Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2010-01-06

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 sjohnson no worries
00:27 _eMaX_1 joined #perl6
00:36 xinming_ joined #perl6
00:50 PacoLinux joined #perl6
00:51 colomon__ joined #perl6
01:02 rgrau` joined #perl6
01:03 payload joined #perl6
01:03 ashizawa joined #perl6
01:07 dbrock joined #perl6
01:25 lichtkind_ joined #perl6
01:46 eternaleye joined #perl6
01:47 eternaleye joined #perl6
01:57 priodev joined #perl6
01:59 rachelBROWN joined #perl6
02:03 priodev left #perl6
02:37 agentzh joined #perl6
02:49 cognominal joined #perl6
03:01 patspam joined #perl6
03:10 dbrock joined #perl6
03:10 Exodist joined #perl6
03:20 galf joined #perl6
03:22 agentzh joined #perl6
03:57 nihiliad joined #perl6
04:04 diakopter hi
04:06 sjohnson diakopter! hi
04:06 sjohnson how's it going
04:18 diakopter ok
04:19 pdcawley_ joined #perl6
04:21 jaldhar joined #perl6
04:25 sjohnson cool
04:28 jaldhar joined #perl6
04:55 soupdragon joined #perl6
05:10 pugs_svn r29472 | diakopter++ | [sprixel] an optimization for gplus. instead of behaving O(n**5), now it's O(n**2.5).  :|
05:31 icey joined #perl6
05:39 meppl joined #perl6
06:29 dduncan left #perl6
06:35 pugs_svn r29473 | diakopter++ | [sprixel] improved correctness and complexity/efficiency drastically. memoization of failed offsets when backtracking. implemented in plus. unclear whether it should be implemented in either.  should go quadratic only on the same classes of patterns Perl 5 does, now...
06:36 JimmyZ joined #perl6
07:03 synth joined #perl6
07:04 kaare joined #perl6
07:16 Su-Shee joined #perl6
07:16 Su-Shee good morning
07:19 szbalint good morning
07:31 brrant joined #perl6
07:42 Nerella joined #perl6
07:46 mberends joined #perl6
07:47 mberends good morning from a classroom via webchat
07:48 soupdragon :)
07:48 Baggiokwok joined #perl6
08:00 iblechbot joined #perl6
08:16 mathw o/
08:18 payload joined #perl6
08:31 payload joined #perl6
08:35 xenoterracide joined #perl6
08:52 jaslts joined #perl6
08:52 jaslts left #perl6
08:53 sjohnson hi mathw
08:53 mathw o/ sjohnson
08:54 sjohnson long time no chat :)
08:59 mathw indeed
09:13 eternaleye joined #perl6
09:16 meneldor joined #perl6
09:18 nadim joined #perl6
09:53 saschi joined #perl6
09:54 saschi hello
10:02 ruoso joined #perl6
10:08 squ joined #perl6
10:10 sjohnson hi saschi
10:15 payload joined #perl6
10:16 cognominal joined #perl6
10:22 sjohnson snuh!
10:22 saschi snuh?
10:22 saschi snow! ^-^
10:23 dakkar joined #perl6
10:29 yath joined #perl6
10:29 yath heya
10:39 squ hola
10:46 riffraff joined #perl6
10:47 BcaT !cucumber
10:47 BcaT !ping
10:47 partisan Pong!
10:47 BcaT !banana
10:47 squ what perl6 makes with people
10:48 BcaT !trout
10:48 squ they become hungry
10:48 BcaT !rubia?
