Camelia, the Perl 6 bug

IRC log for #mojo, 2012-05-31

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

All times shown according to UTC.

Time Nick Message
00:16 gryphon joined #mojo
00:17 GabrielVieira joined #mojo
00:19 gryphon Greetings all. Is this the right place to ask a morbo/hypnotoad config question?
00:43 gryphon Is there a way to define in startup() what files/directories to watch for changes (regardless of using morbo or hypnotoad)?
00:50 jnap joined #mojo
01:02 scott joined #mojo
01:58 d4rkie joined #mojo
02:10 xaka joined #mojo
02:10 gryphon joined #mojo
02:55 asarch joined #mojo
03:26 tempire gryphon: only morbo is relevant when keeping track of changes
03:30 tempire you can tell morbo to watch whatever folders you want
03:31 tempire http://mojolicio.us/perldoc/Mojo/Server/Morbo
03:49 bird_ joined #mojo
03:56 Leandr joined #mojo
04:45 Leandr joined #mojo
05:15 Foxcool joined #mojo
05:16 ovnimancer joined #mojo
05:30 batman joined #mojo
06:04 AmeliePoulain joined #mojo
06:13 Vandal joined #mojo
06:51 rem_lex|pivo joined #mojo
07:21 fhelmber_ joined #mojo
07:31 spleenjack joined #mojo
07:35 AmeliePoulain joined #mojo
07:56 batman joined #mojo
08:00 spleenjack1 joined #mojo
08:00 pau4o joined #mojo
08:12 marcus sri: I was tempted to try sublime text2 again, I heard the vim mode is much better, with ex support as a plugin.
08:20 black joined #mojo
08:26 marcus batman: getbetsty looks cool.
08:26 batman thanks!
08:29 marcus betsy
08:29 marcus batman: you know if Petter develops natively or using something like titanium?
08:29 batman i think he's doing native
08:30 batman don't know what titanium is though :P
08:30 batman have you downloaded his bysykkel app?
08:30 marcus yes, I have
08:30 batman before or after my fb post? :)
08:30 marcus before :)
08:30 batman ok. sweet
08:30 marcus or rather, I didn't see your fb post
08:31 batman aiai.
08:31 marcus saw this one tho https://www.facebook.com/photo.php?fbid=10​151780511110317&set=a.1015176749279031​7.856493.844565316&type=1&theater :p
08:31 batman haha
08:41 xaka1 joined #mojo
08:45 cosmincx joined #mojo
09:31 daxim joined #mojo
09:38 * sri yawns
09:44 batman thanks for the tweet marcus. it's a bit stressing when you call me "talented". it makes faling harder to do :)
09:45 marcus batman: don't plan on failing then :)
09:45 sri that sounds like a good thing
09:45 batman i will fail. i'm almost positive about it
09:45 batman if this hits the fan, i will have trouble scaling
09:45 sri failing is 90% of programming
09:46 batman oh. i thought it was 99% :)
09:46 batman but there's no such thing as "failing". you only learn how "not to do it"
09:46 marcus having trouble scaling isn't failing :)
09:46 marcus just look at twitter...
09:47 batman if the site goes down, i will feel like i failed. even though in some way it's a perfect sign of success though
09:47 sri having trouble scaling is called success
09:47 batman :D
09:53 omega batman: when the site changes the example bets, it jumps up if I am scrolled passed :p
09:53 omega but I like the idea!
09:53 batman omega: yeah i know. it's quite annoying, but i haven't had time to fix it yet
09:56 pau4o joined #mojo
09:58 spleenjack joined #mojo
10:20 batman marcus: petter does objective c and xcode stuff... that's all i know :P
10:20 marcus batman: he answered audungk on twitter :)
10:21 batman aha
10:47 marcus crab: You know there's a illegal burger in Oslo?
10:47 marcus I actually plan to have dinner there today :p
11:07 sri marcus: that reference doesn't work outside of oslo
11:08 sri at first i thought it was slang for drugs :o
11:09 marcus :D
11:09 marcus sri: I think it's cute
11:09 marcus insider references ftw
11:10 marcus would maybe be better with Illegal burger tho
11:10 marcus maybe someone will get curious and google it :)
11:10 marcus it's a very good burger joint.
11:12 marcus I think it would be just as cute for a Berlin-based startup to keep a references to Dolores :)
11:12 marcus damn, now I want a burrito :-/
11:12 batman illegal burgers is really, really good
11:13 batman there was quite a queue yesterday, so i went to revolver instead --- just to get disappointed :(
11:13 marcus yeah, agree, Revolver's burger is much less yummy
11:13 batman good bread, bad meet
11:13 batman meat
11:14 marcus bad sauce, I think.
11:27 ryozi joined #mojo
11:38 jnap joined #mojo
11:47 sri "We need some more secret sauce. Put the mayonnaise in the sun."
11:49 marcus sri++
12:39 d4rkie joined #mojo
12:49 xaka joined #mojo
12:52 marcus hmm
12:52 marcus VACUUM FULL takes quite some time.
13:02 spleenjack1 joined #mojo
13:08 jnap joined #mojo
13:08 Psyche^ joined #mojo
13:28 xaka1 joined #mojo
13:55 inokenty joined #mojo
13:56 andrefs joined #mojo
14:11 Leandr joined #mojo
14:20 stephan48 sri: bad news, IO::Socket::SSL doesnt work with the patch
14:21 stephan48 it doesnt die cause of unexpected args anymore
14:21 asarch joined #mojo
14:21 stephan48 but instead it just ignores them
14:25 sri :(
14:25 stephan48 i can atm just say the ssl connect fails
14:25 stephan48 running the mojo ssl tests should show a problem or?
14:26 sri no
14:27 stephan48 they disable the workaround?
14:27 sri old release didn't depend on it, new release doesn't use IO::Socket::IP at all
14:28 sri remember, our workaround just worked for us, it was 3rd party modules that blew up
14:31 stephan48 excellent
14:34 stephan48 http://paste.stejau.de/?20
14:37 stephan48 segfaults for me#
14:37 stephan48 with his patch
14:37 sri ouch
14:38 sri well
14:38 sri maybe it imports functions/constants from IO::Socket::INET6 that make IO::Socket::IP crash
14:39 sri but i'm just guessing
14:43 stephan48 yea IO::Socket::SSL uses IO::Socket::INET6 if available..
14:47 sri you could try "use IO::Socket::SSL 'inet4';"
14:49 stephan48 nope
14:49 stephan48 doesnt work
14:49 stephan48 seg faults
14:51 gryphon joined #mojo
14:52 stephan48 but IO::S::SSL works also without the IO::Socket::IP workaround with ipv6 here
14:52 stephan48 without any workaround
14:52 sri because it uses IO::Socket::INET6
14:54 stephan48 yup
14:54 stephan48 so how do you solve it now?
14:54 sri we can't
14:54 rafl joined #mojo
14:55 sri as long as IO::Socket::IP is not in core i don't really care about using it
14:55 sri so we have at least another year
14:56 sri at some point IO::Socket::SSL will get support for it anyway, we can wait for that
14:58 andrefs joined #mojo
15:22 andrefs joined #mojo
15:22 asarch joined #mojo
15:22 Leandr joined #mojo
15:22 d4rkie joined #mojo
15:22 daxim joined #mojo
15:22 Xorg joined #mojo
15:22 Gedge joined #mojo
15:22 mikegrb joined #mojo
15:22 keedi joined #mojo
15:22 DaTa joined #mojo
15:22 bobkare joined #mojo
15:22 nuba joined #mojo
15:22 alnewkirk joined #mojo
15:22 ryanc joined #mojo
15:22 amirite joined #mojo
15:22 perlite joined #mojo
15:22 burk joined #mojo
15:22 robinsmidsrod joined #mojo
15:22 jmmills joined #mojo
15:22 kmx joined #mojo
15:22 charsbar__ joined #mojo
15:22 predgtr1 joined #mojo
15:22 kenjiskywalker joined #mojo
15:22 atrodo joined #mojo
15:22 kthakore joined #mojo
15:22 crab joined #mojo
15:22 stephan48 joined #mojo
15:22 tempire joined #mojo
15:22 mattastrophe joined #mojo
15:22 Debolaz joined #mojo
16:24 batman joined #mojo
17:17 mire joined #mojo
18:04 Liam`` joined #mojo
18:40 andrefs joined #mojo
18:44 hide joined #mojo
18:51 JiYu joined #mojo
18:51 JiYu hi
18:52 JiYu some one ever tried to save the websocket information within a redis database?
18:53 JiYu tried to serialize the date before storing, but somehow its not working
18:53 JiYu https://rt.fm/pbw
18:53 JiYu may be someone got any thoughts?
19:09 vervain joined #mojo
19:28 sri JiYu: what made you believe a transaction could just be serialized and restored?
19:28 JiYu sri, I just wanted to try it. is it impossible?
19:29 sri short answer no
19:29 JiYu wooow
19:30 JiYu i will check google for the long version ;)
19:30 sri i can't even think of a reason to try that
19:30 sri maybe you're misunderstanding something about websockets?
19:31 sri it's basically the same as trying to serialize the STDOUT handle
19:32 mire joined #mojo
19:32 JiYu oh
19:32 sri what's even more puzzling to me is that you're not even the first to attempt it :)
19:32 sri how did you get there?
19:32 JiYu haha
19:32 sri was it something in the documentation?
19:32 JiYu no. dont worry :)
19:33 JiYu I just wanted to try to achieve to scale the websocket information over more then one process (e.g. for hypnotoad )
19:33 sri i see
19:34 sri redis pub/sub is very popular for that
19:34 sri but you will always have to relay messages, one connection is tied to a single worker
19:35 sri websocket connections are persistent
19:35 JiYu yeah thats what I got now, when you mentioned it.
19:35 JiYu dont know how I got to that idea ;)
19:36 sri ok, now i feel more comfortable giving you the long answer :)
19:37 sri you could theoretically steal the actual socket handle, and send the file descriptor to other processes via pipes
19:37 sri it's very silly, but theoretically possible
19:37 sri don't do it ;)
19:38 zipkid joined #mojo
19:38 zipkid hello
19:38 sri a backend message but is the correct solution, like redis or zeromq or some other mq
19:38 * sri waves
19:39 sri s/but/bus/
19:39 JiYu yeah but then the connection is still bound on the first process and just "forwarded"
19:39 zipkid is it normal that the content of a 'content_for' block gets rendered where it is defined as well as where it is used?
19:39 JiYu think will have a deeper look at pubsub at redis
19:39 sri nono, you can theoretically actually steal a file descriptor from one process and pass it to another
19:40 sri it's black magic though
19:40 JiYu umm never did that before. sounds dirty but will have a look into it :)
19:41 sri NOOOO! ;p
19:41 JiYu haha. just for curiosity :)
19:41 xaka joined #mojo
19:42 sri search for sendfd
19:44 zipkid https://gist.github.com/2845747  this should make my question clear.... i get 2 x the jquery include
19:44 sri it even works different depending on platform
19:45 sri zipkid: if you use it in an expression
19:46 sri <%= vs <%
19:46 * sri waits for facepalm
19:47 zipkid is just 'oh, thanks' satisfactory?
19:47 sri ok :/
19:48 JiYu yeah seems to be possible, but to make this transaction safe, with references to save in a db to know where the transaction is at the moment and that not another process is "stealing" it. would be a lot of hacking
19:49 JiYu definitely try a different way before that...
19:50 sri reminds me that i wanted to try sendfd for hypnotoad to get rid of the accept mutex
19:52 sri but portability seems like huge problem
19:59 mire joined #mojo
20:19 preflex_ joined #mojo
20:29 jnap joined #mojo
20:49 jnap joined #mojo
21:27 sri http://twitpic.com/9rf3y9/full # sooo pretty
21:31 sri TM2 seems dead again
21:35 tempire someone with python experience needs to add ep syntax highlighting to pygments, so github shows it properly.
21:36 * tempire pokes python people
21:36 tempire I know you're in here.
21:36 * tempire peers
21:55 trek1s joined #mojo
22:01 * sri seconds that
22:15 balazs joined #mojo
23:29 lukep joined #mojo
23:30 jnap joined #mojo

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