»ö« | perl6.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot usage: 'perl6: say 3;' or rakudo: / pugs: / std: , or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by wolfe.freenode.net on 30 October 2009.
00:00 takadonet left
pmurias TimToady: why does viv require both YAML::Syck and YAML::XS? 00:02
00:04 _jaldhar_ joined
pmurias gets the same STD bug in a independent checkout with a different perl... :( 00:05
could someone svn co svn.pugscode.org/pugs/src/perl6 x;cd x;make and nopaste the output of perl viv -e 'self' 00:07
sleep& 00:13
00:13 pmurias left 00:22 FCO joined 00:23 SmokeMachine left 00:29 lest_away is now known as lestrrat 00:45 colomon joined 00:47 ascent joined 00:57 ascent__ left 01:14 ruoso left
colomon Does smartmatch do something smart with hashes and keys? 01:14
01:15 partisan left, parduncia left, ssm left, krunen left, cognominal left 01:16 partisan joined 01:18 partisan left 01:21 zibri left, sjohnson left, ejs joined, zibri joined, sjohnson joined
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; $i ~ "[{for @a {f($_, $i + 1)}}]" }; say f([[], [[]], []]) 01:22
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤Null PMC access in find_method('WHAT')␤in Main (file <unknown>, line <unknown>)␤
spinclad ng: sub f(@a, $i=0) { say "$i: {@a.perl}"; $i ~ "[{for @a {f($_, $i + 1)}}]" }; say f([[], [[]], []]) 01:23
p6eval ng 06c947: Method 'perl' not found for invocant of class 'Proxy'␤current instr.: 'perl6;Array;_block3443' pc 232519 (src/gen/core.pir:18562)␤
spinclad ng: sub f(@a, $i=0) { say "$i: {@a}"; $i ~ "[{for @a {f($_, $i + 1)}}]" }; say f([[], [[]], []])
p6eval ng 06c947: 0: ␤1: ␤Method 'HOW' not found for invocant of class 'Undef'␤current instr.: 'perl6;Perl6Role;ACCEPTS' pc -1 ((unknown file):-1)␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; $i ~ "[{for @a -> $a {f($a, $i + 1)}}]" }; say f([[], [[]], []]) 01:24
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤Null PMC access in find_method('WHAT')␤in Main (file <unknown>, line <unknown>)␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; for @a -> $a {f($a, $i + 1)} }; say f([[], [[]], []]) 01:25
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤1: [[]]␤2: []␤1: []␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; for @a {f($_, $i + 1)} }; say f([[], [[]], []]) 01:26
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤1: [[]]␤2: []␤1: []␤
01:26 krunen joined
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; "[{for @a {f($_, $i + 1)}}]" }; say f([[], [[]], []]) 01:27
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤Null PMC access in find_method('WHAT')␤in Main (file <unknown>, line <unknown>)␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; ~(for @a {f($_, $i + 1)}) }; say f([[], [[]], []]) 01:28
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤Null PMC access in find_method('WHAT')␤in Main (file <unknown>, line <unknown>)␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; (for @a {f($_, $i + 1)}).WHAT }; say f([[], [[]], []])
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤Null PMC access in find_method('WHAT')␤in Main (file <unknown>, line <unknown>)␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; (for @a {f($_, $i + 1)}).PARROT }; say f([[], [[]], []])
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤1: [[]]␤2: []␤1: []␤ArrayIterator␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; (map {f($_, $i + 1)}, @a).PARROT }; say f([[], [[]], []]) 01:30
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤2: [[]]␤3: []␤4: []␤List␤
spinclad aha, so map fails, for is ok. 01:31
rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; my @b = for @a {f($_, $i + 1)}; say @b.PARROT; @b }; say f([[], [[]], []]) 01:32
p6eval rakudo 8dc189: Confused at line 2, near "{f($_, $i "␤in Main (file <unknown>, line <unknown>)␤
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; my @b = (for @a {f($_, $i + 1)}); say @b.PARROT; @b }; say f([[], [[]], []])
01:32 cognominal joined, ssm joined
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤Perl6Array␤1: [[]]␤2: []␤Perl6Array␤Perl6Array␤1: []␤Perl6Array␤Perl6Array␤ 01:32
spinclad rakudo: sub f(@a, $i=0) { say "$i: {@a.perl}"; my @b = map {f($_, $i + 1)}, @a; say @b.PARROT; @b }; say f([[], [[]], []]) 01:33
p6eval rakudo 8dc189: 0: [[], [[]], []]␤1: []␤Perl6Array␤2: [[]]␤3: []␤Perl6Array␤Perl6Array␤4: []␤Perl6Array␤Perl6Array␤␤
spinclad ng: sub f(@a, $i=0) { say "$i: {@a.perl}"; my @b = (for @a {f($_, $i + 1)}); say @b.PARROT; @b }; say f([[], [[]], []])
p6eval ng 06c947: Method 'perl' not found for invocant of class 'Proxy'␤current instr.: 'perl6;Array;_block3443' pc 232519 (src/gen/core.pir:18562)␤
spinclad ng: sub f(@a, $i=0) { say $i, @a.perl; my @b = (for @a {f($_, $i + 1)}); say @b.PARROT; @b }; say f([[], [[]], []]) 01:34
p6eval ng 06c947: Method 'perl' not found for invocant of class 'Proxy'␤current instr.: 'perl6;Array;_block3443' pc 232519 (src/gen/core.pir:18562)␤
spinclad ng: sub f(@a, $i=0) { say $i, @a; my @b = (for @a {f($_, $i + 1)}); say @b.PARROT; @b }; say f([[], [[]], []])
p6eval ng 06c947: 0␤1␤Method 'HOW' not found for invocant of class 'Undef'␤current instr.: 'perl6;Perl6Role;ACCEPTS' pc -1 ((unknown file):-1)␤
01:34 parduncia joined
spinclad ng: sub f(@a, $i=0) { say $i, @a; my @b = map {f($_, $i + 1)}, @a; say @b.PARROT; @b }; say f([[], [[]], []]) 01:35
p6eval ng 06c947: 0␤Could not find non-existent sub &map␤current instr.: '&f' pc 217 (EVAL_1:59)␤
spinclad ng: sub f(@a, $i=0) { say $i, @a; my @b = gather for @a {take f($_, $i + 1)}; say @b.PARROT; @b }; say f([[], [[]], []]) 01:36
p6eval ng 06c947: 0␤1␤Method 'HOW' not found for invocant of class 'Undef'␤current instr.: 'perl6;Perl6Role;ACCEPTS' pc -1 ((unknown file):-1)␤
spinclad all right, enough spam now. finding: map confuses $i's scope somehow. 01:38
01:43 _jaldhar_ left, trondham joined 01:45 jferrero left
colomon spinclad: I don't have time to go back through that all at the moment, but you should know that $_ doesn't work properly with map in ng. 01:49