10:48 BcaT only rubia and tilapia are any good when it comes to fish
10:49 BcaT but lobster and crab and prawns are a different story
10:49 squ crabs which you gathered yourself are twice more delicious
10:51 squ you have to go diving, then get beer
10:51 mariuz joined #perl6
10:51 BcaT i can only do 1 of 3 things in that situation - drink beer
10:51 BcaT i cant dive or cook
10:56 squ just put them in water
10:57 squ == cook
11:01 saschi O_o
11:04 cognominal joined #perl6
11:06 BcaT boiling crabs to death is too much for me
11:06 BcaT they look so sad and innocent
11:06 BcaT and plus they're funny to watch
11:10 nbrown joined #perl6
11:18 sjohnson BcaT: you like crabs?  http://www.youtube.com/watch?v=DjaTLyRcpMM
11:22 saschi that's odd
11:25 sjohnson crabland == odd :)
11:25 colomon joined #perl6
11:32 bluescreen joined #perl6
11:32 payload joined #perl6
11:35 squ BcaT, thats hardest part, true
11:35 squ :-)
11:42 stepnem joined #perl6
11:48 cotto joined #perl6
11:53 Simulacron joined #perl6
11:59 Baggiokwok joined #perl6
12:00 iblechbot joined #perl6
12:03 _Jedai_ joined #perl6
12:18 bluescreen joined #perl6
12:20 ruoso joined #perl6
12:22 SmokeMachine joined #perl6
12:32 fridim_ joined #perl6
12:35 barney joined #perl6
12:35 bluescreen joined #perl6
12:44 colomon rakudo: say slurp;
12:44 p6eval rakudo 3867ff: too few positional arguments: 0 passed, 1 (or more) expected␤in Main (file <unknown>, line <unknown>)␤
12:45 colomon rakudo: say slurp($*ARGFILES)
12:46 p6eval rakudo 3867ff: Contextual $*ARGFILES not found␤Unable to open filehandle from path ''␤in Main (file <unknown>, line <unknown>)␤
12:59 bbkr joined #perl6
13:00 Baggiokwok joined #perl6
13:01 arthur-_ joined #perl6
13:02 bbkr hi. i'm in parrot_install/lib/1.9.0-devel/library (december release or rakudo) and when i do perl6 -e 'use MIME::Base64; say MIME::Base64::encode_base64("test");' i get invoke() not implemented in class 'Undef' error. should .pbc code be included/called in some special way?
13:06 tuxdna joined #perl6
13:20 am0c joined #perl6
13:24 vorner joined #perl6
13:57 pmurias joined #perl6
14:00 snearch joined #perl6
14:00 takadonet joined #perl6
14:01 takadonet morning all
14:03 riffraff joined #perl6
14:06 plainhao joined #perl6
14:10 eric256 joined #perl6
14:13 eric256 the whole negative arguments to substr don't seem to be documented in S32/Str..should i be looking somewhere else?
14:14 PerlJam eric256: the perl5 docs.
14:15 eric256 thanks
14:15 eric256 wouldn't have thought to look there. lol
14:15 eric256 looks like anything outside the string should return undef and a warning...that should be easy enough to fix
14:16 PerlJam eric256: S01 says "Another assumption has been that if we don't talk about something in these
14:16 PerlJam Synopses, it's the same as it is in Perl 5"
14:16 PerlJam so, if you can't find something in the spec, look in the Perl 5 docs for guidance
14:16 eric256 okay, thanks
14:17 soupdragon left #perl6
14:21 eric256 okay....so if the documentation for perl5 says one thing, but it actual does another, should we match the documentation or the implementation? hehe
14:21 eric256 substr("Hello",10,2) should return undef and a warning but i can't seem to get it to through a warning
14:23 eric256 oh ignore me
14:23 eric256 lol
14:31 pmurias PerlJam: shouldn't we try to make the synopsis complete enought to remove the dependency on p5 docs?
14:37 mj41_ joined #perl6
14:40 renormalist joined #perl6
14:46 payload joined #perl6
14:48 PerlJam pmurias: yes.
14:49 PerlJam eric256: if the docs and the implemenation conflict and there's no clear indication of which way to go, then we should discuss it here or on p6l
14:49 PerlJam s/mena/menta/
15:18 justatheory joined #perl6
15:24 nihiliad joined #perl6
15:26 FCO joined #perl6
15:28 pmurias joined #perl6
15:29 ash_ joined #perl6
15:40 payload joined #perl6
15:49 KyleHa joined #perl6
15:52 Psyche^ joined #perl6
16:08 cognominal joined #perl6
16:24 SmokeMachine joined #perl6
16:31 mberends left #perl6
16:34 rachelBROWN joined #perl6
16:57 kaare_ joined #perl6
17:04 cdarroch joined #perl6
17:05 eric256 how can i throw warnings from inside of src/core?
17:10 ash_ rakudo: warn("this is a warning");
17:10 p6eval rakudo 3867ff: this is a warning␤
17:10 takadonet rakudo: warn("Takadonet needs beer")
17:10 p6eval rakudo 3867ff: Takadonet needs beer␤
17:11 eric256 yea, but when i use that in core/Any-str.pm i get Could not find non-existent sub &warn
17:11 ash_ its only 11 am here, its a bit to early for me to have beer, well... realtively speaking
17:11 eric256 rakudo: warn "this is a warning"
17:11 ash_ ah that means its not yet defined
17:11 p6eval rakudo 3867ff: this is a warning␤
17:11 ash_ umm
17:12 eric256 yea, so i don't know how to warn from in there....
17:12 eric256 shouldn't warn be defined pretty early? lol
17:13 eric256 my other problem is trying to defined with (Str $string:  doesn't seem to work....when i do that it says it can't find the right definition of substr, some sort of dispatch issue i think...or me being stupid
17:13 eric256 lol
17:14 ash_ are you in ng?
