pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | <stevan> Moose... it's the new Camel ":P | .pmc == PPI source filters!
Set by Alias_ on 16 March 2006.
ruoso making a daily build and upload (if tests passes) of parrot to debian would be a good test bed 00:01
as it would end up building in another 11+ architectures...
s/daily/something more reasonable/
arcady weekly, perhaps, though even that may be a bit much 00:02
what's actually going with parrot, anyway? 00:03
ruoso hm? 00:04
arcady does anything compile to it, these days?
TimToady Oh, Perl 6 in another week or two... :) 00:05
Actually, pm already has a subset compiling to parrot. 00:06
I think most recent work has been with Tcl though.
arcady well, there's PGE, and there's a perl 6 grammar 00:07
or at least, will soon be
ruoso burning his CPU 00:12
arcady is there actually a the perl 6 grammar anywhere? 00:14
however incomplete
ruoso it seems to be in language/perl6
s/language/languages/ 00:15
that's what I'm trying to play with
arcady hm. seems kinda... incomplete 00:16
00:17 zgh joined
ruoso yeah... there is also the PCR work by fglock... which seems to be writing it too... 00:17
I still don't know how both are interacting... 00:18
arcady well, yes... I think there should be one official "the perl 6 grammar"
ruoso well... that are the synopsis... :)
00:18 amnesiac joined
ruoso there are implementation differences 00:19
that "bottom-up" "top-down" thing...
arcady or rather, the fact that you need both
LeTo TimToady: I've built and tested 0.4.3 on x86, amd64, ppc-darwin, and solaris 10 w sun-cc and gcc ;) 00:20
ruoso arcady, yes... sure... but the choice of where to get one way or another may be different from each implementation... 00:21
arcady still, the idea is that all the multiple implementations have the same grammar 00:22
or should be, anyway
or rather, accept the same language
00:22 Leandro- joined
Leandro- "SELECT codigo, login, id, nome, data FROM kb_clientes WHERE $por LIKE '%$string%' LIMIT 15" 00:22
$por = table
whatōæ½xB4s wrong?
the query is ok 00:23
the problem is with $por
ruoso Leandro-, please join #perl.br
Leandro-, there is likely to be the place for this kind of problem
s/there/that/ 00:25
BTW... the work in languages/perl6 seems really nice 00:26
00:31 stu7 joined
ruoso just found a way to segfault parrot 00:31
avar ruoso: what svn repos?
stu7 - unable to decipher a proper answer from the channel topic, asks stubbornly if perl6 is done ? :) 00:32
buu NO
stu7: It'll be done by christmas.
ruoso svn.perl.org/parrot/trunk
stu7 thx btu :)
avar buu: Will there be ponies? 00:33
00:33 stu7 left
ruoso heh 00:33
buu avar: YES. 00:34
Yay ponies!
avar What can you do with languages/perl6 ? 00:35
arcady parse some small subset of perl6 00:36
ruoso can see pretty tree-like ascii art
avar arcady: without haskell?
ruoso avar, yes 00:37
avar arcady: with ponies?
FurnaceBoy_ omg!!! ponies!!!
avar how does that work?
subset of perl6 in C?
ruoso avar, PGE and Parrot 00:38
avar someone handwrote a Perl 6 compiler in PGE?
*subset of Perl 6 .. 00:39
arcady you mean, in perl 6 rules?
ruoso that's the idea
arcady nothing too tricky about that
someone handwrote PGE in PIR, though
ruoso nothing too tricky until you can emit PIR from this parse
ruoso plans to port lrep to parrot 00:41
arcady hm. PGE makes it much easier to write parsers in parrot
ruoso arcady, far easier... from what I just saw... 00:42
avar unfamiliar with all the acronyms
arcady PIR is the language that parrot compiles and runs, very low level 00:43
ruoso yes
arcady a little bit above assembly
00:43 qu1j0t3 joined
ruoso arcady, a bit more than you're thinking... 00:43
arcady PGE is the Parrot Grammar Engine, an implementation of perl 6 rules written in PIR
00:43 qu1j0t3 is now known as FurnaceBoy
arcady which compiles rules to PIR, I think 00:43
avar svn co
ruoso yes 00:44
00:48 Khisanth joined 00:49 q[uri]_ joined
avar that language/perl6/ stuff isn't the whole thing is it? 00:52
It's like 150 lines
ruoso avar, considering it does'nt include PGE
it's all the stuff
audreyt avar: it's a beginning of the whole thing, must as lrep is also a beginning of the whole thing :) 00:53
also, it's currently just a simple parser
clkao audreyt: yo
audreyt there's no compilation involved.
clkao: yo
clkao so what's with perl6-binding
i don't think it could provide ro alias. but could probably steal the grammar 00:54
audreyt I havn't looked closely but it seems the src filtering is especially good
exactly what you said.
clkao ya. you have some time for it? :P
audreyt this weekend for sure :)
clkao it's easter weekend. i will hopefully be hacking
lots of svk backlog to clear
audreyt woot 00:55
my birthday is coming in 6 days
clkao ya.
obra happy almost birthday, audrey
audreyt heh :) thanks
svnbot6 r9913 | audreyt++ | * Dan Kogai reported that {my $x; $^y}.(42) is broken 00:58
r9913 | audreyt++ | because extractImplicitParameters does not look inside
r9913 | audreyt++ | a Pad (or Sym) node. Fixed.
