Camelia, the Perl 6 bug

IRC log for #mojo, 2010-10-20

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:13 perlrocks Twitter: "RT @kraih: Released #mojolicious 0.999930 (Hot Beverage) with countless new features and bug fixes, enjoy! #perl" --insurgentsoft http://twitter.com/insurge​ntsoft/status/27882549287
00:13 perlrocks Twitter: "Mojolicious interesting. Should I use now" (ja) --shiba_yu36 http://twitter.com/shiba_yu36/status/27882577277
00:45 tempire_1 joined #mojo
00:45 tempire_1 left #mojo
01:37 tempire_1 joined #mojo
01:38 tempire_1 What's easier to understand for beginners?  http://pastie.org/1234344 or http://pastie.org/1234338
01:55 skaurus I have a feeling I asked that before... after moving from daemon_prefork to PSGI and updating mojo from 29 to 30 url /0 doesn't recognized by routes...
01:56 skaurus url /1 is fine
02:33 skaurus dirty fixed it)
02:35 janus left #mojo
02:42 janus joined #mojo
03:00 alnewkirk left #mojo
03:06 alnewkirk joined #mojo
03:34 Alias__ joined #mojo
03:43 perlrocks left #mojo
03:49 yulia joined #mojo
03:57 yulia left #mojo
04:03 tl joined #mojo
04:05 tholen42 left #mojo
04:46 tempire_1 left #mojo
05:03 alnewkirk left #mojo
05:05 alnewkirk joined #mojo
05:18 kvorg left #mojo
05:26 kvorg joined #mojo
05:41 koban joined #mojo
05:57 alnewkirk left #mojo
05:58 alnewkirk joined #mojo
05:59 und3f[A] is now known as und3f
06:30 ltriant left #mojo
06:35 tl left #mojo
06:43 marcus http://bit.ly/a85T26 for the lulz
06:45 tl joined #mojo
06:48 ysyrota joined #mojo
06:51 stephan48 is there a way with submit_button to make it as large as the actual text? http://blog.stejau.de/arti​cles/2010/10/eintrag2.html here the button isnt big enought
07:02 und3f is now known as und3f[A]
07:09 Alias_ left #mojo
07:18 fhelmber_ joined #mojo
07:35 gabiruh_ joined #mojo
07:35 gabiruh left #mojo
08:05 omega can I specify the status code for redirect_to?
08:15 vti DaTa: thanks
08:35 vti DaTa: released
08:37 vti sri: what is a better way to close a connection from a streaming client https://gist.github.com/7d39dc51a1ce5a8f9bb8? And maybe accumulating the body by myself is wrong and $_[0] arg is a content? not clear from the docs
09:02 vti omega: i am afraid not as i see it from the code, but maybe you can set it after you call redirect_to
09:04 spleenjack joined #mojo
09:10 omega ok
09:23 koban left #mojo
09:55 sri moin
09:56 gabb moin
09:57 stephan48 moin
09:59 vti moin
10:01 perlrocks joined #mojo
10:06 perlrocks Twitter: "#Mojolicious 0.999930 http://goo.gl/fb/q3bU9 #mojo #version" --allframeworks http://twitter.com/allfram​eworks/status/27917227336
10:21 skaurus left #mojo
10:21 zakame yay!
10:34 fish_ shouldn't a simple hash in the global scope of a mojo-lite script be available in all requests?
10:35 sri same scope, yes
10:37 perlrocks Twitter: "Mojolicious 0.999930 / Mojolicious / allframeworks.ru - Blog about modern web development and tools http://vk.cc/51KRc" (ru) --xomaa http://twitter.com/xomaa/status/27918853402
10:37 vti sri: have you seen my question and streaming client?
10:37 sri vti: no
10:38 vti sri: what is a better way to close a connection from a streaming client https://gist.github.com/7d39dc51a1ce5a8f9bb8? And maybe accumulating the body by myself is wrong and $_[0] arg is a content? not clear from the docs
10:38 sri thats tricky
10:39 sri $client->ioloop->drop($tx->connection) might work
10:39 sri but i don't see a simple way
10:40 vti hm, but i think that could be a popular use case
10:40 sri patches welcome
10:40 vti and what about the content?
10:40 sri i don't see a good solution though
10:41 sri vti: what do you mean?
10:41 vti what are the arguments passed to the body?
10:41 sri incoming body chunks
10:41 fish_ sri: well, in that case the global %updates seems to lose scope somehow.. the Dumper output sows only the entries for the client which calls the url: http://github.com/discordian​fish/hello/blob/master/hello
10:41 vti $_[0] looks like a content, does it hold *all* received chunks sof far?
10:41 sri no
10:42 sri you are responsible for everything
10:42 vti then what is $_[0] when $_[1] is a current chunk
10:42 sri the invocant most likely
10:42 sri possibly a content object
10:44 sri fish_: wall of text
10:46 sri vti: your biggest problem when reworking that api will be the invocant
10:46 sri the content object can't access the ioloop or transaction, so it has no control over the connection
10:47 vti it is a Mojo::Content::Single object, i just don't understand why it is there and how it can be used
10:47 vti $_[0]->asset->slupr seems to be empty
10:47 sri vti: if you want to clean it up you'll have to read the code
10:48 sri the callback likely lives inside the content object
10:48 crab hi
10:48 * sri waves
10:49 sri vti: the big problem is that to get the body chunks it needs to hook into the content parser
10:50 sri it is right after the chunked filter
10:50 vti sri: ok, that's fine with me, i can concatenate chunks myself
10:50 sri you will understand once you study the code, the problem is vry complicated
10:50 sri *+e
10:51 sri a possible solution could be a double callback
10:51 sri like what we do for websockets in mojolicious
10:52 sri but that would only get you as high as the message layer
10:52 sri still no access to the connection id or ioloop
10:55 sri maybe you can make it work on another layer
10:55 sri websocket transactions have $tx->finish
10:55 crab for production use of mojo i have to use psgi now?
10:55 sri http transactions don't
10:55 vti ye, tried finish and it failed
10:55 crab (sorry, catching up after some time away)\
10:55 sri crab: you *can* use it instead of the old prefork daemon
10:56 sri it is the closest we have atm
10:56 sri until hypnotoad is ready
10:56 sri the old options like fastcgi are of course still there
10:59 crab what is hypnotoad?
10:59 sri vti: the solution could also be very easy
10:59 crab do i care about this if i don't use websockets?
10:59 crab (this == not using prefork daemon)
10:59 sri vti: does $tx->{_state} = 'done'; close the connection from your callback?
11:00 sri hypnotoad is a new prefork server designed for mixing http with websockets and zero downtime software updates
11:00 sri like updating your perl at runtime
11:01 crab cool
11:01 crab until it's done, i can continue to use the old prefork daemon with plain http?
11:01 vti sri: let me check
11:02 sri vti: or possibly $_[0]->{_state} = 'done';
11:03 sri crab: old prefork daemon is gone and unmaintained, but as long as it works i guess you can
11:04 sri behind nginx it is still ok
11:04 sri btw. the prefork daemon is still up for adoption, if anyone has the time to maintain it!
11:05 crab what was the problem with it?
11:05 sri websockets and long polling could kill it
11:05 sri many design flaws
11:06 sri infinite graceful shutdown and stuff
11:06 sri under the wrong circumstances it could run out of workers
11:06 sri and and and
11:08 vti hm, what was that env var that sets chunk size to read?
11:08 sri it works great for plain old http, but every longer lasting http connection can trigger ugly bugs
11:09 sri vti: that constraint got relaxed
11:09 sri MOJO_CHUNK_SIZE does not affect the ioloop anymore for example
11:10 vti well, i get the result too fast, all just in one chunk
11:10 sri why is that bad?
11:11 vti not bad, i just can't test whether my code works :)
11:11 sri see long polling tests then
11:12 vti ok
11:12 sri or just use the chunked transfer encoding
11:14 sri writing chunks with a timer should make it very easy
11:17 koban joined #mojo
11:18 vti sri: ok, $tx->{_state} = 'done'; does the trick
11:18 sri interesting
11:20 MojoGuest817 joined #mojo
11:20 sri now we just need tons of tests
11:20 MojoGuest817 left #mojo
11:20 sri Mojo::Transaction::HTTP::finish might be an option then
11:21 sri also wondering what it does for the server side
11:21 sri so many things to go wrong
11:22 vti wait a sec
11:27 vti i can't get this longpoll thing to work all the way
11:28 vti client disconnects before it gets all chunks
11:28 yakudzo do Mojo::DOM have ability to add xml nodes?
11:30 vti sri: ok, my error
11:31 vti sri: ye, setting to done works
11:32 dotan joined #mojo
11:32 sri yakudzo: limited
11:32 sri with the replace methods
11:33 sri vti: still way too many unknowns for core inclusion
11:33 yakudzo sri, replace inner and add dom node there ??
11:33 sri yakudzo: ye
11:33 sri yakudzo: but more specific methods like ->before and ->after or so would make sense
11:33 sri i would surely accept a patch
11:34 yakudzo (:
11:39 tholen42 joined #mojo
11:40 tl left #mojo
11:52 sri yakudzo: think i'll just add after and before :)
12:05 sri hmm, add_after and add_before might be a smarter choice
12:06 sri before and after might clash with moose
12:08 sri actually, screw moose, i want it to look closer to jquery
12:09 forwardever joined #mojo
12:09 sri it's experimental, lets experiment
12:11 GitHub66 joined #mojo
12:11 GitHub66 mojo: master Sebastian Riedel * 00b30f0 (3 files in 3 dirs): added before and after methods to Mojo::DOM - http://bit.ly/aI5fnn
12:11 GitHub66 left #mojo
12:11 sri yakudzo: there ya go
12:14 GitHub114 joined #mojo
12:14 GitHub114 mojo: master Sebastian Riedel * 1d2c469 (1 files in 1 dirs): fixed typo - http://bit.ly/cz0jpa
12:14 GitHub114 left #mojo
12:20 GitHub173 joined #mojo
12:20 GitHub173 mojo: master Sebastian Riedel * c5ebcca (1 files in 1 dirs): fixed typo - http://bit.ly/bTDG9a
12:20 GitHub173 left #mojo
12:21 spleenjack1 joined #mojo
12:26 spleenjack left #mojo
12:26 MojoGuest377 joined #mojo
12:27 MojoGuest377 sweet:)
12:31 fod joined #mojo
12:35 perlrocks Twitter: "Made working with xml once again a little more fun in #mojolicious. :) http://bit.ly/dAOXmn #perl" --kraih http://twitter.com/kraih/status/27926588587
12:35 fod left #mojo
12:35 fod joined #mojo
12:36 MojoGuest377 left #mojo
12:36 fod left #mojo
12:36 MojoGuest244 joined #mojo
12:36 fod joined #mojo
12:36 MojoGuest244 left #mojo
12:36 spleenjack1 left #mojo
12:38 fod left #mojo
12:38 fod joined #mojo
12:40 avkhozov joined #mojo
12:40 spleenjack joined #mojo
12:40 fod left #mojo
12:41 fod joined #mojo
12:42 fod left #mojo
12:43 fod joined #mojo
12:44 fod left #mojo
12:44 fod joined #mojo
12:47 fod left #mojo
12:47 fod joined #mojo
12:48 fod left #mojo
12:49 fod joined #mojo
12:51 alnewkirk|com joined #mojo
12:51 alnewkirk left #mojo
12:51 fod left #mojo
12:51 alnewkirk|com is now known as alnewkirk
12:51 fod joined #mojo
12:53 fod left #mojo
12:53 fod joined #mojo
12:56 fod left #mojo
12:57 fod joined #mojo
12:58 gabiruh_ is now known as gabiruh
12:58 fod left #mojo
12:59 fod joined #mojo
13:00 fod left #mojo
13:00 fod joined #mojo
13:02 fod left #mojo
13:02 fod joined #mojo
13:03 fod left #mojo
13:35 sri oh, a zombie tv series, thats something new
13:37 sri the walking dead
14:01 tempire_ left #mojo
14:09 meraxes joined #mojo
14:14 hoelzro joined #mojo
14:14 hoelzro left #mojo
14:29 MojoGuest546 joined #mojo
14:31 MojoGuest546 left #mojo
14:31 MojoGuest442 joined #mojo
14:32 MojoGuest451 joined #mojo
15:00 koban left #mojo
15:07 MojoGuest442 left #mojo
15:07 MojoGuest451 left #mojo
15:10 GitHub21 joined #mojo
15:10 GitHub21 mojo: master Sebastian Riedel * ddf7848 (1 files in 1 dirs): improved tests - http://bit.ly/cXUDKn
15:10 GitHub21 left #mojo
15:16 und3f[A] is now known as und3f
15:17 alnewkirk left #mojo
15:17 kvorg left #mojo
15:19 forwardever iterating over array refs or through keys and values in a hash ref in mojo template is really complicated (compared to TT)
15:23 alnewkirk joined #mojo
15:23 sri it's just perl :)
15:25 forwardever I know, but an iterator helper that hides some of the complexity of perl would be great, e.g. in TT it is so simple:
15:25 forwardever http://template-toolkit.org/docs/man​ual/Directives.html#section_FOREACH
15:26 sri i don't see how a helper would make it easier
15:27 forwardever don't know whether it can be done through a helper
15:28 sri how do you think it should look?
15:28 dotan forwardever: you could just use TT...
15:28 sri thats of course true
15:28 dotan I like "it's just perl" - less hidden magic.
15:29 forwardever dotan: I usually use TT, but playing around with mojo templating in a plugin right now
15:29 sri the chances for magic are almost non existing
15:29 dotan TT can sometimes be too smart for our own good, especially when mixed with other clever modules like Class::DBI's iterator.
15:30 sri i guess perl6 would make you more happy
15:30 sri loops there look much nicer
15:34 forwardever e.g. <% foreach my $key ('a','b') {%> throws an exception, but <% foreach my $key ('a','b') {% > works fine: find the difference :)
15:35 forwardever sri: it's { %> vs. {%> , is this something that can be solved?
15:35 sri ummm
15:35 sri {%> is a deprecated tag ending
15:35 sri it will go away in two major releases
15:36 sri {%> used to be the same as begin %>
15:36 sri a relict from before the begin/end times
15:37 sri so yes, it will solve itself over time
15:38 kvorg joined #mojo
15:38 sri <% for my $key (qw/a b/) { %> is not that bad imo, just perl
15:41 dotan it's just perl, except where it's magic block tags ;P
15:41 sri sugestions for non magic alternatives would be more than welcome!
15:42 forwardever it's okay, but if you have array refs saved in array refs, and want to get key/value pairs out of that array, it starts to get ugly
15:42 avkhozov left #mojo
15:42 sri if you don't like perl thats your own little problem ;p
15:43 vti just pass iterators <%= while (my $article = $articles->next) %>
15:43 sri ^^^
15:44 forwardever i like perl, but in templates it's hard to read
15:44 forwardever because of the additional tags
15:44 sri do you have syntax highlighting?
15:44 forwardever yea, all kind of colors on my screen when watching templates :)
15:45 sri which editor?
15:45 forwardever ultra edit
15:45 sri see
15:45 sri so you don't have proper highlighting
15:45 sri wait a minute
15:46 sri forwardever: http://twitpic.com/29y9v5/full
15:47 sri thats how my ep templates look like
15:47 sri embedded perl even has a different background
15:49 forwardever <% iterate my $item => $passed_var %> <% end %> with autodetection of array and hash refs would be cool
15:49 sri not possible
15:51 forwardever probably not, but at least it works in TT somehow (with foreach, not iterate)
15:51 sri tt is a whole different language
15:52 sri i think perl 5.14 will actually make ref handling easier
15:52 forwardever i know
15:52 sri you would benefit from it automatically with ep once it's there ;p
15:53 sri until then you might be better off using tt
15:53 forwardever in most cases, i have to work with much older perls due to hosting providers
15:55 sri but everything is a tradeoff
15:56 sri tag helpers for example are rather unpleasant in TT
15:57 forwardever how does mojo template performance compares with TT performance, mojo should be much faster (because of lack of magic)
15:57 sri it is faster, but thats not important
15:58 forwardever speed of development ?
15:58 sri thats what matters
15:58 sri http://search.cpan.org/~jallwine/Te​mplate-Plugin-HTML-SuperForm-1.0/li​b/Template/Plugin/HTML/SuperForm.pm
15:58 sri this is how proper tag helpers look like in tt
15:59 sri absolutely horrible
15:59 sri no template system is perfect (to my knowledge)
16:02 sri there is always a tradeoff
16:05 und3f Is it possible to easly make asynchronous dns resolve? :|
16:05 und3f with mojo, of course
16:05 forwardever using virtual methods, there is problably a way to get something very similar to tag helpers http://template-toolkit.org/docs/manual/VMethod​s.html#section_Defining_Custom_Virtual_Methods
16:06 sri forwardever: thats pretty much similar to the example i linked to
16:06 sri but you can't easily work with blocks
16:09 sri <% link_to index => begin %><img src="lalala.png" /><% end %>
16:10 sri or a simple <%= form_for add_user => begin %>...
16:12 forwardever in TT its start_form stop_form or something like that, start_link, stop_link ...
16:12 sri right
16:14 sri to me block building feels more natural
16:14 sri matter of taste i guess
16:15 perlrocks Twitter: "Mojolicious and send mail http://ff.im/smVtI" --sharifulin http://twitter.com/sharifulin/status/27945091499
16:18 forwardever the problem that mojo template might have is to much complexity, there are so many tags that it's hard for beginners to understand all this stuff, there is <% and <%= and now even % and %= , than add HTML tags which look very similar because of the < and mix it with perl code, and it get's even tricky for advanced users
16:22 perlrocks Twitter: "Анатолий Шарифулин: Mojolicious and send mail http://ff.im/smWWK" --planetperlru http://twitter.com/planetperlru/status/27945673147
16:25 sri forwardever: i'm not too worried about complexity, it is still simpler than rails templating
16:26 sri would you really rather type form tags yourself?
16:28 sri if you think you can make a more pleasant to use template system, please do and share with us :)
16:28 und3f sri, it is impossible to write udp programs with Mojo::IOLoop?
16:28 sri und3f: i guess
16:29 sri und3f: oh, no we don't do async dns lookups
16:29 und3f and don't have udp support :|
16:29 sri ?
16:30 sri what would be needed to support udp?
16:30 und3f sri, async dns query
16:30 sri ?
16:30 sri 18:29] und3f: and don't have udp support
16:31 sri i mean, what makes you think that
16:31 und3f sri, Mojo::IOLoop - Minimalistic Reactor For TCP Clients And Servers
16:31 und3f sri, thats from pod
16:31 sri ok, that does not rule out udp
16:31 und3f sri, it doesn't have method for udp send/receive
16:32 sri because those are not needed
16:32 sri i'm quite sure everything to support udp is there
16:35 sri at least for the client side
16:37 sri for the server side i guess sockets just need to be flagged for udp and the ->accept call bypassed
16:37 sri which would be a rather small patch
16:38 forwardever not familiar with rails templates (look similar to mojo as far as I can see), think i could live with [% hidden_field( name="foo", value="bar ) %], might not be as perfect as mojo tag helpers, but FOREACH method in TT is a real pleasure
16:38 sri forwardever: use whatever you like most
16:39 sri und3f: if you're trying to implement dns, thats very very very very painful
16:39 und3f sri, i need just client
16:39 sri still
16:39 sri very very very very very painful
16:40 vti und3f: fork, set up socket pair and poll, run dig, write to poll when done
16:40 sri but if you get something working please share :)
16:42 und3f vti, with fork i can use standard gethostbyname but that not very nice decision, as for me
16:42 sri and don't forget ipv6 support
16:43 und3f i wonna go back to AnyEvent :(
16:43 sri a neat little io independent dns library is something i would be very much interested in
16:43 und3f they had everything
16:43 sri aww
16:44 und3f sri, and byte reading was so easyyyy :)
16:44 sri ?
16:45 und3f sri, there are methods in AnyEvent::Socket push_read, witch will run cb when enough data will arrive
16:46 und3f sri, and you just make chains of callbacks
16:46 sri umm, ok
16:49 vti und3f: use a socket connection to a program that will do dns discovery, you can write it in anyevent
16:52 su-bzero joined #mojo
17:13 ysyrota left #mojo
17:15 und3f sri, oh, yes. it is easy to use IOLoop for udp client
17:15 und3f sri, just add some args
17:15 chansen sri: there is always Net::DNS, not sure if qualifies as "little"
17:17 sri Net::DNS would be the sledgehammer :)
17:18 sri i've been tempted a few times to implement only the most minimal feature set required to resolve ipv4 and ipv6 addresses asynchronously
17:18 sri but the specs make me want to poke my eyes out
17:19 su-bzero ;))
17:19 chansen sri: I wouldn't say sledgehammer, but sure read RFC and apply some minimalism using pack/unpack =)
17:20 su-bzero Net::Resolver ;)
17:21 chansen sri: or just for a few worker processes and feed request using a socketpair() ;)
17:22 chansen s/for/fork/
17:22 sri that would be a bit silly :)
17:22 und3f not a bit...
17:23 und3f Thats strange - to write async program and use fork for every resolve...
17:23 chansen sri: I wouldn't say silly, but it depends on the number of resolves
17:23 chansen .. concurrent
17:24 sri but i really wonder how big the most minimalistic recursing resolver would have to be
17:24 chansen isn't even more silly pretending to do async I/O using a blocking resolver?
17:25 sri depends on the use case i guess
17:25 sri but yea, it is
17:26 chansen I can't comment on what it would require to implement a resolver, never done it and just touched the RFC
17:41 stephan48 Would Mojo work as xmlrpc/soap server? or are there "better" tools?
17:41 sri sure would
17:42 sri but you have to implement the protocol yourself
17:42 stephan48 ye
17:42 vti stephan48: i do pingbacks in booty, so it works for me
17:42 sri which in the case of soap just sucks
17:43 stephan48 i am doing a qt app which should contact my app and share some informations over a https connection
17:46 spleenjack left #mojo
17:53 dotan left #mojo
17:57 DaTa wow, this async responses are awesome
17:59 chansen this?
18:01 DaTa these? :)
18:01 DaTa http://github.com/kraih/mojo/blob/​master/lib/Mojo/HelloWorld.pm#L155
18:02 chansen ahh
18:08 Htbaa joined #mojo
18:10 sri hmm, minimalistic resolver might actually not be that big
18:11 sri could be an interesting weekend project
18:14 chansen I don't think it will be big, you just need handle I/O and pack/unpack a few record types
18:14 sri ye
18:15 sri building the first viable request will be the biggest step
18:15 sri from then on it should be fun
18:15 DaTa Mojo::DNS?
18:15 sri naah, maybe a resolver
18:16 sri and i would try to write it in a way to make reuse outside mojo easy
18:16 DaTa Mojo::Resolve::My::Problems
18:17 chansen and a few quirks like misconfigured servers chaining CNAME's
18:19 sri right
18:20 sri but thats the simple part :)
18:21 chansen assuming you know about it and has a recursion limit, I would say yes
18:23 chansen are you doing it by RFC or prior implementation?
18:23 sri still reading up
18:24 sri i'll most likely look for a ruby implementation, since they tend to be lazy enough not implement unneccessary trash :D
18:24 chansen ok
18:24 chansen x2
18:24 sri feel free to join
18:25 sri packet munging is not my specialty
18:26 chansen I would say yes, but I'm deep in project already =(, but if you want a review or feed back I'm here
18:26 sri aww
18:27 sri well, i promise it will be awful :D
18:27 dotan joined #mojo
18:36 dotan left #mojo
18:40 rhaen moin
18:43 sri moin
18:45 sri btw. anyone got a good ipv4/ipv6 address regex?
18:46 sri possibly the perfect regex?
18:48 spleenjack joined #mojo
18:50 chansen sri: I do =)
18:52 chansen you want the obfuscated variant? ;P
18:52 chansen http://nopaste.snit.ch/24631
18:52 chansen just kidding ;P
18:52 chansen sri: http://nopaste.snit.ch/24632
18:52 sri oh pretty
18:53 sri second one looks great
18:53 chansen and is great, because it works as it should ;P
18:55 chansen you should give credit to <Zefram> because he wrote the ABNF
18:55 sri sure if you give me the full name
18:56 sri Andrew Main?
18:56 chansen in yeah, it's mention in URI RFC, I only translated it to regex
18:57 chansen but wouldn't mind a notice ;PP
18:57 sri that regex should go into Regexp::Common
18:58 chansen yeah, the one in http://search.cpan.org/dist/R​egexp-IPv6/lib/Regexp/IPv6.pm is horrible
18:58 sri oh yea
19:00 chansen I have a complete implementation of RFC 3986 in regexp
19:01 Kulag left #mojo
19:01 chansen sri: http://nopaste.snit.ch/24633 if you 3986 compliance in Mojo ;P
19:01 chansen s/you/you want/
19:01 Kulag joined #mojo
19:02 * chansen also got one for JSON ;P
19:03 sri that one i'm actually interested in :D
19:05 chansen I thought you where ;P
19:05 chansen http://nopaste.snit.ch/24634
19:06 GabrielVieira2 is now known as GabrielVieira
19:07 chansen tests is from D. Crockford's own parser JSON_checker
19:07 sri you should put those up on cpan, great reference material
19:08 alnewkirk left #mojo
19:15 chansen someday I might get the tuits ;P
19:18 chansen sri: you do it =)
19:19 kvorg left #mojo
19:20 und3f sri, simple, non-recursive A-record querier http://pastebin.ca/1968332 :)
19:21 chansen copy/paste might work for you but won't work in real use cases
19:22 und3f chansen, that just start
19:22 sri damn
19:23 y0j left #mojo
19:23 sri i've been working on the flags now...
19:23 chansen und3f: ok
19:23 sri of course i only need one version of those...
19:23 * chansen bb in a few minutes
19:24 sri gonna add ->connect(proto => 'udp') btw, to make it easier
19:24 und3f sri, that will be nice
19:24 und3f sri, so you planning to write DNS resolver?
19:25 sri you did most of the actual work it seems
19:25 und3f but it is static
19:26 und3f sri, i'll be  glad if you will improve this )
19:26 und3f seems like we can get nice DNS resolver on mojo today :)
19:27 sri i don't expect too many features, i'll be happy if it resolves a and aaaa records recursively for now
19:27 fhelmber_ left #mojo
19:28 und3f sri, that's all we need too
19:29 und3f sri, i'll start project on github, seems like this is a nice thing
19:31 kvorg joined #mojo
19:34 vti left #mojo
19:48 * chansen back
19:55 sri oh, gotta watch the apple event :)
19:56 und3f DNS resolver on mojo http://github.com/und3f/MojoX-DNS
19:56 und3f with tests :P
20:08 chansen sri: got a live  uri?
20:16 und3f chansen, isn't that copy-pasting anymore? :)
20:29 sri und3f: damn, you're too quick
20:29 chansen und3f: what about Ipv6 or recursive CNAME's?
20:30 sri right, need aaaa support
20:31 chansen und3f: please don't let me discourage you, good start!!!
20:31 sri indeed
20:31 sri und3f++
20:32 sri don't mind chansen, he's always a remorseless perfectionist :)
20:33 chansen I follow RFC's nothing more or less ;P
20:34 sri the inquisition was just following the bible :D
20:34 * sri needs a macbook air
20:34 und3f thanks )
20:34 und3f chansen, ipv6 will add ~tomorrow
20:34 und3f recursion queries - don't know
20:35 und3f chansen, you always can write a patch :)
20:35 chansen und3f++ #  good work so far
20:35 und3f chansen, thanks
20:35 sri it will surely go into the core
20:35 und3f Cool )
20:36 chansen und3f: basically if a CNAME resolves to a CNAME you need to cope with due to misconfigured servers
20:37 chansen just set a global of maximum resoves
20:37 chansen s/resoves/resolves/
20:37 sri once there is recursion
20:42 chansen if you don't support CNAME to CNAME (which is per-factually legal and not supported by RFC (Illegal)) it will unfortunately break servers in the wild
20:48 vti_ joined #mojo
20:50 Htbaa left #mojo
20:55 ltriant joined #mojo
20:57 spleenjack left #mojo
21:21 stephan48 while pulling the changes.... http://abc123.pastebin.com/i6ME2zkr
21:21 stephan48 in boooty
21:28 stephan48 even a reset doesnt work to fix
21:32 forwardever sri: tried to write two helpers to make iterating over hash and array refs easier in templates, maybe you can take a look at it, while api looks good, not so sure about the code: http://pastie.org/1236618 as data is saved on the controller object
21:33 sri busy with async resolver atm
21:33 forwardever okay, will ask later again :)
21:36 forwardever vti: build iterator, as you suggested, and copied from your bootilicious iterator, maybe you can also give some feedback: http://pastie.org/1236618
22:05 und3f good night
22:10 und3f is now known as und3f[A]
22:47 su-bzero_ joined #mojo
22:52 su-bzero left #mojo
23:08 fish_ http://dev.xantus.org/ broken? it says only "disconnected"
23:37 forwardever left #mojo

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary