Pugs 6.2.9 released | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | www.geeksunite.net
Set by autrijus on 4 August 2005.
mikael what are the reasons (if any) for yada-yada-yada-ing subs without giving a prototype, as in "sub foo { ... }"? 00:58
svnbot6 r6210 | fglock++ | * perl5/ List - fixed a portability bug 02:01
r6210 | fglock++ | (Windows "Int" stringification)
r6210 | fglock++ | * perl5/ Array - builds ok, passes a couple of tests
xinming anyone here can tell me if what the differences between := and ::= are? 04:50
autrijus xinming: ::= is := at BEGIN time.
i.e. compile time 04:51
see docs/quickref/namespace
bbiab
xinming dev.perl.org/perl6/doc/quickref/namespace page not found. :-S 04:52
autrijus in pugs 04:53
not p6
& 04:54
xinming autrijus: could you please give me the exact url? :-S 04:58
coral docs or doc? 05:21
xinming: start by locating the pugs svn repository 05:22
i don't know the precise url but the two in the topic are good
second, look in docs/quickref/namespace inside that repository
if the word 'pugs' isn't in the url you're probably in the wrong place.
putter svn.perl.org/perl6/pugs/trunk/docs/.../namespace rt.openfoundry.org/Foundry/Project/...e?rev=5944 05:32
&
svnbot6 r6211 | Stevan++ | Perl6::MetaModel - (p5) 06:22
r6211 | Stevan++ | * begun converting pugs tests (t/oo/) to the metamodel
r6211 | Stevan++ | to see how many we can pass
r6211 | Stevan++ | * fixed Perl6::Object::BUILD to ignore named args which
r6211 | Stevan++ | are not attributes
vkon docs/AES/S28draft.pod is very messy 06:46
castaway_ then fix it, dear henry 06:47
vkon sure... 06:51
06:51 castaway_ is now known as castaway 07:00 ods15_ is now known as ods15
GeJ hello gentlemen 07:11
ods15 all your base are belong to us 07:14
castaway feels left out :) 07:15
ods15 puts() writes the string s and a trailing newline to stdout. - bleh, wtf, why a trailing new line :/ 07:18
castaway because its puts .. use a different method :) 07:22
ods15 it's the only function name for this that is this short 07:24
'printf' is very long, 6 chars.. 07:25
castaway awww.. how terrible
go use assembler :)
ods15 castaway: i'm trying out for ioccc 07:28
castaway ah well :)
ods15 anyway i hacked around it, made it so i only call puts when i actually need a newline
but now i need to re-obfuscate this...
Aankhen`` (quoting 5 relevant lines and 50 irrelevant lines before replying)-- 08:00
ods15 you forgot the '$' 08:03
i think
not sure how perl goes
svnbot6 r6212 | Stevan++ | Mini-MetaModel - Bootstrapped and tested in less then 200 lines 08:06
Aankh|Clone sings the "I Hate Powercuts" song.
Mini-MetaModel? :-o 08:07
Hi pdcawley. 08:08
pasteling "Aankhen``" at 61.246.34.89 pasted "`nmake` fails on Win32" (13 lines, 764B) at sial.org/pbot/12405 08:11
08:11 Aankh|Clone is now known as Aankhen``
Aankhen`` Halp! 08:12
autrijus Aankhen``: hm. what did you change? 08:30
does your PATH contain an alternate ld?
08:43 shachaf_ is now known as Shachaf
QtPlatypus Does anyone have any sort of comment on the Lazy IO edge case I pointed out on p6l? 09:06
Aankhen`` autrijus: I didn't change anything... it built fine last night, it's not building today.
castaway it was the pixies, Aankhen`` 09:07
autrijus Aankhen``: weird... did "make clean" help? 09:08
let me boot to win32 to see if I can duplicate this.
a sec.
Aankhen`` I always run `make clean` before syncing up and building. 09:09
(in other words, it didn't help :-) 09:10
autrijus ok. brb. :)
svnbot6 r6213 | iblech++ | * Usual svn props. 09:16
r6213 | iblech++ | * PIL2JS: PIL::PApp, PIL::Cont: Undefine all params after they have been used,
r6213 | iblech++ | to reduce memory usage of the JS engine. Sadly, it only helps 3 MiB...
autrijus nmaking... 09:19
Aankhen``: it worksforme. 09:30
Aankhen``: try "nmake pil"
17:30 < autrijus> Aankhen``: it worksforme. 09:31
17:30 < autrijus> Aankhen``: try "nmake pil"
Aankh|Clone OK. 09:32
Went to "Linking...", then exited. Is that supposed to happen?
autrijus type "pil.exe"
then "say 123" 09:33
Aankh|Clone Whoa, mondo output.
autrijus then ":q"
Aankh|Clone Done.
autrijus so ghc is fine
linker too
somehow the main pugs doesn't like you
Aankh|Clone :-(
autrijus if you just erase Lexer.o and "nmake"
Aankh|Clone sings of unrequited love.