01:19 lisppaste3 joined 01:32 hcchien_ is now known as hcchien 01:59 mako132_ joined 02:26 scw joined 02:49 r0nny joined
svnbot6 r9914 | fglock++ | PG-P6 - StatementControl.pm passes more tests 03:14
03:16 fglock joined, cjn joined 03:28 jserv-- joined, Soga joined
TreyHarris that went really well... i tried an experiment with my perl (5) class today... i taught them closures immediately after teaching them ordinary subroutines. and they actually got it, most of them 03:29
svnbot6 r9915 | fglock++ | PG-P6 - added failing tests 03:38
03:45 mako132_ joined 03:57 cjn left 03:58 FurnaceBoy is now known as FurnaceBoy_ 03:59 fglock left
gaal audreyt: when does the Big Pad Refactoring take place? 04:42
04:42 nirgle joined
gaal audreyt: alternatively, wanna work on S08? 04:42
04:43 nirgle joined 05:10 justatheory joined 06:02 nothingmuch joined 06:06 buu joined 06:21 f0rth joined 06:46 bsb joined 06:56 tifo joined 07:02 Aankhen`` joined 07:08 drrho joined
azuroth g'day, drrho 07:27
drrho hiho azuroth! 07:28
08:16 dakkar joined 08:18 xinming joined 08:25 Frances joined 08:39 marmic joined 08:53 q[uri] joined 08:56 wilx joined 09:02 pdcawley joined 09:09 chris2 joined 09:36 marcus_ joined
marcus_ seen audreyt 09:36
wolverian ENOBOT 09:37
marcus_ :'(
09:43 bsb_ joined 09:56 avarab_ joined 09:58 cognominal joined 10:01 cognominal joined 10:03 b_jonas joined
rafl Does anyone know who `ruoso', who were here yesterday and complained about the Debian packages of parrot, is? 10:07
10:07 cognominal joined
ayrnieu his nick is registered. 10:09
10:09 Grrrr joined
ayrnieu the nick, rather. 10:09
10:12 avar joined
rafl ayrnieu: How does that help? :-) I'd like to have a email address or something :-) 10:13
10:13 broquaint joined 10:14 wolverian joined, colares joined, PerlJam joined, weinig|sleep joined, kattana_ joined, ko1_away joined, evalbot_9900 joined, chip joined, webmind joined, xX[ReP]Xx joined, chihchun joined, merlyn joined, pasteling joined, revdiablo joined, webmind_ joined, oozy joined, PerlJam joined
ayrnieu rafl - ruoso has MemoNotify and Memosignon active, so you can send a message to him through memoserv, and he'll see it the next he logs in. 10:15
10:17 ko1_away joined
rafl Ah, thank you. :-) 10:21
10:49 broquaint joined
audreyt gaal: pong 10:53
I'll have full cycles for S08 at 36 hours from now 10:54
we can do that before BPR
since I think both involves P.A.Internals change
clkao bpr?
audreyt Big Pad Refactoring
switching from runtime pad allocation to compile time
as well as nail down all the storage classes 10:55
clkao whoot
audreyt (state, my, let)
dakkar how will 'temporary' and 'hypothetical' be implemented? 10:57
'temp' is not a storage class, is it? 10:58
audreyt no... it's installing a POST callback to someone's .TEMP 11:01
ditto "let" vs UNDO
let and temp are not declarators at all
sub f { let $x = 3 } 11:02
this is illegal
unless $x is visible
sub f { let my $x = 3 }
dakkar 'let' is the 'hypothetical' assigner? (I'm getting lost again...)
audreyt yeah... "let" means temporize but only revert in UNDO, not POST
s/POST/LEAVE/ 11:03
wolverian s,let,maybe, # :)
ayrnieu (what? perl6 has backtracking? Nevermind! Carry on!)
dakkar and UNDO is called on fail...
11:03 TimToady joined
dakkar is away: pranzo 11:06
11:07 SamB joined, bsb_ is now known as bsb
bsb audreyt, ingy: I'll be in Taipei 21-26 of April 11:13
for coffee debt settlement
audreyt bsb: ooooh 11:15
excellent
gaal audreyt: ACK, cool 11:16
11:18 colares joined, weinig|sleep joined, kattana_ joined, evalbot_9900 joined, chip joined, xX[ReP]Xx joined, chihchun joined, merlyn joined, pasteling joined 11:20 elmex joined 11:27 Southen joined 11:29 azuroth joined 11:32 nirgle left
azuroth bye bye! 11:33
11:35 azuroth joined, Southen joined 11:53 kanru joined 11:55 Khisanth joined 12:07 ruoso joined 12:08 marmic joined 12:28 kolibrie joined
dakkar is back (gone 01:32:14) 12:38
12:44 Limbic_Region joined 12:49 weinig|sleep is now known as weinig
ruoso playing with parrot/trunk/languages/perl6 12:51
azuroth ruoso: I think rafl was looking for you, for bad mouthing his debian packaging skills ;-) 12:56
ingy bsb: awesome 12:57
ruoso azuroth, people is in a bad mood this days... 13:01
bsb ingy: yep, tickets booked
ruoso azuroth, I just asked him a question and made a suggestion... 13:02
azuroth well, only half of what I said was true 13:03
I'm assuming he'll be able to fix it up all good
good night, all 13:04
rafl ruoso: Hi there! 13:05
ruoso rafl, hi
rafl ruoso: You said some points I want to comment on: 1. I don't think a parrot-svn package make sense. Debian has its experimental branch for that. 13:07
ruoso: I just don't have the time to keep that up to date very often, so help is appreciated.
ruoso: 2. parrot-0.4.3 is packaged and uploaded.
ruoso thanks 13:08
rafl ruoso: That happened on the day it was released. Since then it's waiting in the Debian NEW Queue: ftp-master.debian.org/new.html
That's because parrot changes its binary interface and therefor the SONAME of libparrot on every release and the package names in Debian need to reflect that.
And as soon as you introduce a new package name, your package will be put into NEW for ftp-master approval. 13:09
ruoso rafl, oh...
rafl, ok
rafl There is no possibility to fix this inside Debian. Parrot needs to get a stable API and someone who watches out for changes.
You also said that automatic daily builds should be in Debian and they should be autobuilded on all archs to check parrots quality. 13:10
That's a bad idea as well. Even if it's called "unstable" it's not the place to upload random snapshots of any given software. That's what experimental is for. 13:11
Also the Debian buildds are already stressed enough.
ruoso ok... thanks... 13:12
rafl I think, as I said already, experimental is a good place for parrot svn snapshots.
Experimental gets autobuilded on some archs, so that could help to improve quality as well.
ruoso: So if you have some spare time and want to improve the situation I'd be really glad if you could prepare some svn snapshots from time to time. 13:13
ruoso rafl, ok... sorry if what I said seemed as a complain... it was not my intention...
integral of course, someone could build their own .debs and put them on their own host, if they want to provide very unstable svn snapshots 13:14
13:14 Qiang joined
rafl integral: Of course. I've done that on feather, but I didn't had time to set the autobuilder up again. 13:15
13:16 kakos joined
rafl ruoso: I'm fine with that. I just wanted to comment on some things. 13:18
13:19 kanru joined 13:24 ardder joined, snowstalker joined, SamB joined 13:44 fglock joined, snowstalker joined
svnbot6 r9917 | fglock++ | PG-P6 - moved Pugs-Compiler-Tokenizer into Pugs-Grammar-Perl6 13:59
ruoso needs help with pge... 14:01
14:01 snowstalker joined
ruoso I can match "my $a = (1,2,3)" but can't match "my $a = ()".... 14:02
14:02 vel joined
fglock ruoso: hi 14:03
ruoso fglock, hi 14:06
fglock, I'm playing with parrot+pge+languages/perl6
I think the problem is with 'circumfix:( )' 14:07
which only works if there is actually something inside the ( )
fglock you can see if there are already tests for this 14:10
14:11 justatheory joined 14:14 vytautas joined 14:17 vytautas joined 14:20 nothingmuch joined 14:28 hexmode joined 14:32 webmind_ is now known as webmind 14:33 wilx joined
fglock are there docs on how to parse '$a<2' vs '$a<2>' and division vs. match? 14:42
wolverian afaik infix < needs whitespace 14:43
ayrnieu fare thee well, terse maths. With FORTRAN you came, and you've since always been leaving. Fare the well. 14:45
fglock ?eval 1<2
14:45 evalbot_9900 is now known as evalbot_9917
evalbot_9917 bool::true 14:45
integral ayrnieu: don't dispair! redefine the grammar to be more like haskell!
