented ?
diakopter r: constant $a ::= END { 4 }; say $a; END { say $a };
camelia rakudo efc4ef: OUTPUT«Nil␤Nil␤»
TimToady premature
lizmat does it make sense to keep tests for $*OS and $*OSVER around ? 19:03
TimToady $*KERNEL.name == 'Linux'; $*KERNEL.version = '3.8.0-29-generic'; $*DISTRO.name = 'Ubuntu'; $*VM.name = 'JVM', etc
nobody's really implemented the spec yet 19:04
lizmat the mythical Application object, I gather from the spec
TimToady Application?
"OS" is another one of those underspecified concepts 19:05
conducive to sloppy thinking 19:06
19:06 sjn joined
TimToady probably just about as bad as "length" 19:06
19:07 dagurval_ is now known as dagurval
lizmat S28:86 19:07
synopsebot Link: perlcabal.org/syn/S28.html#line_86
lizmat wrt Application
Variable Spec Type Description 19:08
19:08 cognominal joined
lizmat $?DISTRO S02 Application # Which OS distribution am I compiling under 19:08
19:09 skyheights joined
TimToady I don't remember naming them that, and certainly the name is negotiable 19:09
lizmat seems masak was responsible for that lnie 19:10
line
FROGGS a masakbug! finally! 19:11
19:12 spider-mario left 19:13 spider-mario joined
Faxmachinen Is there any way to X+ lazily? 19:15
FROGGS Faxmachinen: use binding
TimToady those are all, on one level or other, the "Operating System", so maybe OS or System is the type
FROGGS r: my @result := ^Inf x+ ^Inf; say "alive" 19:16
TimToady X is supposed to be lazy in its slower dimension
19:16 skids joined
camelia rakudo fe289b: OUTPUT«Type check failed in binding; expected 'Positional' but got 'Str'␤ in block at /tmp/NNtIWJbR3_:1␤␤» 19:16
FROGGS r: my @result := ^Inf X+ ^Inf; say "alive"
camelia rakudo fe289b: OUTPUT«alive␤»
19:16 skyheights left
FROGGS r: my @result := ^Inf X+ ^Inf; say "alive"; say @result[0] 19:16
camelia rakudo fe289b: OUTPUT«alive␤0␤»
FROGGS r: my @result := ^Inf X+ ^Inf; say "alive"; say @result[42]
camelia rakudo fe289b: OUTPUT«alive␤42␤»
19:16 skyheights joined 19:17 SamuraiJack_ left
FROGGS hmmm 19:17
TimToady S03:4754
synopsebot Link: perlcabal.org/syn/S03.html#line_4754
Faxmachinen I tried "my @foo := ^Inf X+ ^3", but it hangs
19:18 Rotwang1 joined
FROGGS r: my @foo := ^Inf X+ ^3 19:18
camelia rakudo fe289b: OUTPUT«(timeout)»
FROGGS maybe because one side is finite
:/
Faxmachinen That can be fixed 19:19
r: my @foo := ^Inf X+ (^3 xx *)
camelia rakudo fe289b: OUTPUT«(timeout)»
Faxmachinen Hmm
TimToady r: (my @foo) := ^Inf X+ ^3; say @foo[5] 19:21
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling /tmp/XuOJMBRu0N␤Cannot use bind operator with this left-hand side␤at /tmp/XuOJMBRu0N:1␤------> (my @foo) := ^Inf X+ ^3⏏; say @foo[5]␤ expecting any of:␤ postfix␤»
Faxmachinen Now that I think about it, that wouldn't make any sense anyway
TimToady r: constant x = ^Inf X+ ^3; say x[5] 19:22
19:22 domidumont left
camelia rakudo fe289b: OUTPUT«3␤» 19:22
diakopter TimToady: now you done it :)
Faxmachinen Hehe
TimToady n: my @foo := ^Inf X+ ^3; say @foo[5] 19:23
camelia niecza v24-88-g1f87209: OUTPUT«3␤»
TimToady rakudobug on the := being eager
diakopter r: my @foo := BEGIN { ^Inf X+ ^3 }; say @foo[5] 19:24
camelia rakudo fe289b: OUTPUT«3␤»
diakopter r: say 'alive'; my @foo := BEGIN { ^Inf X+ ^5000000 }; say @foo[5]
camelia rakudo fe289b: OUTPUT«(timeout)alive␤»
diakopter wat. 19:25
19:25 jnap left
diakopter how in the world did it print alive 19:25
skids wonders if there is a writeup on buf/blob changes somewhere. The spec seems not to be updated yet.