17:14 ash_ ng doesn't have warn currently
17:15 ash_ ng: warn("shouldn't work");
17:15 p6eval ng 93fa57: Could not find non-existent sub &warn␤current instr.: '_block14' pc 29 (EVAL_1:0)␤
17:15 eric256 lol ahhh
17:15 eric256 yes i'm in ng, sorry didn't realize that it was missing that
17:16 aindilis` joined #perl6
17:17 * eric256 wishes there was a way to fudge + run the fudged file all at once
17:29 colomon___ joined #perl6
17:31 colomon___ eric256: The Str $string: argument thing doesn't work yet in Rakudo, as far as I know.
17:32 ash_ are you doing an indirect call?
17:34 ash_ and, in ng in src/builtin/control.pir they define die, warn is setup the same way but you use a different severity, it doesn't seem hard to implement
17:35 ash_ you could compare that to the old die which is in src/old/builtins-old/control.pir (i don't think this has changed much)
17:37 eric256 ahh i thought it was done already
17:37 eric256 i'll check ash
17:38 eric256 i'll just fudge a few more of the subs that want to use the named args version
17:38 Tene I'll get &warn in, 'sec
17:39 Tene warn and die should be setting dynamic lexicals, not globals...
17:39 Tene That's part of a bigger refactor, though.
17:41 eric256 that bit goes over my head
17:41 eric256 lol
17:42 Tene warn is pushed.
17:43 squ perl6: printf "%0.4194310f", 0x0.0x41414141;
17:43 p6eval pugs: *** ␤    Unexpected "0x41414141"␤    expecting ".", "\187", ">>", "=", "^", operator name, qualified identifier, variable name, "...", "--", "++", "i", array subscript, hash subscript or code subscript␤    at /tmp/IOVdSHNyus line 1, column 27␤
17:43 p6eval ..rakudo 3867ff: Confused at line 2, near ".0x4141414"␤in Main (file <unknown>, line <unknown>)␤
17:43 p6eval ..elf 29473: Parse error in: /tmp/SmLZRFq49n␤panic at line 1 column 0 (pos 0): Can't understand next input--giving up␤WHERE: printf "%0.4194310f", 0x0.0x41␤WHERE:/\<-- HERE␤  STD_red/prelude.rb:99:in `panic'␤  STD_red/std.rb:76:in `scan_unitstopper'␤  STD_red/std.rb:224:in `comp_unit'␤
17:43 p6eval ..STD_red/std.rb…
17:45 ash_ Tene++ thanks for the warn push
17:45 ash_ eric256: you now have warn if you want to use it in ng
17:45 eric256 thanks..pulling and making now
17:45 eric256 ng:warn("test")
17:46 dalek rakudo/ng: 3f06c7e | tene++ | src/builtins/control.pir:
17:46 dalek rakudo/ng: Add &warn
17:46 dalek rakudo/ng: review: http://github.com/rakudo/rakudo/commit/3f06c7ec4eebcbe12f15ce4d3b250e4db4189901
17:46 eric256 ng: warn('test')
17:46 p6eval ng 93fa57: sh: ./perl6: No such file or directory␤
17:46 Tene apparently it's rebuilding now.
17:46 eric256 slow
17:46 eric256 lol
17:47 justatheory joined #perl6
17:47 ash_ ng: warn('test?'); # rdy yet?
17:47 p6eval ng 93fa57: sh: ./perl6: No such file or directory␤
17:47 ash_ nope
17:48 TimToady std: printf "%0.4194310f", 0x0.0x41414141;
17:48 p6eval std 29473: [31m===[0mSORRY![31m===[0m␤Confused at /tmp/b4snQ40i8K line 1:␤------> [32mprintf "%0.4194310f", 0x0.[33m⏏[31m0x41414141;[0m␤    expecting any of:␤  integer␤  postfix_prefix_meta_operator␤FAILED 00:01 109m␤
17:50 ash_ on osx in perl 5.10.1 that instantly seg faults
17:52 stephenlb joined #perl6
17:52 fridim_ Hi, I read that rakudo is well integrated in the fedora distribution, that's great !
17:54 Tene It's available to install as part of the standard repository in Fedora, yes.
17:55 eric256 do we realy need a "test".flip function? seems silly
17:55 eric256 and why isn't it .reverse?
17:56 PerlJam eric256: reverse is for lists
17:56 PerlJam eric256: and yes, we need flip
17:56 colomon___ And it's not like implementing flip was hard work...
17:57 TimToady don't want to accidentally flip the characters when the list happens to contain one element
17:57 colomon___ (Well, I did have to write reverse first to get flip working in ng, as I recall.)
