»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend! | Rakudo Star Released!
Set by diakopter on 6 September 2010.
muixirt need sleep, bye 00:00
00:00 muixirt left
masak blog post! strangelyconsistent.org/blog/novemb...stuff-done 00:00
huf i love your unrelated bits of history 00:03
are you living in china or studying chinese or what's the apropo? 00:04
masak the latter.
Chinese history basically made me appreciate history :) 00:05
diakopter Tene: "There's nobody involved in Perl 6 right now that believes that they can reimplement everything required better, more-accurately, and more-performantly than current VM offerings on the same time-scale as getting a more-complete Perl 6."
Tene: but I thought that's what Parrot .... 00:06
(or even reducing "better, more-accurately, and more-performantly" to "as good, as accurately, and as performantly")
00:07 donaldh left, XaRDaX left 00:08 hercynium left
diakopter Tene: or were you seriously considering Parrot in "current VM offerings"? 00:08
masak seven years, and Parrot *still hasn't been released*! :P 00:09
diakopter ... seeing as how (if I understand correctly) it's all back to the drawing board, throwaway except in the minds of the new rewriters?
masak diakopter: as opposed to how Rakudo has been progressing? :) 00:10
diakopter touche (implicit accent)
masak ng in 2009-2010, jnthn's doing 6model now...
TGE, nqp, nqp-rx... 00:11
seems all we do around here is write and rewrite things until we get proficient at it. 00:12
diakopter well, I was really only referring to the nascent lorito, not the progress of parrot the past 7(-10?) years
00:13 justatheory left
masak it's like those Zen gardens where the gardener softly bends branches until they get the right shape. except we're bending ourself until we get the right model mindset to implement things right. 00:13
"plan to throw 3 or 4 away"
diakopter "plan to throw n-1 away" 00:14
masak :D
that's a tautology!
diakopter so's a tautology!
masak "There's such a mind-numbing lack of competence in any IT field that you don't have to be brilliant to be extraordinary, and you don't have to excel to deliver consistently high quality." -- www.xach.com/naggum/articles/316937...um.no.html 00:17
cotto_work diakopter: I'll become cotto in an hour or so if you'll be around to continue our conversation. 00:18
masak I find there's more than a grain of truth in this quote.
cotto_work: if diakopter won't be around, will you remain at work? :)
cotto_work yes
masak oh :) 00:19
cotto_work I'm really hoping he's around.
dukeleto masak: That quote is especially true in academia. You would think these people would know something. But they don't.
cotto_work ;)
diakopter cotto_work: I'm 'round
00:19 thundergnat joined
masak dukeleto: I know what you mean. I'm in it, too. 00:19
dukeleto: it's only when you find good scientific papers/articles that you realize how bad the rest really were.
cotto_work If you can't dazzle them with your brilliance... 00:20
00:21 stkowski left
Tene diakopter: I didn't really use enough context there. That was part of an answer to muixirt about why we use VMs. There's nobody sufficiently ... [such that they're working on rewriting Rakudo to emit x86 machine code and not use an existing VM]. 00:21
00:23 Chillance left
diakopter Tene: but see my followup question 00:23
dukeleto I think VM's are sufficiently complex that a single person does not have any hope doing it themselves, unless they want to spend a lifetime.
thundergnat Hi #perl6. 00:24
dukeleto thundergnat: welcome!
Tene diakopter: muixirt was saying "VMs are bad; why don't you just emit x86 machine code and leave out the VM." Parrot is a VM, so anyone working on Parrot doesn't count for what muixirt was asking.
thundergnat Non -interdependent modules shouldn't care about their load order in a script should they? 00:25
Tene thundergnat: that sounds right.
thundergnat I'm getting a bizarre error if I change the load order of two completely independent modules.
00:26 sftp left
Tene thundergnat: what error? 00:26
00:26 risou joined
masak hi thundergnat. 00:26
thundergnat: I've had such errors as well. 00:27
there's at least one (still open) bug report in RT about it.
thundergnat For my module Sort-Naturally github.com/thundergnat/Sort-Naturally in the test file t/01-basic.t, if i change the order of "use Sort::Naturally; and "use Test;" It basically doesn't load my module.
fails all tests 00:28
masak huh.
masak tries
thundergnat tested on both windows XP and ubuntu linux
masak ah; here was the one I was thinking of: rt.perl.org/rt3/Ticket/Display.html?id=65738 00:29
see if it's similar...?
diakopter Tene: so, you were including Parrot in "current VM offerings". I honestly don't see how that's a supportable claim, given the pending total rewrite of parrot (again, iiuc).
thundergnat masak: looking... 00:30
00:30 risou left
Tene diakopter: Are any parrot people working on a Perl 6 compiler that emits x86 machine code? 00:30
dukeleto Tene: not that I know of. 00:31
Tene diakopter: If not, I stand by my edit, "... such that they're working on rewriting Rakudo to emit x86 machine code ..." 00:32
diakopter Tene: not that I know of, but I don't follow your argument above about your response to muix
thundergnat masak: Hmm. I'm not sure if its the same issue or not. I'm not overriding core functions (though I am MONKEY_TYPING). 00:33
diakopter Tene: after re-re-re-re-reading, I think I see what you're saying 00:34
masak thundergnat: ran test file unmodified. everything passes.
making prescribed change now.
Tene diakopter: "<muixirt> VMs suck! Do it without a VM!" "<tene> I guess maybe? Nobody here agrees enough with you that they're working on Perl 6 without a VM."
That's all I meant. Perhaps a spoke poorly. 00:35
masak thundergnat: now it passed two tests and then "Could not find sub &nsort"
diakopter no, I did understand you correctly
masak thundergnat: either way, I would definitely say you have a bug here.
thundergnat Yeah, thats what I'm seeing. 00:36
masak thundergnat: care to report it?
thundergnat I'm not sure how small a test case I can golf that down to though.....
masak I'm on it right now :) 00:37
00:39 pmb left 00:41 impious left
masak I'm getting odd results. lib/Sort/Naturally.pm6 compiles fine when I remove everything after the 'augment class Any' block, but t/01-basic.t dies with a parse error. (even an unmodified t/01-basic.t) 00:41
I can't make sense of that.
thundergnat I don't guarantee that I haven't done something completely insane. 00:42
masak no, it looks ok. 00:43
thundergnat++
thundergnat: btw, you've got "conceivably" right, but seem to consistently misspell "yields". 00:46
thundergnat masak: sigh
masak I'm still stumped as to what causes the syntax error. seems to be something on the next line, but beyond that, I don't know. 00:47
oh! must be the operator.
right. I can't just remove that from the .pm file. silly me.
thundergnat: I have it golfed now. hang on. 00:52
thundergnat: gist.github.com/701261 00:53
00:55 QinGW joined 00:57 mikehh joined
mikehh ok how do you build rakudo with the latest parrot release - it reports ===SORRY!=== Parrot revision RELEASE_2_9_1-261-g495feaf required (currently RELEASE_2_9_1-273-gae3500c) 00:59
masak that looks... wrong. :/
mikehh I used perl Configure.pl --parrot-config=/home/mhu/install/lib/bin/parrot_config 01:00
masak mikehh: Parrot recently moved over to git, and Rakudo's configure systems were changed even more recently to accomodate for this.
from what you're saying, I think there's a bug somewhere. 01:01
thundergnat masak: sorry, afk for a bit; Cool I'll go ahead and submit that if you haven't already
masak thundergnat: yes, please do.
mikehh: moritz_, who's the one likely to know more about this, went to bed about two hours ago.
mikehh masak: yeah, I should too, but will check with him later 01:03
masak: was testing parrot for the release on the morrow, but thought I would check rakudo and it gave me that problem 01:05
masak mikehh: thanks for checking.
I'm glad to hear about this too, since I'm the release manager for Rakudo on Thursday :) 01:06
masak heads to bed
'night, #perl6!
Tene 'night
thundergnat night masak, thanks
mikehh cu, mee too
masak oh, right "n8" in German would be "nacht". that makes even more sense. 01:07
n8 :)
01:07 masak left 01:11 jedai joined 01:13 rgrau left 01:39 thundergnat left
colomon ping jnthn 01:52
02:09 Teratogen joined 02:10 knewt2 joined 02:16 jaldhar joined 02:33 molaf_ joined, mariano left 02:34 ch3ck3r joined 02:37 molaf left 02:39 ch3ck left
colomon phenny: tell jnthn nqp-rx / nom branch has spaces where there should be tabs in Makefile.in, then fails during the build. suggestions? 02:39
phenny colomon: I'll pass that on when jnthn is around.
02:43 cdarroch left 02:57 meppl left 03:05 mindos1 left 03:15 whiteknight left 03:18 Alias joined 03:22 leprevost joined 03:27 masonkramer joined 03:37 kanishka joined 03:43 kanishka left 04:01 leprevost left 04:02 _xiaomo_ joined 04:05 justatheory joined 04:15 kanishka joined 04:18 _xiaomo_ left, _xiaomo_ joined 04:19 kanishka left 04:32 nymacro joined 04:36 _xiaomo_ left
sorear good * #perl6 04:37
phenny sorear: 15 Nov 13:34Z <PhatEddy> tell sorear to check back here in irc when he has chance ... irclog.perlgeek.de/perl6/2010-11-15#i_2995501
04:45 kanishka joined 05:01 breinbaas left, breinbaas joined 05:12 masonkramer left
dalek ecza: c135903 | sorear++ | / (4 files):
Implement @*ARGS
 05:16
sorear I hear masak cannot write apps with my dialect
this needs to be fixed
05:18 kaare joined, kaare is now known as Guest20775 05:25 huf joined 05:45 __rnddim__ joined, kaare joined 05:46 kaare is now known as Guest36181 05:47 Lorn joined 05:48 Guest20775 left 05:49 nymacro left 06:04 __rnddim__ is now known as lue
dalek ecza: 3483ebc | sorear++ | v6/tryfile:
[v6] Add command line processing to tryfile
 06:08
06:13 agentzh joined
dalek kudo: f73bf4d | moritz++ | Configure.pl:
fix comparison logic thinko in Configure.pl, mikehh++
 06:21
06:23 perigrin joined 06:37 cjk101010 left, colomon left 06:41 bartolin joined
sorear good news: mono tryfile.exe CORE.setting starts 06:49
bad: it leaks memory. A lot. And it doesn't finish.
dalek ecza: 21f2eac | sorear++ | lib/Cursor.cs:
Instrument Cursor to dump highwater movement
 06:50
sorear ... or, for that matter, make much progress. 06:51
06:56 wtw joined 07:12 ggoebel left
dalek p-rx: c4c31a1 | moritz++ | / (2 files):
[configure] fix two thinkos
 07:18
ch3ck3r moins3n 07:22
07:24 ch3ck3r left 07:25 ch3ck joined 07:29 dukeleto joined 07:30 starcoder joined, araujo joined 07:32 Guest36181 is now known as kaare_
sorear there's probably a better tool than 'less' for studying 0.9GB log files 07:33
dukeleto sorear: cat ? 07:34
sorear: are you searching for a regex? 07:35
sorear: ack usually makes life happier
07:37 jfried joined 07:56 ch3ck left 08:01 colomon joined, cjk101010 joined
sorear dukeleto: I'm more doing "interactive exploration" 08:05
and ack can't handle corpuses larger than 10-100 MB or so without becoming unusably slow 08:06
this one is 875 MB
niecza-compiled STD is falling into an infinite loop but I'm not quite sure *where* 08:10
08:10 justatheory left 08:26 kjeldahl joined 08:28 Raynes joined, satyavvd joined
dalek ecza: 29ec560 | sorear++ | lib/Kernel.cs:
Add a *very* primitive statistical profiler
 08:34
sorear profiling++ 08:36
without this, I would never have noticed the actual problem 08:37
I somehow managed to write an O(2^n) implementation of Cursor.lineof
moritz_ wow 08:38
that's... scary
08:41 icwiener joined, bbkr left 08:42 Trashlord joined 08:55 tzhs joined 09:11 kanishka left, icwiener left 09:15 Woodi joined 09:19 icwiener joined 09:24 mikehh_ joined 09:26 mikehh left 09:29 kanishka joined
dalek ecza: 9912342 | sorear++ | v6/tryfile:
[v6] Stop runaway backtracking in Cursor.lineof
 09:31
ecza: 7b487ef | sorear++ | v6/tryfile:
[v6] Fix readonly value assignment in package_def
jnthn colomon: Use a version of make that doesn't care. ;-) Alternatively, replace the spaces with tabs. (Let me know if you don't have an nqp-rx commit bit.) 09:43
phenny jnthn: 02:39Z <colomon> tell jnthn nqp-rx / nom branch has spaces where there should be tabs in Makefile.in, then fails during the build. suggestions?
moritz_ jnthn: should I give everybody with a mu commit bit also access to nqp-rx? 09:46
jnthn moritz_: I don't think so - or at least, not without asking pmichaud 09:47
moritz_ ok 09:48
I just added colomon++
jnthn Anyone with a Rakudo commit bit is certainly fine though.
moritz_++
moritz_ I can't easily tie the two together, because rakudo has another account
sorear Should all people who have nqp-rx commit bits have them for Rakudo? 09:50
jnthn No, nqp-rx has a more liberal commit policy than Rakudo.
Just that anyone who's trusted enough to have a Rakudo commit bit is certainly trusted enough to have an nqp-rx one :)
dalek ast: e0961ad | moritz++ | S (4 files):
unfudge some tests for rakudo, including tests for RT #64474
 09:51
sorear S (4 files) 09:52
hehehe
how is roast going to look in the future, once niecza is running spectests?
moritz_ sorear: it'll probably have #?niecza markers too 09:53
just like it pugs and rakudo markers for some time
sorear yes, but 09:54
there's currently nothing like a ROAST_REVISION
so if I add new tests, rakudo will break
moritz_ you can be so kind to add them as skipped for rakudo 09:55
or you just break rakudo, and some rakudo devs fudge it
sorear would there be a general rule that new roast tests get fudges for all other major implementations? 09:56
well, implementations that care
moritz_ I guess we'll just see how it works out, and deduct rules when it seems necessary
sorear good plan 09:57
moritz_ I don't expect huge edit wars, or an exponential growth of number of perl 6 compilers :-)
sorear hehehe.
