Camelia, the Perl 6 bug

IRC log for #mojo, 2010-11-04

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

All times shown according to UTC.

Time Nick Message
00:04 und3f is now known as und3f[A]
00:53 tl joined #mojo
00:53 tholen42 left #mojo
01:31 GabrielVieira left #mojo
01:46 memememomo_ is now known as memememomo
01:53 memememomo is now known as memememomo_
01:56 perlrocks Twitter: "Mojolicious meets Xslate - Islands in the byte stream http://bit.ly/da5wK3" (ja) --wPerlnews http://twitter.com/wPerlnews/status/29629487777
02:11 GabrielVieira joined #mojo
02:31 perlrocks Twitter: "How to include Mojolicious in your app's git repository / No time to wait http://htn.to/o1XxcP" (ja) --memememomo http://twitter.com/memememomo/status/29632159308
02:53 tardisx left #mojo
02:53 tardisx joined #mojo
02:57 baton8 joined #mojo
03:42 tholen42 joined #mojo
03:43 tl left #mojo
04:04 tl joined #mojo
04:06 tholen42 left #mojo
06:32 ltriant left #mojo
07:04 und3f[A] is now known as und3f
07:07 perlrocks Twitter: "RT @kraih Released #mojolicious 0.999935 with quite a few performance improvements and bugfixes, enjoy! :) #perl" --xomaa http://twitter.com/xomaa/status/29647353242
07:13 sri moin
07:35 koban joined #mojo
07:37 GitHub13 joined #mojo
07:37 GitHub13 mojo: master Sebastian Riedel * ccb7750 (1 files in 1 dirs): cleanup - http://bit.ly/aXE3N0
07:37 GitHub13 left #mojo
07:38 perlrocks Twitter: "Mojolicious:: Lite The easy wish" (ja) --khirose http://twitter.com/khirose/status/29648628043
07:41 GitHub85 joined #mojo
07:41 GitHub85 mojo: master Sebastian Riedel * f56f468 (1 files in 1 dirs): cleanup - http://bit.ly/aKZ6rG
07:41 GitHub85 left #mojo
08:00 und3f is now known as und3f[A]
08:08 perlrocks Twitter: "Mojolicious meets Xslate - Islands in the byte stream http://htn.to/QvHRX1" --lapis25 http://twitter.com/lapis25/status/29649904400
08:09 MojoGuest297 joined #mojo
08:09 MojoGuest297 From: http://www.google.hu/search?sourceid=ch​rome&ie=UTF-8&q=websocket+demo (1 hits)
08:09 MojoGuest297 left #mojo
08:20 GabrielVieira left #mojo
08:36 arthas joined #mojo
08:36 arthas left #mojo
08:50 yko joined #mojo
08:57 ysyrota joined #mojo
09:49 Rhaen hello world!
10:10 perlrocks Twitter: "Mojolicious:: Lite 'll start! / Mojolicious / allframeworks.ru - Blog about modern web development and .. http://bit.ly/bWSzH8" (ru) --everbslab http://twitter.com/everbslab/status/29655481748
10:20 MojoGuest228 joined #mojo
10:20 MojoGuest228 From: http://www.google.es/search?q=websocket+demo&r​ls=com.microsoft:es&ie=UTF-8&oe=UTF-8&​startIndex=&startPage=0&rlz=&qscrl=1 (1 hits)
10:20 MojoGuest228 left #mojo
10:21 * sri waves
10:21 spleenjack joined #mojo
10:22 memememomo_ is now known as memememomo
10:23 dm_ joined #mojo
10:46 perlrocks Twitter: "Thank you! / Mojolicious meets Xslate - Islands in the byte stream http://htn.to/ucgRAb" (ja) --memememomo http://twitter.com/memememomo/status/29657436410
10:48 perlrocks Twitter: "[Perl] Mojolicious:: Lite with, Text:: XSlate be able to use http://d.hatena.ne.jp/meme​memomo/20100919/1284893899" (ja) --memememomo http://twitter.com/memememomo/status/29657507361
10:50 perlrocks Twitter: "RT! Memememomo: [perl] Mojolicious:: Lite で, Text:: XSlate を 使える よう に する http://d.hatena.ne.jp/meme​memomo/20100919/1284893899" (id) --perlism http://twitter.com/perlism/status/29657620158
10:59 memememomo is now known as memememomo_
11:20 dm_ left #mojo
11:31 dm_ joined #mojo
12:09 perlrocks Twitter: "Mojolicious meets Xslate http://j.mp/azLf0z" (ja) --akiym http://twitter.com/akiym/status/29662631045
12:43 GabrielVieira joined #mojo
12:57 dotan joined #mojo
13:04 daviddelikat left #mojo
13:08 forwardever joined #mojo
13:48 fhelmberger left #mojo
13:52 daviddelikat joined #mojo
14:05 forwardever improved mojox-scaffold, it now has orm support (with dbi and objectdb as example orms) and a step by step interface
14:07 fhelmberger joined #mojo
14:13 und3f[A] is now known as und3f
14:14 Christian joined #mojo
14:14 Christian hi
14:15 dotan forwardever: is mojox-scaffold on github?
14:15 forwardever https://github.com/forwardever/mojox-scaffold
14:16 spleenjack left #mojo
14:18 forwardever some feedback would be great, there are some issues that have to be solved however, e.g. adding validation, make it working when singular und plural name of a resource are the same, and making test files not writing in the main db, but a test db instead (to name the most important ones)
14:20 forwardever another to do: setting custom model namespace in the step-by-step config
14:28 Christian left #mojo
14:50 perlrocks Twitter: "Mojolicious:: Lite you can create a large-scale Web applications http://icio.us/vgw5zf" (ja) --hiro1270jpn http://twitter.com/hiro1270jpn/status/29673277704
15:13 perlrocks Twitter: "Mojolicious:: Lite install it, I wanted to play a bit. What is troublesome Mojolicious spellings." (ja) --syohex http://twitter.com/syohex/status/29675186783
15:48 koban left #mojo
16:01 kimoto joined #mojo
16:03 kimoto sri: i maybe fixed multiple conversion to string in xml escape. i think the performance is faster a little.
16:03 kimoto https://github.com/yuki-kimoto/mojo/commit​/d7e40e4d3a67453a8dbaaaeef3f476c85cdc1c83
16:10 sri kimoto: what does it fix?
16:11 sri especially $v = '' unless defined $v; still being present seems very wrong
16:11 sri disabling all warnings is also questionable
16:12 sri do you mean no warnings 'uninitialized'?
16:13 sri a benchmark would be rather simple i guess
16:16 sri you could very well be right, but i'd like to see a small benchmark :)
16:18 perlrocks Twitter: "Release of asynchronous Redis client on #mojolicious steroids https://github.com/und3f/mojox-redis #perl #redis" --und3f http://twitter.com/und3f/status/29680541517
16:21 dotan left #mojo
16:23 und3f sri, it is not possible to use async capabilities of mojo in fastcgi mode :|
16:34 sri und3f: i know
16:35 und3f sri, don't you thats a little bit strange that async code written in daemon mode will not work in other modes?
16:36 sri well, it can't work with cgi anyway
16:37 sri for fastcgi, well, feel free to add ioloop support :)
16:37 sri (as long as it keeps working on windows)
16:39 sri and psgi has the same limitations as cgi
16:40 sri if this was a simple problem it would be solved already ;)
16:41 sri since i can't test on windows there chance of me rewriting the fastcgi binding with ioloop support
16:41 sri *there's no chanc
16:41 sri *+e
16:41 sri :S
16:42 sri cgi/fastcgi are dead platforms anyway
16:42 sri they have no place in the realtime web
16:44 sri und3f: it might even be simple to patch in, but i simply can't test it well enough
16:46 und3f sri, fastcgi is dead??
16:46 sri it will soon be
16:46 sri it sucks for comet
16:46 und3f till mojo don't work with plackup fastcgi is only possible platform for me
16:47 sri "till mojo don't work with plackup" what?
16:48 marcus sri: most of us don't need comet.
16:48 sri then you don't need async either
16:50 und3f sri, what tools i can use to run mojolicious app in psgi mode?
16:50 marcus well, you could need as http client.
16:50 und3f sri, Plack?
16:50 sri marcus: what for?
16:50 marcus sri: to fetch porn while rendering templates.
16:50 sri in a fastcgi environment it doesn't matter if you block
16:51 sri blocking client can fetch parallel anyway
16:51 marcus yea
16:51 sri und3f: what about it?
16:51 marcus async is Hard, let's go shopping.
16:51 sri !
16:53 sri und3f: we do support Plack perfectly fine
16:53 und3f i am going check again
16:53 und3f tried it not so long ago and got errors
16:53 marcus I just spent the last 3 days debugging lingering issues after changing my ipad/iphone tvguide to use async http/xml parsing instead of blocking :)
16:54 marcus so much fun with threads
16:54 marcus and shit that can go wrong when you update data in the background
16:54 sri und3f: see cookbook for common problems with plack
16:55 sri marcus: USE AN EVENT LOOP!
16:55 marcus sri: IT'S NO GOOD FOR NATIVE APPS
16:55 sri it's clearly the answer to all async problems :)
16:55 marcus clearly :)
16:56 marcus I guess I should rewrite my app in node.js
16:56 sri clearly :)
16:56 marcus just need to port it to iOS first.
16:57 und3f sri, http://paste.pocoo.org/show/286116/
16:57 und3f sri, isn't that right startup of psgi application
16:57 sri und3f: try with a fresh generated app
16:58 sri und3f: maybe your script is old or bugged
16:58 sri it is likely the use lib code
16:59 sri which was once relative to the executable path
16:59 sri that doesn't work for plackup because it is the executable
16:59 sri newer scripts are a bit smarter
17:00 und3f sri, it runs with a new generated lite app.
17:00 sri plack makes easy deployment really hard (ironically)
17:00 und3f sri, but i don't use lib
17:00 sri und3f: well, your problem is somewhere in your script
17:01 und3f sri, is there anything else, except plack?
17:01 sri ?
17:01 und3f to run PSGI apps
17:01 sri non plack psgi compliant servers?
17:03 sri btw. async support is impossible with psgi
17:03 und3f never mind
17:04 sri psgi is built to be blocking
17:04 sri you can only hook into some outer event loop bypassing psgi
17:04 sri which is server specific
17:05 sri and very much non portable
17:06 sri thats the reason mojolicious async support is designed around the assumption of being an optimization
17:10 sri non blocking code is really really hard if you have to play well with others
17:12 daviddelikat left #mojo
17:13 sri this is exactly why i care so much about hypnotoad ;)
17:14 ysyrota left #mojo
17:15 su-bzero_ joined #mojo
17:17 sri oh, fun xmas movie this year... Rare Exports
17:20 su-bzero left #mojo
17:30 achromic left #mojo
18:16 und3f how to keep alive Mojo::IOLoop connections? They are dropping by some timeout
18:20 und3f sri, is it possible? :|
18:20 daviddelikat joined #mojo
18:21 perlrocks Twitter: "@ Kraith Is is therefore possible to use # mojolicious as normal CGI (Apache + Suexec?) Do you have a good example for sortal, editable HTML table?" (de) --disaster123tw http://twitter.com/disaste​r123tw/status/29689838292
18:21 sri und3f: what?
18:21 und3f sri, how to keep Mojo::IOLoop connection alive?
18:22 sri they are kept alive automatically until they time out
18:22 sri ->connection_timeout
18:22 und3f khm..
18:22 yko left #mojo
18:22 und3f sri, what about another idle_timeout ?
18:22 sri ?
18:22 sri what for?
18:23 und3f for keeping connection idle
18:23 sri you need to be way more specific
18:23 und3f now, i'll write all
18:23 sri umm...
18:23 sri thats exactly what connection_timeout does
18:24 sri there is no hard timeout built in
18:24 und3f sri, can i reset connection_timeout for specific connection
18:24 sri just look at the pod
18:24 und3f oh, thanks, missed it
18:25 forwardever added flash messages to MojoX::Scaffold :)
18:27 und3f sri, but i can't disable it?
18:27 sri und3f: why would you ever want to disable it?
18:28 und3f sri, because i need to keep connection to Redis forever
18:28 sri you could just set it to a very large value
18:28 sri set it to ten years then ;p
18:28 und3f looks like a primitive hack :(
18:28 und3f ok..
18:29 sri it's not
18:29 sri disabling a timeout is easy and mostly a very bad idea
18:30 sri a big value makes users think about it
18:30 sri it's encouraging the right behavior
18:31 sri we don't prevent you from shooting yourself in the foot, but we also won't make it easier than neccessary
18:31 sri :)
18:36 jfuller joined #mojo
18:41 MojoGuest134 joined #mojo
18:41 MojoGuest134 From: http://news.ycombinator.com/item?id=1277067 (39 hits)
18:42 MojoGuest134 left #mojo
18:42 achromic joined #mojo
19:10 kimoto_ joined #mojo
19:13 kimoto left #mojo
19:14 jfuller Is this the primary channel for mojolicous?
19:15 und3f jfuller, right
19:16 und3f One more mojolicious-based project released http://check-host.net/
19:16 sri und3f: neat
19:17 und3f sri, just like mojo
19:17 sri :)
19:19 su-bzero_ left #mojo
19:25 sri und3f: i think i might do fastcgi with ioloop, but we need a windows tester
19:25 und3f actually i don't need it anymore :PPP
19:26 und3f sri, just wrote daemon that receives all requests
19:27 sri then someone else will benefit from it ;p
19:36 und3f :)))
19:45 polvo what is the best way to run a lengthy cron script w/o increasing http timeouts?
20:00 GabrielVieira left #mojo
20:06 Htbaa joined #mojo
20:16 gizzlon hey guys ..
20:17 gizzlon have some dumb problems after a git pull .. could someone try this pls:
20:17 gizzlon MOJO_MAX_REDIRECTS=10 perl -Mojo -we 'print g("http://www.stackoverflow.com")->body'
20:17 gizzlon kthnxbye
20:19 gizzlon polvo: lenghty how?
20:21 polvo gizzlon, 5 minutes or more
20:25 gizzlon polvo: dont even understand your question =/ what does cron have to do with http timeouts?
20:27 sri gizzlon: looks like a resolver bug
20:27 sri MOJO_IOLOOP_DEBUG=1 makes it quite obvious
20:27 sri i blame und3f ;p
20:28 gizzlon sri: haha
20:28 gizzlon paralell ftw
20:28 und3f :(
20:30 und3f sri, we need a list of domains that we DON"T resolve :P
20:31 sri und3f: ?
20:31 und3f sri, for example www.stackoverflow.com
20:31 und3f : P
20:31 gizzlon hehe.. great feature
20:33 und3f sri, if i just know about MOJO_IOLOOP_DEBUG before... I could save hours of debugging...
20:36 polvo gizzlon, a http request triggers the job, but the connection closes till the job is done
20:40 polvo gizzlon, hope it's clearer
20:42 JonChisciotte joined #mojo
20:46 gizzlon polvo: so you have a client that does a http get or post to a server, and that triggers a job?
20:46 polvo gizzlon, right
20:46 memowe_ joined #mojo
20:46 gizzlon polvo: and the job just takes soo long the client disconnects?
20:46 polvo gizzlon, right
20:47 gizzlon polvo: dunno, but I guess the client wont disconnect as long as it gets _something_ regulary.. ?
20:47 Alias left #mojo
20:47 polvo gizzlon, no, the client just needs to start the job.
20:48 polvo gizzlon, he doesn't need to wait for response
20:48 sri does the job run inside mojolicious?
20:48 sri that would be wrong
20:49 polvo sri, it does run. why is it wrong?
20:50 sri because everything is bound to the client connection
20:50 polvo sri, i could run the cron job from the command line, but the startup (hello DBIC) takes up to 20 seconds
20:50 sri then use a job server
20:50 MojoGuest433 joined #mojo
20:50 MojoGuest433 From: http://www.google.no/search?sourceid=ch​rome&ie=UTF-8&q=websockets+demo (1 hits)
20:50 polvo sri, job server?
20:50 sri long running tasks should never run inside a web app
20:51 memowe left #mojo
20:51 sri i'm sure google has a lot of infos on job servers
20:51 gizzlon polvo: cron??? Guess you coul fork() or use something like geraman ..
20:52 sri gearman like, yea
20:52 MojoGuest433 left #mojo
20:54 polvo hm, interesting
20:54 sri btw. i won't add ioloop support to the fastcgi binding, it's just not worth the time investment
20:54 gizzlon polvo: or stick the info you need in a file or database and have another scrip to run the job
20:55 jfuller left #mojo
20:55 gizzlon polvo: the web script just does a db insert and another scrips just select out data and runs whatever you want
20:56 gizzlon polvo: but gearman is web-scale ;)
20:59 ltriant joined #mojo
20:59 stephen left #mojo
21:00 polvo gizzlon, may be some day :) for now i'll make a simple daemon.
21:02 stephen joined #mojo
21:02 gizzlon polvo: using mysql sounds easier too me, but I guess that's because I'm used to it
21:06 forwardever [% perl_code %] in templates = holidays for my eyes :)
21:11 und3f is now known as und3f[A]
21:14 xantus ;)
21:16 sri und3f[A]: looks like some of your unpacks are wrong
21:18 * xantus packs a sock full of quarters
21:20 kvorg joined #mojo
21:28 perlrocks Twitter: "Mojolicious:: Lite you can create a large-scale Web apps - Getting Started with Perl code samples http://htn.to/f3ufks" (ja) --memememomo http://twitter.com/mememem​omo/status/298293758263296
21:29 GabrielVieira joined #mojo
21:30 perlrocks Twitter: "RT! Memememomo: Mojolicious:: Lite you can create a large-scale Web apps - Getting Started with Perl code samples http://htn.to/f3ufks" (ja) --perlism http://twitter.com/perlism/status/298760227782656
21:35 sri und3f[A]: non blocking resolver might have to go again if we don't find the bug
21:37 xantus mojo peggs my cpu with a websocket connected, and there is no activity
21:38 sri xantus: make a test case
21:39 sri oh actually we have one
21:39 sri but dns gets priority now
21:39 kvorg left #mojo
21:42 xantus k
21:43 xantus I should be on later to do some hacking
21:43 Htbaa left #mojo
21:44 sri doesn't look good for the async resolver
22:02 sri looks like the name is not compressed for stackoverflow.com
22:11 sri or not :/
22:25 memowe_ left #mojo
22:26 sri TC is also not set
22:26 sri grrr
22:26 fhelmberger left #mojo
22:26 sri no idea why it resolves to 64.34.119
22:27 sri i really don't see why the last part is missing
22:28 kimoto_ sri: sorry, I sleeped soon
22:28 kimoto_ this is profile by NYTProf
22:28 kimoto_ http://pastie.org/1273547
22:28 kimoto_ to_string is maybe many time for stringification by overloading in escape function of Mojo::Template.
22:29 kimoto_ $v = "$v" is to do stringification once.
22:30 sri kimoto_: will take a look at it later, dns bug has priority
22:31 kimoto_ sri: ok. no warnings means no warnings 'uninitialized'
22:32 forwardever left #mojo
22:36 JonChisciotte1 joined #mojo
22:37 y0j left #mojo
22:39 memowe joined #mojo
22:40 JonChisciotte left #mojo
22:44 Alias joined #mojo
22:46 y0j joined #mojo
22:47 DaTa sri: i guess something is wrong with that line: my @packet = unpack 'nnnnnnA*', $chunk;
22:48 DaTa the last byte is missing in $packet[6] for some reason, in $chunk the byte is there
22:49 DaTa but i'm just guessing
22:53 sri DaTa: you're right, thats where the byte gets lost
22:54 DaTa Data::Hexdumper++ :)
22:54 sri i just use unpack 'b*' ;)
22:57 sri yea the A* just doesn't work if there are cnames mixed in
23:04 dm_ left #mojo
23:13 sri ok, that was easy to fix
23:15 DaTa sri**=2
23:16 GitHub189 joined #mojo
23:16 GitHub189 mojo: master Sebastian Riedel * d6e4745 (2 files in 2 dirs): fixed non-blocking DNS resolver bug - http://bit.ly/csfbI6
23:16 GitHub189 left #mojo
23:21 daviddelikat left #mojo
23:23 GitHub151 joined #mojo
23:23 GitHub151 mojo: master Sebastian Riedel * 4b40fd6 (2 files in 2 dirs): improved Mojo::Template performance slightly - http://bit.ly/8Zk1LB
23:23 GitHub151 left #mojo
23:31 daviddelikat joined #mojo
23:31 GitHub139 joined #mojo
23:31 GitHub139 mojo: master Sebastian Riedel * 10f264a (2 files in 2 dirs): fixed a serious WebSocket bug - http://bit.ly/avtn3C
23:31 GitHub139 left #mojo
23:31 perlrocks Twitter: "Mojolicious:: Lite you can create a large-scale Web applications http://j.mp/bbyoqX" (ja) --deris0126 http://twitter.com/deris01​26/status/329428320129024
23:32 sri xantus: should be fixed
23:33 sri gonna release soon
23:49 perlrocks Twitter: "Released #mojolicious 0.999936 with a few important bugfixes. #perl" --kraih http://twitter.com/kraih/status/333948605431808

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