pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
pugs_svn r19770 | rhr++ | [Unicode.pm] AnyChar, Byte, Codepoint 05:05
r19770 | rhr++ | make codepoint level "universal" per S02, respect $?NF
pugs_svn r19771 | rhr++ | [Unicode.pm] StrPos, StrLen, substr 06:39
riffraff hi 09:21
moritz_ hi riffraff ;) 09:29
cosimo hi 09:30
pugs_svn r19772 | ruoso++ | [smop] Had I said YAY? No... I should say YYYYYYYYYAAAAAAAAAAAAAYYYYYYYYYYYY!!!! test/02_stack.t already runs code written in sm0p code. 11:27
ruoso test/02_stack.t leaks 82 objects :) 11:30
good to know the lowlevel mem trace works :) 11:31
moritz_ but at least you know it ;)
pugs_svn r19773 | ruoso++ | [smop] marking a new version, once that the sm0p preprocessor seems to be just fine... 11:38
spinclad smop++!! 11:50
hurrah and huzzah!
moritz_
.oO( can I integrate sm0p into the evalbot? )
 11:52
spinclad .oO ( sure tho it may be a fair few releases before it can take any notice ) 11:55
ruoso moritz_, at least until there is some more types implented it would be a little pointless.. 11:58
there's no much thing you can do with int and bool ;) 12:00
spinclad well, with lambda for continuations and bools for true = -> ($x, $y) { $x } and false = -> ($x, $y) { $y }, you can go quite a way 12:04
no i/o though, that could make it hard
ruoso spinclad, I still don't have a Code type 12:19
:) 12:20
pugs_svn r19774 | moritz++ | [evalbot] fixed rakudo execution in 'perl6:' target 13:36
moritz_ perl6: say("hi"); 13:36
exp_evalbot kp6: OUTPUT[hi␤]
..pugs: OUTPUT[hi␤]
..rakudo: No output (you need to produce output to STDOUT)
moritz_ I don't get it 13:37
bbl&
moritz_ perl6: say "hi"; 14:53
exp_evalbot kp6: OUTPUT[hi␤]
..pugs: OUTPUT[hi␤]
..rakudo: OUTPUT[hi␤]
[particle] seems to be working, what was the trick? 14:54
moritz_ to restart the evalbot *after* svn up'ping 14:56
[particle] heh
ruoso I was thinking better about the autoboxing... 17:36
considering $continuation.has_next()
this method should return true or false,
mofino can it punch you if you use it incorrectly?
ruoso that's the point... how should I deal with it? 17:37
mofino I don't know of any autoboxing devices
ruoso in theory it can return something that is not a proper Bool value 17:38
mofino (I have no idea what you are talking about)
:)
ruoso in the lowlevel implementation, when I call $continuation.has_next() it returns a SMOP__Object*
mofino ok
ruoso should I take additional checking steps? 17:39
[particle] does SMOP__Object have a .true() method? 17:40
ruoso SMOP__Object* doesn't have any method... it have only MESSAGE
mofino backs away slowly 17:40
ruoso the methods are implemented inside the message call 17:41
the way the responder interface decides to
[particle] where is SMOP__Object defined?
ruoso include/smop_base.h 17:42
ruoso the thing is how do I make the autoboxing? 17:43
mofino autoboxing?
ruoso my bool $a = $obj.Bool;
mofino ... 17:44
ruoso bool is a native type... native types are the only ones I know the C lowlevel struct for
mofino forces $a to be a bool?
ruoso all others I only know how to call its responder interface
mofino or rather, predefines $a to be a bool?
ok 17:45
ruoso mofino, I'm talking in much lower level
mofino right
ruoso what makes it really complicated is the stackless thing... 17:46
mofino you mean this isn't complicated in itself?
could've fooled me
ruoso because I need to be able to autobox from the high-level object to the native type without recursing 17:47
mofino I see... I think. 17:48
ruoso because the code that runs the interpreter needs to autobox the return of the $continuation.has_next() call to the native bool
so it can decide if it will run the next operation 17:49
ruoso one thing I can do is to force the signature of this low-level methods to return the native bool and fail otherwise 17:51
hmmm
[particle] that's i think what i was getting at with the .true() comment
ruoso hmmm
[particle]++
cj anyone have a link to an article on fail-soft exceptions? 18:43
ruoso :) :) :) :) I've just saw a DESTROYALL being called from within the garbage collector :) :) :) 18:55
using CPS
:) :) :)
TimToady :)++ 18:57
pugs_svn r19775 | lwall++ | [STD] listops are technically terms, not prefixes 19:06
mofino heh 19:36
pugs_svn r19776 | ruoso++ | [smop] Getting better and better... the garbage collection code is already being called using CPS, but it is still segfaulting at some point. 19:58
ruoso actually... it's not the CPS code that is segfaulting... 20:04
it's the code after that...