A'right. 09:34
Cool, that worked. 09:37
autrijus hrmph. :)
Aankh|Clone Now to install.
Aankh|Clone crosses his fingers.
My kingdom for a build system that doesn't trigger re-compilation when you try to install Pugs! 09:38
Yay, it worked. 09:39
autrijus++
autrijus :)
if you don't want recompilation, edit config.yml
and turn off the prelude compilation. 09:40
Aankh|Clone The Prelude isn't pre-compiled during the regular build?
(i.e. when you type `nmake`)
autrijus it is, but the dependency detection is faulty
so it retriggers no matter what
Aankh|Clone I see.
autrijus ideally it should only retrigger if Prelude.pm is touched.
fixes... welcome.
Aankh|Clone Do you know which part of the build system handles this? 09:41
autrijus yes, util/build_pugs.pl
Aankh|Clone OK.
Let me see if I can do anything at all...
autrijus but only gaal really groks all of it I think.
09:46 Aankh|Clone is now known as Aankhen``
kgftr|konobi clkao: Brownie points for a version of cvsup that understands SVN/SVK 11:04
svnbot6 r6214 | autrijus++ | * Style cleanup on MiniMetaModel.pl to simplify mental effort needed for Visiolization
r6215 | autrijus++ | * even more clarifications.
r6216 | autrijus++ | * more comments
clkao kgftr|konobi: i have a svkup branch 11:19
castaway waves to kgftr|konobi 11:20
kgftr|konobi castaway: yullo! we should be seeing you real soon!
clkao: url? 11:21
castaway yes indeedy
clkao in the svk repository
castaway is flogging off extra junk on ebay (doing well so far :)
clkao it even works across different mirrors. but needs more polishing 11:22
kgftr|konobi clkao: nice... trying to find something to replace CVSup that's nicer... but we use it not just for doing RCS stuff, but also file based push mirroring 11:23
11:38 prefiks_ is now known as prefiks
fglock hi 13:36
Limbic_Region salutations 13:51
fglock bom dia Limbic_Region
Limbic_Region sure is - my wife's bday today and it is Friday 13:52
fglock congratulations to her. 13:53
i didn't notice it was friday. this was a short week for me
Limbic_Region I don't have a tendency to forget what day of the week it is, but I am often unsure of the date (or even the month for that matter) 13:54
fglock :)
i was thinking about starting documenting the runtime API
castaway wishes Jean a happy b'day 13:55
(And hopes shes well too :)
fglock before we have too many implementations
Limbic_Region mostly castaway, I will relay your well wishes
castaway good good 13:56
fglock is helping create a new PM group in Salvador, Bahia. but I've never been there :( 13:59
coke heo. has it been documented if/how system() is going to change in perl6 yet? 14:00
hey, pdc. 14:05
pdcawley_ hey coke
coke you seen anything on how system() might change in perl6? 14:06
mikael is there ever a reason to yada-yada-yada a sub without giving a prototype, as in "sub foo { ... }"?
(except for documentation purposes) 14:07
castaway who needs prototypes?
fglock mikael: I use it as a "TODO" reminder
mikael fglock: ah, yes, that's a good reason :-) kinda like ``foo = undefined'' in haskell 14:09
pdcawley_ coke: I assume it'll fail on failure and return true on success... 14:10
coke (i know someone who's going to write a wrapper in perl5 for CPAN, just wondering if there was something he could use to make it forward compatible.) 14:11
14:28 autark is now known as jp-autark
stevan autrijus: ping 15:18
svnbot6 r6217 | fglock++ | * perl5/Array - t/array-lazy.t passes all tests 15:59
autrijus stevan: pong 16:03
pugscode.org/images/metamodel.png
ods15 THAT'S IT 16:04
rafb.net/paste/results/TSWb8Y12.html
2048 bytes, ends with a smiley, and had 4 out of 5 algorithms
i ain't touching that shit any friggin more. gonna leave it like that for the next entire year till i submit it...