15:18 pmurias joined
pmurias hi all 15:18
fglock pmurias: hi 15:23
15:28 fglock left 15:32 pmurias joined, chris2 joined
pmurias fglock: i get an error from perl6.pl: Pugs::Grammar::List is missing 15:33
15:41 KingDiamond joined 15:59 KingDiamond joined 16:03 justatheory joined 16:19 xinming joined, FurnaceBoy joined 16:20 fglock joined
fglock pmurias: fixing 16:25
svnbot6 r9918 | fglock++ | PG-P6 - added List.pm 16:27
16:36 ghenry joined 16:51 FurnaceBoy is now known as FurnaceBoy_ 16:58 bsb left
kolibrie fglock: do you have an example written out like you propose for your lrep Talk: source text, grammar, compiled grammar, grammar generates AST 17:00
I wrote a small grammar file, now I want to compile it and use it to generate an AST from my source text 17:02
fglock kolibrie: svn.openfoundry.org/pugs/misc/pX/Co...le/Rule.pm 17:08
this is the grammar file used to parse rules in PCR
the AST is generated by the return blocks 17:09
kolibrie fglock: What command do I run to process a source file with a grammar? 17:10
fglock $match = $rule->match( $source ); - the ast tree is in $match->() 17:11
or $match = Grammar::rule( $source ); 17:12
kolibrie fglock: I'm sitting on the command-line with a grammar in p6, and a text file. How do I run the grammar over the text file? 17:13
17:14 zgh joined
kolibrie fglock: do I write a small p6 script that requires the grammar, reads the text file, and runs the match? 17:15
fglock you need to compile the grammar to a p5 package, and then require/use it in a program; then read the source file; apply your main parse rule and get a match object; then get the capture from it ... 17:16
exactly
kolibrie so, to compile the grammar to a p5 package, what do I run?
fglock or, you can split the grammar into rules and compile each one directly in the p5 program 17:17
you can use lrep to compile the grammar
kolibrie I ran 'perl p6compiler.pl grammar_test.p6' 17:18
I can't see that it did anything
fglock if you choose to write in pure p5: $rule = P::C::R->compile( ' <rule><rule> ' );
you may need to supply some command line switches 17:20
p6compiler.pl --print-program xx.pl > xx.pmc 17:23
kolibrie tries
17:24 xinming joined
kolibrie fglock: yeah! I got something 17:25
fglock: now, to use that in my main program, I need to require my generated p5 module 17:26
then how do I call it?
fglock if you compiled a .pmc, you can just 'use' it - perl will load the .pmc 17:27
then: my $match = Grammar::rule( $source ); - use you grammar and rule names
17:28 ruoso joined
kolibrie fglock: and $source is a string from my text file 17:28
fglock sorry: my $match = Grammar->rule( $source ); - rule is a method 17:29
kolibrie begins to get excited
fglock yes
prepare to have Data::Dumper around :) 17:30
kolibrie fglock: no problem
17:31 colares joined
ruoso kolibrie, if everything goes fine, you should be able to recompile the grammar with the generated p5 code 17:34
kolibrie ruoso: that sounds like an adventure for another day 17:36
ruoso :) 17:37
kolibrie fglock: perl parse_text.pl text_file.txt 17:45
Undefined subroutine &ruleop::constant called at grammar_test.pm line 8.
how do I make it require/use the correct lrep files? 17:46
fglock: I added the lines from the p6compiler.pl header subroutine to my parse_text.pl file 17:56
now my errors are gone, and Data::Dumper gets pulled into use 17:57
18:00 lichtkind joined
lichtkind the ! is for things like : if not ....{} ? 18:01
18:06 dduncan joined
lichtkind in the s3 table stands only symbolic unary for the ! but where can i get more info? 18:07
kolibrie fglock: hmm, all my matches seem to be $VAR1 = undef; 18:09
TimToady lichtkind: if anything is undocumented in a synopsis the assumption is that it is handled like Perl 5 does. So yes, unary ! is just a very tight "not". 18:19
lichtkind thanks mr toady
TimToady if !$x == $y is if (!$x) == $y 18:20
lichtkind tim toady and thanks for answering my mail
TimToady so it's often tighter than you like.
you're very welcome
lichtkind tim toady i currently write an tutorial about perl6 and once i wrote you a mail with some questions about beginnings of perl because i maintain nearly all perl related articlres in german wikipedia 18:21
TimToady ah 18:22
lichtkind tim toady but you nearly answered it with your last talk in isreal
TimToady I'll be a little less conferencey for the next month or so, so I should be available to answer any remaining questions.
lichtkind its just a small one :) 18:23
TimToady by Christmas.
:)
lichtkind :)
we had hard disputes about that because i said that you cant seperate an christian attitude from the rest of pearl, some had problems with it , i think i understand that even if im not christian 18:24
dduncan so, did any of you submit proposals to talk at OSCON 2006, and get word back as to whether they were accepted or not? 18:26
18:26 KingDiamond joined
TimToady one cannot help but be who one is (at least without some kind of divine assistance.:) 18:26
dduncan I made a proposal and found out a half hour ago that I won't be giving it ... too many good submissions or something
TimToady or too few good time slots... :-) 18:28
lichtkind TimToady sorry im no nativw speaker 18:29
TimToady sorry
you have to be who you are
unless you get outside assistance
miyagawa dduncan: I made a submission of my talk to OSCON 2006 and I got a reply from them, a couple of weeks ago 18:30
saying the talk was accepted
dduncan hm, makes sense 18:32
they probably would confirm all the yesses first
TimToady me, I have to give a talk whether I want to or not... 18:33
miyagawa that was 2 weeks ago, to be more accurate
dduncan yes, you're spe-shul
miyagawa TimToady: seen this? japan.cnet.com/interview/story/0,20...857,00.htm
18:33 weinig_ joined
Limbic_Region TimToady - did you see the short exhange I had with audreyt concerning the change in the "nature of the list" over say the last 6 months 18:34
TimToady yes, I did. I bl more assiduously than even Audrey. :) 18:35
miyagawa: no, I hadn't. thanks. 18:36
miyagawa doitashimashite.