01:51 _jaldhar_ joined 01:55 vasuvi joined, parduncia left, ssm left, cognominal left
vasuvi Out of curiousity, is Pugs still alive, or has everybody switched to Rakudo? 01:55
02:09 nbrown04 joined 02:10 nbrown04 left, parduncia joined, cognominal joined, ssm joined 02:14 synth left, pugs_svn left, mdxi left, hugme left, jql left, Infinoid left, kcwu left, gbacon left, hicx174 left, cj left, REPLeffect_ left, s1n left, dukelet0 left, mtve left, akl left, RichiH left, tomaw_ left 02:18 RichiH joined, REPLeffect_ joined, synth joined, cj joined, kcwu joined, mdxi joined, hicx174 joined, Infinoid joined, pugs_svn joined, gbacon joined, akl joined, jql joined, tomaw_ joined, dukelet0 joined, s1n joined, hugme joined, mtve joined, irc.freenode.net sets mode: +vv pugs_svn hugme 02:20 RichiH left 02:21 RichiH joined
spinclad colomon: i don't see map at all in ng at present. 02:21
vasuvi: pugs still has some life as a set of haskell modules; in the pugs repo it's gone completely quiet these days. 02:23
colomon: for my finding i should say: in rakudo HEAD, map confuses $i's scope somehow. 02:24
02:27 _jaldhar_ left 02:30 trondham left, _jaldhar_ joined
vasuvi sinclad: ok, thanks for the info! 02:34
*spinclad that is
diakopter vasuvi: it's maintained on hackage
vasuvi okies
02:34 diakopter sets mode: +ooo colomon spinclad TimToady 02:42 partisan joined 03:16 mikehh_ joined, mikehh left 03:23 mberends joined 03:32 ejs left 03:41 araujo left 03:43 justatheory left, araujo joined, araujo left 03:45 araujo joined
lisppaste3 mberends32 pasted "self.VAST for pmurias" at paste.lisp.org/display/92668 03:47
03:48 araujo left 03:49 araujo joined 04:01 Baggio_ joined 04:17 meppl joined
colomon spinclad: watch. :) 04:19
ng: (1...11).map({$^a * 3}).perl.say
p6eval ng 06c947: (3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33)␤
04:21 colomon left 04:42 araujo left 04:43 araujo joined
pugs_svn r29398 | mberends++ | [src/perl6/viv] remove the apparent fossil 'use YAML::XS' from line 11, the 'make test' passes the same number of files without it 04:44
04:45 Baggio_ left
spinclad colomon: ah, .map but not &map. 04:54
ng: sub f(@a, $i=0) { say $i, @a; my @b = @a.map: {f($_, $i + 1)}; say @b.PARROT; @b }; say f([[], [[]], []]) 04:55
p6eval ng 06c947: 0␤Array␤
spinclad ng: sub f($a, $i=0) { say $i, $a; my @b = $a.map: {f($_, $i + 1)}; say @b.PARROT; @b }; say f([[], [[]], []]) 04:56
p6eval ng 06c947: 0 ␤Array␤1Mu()␤
pugs_svn r29399 | Darren_Duncan++ | S06 : typo fix
spinclad ng: sub f($a, $i=0) { say $i, $a.perl; my @b = $a.map: {f($_, $i + 1)}; say @b.PARROT; @b }; say f([[], [[]], []])
p6eval ng 06c947: Method 'perl' not found for invocant of class 'Proxy'␤current instr.: 'perl6;Array;_block3532' pc 233183 (src/gen/core.pir:18815)␤
spinclad finding: something lacking in array argument passing... 04:59
mberends argh! I've broken 'viv', even though 'make test' was happy :( 05:01
diakopter mberends: oh yeah; I should've mentioned 05:02
sjohnson :) 05:03
diakopter for some reason TimToady (or anyone else) can't discern, XS reads the frozens better (without crashing) than Syck
or something somewhat close to that
mberends looks for a way to run viv without two YAML emitters 05:04
diakopter or 3 readers! ;)
just write your own emitter... make it 3
mberends heh, it would be faster, like with your ToJS.pm Occam Razor 05:05
diakopter but srsly, I wouldn't try to remove the XS dependency if I were you... 05:06
mberends :(
diakopter it works
mberends the crash is not with reading, it's viv:93 missing Dump 05:07
diakopter oh?
mberends it only sometimes goes there, not during 'make test' it seems
diakopter I bet if you make clean then make test... 05:08
mberends ok, I'll try that again, thought I had anyway
keeping that mberends64 user out of it this time ;) 05:09
05:16 araujo left
mberends running viv make clean && make test on an eeePC takes some time, the ./tryfile STD.pm and the first few t/spec file have passed 05:18
diakopter: ... you were saying?
05:19 Baggio_ joined, araujo joined 05:21 araujo left 05:22 araujo joined
diakopter hm 05:24
so, I lost the bet :P
mberends we hadn't shaken on anything yet 05:25
YAML::Syck also exports a Dump(), but its use is only on line 1601
05:26 araujo left
mberends moving to the top is obviously worth a try, I'll ask mberends64 then 05:27
05:27 araujo joined 05:29 araujo left 05:30 araujo joined
mberends 'make test' with use YAML::Syck at the top instead of YAML::XS happily chugs away through t/spec 05:33
is there something significant in the fact that pugs/t/spec contains 666 test scripts? 05:38
05:38 araujo left
allbery_b *eyeroll* 05:38
05:45 araujo joined
diakopter hm 05:46
666 in base10, I assume?
mberends yes, it doesn't look so ominous in binary 05:47
er, 10 in which base do you mean? ;)
diakopter heh 05:48
mberends
.oO( there are only 10 kinds of people... )
diakopter in base666, that's 666 people 05:49
mberends ah! :)
diakopter tries to understand for the life of himself, how the gazelle I got jsmeta to detect mutually recursive patterns 05:51
I guess I shoulda written a comment or two
cuz it works 05:52
but I don't... um... remember... how it works.
mberends sniggers
diakopter and the code is..... not useful for that purpose.
mberends the fact that the 'use YAML::Syck;' was in a heredoc does worry me, but moving it to line 11 works everywhere now. Perhaps if someone had commented on the YAML issues in viv... 05:54
diakopter 'course, jsmeta wasn't compiling to a textual language...
hrm
so it encountered the mutual recursion lazily, by name 05:55
<sigh>
mberends: so... it works now?
mberends *lightbulb* I'll write a comment in viv!
diakopter hopes svn blame doesn't show diakopter for that line
mberends diakopter: works, fsvo works 05:56
diakopter oh
mberends the testing may not do the right kind of testing
05:56 Baggio_ left, diakopter sets mode: +o mberends 05:57 diakopter sets mode: +o sjohnson
mberends works == worksforme atm 05:57
diakopter but is it as fast as XS for loading?
