Camelia, the Perl 6 bug

IRC log for #mojo, 2011-11-17

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

All times shown according to UTC.

Time Nick Message
00:00 crab it was only used by the useragent, right?
00:01 sri yes, or more specifically Mojo::IOLoop->connect
00:13 gabriel joined #mojo
00:17 sri http://northstudio360.yb.nl/nimmobay/​embed/YBPlayerLite.swf?c=config_embed
00:17 sri wow
00:32 SmokeMachine joined #mojo
00:49 sri wonder how hard such a 360 degree movie player would be with html5
01:07 xaka joined #mojo
01:35 gshank joined #mojo
01:38 jnap joined #mojo
02:38 D4RK-PH0ENiX joined #mojo
02:42 xaka joined #mojo
03:02 tempire dude
03:02 tempire I appreciate that it's the male symbol that directs movement
03:02 tempire that's right!
03:09 mire joined #mojo
03:12 crab yes, if you grab hold of someone's "male symbol", you can pretty well direct their subsequent movements.
05:34 chansen joined #mojo
05:44 alnewkirk joined #mojo
06:03 chansen_ joined #mojo
06:35 koban joined #mojo
06:44 gabriel_ joined #mojo
06:48 mire_ joined #mojo
06:59 batman joined #mojo
07:12 Vandal joined #mojo
07:19 Foxcool joined #mojo
07:29 sugar joined #mojo
07:45 xoma joined #mojo
08:12 maluco joined #mojo
08:15 gabiruh joined #mojo
08:16 MojoGuest434 joined #mojo
08:16 MojoGuest434 From: http://blog.kraih.com/perl-is-ready-for-html5 (27 hits)
08:20 arthas joined #mojo
08:26 baton8 joined #mojo
08:29 berov joined #mojo
08:55 judofyr joined #mojo
09:03 cosmincx joined #mojo
09:14 kjeldahl joined #mojo
09:18 bj5004 joined #mojo
09:20 smpb joined #mojo
09:25 spleenjack joined #mojo
09:53 spleenjack joined #mojo
09:53 MojoGuest314 joined #mojo
09:53 MojoGuest314 From: http://blog.kraih.com/perl-is-ready-for-html5 (28 hits)
10:00 Eugene joined #mojo
10:10 Alias joined #mojo
10:16 marcus crab: stay away from my male symbol!
10:24 elb0w joined #mojo
11:08 * sri yawns
11:21 osprey joined #mojo
12:01 marcus good morning, sri
12:05 sri good morning, marcus
12:05 batman left #mojo
12:08 batman joined #mojo
12:12 Psyche^ joined #mojo
12:24 kjeldahl I need to gracefully take down a running morbo server, for implementing automatic upgrades and restarts, on a Windows server nonetheless. Any suggestions on how to convince morbo to shut down gracefully, either through a special route or whatever?
12:33 Sjors kjeldahl: not an answer to your question but it might help: did you know Hypnotoad, another Mojo daemon, implements automatic upgrades by graceful take-over by itself?
12:34 Sjors kjeldahl: if you want to restart a running Hypnotoad daemon, just re-run it; the old one will stop listening but will finish its current requests, and the new one will immediately start listening and take over all new requests
12:40 kjeldahl Sjors: I do not think Hypnotoad runs on Windows...
12:41 Sjors kjeldahl: hmm, I don't know, sorry
12:42 kjeldahl Sjors: I think it does not (remember vaguely from looking at the available daemons earlier). Windows, it's not like a box of chocolate, you always not what you will not get... ;-)
12:42 Sjors hehehehe
12:42 purl hehehehehe
12:43 kjeldahl always KNOW what you ....
12:43 Sjors ah nope
12:43 Sjors die "Hypnotoad not available for Windows.\n"
12:43 Sjors if $^O eq 'MSWin32' || $^O =~ /cygwin/;
12:43 Sjors too bad :(
12:43 kjeldahl Yeah, sounds familiar...
13:02 sri morbo was specifically designed not to restart gracefully
13:02 noganex_ joined #mojo
13:02 sri it's for development only
13:03 sri i suppose you could build your own windows optimized hypnotoad somehow by reusing Mojo::Server::Daemon the same way
13:03 sri not something we will include in core though
13:08 MojoGuest685 joined #mojo
13:08 MojoGuest685 From: http://blog.kraih.com/perl-is-ready-for-html5 (29 hits)
13:09 kjeldahl sri: Yes, I\
13:10 SmokeMachine left #mojo
13:12 kjeldahl I do not see the why quitting/restarting gracefully was specifically not designed into a development only server, but that's just me. It does make using Mojo on windows more difficult though, but that could be a good design choice by itself I guess.