some of the constants could be spazzed up a little bit... 16:05
as far as i can tell, there are no memleaks/segfaults, but this code DOES cause 'tcc -b' to crash :(
stevan autrijus++ # pretty pictures 16:09
masak autrijus: ::Point? what's that?
stevan masak: it is just a random user class
and instance
$point
kolibrie autrijus: are the blocks an instance, and the cylindar+cone a package? 16:10
stevan kolibrie: I think so,..
masak stevan: ah, thx
autrijus stevan: please check for sanity 16:11
stevan autrijus: I am
autrijus in particular, I'd like to map the greeks to your 10kft notation 16:12
stevan autrijus: how did you like the mini-metamodel ?
autrijus stevan: it helped a lot :)
stevan :)
autrijus except it bootstraps differently :)
stevan yes
autrijus but I figured out
stevan no Class
autrijus at long as the Universe API remains compatible
the Metaverse bootstrap can take various paths.
stevan it actually started out as an experiment in removing Class
it is also single inheritence 16:14
autrijus yeah, which is why I changed @:super to $:super
kolibrie autrijus: can you add the blocks and cylindar+cone to the legend?
autrijus kolibrie: sure
kolibrie thanks
masak any special reason why the 'theos' and 'pneuma' boxes are larger than the boxes corresponding to ::Point? 16:17
dudley I think because point is "merely" a user level class and not part of the Meatmodel 16:18
At least that's how I interpreted it. 16:19
masak dudley: hmm. doesn't seem very fair
don't know if it matters though
kolibrie of course, point has two sets of boxes, user and meta 16:20
masak you mean they sorta have to split the area between them? :)
kolibrie something like that 16:21
autrijus re-uploading. 16:22
done.
kolibrie ah, point's boxes got upgraded! 16:23
autrijus yeah :)
kolibrie autrijus++ # pretty pictures
autrijus I can't figure out how to show the fact that the MetaClass class on the left is magically mapped to the ::Class on the right. 16:24
so I gave up and merely listed them on the same row.
kolibrie well, isn't that what the purple meta line is for? 16:25
autrijus yeah, although it's kind of subtle 16:27
kolibrie worked for me 16:28
masak autrijus++ # much better box sizes :)
kolibrie the left-hand column, with gnosis, metamorph, and genesis, does that show the bootstrap? 16:29
autrijus yes. 16:31
re-re-uploading.
the sequence is (I think): gnosis -> demiurge -> logos -> pneuma -> metamorph -> genesis 16:33
kolibrie tries to follow the path 16:34
autrijus but is really just gnosis (demiurge) -> metamorph -> genesis (theos), and logos/pneuma is parallel
(which is what the path shows)
(metasyntactic names
)++
kolibrie ok, so gnosis is the class for the demiurge instance 16:35
svnbot6 r6218 | Stevan++ | MiniMetaModel - adding the class_precedence_list to the $Class 16:39
r6219 | fglock++ | * perl5/Array - t/array-fetch-store passes all tests
autrijus_ kolibrie: no, gnosis is just instance data. 16:42
I re-re-uploaded.
metamorph adds more things to gnosis 16:43
(via the add_meth special magic)
16:43 autrijus_ is now known as autrijus
autrijus then genesis add a superclass 16:43
pdcawley_ thinks someone's being a wee bit clever clever here... 16:46
autrijus pdcawley: sadly, without those metasyntactic variables, I'd need to write data(meta(class(::Class))) 16:47
and I quickly lost track with that notation :-/
stevan: I hope I'm not overtly insane 16:51
stevan: note that I did away with the class() layer in your notation
stevan yes
autrijus it's hidden under the dotted wall between the two worlds 16:52
as I think it should be (I think)
stevan yes
I agree
that was why I did the mini-metamodel to experiment with that
autrijus nodnod
also I didn't have class's attr and submeth 16:53
mostly due to space restriction
but also because they are not involved during bootstrapping
I think another picture showing the time-indexed bootstrap process would help too 16:55
stevan yes 16:59
fglock autrijus: ping 17:00
autrijus fglock: pong
fglock I was thinking about adding skeletons for other Containers/Values
and maybe document the runtime API
autrijus fglock: that'd rock, go ahead 17:01
you have the javascript version as reference :)
as well as the "Val" type in svn.openfoundry.org/pugs/src/PIL/Val.hs 17:02
fglock iblech said he could reuse the Perl6 modules, compiled to javascript
autrijus not for intrinsic singular value types 17:03
but otherwise yes
fglock translating Perl6 to Perl5 by hand is harder than it seems :) 17:05
autrijus :) 17:06
fglock array contexts, references are entirely different. it would be nice if someone wrote about that. 17:08
autrijus you mean, aside from quickref and synopses
more like a tutorial?