diakopter un-asks that question; sry
mberends it seemed fast enough
Syck should as fast as they get 05:58
*be
diakopter oh
hm
oh, now I remember how it worked 05:59
mberends but most of these fast YAML libs dive into your native data structure (eg P5's) and don't return until the whole graph is processed
:) quick, write a comment! 06:00
diakopter public record by ilbot2 irclog moritz_++...
mberends convenient 06:01
diakopter speaking of irclog, /me feels drawn to lurk on the past 2 weeks of #parrotsketch and #parrot
diakopter blinks back tears and snorts of rage at purl's noise 06:03
mberends I have little to say there, so I generally lurk those via irclog too
06:03 vasuvi left
diakopter mberends: btw, did you know mono has an llvm backend? it's not complete (some major-ish bugs and holes), but it largely works 06:04
mberends diakopter: yes I'd read about it briefly, but not figured out when it would make sense to use the combo 06:05
VM-on-VM seems wasteful
diakopter as a JIT 06:06
to machine code
mberends JIT-on-JIT as well seems wasteful 06:07
diakopter instead of mono's JIT
mberends ok, only one JIT. phew.
diakopter didn't miss anything in #parrot... now for #parrotsketch 06:08
just F-bombs from two people on #parrot... odd. 06:10
diakopter goes back to typing out-loud my recollection of discovering mutual recursion in jsmeta 06:11
mberends plumage is fairly significant for Perl 6, it may take over from proto
diakopter jsmeta tries to compute a regex pattern-string for the pattern, and it keeps track of the stringifications of the patterns it's entered (HashSet<string>, effectively), and if it enters the same one again, it's recursing, and returns null. 06:13
in the C# edition's case, it would be more streamlined to catch a thrown exception at the proper time 06:14
mberends so cached results from previous evaluations
diakopter instead of doing a bazillion null checks like jsmeta was doing
hrm 06:15
msdn doc fail: HashSet<T>.UnionWith : Modifies the current HashSet<(Of <(T>)>) object to contain all elements that are present in both itself and in the specified collection. msdn.microsoft.com/en-us/library/bb342097.aspx
erm, I think they mean "... that are present in either itself or in ..." 06:16
mberends can you be sure exceptions as flow control does perform better? It's implementation dependent.
diakopter streamlined for code readability, I mean
it's just compile-time 06:17
06:18 SmokeMachine joined 06:19 FCO left
diakopter resignedly but tentatively decides to use the same null-checking/string-returning recursion-detection as jsmeta, since it's known to work 06:30
mberends good. exceptions as a substitute for flow control generally looks suspicious. 06:31
diakopter makes a note to learn that lesson someday ;) 06:33
diakopter has an insightful realization about something 06:34
mberends cringes at a Perl 5 app that used 'die' throughout instead of 'return' for some trivial reason 06:35
06:35 jjx joined
diakopter :) 06:35
sjohnson mberends: maybe the coder thought it was catastrophic if it couldn't return anything 06:37
06:37 jjx left
sjohnson i have written a few dies myself for that reason... but only in those circumstances 06:37
mberends sjohnson: :) 06:38
diakopter mberends: actually I don't need string vs null; I just need true vs false
oh good 06:39
06:43 wolf2k_ubuntu left 06:56 Baggio_ joined 07:05 Baggio_ left 07:14 Baggio_ joined
pugs_svn r29400 | mberends++ | [src/perl6/viv] revert r29398 because of collateral damage, and leave a comment to warn other prospective shavers about this yak 07:22
07:27 Baggio_ left, xinming_ is now known as xinming 07:49 nbrown left 08:18 xinming_ joined 08:19 wayland76 left, wayland76 joined 08:21 mberends left 08:33 Baggio_ joined 08:36 xinming left 08:48 Baggio__ joined 08:52 Baggio_ left 08:53 tenma joined 08:55 tenma left 09:07 FalseVinylShrub joined 09:08 Baggio__ left 09:19 mberends joined 09:25 ejs joined 09:32 kaare joined, wayland76 left, kaare is now known as Guest23543 09:34 soupdragon left 09:37 xinming_ is now known as xinming 09:38 trondham joined 09:46 vorner joined 09:55 ejs left, wayland76 joined 09:56 Su-Shee joined
Su-Shee good morning 09:57
09:57 iblechbot joined 09:58 Baggio_ joined 09:59 xinming_ joined
mberends good morning 10:00
Su-Shee moritz, how have you changed. ;) usally he's the one awake. ;) 10:02
mberends the festive days probably played games with our various sleep cycles ;) 10:03
Su-Shee indeed. I went in one night from getting up at 6.30 to 10.30 ;) 10:04
mberends is already on today's second waking, perhaps there'll be a third 10:05
Su-Shee children or napping? ;) 10:06
mberends napping :) some call it "thinking" ;)
10:07 ejs joined
Su-Shee ah. thinking I do while walking. also ironing sheets is great for concentration and contemplation ;) 10:08
10:18 xinming left 10:29 mkb_ joined 10:30 diakopter sets mode: +ooo moritz_ mberends Su-Shee 10:33 masak joined
masak o/ 10:33
phenny masak: 26 Dec 18:37Z <colomon> tell masak Why did you take the Int constraint off of .pick?
masak phenny: tell colomon I took the Int constraint off of .pick because it means the caller has to do one less explicit conversion if the value sent in happens to be a Num or a Str, or the side effect of numifying a List. 10:35
phenny masak: I'll pass that on when colomon is around.
mberends \o masak
masak mberends: o/
mberends: I made great strides with GGE yesterday on the bus. 10:36
mberends \o/
masak the funniest things to implement are things marked TODO in PGE. :)
10:36 colomon joined
masak rakudo: say "foo" ~~ /<[d..b]>? foo/ 10:37
p6eval rakudo 8dc189: foo␤
masak GGE dies on the reverse range in the enum char class.
ooh!
masak submits rakudobug 10:38
:P
ng: say "foo" ~~ /<[d..b]>? foo/
p6eval ng 06c947: foo␤
masak heh.
spinclad expects one can see through a TODO more readily in rakudo than PIR 10:39
masak spinclad: aye. 10:40
spinclad i read the TreeSpider change over night, it reads very nicely. 10:41
masak spinclad: still, it would be wonderful if either (1) I got the tuits to 'backport' the features/fixes to PGE, (2) someone else did, or (3) the GGE source informed pmichaud++'s decisions about PGE in some way.
spinclad: thanks! :)
I already have a few extra GGE tests, as well as some improvements to the existing tests, that I think can be backported to PGE without hesitation. still, I'm going to ask pmichaud first. 10:43
mberends masak: hopefully you'll keep the valuable bits of the commit message (ie most of it) within the GGE source itself :)
masak mberends: that's a good idea. 10:44
spinclad me thinks it would be good if features/fixes could at least be merged into nqp-rx, as PGE's successor. i don't really expect pmichaud to devote much to PGE henceforth, somehow.
