pugscode.org/ | nopaste: sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse . show $ foldl1 (*) [1..4]
Set by gaal on 18 June 2007.
00:00 turrepurre joined
Juerd audreyt: Can ubuntu-6.06-dvd-i386.iso be deleted? 00:20
00:26 GeJ joined
diakopter Juerd: if it can be redownloaded and if it's not mounted....? 00:28
00:35 weinigLap_ joined
diakopter invite.pugscode.org is back up... more or less. 00:45
Juerd audreyt: And ~audreyt/public_html/tmp? 00:52
01:01 nipotaway is now known as nipotan 01:08 ibrown joined
ibrown what's up peoples. 01:09
diakopter ibrown: you haven't yet added yourself to AUTHORS? :) 01:16
ibrown nope. 01:17
i haven't technically done anything but 1 smartlink
real work's been kicking my ass.
but i fixed my lame Net::FTP problem today, so i can get to finishing up this db work and start on http.pm 01:19
01:21 daxim joined, dolmans joined 01:23 weinigLap_ is now known as weinigLap
Aankhen`` ibrown: May I ask what your plans are for HTTP.pm? 01:39
ibrown haven't even had a second to think about it :( but i'm game for talks. 01:44
you kno?
Aankhen`` Heh.
ibrown it was suggested last night, that i talk to people in perl6 about it to determine what was going on with it 01:45
b/c it needed to be done, and i said, what do you want me to dev besides smartlinks, i feel i could be used elsewhere too lol
and btw: if those smartlinks are important, i can assign the task to some people i know that aren't that technical but might be able to do it
Aankhen`` Might not be a bad idea to look at CGI.pm, and do everything in the opposite way. <G>
ibrown haha see
my biggest beef
is that perl SUCKS via http and html
Aankhen`` I'm not sure smartlinks are so *important* per se, but they're certainly *useful*. 01:46
Okay, what about Perl sucks?
(For HTTP and HTML.)
ibrown in my experience.
whew.
Aankhen`` Actually, as far as HTML goes, I already know the answer: it doesn't use S-expressions. :-P
ibrown S-expressions?
Aankhen`` It's a Lisp thing. 01:47
ibrown oh.
Aankhen`` Lots of parentheses. ;-)
ibrown lol
that sucks.
ok.
Aankhen`` Anyway, about the "Perl sucks" thing, I was basically just trying to probe to figure out what parts of it seem awkward to you for HTTP and HTML.
ibrown here's the biggest thing about cgi.pm that i never understood.
i'm writing a cgi module, for presenting data on the internet.
why did i always have to do the damn inital header? why couldn't CGI do that for me.? 01:48
yeah writing the content type wasn't hard, and i could see maybe needing to change it from it's default one day
but it'd be easier to set an expression for the minor changes, and have major changes be methods, or routines, whatever, gotta understand, i've been doing the internet thing for 15+yrs, building ISPs, and scripting perl/sh
i'm not to smart other than computers and i went to no uni, so i don't know a lot of technical names for things
Aankhen`` Er, you do know about CGI.pm's `start_html()` routine, right? 01:49
ibrown yeah
but why do that?
Aankhen`` Or did you mean using start_html() vs. having it automatically print the header every time?
ibrown exactly.
Aankhen`` Ah.
ibrown there's lots of things that i don't like where it's not intuitive like that
i hate how you set diff methods for example.
Aankhen`` Well, that's probably something that ought to be in Web::Response for Perl 6.
ibrown i think they're methods, that's why i prefaced.
Aankhen`` (A default header with Content-Type: text/html.) 01:50
ibrown what's Web::Response?
i also hate how you do like the "hashes" within the methods for various things.
like.
for example.
tables.
when you do the
start_table(); method
you put all the attributes in the parens 01:51
Aankhen`` See ext/Web/lib/Web.pm for the stub (and I do mean stub) for Web::Response.
It's the other part of the CGI.pm replacement.
ibrown but it's not very like iono. should i really have to do the start_table(width=>"55",height=>"65") i mean it's not that much of a pain.
but on a different scale isn't there another way to assign those types of hash/key val attributes? 01:52
i'm probably not making much sense cause i am a little high.
but i could go on about cgi and my personal thoughts on it's failures
i use the hell out of it though 01:53
i just do it in a totally stupid way.
Aankhen`` First, you shouldn't be passing width and height anyway. Second, that's one of the downsides of unparsed arguments in Perl 5. In Perl 6 you could say any of: start_table(class => "info"); start_table(:class<info>); start_table(:class("info"));
Third, start_table doesn't strike me as a very good idea. HTML and XML ought to be manipulated as trees, not disjoint strings.
ibrown exactly
HTML is not a language 01:54
its a structure
Aankhen`` Yup.
ibrown like
my fav html module is
html::TreeBuilder i think
and HTML::Element
cause when i started using those for parsing data from wget'd mirror'd pages.
i saw html in a diff light
dduncan fyi, concerning HTTP.pm, I think that was intended to work more at the level of stuff like lwp, and doesn't deal with HTML
ibrown oh i know
dduncan Web.pm is more about HTML
ibrown i wasn't planning on doing anything but 01:55
http protocol specs
Aankhen`` dduncan: Yes, I know. :-0
)
ibrown that was the take i was thinking regarding that
01:55 bsb joined
dduncan I was just replying to an earlier comment that mentioned CGI.pm and HTTP together 01:55
no biggie
Aankhen`` Although Web.pm ought to be more intelligent, Web-specific protocol stuff, I think. An HTML generation module ought to obviously be under HTML::
ibrown here's the deal.
i have what i like to call my "pretty-boys"
they're the guys i call when i want something to look pretty 01:56
dduncan though HTTP is the web protocol
its not like it is IP
ibrown i want a module for html that's essentially my "pretty-boys"
Aankhen`` dduncan: Yes, but Web.pm should incorporate all the little touches like sending a default header of text/html.
ibrown yes.
any module, that interfaces with an http server
should call http.pm
and then they should interface
dduncan well, I think we all agree that however it is broken down, Web.pm is a higher-level interface than HTTP.pm
ibrown for example.