fglock yes, more like a language comparison. A few snippets on a language, and how it looks like in the other. 17:09
autrijus docs/other/porting_howto 17:10
fglock specially about data structures and parameter passing
autrijus is probably what you're looking for
feel free to add to it 17:12
fglock ok :)
does '=$fh' means '$fh.to_list' ? 17:14
(I may have asked this before) 17:16
ods15 pugs is a bad name 17:17
i thought it was a program like 'flea' - to deal with perl bugs...
fglock ods15: I have a Westie - it could be "Westie" :) 17:18
ods15 ?
fglock ods15: actually "pugs" sounds a bit like "pulgas", which is fleas in portuguese :) 17:19
spinclad autrijus: in logos, is |%:attr => $.x $.y| cutnpasted from Point? all the rest makes sense 17:21
autrijus++ # pretty pitchers 17:22
kolibrie autrijus: so gnosis is the instance data of the MetaModel instance, and during bootstrapping the instance data is added to by metamorph and genesis, so at the end of the bootstrap the instance of the MetaModel is also an instance of ::Class
spinclad so gnosis, metamorph, genesis are stages in bootstrapping 17:25
and in the picture that's what they've added
kolibrie spinclad: I think that's right 17:26
autrijus: can you change the legend item for 'instance data' so that the dotted line is coming out of the box, rather than underneath it? 17:27
spinclad seconding, that would be clearer 17:28
PerlJam Is it just me or does anyone else have a hard time following Thomas Sandlass' posts on p6l? 17:34
fglock PerlJam :) 17:35
kolibrie spinclad: hmm, yes, pretty sure the %:attr for pneuma and logos should not be the same
fglock l 17:36
(oops) 17:37
17:41 Aankh|Clone is now known as Aankhen``
spinclad pneuma is the meta(::Object), and logos shows its instance data; it's %:attr in inst(meta(C)) for C in <::Object ::Point> (or something) that shouldn't be the same 17:42
just to be pedantically overbearing 17:43
;) 17:44
kolibrie I think the label 'logos' should be under the boxes, such that theos, pneuma, and logos are all instances 17:50
the instance data squares are not labeled 17:51
spinclad i don't think logos has anything to do with Point, but with Object rather. 18:02
leo suggested adding a line separating Point stuff from the stuff above, which would make it clearer what's the metamodel bootstrap and what's a later example instance/class/meta. 18:05
kolibrie autrijus: can you clarify logos for spinclad and me 18:07
stevan kolibrie: logos is the instance data for the meta(::Object)
the metaclass behind the scences for Object 18:08
kolibrie stevan: so logos and pneuma are both for the instance of meta(::Object)? 18:09
stevan no
kolibrie I'm just seeing the same font, etc
stevan pneuma is the instance of MetaClass
kolibrie and think 'logos' should be under the cubes
or a different font 18:10
and where it is
stevan logos is the instance data that pneuma holds
make sense?
kolibrie and the instance cubes next to Point have no name?
stevan no
that is just defined when you do class Point { .. } 18:11
Object has to be built special
actually you might be able to get away with no building it special, but I have not experimented enough
autrijus: I am going to expand the MiniMetaModel into MetaModel 2.0 18:12
it needs to be re-written, the current version is too cluttered with p5 magic to untangle
kolibrie: the key is to understand the bootstrap process
first is gnosis,.. it is a hand constructed instance 18:13
the 'add_method' method is all you need to begin adding more methods to it
which is the metamorph step
once all the methods and attributes have been added in metamorph, you have a (mostly) complete MetaClass 18:14
the next step is to make ::Object it's superclass
but you need ::Object in order to do that
so you create ::Object, this is pneuma 18:15
it is created from the data in logos
once those two are created, you have ::Object
and you can then add ::Object to the superclass of ::Class
completing the cycle
and then you have a fully formed ::Class which is theos 18:16
the cycle is this:
Class is an instance of Class
Object is an instance of Class
Class is a subclass of Object
autrijus: I think MetaClass is just confusing things... maybe we should get rid of it entirely 18:18
kolibrie so gnosis and logos exist at the beginning, and are just data structures
stevan gnosis does
kolibrie where does logos come from? 18:19
stevan but logos is the input which make pneuma,.. so you can create it later
logos comes from Apoc 12 :)
kolibrie right, so it is a constant set of data needed to set up the initial objet
PerlJam stevan: actually, watching you explain it just now and looking at the metamodel.png it occurs to me that the image should be layered.