masak I still haven't really figgered out when to do documentation. right now I'm so caught up in coding that I don't stop to think about it. 10:45
10:45 trondham left
masak spinclad: nod. PGE hasn't reached its end-of-life yet, though. Rakudo still uses it, as well as most other HLLs on Parrot. 10:45
spinclad but (1) or (2) would be good, for projects still using PGE and not ready to switch.
aye, rakudo until ng lands 10:46
masak aye. 10:47
about five days from now, according to pmichaud's estimate ('in about 10 days') five days ago. 10:49
10:49 lestrrat is now known as lest_away
masak unless that's 10 days of actual coding. 10:49
diakopter or 240 hours of actual coding
10:49 Baggio_ left
masak :) 10:50
mberends masak: that triggers a reminder to discuss the status of the proto installed-modules branch. It seems workable enough to come master now, unless you have reservations. 10:54
masak mberends: hold on just a bit. 10:55
mberends fine
masak mberends: irclog.perlgeek.de/perl6/2009-12-12#i_1828488 10:57
far as I know, that last conflict was never resolved.
10:57 abra joined 10:59 lest_away is now known as lestrrat 11:06 ejs left 11:08 iblechbot left 11:13 pmurias joined 11:14 wayland76 left, trondham joined
mberends it seems packages using proto are blocking the installed-modules branch 11:14
masak: three potential solutions: 1) re-path or rename URI.pm in November, 2) notify ihrd++ of the naming collision, and temporarily comment out his uri project in proto's projects.list until that URI.pm is re-pathed or renamed, 3) both 1 and 2
11:14 wayland76 joined
pmurias mberends: hi 11:15
masak mberends: (1) sounds fine to me.
pmurias mberends: could you try svn co svn.pugscode.org/pugs/src/perl6 x;cd x;make and nopaste the output of perl viv -e 'self'?
masak mberends: don't know how reachable ihrd is right now. 11:16
pmurias mberends: i'm trying to hunt down the bug that stopped you from building mildew
masak mberends: the URI.pm in uri looks like a more advanced version of the one in november.
mberends pmurias: irclog.perlgeek.de/perl6/2009-12-27#i_1870447 11:17
pmurias mberends: seen that one, what i would like to try is to check if the bug reappears with a fresh checkout 11:19
mberends ok
pmurias mberends: as on my machine the output looks of that is different 11:20
and i checked with two different perls
mberends right, need just a few minutes..
11:22 trondham left 11:31 xinming_ is now known as xinming 11:48 trondham joined
lisppaste3 mberends32 pasted "viv r29400 -e 'self' output for pmurias" at paste.lisp.org/display/92672 11:52
mberends hope this helps :) 11:53
12:00 pmurias left 12:03 payload joined, RichiH is now known as RichiH_26c3 12:12 Baggio_ joined
diakopter o_o 12:22
o_.
-_.
___
masak now you can all review github.com/masak/gge/blob/master/STATUS for an up-to-date overview of how GGE is doing, and github.com/masak/gge/blob/master/COOLTHINGS for an ambitious look towards the future. 12:25
lunch &
if someone has a name suggestion for one of the futuristic projects, I'm all ears. 12:26
12:28 FalseVinylShrub left, TiMBuS left, Baggio_ left 12:33 payload left
jnthn phenny: tell particle Parrot's r43264 fixes the pbc_to_exe issue that prevented a Rakduo build completing on Win 7 with MS VC++. 12:38
phenny jnthn: I'll pass that on when particle is around.
jnthn: 26 Dec 15:10Z <colomon> tell jnthn Blocks with $_ or implied $_ don't get the correct arity at the moment in ng. It's costing us a good number of tests...
jnthn phenny: tell colomon aye, the reason it's broke is iirc due to a bug in flattening...so .arity keeps coming back with wrong answers. Or something like that. I was waiting on listy stuff getting sorted out. :-/ 12:39
phenny jnthn: I'll pass that on when colomon is around.
12:42 Baggio_ joined 12:52 abra left 12:54 abra joined 13:08 Chillance joined
mberends jnthn: \o 13:18
jnthn mberends: hi 13:30
mberends had/having a nice time with @family?
jnthn mberends: Yes. :-) 13:35
mberends: Been taking some much needed relaxation time.
mberends :-) how's the new laptop?
jnthn Much faster. 13:36
Bit lighter and thinner too. 13:37
Power adapter is a load lighter.
Got things about set up again by now.
Fixed an issue with the Win 7 Rakudo build earlier on today. 13:38
Going to do more @relax for rest of today, I expect, but at least it's ready for me to dig into Rakudo hacking again now :-)
mberends saw that, a Parrot patch. jnthn++ 13:39
13:39 david_f joined
mberends (GGE) we can trust masak++ to keep pushing back the boundaries of what's possible on current Rakudo 13:42
david_f hi - Is there a perl6 equivalent of XML::LibXML (I've done a google search and checked the perl6 IRC logs but I can't find mention of one)? 13:43
mberends 'fraid not yet, david_f. There's some XML related projects in github.com/masak/proto/blob/master/projects.list 13:45
david_f mberends: thanks - I thought I'd seen a blog about it somewhere but couldn't track it down :) 13:47
mberends github.com/krunen/xml looks a bit stale - June 2009 13:48
13:52 orafu left 13:53 orafu joined
david_f is the native call interface(?) mature enough to make calls to libxml2 directly yet? 13:53
13:53 mjk joined 13:59 Su-Shee left 14:00 ssm left, parduncia left, cognominal left, parduncia joined, cognominal joined, ssm joined 14:01 Su-Shee joined 14:02 mjk left 14:11 ssm left, parduncia left, cognominal left, parduncia joined, cognominal joined, ssm joined 14:18 idemal left
AndChat Hi I'm running rakudo on parrot (win32) and so far I am unable to create standalone executables from my perl scripts - can someone give me some pointers, please 14:20
mberends AndChat: it's Not Yet Implemented, sorry 14:24
14:26 iblechbot joined
AndChat Ok thought I was doing domething wrong :) - any idea on timeline for this feature? 14:26
14:27 idemal joined, david_f left
mberends no idea, unfortunately. Rakudo requires the Parrot Virtual Machine anyway. 14:27
14:28 idemal left, idemal joined
AndChat Thanks for the quick response 14:28
mberends :-) please enjoy the bits that do work, eg perl6advent.wordpress.com 14:29
masak mberends++ # trusty spokesguy 14:31
AndChat Just started learning P6 two nights ago - love it. Thx 14:32
masak in Soviet Russia -- and in other places as well -- Perl 6 loves you. :)
14:33 nbrown joined, IllvilJa left
AndChat ;-) lol 14:33
mberends and the Native Interface calls you
14:34 trondham left 14:46 _jaldhar_ left
colomon might as well get his message from jnthn. :) 14:52
phenny colomon: 10:35Z <masak> tell colomon I took the Int constraint off of .pick because it means the caller has to do one less explicit conversion if the value sent in happens to be a Num or a Str, or the side effect of numifying a List.