TimToady by not reifying ^5000000 yet 19:26
benabik It reifies at the binding?
TimToady shouldn't
r: say 'alive'; my @foo := BEGIN { ^Inf X+ ^5000000 }; 19:27
camelia rakudo fe289b: OUTPUT«(timeout)alive␤»
TimToady r: say 'alive'; my @foo := BEGIN { ^Inf X+ ^5 };
diakopter r: say 'alive'; my @foo; BEGIN { @foo := ^Inf X+ ^5000000 }; say @foo[5]
camelia rakudo fe289b: OUTPUT«(timeout)alive␤»
rakudo fe289b: OUTPUT«alive␤(Any)␤»
diakopter O_O
19:28 jnap joined
TimToady r: my @foo := ^50000000; 19:28
camelia rakudo fe289b: ( no output )
grondilu I don't know what you did guys but today compilation of rakudo takes its sweet time
19:29 panchiniak joined
FROGGS grondilu: what do you mean? 19:29
TimToady we sneakily started some other processes on your machine
diakopter we slowed the rotation of the earth
also, its revolution.
grondilu FROGGS: I don't know, but I have had a compilation running for more than twenty minutes now.
FROGGS diakopter: that would be cool actually
grondilu: ohh 19:30
timotimo++ merged a optimization in :/
grondilu (and I had a segfault just before that. Don't know why)
diakopter grondilu: it's your segment's fault 19:32
blame the segment
FROGGS *g*
TimToady blames the mentation of the seg 19:33
grondilu not related: can a multi method dispatch depend on the invocant? If so, what's the syntax? 19:34
std: class { multi method foo($me: where $me ~~ /bar/) {...} } 19:35
camelia std 06636e4: OUTPUT«ok 00:00 47m␤»
grondilu r: class Foo is Str { multi method foo($me: where $me ~~ /bar/) {...} }; my Foo $b = "bar"; $b.foo;
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling /tmp/KUvBCQXrcz␤Cannot do non-typename cases of type_constraint yet␤at /tmp/KUvBCQXrcz:1␤------> multi method foo($me: where $me ~~ /bar/⏏) {...} }; my Foo $b = "bar"; $b.foo;␤ expecting any…
FROGGS r: multi sub foo($me where $me ~~ /bar/) {...}; foo("foo") 19:36
camelia rakudo fe289b: OUTPUT«Cannot call 'foo'; none of these signatures match:␤:($me)␤ in sub foo at /tmp/Hk7urMvPp5:1␤ in block at /tmp/Hk7urMvPp5:1␤␤»
FROGGS r: multi sub foo($me where $me ~~ /bar/) {...}; foo("bar")
camelia rakudo fe289b: OUTPUT«Stub code executed␤current instr.: 'throw' pc 380675 (src/gen/CORE.setting.pir:161576) (src/gen/CORE.setting:10149)␤called from Sub 'sink' pc 416236 (src/gen/CORE.setting.pir:174661) (src/gen/CORE.setting:11519)␤called from Sub 'MAIN' pc 401 (src/gen/perl6.pir:148)…
lizmat TimToady: I just realized that :exists on [] is not implemented
from the spec it is not entirely clear it should, but fwiw: :delete *is* 19:37
FROGGS r: class Foo is Str { multi method foo(\SELF: $me where SELF ~~ /bar/) {...} }; my Foo $b = "bar"; $b.foo;
camelia rakudo fe289b: OUTPUT«Type check failed in assignment to '$b'; expected 'Foo' but got 'Str'␤ in block at /tmp/ekpnwYMkKy:1␤␤»
lizmat *isn't either
19:38 kivutar joined
FROGGS r: class Foo is Str { multi method foo($me where ~self ~~ /bar/) {...} }; my Foo $b = "bar"; <--- that is your problem, not the mmd 19:40
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling /tmp/6w9yTbp8ms␤Unable to parse expression in quote words; couldn't find final '>'␤at /tmp/6w9yTbp8ms:1␤------> ; <--- that is your problem, not the mmd⏏<EOL>␤ expecting any of:␤ post…
timotimo FROGGS: why the :/ ?