13:14 sri for a good user experience morbo needs to restart as quick as possible, graceful restart feature would only get in the way there
13:19 crab and really, nothing should encourage you to use morbo as anything BUT a development server
13:20 kjeldahl crab: So which server on Windows are you suggesting? That's the idiotic choice I'm facing...
13:22 kjeldahl sri: A graceful shutdown, when needed, should not get in the way of anything. I want it to exit; whether it takes a second or ten does not matter. It matters for restarts, but I want the damn thing to stop running so we can do updates, and then restart later...
13:23 Foxcool joined #mojo
13:25 abra joined #mojo
13:29 sri kjeldahl: it gets in the way of maintainability, more code to maintain that's hard to test and like crab said, it only encourages people to use morbo in production
13:29 sri there is Mojo::Server::Daemon
13:29 sri "MOJO_MODE=production ./myapp.pl daemon"
13:30 sri it's the core of both, morbo and hypnotoad
13:31 sri so all features or hooks to implement features are there
13:32 sri i personally don't consider windows a deployment target, but if you do feel free to build an alternative to hypnotoad and morbo around Mojo::Server::Daemon
13:40 Kovensky joined #mojo
13:42 abra joined #mojo
13:42 diegok s/ a deply.+?rget// # :-p
13:42 diegok s/ a deploy.+?rget// # :-p
13:43 sri it is a target for development
13:44 kjeldahl sri: Thanks, I'll look into the M::S::Daemon . As for building Windows specific code for anything, it's not on my current personal development plan, so hopefully some other sponsor will step forward later. :-)
13:44 * sri wonders if he didn't understand a joke diegok was trying to make
13:45 sri the .+? is confusing
13:46 diegok "i personally don't consider windows, but... (as deployment, as development, as anything that involve computers) :-p
13:46 sri ok, then i got it ;p
13:47 marty good morning gang
13:47 diegok but, last week I was defending the position of making ruby gems as much compatible to windows as possible at the rubyconf (buenos aires) :-p
13:48 diegok oh!, and talking about mojo to all the rubysts around :-p
13:50 kjeldahl sri: M::S::Daemon with a route that calls exit works just fine; thanks!
13:51 kjeldahl sri: Eh, which gets me nowhere of course without additional work. Oh well.
13:52 sri kjeldahl: Mojo::IOLoop->singleton->max_connections(0)
13:53 sri server will stop accepting connections and gracefully shutdown once all connections end
13:54 sri you can also use a timer to force a shutdown if graceful shutdown takes too long
13:54 sri Mojo::IOLoop->timer(300 => sub { exit });
13:55 kjeldahl Thanks, will try.
13:59 judofyr diegok: most maintainers of Ruby gems don't have easy access to Windows though :/
14:03 spleenjack joined #mojo
14:13 Eugene joined #mojo
14:15 osprey left #mojo
14:31 smpb_ joined #mojo
14:33 MojoGuest84 joined #mojo
14:33 MojoGuest84 From: http://news.ycombinator.com/item?id=1277067 (123 hits)
14:35 Sjors hmm
14:36 Sjors it seems like when I output something with "%==" in a Mojo template, Mojo inserts a newline right after it
14:36 Sjors is there some way for me to make sure this doesn't happen? :/
14:42 Sjors %== "Test" \n %== "Foo" \n %== "Bar" prints everything on its own line; a fix:
14:42 Sjors <%== "Test" %>\
14:42 Sjors <%== "Foo" %> \     (etc)
14:43 metaperl joined #mojo
14:52 mire_ joined #mojo
15:02 koban left #mojo
15:03 kjeldahl sri: The max_connections(0) hack in a route does not seem to work. It stops the server (doesn't serve anything), but it does not exit. Any addition hints?
15:04 kjeldahl An explicit call to exit after it does not work either.