17:57 TimToady in general, any time Perl 5 overloaded an operator for two different purposes, Perl 6 has two different operators
17:58 TimToady for instance, perl 5's scalar .. is now ff
17:58 TimToady and we differentiate x from xx now
17:59 eric256 i'm just wondering how many times people reverse a string
17:59 PerlJam eric256: you mean flip  :)
18:00 TimToady makes it easy to do pattern matching from the end of the string, for one
18:00 eric256 lol i successfully made "make t/spec/S32-str/substr.t" hang
18:00 eric256 lol
18:02 TimToady another consideration is that a hardwired .flip is likely to be faster than rolling your own (though currently it's probably equivalent)
18:03 TimToady depends on how much support you get from the VM for it
18:03 jackyf joined #perl6
18:03 TimToady but mainly it's a matter of making people distinguish things that need distinguishing semantically
18:04 slavik joined #perl6
18:04 * ash_ out to lunch
18:05 TimToady and we hope that learning a new name is less trouble than learning a new exception rule
18:05 TimToady certainly it makes it easier to look up if it has its own name
18:05 TimToady (admittedly, this can be taken to extremes; see PHP...)
18:07 eric256 there is an irony to getting multiple dispatch and then using different sub names for things that appear to be good candidates for multiple dispatch
18:07 PerlJam eric256: reverse isn't one of those because of the semantic confusion it causes
18:08 PerlJam (if it were to behave as in per l5)
18:08 PerlJam er, s/ l/l /
18:10 TimToady in short, we want Perl 6 to have different FAQs than Perl 5 did  :)
18:10 TimToady we're tired of the old ones...
18:12 TimToady but yes, there are many ironies like that in Perl 6.  we also make it possible to extend the language easily, and try to remove any desire to extend it by putting in most of what you need.
18:13 TimToady we put in a MOP to have world-class single-dispatch OO, then hide most of it behind multi-dispatch so that the language gets to determine the semantics rather than the class.
18:14 nihiliad joined #perl6
18:15 TimToady we put in an intricate type system, and then let the user ignore it much of the time...
18:17 TimToady and to top it all off, we're so full of ourselves that we're completely empty.  :)
18:18 TimToady Contradictions "Я" Us
18:18 cognominal I would expect     rule {  <a>+ }    being equivalent to   token { <a> ** <ws> } but
18:19 cognominal rakudo:  rule a { a+ }; say 'a a' ~~ m/ <a> /
18:19 p6eval rakudo 3867ff: a ␤
18:19 diakopter speaking of.. does type inference need to be spec'd any more than it is?
18:19 PerlJam TimToady: is that the Unicode version of the R meta-op?  :)
18:19 payload joined #perl6
18:19 PerlJam Я seems more appropriate somehow
18:19 cognominal well, it matches  'a' instead of 'a a'
18:19 TimToady cognominal: your expectations of <a>+ are wrong
18:20 cognominal I am not sure S05 is explicit about the expected behavior.
18:20 jnthn <ws> is inserted where there's whitespace in the pattern.
18:20 phenny jnthn: 30 Dec 23:27Z <mberends> tell jnthn welcome back
18:20 TimToady a rule only matchs <ws> where there's whitespace in the rule
18:20 TimToady now, [a ]+ would be a different matter
18:21 jnthn 30 Dec? Heh. I musta created a little too much optimism about when I'd be back. :-)
18:22 TimToady they didn't plant any explosives on you, did they?
18:22 jnthn heh heh
18:22 PerlJam cognominal: having a "deparse" option for regex would be useful to dispell you of your errant expectations.
18:22 jnthn No, I didn't get caught up in that one :-)
18:22 jnthn Just over in Sweden at the moment sorting various @stuff out.
18:23 jnthn Heading back home at the weekend, and then things will return to a vague state of normalish.
18:23 cognominal I should read S05 more carefully
18:23 jnthn Found myself more tired and in need of a break than expected over Christmas / new year, so took lots of rest. :-)
18:24 TimToady that's wise
18:24 TimToady you'll get more done in the long run that way
18:24 jnthn Aye.
18:25 jnthn Anyway...time for me to fine something to eat and drink here.
18:25 * TimToady ponders what a backwards X metaop would look like...
18:26 * jnthn back later, for some definition of later :-)
18:30 cognominal TimToady, if the  {*} #= stuff is slated to disappear, what will be the canonical way to support different action classes for a given grammar. Say for a beautifier versus a compiler.