oh, you think maybe i broke the build? 19:41
FROGGS timotimo: because grondilu mentioned that his rakudo build took ages
I've not tested it btw
timotimo that's weird. it got a bit faster for me for parrot and the same speed for jvm
lizmat grondilu: when testing timotimo's changes, I had to get rid of my install dir 19:42
before it worked
timotimo oh yes, you need to re-run configure*.pl in your nqp 19:43
19:44 kivutar left 19:46 zakharyas left
diakopter nqp: my $a := nqp::ctxlexpad(nqp::ctxouter(nqp::ctx)); my $b := nqp::iterator($a); while $b { nqp::deletekey($a, nqp::iterkey_s($b)) } 19:47
camelia nqp: OUTPUT«Method 'key' not found for invocant of class 'HashIterator'␤current instr.: '' pc 89 ((file unknown):165920556) (/tmp/U06KLlp1Hl:1)␤»
dalek p: 89894b0 | (Donald Hunter)++ | src/vm/jvm/ (7 files):
Added nqp::readfh on JVM. Fixes / cleanup of nqp::writefh impl.
 19:49
Faxmachinen n: say (^Inf X+ ^3).perl 19:51
camelia niecza v24-88-g1f87209: OUTPUT«(timeout)»
Faxmachinen n: say (^Inf X+ ^3)[^5].perl
camelia niecza v24-88-g1f87209: OUTPUT«(0, 1, 2, 1, 2)␤»
diakopter n: say (^Inf XXX+ ^3)[^5].perl 19:55
camelia niecza v24-88-g1f87209: OUTPUT«((0, ).list, (1, ).list, (2, ).list, (1, ).list, (2, ).list)␤»
diakopter r: say (^Inf XXX+ ^3 XXX+ ^3 XXX+ ^3)[^5].perl 19:56
camelia rakudo fe289b: OUTPUT«(0, 1, 2, 1, 2)␤»
diakopter o_O
flussence /<[XZ]>+/ is usually a no-op when dealing with auto-flattening lists 19:57
r: say (^Inf XXX+ [^3] XXX+ [^3] XXX+ [^3])[^5].perl 19:58
camelia rakudo fe289b: OUTPUT«(0, 1, 2, 1, 2)␤»
flussence hm
r: say (^Inf XXX+ [^3] XXX+ [^3] XXX+ [^3])[^15].perl
camelia rakudo fe289b: OUTPUT«(0, 1, 2, 1, 2, 3, 2, 3, 4, 1, 2, 3, 2, 3, 4)␤»
flussence r: say (^Inf XXX+ [^3] XXX+ [^3] XXX+ [^3])[80..100].perl 19:59
camelia rakudo fe289b: OUTPUT«(8, 3, 4, 5, 4, 5, 6, 5, 6, 7, 4, 5, 6, 5, 6, 7, 6, 7, 8, 5, 6)␤»
20:01 stevan_ left, stevan_ joined 20:03 xenoterracide joined 20:06 kaare_ left, stevan_ left
Faxmachinen How would one make whatever the default argument value? 20:11
FROGGS r: sub a ( $p = * ) { say $p.WHAT }; a() 20:12
camelia rakudo fe289b: OUTPUT«(Whatever)␤»
20:12 d4l3k_ joined
diakopter dalek: wb? 20:12
20:13 dalek left
Faxmachinen Hmm, you're right 20:13
20:13 eternaleye left, eternaleye joined 20:14 d4l3k_ is now known as dalek
Faxmachinen No idea how I got it to become WhateverCode 20:16
lizmat r: say Any.exists # how bad would it be if this would work and return False ? 20:17
camelia rakudo fe289b: OUTPUT«No such method 'exists' for invocant of type 'Any'␤ in block at /tmp/szhTn65xzJ:1␤␤»
Faxmachinen Oh, probably the test against whatever within the sub 20:18
I suppose smartmatch against ::Whatever would be the right way?
