-Ofun: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6 Set by apple-gunkies on 11 November 2005. |
|||
nothingmuch | evening | 01:29 | |
dduncan | evening | 01:56 | |
svnbot6 | r8010 | Darren_Duncan++ | r959@Darren-Duncans-Computer: darrenduncan | 2005-11-25 17:38:38 -0800 | 02:00 | |
r8010 | Darren_Duncan++ | /ext/Rosetta-Incubator : reverted version nums on 4 modules to 0.1.0 from 0.10.0, since they never had any X.YY versions | |||
r8011 | autrijus++ | * transcribe some random thoughts re PIL2 before I | 06:07 | ||
r8011 | autrijus++ | accidentally erase it again ;) | |||
luqui | autrijus, ping | 06:25 | |
autrijus | luqui: pong | 06:26 | |
luqui: glad to see you are getting lambdaminded by the day :) | |||
luqui | yeah. I'm getting quite frightened to be honest. | 06:27 | |
"oh", I think, "I just need an RWS monad" | |||
What? How do I know that?! | |||
anyway... | |||
about the notes you just checked in | |||
autrijus | yes? | ||
luqui | A PIL tree is merely an (unordered!) set of static declarations; | ||
autrijus | (as you can see I'm in a more imperative mood.) | ||
luqui | Why not let the PIL tree be the definitinion of the "IMPORT" function | 06:28 | |
or, rather, a wrapper around that, but with the same purpose | |||
autrijus | I don't quite get it. | 06:29 | |
eval 'say 3' | |||
this "say 3" gets compiled with a IMPORT? | |||
please elaborate | |||
luqui | well, that was just an idle thought. It just feels wrong to let the tree be definitions at the top level | 06:30 | |
instead of something that is executed | |||
autrijus | but the definition is expressed as a set of calls to the underlying runtime engine | ||
see example of "say Hello World" below | |||
luqui | ahh, yes, I hadn't finished reading them | ||
I'll get back to you | |||
autrijus | the unorderness is merely noting that we'd like to go thru one level of indirection for sym lookups | ||
so that the order of executing those definition-statements won't matter | 06:31 | ||
it'd be much easier to express with code :) but sure, do finish reading them | |||
luqui | anyway, I read and grokked "colored local type inference" | 06:32 | |
autrijus | excellent. that will come in handy. | ||
luqui | grokked being the important part. so now I get to design a calculus for perl 6 that is inferrable. | ||
autrijus | good. then you need to somehow specify the annotatedness required for union types. | ||
luqui | probably taking some linear combination of their colored F<: and PIL2 | ||
autrijus | that's a Very Interesting research problem. | ||
(and intersection types.) | 06:33 | ||
luqui | oh, yeah, right, union types (as in Foo|Bar), as opposed to tagged unions | ||
autrijus | yup | ||
luqui | I'm hoping I can finesse them out of the core calculus | ||
that is, still let you write Foo|Bar, but only in a limited set of circumstances | 06:34 | ||
autrijus | right, the quest is to define the limit | ||
luqui | make it a pattern matching problem rather than a typing problem | ||
okay.. well, it'll be fun | |||
er, "fun" | |||
autrijus | may be of use is Semantic Typing in www.cduce.org/papers.html | 06:35 | |
B.Pierce noted we need to grok that if we are to include union/intersection typing. | |||
have "fun" :) | |||
sky__ | *yawn* | 06:42 | |
autrijus | hi sky :) | 06:43 | |
luqui | heh, re: KISS. I've found that when designing ASTs KISS works in reverse | ||
autrijus | mmm? | ||
you want to maximize node types? | |||
luqui | you want to get as few nodes as possible, so you overengineer to do so | 06:44 | |
keeping it simple involves putting in a special case node where one *might* not be needed | |||
autrijus | Keep It Simple, Smart | ||
but yeah, I see your point | 06:45 | ||
sky__ | hi autrijus | 06:47 | |
I am debating wheter I should fly to taiwan over christmas | 06:48 | ||
autrijus | that'd be cool | ||
I'll be back from .au on 14th | |||
hcchien | christmas. :) | 06:51 | |
autrijus | .tw president just yesterday proposes to make christmas a off-work holiday | 06:52 | |
but probably wouldn't make it on time this year ;) | |||
obra | autrijus! hello! | 06:57 | |
autrijus | obra: greetings! | 06:58 | |
autrijus is @ taipei.pm meeting | |||
listening gugod's talk on the much improved, certainly impressive Catalyst 5.57 | |||
hcchien | obra: did you see my flickr? | 07:00 | |
Khisanth | autrijus: am I going nuts or did one of your journal entry disappear? | 07:02 | |
autrijus | Khisanth: you are not going nuts -- I had a minor crisis and accidentally rmed bunch of stuff -- don't ask :/ | ||
geoffb | *yawn* Hey all . . . . | ||
Khisanth | hmm almost 2am | 07:03 | |
my cousin was right, this DOES feel like a sunday :( | |||
dduncan | greetings all | 07:06 | |
autrijus | hey dduncan | 07:07 | |
obra | hcchien: no! where? | 07:08 | |
dduncan | I notice things have been quiet with you lately ... lots of reading? | ||
hcchien | obra: www.flickr.com/photos/hcchien/ | ||
obra | that RT looks a lot like trac | 07:09 | |
hcchien | yup. haha | ||
autrijus | dduncan: yeah, cc2e mostly, and pbp | 07:10 | |
obra | how much of RT is tracced? | ||
autrijus | pbp is like a fractal - more you read into it more you find - not surprising coming from dconway | ||
dduncan | I've read every page of pbp as of a few weeks ago | 07:11 | |
hcchien | obra: about 5 or 6 template pages now | ||
dduncan | there's even a new email list for discussing it now ... | ||
its [email@hidden.address] | |||
obra | hcchien: cool | ||
autrijus | yeah, and automatic refactorers -- I bought 4 copies to give to taipei.pm | 07:12 | |
dduncan | started under a week ago | ||
autrijus | .tw only had 15 available copies and oreilly.com.tw isn't going to translate it :/ | ||
geoffb | I wonder why ...? Seems like it would sell like hotcakes in any language. | ||
autrijus | comparing cc2e and pbp side-by-side is very enlightening | 07:13 | |
geoffb | autrijus, do tell . . . . | ||
autrijus | mostly, the parts they agree tend to be about programming | ||
the parts they disagree tend to be about C++/Java/VB vs Perl | |||
dduncan | does anyone know if there will be another YAPC in Canada any time soon, like the Toronto one? ... 'twould be interesting to meet autrijus in person | ||
autrijus | and by focusing on things they agree on, the respective bias can be avoided | ||
dduncan: next hackathon is YAPC::Israel on Feb | 07:14 | ||
dduncan | I can't afford to travel that soon ... summer at the earliest | ||
geoffb | autrijus, makes sense | ||
autrijus | geoffb: people who read advanced computer books prefers English normally | ||
and "perl books doesn't sell except for cookbook/llama" according to ora.com.tw | |||
geoffb | Kinda makes you want equivalent books from other language groups, like a Prolog version. | ||
autrijus | indeed | 07:15 | |
geoffb | I need to pick up Thinking Forth again, I have the feeling it will have a lot of that "best practices" kind of thing as a side effect. | ||
Hmmm, I thought PBP was doing quite well . . . didn't realize it was still far behind the cookbook. | 07:16 | ||
obra | It's also brand new | 07:17 | |
It has had no time to build up a mythos | 07:18 | ||
hcchien: what are you doing the RTrac for? also, when will it be ready? | |||
hcchien | obra: just for fun. ui rules. some ppl in .tw love trac, but I think if rt got the frondend like trac, it would be more popular. :) | 07:20 | |
obra | cool | ||
lots of people love trac | |||
I've been working on something new to make it easy to build lightweight apps. | |||
We'll be releasing the first dev version on Christmas | 07:21 | ||
hcchien | is it on svn repository? | 07:25 | |
obra | hcchien: not yet. still secret | 07:38 | |