Aankhen`` dduncan: I figure HTTP.pm handles the low-level protocol stuff. Web.pm builds on HTTP.pm to make it easier to get going immediately. 01:57
dduncan I agree
ibrown LWP.pm should call http for all communications with the http server, up until the GET/POST
Aankhen`` So, for example, `use Web;` *could* automatically parse HTTP requests and make the args available to you.
dduncan sounds good to me
Aankhen`` Hmm, actually, parsing it automatically sounds like it might be a security hole. 01:58
I don't know enough about it to tell, though. And you'd have to parse it anyway.
dduncan only if you parse it wrong
that is, ...
Aankhen`` Maybe `use Web 'request'; my $req = request();`.
dduncan the only security holes are using unvalidated data
Aankhen`` Yes.
dduncan so done properly, the peer can send you whatever, and it won't cause security problems 01:59
Aankhen`` ibrown: I'm not sure how that'd work. HTTP.pm handles the low-level representation of data in the protocol, doesn't it?
dduncan: Yeah.
ibrown: That is to say, LWP ought to make the connection to the server, then use HTTP.pm to generate requests and parse responses, as it does currently. 02:00
ibrown yes
aanken
web post web get 02:01
however you call classes and sub-classes ;)
Aankhen`` Basically, Web.pm is all the little niceties that let you go from zero to completed script in no time. It's simply a less verbose way of doing things with HTTP.pm. At least, that's what I thought it would be. Juerd might have had a completely different idea in mind. :-)
dduncan gotta go for a bit ... may check web log
02:01 dduncan left
ibrown why wouldn't http.pm create the socket connections and do all the session setups? 02:04
Aankhen`` It only deals with the data structures involved. 02:05
Maybe it should also open the connections and so on. I don't really know.
ibrown i guess we have to ask ourselves this question.
are we going to have like, a socket.pm.
to generate tcp/udp data transfers?
Aankhen`` I figured as much.
ibrown and then http.pm would know the http rfc compliant speak.
and just refer to socket. 02:06
so then, the need to say, retrieve www.google.com/index.html every 20m.
because you wanna look at the img when it changes.
you'd use http.pm
cause that's just a simple get.
i would like to be like.
Aankhen`` Mmm hmm.
ibrown $index.get('domain'->'remainder of string') 02:07
or however you want to syntax it out
Aankhen`` Not sure I follow.
ibrown $index.get('www.google.com':'/index.html') 02:09
Aankhen`` Ah.
But what is $index there?
ibrown the variable to store the result
Aankhen`` Oh.
ibrown my $index ;)
i don't know perl6 syntax yet
Aankhen`` Heh.
ibrown so i would just be making stuff up as i go
;)
real job's been kicking me hard in the head past few to really get looking at it, i just found the project recently 02:10
Aankhen`` Welcome aboard. 02:11
ibrown tru
its a worthy cause.
i was starting my "Ruby" and "rails" research
stumbled on perl6, said if we can fix how perl does html i won't evar switch
Aankhen`` Hey, if HTML is all that's bothering you, there a ton of other options on CPAN. 02:12
diakopter there's more than one way perl does html
Tene ibrown: if you could write up the kind of things that you would like, and concerns that are important to you, it would be awesome.
ibrown most def tene 02:13
Tene I've done a lot of thinking about Web.pm and friends, but don't quite know where to start yet.
I never really used CGI.pm
Most of my web dev stuff has been mod_perl with Template Toolkit 02:14
So a rather different workflow than Web.pm is for
Aankhen`` From the stubs, it seems that Web.pm is aimed at fitting the mod_perl workflow too. 02:17
I dunno about different template plugins though.
Tene Well, that workflow is "get stuff from apache, do everything in Template Toolkit, then pass a formatted chunk of data back to apache" 02:21
Aankhen`` Right. Doesn't seem like it would be difficult for Web.pm to handle that.
Tene Web.pm doesn't do much that would be relevant to that.
Aankhen`` Yup.
All that it would need to do is grab the relevant information from Apache. 02:22
Tene Hmm. I wonder if anyone has started on a mod_pugs for apache...
m'eh
Aankhen`` I believe there was one, but I doubt it's very up-to-date.
ibrown this might be a stupid basic question. 02:28
but with mod_perl in it's fashion today, can you say, setup all the db handles through there, and manage the DB use of the http server through the mod_perl? if that makes sense? 02:29
Tene ibrown: yes, you can do exactly that 02:30
ibrown that's hot 02:31
i don't do a lot of webdev.
it'll probably show. ;)
but i know communication protocols heh
02:35 justatheory joined, KDr2 joined
KDr2 hi 02:35
Tene hi!
Aankhen`` Hullo. 02:36
02:42 KDr2_ joined 02:44 KDr2 joined 02:46 weinigLap_ joined 02:56 jenkinSear left 03:08 thestarslookdown joined
Aankhen`` What does it mean to export a class method? Is it just being able to leave off the colon in `foo $bar: $baz`? 03:18
Or rather, being able to replace it with a comma. 03:20
03:28 justatheory joined 03:42 daxim_ joined 03:43 weinigLap_ is now known as weinigLap 03:53 sunnavy joined 03:55 justatheory joined 03:56 justatheory joined 04:19 dduncan joined 04:30 amnesiac joined 04:31 monomorph joined
diakopter TimToady: does your language talk have a publicly-available slide deck? 04:43
04:46 diotalevi joined 05:04 audreyt joined 05:08 devogon joined 05:12 vel_ joined
TimToady diakopter: not yet. TheDamian++ was going to try to send me something this weekend since he was going to put one together anyway 05:22
Aankhen``: yes, exporting a method produces a multi that can be called with comma in place of colon. 05:32
It does not change the order of the arguments at all.
so if you want a multi variant that is in a different order that has to be declared explicitly. 05:33
05:48 devogon_ joined
gaal Is a slide deck something you would find on a kindergarten ship? 05:51
above the sandbox deck, and the sleepy time hold... 05:52
afk
TimToady speaking of sleepy time... 05:56
zzz &
Aankhen`` TimToady: Ah, thanks. 05:57
G'night.