20:18 btyler_ joined
lizmat r: my %h; say %h<a><b>:exists # Any.exists would fix this 20:18
camelia rakudo fe289b: OUTPUT«No such method 'exists' for invocant of type 'Any'␤ in method postcircumfix:<{ }> at src/gen/CORE.setting:1908␤ in method postcircumfix:<{ }> at src/gen/CORE.setting:1845␤ in block at /tmp/fPGlAbCX5x:1␤␤»
FROGGS r: sub a ( $p = * ) { say $p ~~ Whatever }; a() 20:19
camelia rakudo fe289b: OUTPUT«True␤»
FROGGS Faxmachinen: sometimes it is easier if you paste your one-liner here so we can see what you are trying 20:20
20:20 btyler left
Faxmachinen I must be getting tired or something 20:20
Also yes
_sri developers.slashdot.org/firehose.pl...25/2115204 # "To what degree does language success correlate to beard length? It is absolutely essential. Just look at Perl's fate -- Larry Wall is just too clean-shaven. :-)"
Faxmachinen And you hit the nail on the head anyway 20:21
FROGGS me is tired too fwiw :o) 20:22
but PIR got me :/
Faxmachinen r: say (2 ~~ Whatever), (2 === Whatever); # This is sort of surprising though
camelia rakudo fe289b: OUTPUT«FalseWhateverCode.new()␤»
20:23 btyler joined
FROGGS maybe infix === does something special if Whatever is involved? 20:23
Faxmachinen I would have expected FalseFalse or WhateverCode.new()WhateverCode.new()
20:24 bluescreen10 joined
Faxmachinen It would be ~~ doing something special with whatever, since the default is to make a whatevercode, right? 20:24
20:24 skyheights left
benabik r: sub a ( $p = * ) { say $p ~~ Whatever }; a(); a(1) 20:24
FROGGS ~~ asks the RHS if it accepts the LHF, so yes
camelia rakudo fe289b: OUTPUT«True␤False␤»
benabik r: sub a ( $p = * ) { say $p ~~ * }; a(); a(1) 20:25
20:25 btyler_ left
camelia rakudo fe289b: OUTPUT«True␤True␤» 20:25
benabik Ah. Whatever accepts only *, * accepts whatever.
Faxmachinen r: say 2 ~~ Any; say 2 ~~ Whatever; 20:26
camelia rakudo fe289b: OUTPUT«True␤False␤»
Faxmachinen I probably confused those and assumed it wouldn't work
lizmat r: say Whatever.^mro
camelia rakudo fe289b: OUTPUT«WhateverCode.new()␤»
lizmat r: say WhateverCode.^mro 20:27
camelia rakudo fe289b: OUTPUT«WhateverCode.new()␤»
benabik It feels wrong that Whatever is turning into WhateverCodes.
lizmat r: say Whatever.WHAT.^mro
camelia rakudo fe289b: OUTPUT«(Whatever) (Any) (Mu)␤»
dalek rl6-roast-data: 593a487 | coke++ | / (5 files):
today (automated commit)
 20:31
[Coke] will get that CSV thing working tonight, unless $DAYJOB becomes $NIGHTJOB 20:33
20:33 benabik left 20:34 berekuk left
[Coke] (the nice csv rendering on github, that is) 20:34
Faxmachinen r: say (2 ~~ Whatever, 2 ~~ *, Whatever ~~ *, * ~~ Whatever);
camelia rakudo fe289b: OUTPUT«False True True True␤»
20:37 john____ joined 20:42 panchiniak left
lizmat r: my %h; my $b := %h<a><b>; say %h; $b = 1; say %h # this is so cool 20:42
camelia rakudo fe289b: OUTPUT«().hash␤("a" => {"b" => 1}).hash␤»
Faxmachinen Indeed
20:45 zby_home_ left
TimToady note that [^3] should *not* be flattening in a list context 20:47
20:48 Guillaume2 joined
Guillaume2 hello from Montréal 20:48
20:48 colomon left 20:49 stevan_ joined
sjohnson hi 20:49
FROGGS hi
sjohnson salut mon ami
Faxmachinen r: my @n = 1..10; my @even := @n.grep: * %% 2; say @n; @even[3] = 412; say @n; # Also this
camelia rakudo fe289b: OUTPUT«1 2 3 4 5 6 7 8 9 10␤1 2 3 4 5 6 7 412 9 10␤»
Faxmachinen Hi
Guillaume2 I only know Perl5 20:50
it was my glue language
lizmat that's a start :-)
TimToady Faxmachinen: if you want that, yeah, that's cool. otherwise, it's not so cool... :) 20:51
Guillaume2 I used to be a programmer for 12 years but now I'm studying theology
I think I wanna write bible softwares using Perl
TimToady well, Perl will definitely handle Unicode better than most languages 20:52
Faxmachinen TimToady: True. I figured it was just me who couldn't figure out how to clone an array though
Guillaume2 yes and i wanna output documents in latex
sjohnson it supports snowman bracket technology
Guillaume2 perl + latex i need
what's the GUI defacto for Perl6 ? 20:53
TimToady well, latex is just text--how hard can it be? :)
don't have any yet, and probably won't bless any particular GUI to the exclusion of any other
plus it's rather platform dependent 20:54
and we're trying to be platform independent
so the right GUI on the JVM might be different than the right one on MoarVM
or on Node.js
Guillaume2 I don't wanna write my software in Java 20:55
TimToady the whole point of putting Perl 6 on the JVM is so you don't have to
Faxmachinen Does Perl6 have a .Net compiler yet?