currently there are 2 "layers". I think there's really three (at least as how people think of it) 18:20
stevan PerlJam: 3-d VMRL would be nice too ;)
kolibrie: yes, it is (constant set of data to define ::Object)
PerlJam: this diagram is a little tightly packed
an animation might be a good thing 18:21
for the bootstrap process
stevan ponders a set of spork slides
PerlJam stevan: It would :-) but I'm thinking of something like how the X Window System was explained to me. At the bottom is Xlib which talks to the hardware, above that is Xt which talks to Xlib and can talk to the hardware as needed, above that are things like Motif which talk to Xt or Xlib or the hardware as needed. I think the metamodel is the same way. 18:22
kolibrie well, autrijus has promised another picture showing the timeline of the MetaModel bootstrap
stevan yes, it is layered, but it is also temporal
kolibrie: I think a static picture will loose something though 18:23
PerlJam stevan: At the bottom is the metaverse which does the work, above that are things like ::Class and ::Object, and above that is the real "user space"
stevan: Unfortunately you can only show one view at a time. Have one "structural" picture and one "temporal" picture of the same thing.
masak PerlJam: why? have the layers on the vertical axis and time on the horizontal... 18:24
PerlJam masak: because I believe the lines willbe confusing.
masak PerlJam: you might be right
stevan is going to try animating it 18:26
kolibrie stevan: so in metamodel.png, the label 'logos' should be in the same typeface as 'metamorph' and 'genesis'
stevan I am not sure I understand the signifigance of the typeface
kolibrie that's why I go lost in the first place 18:27
all the labels for the instances are in one typeface, and all the labels for the instance_data are in another
except 'gnosis' is different from them all 18:28
PerlJam stevan: animation would definitely be excellent.
svnbot6 r6220 | fglock++ | * perl5/Array passes 'make test'
r6221 | fglock++ | * per5/List - map() bugfix
gaal hi 18:40
fglock hi gaal
gaal how much of Prelude.pm compiles to PIL2 already?
hey fglock :)
integral is there an AST->PIL2 compiler already?
gaal no idea :) 18:41
what with illness and $work i've been a little out of things
in this diagram: pugscode.org/images/metamodel.png 18:42
what's the character before :super in "metamorph"?
is that a @ ?
PerlJam gaal: yes. 18:43
gaal thanks.
PerlJam (at least that's what I've been thinking it is)
fglock integral: I think PIL replaces AST
integral fglock: err? By AST I mean the result of the parser 18:44
and iirc the AST->PIL translation is done by the Compile monad atm
s/monad/typeclass/ 18:45
PerlJam "monad"'s syntactic similarity to "gonad" makes me want to giggle everytime I see it.
gaal should VERSION perhaps be a method (or an attribute) in logos? 18:47
fglock integral: there's no AST in svn.openfoundry.org/pugs/docs/notes/plan - that's why I thought that PIL replaces AST 18:50
autrijus rehi
I updated the picture
gaal heya
autrijus hopefully integrated all suggestions
integral fglock: oh, sorry, AST == Exp
confusion might also have been from PAST being a former name of PIL
autrijus integral: Exp->PIL2 compiler is not yet ready. working on it. 18:51
kolibrie autrijus: ooooh, ahhhh! very nice 18:52
gaal should the textual result be longer or shorter, autrijus?
(of Exp->PIL)
autrijus gaal: I persume longer 18:53
there are more nodes, in any case.
gaal time for binary representation?
autrijus sure, I plan to go with DriFT "deriving Binary"
drawback is that, like Storable, would be haskel only 18:54
YAML is even longer
but somewhat fast to read 18:55
XML is possible too. I'm open for suggestions for ondisk format
gaal hmm, is that necessarily so? DriFT stores native haskell structures?
what about YAML | gzip?
autrijus no, but rather native byte representations
well, why do we worry about ondisk storage again?
autrijus assumes disk is cheap
integral £30 for 80GB 18:56
PerlJam autrijus: Keep things human readable as much as you can.
autrijus PerlJam: YAML may be attractive then
Aankh|Clone -1 for XML.