18:32 TimToady we still have per-parse action classes, and we still have actions for each complete rule
18:32 TimToady in STD I discovered that the use of {*} almost always meant I should break a rule apart, if I really needed the action
18:32 TimToady but mostly the action gets all the parts labelled well enough in any case without using {*}
18:33 taesoo joined #perl6
18:33 TimToady to the extent that the first thing viv did was throw away all the calls from {*}
18:34 TimToady (except the implied one at the end of each rule)
18:35 TimToady {*} #= was mostly a hack for bootstrapping, and it turns out we don't really need it very much, and it's cluttery looking
18:35 cognominal What is the mechanism to route a method call to the right action-class ?
18:35 TimToady the :action argument to .parse specifies it
18:35 TimToady :actions rather
18:36 cognominal ok, this pard does not change
18:36 cognominal *part
18:36 colomon___ jnthn's back, \o/
18:37 TimToady FSDO
18:37 cognominal I am not sure how an action call looks like in the grammar. Is it a closure?
18:38 uniejo joined #perl6
18:39 TimToady well, you can actually see some in STD.pm in method EXPR
18:39 TimToady push @termstack, $op._REDUCE($op<_from>, 'PREFIX');
18:40 cognominal thx
18:41 TimToady so you could insert <_REDUCE(self.pos, 'SoFar')> to get a similar effect in STD
18:41 TimToady I don't think rakudo supports _REDUCE though
18:42 TimToady the #= probably desugars to something else there
18:42 TimToady and I don't know whether it's accessible from within the grammar
18:43 TimToady but if so, you could probably write a _REDUCE method that calls it
18:44 nihiliad joined #perl6
18:44 snearch_ joined #perl6
18:44 uniejo left #perl6
18:44 TimToady again, though, that probably means you want to refactor into separate rules and get _REDUCE called automatically
18:45 uniejo joined #perl6
18:46 PerlJam TimToady: you know, I've written :action instead of :actions a couple of times myself (while coding) and been momentarily confused when it didn't work.  I wonder if that's going to be a common error.
18:56 cognominal grammar A {  TOP { a } };    class A1::Actions {  method TOP { ... }  };  class A2::Actions {  method TOP { ... }  };   # suppose I want an action for TOP, I still don't understand  the wiring necessary to make it happen.
18:57 slavik1 joined #perl6
18:57 cognominal there is    parse( 'a', :action<A1::Actions)     but what I need in the TOP rule
18:57 cognominal forgot the rule keyword btw.
19:05 PerlJam cognominal: methods in the Actions class that have the same name as regex in the Grammar, get called at the end of the parsing of that particular regex
19:06 PerlJam (if that's what you're asking)
19:06 colomon___ ooooo!
19:06 PerlJam cognominal: also, if the regex has {*} in it, the same-named method in the actions class is called then.
19:07 PerlJam (when the parse reaches the {*} token within the regex)
19:08 PerlJam See S05:3781 or so.
19:08 jan_ joined #perl6
19:09 cognominal but the point of TimToady is the disappearance of the {*}. I am confused.
19:09 PerlJam {*} turned out to not be as useful  (or as a hallmark of a necessary refactor), thus it's not really used much any more.
19:12 PerlJam Yesterday or the day before pmichaud was talking about hooks for LEAVE, NEXT, LAST, REDO, etc. ...  I wonder if this implicit {*} and the end of the regex is a specialization of LEAVE or something.  Maybe it needs its own blast/blorst/whatever we're calling them these days.
19:12 PerlJam oh, phaser
19:12 PerlJam that's what I meant
19:14 PerlJam Or even if the machinery that allows the same-named-method to be called should be exposed to the programmer so that  they can do things like choose another-named-method or set of methods to be called.
19:16 szbalint jnthn: "How many Slovak policemen is needed to make a terrorist attack? Only two. One hides an explosive and the second fails to find it."
19:17 cognominal context: http://www.cbc.ca/world/story/2010/01/06/slovak-explosives-mistake.html
19:28 Tene cognominal: if <a>+ implied whitespace, it wouldn't be possible to say what <a>+ currently means any more, which is a net loss when [<a> ]+ or <a>**<ws> is so easy.
19:28 Tene cognominal: spaces around ** are interpreted as <ws>, though, so you could say <a> ** {1..*}
19:29 Tene iirc
19:33 jferrero joined #perl6
19:39 cognominal hum, curious what <a> ** *   would mean
19:41 jonasbn joined #perl6
19:45 ShaneC joined #perl6
19:45 PerlJam cognominal: I would guess that would be an error because the whatever-* doesn't exist in regex
19:46 colomon___ Tene: wait, are you saying that if sigspace is turned on <a> ** whatever is <a> separated by whatever, and ignore whitespace in between them?