Guillaume2 oh Perl6 runs on a Java VM? 20:56
TimToady niecza runs on .Net
rakudo/jvm still has some rough edges, but passes 99% of the tests rakudo/parrot does already
and the startup time is abysmal, but that's comes standard on the JVM :) 20:57
slava TimToady: are there any implementations that you are aware of that have threading available? rakudo does not :(
TimToady n: say "hello from .Net-land"
camelia niecza v24-88-g1f87209: OUTPUT«hello from .Net-land␤»
TimToady rakudo does
but only on the JVM so far
slava oooooh
awwwwwwwww 20:58
Faxmachinen r: (1..5)>>.say
camelia rakudo fe289b: OUTPUT«5␤3␤1␤4␤2␤»
slava nice
TimToady that's...faking it
Guillaume2 do you absolutely need a Java VM to run Perl6 ?
slava oh :(
no
Guillaume2 ok
TimToady you just need JVM for threads currently
slava Guillaume2: there are two rakudos, rakudo-parrot, and rakudo-jvm
TimToady soon there will be rakudo/moarvm, and that'll support both threading and P5 interop 20:59
diakopter and in a couple months, rakudo-moar
slava TimToady: are there any 'big' features that are still missing? according to feature comparison, rakudo-parrot is missing threading, everything else seems to be there
TimToady well, and speed
FROGGS parrot offers threads too, but these act a bit weird 21:00
diakopter well, and a thousand or two spec'd features from the Synopses
slava FROGGS: do tell. :)
diakopter: is that in response to what major things are missing?
TimToady speed is really the only thing preventing adoption by the general public; CPAN interop is the only additional thing preventing greater adoption by the P5 community
FROGGS slava: you can create os threads, but these are not allowed to write to lexical variables
diakopter or object members 21:01
Guillaume2 ok Parrot is not Java. I understand now,
Faxmachinen Did Quantum::Superpositions actually fork?
TimToady don't think so
FROGGS slava: to do that you create a green thread within the os thread, this is the some sort of message that gets send to the main thread which writes then to lexical variables
I am trying to fiddle with the regex engine and parrot threads atm, but it is hairy 21:02
Faxmachinen Still an awesome feature though, in both languages
21:02 dayangkun left
slava FROGGS: parrot threads are not pthreads? would it make sense to have pthreads? 21:03
Guillaume2 I'm on Mac OS/X but I would like to target mobile platforms, mainly iOs and Android, do you think Perl6 will provide me tools for this soon or later ? 21:05
diakopter slava: they're OS threads like pthreads exposes
it uses pthreads on the platforms it supports I'm sure 21:06
timotimo FROGGS: don't have the concentration to backlog right now. did my nqp optimization thing end up causing the trouble?
FROGGS timotimo: I dunno
21:06 skids left
FROGGS timotimo: but I'll test now 21:06
timotimo thanks
diakopter Guillaume2: what do you mean by target?