RT79242 is wtf-y 09:59
moritz_ indeed 10:00
10:03 QinGW left 10:08 rindolf joined
rindolf PackFile_unpack: This Parrot cannot read bytecode files with version 9.1. ==> what should I do about it? 10:08
Hi all, BTW.
moritz_ rindolf: remove the old parrot, and all files it generated 10:09
and compile a new one instead
sorear that includes all files it installed 10:10
moritz_ and go to #parrot and complain that their packfile format is incompatible across parrot versions (they know, but a bit of complaining helps to remind them that it's annoying for the end user)
sorear go to #parrot and complain that their handling of installing into /usr/local is utterly broken
I've given up on having Parrot in the default $PATH at this point 10:11
moritz_ installs into parrot_install dir that he can easily nuke, and only symlink parrot_install/bin/perl6 into ~/bin
sorear that works? impressibe 10:12
moritz_ it works as long as the install location is known at configure time 10:13
10:14 ddwagnz joined 10:15 ddwagnz left
rindolf moritz_: I did Rakudo perl Configure.pl --gen-parrot. 10:15
moritz_: it should have installed everything into parrot_install
moritz_ rindolf: right. Then you might need to rm -rf parrot_install; make realclean
rindolf moritz_: make realclean where? 10:16
moritz_ in rakudo
and since your parrot copy is probably still svn, rm -rf parrot/ too
rindolf moritz_: I already have a fresh git checkout there. 10:17
moritz_ then you don't need that step.
rindolf moritz_: in ./parrot/ I mean.
Trashlord hi rindolf
rindolf Hi Trashlord
sorear hello 10:18
when in doubt, git clean -dxff
Trashlord I gotta go to work at 2pm to replace someone
rindolf Trashlord: ah.
Trashlord yeah, kinda sucks, but oh well, it's only for like 2 hours or so
moritz_ Trashlord: when is the last time you contributed something Perl 6 related to this channel?
sorear how very uncharacteristic of you 10:19
didn't you write a hugbot once?
rindolf There's a Perl 6 lecture today by szabgab as part of www.lambda.org.il/
Trashlord moritz_: I think a few months ago 10:20
moritz_ Trashlord: please try to keep your signal/noise ratio high
well, raise it, it's not high at the moment :-)
10:21 dakkar joined
Trashlord moritz_: alright then 10:22
10:23 sftp joined
sorear is trying to implement .comb and running into "a" !~~ /./ 10:32
moritz_ huh? 10:33
your . seems pretty selective
sorear some kind of deep brokenness
I had a bug the other day where . wouldn't match unassigned codepoints
10:35 mavrc joined 10:47 timbunce joined
sorear niecza: for [1,2,3] -> $x { say $x.dump } 10:48
p6eval niecza 7b487ef: OUTPUT«1␤2␤3␤»
moritz_ niecza: for (1,2,3) -> $x { say $x.dump }
p6eval niecza 7b487ef: OUTPUT«1␤2␤3␤»
moritz_ the latter is correct, the former not
dalek ecza: cbbbd07 | sorear++ | src/Op.pm:
Fix for loops ignoring argument flattiness
 10:49
moritz_ rakudo: for [1, 2, 3] -> $x {say $x }
p6eval rakudo 015d77: OUTPUT«1 2 3␤»
sorear yeah
that was the actual bug
a loop in .comb was stripping off too many levels of listiness
/./ returns a single match with no positional captures, it's sort of like [] 10:50
previous versions of the regex engine returned ([],) instead, but I've optimized that... exposing the for bug
dalek ecza: 6ca0209 | sorear++ | test3.pl:
Add a prototype of Cool.comb
 10:55
ecza: 78407bb | sorear++ | v6/tryfile:
[v6] Add enum support
10:59 meppl joined 11:04 mavrc left 11:05 mavrc joined
sorear niecza: my %h; %h<a> = (%h<a> = 5) 11:10
p6eval niecza 7b487ef: OUTPUT«Unhandled exception: Autovivification collision␤ at line 650 (SAFE extend @ 15)␤ at line 731 (SAFE G506ANONC @ 6)␤ at line 0 (Assign @ 1)␤ at line 1 (MAIN mainline @ 5)␤ at line 954 (SAFE G818ANONC @ 5)␤ at line 954 (SAFE module-SAFE @ 28)␤ at line 954 (SAFE
..mainline …
moritz_ rakudo: my %h; %h<a> = (%h<a> = 5); say %h.perl
p6eval rakudo 015d77: OUTPUT«{"a" => 5}␤»
moritz_ wouldn't be surprised if rakudo blew up too
11:11 sftp_ joined, sftp left
sorear rakudo doens't blow up because rakudo is missing an internal sanity check 11:12
only, now that STD has hit it I wonder if it may be me that is insane
moritz_ what makes you think that it's missing a check?
sorear in that expression, both instances of %h<a> are evaluated before either assignment
they both create independant variables with WHENCE closures to slot themselves in 11:13
my WHENCE closures check that nobody else has put in a value in the meantime
moritz_ rakudo: my %h; %h<a> = 2 *(%h<a> = 5); say %h.perl
p6eval rakudo 015d77: OUTPUT«{"a" => 10}␤»
moritz_ seems pretty correct to me 11:14
sorear rakudo: my %h; %h<x> += (%h<x> = 5); say %h.perl
p6eval rakudo 015d77: OUTPUT«{"x" => 5}␤»
moritz_ seems correct too
as long as there's no user visible misbehavior, I maintain that rakudo is correct here 11:16
sorear rakudo: my %h; sub foo(\$x) { %h<a> = 5; $x++ }; foo(%h<a>); say %h.perl 11:17
p6eval rakudo 015d77: OUTPUT«{"a" => 1}␤»
moritz_ sorear: please put those in the test suite 11:18
sorear moritz_: what's correct?
moritz_ so far they all looked correct to me
sorear even the last one?
moritz_ wait, no
11:19 daxim joined
moritz_ hm 11:19
not sure
sorear the issue is Perl 6's lazy autovivification
moritz_ it creates a new autovivification closure
sorear %h doesn't know about $x
so %h<a> = 5 sets a new slot, not the nascent one in $x
moritz_ then $x++ is correct to store 1 in %h<a>, no? 11:20
or should it know about the 5?
after double confusion, I think that 1 might be correct 11:21
sorear that's the problem - I don't know
1 is the easy answer
moritz_ but I'm not sure either
sorear but not the POLS one
moritz_ well, POLS does conflict with many optimizations, and Perl has traditionally had a balanced view on how to chose (and not always in favor of POLS) 11:22
dalek ecza: 1f49178 | sorear++ | v6/tryfile:
[v6] Add trait parser
 11:27
11:27 ggoebel joined
dalek ecza: 40ac1ca | sorear++ | v6/tryfile:
[v6] Removed a sanity check that did more harm

  %hash<key> = (%hash<key> = 5) is now accepted, despite being on
somewhat shaky ground in S09. Some other variations violate POLS; see irclog.perlgeek.de/perl6/2010-11-16#i_2998697 .		 11:30
11:31 jedai left 11:36 ch3ck joined
sorear calls it a night 11:36
11:38 rindolf left 11:43 satyavvd left
colomon jnthn: I wasn't clear, sorry. I fixed the makefile problem, then the build failed afterward. 11:50
11:51 masonkramer joined 11:56 jedai joined
colomon and then I cleverly blew away the entire directory instead of keeping it around so I could send you the problem :\ 12:04
I'll try to start again after my walk this morning.
12:15 sftp_ left
jnthn colomon: ah, ok 12:18
Would need to see how it failed
Maybe some *nix silliness. :)
colomon yeah, of course. I apologize for not thinking to save the directory. (I tried to merge latest nqp-rx and failed, then blasted it to get ready of git-messiness) 12:19
moritz_ colomon: fwiw you now have a commit bit to nqp-rx 12:21
colomon moritz_++
github actually let me know.
(and I saw it in the backlog, too.) 12:22
off to walk. 12:27
12:27 masak joined
masak oh hai, #perl6! 12:27
colomon but, err... you installed a curling app on my phone?!
jnthn yayitsmasak! 12:28
12:28 orafu left, orafu joined
colomon wow, big time wrong window there. 12:28
masak has not touched my phone. :) 12:29
moritz_ lolitsmasak!
masak colomon: I'm glad we agree.
sorear++ # gearing niecza to be app-friendly 12:31
moritz_ masak: I've started to submit TODO niecza bug reports to the github issue tracker - feel free to join me in that quest :-)
masak good idea. 12:32
moritz_++
moritz_ $ bzgrep 'masak submits rakudobug' *.bz2|wc -l 12:37
696
that's in the directory of the #perl6 logs, obviously 12:38
masak heh :) 12:41
12:43 [particle]1 joined
masak anyone seen this? search.cpan.org/dist/Whatever/lib/Whatever.pm 12:44
12:47 [particle] left
takadonet morning all 12:51
masak takadonet: \o 12:54
takadonet masak: how are u?
masak takadonet: well rested. I wish I could be both well rested and sleep on the right side of the morning, though. 12:55
and u?
takadonet hacking on my big perl 5 project and made another round of commit on Text-Tabs-Wrap 12:56
masak \o/ 12:57
how's Text-Tabs-Wrap coming along?
I see it has three medals on modules.perl6.org :) 12:58
takadonet well I forgot out how to pass a rx// to it now so that makes a good chuck of the test passing. So out of the 3 'main' cases. 2 work and the third works on most cases
masak nice. 13:00
takadonet: have you thought some more about contributing a Str.indent method to Rakudo? I think it would be fairly low-hanging fruit for you now that you've ported Text::Tabs::Wrap. 13:01
takadonet masak: probably in the future 13:02
masak sure, no pressure. 13:03
13:08 agentzh left
takadonet there is... 13:08
heh
13:15 nothingmuch joined
masak thundergnat++'s rt.perl.org/rt3/Ticket/Display.html?id=79242 is an excellent example of the assertion that modules and applications, just by virtue of crawling the space of all possible combinations and orderings of features, help find rakudobugs. 13:15
also, we're over 800 new/open tickets now. 13:16
takadonet should start closing them?
masak yes, I think the time has come to close at least some of them. 13:17
but I guess there's a bit of a bell-on-cat problem.
and I do respect the core devs' right to choose their priorities. 13:18
13:20 _xiaomo_ joined
jnthn ENOTENOUGHCOREDEVS 13:20
13:21 MayDaniel joined 13:22 Woodi left, PerlPilot joined 13:23 Patterner left
takadonet jnthn: Well for the moment, i'm too far upstream the coding tree to help just yet. I love modifying code, so porting cpan module is the best I can do for now 13:24
13:28 risou joined 13:29 Woodi joined
masak takadonet: and that's just as important, as thundergnat's bug report shows. 13:29
takadonet masak: I know 13:30
side note the program meld is the best thing since slice bread when porting a module 13:32
13:34 Psyche^ joined, Psyche^ is now known as Patterner
colomon meld? 13:34
13:35 MayDaniel left
takadonet meld.sourceforge.net/ 13:36
best way to compare two or third files to see what is different
masak why is a merge/diff tool useful when porting from Perl 5 to Perl 6? 13:40
takadonet you do print statement at different stages of the program. you can tell where things are going wrong
so print the input before a function and after for both version of the module then see what is different 13:41
masak interesting.
sounds fairly complementary to tests. 13:42
takadonet even after if /else
makes it easier to do small steps
well I use it in conjunction with tests 13:43
masak right, that's what I mean.
takadonet now having trouble.... appending a "\n" to every element of an array
masak see if I understand this: you put debug statements in both the Perl 5 and the Perl 6 code?
takadonet rakudo: my @arr= <'b,'a','c'>; @arr>>~>>"\n" ; @arr.say 13:44
p6eval rakudo 015d77: OUTPUT«'b,'a','c'␤»
takadonet yes
I download the module so I don't mess around my system version of it
masak takadonet: you're not assigning to anything there.
takadonet ...
takadonet needs more coffee
masak rakudo: my @arr= <'b,'a','c'>; say @arr>>~>>"\n"
p6eval rakudo 015d77: OUTPUT«'b,'a','c'␤␤»
masak hm. 13:45
oh right. only one element in that array.
takadonet .. 13:46
masak elements inside <> are separated by whitespace.
not commas.
takadonet ya ya.... man really do need that coffee 13:47
13:47 kanishka left
masak alester: re "It won't happen, so there's no point in discussing it." from perlbuzz.com/2010/11/progressing-vs...gging.html -- while agree with both sides of that implication, I don't agree that there's no point in discussing it *because* it won't happen. 13:53
alester: it's more like, there are a small bunch of people actually working on Perl 6 and making it real. then there's a small bunch of people clamoring for a new name. and those two sets of people are mostly disjunct. which, I would say, lends little credibility to the "rename" bunch. 13:54
it's as if I had very little involvement with Google Go, but I was really adamant that the name needed changing, for PR reasons. 13:56
14:00 kanishka joined, nothingmuch left
takadonet is their an 'easy' way to append "\n" to every element of the array but the last one? 14:04
masak what about the edge case, when the array is empty? 14:05
14:05 _xiaomo_ left
takadonet SHOULD not have an array empty 14:06
but if they are, will check beforehand and skip the step
masak rakudo: my @a = <a b c>; @a[0 ..^ *-1] >>~=<< "\n"; say @a.perl 14:07
p6eval rakudo 015d77: OUTPUT«(timeout)»
masak submits rakudobug
colomon rakudobug for that? 14:08
masak rakudo: my @a = <a b c>; @a[0 ..^ @a-1] >>~=>> "\n"; say @a.perl 14:09
p6eval rakudo 015d77: OUTPUT«["a\n", "b\n", "c"]␤»
masak colomon: locally, I get a stack overflow.
takadonet: that work for you?
colomon but do you think it is correct?
masak colomon: no, you're right.
rakudo: my @a = <a b c>; @a[0 ..^ *-1] >>~=>> "\n"; say @a.perl 14:10
p6eval rakudo 015d77: OUTPUT«(timeout)»
colomon you're mixing positional whatever with range whatever
masak oh, I can't do that?
colomon probably shouldn't stack overflow, but...
masak I totally thought I could.
it feels so natural.
rakudo: my @a = <a b c>; @a[0 ..^ *-1] 14:11
p6eval rakudo 015d77: OUTPUT«(timeout)»
colomon well, do you know if it is (0 ..^ *) - 1 or 0 ..^ (*-1) (for starters)
?
masak I'm assuming the latter, for precedence reasons.
colomon and if it is the later, you're making a Range with an Int on one side and a closure on the other
masak so?
that's what WhateverCode is; a closure. 14:12
I don't see the fundamental difference to ordinaty ranging with Whatever.
colomon ranges don't participate in WhateverCode, do they?
masak I don't know.
colomon Certainly 1..* does not generate a closure 14:13
masak S09:245 has @array[0 .. *-1] as an example of valid slicings. 14:14
14:15 mikehh_ is now known as mikehh
masak that line in S09 is due to someone with the nick 'lwall'. 14:15
colomon never heard of
'em
masak :)
14:17 rgrau joined
colomon where is WhateverCode covered in the spec? 14:18
(It does look like it should parse as 0..(*-1)) 14:19
jnthn masak: In [0..*-1] the *-1 would...yes, that. :)
14:19 [particle]1 is now known as [particle]
takadonet masak: perfect. All my 14:19
all my
...
masak jnthn: that's why I expected.
takadonet all my 'meld' test are good now
masak s/why/that/
or rather, I don't care much how it works implementation-wise (as long as it works), but that was the precedence I assumed. 14:20
14:21 plainhao joined
masak takadonet: glad to hear it. 14:21
colomon "For any prefix, postfix, or infix operator that would be curried by a Whatever, a WhateverCode also autocurries it"
to me, that pretty explicitly leaves out the 0..*-1 case 14:22
takadonet takadonet: so got 8 more test passing and 6 fail... not bad start
masak colomon: even though S09 says it should work?