still writing demos ;) | |||
autrijus | movie! movie! | ||
obra | :P | 07:39 | |
code sample? | |||
fsck.com/~jesse/JiftyWiki.tgz | 07:40 | ||
autrijus | movie as in: rawmode.org/~gabbana/movie_title.png | ||
obra | right. no movies yet. | 07:41 | |
certainly not until after the first developer release ;) | |||
sky__ | I HATE MOVIES | ||
they are useless | |||
you can't fucking copy and paste from a movie :) | |||
obra | we'll be lucky if we manage more than two sample apps | ||
sky__ | nor can you skim them | ||
obra | movies are great marketing material | ||
sky__ | or search them | ||
I disagree | |||
good night :) | 07:42 | ||
hcchien | obra: what is the JiftyWiki? | 07:57 | |
a wiki in RT? | 08:00 | ||
obra | no. it's a wiki in Jifty, our new framework | 08:04 | |
that code has everything you need for fastcgi and standalone servers, database setup, etc | 08:14 | ||
hcchien | autrijus just present it. :p | 08:37 | |
autrijus | obra: I presented your NPW slides :D | 08:39 | |
_right_ after gugod's catalyst talk | |||
for maximum dramatic effect | |||
hcchien | obra: so will the rt be rewitten? | 08:40 | |
autrijus | luqui: rehi... I gotta run, be back in ~30min :) | 09:02 | |
dduncan | sky__, who says you can't copy and paste from a movie? it's just a data file like any other | 09:35 | |
gaal | whoa, is that movie actually available for viewing somewhere? | 10:28 | |
nothingmuch | movie? | 10:36 | |
gaal | nothingmuch: you're the star: rawmode.org/~gabbana/movie_title.png (well, your bass-kickin' fingers are) | 10:49 | |
nothingmuch | oh, that one | 10:52 | |
err... =( | |||
merlyn backed out | |||
we're waiting for others to finish audio | |||
r0nny | morning | 10:53 | |
did anyone try to do a pugs plugin for irssi ? | 10:54 | ||
nothingmuch goes to visit his grandma | |||
r0nny | hello | 14:25 | |
are atomic mehtods atomic per instance, or global atomic ? | 14:26 | ||
(and how to check if one of them allready runs) | 14:27 | ||
anyone ? | 14:28 | ||
rafl | Juerd_: Will you be at 22c3? Still need something to sleep? Write [email@hidden.address] and say I already asked for a place to sleep for you. | 15:05 | |
malverian | autrijus: ping | 16:00 | |
DesreveR | yo | 19:48 | |
is there any perl6 thing can search files ? | 19:49 | ||
anyone ? | 19:51 | ||
revdiablo | File::Find | 19:55 | |
I don't know how good it is, or up to date, but it's in Pugs' ext/File-Find | |||
Looks pretty nice | 19:56 | ||
DesreveR | they made it a object | 19:58 | |
revdiablo | Yeah, the interface looks a bit cleaner than File::Find (don't you just love understatement?) | 20:00 | |
DesreveR | i'd like ta use it kinda like map | 20:02 | |
find{ regex }==>sort==>@queue; | |||
revdiablo | Write an alternate interface | 20:05 | |
DesreveR | hmm | ||
revdiablo | Someone would probably give you a commit bit and let you add it to the project | ||
DesreveR | hmm | 20:06 | |
im far from understanding perl6 good enough to do so | |||
well - i will do a sub can be used with array redirection | 20:07 | ||
btw - can i limit the types of a member var in a subclass ? | 20:09 | ||
hmm | 20:10 | ||
20:12
gaal_ is now known as gaal
|
|||
DesreveR | rofl | 20:14 | |
i just realized i can do the insane thing to interpret autofoo files from my build system | 20:15 | ||
20:35
DesreveR is now known as r0nny
|
|||
Aankh|Clone | G'night. | 20:48 | |
Juerd_ | rafl: How soon do I have to decide? I may be in a much better position to do so in a week (re 22c3) | 21:52 | |
r0nny | yo | 22:14 | |
how do i pass a local variable as param to a eval'd function call call ? | |||
dduncan | why is the function call eval'd ... what part is generated code? | 22:17 | |
if no code is generated and you just want to catch exceptions, use a try-block | |||
r0nny | no | 22:18 | |
i got something where i need to generate a function-name at runtime | |||