Guillaume2 well build apps
for iOS and android
using Perl6 :)
timotimo fwiw, i did intensive testing, i did not just throw it in willy nilly. lizmat tested it successfully some time ago, too, but i rebased it
lizmat Guillaume2: in the case of android, that would require a more modern JVM on there then most have 21:07
wrt iOS: Apple does not allow apps not made with XCode 21:08
Guillaume2 thats lame from Apple
I dont know why bought a mac
lizmat that's how it is
Guillaume2 * I bought
FROGGS timotimo: I know that :o) 21:09
Guillaume2 anyone here using weechat ?
diakopter lizmat: well, you can write apps in C# using Xamarin
TimToady next year it'll be YCode instead, and everyone at the conference will be asking for an adapter because they left theirs at home
FROGGS (but some sort of fear is always with me, and that is not limited to a group of devs) 21:10
hehe
DisplayPort3000
21:10 Guillaume2 left
FROGGS but it will be thin, ultra thin 21:10
and all the cables will be made of glass 21:11
timotimo and it will no longer conflate unicode and bytes types!
slava weeee, compiling rakudo star 2013.08 with jvm :D 21:13
FROGGS slava: startup time is a mess, but you get payed back when running larger computations since JVM's optimizer is pretty awesome
slava jnthn: is the thread example as simple as the example in your blog?
FROGGS: I am not very concerned with start-up time atm :) 21:14
FROGGS k, just wanted to prevent disappointment :o9
:o)
timotimo: Stage parse : 132.725
that feels about 1% faster :o) 21:15
21:17 dayangkun joined
slava FROGGS: I see jvm as more of continued use if anythists :) 21:18
timotimo FROGGS: i'm glad ;) 21:20
dalek kudo/nom: 789db4f | (Elizabeth Mattijsen)++ | src/core/Any.pm:
Add Any.exists($key) and Any.delete($key)

This feels like a hack, but it seems the best way to handle cases such as
   my %h; say %h<a><b>:exists; say %h<a><b>:!k # False, "b"
without having to add a large number of candidates for Any:U: in combination with all possible adverbial pairs and combination of adverbial pairs.		 21:25
celesta r: state $foo; 21:26
camelia rakudo fe289b: ( no output )
celesta where can i read about the state declaration?
lizmat S03:4947 21:29
synopsebot Link: perlcabal.org/syn/S03.html#line_4947
celesta thank you
jnthn slava: See my recent talk for a lot more examples (slides linked from latest pot on my blog) 21:30
um, post :)
lizmat
.oO( jnthn been doing too many hashes lately )
 21:31
jnthn Just feeling weedy after a long day... :P 21:32
lizmat r: my %h; my $b=\%h<a><b>; say $b # huh? expected Any, or Capture, but not an empty hash
camelia rakudo fe289b: OUTPUT«().hash␤»
21:32 abnorman left
jnthn hmm :) 21:33
r: my %h; my $b=\%h<a><b>; say $b.WHAT
camelia rakudo fe289b: OUTPUT«(Capture)␤»
jnthn Apparently Capture gists interestingly...
lizmat aha!
jnthn I don't like it gisting that way, I don't think. It makes people think there's a rakudobug :P 21:34
lizmat will look at that tomorrow
:-)
jnthn :) 21:35
21:35 berekuk joined
donaldh r: qp{foo} 21:36
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling /tmp/OO2r9cZpqn␤Undeclared routines:␤ foo used at line 1␤ qp used at line 1␤␤»
donaldh r: qp{/foo} 21:37
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling /tmp/eoj99V4tsl␤Unable to parse regex; couldn't find final '/'␤at /tmp/eoj99V4tsl:1␤------> qp{/foo⏏}␤ expecting any of:␤ argument list␤ postfix␤ subscript␤ semic…
slava jnthn: ty :D
FROGGS NYI
dalek ast: 9442ea7 | (David Warring)++ | S05-capture/caps.t:
a little testing of conjunctive regex captures (&&)
 21:38
jnthn slava: Welcome :)
21:39 berekuk left
Faxmachinen Hmm. Must all ==> be on the same line? 21:39
21:39 berekuk joined
Faxmachinen r: (1..3) ==> map { $^a+1 } ==> map { $^a+1 } ==> say; 21:39
camelia rakudo fe289b: OUTPUT«3 4 5␤»
jnthn Faxmachinen: No, but maybe you are hitting the fact that } terminates a statement.