14:22 KyleHa joined
colomon masak: the spec is known to have errors 14:22
masak KyleHa! \o/
colomon: but it would *suck* if this didn't work!
KyleHa Howdy Masak!
masak colomon: and TimToady wrote that part of S09, so it's not like it's untrustworthy because of many chefs. 14:23
colomon personally, I'm not seeing any reason to have the above quote... it seems like anything hit with a WhateverCode could curry.
masak: TimToady++ is a genius, but he does make mistakes.
masak colomon: well, what I'm saying doesn't depend on TimToady not making mistakes. it depends on what feels right from a user perspective. 14:24
and if I won't be able to write [0 .. *-1], I'll fork and make my own Perl 6.
jnthn
.oO( Who'd be less tormenting to the implementer... )
 14:26
colomon Look, I don't see any obvious harm in allowing that. But presumably the specquote is there for a reason which I do not understand.
takadonet well we will ask TimToady when he is in 14:27
anyway back to hacking...
masak I don't see why this even needs deferral to TimToady. he already wrote in S09 how it should wokr. 14:28
s/wokr/work/
fwiw, I think that rakudobug (the fact that it recurses infinitely) is fairly severe. it's not a terribly uncommon use case to want to slice up to *-1. 14:30
& 14:31
14:33 Raynes left 14:34 kjeldahl left
alester sorear: I'm interested in your assertion that "and ack can't handle corpuses larger than 10-100 MB or so without becoming unusably slow" 14:37
Teratogen does perl 6 handle complex numbers?
alester masak: As far as I'm concerned, any discussion of "Perl 6 should actually be called X" is wasted breath.
Teratogen and does this extend to the math library?
flussence Yes and Yes.
rakudo: atan2(3i) 14:38
p6eval rakudo 015d77: OUTPUT«atan2 is only defined for Reals, you have a Complex()␤»
colomon jnthn: I think I can see where to modify the source to make masak happy
rakudo: atan(3i)
p6eval rakudo 015d77: ( no output )
colomon rakudo: say atan(3i)
p6eval rakudo 015d77: OUTPUT«1.5707963267949 + 0.346573590279973i␤» 14:39
14:40 Raynes joined
colomon jnthn: just not 100% convinced it is a good idea 14:40
14:40 KyleHa left
colomon that is to say, I can't think of any logical justification for why 0..*-1 should work other than, "Man, that would be nice." 14:45
takadonet colomon: is that not a good enough reason? :) 14:46
colomon takadonet: as a for instance, @list xx * is another operator like 0..* which doesn't generate a WhateverCode.
does @list xx *-1 make any sense? 14:47
flussence I rationalise it as shorthand for @a[*].Int - 1
14:50 KyleHa joined 14:53 kaare_ left
colomon bah, compiling rakudo takes too long 14:54
either that or I really really broke it...
oh, builds and passes make test 14:59
but if I try 0..*-1 15:00
"set_number_native() not implemented in class 'Num'"
any ideas what that might mean?
15:06 zby_ joined, zby_ left 15:09 PerlPilot is now known as PerlJam 15:10 PerlJam left, PerlJam joined 15:14 satyavvd joined
colomon jnthn, pmichaud_: ping? 15:16
15:17 timbunce left 15:18 oyse joined, silug left 15:20 satyavvd left, satyavvd joined, wtw left 15:24 am0c joined 15:25 oyse left 15:30 plobsing left 15:32 zby joined 15:38 MayDaniel joined 15:43 plobsing joined 15:48 justatheory joined
masak colomon: here's where I'm coming from. (I should have explained this part from the beginning.) @a[0..4] works. @a[*-1] works. purely by virtue of those two working, I expect @a[ 0 .. *-1 ] to work, and to have the "obvious" semantics that combining the two simpler cases would produce. 15:48
in general, when I combine features in Perl, I have a higher expectation than in other languages that the combination of features will Just Work. this is such a case. 15:49
it's not just a case of "ooh, wouldn't it be nice if...?"
colomon masak: yes, I understand all that.
masak I'm actually fairly surprised that someone would consider arguing against the usefulness of the above case. 15:50
colomon but you're basically expecting two completely different features to magically combine and do exactly what you want
PerlJam colomon: isn't that the Perl Way :)
colomon and you don't seem to be considering that these are completely general features
15:51 jaldhar left
masak colomon: I don't *care* how they combine. 15:51
15:51 jaldhar joined
colomon for instance, the spec says that Feb..* is the same as Feb..Dec 15:52
masak fsvo "the same".
colomon what the heck should Feb..*-1 do?
PerlJam Feb..Nov of course
masak that's a good, but unrelated, question.
colomon it's completely related
masak no, I'm not talking about enums.
colomon because what you've just asked is that Feb..*-1 return closure
return a WhateverCode, specifically 15:53
masak *sigh*
there should be a word for when a user comes to a dev, expecting sympathy, but getting internals-speak instead.
unless you mean to say "it can't be done", I'm not too interested in the internals details in this case. 15:54
colomon how is that an internal detail?
masak if it doesn't work with our current model, we need a new model.
colomon: everything about how .[0 .. *-1] translates into "all elements except the last one" is, by definition, details about the internals. 15:56
colomon but what 0 .. *-1 returns is NOT about internals 15:57
PerlJam masak: given that, I'm having a hard time reconciling how colomon's enum example is "unrelated" 15:58
masak PerlJam: it's not completely unrelated, but it's not about array indices any more. 15:59
PerlJam okay
masak PerlJam: I know colomon++ wants to shed light on what he considers is a problem with making the current implementation do what the spec says, and that he brought up *-1 in the context of enums for that reason. 16:00
16:00 plobsing_ joined 16:01 plobsing left
masak PerlJam: however, my interest is to keep the focus on the very simple issue of slicing an array up to but not including the last element. it's backed up by the spec, it falls out naturally as the combination of two simpler cases. 16:01
colomon it doesn't fall out naturally, that's the problem here!
it looks like it does, that's all.
masak it doesn't fall out naturally in Rakudo. 16:02
that's the problem.
colomon no, it doesn't fall out naturally in perl 6
masak I disagree.
colomon that's the entire point of the enum example.
masak please explain.
colomon Why does [*-1] work? Because it generates a WhateverCode. 16:04
Why does [1..*] work? Because Ranges are intersected with array bounds, more or less.
masak "intersected with array bounds"? 16:05
flussence "clamped to"?
PerlJam colomon: A typical newbie Perl 6 coder isn't going to know those things however. They'll see "this does this and that does that, maybe I can combine this and that"
colomon "clamped to", yes.
masak colomon: what PerlJam said.
colomon PerlJam: yes, it would be absolutely terrific if [0..*-1] worked. 16:06
masak colomon: what you're doing now is explaining internals to me. with my dev hat on, I know those things.
colomon but to build a consistent language, you have to have a REASON it works.
masak colomon: with my user hat on I don't much care.
16:06 MayDaniel_ joined
colomon So, it makes sense to say, well, if you pass a WhateverCode to a Range, it autocurries the Range. 16:07
PerlJam colomon: but the users will have a reason it works. that reason won't have anything to do with WhateverCode or Range however. 16:08
colomon though note that even that has all sorts of problematic issues.
PerlJam s/it works/it should work/
masak colomon: making @a[ *-1 ] communicate from the whatever star to the array indexing via a WhateverCode (as opposed to, say, this being a purely syntactic construct) was a *design decision* taken by Rakudo. you're defending it as if it can't be changed or improved. 16:09
colomon no
16:09 kanishka left
colomon that's not rakudo. that's perl 6 16:09
masak oh?
@a[ *-1 ] not being a purely syntactic thing is in the spec? 16:10
16:10 MayDaniel left
colomon *-1 generating a WhateverCode closure is Perl 6 16:10
masak that's different.
colomon "he final element of an array is subscripted as @a[*-1], which means that when the subscripting operation discovers a Code:($) object for a subscript, it calls it and supplies an argument indicating the number of elements in (that dimension of) the array."
that's S02 16:11
masak huh. ok.
then I agree with you. that part of S02, together with your reasoning about Range, sits in opposition to the line of S09 that I quoted. 16:12
16:13 MindosCheng joined
jnthn I suspect users AND devs can unite around wanting a consistent spec. :) 16:13
colomon and again, it would be awesome if @a[0..*-1] worked.
16:13 bluescreen joined
masak awesome and per spec. 16:14
16:14 bluescreen left
colomon but either we're adding another layer of inconsistent magic to make that work, or there will be consequences for things like Feb..*-1 16:14
and what about something like @a[*/2 .. *-1] ?
masak I don't consider Feb..*-1 that unreasonable, fwiw.
16:15 envi left
colomon Feb .. *-1 returning a WhateverCode, you mean? 16:15
masak colomon: I have objections to */2, but only on a code-maintenance level.
colomon: whatever it *returns*, I'd expect it to *mean* Feb..Nov (like PerlJam said), or be illegal.
if it's illegal, there should be an overwhelming reason for it being illegal, for example that it's unimplementable. 16:16
16:16 tzhs left
masak or that it's implementable, but it can't be made fast. 16:16
diakopter I want @a[but last] to work ;) 16:18
masak all is fair if you predeclare. ;)
TimToady note that there's no guarantee that Foo .. Bar will represent a 0..^$n range
if Jan == 1 then you can't subscript an array with it, assuming Jan is the first element 16:19
colomon hmmm... I guess I mean * div 2, btw. 16:20
PerlJam colomon: */2 would work
masak TimToady: I have always thought of the array indexing being wrapped by some adapter-like thingy in the case of exotic indexings. 16:21
PerlJam for array indexing you get floor(*/2)
TimToady there's also no guarantee of uniqueness in enums--what if Bar - 1 has two enums mapped to it?
colomon jnthn: so what does "set_number_native() not implemented in class 'Num'" mean, anyway? That's not the error I expected to get
16:21 [particle]1 joined, risou_ joined
TimToady likely range operations on enums are on their values, not their keys, so Feb .. * - 1 is really 2 .. * - 1 (assuming your enum is 1-based) 16:22
masak aye. 16:23
colomon masak: As for Feb .. *-1 meaning Feb .. Nov or being illegal -- the change you are effectively requesting to the spec would make it legal and equal to { Feb .. $_ - 1 }
masak this is what I meant by enums being an interesting but unrelated question.
16:23 [particle] left, [particle]1 is now known as [particle]
colomon hopes TimToady++ has a brilliant way of making this all work 16:23
16:23 risou left
TimToady well, I think you folks are misreading S09 16:24
the implication goes the other way
colomon and if you don't like @a[*/2 .. *-1], consider @a[*-3 .. *-1]
jnthn colomon: I'm mostly confused by that error. :/ 16:25
masak colomon: I would expect @a[*-3 .. *-1] to work too.
colomon: and yes, that's inconsistent with how * works in other contexts.
TimToady assuming subscripts can note the arity of the WhateverCode they get
masak hm, that'd help.
jnthn colomon: It feels like an attempt to mutable an immutable thing is happening but I'm still surprised it's that error.
colomon jnthn: context is that's what happened when I tried to allow WhateverCode arguments to autocurry for Range and sequence operators. 16:26
jnthn: gist.github.com/702012
moritz_ TimToady: somehow I thought the arity related to the dimensionality of the array, but I might be wrong about that 16:27
jnthn colomon: OK. I need to decommute now...home in 30-45 minutes and can look more then.
TimToady moritz_: maybe you're thinking of HyperWhatever?
moritz_ yes 16:29
TimToady for @a[*-3 .. *-1] to work, subscripting would notice a 2-arity and feed the size as two arguments
colomon I see the @array[*-3..*-1] example is actually in S09
16:30 rbuels joined
TimToady or, wild idea, or maybe autocurrying should default the second arg to the first 16:30
that might be in the category of six impossible things before breakfast though 16:31
colomon that actually seems kind of sensible to me. 16:32
I guess it might be dangerous in other contexts.
TimToady that wouldn't help the folks who want * * * to square a number in a sequence operator though
16:33 XaRDaX joined
colomon actually, that's the best reason NOT to default the second arg to the first 16:33
because it would only add to the chaos if (* * *)(3) was nine, but 1, 2, * * * ... * was 1 * 2 16:34
(so 1, 2, 4, 8, etc instead of 1, 2, 4, 16, 256, etc) 16:35
TimToady well, the arity detection probably belongs in the subscriptor 16:36
colomon +1
masak I think things are magical enough as they are without things defaulting to things.
colomon TimToady: so, do you have a notion how @a[1..*-1] works?
TimToady exactly as specced
1..*-1 returns a WhateverCode
colomon that's not spec, as far as I can tell 16:37
TimToady if you think S09 contradicts this, you're reading it wrong
colomon no, I think S02 contradicts this
or at least, explicitly says it doesn't have to work in the general case, with no mention of the Range operator specific case
TimToady I don't see that 16:38
colomon "For any prefix, postfix, or infix operator that would be curried by a Whatever, a WhateverCode also autocurries it, such that any noun phrase based on * as a head noun autocurries transitively outward as far as it makes sense, including outward through metaoperators."
There's no mention of what happens for an operator that would NOT be curried by a Whatever, but the very fact they are explicitly not mentioned has to mean something. 16:39
TimToady .. is not in the first category, so the rest doesn't apply; again, you're reading the implication backwards
that's an argument from silence
dalek ecs: c0b0845 | duff++ | S09-data.pod:
Spring before Summer usually :)
 16:40
TimToady all operators autocurry a WhateverCode, whether or not they curry a Whatever
colomon really????
dalek ecs: 8b43df0 | duff++ | S09-data.pod:
random errant "or"
TimToady certainly all binary operators 16:41
colomon then why the heck is "For any prefix, postfix, or infix operator that would be curried by a Whatever," in there?
so $a = *-1 does not assign a WhateverCode to the variable $a?
TimToady assignment isn't really a binary operator 16:42
masak PerlJam: nice catch with Spring and Summer. github.com/perl6/specs/commit/c0b0...cf56f13a20
TimToady neither is ~~
masak PerlJam: but what is that extra 'or' doing there?
PerlJam masak: it's gone now.
masak oki :)
colomon TimToady: perhaps we need a list of the infix operators which are not really binary operators. :p 16:43
TimToady in any case, the intent is that the operators that autocurry Whatever are a subset of the operators that autocurry WhateverCode
and there are operators like .. that autocurry WhateverCode but not Whatever 16:44
colomon and what are those operators?
TimToady the ones that prefer to interpret a bare * as "to infinity and beyond!" 16:45
such as xx
colomon (and so the user will get confused by Feb .. *-1 but @a[0..*-1] works as expected. seems reasonable to me.)
TimToady Feb .. *-1 will only do what is expected if the enum is 0-based 16:46
PerlJam TimToady: For the user defined indices @calendar example in S09, could @calendar{Jan;1;9} also be written @calendar{1;1;9} ?
colomon TimToady: and you're trying to use it as a subscript
if you say @favorite_months = Feb .. * - 1, you're not going to get an array of ten months 16:47
TimToady enums aren't ranges
colomon (as I said, I think that's a reasonable hole)
TimToady though they are often used to represent them
however, if you say Feb .. Dec - 1 you might get 2..11 16:49
or 1..10 if 0-based
using * when you know the final value is a bit silly in any case
why wouldn't they just write Feb .. Nov? 16:50
most enums are declarative
PerlJam :q
colomon TimToady: they could be writing something like Feb .. *-$a
PerlJam looks for generative enums 16:51
.oO(I thought all enums were declarative)
colomon I'm just trying to explore where the magic happens.
16:51 mila_ joined
colomon anyway, my patch for making [0..*-1] work is waiting on jnthn to explain why it fails miserably... 16:52
16:55 mila__ left, sftp joined
PerlJam Perl 6 has this way of making things that I thought were simple suddenly seem very complex. :) 16:57
masak .oO( all I wanted was to recline in this nice waterbed ) 16:58
MindosCheng masak: I finally understand what is .oO( ) notation... 16:59
masak MindosCheng: :) 17:00
PerlJam Does the following work? gist.github.com/702034
masak I can see how it could; don't know whether it does. 17:01
PerlJam I think S09 needs an explicit statement about the relationship between enum values as used in user-defined indexes and how they are mapped to the standard 0-based indexes. I just don't know what that statement should say :) 17:02
TimToady how can that possibly work if all the enums map to 1? 17:03
dukeleto This is pretty cool: search.cpan.org/dist/Whatever/lib/Whatever.pm
masak dukeleto: yeah, found that one earlier today :)
dukeleto masak: I think that is hard proof that Perl 5 and Perl 6 are sister languages, and not competitors. Good stuff flows both ways. 17:05
masak dukeleto: and it's hardly the first such instance.
dukeleto: on the language level, Perl 5 recently got given/when.
moritz_ why can't they both be sisters and competitors? :-)
17:05 stkowski joined
masak moritz_++ :) 17:06
moritz_ masak: and say(), ~~, // and others
masak right.
moritz_ I mean, I used to compete with my brother and sister all the time
takadonet well it's time to 'steal' their cpan module!
masak and Regexp::Grammars.
PerlJam TimToady: I don't understand how "mapped to 1" related to their use as indices.
TimToady that essentially gives them the same identity
unlike a Set, where the 1 would indicate existence 17:07
PerlJam okay, that makes sense then.
masak TimToady: so the mapping in enums essentially allows the user to give the same value several names?
TimToady that is how enums are typically defined
masak right. 17:08
TimToady they're really just a shorthand for constant declarations
PerlJam so, could @calendar{Jan;1;9} also be written as @calendar{1;1;9} ?
TimToady yes, if Jan == 1 17:09
frettled I'm no. 1, yes. 17:11
What? Oh, you were monthing. :)
TimToady
.oO(nick encapsulation violation)
I find it very strange that English allows "no." as an abbreviation for "number"; obviously some history there... 17:12
frettled It makes for some excellent ambiguity. 17:13
TimToady no.
masak :P
frettled Up is my no. 17:16
diakopter o. no. no. no.
17:22 risou joined
TimToady I'll be going down so. for the holidays. 17:24
17:24 risou_ left
masak diakopter: your pun is so.ewhat obscure :P 17:25
17:26 Bzek left 17:27 silug joined 17:28 Bzek joined
dalek ecs: 977d920 | TimToady++ | S02-bits.pod:
[S02] clarify * vs *-1 semantics for globbish ops

Globbish ops like .. and xx do not autocurry on Whatever but do autocurry on WhateverCode. Also mention that assignment and smartmatching don't autocurry because they're primitive pseudo-operators.		 17:29
masak \o/ 17:30
17:30 kanishka joined
diakopter yay 17:34
"laid out in memory contiguously without pointer indirection"
yay "assignment ... syntactic sugar for underlying primitive" 17:35
colomon TimToady++ 17:38
#phasers in 80 minutes? 17:40
diakopter mruhaha; this means my perlesque subset language is a lot (like, nearly all the way) closer to correct than [I] previously thought.
17:41 takadonet left 17:42 takadonet joined
masak oh dear, phasers. gotta go nom. & 17:42
17:42 masak left 17:44 Axius joined
jnthn home 17:48
17:49 Axius_ joined 17:50 Axius left 17:51 am0c left 17:53 zb joined, kanishka left
colomon jnthn: I'm hoping you have some time for me to pick your brains here tonight. (And I'm hoping I have time to do so!) 17:54
jnthn colomon: OK, I should be about a bit. 17:55
colomon gist.github.com/702012 is the code change that lead to the set_number_native error 17:58
anything obviously wrong with it?
17:58 MayDaniel_ left 17:59 cjk101010 left 18:00 dakkar left
jnthn colomon: Any chance of a diff? 18:00
<-- lazy ;)
18:01 Axius_ left
jnthn Also when does teh error occur? 18:01
colomon it's not going to do you any good...
jnthn Runtime? Compile time?
colomon gist.github.com/702174
18:02 cdarroch joined, cdarroch left, cdarroch joined
colomon getting the error in the REPL when I send it 0..*-1 18:02
(That's the diff for the code, the other diff is that operators which should never autocurry have their not_curried value set to 2) 18:03
jnthn ah, I see the aim... 18:04
But if you do e.g. --target=pir does it produce output? 18:06
e.g. is this actually dying because of the generated code, or while generating it?
18:07 impious joined
jnthn colomon: I understand what the patch is doing now 18:07
colomon: And it doesn't look wrong in any way that's obvious to me.
(The checking logic looks right, anyways.) 18:08
colomon \o/ 18:10
of course, an obvious error causing problems would be easier to debug
Wynne:rakudo colomon$ perl6 --target=pir 18:11
> 0..*-1
set_number_native() not implemented in class 'Num'
so dying generating code, I think?
jnthn Looks like it. :S 18:14
sorear good * #perl6
jnthn o/ sorear
MindosCheng \o sorear 18:15
colomon jnthn: any easy way of adding a "say" in there, so we can see if it is getting past the conditionals or not?
jnthn pir::say("# lolz fail"); will work in Actions.pm easily enough. 18:17
colomon can I make it "# lolz fail" ~ $/ ? 18:18
jnthn aye
maybe ~$/
(e.g. coerce match object to a string) 18:19
just to be on the safe side
sorear @a[1..*] doesn't work because of a critical syntactic ambiguity
jnthn Going to sort out some noms pre-#phasers, bbiab 18:20
18:20 _kaare joined
sorear just use @a[{1..$^a}] 18:20
oh masak is gone
colomon sorear: you've got a lot of backlogging to catch up on there. :)
takadonet sorear: you will also see that I was the cause of it 18:23
18:23 risou_ joined 18:24 risou left 18:25 Chillance joined
colomon jnthn: looks like it is indeed properly getting past the conditionals and into the currying body 18:25
sorear whee, done 18:27
18:28 molaf_ left
sorear alester: ack was designed for trees of code; it can't optimize "ack foo /tmp/one-gig-logfile" by throwing away files 18:30
alester sorear: oh yeah, if you're just doing a straight grep, use grp.
grep.
18:30 PerlJam left 18:33 impious left 18:34 kjeldahl joined 18:36 lichtkind joined
lichtkind hai 18:36
sjohnson p6 masters, i was wondering if this very nice python example can be written as one complex statement as well. 18:37
www.pastie.org/1303259
i remember asking if there was a function based way to do string replaces a while back, instead of s///, which might make it easily doable 18:38
in this python example, it is quite certain the int will be a BigNum too
18:39 wamba joined
takadonet colomon: your up sir :) 18:39
18:40 wamba left, wamba joined
colomon sjohnson: seems like you're talking about subst? 18:41
flussence rakudo: say :16('09:f9:11'.comb(/<xdigit>+/).join)
p6eval rakudo 015d77: OUTPUT«653585␤»
sjohnson wow
:16
i didnt know about that
colomon but yeah, unless I'm missing something, flussence++'s approach is much better than that python code there 18:42
errr, or maybe 18:43
rakudo: say :16('09:f9:11'.subst(':', '', :g).join) 18:44
p6eval rakudo 015d77: OUTPUT«653585␤»
colomon whoops
rakudo: say :16('09:f9:11'.subst(':', '', :g))
p6eval rakudo 015d77: OUTPUT«653585␤»
MindosCheng Cool. 18:45
sjohnson colomon: im surprised no one told me about subst
maybe i asked on a different channel
but the answers were mostly "you don't need it. use s///" :(
so thanks for that
colomon sjohnson: I'm surprised you got this far without hearing about it, it was the only way to do it until moritz_++ got s// working in spring
flussence the :16 is pretty non-obvious too
(is that part of the :param syntax, or is it a special case?) 18:46
sjohnson is there a nice way to do it in p5 too?
TimToady it's special
18:46 hudnix joined
TimToady sjohnson: recent P5 added /r to return the value 18:46
but the .subst vs .=subst is a P6ism 18:47
sjohnson im glad subst is around
one of the few things i actually like about PHP
another good thing about PHP is being able to push onto arrays using $array[] = $val; (square brackets)
.oO(i like .push better tho)
MindosCheng sjohnson : I like the $array[] = $val; thing :) 18:48
sjohnson MindosCheng: don't forgot to do $a = array() before, otherwise it won't wokr 18:49
rakudo: my $house = "0843ffffffffffffewm".subst('f','aA'); say $house 18:50
p6eval rakudo 015d77: OUTPUT«0843aAfffffffffffewm␤»
MindosCheng sjohnson : Soga.
sjohnson oops
rakudo: my $house = "0843ffffffffffffewm".subst('f','aA', :g); say $house 18:51
flussence That house isn't up to building code.
p6eval rakudo 015d77: OUTPUT«0843aAaAaAaAaAaAaAaAaAaAaAaAewm␤»
sjohnson thats better
rakudo: say :g
p6eval rakudo 015d77: OUTPUT«Unexpected named parameter 'g' passed␤ in 'say' at line 5505:CORE.setting␤ in main program body at line 22:/tmp/g3974lHYMO␤»
sjohnson
.oO( ????? ! )
moritz_ rakudo: $*OUT.say(:g) 18:52
p6eval rakudo 015d77: OUTPUT«␤»
moritz_ sub vs. methods
sjohnson is :g just a symbol that stands for some perl recognized integer on a big list of special numbers? 18:53
or some non-printable character that i can't see
TimToady it's a named arg that is being ignored by the method
see S12:2099 18:54
18:55 KyleHa left, KyleHa joined, KyleHa left
sjohnson rakudo: say 0x65 18:56
p6eval rakudo 015d77: OUTPUT«101␤»
sjohnson whoops
moritz_ rakudo: say :g.perl 18:57
p6eval rakudo 015d77: OUTPUT«"g" => Bool::True␤»
MindosCheng rakudo: $_='FFF'; s/F/Aa/g; say
p6eval rakudo 015d77: OUTPUT«===SORRY!===␤Unsupported use of /g; in Perl 6 please use :g at line 22, near "; say"␤»
sjohnson rakudo: say "mamama".subst('ma', 'mi', 1); 18:58
moritz_ rakudo: $_='FFF'; s:g/F/Aa/; .say
p6eval rakudo 015d77: OUTPUT«No applicable candidates found to dispatch to for 'subst'. Available candidates are:␤:(Mu : Any $matcher, Any $replacement, Any :samecase(:ii($samecase)), Any :samespace(:ss($samespace)), *%options)␤␤ in main program body at line 22:/tmp/HErOqHp08U␤»
rakudo 015d77: OUTPUT«AaAaAa␤»
colomon #phasers in t-2
sjohnson rakudo: say "mamama".subst('ma', 'mi', Bool::True);
p6eval rakudo 015d77: OUTPUT«No applicable candidates found to dispatch to for 'subst'. Available candidates are:␤:(Mu : Any $matcher, Any $replacement, Any :samecase(:ii($samecase)), Any :samespace(:ss($samespace)), *%options)␤␤ in main program body at line 22:/tmp/9lOqx30EMx␤»
moritz_ sjohnson: what do you want to express with the third argument?
sjohnson moritz_: just trying to reverse engineer p6 guts for learning purposes 18:59
MindosCheng moritz_++
18:59 hudnix left
TimToady you can't pass a positional argument to a parameter that is declared named only 18:59
moritz_ sjohnson: you know, there's also source code you can look at :-)
MindosCheng For blackbox testing :P 19:00
sjohnson trying to bone up on my p6 obfuscated code for future contests too!
19:00 [Coke] joined
colomon jnthn: error is happening somewhere in the last two lines here: 19:00
if $right.returns eq 'WhateverCode' {
$right_new := PAST::Op.new( :pasttype('call'), :node($/), $right);
my $total_arity := $counter + $right.arity;
19:03 MayDaniel joined
jnthn colomon: ...weird. 19:03
colomon yes
but "say" doesn't lie. ;)
guess I can narrow it down to the line... 19:04
19:04 hercynium joined
sjohnson TimToady: what are your current thoughts on p6 these days 19:05
in a broad "how's it going?" kind of sense
MindosCheng reading historical documents in the middle night.
19:05 [Coke] left
hercynium needs to spend more time here :) 19:06
sorear hello
MindosCheng s/middle/mid/
TimToady sjohnson: it doesn't matter how I think it's going; it only matters how it's going :) 19:07
hercynium "At that moment, the student was enlightened"
sjohnson haha 19:08
p6koans
:) 19:09
MindosCheng hercynium++ 19:10
jnthn colomon: Does this help: $total_arity := $counter + +$right.arity; 19:12
(if yes then, srsly eww.)
colomon my $total_arity := $counter + $right.arity; is definitely the line that is failing
diakopter +($right.arity)
er
jnthn Same ;)
colomon not needed in p6, maybe needed in nqp? 19:13
jnthn the parens?
diakopter find coffee
TimToady what happens if $right doesn't have a .arity?
jnthn Oh wait, that's .arity on the PAST::Block
MindosCheng was enlightened with the koan of lazy evaluation.
jnthn OK, REALLY wtf. :/ 19:14
TimToady hopes to be enlightened lazily...
dalek ecza: 13d3657 | sorear++ | v6/tryfile:
[v6] fix order-of-operations bug in package_def
 19:15
19:15 hudnix joined
hercynium TimToady: I tried that but without a restricted calorie diet it's impossible 19:15
diakopter sorear: v6 ?
sorear where I'm trying to build the bootstrapped niecza 19:17
as opposed to the Perl 5 one
I'm still just chasing STD.pm6 though
*almost* there
diakopter oh
colomon jnthn: +$right.arity doesn't help 19:19
dalek ecza: 4f1401b | sorear++ | docs/lhf.txt:
Remove two stale LHF items
 19:20
19:26 MayDaniel left, masak joined
jnthn colomon: Um. Hmm. 19:28
jnthn is confused
Util Does NQP support .match and .subst as methods? I have only gotten them to work as subs.
colomon jnthn: not half as confused as I am 19:29
jnthn What is $counter?
colomon 0
now counter 0
now right.arity 1
jnthn Initialized as my $counter := 0;
? 19:30
colomon (those are really +$counter and +$right.arity)
recompiling to get ~$counter and ~$right.arity 19:31
jnthn maybe pir::typeof__SP($counter) too 19:34
I'm curious what the two things actually are
colomon still say 0 and 1 when ~ified 19:36
Lorn repo?
19:36 risou_ left, rindolf joined
colomon jnthn: recompiling with typeof now 19:37
19:37 TheHarlot joined
TheHarlot okay... a few questions. Is it possible to do static variables in Perl 6 classes? If so, what sigil do I use? 19:37
jnthn TheHarlot: Just declaring a lexical in the class body should have that effect. 19:38
(e.g. one of them, visible just inside the class) 19:39
TheHarlot second: is reading the contents of a directory any different from reading from a file? If so, how different? Or can I just do @list = slurp("/var/my_file_pool"); and Perl 6 will figure it out.
jnthn e.g. use my
No, slurp is only for files.
TheHarlot jnthn, ah, I was going to jump over to "our".
jnthn my @files = dir('/the/path')
TheHarlot ah, thank you. 19:40
jnthn TheHarlot: Well, that is visible from outside teh class.
TheHarlot: Depends how much encapsulation you want :)
TheHarlot well, I have a weighted count in this class. So, one class will have 50, others will be weighted at 30, 41, 25, etc. The variable is mostly just to hold a cached value of the total weight of these objects.
I will be reading a directory in, making objects based on that directory--with an initial weight (that changes over the program). The next step is that I randomly choose a entry based on the total weight, and shoot it off to a program, to play until finished. 19:42
masak TheHarlot: see strangelyconsistent.org/blog/i-can-haz-constant
TheHarlot which gives the next issue: how do I call an outside program in Perl6, and return when that program exits.
TimToady I guess the only problem with "my" is if you need it modifiable from subclasses
TheHarlot I am thinking something like how the static keywork can work in C for variables. 19:43
masak TimToady: then it should be public anyway.
TheHarlot er... c++
masak s/should/should probably/
TheHarlot thanks masak, jnthn and Larry ~.n
jnthn TimToady: Well, that's consistent with has attributes, which are also only visible inside the class :)
colomon jnthn: both types are Integer
TimToady TheHarlot: we have a "state" keyword that is like static 19:44
it makes a variable persistent across invocations
TheHarlot yeah, that is probably what I want.
TimToady but a class body is only invoked once, so it doesn't buy you anything
TheHarlot uh--does rakudo have a different version of perldoc, btw?
TimToady if you put it into a method body, though, it would persist 19:45
19:45 hanekomu joined
TheHarlot TimToady, a bit counter intuitive--but I could put it in a weight tracker method, I guess? 19:45
TimToady you could
TheHarlot hmmm... or when updating the weighting, I could have that function return the new total weight. 19:46
sorear wants to know why it's not possible to use semantic tags with code
well, intermixed I mean
TheHarlot which is not a pretty solution.
TimToady I don't think rakudo comes with a perldoc
masak TheHarlot: see S26 for Perl 6's variant of Pod. only a small part of it is in Rakudo so far.
jnthn colomon: As they should be. :S 19:47
masak oh, the program? no, it doesn't.
colomon jnthn: I was afraid of that
masak re-remembers u4x
TheHarlot masak, TimToady: I am mostly look for an easier way to rtfm on that one.
colomon jnthn: anything else you can think of trying here? 19:48
TimToady at the moment the synopses at perlcabal.org/syn/ are still the most comprehensive docs, though they're written more for implementers than users 19:49
masak TheHarlot: not much beats the synopses right now in terms of rtfm. for all else, I'd recommend coming here.
sorear Is Darren Duncan insane?
jnthn colomon: I'm just finding it incredibly bizzare that adding two Integers together could lead to the error we're seeing...
TimToady sorear: not really, but he's pretty far out on the Spectrum 19:50
colomon jnthn: I couldn't agree more!
sorear TimToady: what is the Spectrum?
jnthn Why, the good old ZX of course. 19:51
masak :)
flussence k
whoops 19:52
TimToady the Autistic Spectrum, which most of us here are on somewhere
diakopter some of us even move along it over time 19:53
masak dAut
er, dAut/dt
TimToady we all have our little failures in pragmatics from time to time
jnthn I never do that!
;)
TimToady jnthn: you are anchoring the Neurotypical End :) 19:54
masak jnthn: don't you dare leave! :P
MindosCheng rakudo: map { $_++ } <TimToady diakopter> 19:55
p6eval rakudo 015d77: OUTPUT«===SORRY!===␤Confused at line 22, near "map { $_++"␤»
sjohnson heh
TheHarlot hmmm... just found my method was a pretty bad.
TimToady MindosCheng: needs a comma after the block to parse
TheHarlot though, I did figure out more on this issue.
TimToady, I technically was only on the Autistic Spectrum due to a misdiagnosis. 19:56
TheHarlot scratches her head
masak replied to Darren Duncan
TheHarlot: were you really supposed to be diagnosed with some other spectrum? 19:57
TimToady TheHarlot: that just means you're on the normal end :)
TheHarlot long story short, lazy doctors ended with me being on the chemical equivalent to PCP for about twenty years, as after being medicated for Aspergers for two years, they decided I was Bipolar.
And gave bipolar meds to somebody who is not actually bipolar... 19:58
MindosCheng rakudo: say map { $_++ }, <TimToady diakopter>
p6eval rakudo 015d77: ( no output )
TheHarlot TimToady, pfft! that is like saying gamma radiation or AM Radio is part of the visible light spectrum ~.n
19:59 jaldhar left
TimToady I think the Autistic Spectrum is more like the universe, which has no "outside" 19:59
at least on the neurotypical end, you can't be more normal than "Normal" :) 20:00
you can find other dimensions to go off into though...
20:01 mj41 joined
TheHarlot well in full honesty. There are only two definitions of normal I have been given. One is Situation Normal (SNAFU), and the other is the act of washing dishes. 20:01
MindosCheng It's multi-dimensional, more than that in the universe. 20:02
TimToady sometimes normal just means "at right angles", which implies another dimension...
hercynium in another dimension, my dishes are all clean
colomon jnthn: here's a gist, just in case there's something stupid I'm doing: gist.github.com/702404 20:03
TheHarlot In the case of the Autistic Spectrum is is a set of issues in one are of human life. Like Visible Light. You can still jump over to radio waves. Which have their own divisions (AM, FM, UHF, etc)--and work their own way... but they are no visible light. They however are the same universe. Just a different subset of that universe.
Unless you imply stuff like Schizophrenia, Bipolar--and quite a bit of mental disorders that are not part of the Autistic Spectrum are "normal". 20:04
masak TheHarlot: I'm just waiting for the moment when you tie all this back to Perl 6 somehow.
TheHarlot ah... I was commenting on something somebody else said.
TimToady so what's "normal" Perl 6 code :)
TheHarlot That everybody here is part of the Austism Spectrum. 20:05
masak :)
TheHarlot TimToady, by my two definitions, either a jeph or a dish washer program.
TimToady :)
sorear austism, eh? 20:06
rakudo: say $*IN.slurp
p6eval rakudo 015d77: OUTPUT«Land der Berge, Land am Strome,␤Land der Äcker, Land der Dome,␤Land der Hämmer, zukunftsreich!␤Heimat bist du großer Söhne,␤Volk, begnadet für das Schöne,␤vielgerühmtes Österreich,␤vielgerühmtes Österreich!␤␤Heiß umfehdet, wild umstritten␤liegst dem Erdteil du inmitten,␤einem
..stark…
masak snaps at attention 20:07
TimToady attention snaps back
masak that's Austrism, not Autism. :P
attention whiplash.
TimToady and vice versa 20:08
TheHarlot I mean, if we were to put Perl 6 code to the DVM (or whatever the TLA is), as an analogy, each and every piece of code has problems it can have. With Perl 6 code that can be attributed to any mental illness possible.
TimToady provably, we don't have ADD here
TheHarlot Autism is just one spectrum the code could be screwed up. You can have Schizo Code, and Bipolar code as well.
20:08 fhelmberger joined
TheHarlot The code would only be considered ill if it has those issues to levels that it no longer functions at its goal. 20:09
TheHarlot blinks
sorear masak++
TimToady as we get closer to true AI, this may become truer than we'd like to think...
MindosCheng rakudo: say <Autism not-Autism>.pick 20:10
p6eval rakudo 015d77: OUTPUT«Autism␤»
masak rests his case
TimToady wrests his case
TheHarlot right--and my final question for the moment: how do I call an outside program in Perl6, and return to the Perl6 code, once done being called?
hercynium crests his race
masak TheHarlot: run() 20:11
TheHarlot thank you masak
now I can go trying to code this piece.
masak you're welcome, harlot.
jnthn But it's a good idea to pass it an argument ;)
masak run("ls")
TheHarlot jnthn, I dunno--maybe I could use that to get the voices to talk to my program ~.n
hercynium code in piece
TimToady the return value is the opposite of p5's system()
TheHarlot TimToady, returns true on success then? 20:12
TimToady correct
masak TimToady: only Rakudo doesn't do that yet.
TimToady that's spec, anyhoo
colomon boy, that seems like it ought to be an easy bug to fix...
TheHarlot Not really concerned about return value. I am mostly writing a music player right now. All I care about is way to randomly play songs based on a directory listing.
masak nice! 20:13
TheHarlot I will post a blog, as I start to work on being at least an EPO Tinman.
TimToady should do for htat
*that even
hercynium sane return values from system commands? NOE!
(as opposed to 0E0?)
MindosCheng TheHarlot++
TheHarlot MindosCheng, the name in the system would be KatrinaTheLamia 20:14
I am just running this, as my shell is currently down. 20:15
sjohnson rakudo: say
95 # Maybe in the future... I will add something similar to cfvi, where it gets rid of template_c and .swp files
p6eval rakudo 015d77: OUTPUT«␤»
sjohnson oops
note to self: always check paste buffer
masak sjohnson: that's actually a known rakudobug.
std: say
sjohnson haha
p6eval std a194beb: OUTPUT«Potential difficulties:␤ Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/zFrPifZjqj line 1:␤------> say⏏<EOL>␤ok 00:01 117m␤»
sjohnson i'm good at finding bugs by accident 20:16
jnthn dresses masak's say
sjohnson rakudo: say 0b1100101 # what i really meant to do
p6eval rakudo 015d77: OUTPUT«101␤»
TheHarlot rakudo: Any() + 50
p6eval rakudo 015d77: OUTPUT«Could not find sub &Any␤ in main program body at line 22:/tmp/2G2R9k1iMZ␤»
masak rakudo: say Any + 50
p6eval rakudo 015d77: OUTPUT«Use of uninitialized value in numeric context in <anon> at line 22:/tmp/e9xf08xgKn␤50␤» 20:17
TheHarlot rakudo: my $var; say $var + 50;
p6eval rakudo 015d77: OUTPUT«Use of uninitialized value in numeric context in <anon> at line 22:/tmp/nhOJnqa9Gm␤50␤»
moritz_ doesn't understand why it dies instead of just warning
jnthn moritz_: It printed the 50 too
TheHarlot right... going to need to do something to initialise a variable in a class.
moritz_ oh, right
TheHarlot wait--it does not die then? 20:18
rakudo: my $var = 50; say $var;
p6eval rakudo 015d77: OUTPUT«50␤»
TheHarlot compares output
ah... that character is a newline... I guess? 20:19
sjohnson rakudo: say rand
p6eval rakudo 015d77: OUTPUT«0.663876019043045␤»
TimToady .u ␤
phenny U+2424 SYMBOL FOR NEWLINE (␤)
TimToady you can use it on input too 20:20
std: { 1 } { 2 } 20:21
p6eval std a194beb: OUTPUT«===SORRY!===␤Unexpected block in infix position (two terms in a row, or previous statement missing semicolon?) at /tmp/CSMpbfLQYw line 1:␤------> { 1 } ⏏{ 2 }␤ expecting infix or meta-infix␤Parse failed␤FAILED 00:01 119m␤»
TimToady std: { 1 }␤{ 2 }
p6eval std a194beb: OUTPUT«ok 00:01 118m␤»
TimToady TheHarlot: if you have a utf-8 aware irc client, and if you have good font support, it should come out as a little NL of some sort 20:22
MindosCheng using ChatZilla now. 20:24
I am still confused with the closure example. 20:25
sorear gets some popcorn and watches masak and dduncan
MindosCheng rakudo: my $a = -> { my $c=10; return { say $c-- } }(); $a(); $a();
p6eval rakudo 015d77: ( no output )
masak sorear: I don't intend to continue that thread.
MindosCheng Why it has no output? 20:26
sorear because the 'return' statement ended the program
masak MindosCheng: something looks broken.
oh, right.
sorear pointy blocks don't catch return, so the exception bubbles up and caused the eval() frame to return
masak sorear++
MindosCheng sorear++ 20:27
rakudo: my $a = -> { my $c=10; -> { say $c-- } }(); $a(); $a();
p6eval rakudo 015d77: OUTPUT«10␤9␤»
TimToady rakudo: my $a = sub { my $c=10; return { say $c-- } }(); $a(); $a(); 20:28
p6eval rakudo 015d77: OUTPUT«===SORRY!===␤Confused at line 22, near "my $a = su"␤»
masak TimToady: needs parens around sub {...}
TimToady rakudo: my $a = (sub { my $c=10; return { say $c-- } })(); $a(); $a();
p6eval rakudo 015d77: OUTPUT«10␤9␤»
TimToady std: my $a = sub { my $c=10; return { say $c-- } }(); $a(); $a();
p6eval std a194beb: OUTPUT«ok 00:01 122m␤»
jnthn Eww!
TheHarlot TimToady, it comes out as two hex quadlets in my client's font.
rakudo: say rand(9001); 20:29
p6eval rakudo 015d77: OUTPUT«===SORRY!===␤Unsupported use of rand(N); in Perl 6 please use N.rand or (1..N).pick at line 22, near "(9001);"␤»
TheHarlot rakudo: say 9001.rand()
p6eval rakudo 015d77: OUTPUT«691.640323830381␤»
jnthn TimToady: Granted Rakudo should follow STD, but I'd rather STD didn't let you do that tbh.
20:29 plainhao left
TheHarlot rakudo: say Interger 9001.rand() 20:29
jnthn TimToady: Just visually looks...weird.
p6eval rakudo 015d77: OUTPUT«Could not find sub &Interger␤ in main program body at line 22:/tmp/kgjWXbY5_v␤»
TheHarlot rakudo: say Integer 9001.rand()
p6eval rakudo 015d77: OUTPUT«Could not find sub &Integer␤ in main program body at line 22:/tmp/6TB35dbiOg␤»
TimToady TheHarlot: do you get french quotes here: « Oui! »
masak TheHarlot: read the error message.
TheHarlot TimToady, I get the unicode version of << and >> 20:30
20:30 daxim left
jnthn TimToady: I get the same dissonance in C# where you can do like "var x = new Foo().Bar()" - my brain always groups that wrongly. Maybe that's just my brain though :) 20:30
TimToady good, then it's probably just a font issue
TheHarlot masak, yeah... trying to figure out howto type cast... then I just decided to my code does not need it.
masak rakudo: say 9001.rand 20:31
p6eval rakudo 015d77: OUTPUT«8506.55472686801␤»
masak rakudo: say 9001.rand.int
p6eval rakudo 015d77: OUTPUT«Method 'int' not found for invocant of class 'Num'␤ in main program body at line 22:/tmp/ONztBUqdoE␤»
masak rakudo: say 9001.rand.Int
p6eval rakudo 015d77: OUTPUT«8308␤»
TimToady rakudo: say (^9001).pick
p6eval rakudo 015d77: OUTPUT«4014␤»
TimToady rakudo: say (^90010000).pick
TheHarlot rakudo: say Int(9001.rand)
p6eval rakudo 015d77: OUTPUT«(timeout)»
rakudo 015d77: OUTPUT«Could not find sub &Int␤ in main program body at line 22:/tmp/0E7b1ouLyj␤»
TimToady tell me it's not generating a list from ^90010000... 20:32
TheHarlot hmm... okay, I will note that behaviour on objects for the future.
masak TimToady: seems it is :(
sorear TheHarlot: め # Here, have a very common Unicode character >256
masak TimToady: or maybe .pick is flattening it out.
TheHarlot rakudo: say (^100000000000).pick
p6eval rakudo 015d77: OUTPUT«(timeout)»
sorear « is <256 so it's not necessarily a fair indicator of Unicode support 20:33
TimToady troo
but I think there would at least have been some garbage
20:33 lue left, lue joined
XaRDaX Hi all. I've a question: Is there a Module in p6 like IO::Socket in perl5? 20:33
TimToady extra  and such
sorear alpha had an IO::Socket, iirc 20:34
TheHarlot sorear, the first one is a japanese character... my mind is spacing out on the charset (I keep wanting to say "hanashita"... but that is wrong... so bloody wrong)
sorear .u め
phenny U+3081 HIRAGANA LETTER ME (め)
TheHarlot Hiragana! Haha!
colomon .pick is flattening it out, because 100000000000 isn't really an Int in Rakudo. 20:35
rakudo: say (^1000000000).pick
p6eval rakudo 015d77: OUTPUT«(timeout)»
MindosCheng rakudo: say <め 安>
colomon rakudo: say (^100000).pick
p6eval rakudo 015d77: OUTPUT«め安␤»
rakudo 015d77: OUTPUT«(timeout)»
TimToady rakudo: say 90010000.WHAT 20:36
p6eval rakudo 015d77: OUTPUT«Int()␤»
TimToady that is
dalek odel: 021ee7e | jonathan++ | dotnet/ (4 files):
[dotnet] Add an LHF entry.
colomon yeah, something weird is going on here.
TheHarlot well, seeing as how my random music player is using NIALL based weighting, I do not need it to be an Int(). Though, which takes up less space, Int() or Float()?
Well, less space, and less cycle abuse? 20:37
sorear TheHarlot: hanashita is at least a word, and has the right initial consonent
consonant.
TimToady "hanashita" wo hanashita!
XaRDaX sorear: do you mean v6-alpha?
TheHarlot sorear, it means "speak" with tense of past but continuing on.
colomon rakudo: say (^1000000).max.WHAT
p6eval rakudo 015d77: OUTPUT«Int()␤»
colomon rakudo: say (^1000000).min.WHAT
p6eval rakudo 015d77: OUTPUT«Int()␤» 20:38
TheHarlot That is why I knew it was the wrong word for hirogana.
sorear alpha: say 2 + 2
p6eval alpha : OUTPUT«4␤»
colomon ooooo.
sorear alpha: say IO::Socket.^methods
p6eval alpha : OUTPUT«sendcloserecvsqrtrindexasinintatanhcoshcosecgrepvaluescharsrootsreverseisaucfirstkvsamecasetanlog10coscapitalizefmtbytesflipdoesp5choplcasinhacotanreducecotanhchracotanh:d:e:fcombfloorasechatanacos:levalroundchoptanhsplitmatchexp:sacoshwordscancotanatan2lcfirstucsrandkeyspolarmapci…
colomon I think the Range.pick patch a few weeks ago broke it.
TheHarlot wait... there is more than just rakudo in p6eval?
pugs: say "I am a perl 6." 20:39
p6eval pugs: OUTPUT«I am a perl 6.␤»
TimToady TheHarlot: that's okay, most of use are a bit weak on our 日本語 here.
s/use/us/ # our English too
sorear XaRDaX: 'alpha' is a retronym for Rakudo 2010.01
yapsi: say "hi"
p6eval yapsi: OUTPUT«===SORRY!===␤Unable to find module 'Yapsi' in the @*INC directories.␤(@*INC contains:␤ lib␤ /home/p6eval/.perl6/lib␤ /home/p6eval//p2/lib/parrot/2.9.1-devel/languages/perl6/lib␤ .)␤»
20:39 wamba left
sorear nqp: say("Hi") 20:40
p6eval nqp: OUTPUT«Hi␤»
TheHarlot TimToady, well, I am strong enough, that I barely use subtitles in anime... mostly for harder words or when I do a double take on what a character said.
sorear perlesque: say("Hi")
p6eval perlesque: ( no output )
20:40 jedai left
TheHarlot TimToady, mostly with the hanashita/hiragana thing is with a similar issue in English. Where something is on the tip of my tongue--and I am trying to figure it out. 20:40
jerl6: say "I should not be here"
ah... it returned the proper output. 20:41
20:41 jedai joined
sorear almost everyone here is at least learning Japanese 20:41
a funny coincidence
MindosCheng pugs: my $a = -> { my $c=10; return { say $c-- } }(); $a(); $a();
p6eval pugs: ( no output )
XaRDaX sorear: thank you
TimToady
.oO("Know your enemy...Ruby")
masak goes offline for battery reasons
20:41 masak left
MindosCheng sorear: I only know a bit Japanese... 20:42
pugs: my $a = -> { my $c=10; -> { say $c-- } }(); $a(); $a();
p6eval pugs: OUTPUT«10␤9␤»
TheHarlot ah--I am fairly solid in it... not fluent. But solid. I am also working on Mandarin and soon German. I also want to look into Spanish, so I can finally watch a soccer game, and not hate the announcer.
MindosCheng masak: Cya.
TheHarlot HATES the british football announcers. sooo much.. hate...
20:43 fhelmberger left
MindosCheng TheHarlot: Mandarin is my native language... zh-TW 20:43
TheHarlot ah, well, from what I have seen of it, it is a brilliant language to know.
20:44 stkowski left
sorear Audrey was zh-TW; I wonder if she started this trend 20:44
20:44 PerlJam joined
TheHarlot I mean, if I could choose a language to be spoken world wide instead of English, it would be German, Mandarin or Sanskrit. 20:44
MindosCheng sorear: I happen to know Audrey long time ago.
TheHarlot: It's nice to know so many language. 20:45
TheHarlot The notion of making English a standardised academics language or language of commerce, makes about as much sense to me, as doing the same with Ebonics, Welsh, Newfie or Cockney Slang. 20:46
MindosCheng, it is. Though, I do not really consider English a language.
TimToady it's really a creole, like Japanese
but creoles can become real languages, for some definition of "real" 20:47
if communication happens, it's a language
TheHarlot TimToady, though, Japanese at least has slightly better tenses to deal with it.
MindosCheng sorear: That's why I started to learn Perl.
20:48 takadonet left
hanekomu and with Perl 6, Perl becomes a creole as well 20:48
TimToady TheHarlot: depends on what you're trying to express
you can't say "will have been going" in Japanese, fer instance
PerlJam hanekomu: you don't think Perl 5 is a creole?
TheHarlot I hate making use of tenses in English. Hate, hate, hate! Especially since I have started looking into German, Mandarin and Japanese 20:49
TimToady every language has its warnts
hanekomu PerlJam: there's more than one way to do it, but there's only one perl interpreter
TheHarlot TimToady, well, try explaining temporal mechanics in English.
TimToady *warts
moritz_ the usage of tenses in English isn't so much different from German
TheHarlot moritz_, my understanding was German has more tenses though
TimToady since I have not idea what temporal mechanics are, it's will be difficult for me to explain them in any language 20:50
*it
if you want lots of tenses, try classical Greek
hanekomu TimToady: English isn't my native language, so I'm not even sure what "will have been going" means... Does it mean "before X, he will have been going to Y" or some such?
TimToady (using the word "tense" loosely)
TheHarlot TimToady, uh, well, lets say I enjoy analyzing and thinking about plot elements of Dr. Who, Torchwood, Back To The Future--and most stuff from the Discordians.
Though, conveying what exactly happened, is next to impossible in english. 20:51
Or, if it is possible, only in the least elegant manners.
It also sucks for planning stuff.
As I tend to plan for different opportunities. And, the inability to give solid tenses in those plans is irritating in English.
colomon Range.pick fixed locally, starting spectest ASAP. 20:52
TimToady hanekomu: by point X in the future, it will be the case that he is already going, and is still going at point X
PerlJam The world needs less tenseness, not more :)
sorear when you call English a creole, are you referring to the Norman Invasion and related events?
TimToady but English mixes up its tense system with its aspect system rather badly
sorear: among other things 20:53
TheHarlot PerlJam, well, by adding proper tension to the English language, you will remove some chaos.
PerlJam TheHarlot: but I *like* chaos!
TheHarlot PerlJam, well, yeah--but that chaos will appear in other, better uses of it.
TimToady the craziest thing about English is that it borrowed its *pronoun* system from the Vikings
TheHarlot As now, you no longer have all that perfectly good chaos being wasted on order of events, you can now spend it on better things. 20:54
too many people are worried about the conservation of energy, the conservation of matter... what about the conservation of insanity? 20:55
hanekomu looks at a table of english tenses
The horror...
TimToady rosettacode.org/wiki/Discordian_date#Perl_6
PerlJam Hmm. I don't think I want to conserve insanity. Seems better to spend insanity freely.
TheHarlot we need a way to more inefficiently use the insanity we all have. 20:56
hanekomu But it seems to me that in Japanese you can use auxiliary verbs to get to some of those tenses
TheHarlot PerlJam, the thing is, we are only given a small amount of madness. We need to make the best use of it.
hanekomu though IANAL (I Am Not A Linguist)
TimToady well, Japanese slices their verb system much differently than English
TheHarlot We lose our madness by wasting it on describing order of events--we do not have that madness for other truly crazy things! 20:57
20:57 donaldh joined
sorear wonders if hanekomu is a native 20:58
20:58 donaldh_ joined
hanekomu native German speaker, yes 20:58
:P
TimToady hanekomu: it's true that you can convey any idea if you work at it; the differences between languages are more in what you have to say, not it what you can say
hanekomu true
20:58 donaldh__ joined
PerlJam wonders what you have to say to talk about future events in a language with no future tense 20:59
TimToady wonders why hanekomu's feathers are crowded...
sorear PerlJam: that depends on what you mean by "no future tense"
Japanese has neither a present tense nor a future tense, but rather a single not-past tense 21:00
TheHarlot TimToady, well, newspeak and duckspeak do have things you cannot say in them. But that only applies to people in Great Britain, and other Orwellian countires.
Tene I expect people in here will find this interesting; www.guardian.co.uk/books/2010/nov/0...r-language
hanekomu TimToady: feathers? crowded? what? :)
sorear if you want to say that something will happen tomorrow, you use the word for tomorrow + not-past tense
TimToady 羽はねhanefeather; plume; wing
TheHarlot PerlJam, oh, you set the stage. "this is tomorrow", and work from there in your sentence.
hanekomu ah that :)
TimToady 込む 混むこむkomuto be crowded
hanekomu it's a Go term
senseis.xmp.net/?Hanekomi 21:01
sorear TimToady: oooh, what's that from?
hanekomu sorear: from this: search.cpan.org/~marcel/?R=D
TimToady a version of Jdict that has been rendered greppable
hanekomu oh
nice
21:02 donaldh left, donaldh__ is now known as donaldh
hanekomu other "hane" actually 21:02
TheHarlot hanekomu, ah! You may be the person to talk to in a month or so, when I have Jerl 6 at a point, that I can use it to put stuff onto the Android Market.... as one think the Driod market is missing, is a good Go game.
21:02 donaldh_ left
sorear TheHarlot: Jerl 6? 21:02
TimToady more like a "wing" kind of hane 21:03
TheHarlot sorear, Perl 6 for the JVM.
sorear TheHarlot: how far along are you?
MindosCheng 羽翼
TheHarlot sorear, sorry, I have not mentioned it in here in forever, and I have done nothing on it.
What being normal and all (my SNAFU life)
TimToady or maybe from haneru
hanekomu TimToady: yes 21:04
TheHarlot sorear, however, I am in a state, that I am getting stuff pushed out on the two main components for it. libNIMh and Java Kitty. Java Kitty being a JVM Dead Cat Hub Thingy.
TimToady that's one of those verbs that, though it uses several different kanji, I rather suspect are related etymologically 21:05
TheHarlot (I have no better way to describe how Java Kitty relates to the JVM)
hanekomu A "hane" move in Go is usually translated as a "bend"; it means to bend around the opponent's stone. But I guess it's more like "jump in front of the opponent, blocking his way". This would fit well with the 込む part.
TheHarlot rakudo: my $x = 0; while $x < 10 { $x++; next} 21:07
p6eval rakudo 015d77: ( no output )
TimToady I'd guess it's 跳ねる, to jump; to leap; to prance; to spring up; to bound; to hop
TheHarlot rakudo: my $x = 0; while $x < 10 { $x++; continue; }
p6eval rakudo 015d77: OUTPUT«Could not find sub &continue␤ in main program body at line 22:/tmp/WIwtDm64Di␤»
hanekomu TimToady: yes
TheHarlot ah... it is next I want, definitely.
TimToady as in a "jump" in checkers, kinda
TheHarlot: if you're partly insane, you've come to the right place 21:09
hanekomu :)
TimToady it's part of our mandate here to figure out how to put our collective insanities to good use
and if you're differently insane from the rest of us, all the better 21:10
diakopter warnts a nap 21:11
TheHarlot TimToady, I have been her before. Under the name "KatrinaTheLamia"... I am just using this, because my shell is currently not here (for about ten days). I may^H^H^H should update irssi to join this channel again. 21:12
been here before* 21:13
I am going with this, as I cannot be bothered to set up this IRC client properly. Laziness really.
sorear hane is also translated as hook or bend
21:13 timbunce joined
hanekomu "bend" fits 21:13
sorear my teacher used that word quite a lot when talking about drawing hiragana
TimToady probably because wings bend 21:14
sorear TheHarlot: what do libNIMh and Java Kitty do? what kinds of stuff can Jerl6 programs do?
sorear wants to compare niecza against it now 21:15
TimToady now warnts a nap, and blame diakopter... 21:16
*s
diakopter boorish/booring 21:17
hanekomu pretty sure now it's the "bend" meaning, since 跳ね (hane) is 젖힘 (jeotchim) in korean, coming from 젖히다 = bend (back). 21:18
hanekomu is satisfied now. :)
diakopter er, and I don't mean the 2nd usage listed on urbandi********.***
sorear error: quantifier quantifies nothing
diakopter std: ???.?++ 21:20
p6eval std a194beb: OUTPUT«ok 00:01 118m␤»
diakopter fix THAT
;)
oooo 21:21
rakudo: ???.?++
p6eval rakudo 015d77: OUTPUT«===SORRY!===␤Method 'isa' not found for invocant of class 'Undef'␤»
21:21 timbunce left
TheHarlot rakudo: run("ls -F") 21:22
p6eval rakudo 015d77: OUTPUT«Operation not permitted in safe mode␤ in 'Safe::forbidden' at line 2:/tmp/I8nNGZFTB0␤ in main program body at line 22:/tmp/I8nNGZFTB0␤»
diakopter there are ways around that, but please don't :)
TheHarlot sorear, libNIMH is a set of frameworks for dealing with the OS, that I have gotten irritated with programming over and over. 21:23
sorear, Java Kitty is a set up for developing onto the JVM. It will allow debugging tools for Android, Applets, Servlets and what not. Running as a JVM, with multiple entrances.
sorear, Jerl 6, will be Perl 6, but targetting the JVM to compile onto. Having access to all the classfiles and what not a Java program would. With two versions, Jerl 6 written in C--which produces .class files and links them to Jerl6.jar. Jerl 6 will also have a version written in Jerl 6, that does not require a prior compile step to deploy. 21:25
21:26 lichtkind left
TheHarlot The steps to the final goal, of Jerl 6, compiling Perl 6 code a run time, will be to get a Jerl 6 compiler done in C--then produce another Jerl 6 compiler--but done in Perl 6. 21:26
21:26 dual left
sorear TheHarlot: what about the present tense? 21:26
TheHarlot it is mostly me blasting hot air. 21:27
diakopter jnthn: do you have AutoCrlf or SafeCrlf (or neither) set in your .git/config
TheHarlot libNIMH has a set of prototype header functions--but the have not been defined. Though, that is mostly a matter of pointer arthemetic for its entirety to define 21:28
moritz_ colomon++
rindolf rakudo: 10000000000000000000000000000000000000000000003 21:29
TheHarlot All the structures and prototypes are there... it is mostly about having the functions get/set and what not.
p6eval rakudo 015d77: ( no output )
rindolf rakudo: say 10000000000000000000000000000000000000000000003
p6eval rakudo 015d77: OUTPUT«8023796054858137603␤»
rindolf What is needed to add bignums to Rakudo?
dukeleto TheHarlot: you will have some fun representing dynamic language semantics in a mostly static virtual machine. Do you have any code in that direction? 21:30
sjohnson rindolf: that's a big number!
rindolf: time
rindolf sjohnson: well, Math::GMP can handle it fine.
diakopter rindolf: if it were that simple to answer, someone would've written down the steps by now, I'd think :)
sjohnson sorear told me that they are comin around the mountain when they come.
rindolf sjohnson: I need this if I am to use Rakudo for Project Euler.
diakopter and riding six white horses.
sjohnson diakopter: :) 21:31
TheHarlot dukeleto, well, it should be easy (for me) to figure out. I mean. the JVM already has Clojure, Jython and JRuby on it.
21:31 timbunce joined, timbunce left
dukeleto TheHarlot: of course. I am sure it will be trivial. 21:31
TheHarlot dukeleto, there will be two versions, like I said, the compile object version, which will use CJerl6, and link it to Jerl6.jar. Then the one that compiles as your run the Jerl6 code. Which will be the J6Jerl6 compiler. 21:32
diakopter dukeleto: :) 21:33
TheHarlot My understanding is that Perl 6, is compiled, at some state in its running.
First I seperate that--which will cause issues, in the over all matter. Then, work to bring it back together, by the matter of writting a Perl 6 compiler in Perl 6. 21:34
PerlJam TheHarlot: just like Perl 5
TheHarlot Which, my understanding is, that Perl 6, is a "real" enough of a language, that it could be used to write a compiler for itself. 21:35
My other option, is to write a Perl 6 compiler in Java, Jython or JRuby... I think I made the best decision here.
dukeleto TheHarlot: i am sure you will find some holes in the spec, however you decide to implement it. 21:36
TheHarlot dukeleto, well, yeah... that is why the spec is not a waterfall spec, but a washing machine spec.
or something to that effect.
I am aware Perl 6 does have a .net port in some stage. And Rakudo, is a port to parrot... which parrot is a more FlOSS version of .net and JVM (well, an equivalent idea). 21:37
Though, once I can get the J6Jerl6 compiler, there is little reason to need the CJerl6 one--as that mostly exists as a bootstrap for the most part. 21:39
dukeleto TheHarlot: perl 6 and parrot were the same project in 2000, and later split.
diakopter [dish-]washing machines are often connected to garbage disposals
TheHarlot well, yes... my understanding is parrot was more of a virtual machine.
diakopter++
dukeleto TheHarlot: if your work could provide a layer of converting JVM bytecode to Perl 6 and back, the parrot project would be very intereted in that 21:40
diakopter TheHarlot: "once I can get the J6Jerl6 compiler" - I'm just curious... what's your "person-hours" estimate on this... decimal order of magnitude
21:41 rindolf left
TheHarlot dukeleto, I am aware, that Parrot started off nearly complete connect to Perl 6. However, the JVM did start off as part of Java. And, I will look into that... however ClassFile to Perl 6 is not going to look pretty in the least bit. 21:41
diakopter, that... well, lets say working five to eight hours a day. five days a week--not expecting it in less than ten to fifteen weeks.
and that is on the J6Jerl6 compiler alone. 21:42
diakopter let's break it down a little
how much of that is the parser/grammar engine
TheHarlot Nothing on CJerl6, nothing of Java Kitty, nothing on libNIMH.
21:42 _kaare left
TheHarlot diakopter, I was not going to break it down, until I got close enough to start the J6Jerl6 portion. 21:42
diakopter, I am still working on the Java Kitty portion. 21:43
diakopter ... which begs the question: how can you estimate 15 weeks ...
dukeleto TheHarlot: i am not worried about how ugly the code is, but Parrot is interested in Parrot Bytecode <--> JVM bytecode interop
TheHarlot diakopter, well 8 * 5 * 15. 21:44
diakopter TheHarlot: 15 weeks.
TheHarlot diakopter, this is based on other projects I have put code into, and how quick the code base filled up.
dukeleto TheHarlot: so if you do something that provides that, parrot is interested. If you can give us perl 6 from jvm bytecode, we can turn that perl 6 into parrot bytecode or PIR/PASM/NQP-rx/whatever
diakopter ok
TheHarlot diakopter, I should have reasonable knowledge of JVM from Java Kitty at that point, and Perl 6 from CJerl6 at that point. 21:45
It is mostly a matter of implementation, and I doubt that there is more than 600 hours worth of typing, to get up to what the specs say on that date. 21:46
diakopter uh-huh
dukeleto TheHarlot: i am very interested to see this happen. Good luck on this important journey. 21:47
TheHarlot diakopter, look, if it sounds like I am full of shit here, it is because you asked the wrong question.
diakopter and nobody ever has to think while not typing
TheHarlot diakopter, again, if I sound full of shit here, it is because you are asking the wrong questions.
diakopter so answer the right ones :P
TheHarlot there are two steps before that, you can ask for time frames on.
sorear I've been hacking niecza for 20 weeks now, and the road ahead is at least as long as the road behind me 21:48
PerlJam 600 person hours spread across 5 years? :)
diakopter I don't think you're full of shit; I was curious what your expectation is about the project's size.
dukeleto TheHarlot: I think everyone wants your ideas to work. We are just skeptical of the timeframes. But please, prove us wrong :) 21:49
21:49 spinclad joined
TheHarlot diakopter, PerlJam: well, the back end to my C JVM compiler setup, that reasonably any language could use (plans for MUMPs, Python 3, Python 2.6 are in the future), are still variable in time. I am hoping to get something out in a month--maybe 2... but that is assuming things do not go horribly wrong here. 21:49
Even then, that time frame could be horridly wrong.
The next step, CJerl6, I honestly could not give a reasonable time frame, and can only promise to shoe out code over the while I figure that out. 21:50
Then, J6Jerl6, which requires me to have completely both Java Kitty and CJerl6, should be no problem.
The J6Jerl6 step is not the problem. It is the two directly before it that are the doozy. 21:51
dalek kudo: 94b52d1 | colomon++ | src/core/Range.pm:
Add back in Range.pick multi which takes no arguments, as it is a huge optimization for large Ranges.
diakopter TheHarlot: it sounds like you have a reasonable overall architecture planned out; might I interest you in contributing to the JVM backend of 6model?
TheHarlot 6model? I will look into that then. 21:52
diakopter mberends has been working on it as of late
TheHarlot anyways--right now, I am hungry, and will be heading to the store for food. 21:53
and, in full honesty, none of you accused me of being full of shit--but if I saw somebody responding the way I did, I probably would suspect the person was full of shit. If only due to a flawed line of questioning. 21:55
XaRDaX Another question for you... How can I write in a SOCKET? In perl5 was print SOCKET ARGS but in p6 doesn't work. Any suggestions? 22:00
moritz_ probably $socket.print('foo') 22:04
PerlJam are sockets even implemented?
jnthn Yes 22:05
XaRDaX yes
found it! $socket.send(); 22:06
PerlJam shows how much I've paid attention to such things :) 22:07
jnthn Well, you should listen() more :P 22:08
22:15 kjeldahl left
MindosCheng rakudo: say #`[早安]"Morning!" 22:17
p6eval rakudo 015d77: OUTPUT«Morning!␤»
22:19 Cyberhawk joined, Cyberhawk left
dalek ecza: 5c951ff | sorear++ | / (3 files):
Re-merge
 22:21
ecza: fe957ea | sorear++ | lib/ (2 files):
Implement lexer subclass reuse optimization
sorear so now niecza-compiled STD can process most of (STD's) CORE.setting without hitting swap 22:23
MindosCheng sorear++
sorear I see I still need 'import' 22:24
22:26 jlaire joined 22:32 mj41 left 22:34 masak joined
masak blog post! strangelyconsistent.org/blog/novemb...ite-revolt 22:35
rakudo: sub foo($x? is rw) { $x = "OH HAI" }; foo()
p6eval rakudo 015d77: OUTPUT«Cannot modify readonly value␤ in '&infix:<=>' at line 1␤ in 'foo' at line 22:/tmp/rnAMWYBIfS␤ in main program body at line 22:/tmp/rnAMWYBIfS␤»
masak discuss.
masak submits rakudobug
Tene masak: www.guardian.co.uk/books/2010/nov/0...r-language
masak looks
Tene masak: A third of the way through or so, I was thinking "Could he use 'is copy' instead?" :) 22:37
masak: I agree in that I expect rw defaults to be a copy/assign. 22:38
masak: Shouldn't be too hard to fix, I expect. 22:39
masak Tene: oh, good.
22:39 hanekomu left, satyavvd left
Tene I mean, the functionality is obviously already there, as 'is copy' can do it fine, so you'd just look at what's different in the PAST, etc. 22:40
masak I had the same thought. 22:41
colomon rakudo: say "hi"
p6eval rakudo 015d77: OUTPUT«hi␤»
jnthn akshually probably an easy-ish binder patch. 22:42
masak \o/
colomon errr... p6eval's rakudo is from Nov 13th. :\
22:43 lichtkind joined
lichtkind is .defineda method any scalar has? 22:43
jnthn It's in Mu
So everything has it pretty much.
lichtkind thank you 22:45
masak Tene: so far, the article is sad and intriguing at the same time.
lichtkind jnthn: i now start to really work on my grant
masak (Dan Kogai)++ # "Real Life on Earth is base-4 coded :-p" 22:47
lichtkind can i say in english Hashmethods ? 22:48
or hash methods
22:49 satyavvd joined
lichtkind masak: i didnt got it :) 22:53
masak lichtkind: the joke about Real Life and base 4?
lichtkind yes 22:54
masak lichtkind: didn't your parents tell you about the birds and the genes?
lichtkind no :)
masak sorry, as a biologist I sometimes take these things for granted. 22:55
lichtkind: every cell in your body contains gigabytes of DNA information. it is base-4-coded. 22:56
lichtkind i alway thought since they are only 2 pairs of acids its actually digital
masak so maybe giganibbles is more accurate...
lichtkind but you say the order matters?
masak it's "digital" regardless.
digital only means that it's isomorphic to numbers. 22:57
lichtkind with digital i mean base 2
masak lichtkind: well, stop that :)
lichtkind you mean a g-t is different to an t-g pair?
masak lichtkind: yes.
lichtkind: the "meaning" emerges during translation, which looks at only one strand at a time. 22:58
lichtkind im more a fan of epigenetics anyway
masak which means that all four of the acgt bases are significant.
lichtkind yeah 22:59
masak is a bit exhausted from the misunderstanding density in the above few lines 23:00
MindosCheng If one string is AAATCG it can encoded as 000123, and the complementary string is CGATTT 230111
, for example.
masak MindosCheng: that's not complementary, that's reverse-complementary.