dduncan | is this function package-qualified or object oriented? | 22:19 | |
r0nny | package-qualified | ||
dduncan | is the package name constant or generated? | ||
r0nny | basically i need to generate a special kind of object depending on a variable | ||
its generated | |||
read from a config file | 22:20 | ||
dduncan | is the function name constant or generated? | ||
what about the arg list ... constant or generated? | |||
fyi, I have a working example in Locale-KeyedText where the package name comes from a config variable, but everything else is constant ... | |||
this example works without 'eval' ... | 22:21 | ||
r0nny | can i look at it ? | ||
dduncan | if anything but the package name is generated, an eval may be needed | ||
its in /ext | |||
er, its in /ext/Rosetta-Incubator | |||
r0nny | where exactly | 22:22 | |
dduncan | the line says: $text = &::($module_name)::get_text_by_key( $msg_key ); | ||
line 229 | |||
the $module_name is a package name from a config var, everything else is constant | 22:23 | ||
r0nny | hmm | 22:24 | |
Khisanth | ?eval 0... ==> @;foo | 22:25 | |
22:25
evalbot_8004 is now known as evalbot_8011
|
|||
evalbot_8011 | Error: unexpected ">" expecting term | 22:25 | |
dduncan | I'll be going outside in a minute, ... | 22:26 | |
r0nny | darn | 22:30 | |
i got a problem loading a package with a generated name | |||
darn | 22:31 | ||
dduncan | I have to go out for a bit, but if you quickly post the code that isn't working (use nopaste if more than 1-3 lines), I may be able to help you | ||
r0nny | mom | 22:32 | |
there are multiple files | |||
dduncan | to put it in perspective, Locale::KeyedText works by invoking modules whose package names are supplied by the user but that otherwise have a known interface | 22:33 | |
it tries to invoke the known function on a var-supplied package name, first loading the package if necessary | 22:34 | ||
the line I showed you assumes package already loaded | |||
r0nny | rafb.net/paste/results/kCoS2w20.html | ||
i think i did something wrong on the call | 22:36 | ||
dduncan | so you have a compile problem in a different file ... unless the one you pasted is Builder.pm | ||
nevermind | 22:37 | ||
r0nny | p pasted both files | ||
found the problem | |||
dduncan | good | ||
r0nny | space betwen function name and op () | ||
dduncan | if I were you, I would break up line 20 into more lines | ||
r0nny | well - i will do now | 22:38 | |
dduncan | eg. slurp into a variable before evaling it | ||
r0nny | im going to use my eyml sub | ||
dduncan | then make sure the variable contains valid perl code by printing out its contents ... eval requires valid perl code as input methinks | 22:39 | |
a yaml file isn't perl code | |||
r0nny | btw - it use just a value from the file | ||
so it genrates the package name | 22:40 | ||
dduncan | also I'll warn you that CATCH isn't working yet ... or wasn't last week, but checking the value of $! like in perl 5 (check $@) does work | ||
r0nny | wont it die if it cant catch ? | ||
dduncan | no, because you have a try block | ||
a try block without a catch is a no-op when exception thrown | 22:41 | ||
see the Locale-Keyedtext code again for what I refer to | |||
it comments out the CATCH solution and uses the other for now | |||
r0nny | hmm | 22:42 | |
dduncan | if the file slurp results in a package name, you usually first have to 'use' it before invoking a function | ||
r0nny | i do allready | 22:43 | |
dduncan | good | ||
anyway, I have to go out now for a bit ... yard work | 22:44 | ||
r0nny | hf | 22:51 | |
23:15
SamB is now known as stupid_2d_gamepl,
stupid_2d_gamepl is now known as sidescrolling,
sidescrolling is now known as SamB
|
|||
svnbot6 | r8012 | fglock++ | * /docs/notes/laziness.txt - "List" vs "Tuple"; | 23:25 | |
r8012 | fglock++ | reorganization of sections. |