15:04 GitHub103 joined #mojo
15:04 GitHub103 [mojo] kraih pushed 1 new commit to master: http://git.io/j4ue0w
15:04 GitHub103 [mojo/master] fixed small IPv6 portability issue in Mojo::IOLoop::Client - Sebastian Riedel
15:04 GitHub103 left #mojo
15:07 kjeldahl Gotta run. I'll ask again tomorrow, unless I find something in the meantime...
15:08 sri umm
15:08 sri perl -Ilib -Mojo -E'a("/" => sub { Mojo::IOLoop->singleton->max_connections(0); shift->render_text("works")})->start' daemon
15:08 sri it works flawless
15:13 crab maybe not on windows?
15:14 crab though it's hard to believe that "exit" doesn't work right on windows.
15:14 crab (but i'll believe it if i have to!)
15:17 Eugene1 joined #mojo
15:18 crab heh, "Scalatra"
15:21 sri what the hell
15:22 sri we have dozens of tests containing $ENV{REQUEST_URI} but not a single one actually tests it
15:23 sri i'm gonna throw that apache only shit out
15:24 sri also bye bye mod_rewrite tests… i'm convinced they are garbage
15:25 metaperl|2 joined #mojo
15:25 * sri wants to strange the person who gabe him those tests cases
15:25 sri *strangle
15:26 sri argh
15:26 sri *gave
15:44 fmerges joined #mojo
15:45 fmerges joined #mojo
15:56 GitHub23 joined #mojo
15:56 GitHub23 [mojo] kraih pushed 1 new commit to master: http://git.io/LN5OOw
15:56 GitHub23 [mojo/master] removed incomplete mod_rewrite tests from request_cgi.t - Sebastian Riedel
15:56 GitHub23 left #mojo
15:57 crab at one point, i used to be rather expert at mod_rewrite and its interaction with other things.
16:00 sri crab: you might want to join the list discussion then :)
16:01 sri and possibly save the day with a great solution
16:02 crab what i learned from my time with mod_rewrite was: when people start asking questions, head in the other direction.
16:02 sri there is so much garbage in apache cgi environments
16:02 sri heh
16:02 sri that does sound wise indeed
16:07 abra joined #mojo
16:18 kaare joined #mojo
16:22 jnap joined #mojo
16:25 Eugene joined #mojo
16:35 Maximdd123 left #mojo
16:37 xaka joined #mojo
16:49 gshank joined #mojo
16:56 mire_ joined #mojo
17:01 MojoGuest688 joined #mojo
17:16 mercutioviz joined #mojo
17:30 marcus When The going gets tough.
17:35 grim_fandango joined #mojo
17:38 abra joined #mojo
17:47 gshank_ joined #mojo
18:09 mercutioviz joined #mojo
18:09 marty The tough GO SHOPPING!
18:33 marcus joined #mojo
18:45 smpb joined #mojo
19:58 batman i'm going to discuss security tomorrow with a coworker, who wants the code to be readable and understandable. reason for this is that he feels that if he's in control of the complete code, it's more secure...
19:58 batman what i try to convince him of, is that it's better to stand on the shoulders of others instead of trying to fix the same issues over again...
19:59 batman got any input on how to convince him? :P
19:59 marcus http://imgur.com/0FAqT
20:00 batman got an example: he asked me how many lines of code net-server was, and i told him it was about 9000 lines long, he didn't respond well... he wants to know how the data flows from the client and to the application core, to be sure it doesn't do anything insecure...
20:00 batman marcus: haha!
20:00 batman i thinkt that was for me? :)
20:00 marcus batman: for your coworker ;)
20:00 batman yeah, yeah
20:00 batman excellent
20:01 batman but i think he's looking for some hard facts...
20:01 marcus security is hard, let's go shopping
20:01 sugar joined #mojo
20:01 marcus of course a big closed source code is much more vulnerable than a hardened open source code base that has had actual people trying to break it.
20:02 batman but i mean... what's the worst thing that can happen, until the data actually appears in your controller?
20:02 batman marcus: yeah, i tried explaining him that as well...
20:02 marcus batman: ask him if he thinks security through obscurity is a good idea.
20:02 batman marcus: i told him that plack, net-server, mojolicious ++ was used by thousands, so if something was messed up, it would probably already be fixed
20:03 batman marcus: i'm afraid of the answer :P
20:03 marcus there's already been vulnerabilities detected in the mojolicious code base and quickly fixed in the early days. which means that there are actual people looking for it. which means that it's already better than writing it yourself.