MindosCheng masak: sorry.
dukeleto dna and relatives are in base 4. Proteins are in base 20, which makes more lots of fun combinations.
s/more lots/lots more/ 23:01
dukeleto was using stack-based thinking for bit there
masak :)
and the mapping is from (RNA) base triplets to amino acids, so there's a theoretical max of 64 possible amino acids. but the mapping is "degenerate" which means that (just as with Perl 6 enums) you sometimes end up on the same output starting from different inputs. 23:02
the mapping is called "The Genetic Code" and was successively uncovered in the sixties. 23:03
since then, people have consistently mistaken the DNA sequence for the Genetic Code, so it was probably a bad name in retrospect. :P 23:04
TheHarlot okay--final question, then I can head off: how do I properly escape variables before shoving them off to run("")? 23:05
masak rakudo: my $variables = '$x $y $z'; say $variables.perl 23:06
p6eval rakudo 015d77: OUTPUT«"\$x \$y \$z"␤»
masak you mean like that?
jnthn masak: Making 'is rw' assigny with default values does kinda break the "it's binding if there is a value" symmetry. 23:08
masak: Though maybe usefulness beats that here. 23:09
masak jnthn: yes, I touch upon that in my blog post.
jnthn masak: I don't feel especially strong either way at first blush.
Would be interested to see TimToady++'s opinion.
masak aye.
but I agree that it's a question of model purity versus convenience. 23:10
jnthn masak: The one to consider is if we can ever get something weird out of
XaRDaX too few positional arguments: 2 passed, 3 (or more) expected -> Is that a bug?
masak XaRDaX: yes. were you using .subst?
jnthn my $a = ...stuff...; sub foo($x is rw = $a) { ... }
XaRDaX no 23:11
jnthn Is there a case where binding vs assignment of $a here would cause issues?
masak jnthn: I was thinking there might be a good reason for it being binding, for example when the default is a sub call.
jnthn Well, or a variable like in the above case.
masak maybe those boil down to the same thing, yes.
jnthn It may have interesting implications if it's an l-value sub too.
masak XaRDaX: then perhaps share a bit more of what you're doing in a nopaste?
XaRDaX ok 23:12
jnthn masak: I think leaving it as binding is the conservative thing to do and - to me - the least surprising. But it seems you've got surprised the other way around. And my views on the binder doing surprising things are doubtlessly tainted by having written the thing. :)
XaRDaX pastebin.com/FDwzuiqD 23:13
masak jnthn: this is exactly the feedback I hoped to get from you.
jnthn: how do you feel about a special case for literals, as Tene++ suggested?
masak looks at XaRDaX's paste
jnthn Oh, I thought we fixed the "variable not declared gives the argument error" bug? 23:14
TheHarlot masak, more, I have a variable that contains characters like () and " ". I want to have those escape, so they can be fed as a file name.
jnthn XaRDaX: What Rakudo are you running?
23:14 envi joined
jnthn XaRDaX: Looks like $sock is undeclared. 23:14
XaRDaX: Not at all keen.
gah
masak: ^^
XaRDaX rakuto 2010.10
jnthn XaRDaX: OK, I think we fixed it in master. It crashes when reporting the error. 23:15
XaRDaX downloaded 3 hours ago
jnthn Yes, but released last month :)
XaRDaX :D
TheHarlot like a file named "12 - Frank Zappa Stink Foot(The pood slippers).mp3" for a fictional example. If that was in $track and I ran run("mplayer $track"); mplayer is going to WTF?!
jnthn XaRDaX: Anyway, I think you need a "my" at line 11
masak TheHarlot: try using .trans with the substitutions you want to apply to the string.
TheHarlot: S05 has the scoop on .trans, right at the bottom. 23:16
TheHarlot ah.
masak .trans is nice, because it lets you do several substitutions at once, and you don't have to worry about the order they're to be made in.
jnthn rakudo: $x = 42
p6eval rakudo 015d77: OUTPUT«===SORRY!===␤Symbol '$x' not predeclared in <anonymous> (/tmp/Yez_Wcgutn:22)␤»
jnthn XaRDaX: It should do something like the above. I bet if you try that program with the undeclared variable, you'll get the same error. 23:17
er, as in, the same one you got before
XaRDaX yes same error
TheHarlot rakudo: my $the_answer = 5 23:18
p6eval rakudo 015d77: ( no output )
TheHarlot XaRDaX, put my in front of your undeclared variable.
rakudo: my $the_answer ::= 5
p6eval rakudo 015d77: ( no output )
XaRDaX TheHarlot: I did but i got the same error 23:19
masak XaRDaX: are you running a non-new version of Rakudo?
TheHarlot XaRDaX, what line is the issue being reported on?
XaRDaX This is Rakudo Perl 6, version 2010.10 built on parrot 2.9.1 23:20
jnthn masak: October release, was mentioned above. 23:22
masak: I'm pretty sure the bug was fixed since the last release.
XaRDaX I've added my ($host,$port,$nick); in the line before the method
TheHarlot is the error now being reported on the line you added that "my" statement? 23:23
XaRDaX no line specification in the error 23:24
that's strange
TheHarlot okay...
jnthn XaRDaX: The error is totally bogus. It's an error that occurs when trying to report the real error.
XaRDaX: The latest git master has it fixed, and 2010.11 will be out in not so long :) 23:25
23:26 mavrc left
XaRDaX jnthn: this mean that there's a real error in my code. 23:26
TheHarlot Symbol '@dir_list' not predeclared in load_songs (musical.p6:39) << btw, the line number is directly after the file name.
TheHarlot removed a "my" from her code.
jnthn XaRDaX: Correct
XaRDaX: You need to write my $sock = IO::Socket::INET.new(); 23:27
you have just $sock = IO::Socket::INET.new();
Without the my
If you add that my to what you pasted at pastebin.com/FDwzuiqD it should be fine.
XaRDaX pastebin.com/gm041Ukw -> still not working 23:29
jnthn What error now? Same one? 23:30
XaRDaX yes same old error
MindosCheng try to install Rakudo Star
jnthn XaRDaX: Oh! 23:31
You use $sock in method join too
XaRDaX: Did you not want to make it an attribute?
has $!sock;
So you can share it between the methods
XaRDaX argh i didn't saw that 23:32
TheHarlot masak, is doing $blah.trans( / <-alpha> | <-digit> / => "\$/"); going to return the results I want? Or am I doing something buggy?
XaRDaX hahaha
yaww!!! It's working now! 23:33
jnthn \o/
XaRDaX++
masak TheHarlot: that regex looks fishy.
XaRDaX what a stupid error... thank's jnthn
masak TheHarlot: first, use C<||>, not C<|> for now.
XaRDaX jnthn++ 23:34
masak TheHarlot: second, seems your two alternatives cover all possible characters between themselves.
TheHarlot rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.trans( / [ <-alpha> || <-digit> ] / => "\\$/");
p6eval rakudo 015d77: OUTPUT«I want ty (ypen) yyu and EAT yyu␤»
masak heh :)
that [] is entirely redundant.
also -- wtf happened there? 23:35
23:35 Chillance left
masak "yyu"? 23:35
TheHarlot rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.trans( / <space> || <symbol> / => "\\$/");
p6eval rakudo 015d77: OUTPUT«I want ty (ypen) yyu and EAT yyu␤»
TheHarlot okay... this... this is an error report.
colomon that's o => y, isn't it?
TheHarlot rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.trans( / \s / => "\\$/");
p6eval rakudo 015d77: OUTPUT«I want ty (ypen) yyu and EAT yyu␤» 23:36
TheHarlot rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.trans( / \s / => "\\\\$/");
p6eval rakudo 015d77: OUTPUT«I want tn (npen) ynu and EAT ynu␤»
TheHarlot colomon, I want to take anything that is not a letter or number, and put a "\" infront of it.
that is what I asked masak prior. He suggested trans... I have never been able to use tr/// even in the Perl 5 days... so... yeah. 23:37
masak TheHarlot: I mean .trans with strings even on the lhs. 23:38
TheHarlot: I'm guessing the results we're seeing is due to the regex stringifying.
rakudo: say ~//
p6eval rakudo 015d77: OUTPUT«===SORRY!===␤Null regex not allowed at line 22, near ""␤»
masak rakudo: say ~/./
p6eval rakudo 015d77: OUTPUT«_block124␤»
TheHarlot rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell ~~ trans( / \s / => "\\$/");
p6eval rakudo 015d77: OUTPUT«Could not find sub &trans␤ in main program body at line 22:/tmp/JC1F9MNi4g␤»
TheHarlot rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell ~ trans( / \s / => "\\$/"); 23:39
p6eval rakudo 015d77: OUTPUT«Could not find sub &trans␤ in main program body at line 22:/tmp/IrQNw6RikJ␤»
jnthn rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.subst(/\W/, { "\\$/" })
p6eval rakudo 015d77: OUTPUT«I\Any()want to (open) you and EAT you␤»
jnthn uh
rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.subst(/\W/, { "\\$_" }, :g)
p6eval rakudo 015d77: OUTPUT«I\ want\ to\ \(open\)\ you\ and\ EAT\ you␤»
jnthn there
masak right.
TheHarlot thank you jnthn
colomon rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.comb(/./).map({when /<alpha>/ { $_ }; when / <digit> / { $_ }; "\\" ~ $_ }).join 23:40
p6eval rakudo 015d77: OUTPUT«I\ want\ to\ \(open\)\ you\ and\ EAT\ you␤»
masak rakudo: my $bloody_hell = "I want to (open) you and EAT you"; say $bloody_hell.trans( [" ", "(", ")"] => ["\\ ", "\\(", "\\)"] )
p6eval rakudo 015d77: OUTPUT«I\ want\ to\ \(open\)\ you\ and\ EAT\ you␤»
colomon jnthn++'s version is better, but I already had mine halfway typed at the time.
masak that's what I meant by "use .trans".
TheHarlot masak, that mostly requires me to think of all the scenarios I may come across in file names 23:41
which would have me reading the bash, sh, tcsh, ksh, etc manpages...
jnthn Having to read sucks. 23:42
;)
masak TheHarlot: right. jnthn++'s solution is more robust in that sense.
TheHarlot which would probably have me find writing this in bash may be quicker =P
masak at least if backslashing all \W is harmless.
TheHarlot jnthn, or, I could just do anything not a letter or number--and do it the easy way.
masak, it is harmless, and makes it harder to get stuff in via my own oversight 23:43
jnthn \W is "not a letter or a number" pretty much...
florz TheHarlot: just don't call a shell?
TheHarlot well--unless people get rather creative with their file names. At which point an errantly named file could wipe my system. 23:44
florz, ah, I am saving time, by sending this stuff off to mplayer to deal with.
florz TheHarlot: yeah, just don't call a shell
TheHarlot: if you want to use mplayer, why are you calling a shell? 23:45
jnthn Yeah, I love that track. "Da Hakaz - I love to $(rm -rf /*) you all nite long"
TheHarlot florz, that would mean setting it up, so my Perl 6 script can handle anything...
...
florz, other options?
florz TheHarlot: call mplayer?
TheHarlot how?
call("mplayer tracke")?
florz I don't have a clue how to (in perl 6, that is ;-), but I guess someone will 23:46
TheHarlot I got the notion to run("mplayer track"); by asking, "how do I run a command, then return to Perl 6 once done."
florz in perl 5 it would be system('mplayer','mplayer',$track); 23:47
dunno whether things work similarly in p6
Tene jnthn: whould there be a binding failure when you try to pass a ro value to an rw param?
jnthn Tene: I think there's meant to be. 23:48
Tene rakudo: sub foo($x is rw) { say $x }; foo(42);
p6eval rakudo 015d77: OUTPUT«42␤»
jnthn Tene: I'm not sure we do it...
XaRDaX while $irc.sock {say $irc.sock.recv();} Is that the right way to read the socket?
Tene jnthn: then allowing a constant as a default value for an rw param should be an error if it means binding.
jnthn Tene: Yes, there is that
Tene: We could even pick up the obvious cases at compile time.
TheHarlot jnthn, yeah... at least with the current escaping, that (the song "Da Hakaz - I love to $(rm -rf /*)" is not a big issue. Though, even adding slashes does not remove the possbility of that happening via other ways. People have gotten creative enough in the XSS world, that I do not image it is still not an issue. 23:49
florz, well, run does work similarly to system. I suppose I could do run('mplayer','mplayer',$track); 23:50
florz TheHarlot: doesn't look like it, just tried it
23:51 masonkramer left
masak MindosCheng: oh, you're from .tw? nice! it's one of the Mandarin-speaking areas I feel interested in visiting. 23:51
23:51 masonkramer joined
TheHarlot I never bothered with system in Perl 5. By the time I started coding in it (about ~2000... before then, I was still doing Perl 4), people had made enough bindings that I did not need to. 23:51
Tene TheHarlot: you should always do it the multiple-args way.
TheHarlot Tene, florz just said that did not seem to work. 23:52
florz at least in R* 10.09
Tene If you really want to allow injection, you should ask for it explicitly. sub inject-me-harder($cmd) { system('bash', '-c', $cmd); }
IMO
23:52 donaldh left
masak Tene: :P 23:53
jnthn Tene: Sounds like a good fit in S16... 23:54
;)
TheHarlot giggles
right... I am having issues finding the S... S16 you say? 23:55
colomon jnthn: my nqp-rx/nom error: gist.github.com/702754
masak TheHarlot: lucky you.
jnthn colomon: oh ouchee. 23:56
colomon jnthn: unfortunately (debugging-wise) I'm running off to the pub now...
Tene I would be pleased if Perl 6 didn't contain any simple 'pass this off to the shell' functions, but I expect others would disagree with me in the name of "Make easy things easy"
jnthn colomon: The pub is always fortunate!
Tene I'd really prefer it if you had to work to allow injection failures like that, though.
TheHarlot okay... S16 does not have what I am looking for.
Tene, yeah... I am not wanting to just pass it off. 23:57
colomon afk
TheHarlot I understand the dangers there in.
Tene TheHarlot: Lemme update my rakudo and see if I can help you. 23:58
TheHarlot oh, get the error that run() only accepts one argument
Tene ><
yeah, looks like it does. That's unfortunate. 23:59
TheHarlot This is Rakudo Perl 6, version 2010.09 built on parrot 2.8.0 << wonderful. We got an old version in the Fedora 14 repos.

Powered by Cro and the Raku® Programming Language.

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

Thank you!