colomon: 12:39Z <jnthn> tell colomon aye, the reason it's broke is iirc due to a bug in flattening...so .arity keeps coming back with wrong answers. Or something like that. I was waiting on listy stuff getting sorted out. :-/
colomon oooo, one from masak too! 14:54
masak winks
colomon why .floor instead of prefix:<+> ?
masak I guess both would work... but .floor feels more explicit, since we actually want to pretend that we're working on an Int inside the method. 14:55
14:59 AndChat left 15:01 pmurias joined, vorner left
pmurias mberends: thanks for the paste 15:02
mberends pmurias: did it match your local AST? 15:03
15:07 patspam left
pmurias mberends: no 15:09
15:10 colomon left
mberends :( does it seem to be a difference between Perl 5.10.0 and 5.11.0? 15:10
lisppaste3 pmurias pasted "the ast i get" at paste.lisp.org/display/92681 15:11
pmurias mberends: i get it both with 5.10.0 and 5.11.0
15:14 masak` joined 15:15 pmurias left, Chillance left, masak left, xinming left, mberends left, mikehh_ left, partisan left, y3llow_ left, PacoLinux left, dalek left, avar left, slavik left, cotto left, pjcj left, buubot left, stepnem left, omega_ left, lisppaste3 left 15:16 partisan joined, dalek joined 15:18 pmurias joined, y3llow joined, masak` is now known as masak, pjcj joined
masak by the way, my latest rakudobug bounced. 15:18
I don't know if it's a one-off occurrance or a sign of a larger problem. 15:19
if it happens again, I guess I'll try to reach someone.
the error was '<rt-perl6@netlabs.develooper.com>: Sorry, I couldn't find any host named localhost. (#5.1.2)' 15:20
15:20 pmurias left, pmurias joined
rjbs guesses that your script said "HELO localhost" 15:21
masak my script?
I just sent in a bug report as usual.
15:22 pmurias left, mikehh_ joined, mikehh_ is now known as mikehh 15:23 xinming joined 15:24 mberends joined, pmurias joined, Chillance joined, PacoLinux joined, slavik joined, cotto joined, buubot joined, omega_ joined, lisppaste3 joined, irc.freenode.net sets mode: +o mberends, patspam joined 15:26 stepnem joined
rjbs "the script" then 15:26
15:26 omega__ joined
masak I'm sorry, I'm not following. maybe I know to little about email protocols to be able follow. 15:27
s/able/able to/
rjbs At the beginning of an SMTP conversation, the sender says "HELO $sender-hostname"
masak I'm certain that the email body didn't contain 'HELO localhost'. why would it? 15:28
rjbs and if your program is saying "localhost" instead of "my-actual-name.domain.com" it may be bounced as bogus
Where is the source of rakudobug?
15:28 literal joined
masak from what I can tell, the rest of the SMTP conversation doesn't contain 'HELO localhost' either. 15:28
rjbs: it's RT.
15:28 avar joined
masak rt3, to be exact. the one at perl.org. 15:29
rjbs So you submitted via the web?
masak no, that's turned off.
I sent an email to [email@hidden.address]
rjbs Okay, how did you send the email?
masak via Gmail, as I always to.
s/to/do/
15:29 Baggio_ left
masak I got the whole SMTP conversation back. would it help if I nopasted that? 15:30
rjbs Yes!
15:30 pmurias left
masak gist.github.com/264298 15:31
rjbs That contians none of the SMTP conversation.
masak oh. :/
that's all I got back, though.
rjbs It's still potentially useful, looking.
My guess, reading that, is that something has been misconfigured at Develooper (perl.org) 15:32
The message is accepted by them from Gmail, then routed around internally a bit before bouncing. 15:33
I'm guessing they used "localhost" somewhere that it isn't allowed, but it's not clear. Sorry I can't help more
masak it does help -- thanks. 15:34
obra: ping.
15:34 wolf2k_ubuntu joined 15:38 Baggio_ joined 15:40 avar left, literal left, omega__ left, patspam left, slavik left, mberends left, omega_ left, cotto left, buubot left, Chillance left, PacoLinux left, lisppaste3 left 15:45 mberends joined, patspam joined, Chillance joined, PacoLinux joined, slavik joined, cotto joined, buubot joined, lisppaste3 joined, irc.freenode.net sets mode: +o mberends 15:46 omega_ joined 15:48 am0c joined, avar joined, literal joined, abra left 15:54 abra joined 15:56 Su-Shee left, Baggio__ joined, Su-Shee joined 15:57 colomon joined, jaldhar joined 15:58 Psyche^ joined 16:00 Patterner left, Psyche^ is now known as Patterner, payload joined, Baggio_ left
masak colomon: I just backlogged over the reason for your question about .pick and Int. generalizing, fewer core methods/subs should have nominal types than one would perhaps first assume. when we get support for 'as SomeType' auto-conversion, we can encode things like '.= floor' that way. 16:01
hejki any means to figure out filesize apart from qqx{ls -l} ? 16:06
masak hejki: in Perl 5, it's -s 16:07
mberends hejki: Parrot programmers may be able to help you with stat(), which can return all fields in an array from a single call, japhb++ told me
masak hejki: the Perl 6 version of `-X $file` is `$file ~~ :X` 16:08
hejki: I've implemented a few of those. it's not very difficult.
hejki: I don't think :s is implemented, but if you want, I can guide you through it. 16:09
16:13 justatheory joined
masak confirms: only :d, :e and :f are implemented so far 16:14
but it seems to be a piece of low-hanging fruit to implement :s
hejki: so, what do you say?
hejki hmm..
hit me
:>
masak excellent. 16:15
so here's the deal. you take the clues I give you, produce a patch, and I apply it. instant win.
16:15 abra left
hejki :D 16:15
masak here are the clues: github.com/rakudo/rakudo/blob/maste...ny-str.pir www.parrotcode.org/docs/ops/io.html (the part about 'stat')
16:15 Baggio__ left
masak feel free to ask further questions if you have any. 16:16
16:16 colomon left
Su-Shee a "is size null" would be nice aka "is empty" 16:19
masak Su-Shee: then I formally extend the above deal to cover you as well. the clues are the same as above.
Su-Shee already reading it.. :) 16:20
masak :)
Su-Shee really doesn't look that difficult.
masak it really isn't.
Su-Shee a low hanging fruits list would be nice, I wouldn't even know what has to be done some day and is perfect to get into "implementing things" 16:21
masak the one thing that's perhaps easy to miss is the !EXPORT list at line 26.
hejki : 16:23
:)
16:24 payload left
Su-Shee I've tried something blindly: sushee.no-ip.org/chop.txt 16:29
masak you don't need the 'lastchar' line.
also, src/setting/Any-str.pm already contains a pure-Perl6 version of .chop 16:31
hejki how about heikki.mehtanen.fi/any-str.-s.patch
Su-Shee just playing.
masak looks
hejki (should i do any error checking?
masak Su-Shee: well, apart from that, it looks fine. 16:32
hejki (i'm not sure what language that even is :P)
masak hejki: it's PIR, Parrot Intermediate Representation. one step above Parrot assembly code.
hejki kk 16:33
masak hejki: hm, you forgot line 26...
Su-Shee :))
hejki the push_eh-part?
16:33 vorner joined
Su-Shee the export your actual stuff part. 16:33
hejki oh no.. the export :P
ye :P
masak aye.
hejki i'll fix it
done 16:34
masak hm...
I'm wondering whether we should box the Parrot int before returning it...
we probably should.
hold on, I'll see if I can find a similar case in existing code.
hm, .chars doesn't box, so we'll skip that. 16:35
16:37 jaldhar left
masak hm, the patch didn't apply cleanly... 16:37
investigating.
hejki might be with the filesnames oslt
masak hejki: I notice you didn't write any documentation... 16:38
oh wait.
it's just me being blind. :/
hejki i did.. it says Returns file size.
masak hejki++
hejki should i write more? :)
go on push it.. i need it for a project ;P 16:39
oh wait.. i guess i could apply the patch locally as well :D
masak no, this is fine.
I've applied the patch now.
I'll just take it for a spin, and then I'll push.
hejki :)
masak ah. line-ending whitespace. git++. fixing. 16:40
hejki dang
masak probably nothing you did -- you might just have a sucky editor. :P
hejki i like my vim 16:41
masak vim is OK.
I bet it can be configured not to leave garbage whitespace, too. :)
hejki it might've been cause of visual mode 16:42
copypasted the basic structure.. :)
masak ah.
sometimes autoindent causes that, too, I think.
hejki ye might be 16:43
but hope it works. i think i'm off for some tv and food. laters :)
Su-Shee is there a "trim" already?
masak rakudo: say ' foo '.trim
p6eval rakudo 8dc189: foo␤
masak hejki++ # pushed 16:46
16:47 jaldhar joined
dalek kudo/master: 77bf8cf | masak++ | src/builtins/any-str.pir:
[src/builtins/any-str.pir] implemented :s
 16:52
masak does Perl 6 have a business case? an implicit one, I mean. 17:00
Su-Shee how do you mean? 17:02
masak I was explaining the raison-d'être of Perl 6 to my father. I explained how mathematical notation had evolved over the centuries. 17:03
Su-Shee and he was asking what all this will be good for? 17:04
masak how cumbersome it was to express things that are dead simple today, before the invention of the zero, or the x^2 notation, or the equals sign.
and he asked about the business case of Perl 6.
it made me realize that he perhaps doesn't grok the economic model of open-source development. 17:05
but somewhere beneath his misconceptions about it, there's a good kernel of a question.
moritz_ masak: the question is more if *you* have a business case that involves Perl 6
masak yes, but that's another question than the one I think I'm hinting at. 17:06
Su-Shee so stop hinting and ask ;)
masak well, FOSS doesn't operate under an money-based model of profit. 17:07
still we're definitely building up a kind of economic good here.
but it's measured in some kind of community units, rather than in money.
the real worth of Perl 5 is in CPAN, and PerlMonks, and freenode. 17:08
Su-Shee yes. because using open source has very concrete and tangible advantages in the company environment. starting with not wasting money on licences and being able to adjust/fix "the software"
no, not just that.
masak oh, definitely. the freedome/libre/gratis aspects are important, too.
Su-Shee it's not about the really "nothing to pay", it's its surrounding "how to pay". 17:10
masak hm... please elaborate.
Su-Shee like "paying red hat for 5 years security releases backwards - guaranteed" 17:11
masak ah.
Su-Shee or, you can buy a DVD with handbook with one of those typical shop systems based on LAMP - costs like 300 euro. 17:12
you pay for the intelligent compilation.
masak but I'm not out to get paid. that's not why I do Perl 6. 17:13
granted, I wouldn't mind if someone threw a heap of money on me. that would probably motivate me somehow.
but I'm not in it for the money.
17:14 payload joined
Su-Shee open source always was _explicitly_ not about free beer - so the question for money earning use cases of perl 6 is not just legitimate, but also a good and clever question. 17:14
and you want the "business world recognition" 17:15
masak I wasn't implying making money is not a legitimate motivation. I'm just saying that hasn't been my motivation so far. 17:16
Su-Shee na, that's not what I'm trying to say. give me 5 sentences to explain. 17:17
masak :)
Su-Shee one of the strong suits of OSS is the real world usability in specific business scenarios where OSS is gaining different realms every decade or every five years or so.
first, it has been "web" (servers, programming languages) - that's still a real stronghold. 17:18
then came data bases - more and more companies made their db available to linux or published even OSS databases. then came the "let's make something for users" era with KDE and GNOME. 17:19
for a few years - besides web 2.0 - it's now "all things tiny" - cell phones, pdas, netbooks, younameit. 17:20
so, what's coming next? and how does perl 6 might play a role in it? that is essentially the question you'll have to explain your dad.
masak hm. 17:21
yes.
what you're saying makes a lot of sense.
Su-Shee if we look at python, than python took over from perl the realm of nlp, bio inf. and specifically the acadamic realm, where perl never was really strong. php took over "the web". so, what's the realm we would place perl 6 next year? what would be a future use case to gain users, topics and later on businesses? 17:22
masak I can see Perl 6 making the web world a better place. that's why I'm helping with the Web.pm project. 17:23
vorner Su-Shee: May I point out that one of the strongest places for OSS is development - it has much lower entry barier for people who learn, who know a little, or just want to build tools for other tasks with minimum effort
masak but I'm sure there are more realms.