YAML++
18:56 Aankh|Clone is now known as Aankhen``
autrijus it's still rather verbose but at least somewhat readable. 18:56
gaal it is but pugs -CPugs Prelude.pm is like 1 mb 18:57
autrijus and we can even validate with Kwalify.
gaal: with yaml it'll be like 10mb or more.
yaml+gz is a possibility.
gaal that's too big to work with, isn't it? 18:58
autrijus well, syck is fast
stevan autrijus: I animated the metamodel visiolization
no.perlcabal.org/~stevan/metamodel.gif
autrijus oh wow.
stevan this is the timeline for the bootstrap
it needs annotations
but it is basically there
autrijus you rock 18:59
I fixed the font and legend
Aankhen`` msdn.microsoft.com/library/default....202004.asp
autrijus you may wish to copy the legend over if the rest is too painful
stevan yes, I saw, I will remake
Aankhen`` # When to use XML
stevan autrijus: the whole thing took me about 5 min
autrijus oh ok, remake then
Aankhen``: we are storing tree data, obviously yaml is superior :) 19:00
stevan autrijus: I am wondering if spork slides might be better
autrijus stevan: well, having separate frames available helps too
stevan allow the user to control back and forth
autrijus then we can add voice and subtitle
Aankhen`` I already made my preference for YAML clear. :-P
stevan PowerPoint!!!
autrijus yeah why not, seeing I'm using visio already ;)
autrijus really tried Dia but really can't make it work 19:01
stevan cant either,.. but I use OmniGraffle :)
kolibrie stevan: very nice animation
autrijus powerpoint embeds visio directly.
that's a good thing 19:02
stevan kolibrie: thanks :)
kolibrie svg does animation, but browsers don't support svg that well yet
stevan autrijus: I will work on this over the weekend, complete with annotations
autrijus: did you get my note about MetaModel 2.0?
autrijus stevan: yes 19:04
stevan: I think MMM is a good start 19:05
and I got the basic idea I think
Aankhen`` ?eval my $foo = "init"; sub bar () { UNDO { $foo = "init"; }; $foo = "abcd"; die "HAHAHAHAHA" }; bar()
evalbot6 Error: No compatible subroutine found: "&UNDO"
autrijus so I'll push ahead with the haskell one
till I get roles working
(which is 6.28.0 showstopper)
and we can sync from time to time 19:06
stevan autrijus: I really think roles need to be attributes-less 19:13
butI am not sure it's a battle worth fighting
kolibrie a role is a set of methods, which does not have instance data? otherwise it is like a class? 19:15
stevan kolibrie: no, it has attributes, which will become instance data for the class the role is flattened into 19:16
kolibrie stevan: just trying to picture the difference between a class and a role 19:17
still a little fuzzy for me 19:18
stevan kolibrie: a role is just a collection of methods and attributes 19:23
svnbot6 r6222 | fglock++ | * updated perl5/ TODO list, MANIFEST
stevan to be included into the class
autrijus stevan: why does role-defined instance attrs bad? 19:26
s/does/is/
or do you mean role-defined _class_ attrs? 19:27
or both?
kolibrie stevan: so roles are a definition, and don't have user instances, whereas a class may have user instances 19:31
stevan kolibrie: yes 19:32
autrijus: the problem is with the conflict rules 19:33
if methods are in conflict between two roles, neither is included...this makes it easy to treat role composition an non-order-dependent
autrijus right... so why can't attrs be treated like rw methods? 19:34
stevan if the class does not supply a method of the same name as the conflict,.. then an exception occured
occurs
autrijus: the problem is in the interelation of the methods/attributes
autrijus how is it different from methods/methods?
stevan hmmm
actually it is not very different
autrijus :) 19:35
stevan however, that makes it worse IMO :)
ok, maybe I will introduce the alias idea from the traits paper as a means of disambiguation
fglock autrijus: is 'id()' a container object method? 19:36
spinclad this would be a map from role namespace to class namespace?
