Camelia, the Perl 6 bug

IRC log for #mojo, 2010-10-26

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

All times shown according to UTC.

Time Nick Message
00:06 spleenjack left #mojo
00:12 Akron Good night!
00:13 Akron left #mojo
01:34 alnewkirk left #mojo
01:37 alnewkirk joined #mojo
01:46 ltriant left #mojo
02:04 ltriant joined #mojo
02:28 GitHub12 joined #mojo
02:28 GitHub12 mojo: master Sebastian Riedel * 62b75aa (1 files in 1 dirs): cleanup - http://bit.ly/9yXyzw
02:28 GitHub12 left #mojo
02:29 sri think i'll push a release later today
02:48 xantus \o/
02:54 xantus what version will that be
02:55 xantus 0.999931?
03:02 xantus matches version history, so I'm adding my prereq for that
03:40 MojoGuest67 joined #mojo
03:43 MojoGuest67 left #mojo
04:03 tholen42 joined #mojo
04:05 tholen left #mojo
04:31 GitHub66 joined #mojo
04:31 GitHub66 mojo: master Sebastian Riedel * e7532f0 (1 files in 1 dirs): cleanup - http://bit.ly/cUlIVF
04:31 GitHub66 left #mojo
04:32 sri release is tagged, uploading :)
04:47 perlrocks Twitter: "Released # mojolicious 0.999931 with the new non-blocking DNS resolver, enjoy. :) # Perl" (es) --kraih http://twitter.com/kraih/status/28760726075
04:58 GitHub98 joined #mojo
04:58 GitHub98 mojo: master Sebastian Riedel * fa178aa (3 files in 3 dirs): fixed typos - http://bit.ly/cMvuvd
04:58 GitHub98 left #mojo
05:01 sri hope we'll be 100% green again with the new release :)
05:01 sri http://www.cpantesters.org/distro​/M/Mojolicious.html#Mojolicious-0.999930
05:16 ysyrota joined #mojo
05:46 kvorg left #mojo
05:52 koban joined #mojo
06:02 ltriant left #mojo
06:07 kvorg joined #mojo
06:20 achromic left #mojo
06:21 achromic joined #mojo
06:28 perlrocks Twitter: "RT @ kraih: Released # mojolicious 0.999931 with the new non-blocking DNS resolver, enjoy. :) # Perl" (es) --nerform http://twitter.com/nerform/status/28765754812
06:51 perlrocks Twitter: "RT @ kraih: Released mojolicious # 0.999931 with the new non-blocking DNS resolver, enjoy. :) # Perl" (de) --SaschaDroste http://twitter.com/SaschaDroste/status/28766752842
07:06 arthas joined #mojo
07:30 omega beginning to like Test::Mojo::JSON now
07:42 und3f[A] is now known as und3f
07:51 perlrocks Twitter: "RT @ # mojolicious 0.999931 kraih Released with the new non-blocking DNS resolver, enjoy. :) # Perl" (es) --xomaa http://twitter.com/xomaa/status/28769331676
07:54 pid joined #mojo
07:59 und3f is now known as und3f[A]
08:21 tab joined #mojo
08:23 tf joined #mojo
08:23 tf left #mojo
08:23 dotan joined #mojo
08:53 vti_ sri: i guess dns resolver is a bit slow when there is no internet connection
09:08 omega can I specify nested routes in ::Lite easily?
09:09 omega I presume under is what I want
09:13 marcus omega: yes
09:13 marcus if you want auth, or something, you want under
09:14 yakudzo http://yoshinorimatsunobu.blogspot.com/20​10/10/using-mysql-as-nosql-story-for.html
09:14 yakudzo Wow
09:18 kvorg left #mojo
09:41 spleenjack joined #mojo
10:11 spleenjack left #mojo
10:17 omega are under's "inherited?
10:17 omega hms, no
10:28 forwardever joined #mojo
10:37 marcus omega: inherited?
10:37 marcus can you inherit lite files?
10:37 omega no, not really inherited
10:37 omega more like "extended" or "nested"
10:37 omega but I found a but ugly solution :)
10:37 marcus yay
10:38 marcus maybe you don't want a lite app anymore
10:38 marcus then you can use a bridge
10:38 omega code-refs ftw!
10:39 omega also the myapp routes is handy, but is it easy to make it output the method as well?
10:39 omega like I have specified get '/', post '/', they both look the same in the output
10:46 omega more test-mojo-json fixes pushed, yay :)
10:49 spleenjack joined #mojo
10:54 omega yay, all tests pass :p
11:00 tab left #mojo
11:32 kvorg joined #mojo
11:33 sri vti_: should be faster in the release
11:52 MojoGuest626 joined #mojo
11:52 MojoGuest626 From: http://www.google.co.in/url?sa=t&source=​web&cd=3&sqi=2&ved=0CCEQFjAC&amp​;url=http%3A%2F%2Fdev.xantus.org%2F&rct=​j&q=websocket%20demo&ei=NMDGTNbPKYHR​cd_B-IkO&usg=AFQjCNGw6nZv0wO42uzUnHhWcD0​vcKNULg&sig2=b2AyALgd4DU_C0fbCQ0DVw (1 hits)
11:52 MojoGuest626 test
11:52 sri fail
11:58 MojoGuest626 left #mojo
12:31 su-bzero joined #mojo
12:38 su-bzero left #mojo
12:44 MojoGuest805 joined #mojo
12:44 MojoGuest805 /help
12:44 dotan sri: except for websockets, what's "html5" about mojolicious?
12:45 dotan or is html5 just the current term for new hawtness, like web2.0 used to be?
12:45 MojoGuest805 left #mojo
12:52 dotan I have a feeling I already asked that question... let me google it for myself.
12:55 dotan oh yeah: http://irclog.perlgeek.de/​mojo/2010-07-14#i_2552764
13:21 alnewkirk left #mojo
13:28 Akron joined #mojo
13:46 marcus .co.in is pretty awesome
13:52 MojoGuest961 joined #mojo
13:53 MojoGuest961 is now known as hallo
13:53 hallo left #mojo
14:11 und3f[A] is now known as und3f
14:25 forwardever vti_: you did it :)
14:28 vti_ forwardever: ye :)
14:29 vti_ now time for some cleanup
14:29 vti_ :)
14:29 forwardever okay, go ahead, I will review and comment :)
14:39 MojoGuest997 joined #mojo
14:39 MojoGuest997 From: http://news.ycombinator.com/item?id=1277067 (20 hits)
14:41 MojoGuest970 joined #mojo
14:41 MojoGuest970 From: http://www.google.de/search?hl=de&sa​fe=off&q=websocket+demo&aq=f&amp​;aqi=g10&aql=&oq=&gs_rfai= (1 hits)
14:41 MojoGuest970 hello
14:43 MojoGuest970 left #mojo
14:43 GabrielVieira joined #mojo
14:45 MojoGuest997 left #mojo
14:48 fhelmber_ joined #mojo
14:54 vti_ sri: since app has a helper method, maybe m::lite should have it too?
15:07 su-bzero joined #mojo
15:10 alnewkirk joined #mojo
15:11 vti_ sri: never mind
15:13 koban left #mojo
15:23 su-bzero left #mojo
15:38 GabrielVieira left #mojo
15:53 kvorg left #mojo
15:54 su-bzero_ joined #mojo
16:08 spleenjack left #mojo
16:11 spleenjack joined #mojo
16:19 sri ;p
16:24 marcus /win 6
16:27 sri /fail 6
16:28 su-bzero_ left #mojo
16:28 marcus it all evens out in the end
16:29 vti_ sri: can i pass static_class as i pass template_class?
16:29 Akron left #mojo
16:29 sri vti_: could be
16:31 GabrielVieira joined #mojo
16:34 pid left #mojo
16:36 forwardever sri: updated ResourcefulRoutes plugin today, so it is prepared to support different orms (some works left) as well as different template engines, however, not sure about the namespace, right now, I have located classes under MojoX::Scaffold (MojoX::Scaffold::Template, MojoX::Scaffold::Model), does that make sense?
16:37 sri sounds like your whole project needs a new name
16:38 forwardever Scaffold?
16:38 sri maybe something less generic
16:38 forwardever like...
16:39 sri SimpleCRUD or so
16:42 forwardever it's all about scaffolding (templates, models, validation ...), SimpleCRUD sounds like a plugin for an ORM
16:51 armicron joined #mojo
16:53 vti_ sri: how about adding MyApp::Plugin namespace to plugin namespaces by default?
16:54 vti_ sri: look what i get on 'Bad signed cookie "mojolicious", possible hacking attempt.', looks like a dispatching loop https://gist.github.com/3f21ffa0e4dd52f27b26
16:54 mudpit joined #mojo
16:56 mudpit Plack distribution has a nice benchmark comparing the various psgi servers, but Mojo::Server::PSGI doesn't appear to work with it
16:57 dotan left #mojo
16:59 vti_ sri: dispatching loop was my fault :)
17:00 forwardever mudpit: as far as I understand it, mojo::server::psgi allows mojo apps to run on psgi servers, but it's not a real psgi server that allows all kind of psgi apps to run on mojo
17:01 mudpit oh
17:02 forwardever according to the docs: Mojo::Server::PSGI allows Mojo applications to run on all PSGI compatible servers
17:02 mudpit so it' Mojo::Server::Daemon that's the full-fledged server
17:02 mudpit anyway to include that in the benchmarks?
17:03 forwardever mudpit: i think sri is working on a new prefork server that can be used in production, called hypnotoad, not finished so far
17:04 forwardever daemon is rather a development/test server
17:04 mudpit ok
17:05 forwardever you might use it for smaller projects, but probably not recommended, so benchmark would not make sense
17:05 mudpit i know benchmarks are a taboo subject in the mojo camp, but i'm curious how it would stack up to Feersum and Starman when it's finished
17:05 forwardever mudpit: where can i see the benchmarks ?
17:07 mudpit http://cpansearch.perl.org/src/MIYA​GAWA/Plack-0.9951/benchmarks/ab.pl
17:08 mudpit http://pastie.org/1250548
17:09 forwardever interesting
17:11 forwardever starman in combination with mojolicious seems to be a good choice right now
17:13 forwardever would be interesting to see such benchmarks using a mojo app (Mojo::Server::PSGI) with various numbers of routes defined
17:15 mudpit yup
17:16 mudpit left #mojo
17:25 sri i will never support such stupid benchmarks
17:25 sri ever
17:25 sri they are absolutely meaningless
17:26 sri if you choose a web server based on some micro benchmark result you deserve what you get
17:26 sri everybody can optimize a web server for ab
17:26 sri in the end what you will get is examples/microhttpd.pl
17:27 sri i've added that example to illustrate the stupidity of it :)
17:27 kvorg joined #mojo
17:28 sri http://github.com/kraih/mojo/blo​b/master/examples/microhttpd.pl
17:30 forwardever sri: not an expert, but benchmarks on running full fledged mojo app on different servers would be interesting, no??
17:31 sri what would you benchmark?
17:31 sri and why?
17:31 forwardever speed of dispatching using 10, 100, 1000 routes
17:31 sri mojolicious is much slower than any half decent web server
17:32 sri so you benchmark mojolicious against mojolicious
17:32 sri no point
17:33 sri parsing http is never the bottleneck in a mojolicious app
17:34 sri if you have a database backend, that will always be your bottleneck
17:35 forwardever okay okay, but just to fully understand it: is mojolicious performance the same on all servers like starman...?
17:35 sri pretty much yes
17:37 sri (for half decent implementations)
17:44 forwardever another question: how big is the influence of the number of routes on mojolicious performance, is it noticeable?
17:44 sri of course
17:46 sri everything costs
18:02 forwardever in rails docs, i have read wild controller routes (like :controller/:action are not recommended for RESTful applications as all actions in every controller get accessible via GET requests, however, this seems to be the only way to keep number of routes low, is there a way to prevent methods from beeing called (e.g. no dispatch to _private_methods ?
18:11 forwardever so I should probably rethink my routing strategy and make stronger use of regular expressions, as with my routes plugin, the number of routes is growing rather quickly...
18:11 sri no private methods
18:12 sri convenience always comes with a cost
18:13 sri but thats fine as long as it allows you to get stuff done faster
18:13 sri in the end your only limited resource is time
18:13 sri everything else can be bought
18:14 Htbaa joined #mojo
18:15 forwardever just think there might be a smarter way to generate routes using resourceful_routes plugin, will have to think about it...
18:18 ysyrota left #mojo
18:22 sri http://techcrunch.com/2010/10/26/wepay-ice-paypal/ # wow
18:22 sri best marketing stunt ever
18:22 forwardever sri: from a performance point of view, would it be better to create a $r->waypoint('/users') instead of 7 routes that match /users ?
18:23 sri possibly
18:23 spleenjack left #mojo
18:25 stephen joined #mojo
18:27 forwardever okay, will use waypoint, hoping that it performs better (... and do benchmarks later :)
18:37 jfuller joined #mojo
18:50 forwardever sri: using methods like via and name on $w after my $w = $r->waypoint("/$path")->to(controller => $ctrl, action => "index"); doesn't seem to work, is this the expected behaviour?
18:59 Htbaa1 joined #mojo
19:00 marcus heh
19:00 marcus please d't optimize your routes until it is an actual problem
19:00 marcus that is fucking silly
19:00 marcus there is soo much more to be gained by optimizing your model and setting up a proper proxy server
19:00 marcus like varnish
19:02 forwardever it should have taken only a couple of minutes, however waypoints do not work as expected
19:02 Htbaa left #mojo
19:03 * marcus starts setting up gitorious for nordaaker
19:03 MojoGuest398 joined #mojo
19:03 MojoGuest398 From: http://news.ycombinator.com/item?id=1277067 (21 hits)
19:11 MojoGuest949 joined #mojo
19:11 MojoGuest949 From: http://blog.kraih.com/ (4 hits)
19:11 MojoGuest949 /join
19:12 MojoGuest949 left #mojo
19:22 robinsmidsrod left #mojo
19:25 sri wise words from marcus
19:25 robinsmidsrod joined #mojo
19:25 sri performance is not a problem until it is a problem
19:26 forwardever sri: nonetheless, waypoints do not work correctly (I think)
19:27 sri quite sure they work as they should
19:28 forwardever as mentioned earlier, it's not possible to apply name and via method on $w, $w = $r->waypoint("/path")->to...
19:28 sri of course it is
19:28 sri you might be misunderstanding something
19:29 forwardever maybe
19:29 forwardever but what ?
19:30 sri step back, grab a beer, and take your time to read up on it
19:31 sri don't rush it, learn routes inside out first
19:31 kvorg left #mojo
19:31 forwardever at least I couldn't find any tests in that regard (applying name and via on $w)
19:36 spleenjack joined #mojo
19:37 dotan joined #mojo
19:39 MojoGuest398 is now known as jambalaya
19:44 jambalaya left #mojo
19:47 dotan $app->types->type(vti => 'help/vampire') ?
20:01 sri !
20:08 marcus poor vti.
20:08 marcus wow, installing java stuff on freebsd is a nightmare
20:08 marcus 3 different web pages where I had to click accept on the license and download stuff and copy in to the server so far
20:10 marcus for bonus points, the first two expected me to stuff the file in /usr/ports/distfiles, and the 3. one expected it in /tmp
20:13 sri ouch
20:13 sri what do you need java for?
20:16 marcus gitorious
20:17 marcus it uses activemq for some reason
20:18 dotan left #mojo
20:18 Akron joined #mojo
20:20 GabrielVieira left #mojo
20:23 marcus just realized there is something called Apache Camel...
20:23 marcus and thatt's java based
20:35 armicron left #mojo
20:38 fhelmber_ left #mojo
20:51 kvorg joined #mojo
21:02 ltriant joined #mojo
21:31 Akron Is there a reason, why Mojolicious::Controller::redirect_to defaults to 302?
21:40 arthas left #mojo
21:47 dotan joined #mojo
21:47 spleenjack left #mojo
21:51 Akron Sear for "Mojo::URL" in the german Wikipedia and the first suggest is "John Rambo".
21:51 Akron Search
22:09 Htbaa1 left #mojo
22:22 stephen left #mojo
22:27 stephen joined #mojo
22:32 forwardever left #mojo
22:41 sri lol
22:41 sri Akron: why not 302?
22:41 Akron It may be misused for permant redirection.
22:42 Akron Which is - I think - the more common case.
22:43 sri 302 is pretty much standard
22:43 sri it just works, we would need a very good reason to change it
22:44 sri 303 is http 1.1
22:45 sri if you tell me "but everybody supports 1.1" i'm going to slap you :)
22:45 kvorg left #mojo
22:50 Akron But I think 301 (and 303) are more supported by search engines for permanent redirection - leading to the deletion of the redirecting url. and I think permanent redirection is more common than temporary.
22:51 Akron It's a semantic question. I use 302 if I really want to keep the old URL alive.
23:03 sri umm
23:03 sri Akron: the most common use case is a redirect after a form post
23:04 sri thats surely not a permanent one
23:04 sri quite sure thats also the only documented use case
23:05 Akron That's right, but why do you think that's the most common case?
23:05 sri experience
23:06 Akron Hm.
23:09 Akron Well - in fact I have no problem with it. I was just wondering. But yeah - in a controller it makes sense to return a 301. But maybe you should document that in the POD ...
23:11 sri document what?
23:14 GitHub5 joined #mojo
23:14 GitHub5 mojo: master Sebastian Riedel * f22327c (1 files in 1 dirs): cleanup - http://bit.ly/dg8gwH
23:14 GitHub5 left #mojo
23:15 dotan left #mojo
23:17 Akron sri: ;) *thumpsup*
23:17 Akron thumbs i mean ...
23:18 kthakore Akron: so you are from india?
23:19 Akron Am I? Not really ...
23:22 und3f is now known as und3f[A]

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