Su-Shee vorner: in germany, it also has a very strong position in administration due to the independance from microsoft. 17:24
masak vorner: yes, that's a good point. most people who get interested in Perl 6 are programmers. the vast majority. tiny scripts can function as a sort of 'gateway drug' to more complex use cases.
vorner I'm not sure if perl can be good at the „learning“ part, but as I see, perl 6 will be a really powerfull tool at the tool building, system scripting, etc 17:25
Su-Shee masak: without excellent and broad and sexy web support, perl 6 won't gain anything. (or any other programming language, for that matter)
masak Su-Shee: agreed. 17:26
we're aiming towards excellent. :)
and broad.
vorner still, perl6 as is today is little impractical for „real live“ problems
masak mainly because it's slow and unstable. 17:27
mdxi what does "sexy web support" in a programming language mean? *the browser* is where all "sexy" web development is going, from my point of view
Su-Shee where I see perl 6 - and I'm really sure, that this realm will matter in very expensive and important terms - is data mining, text processing and all kinds of "analysis stuff". from "mashups" to banking to "social physics"
mdxi on the programming side, it's been reduced to "do you speak JSON?"
masak I have a hunch we'll see improvements in speed and stability during 2010.
mdxi s/programming/server/
Su-Shee mdxi: it means if it isn't fun and smooth and easy and somewhat cool, I don't use it.
masak mdxi: how about Perl 6 that can compile down to Javascript? 17:28
mdxi: then you can write your whole web application in Perl 6.
vorner masak: and it lacks libraries still, but yes, the language design is really nice
Su-Shee it'll be sexy, if it takes the _future_ of how web is done into account.
masak vorner: agreed, it lacks libraries. but that's a problem that won't be solved overnight, and not without wider adoption.
vorner masak: no argue there, sure 17:29
masak Su-Shee: could you be more specific?
Su-Shee hm, how do I say that.. 17:30
vorner Another problem I see with perl5 ‒ it can't be easilly compiled into one compact unit that a windows user could just download and run 17:31
the problem somehow does not exist on linux with distros and package managers, but many people develop applications for users in C++ just because of this 17:32
mdxi Strawberry Perl
Su-Shee masak: in 199x, noone did web MVC style. then came RoR and took the smalltalk paper (which is very good, btw) and poured it into code. and they did it "the right way" - it spoke to all the web designers (!) which had been left out by perl and php and took them at their hands and gave them something easy to deploy.
masak nod.
Su-Shee masak: so, what will be necessary, if you want to play in advanced web in 2012?
masak Success By Being Opinionated In The Dimensions Where It Matters. 17:33
Su-Shee masak: I (and this is really just my feeling) see an even more strict seperation of "gui" and "middleware". I don't see the need for handling sessions in perl anymore or to deal with templating systems.
masak nod. 17:34
Su-Shee (former due to window.name and DOM sessions, latter due to great JS frameworks which all deal with this stuff, if you do it right)
aeh, DOM storage.
vorner mdxi: does this allow a developer to create single exe that does not need to be installed, no modules spread over directories? 17:35
Su-Shee masak: so, how would you do a web framework which still has to support ancient stuff in technology terms (it still is CGI and stateless http, there's no way around it :)
17:35 jaldhar left
mdxi vorner: oh, you meant applications. i thought you were talking about the CPAN and such from a developer's point of view. 17:36
Su-Shee masak: but has to support future cool stuff like CSS3 and HTML5 and Javascript and the possibilities _those_ technologies give you?
mdxi vorner: no, it doesn't.
masak Su-Shee: good question. there has to be a staircase of opt-in complexity levels somehow. 17:37
from 'bare-bones' to 'the whole AJAX thingamajig'.
Su-Shee masak: it has more to do with "to what people will perl 6 appeal"
vorner mdxi: I know people who code in C++ just because they can deliver a single exe that will just run no matter where it is, how it got there, what libraries are installed. 17:38
Su-Shee masak: because that's what decides about style and projects and what will work first.
mdxi vorner: of course. C++ is the /lingua franca/ of Windows development
masak Su-Shee: people who expect the same flexibility as with Perl 5, but more expressive power and convenience.
Su-Shee masak: appeal to the admins, and you probably get great networking support first. appeal to the web folks and you'll get cool web stuff. 17:39
masak I'd welcome a project that would appeal to admins.
Su-Shee masak: no. brutally speaking "people" don't know perl 5 anymore nowadays.
mdxi "people" never did
masak how so?
vorner mdxi: Yes, right. Time to change it. It would be great if there was a way to compile perl6 to an .exe, bundle libraries 17:40
Su-Shee mdxi: they did massively up until 2000. the old web was made entirely of perl.
vorner GUI library that is easy to use, and so on
Su-Shee masak: because web is done with php, academics use java and python and the rise of embedded stuff has yet to chose its "scripting language". 17:41
vorner People prefer languages where it is less easy to shoot yourself into the foot. Perl requieres you know what you are doing
Su-Shee masak: in germany, there aren't even perl job besides some admin scripting anymore. 17:42
(sorry, moritz :)
masak vorner: I disagree. some people think like that. others prefer expressiveness to safety from your own mistakes.
17:42 payload1 joined, payload left
masak s/your/their/ 17:42
vorner masak: Yes, I know. That's why I use perl.
masak and the caveat is less relevant with Perl 6, too. 17:43
Su-Shee vorner is right in certain terms. technians as a mass aren't really the first folk if it comes to expressiveness :)
masak if you want, you can program quite conservatively in Perl 6.
Su-Shee technicians.
masak: yes, but you're not forced to.
masak right.
Su-Shee and you won't with p6.
vorner But many people think perl to be cryptic and hard to read
masak opt-in masochism. :)
vorner: bad Perl is. 17:44
Su-Shee vorner: I find it hard to type.
vorner There are even jokes around here that say that perl is write-only language, perl should be used to talk, not to program and that it looks the same before and after encryption
huf it's not like bad perl is worse than bad c
Su-Shee look, I wouldn't even care for the usual perl critic - focus on what the future of it will bring and place perl 6 there. 17:45
mdxi bad perl (aka "PERL", aka "CGI") is one reason i'm actually kinda glad Perl and the web aren't synonymous anymore
vorner I agree that people who want to write bad code always find a way
mdxi also, honestly, i don't think any language has, or is going to "win" the web
Su-Shee my favorite perl joke is that it is the only language which you can write programs in only made up by punctuation :)
mdxi: perl did and afterwards, php did. now ruby is somewhat dominating the pace in terms of "ideas" (not in massive jobs, interestingly) 17:46
vorner However, if perl is to be adobted widely, people must learn it. And to learn it, the myth about near-unlegibility must die, they fear it. 17:47
mdxi i disagree, except that perl definitely did early on. there's an *awful* lot of Java and even stuff like Cold Fusion out there
Su-Shee what's for example really not good in terms of "reaching the masses" is the lack of contemporary perl books. nowadays books like "xy hacks" or "foobar for beginners" plainly doesn't use perl anymore.
mdxi: that's true, I forogt java. 17:48
which means that perl 6 has to gain a foot in the business world.
mdxi Java feels like the COBOL of the web. it's what all the big players latched onto, adn now it, in its various forms, is a huge bulk of legacy code. 17:49
the modern world is a little less entrenched than in the mainframe days... but... those mainframes are still out there :)
Su-Shee yeah, well, as long as universities are churning out young, cheap computer scientists which all went through the school of java... 17:50
17:50 meppl left
mdxi which is the really sad part 17:50
vorner is glad to be on the one czech university that still has java as optional language only :-)
Su-Shee luckily, I didn't even study C.S. ;) would have been lisp and pascal and such in my days anyway. 17:51
17:51 partisan left 17:52 partisan joined
moritz_ nothing wrong with learning either language 17:52
vorner Hmm, seems like the time has stopped here, I went trough pascal, lisp, prolog and haskell (and C++) as the languages I was forced to learn
Su-Shee moritz_: comp sci in 1990 would have been sheer old school boredom. (from my perspective ;)
mdxi the whole point of languages is that they don't all solve the same problems the same way. they represent different ways to *think*, basically. a monolingual cs program is a travesty (says the guy who collects programming texts from the 1960s and 1970s)
moritz_ Su-Shee: :-)
Su-Shee anyway. luckily, nowadays we can have something nice. ;) 17:53
masak mdxi: I agree completely.
moritz_ I'm offline most of the time these days, if you want to reach me it's best by email (moritz at faui2k3.org)
happy new year if I don't manage to drop in before the 1st! 17:54
masak moritz_: happynewyear!
mdxi: but while pluralism is indeed very good, some languages are clearly better suited as tools for solving some classes of problems.
mdxi masak: of course. it follows naturally :) 17:55
and, in some cases, by explicit design
Su-Shee yeah, that was never really a decision which "businesses" took into account. ;)
masak Perl 6, while not being revolutionary in any part except perhaps grammars, definitely gives me a warm, fuzzy feeling in many respects, just by being very convenient. 17:56
Su-Shee masak: sadly, that's not what your CEO is interested in. (take CEO as a metaphor here, please)
masak nod. 17:57
the CEO sees the killer apps, but not the underlying tech.
vorner I fear one thing about it ‒ its unicode support. Perl 5, when you launch it on not-properly formated codepoints, just passes them trough not complaining. I hope this will be possible with perl 6 too.
Su-Shee and rightly so IMHO. since the web reached "the masses", they have the right not to be interested wether the cool (or shitty) UI is served by perl or something else. 17:58
masak: do you read german by any chance?
masak Su-Shee: not nearly as well as I'd like. 17:59
vorner: you'd better boil that down to a small use case in order to clarify what you mean. 18:00
diakopter invalid surrogate pairing or combining chars 18:01
Su-Shee masak: over the last months, there's popping one article after another about tech which essentially means a massive rise of data mining "for us common folks". and that's where I would place perl 6 first.
vorner masak: Imagine an IRC bot and someone just sends something that is not utf8, but something else, lets say iso-8859-2 (happens all the time)
I noticed a python bot that just drops dead on one channel when this happens
Su-Shee *hihi* 18:02
Su-Shee searches for some polish letters on her keyboard.. ;)
vorner I wouldn't like this to happen to a perl 6 bot, if I wrote one
18:06 trondham joined 18:07 jaldhar joined
masak vorner: encodings are non-trivial. Postel teaches that we should try not to die on bad input, though. and Perl has a tradition of being lenient and workaroundy. 18:09
Su-Shee I second that. great tools like "browsers" and mplayer and such show how important that is. 18:10
vorner Or another thing, imagine you want to count lines. Then you read it as a text file. Would be bad if it failed because someone didn't do their homework when setting correct encoding.
masak nod. 18:12
vorner: do you have Rakudo? have you tried to read a badly encoded file?
I can write the line-counting program for you: `say +$*IN.lines` :)
vorner I tried it some time ago, now it compiles updates from git, but I'll try soon 18:13
masak when you tried it some time ago, did it work? 18:14
vorner no, it failed 18:15
throwing an exception
I didn't make it to read it as bytestrings
it was attempt to read and parse mbox files, which turned to be really resource-consuming 18:16
and some of the files contained 8bit transfered mails with other encodings
18:17 soupdragon joined
vorner yes, it still fails 18:18
recode utf8..iso-8859-2 | ./x.p6
ž
Malformed UTF-8 string
in Main (file src/gen_setting.pm, line 324)
masak vorner: care to submit a rakudobug? 18:19
vorner Is it really a bug? Or is it a feature?
I could submit it, of course 18:20
masak when in doubt, submit. :)
vorner ok
masak make sure you include the above discussion about it possibly being a bug.
18:23 jaldhar left, jaldhar joined 18:25 payload1 left, IllvilJa joined
vorner I don't seem to be able to find a place where to put the bug. I managed to log in into some perlbug web, it allows me to search bugs, but I see no „new bug“ button anywhere :-| 18:28
masak it's all done by email. [email@hidden.address] 18:32
last time I tried that though, I got a sad message back from some mail-daemon. that may or may not have been fixed since.
vorner Is there a hepl of how such mail should look like? Or, it can be just normal email with text? 18:33
masak just a normal email with text. 18:39
the subject might have the text '[BUG]' (sans quotes) in it. that gives it the right internal tag in RT. 18:40
also, it really helps if the subject is descriptive.
18:40 mkb_ left
vorner Ok, thanks, I'll write it once I get from shower 18:41
18:41 payload joined
masak food & 18:56
18:56 masak left 19:20 colomon joined
vorner Hm, the mailer daemon gives me a really nice error :-D. „Sorry, I couldn't find any host named localhost.“ 19:25
19:29 ispeak joined 19:30 ejs joined 19:31 jferrero joined 19:37 kensanata joined 19:46 trondham left 20:34 payload left, payload1 joined 21:07 vorner left 21:12 kensanata left 21:18 ispeak left, colomon left 21:37 nbrown left 21:41 ispeak joined 21:42 mberends left 21:53 ruoso joined 21:54 trondham joined 22:14 fridim_ joined 22:17 donaldh joined, donaldh left 22:23 trondham left 22:24 ejs left 22:28 wolf2k_ubuntu left 22:29 iblechbot left 22:36 colomon_ joined 22:41 wayland76 left 22:53 ispeak left 23:08 ispeak joined 23:11 ispeak left 23:13 Su-Shee left 23:30 colomon_ left, soupdragon left 23:36 soupdragon joined
Tene anyone available to confirm some ng spectest failures? 23:36
diakopter hi 23:38
on latest parrot or something? 23:39
Tene No, on $PARROT_REVISION I'm seeing a failure in t/spec/S32-array/end.rakudo 23:40
23:42 fridim_ left 23:52 colomon joined
dalek kudo/ng: daa5600 | tene++ | src/Perl6/Compiler.pir:
Add a "load_module" method to Perl6::Compiler for HLL interop
 23:53
kudo/ng: 5e19f61 | tene++ | src/ (2 files):
Change the language the compiler is registered under to match both the HLL namespace and the filename.

Powered by Cro and the Raku® Programming Language.

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

Thank you!