Jmax ja, ja, ja, ja 06:01
weisst nicht wie gut ich der bin
oh it's "wir" 06:04
06:26 diotalevi joined 06:38 vel joined 06:47 theorbtwo joined 06:52 Entonian joined 07:09 franck__ joined 07:12 iblechbot joined 07:17 vel joined 07:32 diakopter joined 07:33 bsb left 07:34 jisom joined 08:19 weinigLap joined 08:20 weinigLap joined
moritz_ good morning 08:26
08:26 moritz_ is now known as moritz, Alessio334443 joined 08:27 Alessio334443 left
diakopter moritz: hi 08:31
moritz hi diakopter ;) 08:33
08:34 KDr2 joined, buetow joined
Juerd Aankhen``: That description of Web fits well 08:35
Aankhen``: I was thinking that HTTP::* were just object classes, with no way to automatically create instances, where Web knows how to create those instances 08:36
Aankhen`` Juerd: Good to know I wasn't off-base then.
Right.
Juerd Aankhen``: Web should detect whether mod_perl was used or CGI or FastCGI or whatever, and create the appropriate $request and $response objects to work with that.
Aankhen`` nods.
What about what ibrown said regarding the actual connections and so on?
Does that fall under the purview of an HTTP.pm module? 08:37
Juerd Aankhen``: Haven't read that. Your line matched /Juerd/ :)
Aankhen`` Hehe.
Juerd Could you summarize?
Aankhen`` Er, let's see.
08:38 dduncan left
Aankhen`` I think it boiled down to having a generic Socket.pm module; then HTTP.pm would be a module that specifically spoke the HTTP protocol, and could create HTTP connections and conduct conversations in the protocol. 08:38
moritz do we want to implement full HTTP server/client in HTTP.pm? 08:39
Aankhen`` On the one hand, I thought HTTP.pm was only for representing the data structures involved. On the other hand, it *is* HTTP.pm, so maybe some expectations come with that name. I dunno.
Juerd The Web toolkit needs two things: Web server tools (like CGI), and Web client tools (like LWP). There's a great overlap, but the former is more important IMO.
Aankhen`` Certainly.
moritz Aankhen``: we could have HTTP.pm hold the data structures, and HTTP::{Client,Server} do the actual work 08:40
Juerd The web server tools don't need actual protocol implementation, it can ride on top of CGI etc.
Aankhen`` moritz: That was my initial response too, only not quite so well-phrased.
Juerd moritz: I was thinking of HTTP::Request objects that can be "done"
Consider a CGI script. You have this automatically generated $request object.
daxim_ what is "done"? 08:41
Juerd If you want to re-execute the request, you should be able to do that with not much more than $request.do() or $request.execute()
Aankhen`` daxim_: I guess the HTTP::Request object represents a *potential* request, which can be executed.
Juerd "do" is confusing because of roles.
Aankhen``: Potential and/or current and/or past.
Aankhen`` Right, sorry.
Aankhen`` is a little sleepy.
Juerd Aankhen``: The query string of the CGI script's request should be accessible through $request.query 08:42
Aankhen`` Are POST parameters accessed separately? 08:43
Juerd And form elements through $request.get<foo> and $request.post<foo> (alternatively, because get and post are verbs, $request.query<foo> and $request.body<foo>)
Aankhen`` Ah.
Juerd I sure hope so!
Aankhen`` Hehe.
Hmm, let me see if I understand.
Juerd CGI.pm made the Perl world think weird things about get versus post, when they're combined.
Aankhen`` Oh.
The query string is the raw string, .get and .post are the variablesā€¦ right.
s/variables/parameters/ 08:44
Juerd There could be $request.fields<> that picks .post{$key} // .get{$key}
Yes, but it isn't necessary
Someone complained that .get and .post are verbs and should *do* something. I tend to agree.
But the design has .query and .body of course, the actual strings in which the rest is encoded 08:45
Those strings don't have to be raw. They can be objects that stringify to the raw string, but also allow hash-like access.
Aankhen`` Yeh.
Juerd They could start out as raw strings, and be upgraded by assigning roles to them.
And if at all possible, .body should be lazy 08:46
08:46 franck__ joined
Juerd $request.body does MIME::MultiPart if $request.header<Content-Type> ~~ /multipart/; 08:47
Aankhen`` That's another thing. The stubs you posted have a lot of `does Foo`. Shouldn't that be a runtime thing, e.g. my $args = HTTP::Argument.new() but HTTP::Argument::Upload 08:48
Juerd $request.body<filearg> does Web::FileUpload if $request.body<filearg>.header<Content-Disposition> ~~ /filename/;
Aankhen`` Since otherwise every HTTP::Argument object also does HTTP::Argument::Upload.
Juerd To be honest, I have no idea :) 08:49
Aankhen`` Good to know. ;-)
Juerd So far there's just the synopses, which aren't enough for me to really know the language yet.
Aankhen`` I guess so.
Juerd For example, I have no idea how to actually give an object a hash-like interface 08:50
But I'm determined that it should be used, a lot.