CPS DESTOYALL is doing just fine...
home & 20:05
cj wp.colliertech.org/cj/?p=175 20:39
lambdabot Title: The Pædantic Programmer 􏿽xBB Blog Archive 􏿽xBB Perl 6 on the CLI via DLR; A summary
cj I'm building .deb packages of the DLR, so you debian folks in the audience should have an easier time of contributing, if you're interested. 20:41
anyway, I've got /work/ work I need to get to... after lunch.
wolverian cj++ # nice 20:59
pugs_svn r19777 | rhr++ | [Unicode.pm] allow writing to substrs at a different unicode level than the substr call 21:44
rakudo_svn r25388 | jonathan++ | [perl6] Fix breakages to attributes. 22:06
rakudo_svn r25389 | particle++ | [rakudo] added perl6doc 22:24
r25389 | particle++ | this is a working partial implementation of perl6doc,
r25389 | particle++ | intended as a beginning for anyone who wants to hack the pod grammar
r25389 | particle++ |
r25389 | particle++ | svn up && cd languages/perl6 && make Makefile && cd src/utils && make perl6doc(.exe)?
moritz_ "make Makefile"? 22:31
either "make" or "make -f Makefile"
pugs_svn r19778 | rhr++ | [Unicode.pm] make samebase use a consistent unicode level 23:01
rhr_ TimToady: can we have :nf<xx> and :enc<xx> regex modifiers to go along with :codes etc.? e.g. token :codes :nf<d> {...} 23:05
moritz_ why not :codes<nfd> and :bytes<utf-8>? 23:06
TimToady :bytes would also need to know normalization
rhr_ bytes needs nf also
moritz_ right
TimToady though I suppose one could make a case for :bytes<nfd utf-8> 23:07
or vice versa
but I don't see a strong argument either way for bundling/unbundling 23:08
rhr_ me either 23:09
TimToady bundling associates it more clearly in people's minds
and I don't know that :enc makes any sense without :bytes 23:10
moritz_ same with :nf<d> without :codes
TimToady (or :bytes)
on the other hand it could be viewed as setting a lexical default 23:11
rhr_ it seems unlikely you'd want a particular nf but not care about codes/bytes
TimToady "whenever I say :codes below I mean :nf<d>"
I dunno, whichever way we do it will be wrong :) 23:12
moritz_ allow all ways? ;-)
in good perl tradition
TimToady well, in perl tradition, anyway...
moritz_ pugs: my @arr = <ab cd>; my $scalar = @arr; say $scalar.perl 23:13
exp_evalbot OUTPUT[["ab", "cd"]␤]
rakudo_svn r25390 | jonathan++ | [rakudo] Attributes should always be named $!foo when actually stored, so implement this. If it's declared $x, make $x an alias to $!x. This should be somewhat more in line with S12.
moritz_ pugs: my @arr = <ab cd>; my $scalar = @arr; say (+$scalar).perl
exp_evalbot OUTPUT[2␤]
rhr_ S02:702 has the example substr($string, 42.as(Bytes), 1.as(ArabicChars)) how can 42.as(Bytes) know to create a StrPos (as opposed to StrLen) object associated with $string? 23:22
rhr_ should there be a substr variant that takes two StrLens? 23:23
rakudo_svn r25391 | jonathan++ | [rakudo] Start to handle nested classes. We don't get the nested namespaces right yet, but at least attributes and methods get associated with the correct class now. 23:25
pugs_svn r19779 | cosimo++ | [t] Test suite refactoring. 23:28
r19779 | cosimo++ | Moved various builtins/io under t/spec/S16-fh.
r19779 | cosimo++ | Fixed pod syntax.
r19779 | cosimo++ | Added S16 smartlinks.
mncharity 23:31
lambdabot mncharity: You have 1 new message. '/msg lambdabot @messages' to read it.
pugs_svn r19780 | cosimo++ | [t] Moved and smartlinked `connect' builtin tests (t/spec/S16-fh)
mncharity TimToady: thanks 23:34
cosimo_ hi! 23:37
mncharity hmm. that different grammar rules might or might not wish the grammar to have been .new()ed... hmm. maybe it's time to make kp6-ruby more true to the p6 prototype-like oo semantics. or maybe that's mission creep for kp6 bootstrap (kp6 does non-instance rules... I don't think it does instance rules(?)). 23:38
but that's a puzzle for later. bbl &
cosimo_: hello :)
(I'm just now disappearing) & 23:39
cosimo_ anyone in channel, I'm refactoring S16 tests. I created also "S16-unfiled". Is that ok for now? 23:41
pugs_svn r19781 | cosimo++ | [t] Moved slurp tests under t/spec/S16-unfiled 23:42
r19781 | cosimo++ | Added smartlink to S16/Unfiled/IO.slurp
r19782 | cosimo++ | [t] Refactored IO.getpeername tests under t/spec/S16-unfiled. 23:45
r19782 | cosimo++ | Added smartlink to S16
cosimo_ ok... got it. :-) 23:48
pugs_svn r19783 | rhr++ | [Unicode.pm] more StrPos/StrLen arithmetic 23:56

Powered by Cro and the Raku® Programming Language.

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

Thank you!