stevan I think private methods/attributes can be dangerous too
spinclad: no, the contents of the role namespace is composed into the class namespace
spinclad without renaming then? 19:37
autrijus fglock: a private one maybe
stevan no, the alias would force a rename
autrijus fglock: the container only exposes =:=
which is then implemented by comparing ids
but it does not publicly offer .id
spinclad ok
autrijus fglock: on the other hand, if we expose them as full perl6 objects (which I'm not sure there is a way to do that yet) 19:40
then naturally their .id would be their object id.
i.e. the "Cell" are perl6 objects
but I don't think AES has them exposed yet.
in particular, I'm not sure @array.id returns anything.
kolibrie stevan: light bulb! so a role is like a base class, never used directly, but useful for giving functionality to many classes 19:41
stevan kolibrie: that is one way to look at them yes 19:42
except they can be injected at any point
autrijus and does not participate in ->SUPER calls
stevan yes, the SUPER:: calls in a Role method are late bound,.. so they will call the SUPER:: of the class they are composed into 19:43
autrijus exactly.
kolibrie ok 19:44
TheMaaaa seen iblech 19:54
jabbot TheMaaaa: iblech was seen 21 hours 48 minutes 31 seconds ago
autrijus journal up. see you tomorrow! 20:06
autrijus waves &
nothingmuch hola. 20:21
fglock hola nothingmuch 20:22
shabbath shalom 20:23
nothingmuch not yet shabbath as far as
i'm concerned
i still have friday night which is morally friday =)
fglock (I forgot about the time zone difference) 20:24
I'm still confusing the MetaModel and the Runtime a bit 20:26
nothingmuch hasn't any time to grok that now =( 20:30
ods15 hello nothingmuch 20:31
nothingmuch hi ho 20:32
say - you know mplayer... I have a broken movie, avi/divx.... can I get it to skip the missing frames and try to find the mpeg sync bytes (assuming MPEG-4 has those) and continue from there? I would like to use mencoder to save a movie which might have some stuff missing, but that is still playable 20:33
ods15: ? 20:35
ods15 hmm? 20:36
nothingmuch scroll back one line =)
ods15 oh sorry, was busy hacking :) /me reads Q
if you don't have the index, use main/TOOLS/avi-fix (don't forget '-fix'!)
if you do have the index, just play it back
nothingmuch no index... -forceidx didn't work either. I'll grab the source and see if said avi-fix works 20:37
ods15 regarding avifix, you should probably >/dev/null it's output, it takes more cpu time displaying it than the program to actually work
nothingmuch heh
ods15 avifix always works, but it's HELLA slow
the length of the movie is basically the amount of time it takes
nothingmuch well, i can leave it overnight or something
ods15 yup
nothingmuch even on very mean hardware? I can do it work.. =) 20:38
ods15 after it works you need -idx, i think, but not sure
nah, it's hd bottleneck, not cpu
pc is defenetaly useable while it's working..
nothingmuch ah, in that case I have some very mean hardware for that tooo
6-way sata raid 20:39
nothingmuch benchmarked it to around 300MB (yes, bytes) per second, write speed from /dev/zero
and slightly more than that for reading the file into /dev/null
ods15 lol, hehe
my sata sucks :( pci
nothingmuch it's got an external scsi interface, so it can handle that load =) 20:40
ods15: it's outputting 6 lines and exiting with status 1 20:43
ods15 i dunno. i don't remember much of that prog 20:45
20:48 PerlJam_ is now known as PerlJam 20:51 chip_ is now known as chip
Aankhen`` goes to sleep. 21:06
G'night. 21:07
renormalist Hi all! 21:54
I currently don't quite understand with which kind of regular expression I can work in Pugs. It feels like the half of Apo/Ex/Syn 05, but without the subpatterns and the new stuff? 21:55
is there a description what the current useable regex language is in Pugs? 21:56
I'm also interested in modifiers like :overlap, :exhaustive and the like. 21:57
anyone with a hint awake? 21:58
QtPlatypus renormalist: You can use the grammer parsing capabilities of the regex rules. 21:59
An I think they have just resently got :w to work. 22:00
renormalist Who is working on this, i.e. whose logs do I have to read to see what happens in that rules-area? 22:07
or will it just appear in autijus journal entries?
QtPlatypus Most of the PGE is done on the parrot side of the world.
renormalist also things like <$subrule> , the stuff that feels like Parse:recdescent? 22:08
QtPlatypus <subrule> works 22:59
{ } doesn't work yet.
fglock_ hi 23:13

Powered by Cro and the Raku® Programming Language.

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

Thank you!