Aankhen`` I used to know, but I've forgotten.
does Hash; # I know this does some of it automatically, just can't remember whether there was more needed
Juerd Because .param(), a method, just isn't Perlish, and makes every module implement its own interface w.r.t. gettinga list of all the keys, etc. 08:51
Aankhen`` Aye.
Juerd While Perl already has wonderful standard interfaces for that
.keys, .values, .kv, .<>
A shame not to use it.
And raw strings that represent structured data are useful too, especially if you can access both. 08:52
Aankhen`` Great, can't find the does Hash bit in the synopses.
Juerd $req.query ~= "&foo=bar";
push $req.query<foo>, "bar";
Where $req.query<foo> is a string/array-hybrid, of course.
Aankhen`` Ah, here we go. 08:54
S13/"Type Casting"
A class may define methods that allow it to respond as if it were a routine, array, or hash. The long forms are as follows:
[...]
Those are a bit unwieldy, so you may also use these short forms:
method %.{ *@@slice } {...}
Juerd barfs
Whoa.
Aankhen`` Hey, use the long version if you like. 08:55
method postcircumfix:<{ }> (*@@slice) {...}
08:55 thoughtpolice joined
Juerd Compare that to Ruby's: def foo[] 08:55
Aankhen`` Does Ruby include multidimensional arrays and hashes?
And what is `foo` there? The class name?
Tene Juerd: the synopsis is referring to defining {}, not foo{}
Juerd No. Personally, I don't care about those at all. I think they make the language overly complex while nested data structures work very well in common circumstances. 08:56
Tene: Still, though, a somewhat nicer syntax might be... somewhat nicer. 08:57
Aankhen`` Personally, I like the notion of multidimensional arrays and hashesā€¦ remains to be seen whether I'll actually use them. :-)
Tene Juerd: method %.{} {...}
Aankhen`` Tene: Er, what's the distinction? I think I missed something. :-S
Juerd method % { ... }
method @ { ... } 08:58
Tene Juerd: ruby puts block parameters inside the block.
Juerd method = { ... }
Tene Perl doesn't.
(Although with placeholder vars, it might...)
Juerd Tene: *@@ really hurts IMO
Aankhen`` Tene: You could use placeholders. :-)
Juerd No matter how logical *@@ may be 08:59
Tene Hmm. I don't remember what the * means there. Aankhen``?
Juerd Tene: Slurp
Tene Ahh, right.
Aankhen`` See, I love Lisp.
(defun $#@! () 1) 09:00
($#@!) => 1
Juerd But there's only one signature that makes sense for .{} -- instead of copy/pasting it around all the time, it could just be the one and only possibility and thus default...
Tene nods.
Although you *could* do multiple dispatch on it
multi method @ (Even $n) {...} 09:01
multi method @ (Odd $n) {...}
Juerd I guess
09:02 BooK joined
Juerd Still that looks better that @.{ Even $n } 09:02
I want Perlish interfaces to be the standard thing. If it's ugly or hard, I'm afraid people will write method param { ... } instead.
Aankhen`` 'S true. 09:03
Juerd admits he doesn't understand slices.
I don't even know how to iterate over a real multidimensional array.
Juerd wouldn't mind at all if slices didn't make it to 6.0.0 09:04
They really feel like a second system symptom...
moritz they feel like the price we have to pay for all the list magic we use elsewhere 09:05
waterbed theory and all
Juerd I question that... 09:06
moritz Juerd: maybe there _is_ a better way around...
Juerd: but we have to find that
Juerd What kind of magic used elsewhere requires a slice data type?
Or requires multidimensional structures?
moritz I'm speaking of magic like autoflatting zip operator 09:07
Juerd That's reversed logic, I think.
Aankhen`` I can't figure out the autoflattening stuff myself.
Juerd I think autoflatting zip is required because we have these slices :)
moritz no, it's not 09:08
Juerd Oh
moritz if you want to be able to do some for (@a Z @b) -> $a, $b...
@a Z @b has to be flattened
lambdabot Maybe you meant: activity activity-full admin all-dicts arr ask . v
moritz and if you "magically" omit the Z, that's still the same 09:09
otherwise you'd have to invent at for (@a Z @b) -> [$a, $b] syntax
which really sucks
Juerd I still don't see what was wrong with the Ā„-era zipping 09:10
Where LIST Ā„ LIST yielded elA1, elB1, elA2, elB2, ... and for loops just took N elements at a time, where N is the number of arguments to its block. 09:11
moritz it's just a workaroudn for the zip problem, not a solution to the generall flattening vs non-flattening problem 09:13
and having two different ops to produce these two cases feels clumsy as well 09:14
09:16 isaacd joined
moritz Aankhen``: re default Content-Type text/html: it encourages people not to specify a charset... 09:19
Aankhen`` moritz: The charset defaults to UTF-8.
moritz Aankhen``: maybe we should also default to UTF-8, because perl 6 has so cool unicode handling?
Aankhen`` (According to the stubs.)
moritz Aankhen``: ah, you were faster ;)
Aankhen`` So Content-Type: text/html; is actually an abbreviated version of it. The actual header would be: Content-Type: text/html; charset=utf-8 09:20
moritz and I think the interface for content-Type and charset should be separate
they are a common http header line, but sometimes you want to modify them separately
Aankhen`` They are separate.