19:47 zibri joined #perl6
19:51 TimToady std: /<a> ** */
19:51 p6eval std 29473: [31m===[0mSORRY![31m===[0m␤quantifier quantifies nothing at /tmp/mVdvzAKinf line 1:␤------> [32m/<a> ** *[33m⏏[31m/[0m␤    expecting quantmod␤FAILED 00:01 108m␤
19:54 PerlJam colomon___: regardig C<**> ... "Space is allowed on either side of the complete quantifier.  This space is considered significant under C<:sigspace>, and will be distributed as a call to <.ws> between all the elements of the match but not on either end."  (S05:839)
19:54 colomon___ sweetness
19:55 PerlJam though rakudo doesn't quite seem to work correctly.
19:55 colomon___ PerlJam: what else is new?  ;)
19:56 PerlJam rakudo:  say "a,a,a,a" ~~ / 'a' ** ',' / ?? "yes" !! "no"
19:56 p6eval rakudo 3867ff: yes␤
19:56 PerlJam rakudo:  say "a,a,a,a" ~~ regex { 'a' ** ',' } ?? "yes" !! "no"
19:56 p6eval rakudo 3867ff: yes␤
19:56 PerlJam rakudo:  say "a  ,   a   ,   a   ,   a" ~~ regex { 'a' ** ',' } ?? "yes" !! "no"
19:56 p6eval rakudo 3867ff: yes␤
19:57 PerlJam rakudo:  say "a  ,   a   ,   a   ,   a" ~~ rule { 'a' ** ',' } ?? "yes" !! "no"
19:57 p6eval rakudo 3867ff: yes␤
19:57 PerlJam As long as you weren't looking for a false positive  ...  :)
19:58 cognominal std:   say 1 ?? 2 :: 3  # curious to see if std protects me against that one
19:58 p6eval std 29473: [31m===[0mSORRY![31m===[0m␤Please use !! rather than :: at /tmp/Kd0b7tahqj line 1:␤------> [32msay 1 ?? 2 [33m⏏[31m:: 3  # curious to see if std protects m[0m␤    expecting any of:␤        bracketed infix␤  dotty␤  infix or meta-infix␤        infix stopper␤    postfix␤
19:58 p6eval ..standard stopper␤  terminator␤F…
19:58 cognominal cool
19:59 eric256 is there anyway to fudge just a single test file?
19:59 PerlJam run fudge on just that file?
20:00 TimToady phone
20:00 eric256 i dunno how to do that though
20:00 ash_ eric256: look in t/spectest.data
20:00 ash_ thats how it knows which files to run
20:01 ash_ if you want to add/remove a file from the fudge
20:01 eric256 no i just want to fudge one...the fudging on this test file doesn't appear to be working right, so i need to keep rebuilding it
20:02 ash_ fudge 1 test?
20:02 eric256 one file
20:04 ash_ i think i am not understanding what your asking
20:05 eric256 i have a test file with fudge commands in it, they arn't all working, so i need to fudge the file and look at the output, rinse and repeat until the out put is correct
20:06 PerlJam eric256: have you read t/spec/README ?
20:07 ash_ is it a file you wrote? the fudge command checks all the files listed in t/spectest.data and fudges the files listed in there
20:07 Lorn joined #perl6
20:07 ash_ after its fudged there will be a t/spec/S##/filename.rakudo instead of .t
20:08 colomon___ eric256: make t/spec/S##/filename.t will make the .rakudo file and then run it.
20:08 eric256 right, and thats working, but its slow, i just wanted to fix this one, and i can't even seem to find fudge, and its an exeisting spec
20:08 eric256 colomon___:  in this case its not makeing the .rakudo file....thats kinda my problem.....mmm no errors either
20:09 colomon___ eric256: ooo, tricky.
20:09 PerlJam eric256: t/spec/fudge
20:09 eric256 oh duh found fudge...thanks perljam... didn't expect progams in there cause i'm stupid
20:21 ash_ joined #perl6
20:24 zamolxes_ joined #perl6
20:26 elmex joined #perl6
20:37 bluescreen joined #perl6
20:43 Tene colomon___: depends on whether there's whitespace around the ** or not.
20:44 Tene <digit> ** ',' # whitespace allowed
20:44 Tene <digit>**',' # no whitespace allowed
20:44 Tene is my memory
20:44 mberends joined #perl6
20:48 colomon___ Tene: right, and :sigspace.  PerlJam explained back up there a bit.  :)
20:48 colomon___ every new thing I learn about grammars makes me want to go \o/
20:50 cognominal TimToady, btw, actions object should be probably  stackable:  Grammar.parse($str, :actions(  <A1 A2>.>>.new )
20:50 PerlJam cognominal: What does that mean?
20:52 KyleHa joined #perl6
20:52 cognominal that would mean that one should call a method of A1 then a method of A2 when an action method is to be called
20:52 cognominal s/one/the implementation/
20:53 PerlJam cognominal: so when a regex foo in the grammar matches, the foo method of A1 should be called and then the foo method of A2?