21:40 BenGoldberg joined
jnthn Faxmachinen: BUT TimToady may have rearranged the grammar to make that not an issue. I forget. 21:40
Faxmachinen I tried that on multiple lines, then I get bogus statement
21:41 lowpro30_ joined 21:42 kbaker left
jnthn Faxmachinen: Sounds familiar. 21:42
donaldh I just have some test regressions for dir with chdir emulation. dir an absolute path, get absolute paths in result
dalek ast: cb5beb6 | (Elizabeth Mattijsen)++ | S02-types/autovivification.t:
Some more test passing now that we have Any.exists
ast: ae3f5f8 | (Elizabeth Mattijsen)++ | S02-lexical-conventions/unicode.t:
Unfudge tc() tests, now that we have tc()
ast: 81aeb68 | (Elizabeth Mattijsen)++ | / (4 files):
Unfudge tests / marked as todo that seem not to blow up anymore
donaldh fix that and it could be ready to merge.
lizmat gnight #perl6! 21:43
Faxmachinen Good night
celesta night
good bye #perl6 21:44
masak 'night, cel.
21:44 celesta left, abnorman joined
BenGoldberg rn: sub postfix:<!>(Int $a) { [*] 1..$a }; say 5!; say 5.! 21:44
camelia niecza v24-88-g1f87209: OUTPUT«120␤120␤»
..rakudo fe289b: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&postfix:<.>' called (lines 1, 1)␤»
21:45 lowpro30__ joined
masak submits rakudobug 21:45
std: sub postfix:<!>(Int $a) { [*] 1..$a }; say 5!; say 5.!
camelia std 06636e4: OUTPUT«ok 00:00 48m␤»
21:46 lowpro30 left
Faxmachinen r: eval("(1..3)\n\t==> map \{ \$^a+1 \}\n\t==> map \{ \$^a+1 \}\n\t==> say;"); 21:46
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling eval_0␤Bogus statement␤at eval_0:3␤------> ⏏==> map { $^a+1 }␤ expecting any of:␤ postfix␤ statement end␤ statement modifier␤ statement modifier loop␤ pre…
21:46 Mouq joined
Faxmachinen r: eval("(1..3) ==> map \{ \$^a+1 \} ==> map \{ \$^a+1 \} ==> say;"); 21:46
camelia rakudo fe289b: OUTPUT«3 4 5␤»
flussence rn: sub postfix:<!>(Str $a) { uc $a ~ "!!!!" }; say 'hello'!; say 'hello'.!; 21:47
camelia rakudo fe289b: OUTPUT«===SORRY!===␤CHECK FAILED:␤Undefined routine '&postfix:<.>' called (lines 1, 1)␤»
..niecza v24-88-g1f87209: OUTPUT«HELLO!!!!␤HELLO!!!!␤»
flussence (just making sure niecza wasn't reading "5." as a number)
BenGoldberg rn: sub postfix:<!>(Int $a) { [*] 1..$a }; sub postfix:<!>(Str $a) { uc $a ~ "!!!!" }; say 5!, 'hello'!;
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling /tmp/WqYL4op97K␤Redeclaration of routine postfix:<!>␤at /tmp/WqYL4op97K:1␤------> b postfix:<!>(Str $a) { uc $a ~ "!!!!" }⏏; say 5!, 'hello'!;␤ expecting any of:␤ postfix␤ stat…
..niecza v24-88-g1f87209: OUTPUT«===SORRY!===␤␤Illegal redeclaration of routine 'postfix:<!>' (see line 1) at /tmp/s3oKhK4wss line 1:␤------> >(Int $a) { [*] 1..$a }; sub postfix:<!>⏏(Str $a) { uc $a ~ "!!!!" }; say 5!, 'he␤␤Unhandled exception: Check …
BenGoldberg rn: method postfix:<!>(Int $a) { [*] 1..$a }; method postfix:<!>(Str $a) { uc $a ~ "!!!!" }; say 5!, 'hello'!; 21:48
camelia rakudo fe289b: OUTPUT«Useless declaration of a has-scoped method in mainline␤Useless declaration of a has-scoped method in mainline␤===SORRY!=== Error while compiling /tmp/BKiDqai5aM␤Cannot negate , because it is not iffy enough␤at /tmp/BKiDqai5aM:1␤------> :<!>(S…
..niecza v24-88-g1f87209: OUTPUT«===SORRY!===␤␤Methods must be used in some kind of package at /tmp/jYkfHhATYv line 1:␤------> method postfix:<!>⏏(Int $a) { [*] 1..$a }; method postfix:<␤␤Methods must be used in some kind of package at /tmp/jYkfHhATYv …
21:48 lowpro30_ left
BenGoldberg rn: multi postfix:<!>(Int $a) { [*] 1..$a }; multi postfix:<!>(Str $a) { uc $a ~ "!!!!" }; say 5!, 'hello'!; 21:49
camelia rakudo fe289b, niecza v24-88-g1f87209: OUTPUT«120HELLO!!!!␤»
donaldh code reviews welcome: github.com/donaldh/rakudo/compare/chdir-emu
Faxmachinen jnthn: Yeah, it works when ==> is last on the previous line, rather than the map block
Mouq ding
yoleaux 07:46Z <FROGGS> Mouq: What if <unsp> unspaces more than one char? It could unspace \s+, wouldn't that be better?
21:49 berekuk left 21:50 lowpro30__ left
FROGGS Mouq: dont read that message :o) 21:51
jnthn Faxmachinen: Yeah, it bugged me too. I vaguely recall the standard grammar may have been tweaked after I last ran into it, I'd have to check the history to be sure.
Faxmachinen: If it did, Rakudo didn't catch up yet.
Mouq :) kk
donaldh 'night #perl6 21:52
21:53 donaldh left 21:54 SmokeMachine joined
notjack /quit 21:57
21:57 notjack left 21:59 colomon joined 22:00 bruges_ joined 22:01 _jaldhar left 22:02 bruges left 22:03 Rotwang1 left
Faxmachinen r: ("foo bar baz".split(/ /)>>.comb(/./)).perl.say 22:04
camelia rakudo fe289b: OUTPUT«===SORRY!=== Error while compiling /tmp/vV4gHu77up␤Null regex not allowed␤at /tmp/vV4gHu77up:1␤------> ("foo bar baz".split(/ /⏏)>>.comb(/./)).perl.say␤»
Faxmachinen r: ("foo bar baz".split(/\s/)>>.comb(/./)).perl.say 22:05
camelia rakudo fe289b: OUTPUT«(("f", "o", "o").list, ("b", "a", "r").list, ("b", "a", "z").list)␤»
Faxmachinen r: ("foo bar baz".split(/\s/)>>.comb(/./)).tree.perl.say
camelia rakudo fe289b: OUTPUT«LoL.new(("f", "o", "o").list.item, ("b", "a", "r").list.item, ("b", "a", "z").list.item)␤»
Faxmachinen Why does that make a LoL instead of a tree?
jnthn There's no tree type; .tree is for making a LoL 22:06
diakopter you're a tree
jnthn so's your mom 22:07
masak 'night, #perl6
jnthn OK, teaching tomorrow, I should sleep too
22:07 spider-mario left
jnthn & 22:08
Faxmachinen r: ("foo bar baz".split(/\s/).map({ [$^a.comb(/./)] })).perl; 22:10
camelia rakudo fe289b: ( no output )
Faxmachinen r: ("foo bar baz".split(/\s/).map({ [$^a.comb(/./)] })).perl.say;
camelia rakudo fe289b: OUTPUT«(["f", "o", "o"], ["b", "a", "r"], ["b", "a", "z"]).list␤»
Faxmachinen That makes a different thingy
This is so confusing
22:10 PacoAir left
Faxmachinen So is that a LoA? 22:11
22:11 john____ left
Faxmachinen r: say ((1..3) Z, (1..3)).tree.perl; # And this makes a LoA instead of a LoL 22:13
camelia rakudo fe289b: OUTPUT«([1, 1], [2, 2], [3, 3]).list␤»
Faxmachinen Oh well, bedtime 22:14
22:14 Faxmachinen left, bluescreen10 left 22:18 Rotwang1 joined, Psyche^ left 22:19 abnorman left 22:23 colomon left 22:32 abnorman joined 22:42 abnorman left 22:44 Rotwang1 left 22:45 _jaldhar joined 22:46 berekuk joined, abnorman joined 22:47 dmol left, btyler left 22:57 jnap left 23:07 skids joined 23:08 btyler joined 23:10 colomon joined 23:13 FROGGS left 23:23 FROGGS joined 23:32 colomon left 23:35 ajr_ left, berekuk left 23:36 btyler left, virtualsue joined 23:37 Teratogen joined
Teratogen howdy howdy howdy 23:37
23:42 berekuk joined, Mouq left 23:44 berekuk left 23:45 cosimo left, cosimo joined 23:47 Ayiko left 23:48 FROGGS left, Util joined 23:51 Mouq joined 23:59 FROGGS joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!