09:21 franck___ joined
Aankhen`` (Separate properties of the response.) 09:21
moritz ok 09:22
Aankhen`` heads for lolsleep.
moritz sleep well ;) 09:24
diakopter Juerd: you around? 09:38
Juerd I am 09:40
diakopter got trac running in mod_python... two little apache config "issues" remain. 09:41
diakopter waits for Juerd to discover the issues 09:42
Juerd Tell me more please :) 09:43
Wow, dev.pugscode.org loads in "just" 1.5 seconds
Hm, is that singular or plural, 1.5? 09:44
diakopter plural, I've always heard
trac returns a Location: :9999/ header upon successful login, and in fact the succesful authenticator is "logged in" if he browses to the :9999/ url 09:45
09:47 polettix joined
diakopter of course, dev.pugscode.org/ merely redirecting to dev.pugscode.org:9999/ would solve that problem. 09:48
lambdabot Title: Pugs - Trac
diakopter but that's definitely not the elegant solution? I mean, we *want* this to run through the port 80 reverse proxy.... right? 09:49
Juerd I don't really know why the proxy was used in the first place 09:50
I haven't installed any of this
diakopter oh, I see the problem 09:51
Juerd I think redirecting is a good solution if that fixes problems. It's less elegant, but I think productivity is more important :)
diakopter default-handler is dav probably...
wait, that's for another problem I haven't mentioned to you yet
09:52 rho joined
diakopter hm; um; hang on 09:55
actually, I don't see why apache (1.3) couldn't be shut off and all of this could be run through 2.2.... 10:01
moritz which parts run through 1.3 atm? 10:02
10:02 franck__ joined
diakopter everything, in a manner of speaking... perlcabal.org, trac, svn, some reverse proxies, some redirected virtual hosts... 10:04
Juerd diakopter: Because I don't have the tuits, and others haven't offered to do it :) 10:05
diakopter I mean, trac and svn are reverse proxied to 2.2...
Juerd There's nothing on feather that couldn't run on 2 AFAIK 10:06
diakopter and invite.p is reverse proxied to jifty:6666 (which needs a runit or daemontools service dir, imho)
10:07 ofer joined 10:37 mico joined 11:15 vel joined 11:18 wolverian joined 11:22 crashmatrix joined
diakopter Juerd: ping 11:22
11:26 franck___ joined 11:27 diakopter joined
Juerd diakopter: pong 11:29
diakopter I think I got everything transferred... 11:30
to 2.2 11:32
11:36 fglock joined
Juerd diakopter: Great 11:38
11:43 nipotan is now known as nipotaway 11:48 agentzh joined
agentzh ibrown: using Test.pm instead of printing TAP yourself will help smartlinks.pl to mark the tests with ōæ½xA1ōæ½xCC and ōæ½xA1ōæ½xC1 using the smoke data :) 11:50
lambdabot agentzh: You have 1 new message. '/msg lambdabot @messages' to read it.
agentzh @messages
lambdabot Aankhen`` said 1d 15h 20m 43s ago: I didn't fix itā€¦ I couldn't even reproduce it. :-S
agentzh Aankhen``: :)
moritz agentzh: I hope it pleases you to learn that I implemented paged results to the irclog search 11:52
agentzh yes, I saw that but forgot to mention. moritz++ 11:53
and i love the new font too
moritz tomorrow I'm moving roughly 2000km, so no time for hacking 11:54
that was Aankhen``++ ;-)
agentzh wow
2000km !?
on foot?
by bike?
moritz agentzh: I'm currently in Edinburgh, Scotland for studying, and I'll move back home to germany this weekend
on sunday actually
agentzh: no, air plane ;) 11:55
agentzh one of my classmates will go to that university for studying soon :)
sadly he is a Java/Python guy ;)
moritz agentzh: which university? there are 4 in edinburgh ;)
agentzh Edinburgh University...it seems 11:56
moritz lucky guy ;) 11:57
agentzh indeed :)
moritz heriot watt just plain sucks
agentzh ah 11:58
moritz but I could buy a master degree there - for 40 Pound ;) 11:59
agentzh heh
it seems that i will only get a B.S. degree in my life ;)
not sure if i'll get a chance to continue my studying. 12:00
moritz :(
why? money? 12:01
agentzh no, just got bored ;)
moritz what are you studying btw? CS? 12:02
diakopter and now commitbit works again, too.
agentzh yup
diakopter silly debian
moritz silly usage of unstable on a "production" server ;) 12:03
12:05 mico joined
agentzh the "feather puzzle" has now finally been worked out? 12:05
moritz agentzh: yes, it was a half damaged RAM component 12:06
agentzh wow, that's very interesting
moritz agentzh: which didn't give wrong results, but slowed down everything
agentzh heh
moritz agentzh: Juerd discovered that is was steaming hot iirc ;)
agentzh that's a wonderful news 12:07
agentzh tries to login to feather again.
moritz commitbit now respons within a few seconds, not within a few dozens of minutes ;) 12:08
agentzh yay!
and i'd expect autosmoking to complete within 1 hour .oO(it used to be 288 min) 12:09
daxim_ ah, so that's why brian d foy thought pugs would need a few days worth of compilation 12:13
agentzh lol
feather is not as fast as my core2duo machine, or 6-min-smoking-cycle will allow per-commit-smoking. 12:15
daxim_ even on my 9 year old pentium3 it always finished under two hours 12:16
agentzh 9 years old? wow. 12:17
agentzh has been considering to throw away his 8 years old machine. 12:18
testgraph.pl failed to convert smoke.yml to HTML in the last autosmoke. 12:20
"Could not convert .yml to testgraph: at util/run-smoke.pl line 94."
diakopter I think you can blame me for that. upgraded YAML::Syck... you might start looking there... 12:21
agentzh but the problem seems to be transient. 12:22
diakopter oh
agentzh diakopter: ah, okay
that explains the problem :)
moritz and I thougth the problem could be _solved_ by upgrading YAML::Syck ;) 12:23
agentzh now calling "util/testgraph.pl smoke.yml smoke.html 12:24
manually works without any problems
diakopter manually as what user with what env 12:27
agentzh i believe at least PATH is almost the same :) 12:28
the perl is the same perl... 12:30
moritz agentzh: try a make realclean
agentzh there has been a long long debate regarding perldoc on p6l...
moritz agentzh: it helps often ;) 12:31
agentzh moritz: will do if i get a second failure cron report... :)
diakopter agentzh: what time did it fail last night
agentzh diakopter: let me see...
diakopter I might have been, ah, brazenly 'killall perl' at the time or something. 12:32
agentzh 20 Jun 2007 22:32:07 +0200
yes, it seems like that...
diakopter or, better, in the middle of upgrading stuff. 12:33
agentzh heh, i'll ignore it for now.
diakopter apt-get upgrade says everything's the latest. in sid.
agentzh feather is now suitable for real-world development. 12:34
feather++
i've just completed the last school assignment today and i'm now merely waiting for graduation. hopefully i'll go back to pugs hacking pretty soon :) 12:37
& 12:44
12:44 agentzh left 12:49 iblechbot joined 13:00 Loro_ joined 13:04 arec joined 13:06 arec joined
arec #login 13:08
13:09 arec joined 13:11 isaacd_ joined 13:12 kanru joined 13:18 mico left, mico joined 13:19 jefflee joined 13:23 cognominal_ joined 13:55 arec joined 14:11 penk joined 14:24 [particle] joined 14:31 ruoso joined 14:50 rfordinal_athome joined 14:54 obvio171 joined 14:56 rfordinal_athome joined 15:17 rindolf joined 15:18 theorbtwo joined 15:43 diakopter is now known as diakopter_afk 15:52 franck__ joined 15:53 barney joined
moritz mooses idly 15:54
TimToady squirrels backwards 16:05
moritz squirrels taste good ;)
statico (dramatic squirrel)++ 16:09
16:20 Entonian joined 16:25 xinming joined 16:29 weinigLap joined 16:38 mico joined 16:51 justatheory joined 16:58 pmurias joined, justatheory joined
pmurias hi 16:59
Tene Hi!
moritz hi Tene ;) 17:00
pmurias moritz: ever eaten squirrels?
Tene hi moritz
moritz pmurias: no, I was just joking ;)
pmurias ;)
17:01 ruoso joined, GabrielVieira joined 17:07 ruoso joined 17:08 justatheory_ joined 17:13 weinigLap joined 17:21 Psyche^ joined 17:26 r0bby joined 17:27 Psyche^ is now known as Patterner 17:30 ashleyb joined 17:32 vel joined 17:34 drupek12 joined
rindolf Hi all. 17:38
moritz hi rindolf ;)
rindolf moritz: ahoy!
moritz: what's up?
17:39 justatheory joined
moritz rindolf: random hacking, and preparing to move back to good ol' Germany 17:41
rindolf moritz: I see.
moritz: where are you located now?
moritz rindolf: Edinburgh, Scotland
rindolf moritz: I see.
17:41 franck___ joined
fglock pmurias: ping 17:52
mncharity hi fglock 17:54
fglock mncharity: how are you doing? 17:55
mncharity if some time you would like to take a pass through the yet_another...whatever it was called, regexp engine, explaining/documenting and cleaning, just let me know. maybe we could find a collaborative editing website to work on the file simultaneously.
ok. still doing doing jobsearch avoidance. mostly banging on olpc. 17:56
Tene 'gobby' is a gtk-based multi-user editor.
fglock mncharity: I've got some new ideas for p6regex-in-p5regex
mncharity looks
Tene cross-platform, connects over the internet, etc. 17:57
it's not vim, though. ;)
[particle] also see writewith.com
fglock mncharity: I'm about to change jobs
moritz fglock: do you already have a new one? 17:59
fglock: or the prospect at least?
fglock yes, I'm taking care of the visa
mncharity gobby.0x539.de/trac/ neat. that could be fun. 18:00
lambdabot Title: obby - Trac
mncharity gobby.0x539.de/trac/wiki/Screenshots
lambdabot Title: Screenshots - obby - Trac
mncharity Tene: nor emacs :) 18:01
fglock mncharity: I'm working out a plan for kp6; I'd like to discuss some ideas on that too
mncharity fglock: I'd like that
re writewith.com, looking... 18:02
fglock cool - I'll start the gobby download and get a coffee 18:03
pmurias fglock: pong 18:07
18:07 buetow joined
mncharity fglock: oh, I'm not sure I can do it today (though I too will start a download). have to head out in ~2 hrs, and lots to do before. um, maybe... 18:07
this weekend may be a no-sleep all-nighter before deadline special. if not, this weekend. else, early next week? 18:08
18:08 jisom joined
fglock mncharity: sure, that's ok 18:08
pmurias: I did a lot of cleaning on the visitors and kp6-perl5 18:09
mismatched accessors are more likely to warn now 18:10
mncharity re writewith.com, looks interesting. though much less realtime than gooby. send email invite? each day to read my email, it takes assorted filters do deal with 10^4 spams. very not realtime. :/
great for business though 18:11
[particle] i haven't tried realtime collaborative editing with writewith, but yeah, great for business 18:12
18:14 vel joined
pmurias fglock: you might want to look at Lexical::Persistense 18:14
fglock I've got gobby 18:15
pmurias: looking
pmurias saw it while reading the Writing a perl REPL atricle
mncharity re gobby, fedora has an rpm :) 18:16
Tene it's in yum, too
pmurias and portage
18:16 REPLeffect joined
mncharity :) 18:17
theorbtwo ...and debian.
fglock pmurias: it looks like an implementation of state variables; I think recursive calls would share the variables 18:18
mncharity could someone not behind a firewall start a test gobby session? 18:26
i'd like to try joining...
fglock Lexical::Persistence is much more complex than Pad.pm, and seems to do less :(
mncharity: audreyt used to start gobby in feather, but I don't know how to do that 18:27
Tene fglock: ssh -Y feather gobby
mncharity notes that you really have to trust feather security to -Y it... 18:28
or not care
Tene I was more presuming the latter. 18:29
(you could always run it in Xnest ;) )
mncharity re Xnest, oh, that's very neat. thanks for the pointer. :) 18:30
fglock: so where is the new job? 18:32
(/me lacks feather account)
Tene But, yes, mncharity is right, -Y is a security issue. -X might be more appropriate. 18:33
fglock mncharity: in Lisbon
Tene And, even then, there are potential security issues.