20:04 batman another thing... am i wrong when i say that probably 99% of all mistakes happen -after- the frameworks has handled the data over to the code in the controller?
20:04 marcus http://blog.kraih.com/mojolicious-​116-emergency-release-please-upgr
20:04 marcus batman: yupp
20:04 batman yep. i saw that.
20:04 batman i'm loving it when people twitter about mistakes, hickups and vulnerabilities
20:04 marcus that was an issue in the static server tho
20:05 batman like github, dotcloud, cloudflare, ...
20:05 batman still. it was an issue. and it got fixed :)
20:06 batman i'm guessing my estimate of 99% isn't accurate enough though... i'm guessing more like 99.9999% :)
20:06 batman tomorrow i'm going to explain to him how the data flows though a socket to the core of the app (the controller) ...
20:06 marcus There are a *lot* of XSS holes out there.
20:07 marcus at least mojolicious mitigates that a bit by escaping all your template data by default.
20:07 batman i'm going to make some big gaps in the drawing, where the kernel and perl does its handling......
20:08 batman i think one of the good things about mojolicious is that at least there's no xs holes, since it's pure perl right?
20:08 batman (yeah, i read your "xss")
20:08 sri phew, almost 1000 lines ofcode removed during the last few weeks, feels good :)
20:08 batman sri: yeah, that always feels good...
20:08 marcus judofyr who works for me found a XSS exploit in bokskya and used it to create a SSO mechanism that redirects people to their login then posts the password back to him :D
20:09 marcus so that he could use the credentials to convert their books to kindle :D
20:09 batman sweet
20:09 marcus sri: does that include the async resolver? :)
20:10 sri it does
20:10 sri it's about 300 of those ;p
20:11 sri oh, and i did not include stuff i deprecated, that's another 300 or so
20:11 marcus \o/
20:12 sri gotta love perls expressiveness, lines just melt away when you pay a little attention
20:14 sri 11331 lines in lib and 16280 lines in t
20:16 marcus '"g. I don't even see the code. All I see is blonde, brunette, redhead."
20:17 sri Oo
20:17 marcus -- the matrix
20:18 marcus https://metacpan.org/favorite/leaderboard # dancer is ahead now.
20:18 sri QUICK SOMEONE MAKE 10 NEW GITHUB ACCOUNTS!
20:19 sri how can i not recognize a matrix quote… WHAT IS WRONG WITH ME? :o
20:21 sri metacpan +1 really went downhill once they started allowing non cpan ids to vote
20:21 Shaeto joined #mojo
20:23 marcus it's still just a prototype, it's not even linked up yet afaik.
20:25 marcus suggested a new leaderboard that you might do well on - https://github.com/CPAN-AP​I/metacpan-web/issues/392 :p
20:25 sri lol
20:26 sri tags would also be cool
20:27 sri many fun statistics
20:28 batman starman is one request pr child, or..?
20:28 batman since it's built on net-server...
20:28 marcus batman: starman is preforking
20:28 batman and?
20:28 sri but it is one request per child
20:28 marcus ah, you mean it's not async.
20:29 batman so if you have a pool of 20 children, you can handle 20 requests at a time...
20:29 sri correct
20:29 marcus This module actually includes the longer alias and an empty subclass HTTP::Server::PSGI::Net::Server::PreFork for those who like to type more ::'s.
20:29 batman but hypnotoad can handle x number of requests pr child, but if course each code block will...uhm...block... :)
20:29 marcus <3 miyagawa
20:30 batman s/if/of/
20:30 marcus batman: yes
20:30 sri you can configure hypnotoad for 1 request per child too though
20:31 marcus urgh, so tired of probabilities.
20:31 batman so if i want to handle many connections, i need to apply my POE mojo to mojolicious?
20:31 marcus hidden markov chains in AI-Class this week
20:31 marcus batman: async only makes sense if you do async stuff
20:31 marcus like waiting for http responses
20:31 batman so the answer is yes.
20:31 marcus or using anyevent modules
20:32 batman i don't like AE :(
20:32 batman https://metacpan.org/module/​AnyEvent::TFTPd#DESCRIPTION <--
20:32 marcus batman: you can use it with EV tho, which integrates into mojolicious
20:32 Shaeto i have started to polish startup sub and still don't understand why ->route('/mypath')->namespace('my::class::path')-> doesn't work. if i embedd namespace => to every ->to( ..) construction it works fine but why route node can't apply namespace to all chilldren ?