20:53 PerlJam cognominal: if so, why?
20:55 cognominal because A2 foo may need info generated by A1 foo.
20:55 TimToady if you need such semantics, just create an actions modules that delegates appropriately
20:56 TimToady *module
20:56 jackyf joined #perl6
20:56 cognominal yes, I could do that too.
20:56 TimToady you certainly shouldn't be relying on hyperops, whose order is officially undefined
20:57 TimToady (though in this case the order is preserved by the time it gets into :actions())
20:58 TimToady *postserved might be more accurate...
20:58 TimToady *curcumserved maybe...
20:58 TimToady s/u/i/ #gah
20:59 cognominal :)
20:59 bbkr joined #perl6
21:00 TimToady anyway, if it's a common operation to combine A1 and A2 in a standard way, there should probably be a named combinator to generate a delegational class around them
21:00 cognominal I note that when I learn something I am incapable to relate it to something I already know when it would be most useful.
21:01 TimToady not just relying on the list to apply an implicit combinator
21:01 cognominal grammar being classes, what I ask obviously fall in the realm of classes.
21:01 cognominal *falls
21:02 TimToady though in this case the action class isn't a grammar
21:02 cognominal I should learn japanese instead being tripped with these stupid flexionnal languages.
21:02 cognominal yes, but my mind was focused about grammars :)
21:03 TimToady those Indo-Europeans have a lot to answer for...
21:03 cognominal btw you should call given wa instead :)
21:05 frettled We should have a "freeze" function in Perl 6.  It will "freeze" the specification and functionality for the programmer using it.  :D
21:06 mberends frettled: the Linux command for that is 'ifconfig eth0 down'
21:06 TimToady cognominal: it was considered :)
21:07 cognominal I have a nice app called "human japanese" on my iPhone. It tought me a thing or two.
21:08 frettled mberends: :)
21:08 frettled cognominal: thanks for the tip!
21:11 uniejo joined #perl6
21:11 cognominal In France, they want to pass a law to prohibit phones in school. It makes sense because kids play with them instead of listening to the teacher, but app phones are the best pedagical tools I have ever seen and it can only get better.
21:11 frettled cognominal: I see that it's English-based; how is the pronunciation presented?
21:12 colomon_ joined #perl6
21:14 cognominal it is english based,  you click in a sentence to hear it.  It is very progressive. It starts with hiragana.
21:15 frettled I was thinking about the romanization, which might lead to interesting pronunciation misconceptions.
21:16 cognominal frettled, there is light version of it if you want to try it before buying.
21:17 cognominal The author says he works on a more advanced version which will certainly be reday by the time I finish the present one.
21:17 ash_ it would be kinda cool if more schools did something different like having app's to teach
21:17 frettled cognominal: Yep, I started downloading it, but it seems to take a while, even over wifi.  :)
21:17 ash_ in one of my software engineering electives next semester we are making andriod phone apps, so maybe i can try doing something like that
21:17 ash_ it can teach perl 6 :-D
21:18 frettled \o/
21:18 ash_ hmmm wonder if i could get parrot running on the andriod...
21:18 cognominal I think pupil would be more mature if they could learn by themselves using such device and the teacher would have more time to interact with pupils.
21:20 xalbo The problem is that if the pupils are not more mature, they'll learn even less.
21:20 cognominal I don't expect much from educational institutions, I was taught  using a sliding rule at a time when cheap calculators were already available.
21:21 ash_ my grandfather has a sliding rule still, he's an architect, i worked for him a few summers ago taking his hand drawings and putting them in CAD
21:24 frettled cognominal: I think I may like that app already.  Thanks again.
21:25 ash_ hmm, i can use arm-eabi-gcc to cross compile from x86 to arm, wonder if that works with parrot, i'll go ask
21:26 bbkr hi. i'm in parrot_install/lib/1.9.0-devel/library (december release or rakudo) and when i do perl6 -e 'use MIME::Base64; say MIME::Base64::encode_base64("test");' i get invoke() not implemented in class 'Undef' error. should .pbc code be included/called in some special way?
21:28 diakopter I didn't know "use " was working in rakudo?   /me must be behind
21:28 mberends bbkr: the 'use' command is not implemented for .pbc files yet, afaik.
21:28 cognominal frettled, you are free to /msg me the name of your favorite apps.  It is difficult to find one's way in the app store when rating are done by semi-illiterate kids.
21:29 mberends 'use' does work with .pm files and .pir files that are compiled from .pm files
21:29 bbkr it loads module without error. problem is on method call.
21:30 bbkr ok, i'll write my own pure p6 Base64 or i'll wait till it will be working. thanks.