18:33 isaacd__ joined
Tene refers to 'ForwardX11' in ssh_config(5) 18:34
moritz fglock: quite a distance ;) 18:35
18:37 [particle1 joined
fglock moritz: I enjoy travelling 18:39
mncharity re X, really want a display-free server mode. picture feather costing a continuous full-pugs-tree editing jam :)
18:39 [particle2 joined
mncharity re Lisbon, oo, pretty 18:39
pmurias sobby is the standalone gobby server 18:42
18:45 weinigLap joined
pmurias mncharity: there is now an obby session on port 6552 on feather (getting a newer gobby on my box to test it) 18:47
mncharity re sobby, so the remaining issure would be svn integration. ie, updating an obby session from the repo, and choosing sets of changes to commit. 18:49
re session, oh, great, trying...
ok, so it's 6522 on feather.perl6.nl 18:58
anyone else care to join to play for a moment? 18:59
moritz wait a second, I'll install gobby 19:00
s/a second/five minutes/ # very slow network :(
19:02 dduncan joined
mncharity ;) 19:03
it seems happy with Test.pm
what is some really groddy challenge for editors p6 file? 19:04
pmurias mncharity: damned gnutls reports compile time error here 19:06
19:06 Limbic_Region joined
mncharity :( 19:06
pmurias trying the keyword'ed version 19:07
moritz how do I start a network session?
fglock I get "protocol version mismatch" 19:08
pmurias what's your gobby version
?
moritz ok, I joined, and opened Test.pm
fglock 0.3.0
mncharity recalls something about the protocol version n not being compatible with the version n-1 19:09
moritz 0.4.1-2 works for me
that's the one in debian etch
fglock anyone working on a mac?
pmichaud ?eval my $a = '!'; say ++$a; 19:12
fglock upgrading...
pmichaud ah
[particle2 pmichaud: try run.pugscode.org
unless that's out too 19:13
19:13 [particle2 is now known as [particle]
pmichaud it's running 19:13
19:13 justatheory joined
pmichaud okay, that helped. Thanks. 19:15
mncharity moritz: thanks, that was fun 19:16
moritz mncharity: yw 19:17
mncharity summary: syntax highlighting was fine. couple of second latency on letter hits. letter rather than line based edits.
desire for emacs/vim backends 19:18
pmurias: what did it take to start it sobby on feather?
yipes, later & 19:21
thanks everyone, it was fun! :)
pmurias sobby
a simple command 19:22
19:22 rfordinal_mobile joined 19:24 shay_ joined
pmurias should i run it in the background? 19:24
19:24 shay_ is now known as shay|p6devel
shay|p6devel this sparc64 is ready for perl6 development :) 19:24
pmurias maybe this would be a waste of resources as it can be started on demand 19:25
19:39 diakopter_afk is now known as diakopter 19:41 prism joined 19:49 diakopter is now known as diakopter_afk
Limbic_Region anyone know where I can get a good tutorial on SML for extreme beginners that is free? 19:54
19:55 diakopter_afk is now known as diakopter
diakopter LR: www.cs.cmu.edu/~rwh/smlbook/online.pdf 19:56
or did you mean a different SML? 19:57
19:57 msphix joined 19:58 xinming joined 20:03 rfordinal_mobile joined
Limbic_Region no, that SML diakopter though I am not sure that is geared towards extreme programmers 20:06
I will have a look though
s/programmers/beginners/
moritz at first look it's a bit like OO-Haskell ;) 20:07
Limbic_Region well, the reality of the matter is I want to learn Alice 20:08
Alice is based on SML with lots of extra doodads and whizbangs
unfortunately, Alice doesn't provide a beginner's guide (that I can find) - it assumes knowledge of SML
ibrown what's good people? 20:10
Limbic_Region google beginner guide programming ML 20:12
look at the very first hit
Limbic_Region really laughed
moritz that hurts ;) 20:14
ibrown ugh why do people use benchmark comparisons against programming languages to prove their point? isn't that a little, un-realistic? 20:22
Patterner That's because I use assembly and therefore win every time. 20:24
ibrown lol
humor 20:25
Patterner Well, it's either laugh or cry...
And I prefer to laugh.
ibrown i've always ran away from ASM. except basic concepts heh. 20:26
moritz I use assembler rather often; I don't write it myself though ;)
that's what compilers are there for ;)
ibrown touchee 20:27
20:32 weinigLap_ joined 20:34 weinigLap joined
Aankhen`` 80,6985167So everybody reading this is now condemned to write an OO Perl 6 module, 20:35
80,6985167and document it with Pod 6 ;-)
Argh.
moritz ;)
Aankhen`` Stupid retarded paste mechanism.
Limbic_Region Aankhen`` - I wrote a Perl 6 OO module before pugs supported OO but it is documented with POD 20:36
moritz Aankhen``: time enough to summon an op and make him/her kick you ;)
Aankhen`` Heh.
Let's try that again.
What I was trying to say is:
80,6959317So everybody reading this is now condemned to write an OO Perl 6 module,
80,6959317and document it with Pod 6 ;-)
NO
moritz *LOL*
Aankhen`` WHO TOLD THIS BLOODY PIECE OF JUNK TO REPLACE MY TEXT WITH THE CLIPBOARD CONTENTS
moritz did
Aankhen`` moritz++ # "So everybody reading this is now condemned to write an OO Perl 6 module, and document it with Pod 6 ;-)" 20:37
I win.
Sort of. :-|
moritz Aankhen``: now I got you to quote me three times instead of only once as intended by you ;)
Aankhen`` Hehe. 20:38
moritz maybe I should have said "sentenced" instead of "condemned"
but I couldn't remember that word
Aankhen`` Nah, it's more humourous like this.
Anyway. I think what you said in your mail is extremely true.
Some people appear to be arguing against Pod without even understanding most of it. 20:39
Juerd mncharity: If you want a feather account, mail a request to [email@hidden.address] including your real name and preferred login name. 20:40
20:42 devogon joined
rhr trac is acting funny, dev.pugscode.org/log goes to the wiki, but dev.pugscode.org/anyrandomtext/log goes to the revision log 20:48
lambdabot Title: Pugs - Trac
moritz nice one ;) 20:50
that _might_ have to do with the apache 1.3->2.2 transition
ibrown what's the biggest supported alternative to POD right now? 20:53
pmichaud is there a way to get email notifications of commits to the pugs repo?
moritz I no of none
but there is an RSS feed in trac IIRC
and there is a RSS2Mail perl script somewhere
pmichaud hmmm, the links from www.pugscode.org to RSS feeds don't seem to work 20:54
at least, they don't give me something that looks like an RSS feed.
Oh well, I'll explore later. Thanks. :-)
afk # food 20:55
moritz pmichaud: that's probably related to the trac issue rhr++ just reported
21:05 gene9 joined 21:17 weinigLap joined 21:18 avar joined 21:25 drupek12 joined
ibrown you know. sometimes i'm a friggin moron. 21:44
21:51 buetow joined 22:02 [particle1 joined 22:03 REPLeffect joined 22:15 weinigLap_ joined 22:26 buubot joined 22:33 buubot joined 22:40 octagon joined 22:56 buubot2 joined 22:57 buubot joined 22:59 octagon left 23:03 REPLeffect joined 23:15 xinming_ joined 23:19 prism joined 23:21 REPLeffect joined 23:22 REPLeffect joined, prism joined 23:23 REPLeffect joined 23:31 REPLeffect joined
moritz ?eval <cheap fast right>.pick(2) 23:38
Tene no evalbot :(
moritz ERRNOEVALBOT
23:41 p6evalbot joined
Tene runs an evalbot on his laptop. 23:41
ibrown heh 23:44
Tene ?eval <cheap fast right>.pick(2)
p6evalbot ("fast", "cheap")
ibrown that's weird.
moritz ?eval <cheap fast right>.pick(2)
p6evalbot ("fast", "cheap")
moritz ?eval <cheap fast right>.pick(2) 23:45
p6evalbot ("fast", "cheap")
moritz ?eval <cheap fast right>.pick(2)
p6evalbot ("right", "cheap")
ibrown interesting.
Tene ibrown: what's weird about it?
ibrown the order it printed the words in the first example.
Tene ibrown: what's weird about that?
ibrown well. it poses the question inside my head of A: why'd it pick fast first, then cheap? when fast was the 2nd element, and cheap was first, then i wondered why fast was always before cheap. 23:46
moritz ?eval ('a' .. 'f').pick(*)
p6evalbot ("b", "a", "e", "d", "c", "f")
moritz ?eval ('a' .. 'f').pick(*)
p6evalbot ("f", "c", "b", "e", "d", "a")
moritz ?eval ('a' .. 'f').pick(*)
p6evalbot ("a", "b", "d", "f", "c", "e") 23:47
moritz ibrown: pure coincidence
ibrown ?eval ('a'...'z').pick(2)
p6evalbot Error: ā¤Unexpected "'"ā¤expecting operator or ")"
ibrown oh my bad
Tene too many .s
ibrown ?eval ('a' .. 'z').pick(2)
p6evalbot ("p", "l")
ibrown ?eval ('a' .. 'z').pick(2)
p6evalbot ("c", "e")
ibrown ?eval ('a' .. 'z').pick(2)
p6evalbot ("w", "u")
ibrown awesome
kinda what i had figured.
the reason it was sticking out, was because most likely i guess, the list had 3 vars to choose from heh. 23:48
and coincedence
Tene *nodnod*
ibrown so. question tene.
practical non-web example of using .pick
moritz writing contracts ;) 23:49
<fast cheap right>.pick(2) # didn't I tell you?
ibrown hahaha
Tene ibrown: @cards[(1..52).pick(*)] 23:50
ibrown tru for dealing cards, rolling dice 23:51
Tene Any place where you want a random item from a set
moritz which is the same as 'use List::Util qw(shuffle); @cards = shuffle @cards;' in p5
ibrown i haven't ran across the need to ever do that in a script heh. is that odd? 23:52
moritz has 23:53
ibrown except web ;) figured it might have been one of those things i was like, missing out on
for example: i learned today.... 23:54
that in p5... you can create a scope w/out the requirement of a conditional
or routine.
never knew that.....
moritz most languages allow that
ibrown aye
moritz even c (C99 at least... )
ibrown cause i knew for example c could do that
and i never really thought, i guess abt using that in C. 23:55
so for example: my file slurp would be too many lines... i just needed a scope and local
moritz you can generally assume that perl can do everything that C can ;)
ibrown i guess that's what i get for looking stuff up on the intarnet, and not RTFM so much
i'm no good at condensing my code to fewer lines.
moritz use File::Slurp; # there's a module for every task ;) 23:56
anyway, I'm off to bed ;)
ibrown tru. but local $/ is just as easy... nite mang.

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!