20:33 * marcus just noticed https://metacpan.org/module/EV::ADNS
20:34 marcus sri: might be a reasonable alternative for people still wanting async dns with mojolicious?
20:34 sri there is also AE::DNS
20:35 sri i might actually have added the AE backend to mojolicious core if the 6.1 release didn't happen
20:36 marcus I don't understand why you shouldn't use async with UDP?
20:36 marcus AE author on crack?
20:36 sri context?
20:37 marcus sri: https://metacpan.org/source/MLEHMANN/An​yEvent-6.1/lib/AnyEvent/Handle.pm#L573
20:37 batman marcus: it was after MLEHMANN looked at my module...
20:37 sri marcus: actually makes sense in that context
20:38 batman sri: then please enlighten me...
20:38 xaka Shaeto: long time ago i was catched by the same "issue" :) sometimes mojo works in a bit different way, not how you expect if u dont read docs. Your solution is ...->to(namespace => ...)-><children>
20:39 sri the module was designed for streams, not messages
20:39 batman sri: but it worked nicely.
20:39 xaka Shaeto: just replace your namespace(...) with to(namespace => ...)
20:39 batman he changed the code + added the comment so my code *should* not work...
20:39 sri i've been lazy and used my stream abstraction layer for udp too, doesn't make it right
20:39 sri always felt wrong
20:40 xaka sri: is doesn't fit very well for udp?
20:40 sri of course he's a jerk for actively preventing it
20:40 sri but it makes sense
20:40 Shaeto xaka: that will work but i have to put namespace to every ->to (..) :(
20:40 Shaeto or just ->to for upper node ?
20:40 batman i'll buy you beer sometime, if you would like to explain it to me :/
20:41 xaka Shaeto: just ->to as i said
20:41 marcus sri: I bought a Beck for miyagawa this weekend ;-)
20:41 Shaeto damn that works, thanks, very unclear in docs
20:41 xaka Shaeto: children = ...->to(namespace => ...); children->route(...), ...
20:41 marcus beck's even
20:42 Shaeto thank you xaka
20:42 xaka sri: one more dude catched ^^^ )
20:42 sri batman: in fact i think IO::Socket::INET is wrong for allowing you to do UDP
20:42 xaka put the cross (mark) on to the wall :-p
20:42 batman sri: then i really think you need to give up on me :P
20:42 marcus udp isn't really a socket, is it..
20:42 xaka why not?!
20:43 xaka it is
20:43 marcus it's connectionless
20:43 sri it is
20:43 sri both
20:43 batman marcus: still a socket
20:43 purl Socket to me! or pretty nasty
20:43 xaka marcus: SOCK_DGRAM      Supports datagrams (connectionless, unreliable messages of a fixed maximum length).
20:43 sri udp sockets are bound to a port
20:43 marcus true
20:43 xaka connectionless but still socket
20:44 * marcus goes back to his markov chains and laplacian smoothing
20:44 xaka free stanford courses? ;)
20:44 marcus yeah
20:44 sri using them like a stream only works when you don't care about host/port information… which is generally an indicator that you're lazy and doing it wrong
20:45 xaka there is 5 more have been added today (or yesterday)
20:45 xaka for a 2012
20:45 marcus nice
20:45 marcus I'm phearing midterms this weekend
20:45 marcus for ai-class.com
20:47 xaka sri: you should mention somewhere in docs that if you want to use different namespace for all childrne, you have to do "to(namespace => ...)" instead of "namespace(...)"
20:48 sri xaka: it is mentioned
20:48 xaka Shaeto catched like i i'm long time ago
20:48 sri http://mojolicio.us/perldoc/Mojolicious/Gui​des/Routing#Route_to_class_28_namespace_29
20:49 Shaeto $r->namespace('MyApp::Controller');
20:49 Shaeto from the doc
20:49 xaka aha
20:49 sri read the description too!
20:49 Shaeto doesn't work for sub routes
20:49 sri IN THE WHOLE APPLICATION!
20:50 xaka somehow it's unclear, sri. we dont push you, just let you know, dont be so angry when people trying help you
20:50 sri i don't care about problems, just solutions
20:51 marcus patches welcome?
20:51 purl i think patches welcome is always true or unless for search.cpan.org or swahili for "Put up or shut up."
20:51 xaka add one more section: if you want to change namespace for all children, use "to(namespace => ...)"
20:51 Shaeto +1
20:51 purl 1
20:51 xaka "From time to time you might want to dispatch to a whole different namespace." - doesn't say that namespace inherited
20:51 sri that's implicit
20:52 Shaeto actually no :)
20:52 sri *all* to() values are inherited
20:52 sri ALWAYS
20:52 marcus I'd rather inherit a pony.
20:52 * sri demans a unicorn
20:52 sri *+d
20:52 marcus I wonder how unicorn tastes
20:52 marcus probably like chicken
20:53 xaka doc is really big (and well written) so you cant remember everything, especially that everything inherited )
20:53 marcus xaka: I suggest a pull request
20:53 xaka then tastes lake a horse, magic horse
20:53 xaka *they
20:53 marcus horse is pretty yummy.
20:53 marcus we have horse shashlik sometimes. :)
20:53 xaka magic means coca, marijuana, ... P:)
20:53 sri tastes like rainbow
20:54 marcus horse shashlik where you don't need to bring beer \o/
20:54 xaka mmmm...shashlik
20:54 xaka what the analog in english?
20:55 marcus barbeque I guess
20:55 xaka noooo, it's to common
20:55 xaka *too
20:55 sri also, why the hell do people switch namespaces so frequently?
20:55 marcus http://en.wikipedia.org/wiki/Shashlik
20:55 xaka usually barbeque means the beer )
20:55 sri what is wrong with you people! ;p
20:56 batman this used to be my module, before leont *fixed* it: https://metacpan.org/module/AnyEvent:​:Handle::UDP#BACKWARDS-COMPATIBILITY
20:57 sri the whole idea that you connect a udp socket is a little irritating imo
20:58 sri you bind, send and receive
20:58 batman that's strange... i think he was at the yapc. too bad i didn't recognize him :/
20:58 batman sri: agree
20:58 marcus wonder if he was at lpw, face seems familiar
20:59 batman marcus: makes sense
20:59 purl makes sense is, like, http://xrl.us/bjq3kq
21:01 batman sri: i only implemented the server part, and my ::UDP module only had the "on_read" event.
21:02 batman also: "Does not make any sense for UDP: on_prepare, on_connect, on_connect_error, keepalive, oobinline and on_eof."
21:02 sri i think the best non-blocking udp api i've seen so far was in node.js
21:02 sri http://nodejs.org/docs/v0.6.1/api/dgram.html
21:04 sri IO::Socket::INET really encourages bad udp apis, doing it properly is really hard, try sending to a ipv6 host...
21:05 batman really annoying... i don't get Mojo/Server/Hypnotoad.pm *reading more code over again*
21:06 marcus ALL GLORY TO THE HYPNOTOAD
21:06 sri ALL GLORY TO THE HYPNOTOAD!
21:07 sri hypnotoad is a little tricky due to zero downtime restarts
21:08 sri HERE BE DRAGONS
21:09 batman but is all the read/write socket stuff handled inside IO::Poll...?
21:10 gabriel__ joined #mojo
21:10 batman nevermind...
21:10 batman i don't bother to understand everything :P
21:10 batman i'm going to kick ass in the discussion tomorrow nomatter :)
21:11 marcus batman: ioloop
21:11 batman i don't see it...
21:11 batman marcus: don't bother. i'm off :)
21:12 * marcus doesn't bother
21:12 marcus time to find my robot!
21:12 batman catch it before it catches you :)
21:12 marcus http://nordaaker.com/docs/skitch/​Intro_to_AI_--20111117-221246.jpg
21:23 lammel2 joined #mojo
21:26 sri hmm, github bot fail
21:28 sri anyway, new namespace description http://mojolicio.us/perldoc/Mojolicious/Gui​des/Routing#Route_to_class_28_namespace_29
21:28 sri i've also promoted the nested routes description with stash inheritance
21:38 GitHub94 joined #mojo
21:38 GitHub94 [mojo] kraih pushed 1 new commit to master: http://git.io/5b52cA
21:38 GitHub94 [mojo/master] bridges are more important than waypoints - Sebastian Riedel
21:38 GitHub94 left #mojo
21:41 * marcus signs up for HCI and game theory classes in january . Thanksxaka
21:42 * sjn is impressed by marcus' seemingly bottomless access to time
21:42 sjn :)
21:43 marcus sjn: I'll be done with AI-class by then
21:43 marcus and hci should be much easier for me ;-)
21:45 sjn assumption is the mother of all fuckups ;)
21:45 sjn but sure, ome class should be doable if you spend your time right :)
21:45 sjn now*
21:45 sjn uhm. one*
21:46 sri once he's done he can make the AI do all the real work for him
21:46 marcus sri++
21:47 marcus sjn: yeah, I tried for two classes this fall and failed, so I might drop out of one after christmas as well, but I'm not sure which of game theory and hci I want more, so I'll start both
21:48 sjn game theory can be interesting if you can handle the math
21:48 sjn hci is something you can play with at hackerspaces :D
21:49 sjn kinect ftw
21:51 marcus sjn: well, I'm barely handling the math for aiclass at the moment :-) But I had something like HCI the first year of college. It's like 15 years ago tho, so I hope the field has moved a long a bit.
21:52 marcus but at least most of the math I'm struggeling with now is calculation of probabilities, so it should be a nice preparation for game theory ;-)
21:52 marcus struggling
21:54 metaperl joined #mojo
21:55 marcus sri: shouldn't Mojolicious::Plugins pod say something about which events run_hook supports out of the box?
21:55 marcus want a pull request for that?
21:57 sri marcus: Mojolicious.pm does that
21:58 sri it's not really events in the case of Mojolicious::Plugins
21:59 marcus sri: true, the pod I was looking at has changed to emit_ stuff in any case :)
21:59 marcus which is a bit clearer
22:00 marcus was looking at an older codebase
22:00 sri ah
22:05 marcus sri: what do you think about X-Proxy-Path header? For my plugin.
22:06 sri why not X-Forwarded-Path?
22:07 marcus sri: thought about that, but it's not the full path. for a url /foo/bar/baz where myapp serves /bar/baz proxy-path is /foo
22:07 sri think that's what you said last time
22:07 marcus I'd expect -forwarded-path to be /foo/bar/baz
22:07 sri true
22:08 sri X-Forwarded-Script :)
22:08 marcus looks like catalyst uses X-Forwarded-Path
22:08 marcus with {using_frontend_proxy}
22:09 marcus there's also https://metacpan.org/module/Catal​yst::TraitFor::Request::ProxyBase
22:10 marcus maybe just support  X-Request-Base  then
22:12 * sri shrugs
22:31 lammel2 left #mojo
22:44 tholen_ joined #mojo
22:48 metaperl joined #mojo
22:50 SmokeMachine_ joined #mojo
22:53 SmokeMachine___ joined #mojo
23:07 marcus I miss perlrocks :-/
23:07 marcus can't we get it back without the translations?
23:08 marcus I find it funny that the Mojolicious version of https://metacpan.org/source/BOBTFISH/Catal​yst-TraitFor-Request-ProxyBase-0.000005/li​b/Catalyst/TraitFor/Request/ProxyBase.pm was 2 lines of actual code.
23:08 sri it was broken, refused to stay connected
23:09 sri feel free to take it over
23:09 marcus https://metacpan.org/source/MRAMBERG/​Mojolicious-Plugin-RequestBase-0.01/l​ib/Mojolicious/Plugin/RequestBase.pm
23:09 sri marcus++
23:09 marcus I might, but not now.
23:09 marcus bedtime right now :)
23:09 sri give it AI! :D
23:09 marcus You *will* phear it :)
23:09 marcus I might write it as a hubot plugin :p
23:10 marcus that way I can use it on nordaaker as well
23:10 marcus but it would be a bit sad if perlrocks ran coffeescript
23:10 sri haha, that would be cool
23:10 sri the first part, not the second
23:11 marcus nite
23:11 sri nn
23:20 Shaeto left #mojo
23:30 mire_ joined #mojo
23:40 GitHub172 joined #mojo
23:40 GitHub172 [mojo] kraih pushed 1 new commit to master: http://git.io/qBzV_w
23:40 GitHub172 [mojo/master] small routing guide tweaks - Sebastian Riedel
23:40 GitHub172 left #mojo
23:51 SmokeMachine joined #mojo
23:57 mire_ joined #mojo

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