21:37 orafu joined #perl6
21:44 justatheory joined #perl6
21:44 douglashunter joined #perl6
22:18 * eric256 got fudge to hang
22:19 eric256 adding ""#?rakudo todo 'exception'" to the line right before a block stops it from working at all...
22:27 pjcj joined #perl6
22:28 * mberends looked at t/spec/fudge and gave up trying to understand it
22:29 orafu joined #perl6
22:33 diakopter moritz_: where you
22:38 szbalint I've must have scared him off :)
22:42 buu joined #perl6
22:42 buubot joined #perl6
22:46 eric256 yay, figured out this fudging thing....now to finish actualy fixing the file
22:53 wayland76 joined #perl6
22:54 wayland76 Hi all.  Is the -ng branch at all likely to land before the January release?
22:56 japhb joined #perl6
23:01 Chillance joined #perl6
23:04 ash_ maybe? i think jnthn++ and pmichaud++ are the two that would know the best to be honest
23:04 cognominal_ joined #perl6
23:05 patspam joined #perl6
23:27 diakopter wayland76: haven't heard anything on that in few weeks.  my conjecture would just be (more) FUD (unless you trusted my opinion, of course).
23:31 eric256 rakudo: my $x = undef;
23:31 p6eval rakudo 3867ff:  ( no output )
23:31 eric256 ng: my $x = undef;
23:31 p6eval ng 3f06c7: Could not find non-existent sub &undef␤current instr.: '_block14' pc 29 (EVAL_1:0)␤
23:31 eric256 ng doesn't stand for not-good does it?
23:31 eric256 hehe
23:33 TimToady wayland76: yes, I think the intent is to get it in the January release even if there are still a few minor regressions
23:34 IllvilJa joined #perl6
23:35 Tene eric256: 'undef' doesn't exist anymore.
23:35 TimToady std: my $x = undef;
23:35 p6eval std 29473: [31m===[0mSORRY![31m===[0m␤Unsupported use of undef as a value; in Perl 6 please use something more specific:␤     Mu (the "most undefined" type object),␤   an undefined type object such as Int,␤    Nil as an empty list,␤    *.notdef as a matcher or method,␤ Any:U as a type
23:35 p6eval ..constraint␤       or fail() …
23:36 eric256 ng: my $x = Nil;
23:36 p6eval ng 3f06c7:  ( no output )
23:36 eric256 my $x = Str;
23:36 eric256 ng: my $x = Str; $x.perl.say;
23:36 p6eval ng 3f06c7: Str␤
23:37 eric256 so substr, where it would have returned undef should return Str?
23:37 TimToady ng: my Str $x = Nil; say $x.WHAT
23:37 p6eval ng 3f06c7: List()␤
23:37 TimToady heh, should be Str, oh well
23:37 eric256 the error above sas "Nil as an empty list" so it did what i expected for you
23:39 frettled Mostly off-topic question: I think I might go to YAPC::NA this year, since I've promised to visit a friend in NYC anyway.  How insane am I for wanting to drive from NYC to Columbus?  :)
23:39 wayland76 thanks people.
23:39 wayland76 diakopter: I saw one piece of info: http://lith-ology.blogspot.com/2009/12/seven-days-between-parrot-and-camel_28.html
23:40 eric256 thats only 500 miles, thats nothing
23:40 frettled eric256: I wasn't thinking about the distance :)
23:40 eric256 my wife drives 1000 (denver to san diego)  all the time
23:41 frettled eric256: this fall, I drove about 1000 miles to visit Abigail, then took a 500 mile detour to France to pick up some wine, and drove back home again.  :)
23:42 frettled No, I'm more concerned about traffic conditions.
23:43 eiro 'lo guys
23:43 frettled lolo
23:43 TimToady wayland76: my info is as of the phone call today
23:43 eiro glob() isn't yet implemented, right ?
23:43 diakopter oh
23:46 orafu joined #perl6
23:46 TimToady eiro: doesn't look like it
23:47 eiro just to be sure :)
23:47 eiro thx TimToady
23:51 TimToady eric256: S02:1872 specs that it should return Str, actually
23:52 TimToady the error is slightly misleading, insofar as Nil is really the empty parcel, not the empty list
23:52 eric256 thats what i said... lol
23:52 eric256 well not based on a spec, just judgment hehe
23:52 TimToady Nil and () only turn into the empty list in list context
23:53 eric256 and the file that i just spent ours fixing, i rm'ed by accident
23:53 eric256 that sucked
23:57 eternaleye joined #perl6
23:57 wayland76 TimToady: Thanks again! :)
23:57 colomon__ joined #perl6
23:58 frettled eric256: quick, recover the inode!

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo