Camelia, the Perl 6 bug

IRC log for #mojo, 2010-07-30

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

All times shown according to UTC.

Time Nick Message
00:21 tempire vim syntax highlighting++ !
00:34 ashleydev joined #mojo
00:35 sri hmm, anyone want bonjour support for the daemon?
00:46 GitHub46 joined #mojo
00:46 GitHub46 mojo: master Sebastian Riedel * 238fc08 (3 files in 2 dirs): added experimental Bonjour support - http://bit.ly/bTRwFS
00:46 GitHub46 left #mojo
00:46 sri it's quite convenient
00:50 GitHub74 joined #mojo
00:50 GitHub74 mojo: master Sebastian Riedel * 55bbdc9 (1 files in 1 dirs): updated cheatsheet - http://bit.ly/9lbiqd
00:50 GitHub74 left #mojo
00:55 xantus_ sri: hell yeah
00:58 y0j joined #mojo
01:30 GitHub110 joined #mojo
01:30 GitHub110 mojo: master Sebastian Riedel * b498222 (4 files in 4 dirs): cleanup - http://bit.ly/d1LNG6
01:30 GitHub110 left #mojo
01:30 sri interesting, bonjour for https doesn't get picked up by browsers
01:33 GitHub27 joined #mojo
01:33 GitHub27 mojo: master Sebastian Riedel * c541084 (1 files in 1 dirs): cleanup - http://bit.ly/97ny0f
01:33 GitHub27 left #mojo
01:34 sri marcus was also asking about bonjour recently i think :)
01:42 GitHub73 joined #mojo
01:42 GitHub73 mojo: master Sebastian Riedel * 5594a2d (2 files in 2 dirs): pod cleanup - http://bit.ly/agqcXI
01:42 GitHub73 left #mojo
01:45 Alias joined #mojo
02:29 achromic joined #mojo
02:32 GitHub20 joined #mojo
02:32 GitHub20 mojo: master Sebastian Riedel * 65d7f91 (2 files in 2 dirs): cleanup - http://bit.ly/9XH27U
02:32 GitHub20 left #mojo
03:11 Foxcool joined #mojo
03:19 janus joined #mojo
03:41 GitHub51 joined #mojo
03:41 GitHub51 mojo: master Sebastian Riedel * 26af459 (3 files in 3 dirs): added trim method to Mojo::ByteStream - http://bit.ly/clRU54
03:41 GitHub51 left #mojo
04:04 tholen42 joined #mojo
04:15 Foxcool_ joined #mojo
04:15 ask joined #mojo
04:39 ashleydev joined #mojo
04:39 Foxcool joined #mojo
06:34 ashleydev joined #mojo
06:53 sri http://github.com/kraih/mojo-tmbundle/commi​t/3112e35a2926451a4c69692679ddb9b577ad48c2 # now with a few snippets :)
06:53 garfield [ Commit 3112e35a2926451a4c69692679ddb9b577ad48c2 to kraih's mojo-tmbundle - GitHub ]
06:53 garfield http://xrl.us/bhurp2
07:15 MojoGuest670 joined #mojo
07:27 yko joined #mojo
07:29 Foxcool_ joined #mojo
07:49 MojoGuest145 joined #mojo
07:50 MojoGuest145 ...
07:58 Alias_ joined #mojo
08:01 sri yay, think i got highlighting right now :)
08:01 yko :D
08:02 vti sri: how do i use bonjour? :)
08:02 sri it just works
08:02 sri (if you have the xs module it needs installed)
08:03 vti every started daemon?
08:03 yko vim users have Mojo highlighting since Jan 08, 2010 :p
08:03 sri mojo version will show you
08:03 sri every daemon yes
08:03 sri it doesn't cost anything as far as i can see
08:04 vti yesterday before my internet died i noticed that i can't handle 'connection refused' errors again =/
08:04 sri just triggers a service thats running on os x anyway
08:04 vti there is always $tx->success
08:04 vti sri: ok, i will try it :)
08:04 sri bonjour is really damn convenient
08:09 sri http://twitpic.com/29y9v5 # latest tmbundle
08:09 garfield [ Almost there now with #mojolicious highlighting. on Twitpic ]
08:10 sri also have snippets for ep tags
08:10 sri x<TAB> and you get <%=  %>
08:11 vti nice, might want to copy it for vim ;p
08:12 sri it's actually even cooler, most snippets have multiple insert positions, so you tab multiple times to get to different positions
08:14 sri and those snippets are context sensitive, you can use the same trigger word/character in your script and template for different things
08:14 yko hmm. nice idea. we can implement most helpers in vim ye
08:16 sri vti: the safari bonjour support is hidden in the bookmark drawer btw
08:28 esskar joined #mojo
08:33 esskar moin
08:33 esskar memowe: http://perlpunks.de/paste/show/4c528e46.2ff2.1c4
08:33 * sri waves
08:33 esskar sorry, no diff installed here
08:38 memowe esskar: thanks.
08:38 esskar np
08:38 memowe sri: I'm working through the client tests. You'll get patches soon :)
08:38 memowe sorry, afk :)
08:38 sri \o/
08:39 yko sri: http://alexiuss.deviantart.com/gallery/#/d2u7p4v
08:39 garfield [ alexiuss's Gallery ]
08:40 sri lol
08:41 yko what did i told about developer in a box? :D
08:42 Foxcool_ joined #mojo
08:47 vti don't ignore my connection refused errors!
09:10 sri vti: who are you yelling at?
09:13 vti 10:04 <@vti> yesterday before my internet died i noticed that i can't handle 'connection refused' errors again =/ there is always $tx->success
09:15 sri i don't believe in errors without unit tests
09:16 Foxcool_ joined #mojo
09:27 Alias_ joined #mojo
09:31 vti mojo get 'localhost:1234' finishes silently
09:32 vti ok, ok writing
09:34 vti sri: https://gist.github.com/3689e215ecb3026ccf77
09:47 MojoGuest433 joined #mojo
09:52 Foxcool_ joined #mojo
09:54 GitHub81 joined #mojo
09:54 GitHub81 mojo: master Sebastian Riedel * 558d224 (2 files in 2 dirs): fixed client error reporting bug - http://bit.ly/atnslL
09:54 GitHub81 left #mojo
09:54 sri vti: you know you are stealing memowe his work? ;p
09:58 vti i remember that it worked some day
09:58 vti i am writing oauth provider! ;p
09:58 y0j joined #mojo
09:59 sri \o/
09:59 sri you know... things without unit tests break :)
10:00 sri thats exactly why we have a gsoc student!
10:00 * sri cracks the whip
10:05 sri vti: what do you think about bonjour?
10:05 marcus We won't be doing any rendezvous publishing, please install a Net::Rendezvous::Publish::Backend:: module
10:05 vti have this too
10:05 sri marcus: try the apple one
10:05 marcus am on it
10:05 sri those are sadly platform specific
10:06 vti hehe, works :)
10:06 sri :)
10:07 marcus works here too
10:07 sri \o/
10:07 marcus very hidden placement in safari
10:07 sri yea
10:07 sri i expected something in the finder or so
10:07 marcus you can pull it to the toolbar tho
10:07 sri Oo
10:08 marcus I expected it to be in the bookmarks menu
10:08 sri tried that earlier but failed
10:08 marcus from the sidebar
10:08 marcus worked fine here
10:08 sri oh cool
10:10 sri was just like 5 lines of code, quite nice
10:11 sri fire and forget
10:12 marcus the new sherlock series is awesome
10:12 marcus must see
10:12 sri have it somewhere already
10:13 marcus saw it with my brother yesterday
10:13 marcus now looking forward to next episode
10:14 sri new futurama episode too
10:14 marcus yeah, put.io is getting it for me
10:15 marcus <3 rss subscriptions
10:15 marcus + boxee app
10:16 marcus just hope they add convert to ipad format soon.
10:16 marcus iphone one is a bit lowres
10:16 * sri never converts
10:17 marcus put.io converts on the server side
10:17 marcus So I can just mark stuff I want to stream to the phone, and it'll be available in mp4.
10:21 GitHub62 joined #mojo
10:21 GitHub62 mojo: master Sebastian Riedel * 235998b (1 files in 1 dirs): cleanup - http://bit.ly/9oefvr
10:21 GitHub62 left #mojo
10:21 sri cool
10:25 sri not even seen true blood yet :/
10:30 murakoma joined #mojo
10:32 sri marcus: btw. try the textmate bundle :)
10:51 ask joined #mojo
11:26 vti http://github.com/vti/mojox-oa​uth/blob/master/example/client
11:26 garfield [ example/client at master from vti's mojox-oauth - GitHub ]
11:26 garfield http://xrl.us/bhur29
11:26 vti http://github.com/vti/mojox-oa​uth/blob/master/example/server
11:26 garfield [ example/server at master from vti's mojox-oauth - GitHub ]
11:26 garfield http://xrl.us/bhur3b
11:26 vti now you can run oauth client and server and test them agains each other!
11:32 sri \o/
11:34 sri looks good
12:43 Akron joined #mojo
13:12 zakame .j rakudo
14:16 MojoGuest607 joined #mojo
14:17 MojoGuest607 is there a way to connect to other IRC servers that's just not obvious to me?
15:42 stephan48 joined #mojo
15:48 stephan48 joined #mojo
16:16 stephan48 joined #mojo
16:19 stephan48 joined #mojo
16:38 Akron left #mojo
16:54 _mpu joined #mojo
17:43 diegok joined #mojo
17:53 xantus_ I started seeing the $fd issue
17:56 ashleydev joined #mojo
18:14 siegfri3d joined #mojo
18:14 sri xantus: can you turn it into a unit test?
18:15 sri it's really annoying to have people talk about a bug without giving me the chance to replicate it...
18:23 xantus_ Its coming from client from within mojox couchdb
18:23 xantus_ I'm wondering if its due to some error
18:26 sri are you using the latest github version?
18:27 sri it smells like people are just cloning clients without caring about keep alive queues
18:31 xantus_ yeah
18:32 xantus_ I do think its mojox couchdb
18:32 xantus_ creating too many clients
18:36 sri just wrote a little stress test and i can't make it run out of fds if used right
18:36 sri i'm almost certain now
18:39 vti almost doesn't count!
18:40 sri ?
18:41 vti just a joke... ;p
18:47 Akron joined #mojo
18:48 GitHub33 joined #mojo
18:48 GitHub33 mojo: master Sebastian Riedel * 7def613 (1 files in 1 dirs): added stress test for client - http://bit.ly/a4pGDB
18:48 GitHub33 left #mojo
18:52 sri -.-
18:53 sri maybe i should add a global idiot proof mode disabling all features that can be dangerous :)
18:53 sri MOJO_NO_ROPE=1
19:00 ashleydev joined #mojo
19:00 sri ok, there is one more thing i could do
19:00 sri actve leak prevention
19:00 sri *+i
19:00 xantus_ MOJO_IGIT_MODE=1
19:42 GitHub14 joined #mojo
19:42 GitHub14 mojo: master Sebastian Riedel * ffe6b64 (4 files in 3 dirs): added automatic file descriptor leak prevention to Mojo::Client - http://bit.ly/a1F0UB
19:42 GitHub14 left #mojo
19:43 sri should be a lot harder to use the client wrong now
19:49 sri i wanted to make automatic keep alive easy, and in turn made fd leaks easy too
19:50 sri this is better
19:52 sri now you would have to get really creative to leak, like for (1..100) { Mojo::Client->new->async->get(...)... }
19:52 stephan48 ;)
19:57 xantus_ ;)
19:58 xantus_ ok, installed.  will test later
19:59 xantus_ reorganizing how I deal with couch
19:59 xantus_ each client session gets a db
19:59 xantus_ _changes filter replication can propigate events to the client db
20:00 xantus_ you can do partitioning too!
20:09 sri couchdb replication is absolutely amazing
20:36 bedi joined #mojo
21:51 cognominal joined #mojo
21:59 GitHub159 joined #mojo
21:59 GitHub159 mojo: master Sebastian Riedel * 9c2c3fa (3 files in 2 dirs): add tests and warnings - http://bit.ly/dw0KTQ
21:59 GitHub159 left #mojo
22:14 ask joined #mojo
22:42 GitHub116 joined #mojo
22:42 GitHub116 mojo: master Sebastian Riedel * 771921a (6 files in 4 dirs): disabled keep alive by default in Mojo::Client - http://bit.ly/d4NJDo
22:42 GitHub116 left #mojo
22:43 sri overkill is underrated
22:46 Akron :)
22:49 tempire before_dispatch hooks are app-wide, right?
22:52 marcus sri: what is the vector software you used?
22:53 sri tempire: all hooks are app wide
22:53 sri marcus: Opacity
22:53 tempire just making sure.
22:54 Akron Is there a possibility to rely PATH_INFO for redirect_to, url_for etc. instead of REQUEST_URI?
22:54 Akron When using mod_rewrite it's no fun.
22:57 Akron Or is there something I'm missing in mod_rewrite? My RewriteCond %{REQUEST_URI} !/foo\.fcgi \n RewriteRule ^(.*)$ /foo.fcgi/$1 [L,QSA] Always adds the scriptname in mein URI.
23:00 sri think you can do all that with mod_rewrite, but i always forget how
23:00 sri in the cookbook we have easier solutions i think
23:01 sri ScriptAlias or so is enough
23:03 Akron Shared webhost. :(
23:03 Akron It's amazing that FastCGI is working! But I don't think I can set script alias via .htaccess.
23:04 diegok joined #mojo
23:07 marcus sri: sick pricing
23:07 marcus although cheap compared to adobe I guess
23:07 sri was about to say it, better than adobe
23:08 sri Akron: you can always register a before_dispatch callback and modify ->req->url yourself
23:08 sri s/callback/plugin hook/
23:08 Akron Ah - that's a good idea!
23:09 sri thats ugly though, so you should try to find a webserver specific solution
23:09 sri to keep your app portable
23:09 sri oh shit...the whole keep alive issue has another reason :S
23:10 sri it's just a normal leak
23:10 sri !@#$%
23:13 achromic is there a todo list for the mojo project? i'm curious what to expect from it in the future.
23:13 sri feature wise i'm quite happy
23:13 sri now it's time for cleanup
23:14 sri stunning people with usability improvements :)
23:14 achromic so more sugary syntax
23:14 achromic ?
23:14 sri ye
23:14 achromic what about performance/benchmarks?
23:15 sri of course
23:15 achromic cool
23:16 sri it's not the main focus though, if we wanted to optimize for performance we would be using C ;p
23:17 sri my main goal is to optimize for *developer performance* :)
23:18 sri doing cool stuff with the least amount of work
23:18 achromic true, but i'm sure mojo would still kick catalyst's bloated butt
23:18 sri it already does imo :)
23:19 achromic are there any published numbers for outsiders to compare?
23:19 sri numbers?
23:19 achromic benchmarks
23:19 sri don't go there
23:20 sri you don't ever pick a web framework because of a benchmark
23:20 achromic some users need a bit more than sugar to win over
23:20 sri if you do you deserve what you get ;p
23:20 achromic besides the syntax guides, the project is a bit opaque
23:21 sri i will never publish benchmarks, because they are meaningless
23:21 sri the whole idea of picking a framework means that you've decided to trade raw performance for convenience and agility
23:22 sri *if* you ever become the next twitter you'll just replace components and grow out of it
23:22 achromic but it puts you in a ballpark
23:22 achromic if one framework is an order of magnitude faster in rps, or startup time...
23:22 sri when i wrote catalyst i never did benchmarks and i will not start now ;)
23:23 sri maybe we are faster, it doesn't matter
23:26 sri achromic: if you want to benchmark frameworks you do it with events like http://r09.railsrumble.com/entries
23:26 garfield [ Rails Rumble :: Rails Rumble Leaderboard ]
23:27 achromic because all startups are created in a weekend :-P
23:27 sri twitter started as a hello world blog
23:28 achromic ok, i see a similar question to mine is in the mojo faq, but has no answer: "Is anybody using Mojo for high-traffic sites? What's the performance like?"
23:28 sri it's enough :)
23:29 sri people misunderstand scaling all the time
23:29 sri let me requote something i wrote to the list
23:29 achromic hah, found something: http://www.slideshare.net/hidek/benchmarks-of​-perl-web-application-frameworks-presentation
23:29 garfield [ Benchmarks of Perl Web Application Frameworks ]
23:29 garfield http://xrl.us/bhutm3
23:30 sri "Take reddit for example, they handle about 800 req/s during peak hours with 80 servers, thats just 10 req/s per server.
23:30 sri People often mistake mistake micro benchmarks with scalability."
23:31 sri if you would pick the *fastest* framework, you would always pick the one with the least amount of features
23:31 sri it's flawed logic
23:32 sri if thats how you tick do yourself a favor and use C right from the start
23:32 achromic i don't care about pure numbers, i just want to know i'm in the right ballbark
23:33 achromic i'm sure others feel the same way
23:33 sri i have a little experience in this area, trust me when i say it's enough :)
23:34 sri let me assure you we are many times faster than rails :D
23:35 achromic yeesh, i could never get a piece of technology apporove by saying something like that
23:36 achromic (don't mean to troll)
23:37 sri it's faster than rails doesn't work? Oo
23:38 Akron It is? Wow!
23:38 achromic that's just a vague assurance. how was rails setup (mongrel, etc.)?
23:39 Akron It's Perl. That's enough to me. ;)
23:39 sri any numbers i published would be just as meaningless
23:40 sri because thats just not how you can measure web framework performance
23:41 sri how much more servers could you buy if app development took 2 months instead of a year?
23:42 sri how much is the ability to add features quicker than the competition worth?
23:43 sri scaling is an ongoing process, there is no single solution
23:44 achromic that's a good point. is that illustrated anywhere? side-by-side examples of apps created with different frameworks. i saw Alias start a series on Mojolicious and Dancer on his use.perl journal a while back
23:44 sri achromic: http://al3x.net/2010/07/27/node.html # awesome article you need to read
23:44 garfield [ Alex Payne — Node and Scaling in the Small vs Scaling in the Large ]
23:47 Akron I think modules should scale well for apps - but apps should scale by needs.
23:47 Alias_ joined #mojo
23:49 Akron But - there is something wrong when scalability is impossible due to design.
23:50 sri right
23:50 ashleydev_ joined #mojo
23:50 sri as long as you can throw money at a problem to make it go away it scales :)
23:51 cognominal joined #mojo
23:51 Akron :)
23:54 GitHub84 joined #mojo
23:54 GitHub84 mojo: master Sebastian Riedel * 8e222cc (6 files in 4 dirs): fixed memory and file descriptor leaks in Mojo::Client - http://bit.ly/bpzRe3
23:54 GitHub84 left #mojo
23:55 sri was so easy to fix...

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