dduncan anyway, I have a few questions or discussion points that I'm not sure whether I should bring up here or on p6l
maybe here first
to setup a context, this has to do with my Rosetta relational DBMS project ... 18:39
18:40 weinig_ joined
dduncan my current plan is to have its native interface language look like or be Perl 6, since in large part Perl 6 seems to include all the relational algebra operators or close analogs to them 18:41
18:41 FurnaceBoy_ is now known as FurnaceBoy, weinig_ joined
Limbic_Region TimToady - at the risk of opening up a can of worms - what is the current process for marrying Parrot with Perl6 (IOW - what is being done to ensure Parrot supports the language) 18:42
lichtkind TimToady i even read english books but some native phrases are still new to me, would you say that the way perl springs directly from your christian attitude or was it just christian to release it so does everybody else could use it too? 18:43
Limbic_Region seldom sees the VM camp agreeing with the language camp
ayrnieu you don't have to preface wormcan-opening with "at the risk of". If worms are freed, you can just say "I *knew* that this would happen!"
dduncan or initially, I will have the language be a very small subset of Perl 6 grammar and would parse it myself, so I have something which can be implementable in multiple languages and/or translatable/mappable to multiple languages
eg, all operators would simply be prefix format with parens around their arguments 18:44
TimToady Limbic_Region: I think once we get Perl running on top of Parrot in *any* form, there will be plenty of pressure to reduce any impedance mismatch between Perl and Parrot. 18:45
dduncan a very simple grammar is easy to make a simple parser for
TimToady so I'm not too worried about "Parrot supporting the language".
Limbic_Region ok - fair enough
I am just concerned about the number of things that have been relegated to the VM - coroutines, threads, async io, etc 18:46
well, concerned is probably not appropriate there
ayrnieu (audreyt seemed keen in her perlcast about multiple implementations, however, so 'parrot supporting the language' may also have the pressure of 'this specced thing works /over there/'.) 18:47
Limbic_Region decides to shut up
TimToady lichtkind: I'd say both of those are true, for some definition of "christian". certainly there are a lot of people who would claim to be christian who behave differently, but to me, I see a God who simultaneously can have a central plan but also likes a lot of diversity around the edges. 18:48
Limbic_Region takes refuge in the fact that smarter people then him are handling it
dduncan but I also wanted to make it such that one could essentially code things in plain native perl 6 such that it looks and works about the same
er, I'll come back to this another day, when I actually have documents to show
TimToady dduncan: okay 18:49
dduncan or maybe I'll still ask a few questions
1. part of the perl 6 design is that one can make custom grammars for the language itself ... 18:50
if I wanted to restrict things in some way, such as say that one can only code using strong data types, can I do that?
I prefer to push some kinds of checking to compile time, so I don't have to check at run time that pieces are giving each other valid formatted data 18:51
TimToady you can do anything you like with a given lexical scope.
dduncan 2. to clarify, it is now possible to have multi-member hash keys, so I can look things up by, say, 2 or 3 distinct values rather than one? 18:52
I thought I saw something like that in S9
TimToady it's just that, the more you diverge from the expectations of the naive programmer, the less they'll understand your code without studying up for it.
%foo{slice1; slice2; slice3} is allowed, where each slice is a lazy list. 18:53
dduncan eg, %foo{'bar','baz'} = 17;
TimToady You'd have to use >>=<< to set them all. 18:54
dduncan and $quux = %foo{'bar','baz'}; # contains 17
TimToady or have a pragma that turns all binops into hypers.
you're confusing , with ;
dduncan yes 18:55
TimToady ; separates dimensions, comma only separates list items.
dduncan sorry, I wrote that before seeing your previous line
TimToady but yes, with ; that should work.
dduncan for context, I want to have a data type that is like a Relation
TimToady it might be naively implemented as %foo<bar><baz> 18:56
but you can declare multidimensional containers.
dduncan now, the nature of a Relation is that you can have multiple keys over the same data 18:57
lichtkind tim toady mille gracie (1000 thanks) thats exactly how i felt about it but you know, i cant write it if i cant quote you on that. personally i also believe in god and agree with you in most things, maybe thats why i like perl but i always have problems with christians when it comes to the interesting questions :)
TimToady lichtkind: yes, well, the interesting questions are of the form "what's really at the center?" 18:58
dduncan eg, for a person relation, one attribute social security num could be a key, and in limited senses, a name could also be a key; those don't overlap
TimToady dduncan: go on 18:59
dduncan if a hash is like a key, its like multiple hashes that have common values, and adding a value to one adds it to the other too; this hash multiplicity being encapsulated in the one relation container
TimToady or like a single hash that the first dimension selects the "real" key underneath. 19:00
dduncan a proper relation definition is a set of distinct tuples that all have the same set of attributes
everything is addressed by name, not ordinal position, like a hash
the name being a key value when selecting tuples from a hash 19:01
the name being the attribute names when selecting what people call columns in sql
TimToady right. so you'd kind of like %hash{index; key} to let you have multiple entries be the same entry under different indexes. 19:02
dduncan in a way, yes
except that you first specify which index set, and then which values in it
and each index may have different numbers of values in it than others 19:03
TimToady right, the type of the subsequent dimensions varies depending on the index
dduncan and the values are each like hashes themselves
lichtkind tim toady thats easy in the center is loving consciousness i ment more the intelectual interesting questions like was abraham visited by aliens and like that, but i wont be here too OT
dduncan aka, each value is a tuple
I don't know whether it would be easiest for me to just define a perl 6 class which provides the functionality I'm thinking of, then show it to p6l for consideration of something analagous being in the language itself 19:04
19:04 _bernhard joined
dduncan fundamentally, relations and tuples are basic and very commonly useful types of containers, and something akin to them deserves to be in a full language 19:05
TimToady it's usually easier to discuss a concrete proposal that is certainly wrong than an abstract proposal that is of uncertain merit. :)
dduncan and if so, you could have a lot of people using Perl 6 itself for things they might otherwise use a separate database for 19:06
TimToady I agree.
xinming TimToady: I have a question on accessing kv pair in hash, in perl 5, eg: my %h = (abc => 1, xyz => 2); $h{abc}; <--- here , abc in $h{abc} is acted as a string, will this be changed in perl 6? eg, abc will become a method call instead of a string
TimToady the definitions in S09 are currently biased toward PDL-think, but if there are places where relational-think are more appropriate, I'm all ears.
dduncan so I think I will try making a concrete Perl 6 package soon that does what I'm thinking of, and then we could use that as a point of departure 19:07
TimToady dduncan++
dduncan incidentally, this is meant to be as fundamental as a Set type
and Set is currently a separate package in the Pugs tree
xinming This isn't discussed in Synopsis. And if it is the same as perl 5, what's the difference between %h{abc} and %h<abc> ?
dduncan Relation theory is essentially Set theory 19:08
TimToady if abc isn't a function that can take 0 args, the first is a syntax error.
or perhaps a semantic error detected when the end of the current compilation unit is reached without a valid definition of sub abc. 19:09
xinming ok, so, abc will be a method call.
oops. function call.
TimToady abc is abc()
19:10 justatheory joined
TimToady the only remaining autoquoter from Perl 5 is the left side of =>. 19:10
dduncan fyi, I don't know if you've heard of thethirdmanifesto.com , but I currently subscribe/incorporate their definitions of what the relational data model is, and not SQL's definition
just to put things in context
TimToady haven't had a chance to look at it yet, though I've seen your refs to it. 19:11
xinming TimToady: Is there a way to make variable name contains space?
or, did you ever consider it. :-)
TimToady ::<$I am a variable name> = 1; 19:12
Maybe require a MY on the front...
dduncan would the $ best go inside the <> or outside it?
cognominal TimToady, in your postcasted conference, you said that French are reluctant to borrow word. This is half true, this not because the academy sneer at it, or that a minister pass a law that we don't do it. 19:13
TimToady depends on whether you're interrogating the symbol table hash or trying to do a symbolic name.
xinming dduncan: It should be inside IMHO. and $<abc> === $/<abc>
cognominal you should not confuse the institution and the people :)
TimToady for the latter, you can do $::("I am a variable name")
cognominal s/borrow word/borrow words/
TimToady cognominal: I think everyone understands that it's just a stereotype. :) 19:14
dduncan regarding my concrete example, it may be 1-2 weeks before I produce it, since I have to devote a lot of time to unrelated $job
cognominal :)
dduncan but it will come
but I'll keep this example simple, try to 19:15
TimToady dduncan: I don't doubt it. You have the gift of persistence, and I mean that in a positive way.
dduncan thanks
fyi, I'm actually relieved my OSCON talk wasn't approved, as I would have preferred more time than I would have had for preparation 19:16
lichtkind why is there no ?!
TimToady um, what is ?! supposed to do? 19:18
ayrnieu the ternary operator is now named ?? !! , if you mean that. 19:19
lichtkind i know i mean as oposite of ?
xinming hmm, could we make @array ordered hash? Since, in @a[1], 1 here is something like a key in %a<1> :-P 19:20
ayrnieu lichtkind - you mean 'where did the zero-width negative lookahead assertion go?', in regex? 19:21
lichtkind nono
TimToady xinming: certainly. there's no reason you can't have a container that does both Hash and Array roles.
xinming doesn't think that is a good idea for now... :-P
lichtkind i mean boolean context operator
TimToady ?eval ?"yes" 19:22
ayrnieu you don't like !?"yes" ?
19:22 evalbot_9917 is now known as evalbot_9918
evalbot_9918 bool::true 19:22
TimToady ?eval !"yes"
evalbot_9918 bool::false
xinming bool:false
s/:/::/ 19:23
TimToady :)
er, ::)
lichtkind thank again # getting this tut completed
TimToady ?eval true "You're welcome" 19:24
evalbot_9918 bool::true
xinming ayrnieu: ? means it is bool context, and ! means "not", so, normally ! is enough, If you wish to know which is the same as force a bool context, you can use !! IMHO. :-P
lichtkind maybe i shoul translate it in english when im done
ayrnieu xinming - I would hesitate to do that, with ternary syntax. 19:25
xinming ?eval !! "yes?"
TimToady well, Perl won't get confused, it being a term rather than an operator, but a human might.
evalbot_9918 bool::true
TimToady but why would you use !! when ? does exactly the same thing?
xinming TimToady: for your moto, TIMTOWTDI :-) 19:26
TimToady TMTOWTDI is how I spell it. :-)
cuz i always write there's 19:27
xinming TimToady: audreyt ever mentioned, that in perl 6, block is not an expression, do you ever considered to changed this? eg: my $a = do if ?right { "right" } else { "nonono" }; 19:28
19:28 pawel joined
xinming hmm, though, I think, we should declare $a before we use that form. 19:31
lichtkind does ~~ impose sting kontext? 19:32
xinming also doesn't like the parenthesis in method call, as method in xinming's opinion is a kind of sub call.
lichtkind: no, It's smart match operator
lichtkind: ~ is string context.
lichtkind k thx
xinming lichtkind: yw 19:33
TimToady xinming: how would you want to write a method call then?
19:34 q[uri]_ joined
lichtkind if anybody here can read german, thats the tutorial im writing on: wiki.perl-community.de/bin/view/Wis...l6Tutorial 19:34
xinming TimToady: $obj.meth "p1", p2, p3; 19:35
TimToady what should $obj.meth + 1 do?
or $obj.meth +1
or $obj.meth / foo
that's why we put in the disambiguating : (like Haskell $) 19:36
so you can say $obj.meth: "p1", p2, p3
and Perl knows to expect a term after the : rather than an operator.
xinming ah
thanks 19:37
TimToady yw
q[uri]_ anyone seen allison around?
TimToady doesn't generally hang out here.
q[uri]_ TimToady: thanx. 19:38
19:38 q[uri]_ left
Limbic_Region wonders if allison has a stalker 19:38
TimToady if that was the uri I know, i'd say, no... :) 19:39
xinming hmm, what about the my $a; $a = do if ?test { "true" } else { "false" } example? 19:40
TimToady should work now if you want $a to end up with a string rather than a closure. 19:41
dduncan okay, I have another question
Limbic_Region oh - I didn't even see the uri in the <q[uri]_> 19:42
dduncan is there a defined generic equality operator that takes values as they are and doesn't impose a string or numeric context?
TimToady presuming ?test really means one thing.
and isn't trying to slurp arguments
19:42 david_ joined
xinming TimToady: hmm, I just mean the test statement. sorry for my poor English. :-) 19:42
dduncan eg, where you can say "Any comp Any", and it would return true if the 2 are of the same actual type and are equal within that type? 19:43
TimToady but <space>{ at the statement level is forced to be a block regardless of expectations.
yes, that's what === is supposed to do.
dduncan okay 19:44
xinming hmm, what I want to say might be... my $a; $a = loop(;test_statement;) { more_statements; $a }; 19:45
oops.
my $a; $a = loop(;test_statement;) { more_statements; $b };
then, $a will be $b when the loop finished 19:46
cognominal For those who read French, in the French GnuLinuxMagazine of April I wrote about the next Sun processor called Mehari that runs MiniPerl6 . rafb.net/paste/results/Bq6ToS88.html
TimToady we require "do" in front of any statement_control that is to be used within another expr for its value.
otherwise we get ambiguities with statement modifiers like "if" and "unless" and "for" 19:47
so you want my $a; $a = do loop(;test_statement;) { more_statements; $b };
except for the fact that loop is probably going to return the last value of the test statement by default, if it's like Perl 5. You'd need to leave() with an explicit value. 19:48
so by current definition that'd have to be 19:50
my $a; $a = do loop(;test_statement;) { more_statements; leave <== $b }
but I'm not entirely happy with leave's definition yet.
xinming hmm, does "if else" needs "leave" for returning a value? 19:51
TimToady no 19:52
doesn't test the expr again at the end of the loop
loops return conditional by "last expression evaluated" rule, but it might be possible to exempt the conditional expr, I suppose. Still, it's ambiguous, and I'd rather have an explicit loop exit if that's what is meant. 19:53
maybe "last <== $b" is possible.
or "last :value($b)" 19:54
19:57 bsb joined
xinming still has 2 questions about perl 6, but hesitate to ask... :-P 19:57
FurnaceBoy senses deadlock
xinming It's about class, I really like the word "is" used in the language... 19:59
eg: class Foo { has $.a is rw }; so, I like class Foo { method private_one is mine { ... } } instead of class Foo { submethod private_one } :-) 20:00
class Foo { submethod private_one { ... } }; 20:01
20:08 KingDiamond joined
TimToady xinming: different storage classes work better with different declarators out front. We could do "my $foo is constant", and in fact we used to, but we pulled out "constant" as a declarator when we realized the variable was scoped to the compiler. 20:11
similarly, "submethod" is something with scoping rules that are both like subs and like methods.
so it's better to pull that out front where it's obvious. 20:12
xinming thanks for your clearly description. :-) 20:16
my English sucks. :'( 20:18
s/clearly/clear/
lichtkind thanks for all 20:19
Limbic_Region does the bit in S5 about rule matching against non strings (streams and arrays) currently work in Pugs? 20:22
TimToady none("clue") 20:23
xinming perl is the worst language to learn, and best language to use. :-)
20:23 fglock joined
xinming I don't know if I'm right, But for reading Synopsis, I've learnt a lot. 20:23
TimToady I suspect English is worse to learn. I'm still working on it... 20:24
mugwump xinming: you mean "hardest", presumably
xinming mugwump: No, I mean worst
20:24 weinig_ is now known as weinig
mugwump Well, I found it the most fun when I first picked up the Camel book about 8 years ago 20:26
xinming It's also a fun when learn perl 5.
dduncan I also first was introduced to Perl 8 years ago 20:27
mugwump must have been "OO nut learns Perl" season 20:28
dduncan so the 2nd ed Camel book was my first, gotten then ... it was a course text book
Limbic_Region dduncan - you can't have been introduced to perl8 years ago
unless of course you are a time traveler
we are still working on p6
Limbic_Region intentionally misparsed
fglock kolibrie: ping - sorry, I lost the connection
dduncan it was January of 1998 to be specific
kolibrie fglock: pong (I saw it) 20:29
Limbic_Region 2002-07-11 for me
yep, I know the exact date
dduncan I am also fortunate to have started straight with version 5, and not have to deal with <= 4
Limbic_Region wanders off homewards
kolibrie fglock: not getting errors any more, but every $match comes back as undef
any ideas?
dduncan I know the exact date too, more or less ... it was around January 5th, whenever the winter quarter at college began 20:30
if I look up my course time table, I could know the exact hour too
fglock kolibrie: can you try a very simple match, like '.'? you can also try using PCR directly (see the pod) 20:31
pawel xinming: perl5 was the first programming language i learned
20:31 pawel is now known as pmurias
dduncan it was the computer science program's 1st year networking course (called 'comp 170' locally) 20:31
kolibrie fglock: so 'rule anything { . }' 20:32
20:32 kanru2 joined
dduncan our first homework assignment was to download and install perl 5, so that 'print "hello world\n";' executed 20:32
fglock yes
dduncan in my case, being on a classic mac, it was MacPerl 5.20 or so, embedding Perl 5.004
so 5.004 was my first version 20:33
xinming I love perl 5, because It has built-in hash, And It is the first language I learnt which contains built-in hash. :-P
pmurias using perl5 was a pain on Windows, so i always went to my fathers office to program on the NetBSD they used as a server 20:34
20:34 KingDiamond joined, claudio_ joined
dduncan while Perl is my favorite language now, and since 1999, ironically I started out strongly disliking it on first introduction because it was so un-strict (I was previously used to Pascal and Java) ... I was surprised that a program containing just a print statement would actually work 20:34
I ended up failing that course through neglect 20:35
then in summer of 1998, I decided to try using Perl to solve a practical problem of text processing, first by rewriting a Pascal program, and from that point I actually started to like it, and then in retaking the course in january of 1999, got an A+
Perl's been my favorite language since 20:36
kolibrie fglock: hmm, the match on 'anything' returns a match object containing the first letter of the package name 20:37
fglock: switching to Grammar::rule( $source ) gets the first character 20:38
dduncan er, when I say 'strict', I meant that I was used to programs having all their statements inside functions/ subroutines ...
I couldn't understand how a bare statement outside a procedure could execute
but that quickly became water under the bridge
kolibrie fglock: now I just will have to take baby steps to build up my grammar 20:39
fglock: can P::C::R take a whole grammar at a time, or only one rule at a time
fglock kolibrie: right - the 'old lrep' grammars were not OO 20:40
PCR only works on single rules, which belong to a grammar (p5 package)
dduncan and there was that we could just use variables without declaring them
those 2 things 20:41
but in my modern Perl coding, I put everything reasonably possible in subroutines and use strict
kolibrie fglock: so I could read my p6 grammar file rule-by-rule and spoon-feed it to PCR
fglock sure 20:43
xinming dduncan: Most script you see which are written in perl without use strict IMHO is all modified after use strict; debugging is done. :-P
pmurias i used to write my programs on paper at when first learning, and only type them in after the were complete
fglock you can easily write a grammar compiler this way - PCR can emit perl5 code 20:44
svnbot6 r9919 | fglock++ | PCR - fixed hash ordering
r9919 | fglock++ | PG-P6 - added a failing test - there is a bug related to expression-parser reentrance
kolibrie fglock: ok, I'll look at those docs
(probably tomorrow) 20:45
20:47 justatheory joined
fglock in the p6 parser - is the bottom-up parser supposed to be used for expressions only? 20:49
ruoso fglock, in parrot? 20:51
fglock, parrot/languages/perl6, i mean... 20:52
fglock in the parser 'as-specced'
ruoso I always thought it was more implementation-oriented than spec-oriented... 20:53
fglock I mean, which parts actually 'need' to be parsed with rec-descent 20:54
bsb I thought the parsers overlapped, bottom up for speed, rec-descent for sane error messages 20:58
21:01 blb30346 joined
fglock bsb: I was thinking about trying to use bottom-up in more places, but I wonder what things it would break (error messages may be one of them) 21:02
bsb I don't think there's any limits defined 21:07
so use it if it works
ruoso I think the best combination will be the one that is written first... 21:13
avar fglock: go crazy with it;) 21:14
21:15 feng joined 21:16 Limbic_Region joined 21:19 nirgle joined 21:25 david__ joined 21:35 lichtkind joined 21:37 KingDiamond joined 21:38 lumi joined 21:44 frederico joined 21:51 nicbrown joined 21:55 weinig is now known as weinig|out, bsb left 21:56 fglock left
avar will objects methods in Perl 6 be fairly complete? 22:06
I.e.
In Perl 5 you didn't have them, and hence you'd pollute the core too much
you you can have lots of object methods which make sense, like they're in their own namespace
one for lists, one for ints, etc. 22:07
But Perl 6 should be able to have (2,3,1).max 5.hex, 10.bin(ary) (or something), i.e. a fairly complete set of object methods 22:08
@list.uniq @list.top (the "biggest" element, or something)
Has there been some discussion or spec on this?
$cont = $file.read/slurp 22:09
TimToady avar: that's what MMD is all about. You can have a slurp method without requiring the Str class to know about files. 22:14
avar TimToady: Exactly, I was wondering if the direction for std Perl 6 would be a fairly minmal set of methods for str, int, num etc (5-10) or "go crazy" in the knowledge that it wouldn't pollute the rest of the knowledge 22:15
I.e. be about ruby's level in methods per type 22:16
or, perhaps a bit higher;)
ruoso I think this is a point where OO starts getting weird... 22:17
what if I want to implement a new operation with files
and I dont want to mess with the standard types
open() returns the type that doesn't know about my new operation 22:18
and then I have to "decorate" the object
(java does this a lot... every time... it sucks...)
avar why wouldn't they be compatable? 22:19
-5.abs => 5, and everyone understands five
Or am I misunderstanding you?
ruoso maybe
TimToady -5.abs yields -5, actually...
avar TimToady: What do you know about Perl, just some random yahoo on IRC 22:20
ruoso what if you want a new operation 5.foo
avar >;)
TimToady multi foo (5) {...}
avar ah, because methods call are so high up the presidence table, I was reading about that the other day 22:21
yes of course
ruoso so... foo is not part of 5 interface... 22:22
you just add this feature to it...
TimToady I seriously doubt the 5 type knows about foo...
ruoso $foo = 5; $foo.foo; 22:23
TimToady the surprising precedence thing right now is that, if we make non-variable $ a unary, $$foo.bar is $($foo.bar).
ruoso: still works, because SMD will fail over to MMD. 22:24
ruoso needs to study more on Perl 6 method resolving... 22:25
avar If I want to help with Perl 6 now without it taking up all my life what's my best bet?
writing Perl 6 testcases?
ruoso avar, the one you have more fun :) 22:26
nah... actually, that will end taking up all your life..
22:26 nirgle left
avar writing test cases would be fun;) 22:27
LeTo $ perl -le 'print (-5)->abs' 22:31
-5
Can't call method "abs" without a package or object reference at -e line 1.
$ ./pugs -e'say (-5).abs'
5
what's the problem ;) 22:32
avar -> is not the method dispatches anymore 22:33
It make some kind of anon sub (IIRC), e.g. for (%hash.kv) -> $k, $v { ... }
LeTo err $ perl -le ...
that was perl5 of course 22:34
avar fleh, failed to see that,)
LeTo and the second pugs example certainly did work 22:35
22:35 larsen joined 22:37 broquaint joined
LeTo avar: writing tests is the best start to dig into things 22:41
avar What project should I do it for?
aren't there three seperate projs. or something, parrot, pugs ...
LeTo whatever you prefer 22:42
projects are distinct now still, and have common parts, increasingliy so (I hope) 22:43
avar no I was hoping for a list of projects and which one was best suted ;)
suited for what I'd want to do;)
and a pony 22:44
LeTo sure but compiler and runtime are separate parts in the perl 6 worldd 22:45
s/dd/d/
22:45 Qiang joined
LeTo re pony: this is still seeking a new rider 22:46
avar you mean parrot / perl ?
LeTo the ponie, yes 22:47
ruoso there is work going on on languages/perl6 too... 22:48
and it looks good..
avar languages/perl6 is like 300 lines of code
ruoso avar, that means it will be easier to start playing with it... 22:49
LeTo use.perl.org/article.pl?sid=05/12/21/2026205
use.perl.org/article.pl?sid=05/12/1...amp;tid=33
avar They're just switching backends, not programming the whole thing again, right? 22:50
ruoso BTW... is it possible to say: Int does MyRoleFoo; 5.foo; ? 22:51
LeTo if MyRoleFoo has a method 'foo' that takes an Int invocant, sure 22:52
ruoso is it scoped? 22:53
or just *can be* scoped?
LeTo it's scoped to the rules of P6 scoping, i.e. whatever you want to achieve 22:54
ruoso constantly gets even more impressed with all this stuff.. 22:55
LeTo that is - you can scope it to any(global,file,lexical)
and class 22:56
ruoso that's a really smart way of re-implementing the "decorator" pattern... 22:57
ruoso just hates this in the java api...
new LineReader(new Reader(new FileInputStream(new File(name)))) 22:58
this is the java way
LeTo python seems to go that direction recently
ruoso which one?
LeTo decorators 22:59
ruoso gah
this is a PITA for the programmer...
ruoso is working with Java(arg..) in the last 3 years... he really hates the ultra-verbosity of the java API 23:01
LeTo well, "there's no way to do it" - find some extra syntax ;-)
that was of course py* related 23:02
ruoso heh
fits very well for java
Iterator i = v.iterator(); while(i.hasNext()) { Hashtable h = (Hashtable)i.next(); } 23:03
I keep writing this all day.
LeTo dunno java - I like to drink *good* coffee though
above code contains a cast, isn't it 23:04
Hashtable h = (Hashtable)i.next();
that one
ruoso yes
I just hate it 23:05
LeTo that's just b0rked
ruoso realizes he will use "does" a lot... 23:06
LeTo what does the runtime, when you code '(String)i.next();' for that? # and s/String/*/ - any other dunno type 23:08
23:09 chovy left
ruoso you mean: "Hashtable h = (String)i.next()"? that's a compile-time error... 23:10
LeTo ok - and when (xxx) is user defined?
ruoso (OtherClass)object just checks... 23:11
the object doesn't need to change...
it does not change
LeTo e.g. 'xxx' isa 'Hashtable'
23:11 kanru2 is now known as kanru
ruoso it works... 23:11
but You can't access any methods defined by xxx
wait...
arcady btw, the new java lets you get rid of the most common use of iterators 23:12
ruoso you mean in "Hashtable h = (xxx)i.next()"
arcady for(Hashtable h : v) { ... }
ruoso keeps with java 1.4.2 until java is free...
arcady 1.5 actually is better
fewer casts too
and... it will never be free
ruoso and what about claspath and kaffe and gcj 23:13
avar ruoso: does gcj implement 1.4.2 ?
ruoso almost...
avar Cool
does gcj use classpath?
ruoso yes and no... 23:14
they're brother-projects
avar But gcj can't compile normal code without classpath right?
since it's the stdlib
kind of like gcc / libc
...
ruoso www.kaffe.org/~stuart/japi/htmlout/...spath.html
ruoso is going home^Wto the pub 23:23
23:29 lichtkind joined
lichtkind ?eval my $a := my $b := my $c; 23:34
23:34 evalbot_9918 is now known as evalbot_9920
evalbot_9920 \undef 23:34
lichtkind ?eval my $a := my $b;
evalbot_9920 \undef
lichtkind ?eval my $a := my $b := my $c; $a = 2; $c; 23:35
evalbot_9920 Error: Can't modify constant item: VUndef
lichtkind ?eval my $a := my $c; $a = 2; $c;
evalbot_9920 Error: Can't modify constant item: VUndef
lichtkind ?eval {my $a := my $c; $a = 2; $c;}
evalbot_9920 Error: Can't modify constant item: VUndef 23:36
lichtkind do anybody here know if i can $a := $b := $c ?
arcady ?eval my ($a, $b, $c); $a := $b := $c; $a = 2; $c 23:38
evalbot_9920 \2
lichtkind ahh thanks
arcady you were binding $a not to $b but to the value of (my $b), which is undef
23:39 Khisanth joined
lichtkind arcady why is my $b only a value ? 23:41
audreyt because it's a bug, that's why 23:45
this weekend's GPR (Great Pad Refactoring) should fix that 23:46
clkao hola audreyt
arcady excellent... now that I think about it it does seem wrong
lichtkind hello audrey
audreyt yo 23:47
final day of $job this week 23:48
audreyt is _so_ looking forward to weekend
obra aw. sad.
clkao me too, om
obra <- greedy
clkao one more day to easter weekend
audreyt heh
obra: yeah, each $job day so far results in multiple jifty commits and LML releases and whatnot :) 23:49
mugwump looks at clock
avar Maybe you guys will know this (of all people), I'm building a an application at wrk that consists of pure perl modules and (well, will) consist of some XS code as well. Now, my main namespace is Wrk::, I want Wrk.pm to be pure-perl, Wrk/Some/C to be xs etc. Problem is, I can't convince ExtUtil::MakeMaker to properly make my xs stuff if I put it anywhere other than in /Wrk.xs and /lib/Wrk.pm
mugwump about 4 hours until end-of-week here
obra is optimizing Jifty
avar I tried putting my XS stuff deeper in the tree and create a Makefile.PL there like perl the interpeter and its stdlib do, but the main makefile didn't pick up on it
audreyt avar: /Wrk.xs and /Wkr.pm works too.
TimToady ?eval system q/perl -le '(my $a) = 42; print $a'/;
avar I know it's possible somehow, but I can't find anything in the docs
evalbot_9920 Error: No such method in class Str: "&system" 23:50
lichtkind audreyt i was that guy at gpw in bochum with the editor in perl i wanted aboriginal to ask you if you have special needs for features because your known to have some extreme programming style :)
TimToady heh.
avar audreyt: yeah of course, the point is that I want it in /ext/Wrk/XSStuff.pm/xt etc.
I was trying to poke the perl build process to find out how it was done;)
audreyt TimToady: you expect otherwise? :)
avar: WriteMakefile(XS => { '/ext/Wrk/XSStuff.xs' => 'XSStuff.c' }) 23:51
arguably there should be a cc_xs_files() or xs_files() command in Module::Install to do that for you
TimToady works in my pugs, and prints 42.
audreyt TimToady: right, it's just the bot has safe mode on. 23:52
TimToady I just find it odd that system doesn't exist at all.
avar audreyt: That'll work for doing that?
audreyt avar: I think so. it's even documented in EU::MM
TimToady: *nod* maybe bind it to something that throws an exception?
TimToady always in favor of more accurate error messages... 23:53
avar audreyt: I read about it before actually, but it doesn't explain what it's for which didn't make me relate it to the problem I was having.
TimToady and even more precise error messages, as long as they're accurate... :)
avar audreyt: so that'll find stuff in ext/..., xsubpp it to XSStuff.c which'll then get picked up in the compilation 23:54
cool;) 23:55
mugwump avar: you'll probably need to set the 'C' and maybe 'OBJECT' key, too. I could never get .xs compilation in a sub-directory working, though. best perhaps to use a whole sub-module (ie, dir with its own Makefile.PL)
avar I tried using a seperate Makefile.PL, but the main file didn't pick it up and I'm not sure how to make it..
audreyt TimToady: implemented. testing... 23:56
avar ah, there it is;)
DIR =>
audreyt lichtkind: extreme programming has some good ideas...
avar audreyt: Like testcases;)
audreyt I'm not sure how/why my style is considered extreme though :)
lichtkind audreyt i thought more in that way that you handle lot of stuff at once, i dont ment xp literally :) 23:57
avar Extreme Programming sounds scary
lichtkind ok if you have crazy ideas let me know :) 23:58
avar defenitly so
avar I always imagine some guy typing after having had 1 litres of coffee when I read it
TimToady a liter here, a liter there, pretty soon you're talking Real Coffee...
avar is reminded of a Bill Hicks quote 23:59
audreyt or Double Espresso...
avar "2 packs a day? You pussy! I go through two lighters a day"
stevan heya audreyt :)
audreyt with Floating Creams

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!