The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2014-10-15

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

All times shown according to UTC.

Time Nick Message
00:03 D4RK-PH0ENiX joined #mojo
00:08 sri ouch, looks like the ssl 3.0 attack is pretty serious http://googleonlinesecurity.blogspot.ru/2014/10/this-poodle-bites-exploiting-ssl-30.html
00:19 powerman sri: interesting, but unlikely will affect mojo
00:20 powerman in mojo https clients attacker unlikely will be able to control body of requests and be able to send a lot of requests
00:20 powerman in mojo https servers actual https usually terminated on nginx frontend and mojo work by http only, so this issue should be mitigated in nginx
00:21 Nei thanks for your analysis ;)
00:23 powerman unlikely doesn't mean never, and it's security after all :) I'd like to see TLS_FALLBACK_SCSV implemented in IO::Socket::SSL someday, but I'll be surprised if it happens soon
00:31 Averna joined #mojo
01:08 cpan_mojo Mojolicious-Plugin-HostMeta 0.15 by Nils Diewald - http://metacpan.org/release/AKRON/Mojolicious-Plugin-HostMeta-0.15
01:09 Mso150 joined #mojo
01:25 klapperl joined #mojo
01:25 cfedde joined #mojo
01:56 nicomen joined #mojo
02:02 r0b3rt_ joined #mojo
02:11 r0b3rt joined #mojo
02:14 nicomen joined #mojo
02:15 woz joined #mojo
02:36 zivester joined #mojo
02:48 noganex_ joined #mojo
03:01 davido___ joined #mojo
03:17 woz joined #mojo
03:22 basic6 joined #mojo
03:51 KCL joined #mojo
04:23 irq joined #mojo
04:31 Adura joined #mojo
04:51 preaction joined #mojo
05:56 dod joined #mojo
05:59 Adurah joined #mojo
05:59 dod joined #mojo
06:01 franzkafka joined #mojo
06:02 franzkafka Assume that I want to route to a login page if not authenticated, and a different page if you are. Is this best accomplished with under or with a callback?
06:03 franzkafka And this is for / (i.e., mypage.com/)
06:04 dod joined #mojo
06:06 dod joined #mojo
06:07 Adura joined #mojo
06:10 hesperaux joined #mojo
06:14 neyasov_ joined #mojo
06:15 Insane if controller calls some helper and code inside helper fails, how to force helper to render exception page and not render anything from controller?
06:22 franzkafka I want index to either take me to the login page or the dashboard (if logged in)
06:55 rawler joined #mojo
06:56 irq joined #mojo
07:10 dp_ joined #mojo
07:29 fhelmber_ joined #mojo
07:37 trone joined #mojo
07:50 woz joined #mojo
07:54 denis_boyun joined #mojo
07:58 denis_boyun joined #mojo
08:17 Mso150 joined #mojo
08:26 zivester joined #mojo
08:45 zivester joined #mojo
08:55 Mso150 joined #mojo
09:11 arthas joined #mojo
09:36 basiliscos joined #mojo
10:27 neyasov joined #mojo
10:33 Vandal joined #mojo
10:40 aleksey joined #mojo
10:43 d4rkie joined #mojo
10:50 D4RK-PH0ENiX joined #mojo
11:05 D4RK-PH0ENiX joined #mojo
11:10 basiliscos joined #mojo
11:43 Insane have this code: http://pastebin.com/6ZeHbU6z if inserted into html template, render fails ith error: Can't find string terminator '"' anywhere before EOF at template issues/issue.html.ep line 62. 62 is the last file line number. latest Mojolicious. what the... ? o.O
11:44 Insane oh... my mistake
11:49 neilhwatson joined #mojo
11:52 ryozi joined #mojo
11:52 marty joined #mojo
11:53 marty o/
12:00 mishantil \o
12:01 mishantil Btw; anyone here live in/near Berlin?
12:01 mishantil Know of anything particularly interesting happening around the 20th of november?
12:06 marty postgres plperl - yum
12:08 marty postgres has just enough json capability to make a conversion from mongo fairly painless. *I think*
12:17 Akron joined #mojo
12:27 Akron Hope "SSL connect attempt failed error:14077410:SSL routines:SSL23_GET_SERVER_HELLO:sslv3 alert handshake failure" is a cpan/puddle issue and doesn't mean there's something wrog on my side. Happens with "mojo cpanify" currently.
12:46 ignacio_ joined #mojo
12:59 tbushell_ joined #mojo
13:00 tbushell joined #mojo
13:00 tbushell_ joined #mojo
13:02 tbushell joined #mojo
13:05 markov Akron: http://it-beta.slashdot.org/story/14/10/15/000239/google-finds-vulnerability-in-ssl-30-web-encryption
13:05 tbushell_ joined #mojo
13:05 zivester joined #mojo
13:06 markov New systems will not use SSL3, afaik is the Pause server quite old.  So, if patches disable your SSL3, then you cannot access those servers anymore.
13:07 alanminter joined #mojo
13:09 alanminter perl -MPOSIX=strftime -E"say strftime(q{%T}, localtime)"
13:09 * alanminter wonders if that works for any Windows users
13:10 alanminter appears it fails on Strawberry 5.20.1/Win8.1
13:10 * alanminter likes http://strftime.net/
13:13 Adura Works on Cygwin's.
13:14 markov alanminter: according to the Linux manpage, %T comes from the "Single UNIX Specification" extension to the ANSI.  It is included as standard for POSIX:1003-2003
13:14 alanminter ah, thanks markov
13:14 alanminter That'll be what the 'SU' signifies :)
13:16 alanminter (Some other 'SU' features kind of work, but quite hit-and-miss)
13:17 sri batman: something is wrong with convos
13:18 neyasov_ joined #mojo
13:31 HtbaaPi joined #mojo
13:36 HtbaaPi joined #mojo
13:38 HtbaaPi hello
13:38 sri o/
13:38 HtbaaPi I'm having some issues with Mojo::UserAgent post calls resulting in premature connection close errors
13:38 HtbaaPi I've already set the timeout to 0 and have read the part about the error in the FAQ
13:39 HtbaaPi I run my app with hypnotoad and a part of it (a form) calls my REST API when the form is being submitted.
13:39 HtbaaPi usually this works, but it seems that half of the time this premature connection error occurs
13:41 HtbaaPi I also see messages from time to time that a hypnotoad worker has no heartbeat so it's restarting. Not sure if these are related though
13:42 HtbaaPi ah, the api call from within my controller is also a blocking one.
13:47 HtbaaPi my controller is actually uploading stuff as well so I guess it's possible it's taking too long. Perhaps time to read up on non-blocking at https://github.com/kraih/mojo/wiki/Blocking-vs-non-blocking-101
13:52 tbushell left #mojo
14:02 franzkafka joined #mojo
14:03 franzkafka Question: Does anyone else have to stop/start morbo everytime they make a change?
14:03 franzkafka For example, if I change an HTML file, I have to stop/start morbo to see the change.
14:04 HtbaaPi franzkafka: I used to have that a long time ago when on windows
14:05 franzkafka I'm using CentOS 7, and for every change I make I have to do this. I change my SQL query, I have to stop/start. I change any CSS/HTML, I have to stop/start.
14:07 hernan franzkafka: how are you running morbo ?
14:07 hernan it should reload itself when you tell it which dirs to watch
14:07 franzkafka morbo script/myapp
14:08 hernan type morbo
14:08 hernan and read the options
14:09 hernan and then run your mojo again with the proper option to watch directorys
14:10 hernan HtbaaPi: try delaying the response
14:10 hernan render later i think
14:11 HtbaaPi yeah I'm trying that now. But now I'm having issues with performing other uploads inside my callback
14:11 franzkafka awesome, there we go
14:11 franzkafka I wonder why it doesn't just watch lib/ public/ and templates/?
14:12 hernan because why it doesnt watch /bla /ble and /bli also ?
14:12 HtbaaPi hernan: I'm now calling render_later and using Mojo::UA non-blocking. First call is to create a new ticket, then up to 4 other calls are made to upload previous uploaded files. This works, but some file uploads don't seem to come through
14:12 hernan you might not use lib public or template...
14:12 franzkafka because those directories are not created as a default structure when you do a 'mojo generate app AppName'?
14:13 hernan i dont see why it should have hardcoded directories to watch
14:13 hernan i think the current behaviour is whats expected
14:13 hernan you configure it according to your needs, that  is, if you want to use it...
14:14 franzkafka It certainly didn't used to be this way. I did not have to specify what directories to watch after doing a 'mojo generate app'
14:14 hernan why its not enabled by default... whats the correct default ?
14:14 franzkafka the correct default is the same default schema used when generating an app
14:14 franzkafka I see where you're coming from though
14:14 hernan correct default for me might not be what you consider default... thats why i think its correct
14:15 hernan HtbaaPi: you must be makeing some mistake..
14:15 HtbaaPi probably :)
14:15 hernan try using nonblocking with render later
14:15 HtbaaPi I am now :)
14:15 elik joined #mojo
14:15 hernan try to make a big sleep of 10 seconds and then make a request somewhere to test
14:15 hernan see if you can get that working first
14:16 HtbaaPi I don't understand. Where should I put the sleep? Inside the callback?
14:16 hernan yes, its my sugestion for your testing purposes
14:17 HtbaaPi ok, and then see if it still works with the sleep in place? Just tested it now without the sleep call and all attachments have been uploaded. But I'll test it anyway
14:18 hernan HtbaaPi: great then you've made it
14:18 Claudio joined #mojo
14:18 hernan no need for sleep.. it was just a test
14:18 hernan HtbaaPi: ^^
14:19 hernan a test to simulate some long running processing
14:19 Akron joined #mojo
14:19 HtbaaPi ah well, I'm still not sure if it's working properly yet :)
14:20 HtbaaPi another side effect I'm running in to was that sometimes a ticket gets created more than once. Even with a mechanism in place that prevents a user from submitting the form multiple times.
14:20 hernan it shuold..
14:20 Akron markov: That's why I was refering to poodle - but said "puddle", don't know why. I understand the problem, but don't know if it's on my side or on cpan's side.
14:20 hernan HtbaaPi: debug.. add some warnings
14:21 HtbaaPi so I was thinking maybe, with the hearbeat messages as well, something is going wrong with the blocking stuff and workers timing out. Which perhaps causes some retries or something
14:21 HtbaaPi hernan: yeah I'll be logging the submissions now as well. Pesky bug :)
14:22 hernan yeah , you will find that error >)
14:24 HtbaaPi ah I think the missing attachments was due to Morbo not accepting the size of my requests.
14:25 HtbaaPi with the sleep calls put in it also works, so all is well me thinks :)
14:34 sri it appears there's a way to have relatively secure cpan installations, wonder if we should mention that on the frontpage
14:34 marty2 joined #mojo
14:34 sri cpanm --mirror https://cpan.metacpan.org --mirror-only Mojolicious
14:35 sri (If you already have a secure Perl environment with cpanm, you can also use this one-liner.)
14:39 magarwal joined #mojo
14:41 sri i guess this is the best we can do
14:41 sri cpanm --mirror https://cpan.metacpan.org --mirror-only --verify Mojolicious
14:42 sri apparently fastly is responsible for the secure mirror
14:44 asarch joined #mojo
14:44 HtbaaPi hmm this is really weird. For some reason an action is being called multiple times with hypnotoad
14:50 sh4 joined #mojo
14:53 HtbaaPi huh? My non-blocking call works in Morbo. But when I run hypnotoad it simply won't get called and instead the request times out or it retries after 15/16 seconds :S
14:54 sri maybe like this https://gist.github.com/anonymous/ca4ce5ca98727a8e27b4
14:58 Jonis looks good :)
15:01 Akron That's a really nice advice, but maybe overkill for the starting page ... hmmm ...
15:02 good_news_everyon joined #mojo
15:02 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/8mnbKw
15:02 good_news_everyon mojo/master 1ba29bd Sebastian Riedel: mention how to install Mojolicious securely
15:02 good_news_everyon left #mojo
15:04 HtbaaPi would there be a reason why hypnotoad wouldn't be able to run an action inside a controller that uses Mojo::UserAgent in a non blocking way? In Morbo it works, in Hypnotoad the post call by the ua isn't being executed...
15:07 basiliscos joined #mojo
15:09 sri wonder if this would make more sense with two one-liners https://gist.github.com/anonymous/a94782272607490c7134
15:09 HtbaaPi oh wow, didn't expect that... When non-blocking I can't call /api/whatever. It has to include the full domain, e.g. url_for()->to_abs
15:09 sri and get rid of the "$ curl get.mojolicio.us | sh"
15:12 sri jberger, marcus, batman, crab, tempire: thoughts?
15:13 Akron Then the user is forced to use cpanm?
15:13 sri what?
15:14 Akron If you want to get rid of "$ curl", the user needs cpanm already installed, right? Or am I misunderstanding your proposal only promoting the secure way.
15:16 sri did you read the gist?
15:17 alanminter It think it's about using perl rather than sh (which might be mapped to bash)
15:17 alanminter I like that cpanm line
15:17 sri it's about making this more obvious https://gist.github.com/anonymous/0a369768a5c91d2dffea
15:19 sri right now the two one-liners seem rather disconnected, when in reality there are just tiny variations
15:21 Akron Ah - sorry - I just read "get rid of $curl ..." and sidn't see the difference to cpanmin.us.
15:21 davido___ joined #mojo
15:24 good_news_everyon joined #mojo
15:24 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/EAxZ-g
15:24 good_news_everyon mojo/master 0915ed6 Sebastian Riedel: use more similar installation one-liners
15:24 good_news_everyon left #mojo
15:26 sri like that
15:29 good_news_everyon joined #mojo
15:29 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/RTwk1w
15:29 good_news_everyon mojo/master 3240959 Sebastian Riedel: the slash is not necessary
15:29 good_news_everyon left #mojo
15:34 Akron That's nice. But the "And if" sentence doesn't sound like it's finished - may be end with a colon?.
15:34 neyasov_ joined #mojo
15:36 sri batman: convos keeps eating my sentences
15:36 sri something is very very wrong with the server :(
15:40 sri hmm, looking through the irclog... it appears i only see half of all messages today :S
15:40 sri BATMAN!!!!1
15:40 sri where is the batsignal when you need it
15:42 fhelmbe__ joined #mojo
15:48 neyasov_ joined #mojo
15:50 KCL_ joined #mojo
15:51 franzkafka Within Mojolicious::Controller is there any way to get all params as a hash ref?
15:53 alanminter Mojo::Parameters
15:56 magarwal i am using mojolicious + websocket, is there a way that the message can be broadcast to all the clients, except of sending one by one to each client.
15:56 magarwal https://groups.google.com/forum/#!topic/mojolicious/HkUaR6dFfak
15:57 batman the bat is not feeling to good :/
15:57 batman sri: do you have an example?
15:57 batman is it half a sentence or half of the sent messages?
15:58 disputin joined #mojo
15:58 sri batman: messages missing
15:58 sri now it works again
15:58 batman and it doesn't tell you that you lost connection to the server?
15:58 sri but every now and then it stops working completely, i can send messages still, but don't see anything
15:58 Akron batman: Then get well soon!
15:58 batman it should give you a timeout message after a while, if you don't refresh too often...
15:59 sri nope, even if i reload a dozen times
16:00 batman i don't have any other explanation than if the connection is bad...
16:00 sri actually, there are messages still missing from my history
16:00 batman are you at a cafe or some on 3G?
16:00 sri like everything between http://irclog.perlgeek.de/mojo/2014-10-15#i_9513447 and http://irclog.perlgeek.de/mojo/2014-10-15#i_9513628
16:00 sri it's the server 100%
16:00 sri there are messages missing
16:00 sri like right now
16:01 batman i will have a look...
16:02 marcus Anyone have join/ quit logs for that period?
16:04 sri marcus: joins are there
16:04 sri in the logs above
16:05 batman sri: i will try to restart and turn on more logging
16:05 sri and i did not rejoin
16:08 sri btw. i would still like opinions on the installation section https://github.com/kraih/mojo#installation
16:09 tempire I prefer the get.mojolicio.us, personally.
16:09 sri i don't think it works as well now
16:09 tempire Though the secure toolchain addition isn't a bad idea, for the naysayers.
16:10 sri not just naysayers, i think we should be pushing a bit there, to get toolchain secure by default
16:10 magarwal sri, any comments on "i am using mojolicious + websocket, is there a way that the message can be broadcast to all the clients, except of sending one by one to each client."
16:10 marcus I agree with sebastian, we should push perl forward wrt ssl
16:11 magarwal broadcast at once, instead of using the loop
16:11 marcus magarwal: emit
16:11 batman *restart convos* # sorry
16:14 mst magarwal: I don't understand. what would you expect some "broadcast" feature to do other than a loop?
16:15 magarwal mst, https://groups.google.com/forum/#!topic/mojolicious/HkUaR6dFfak
16:16 magarwal mst, the same message has to be delivered to all the clients and as the number of clients increases, loop will take more time and defeat the purpose of realtime updates
16:16 magarwal marcus, thanks
16:17 mst that doesn't answer the question
16:18 mst also you never told me what that client object is
16:18 mst magarwal: please be less vague
16:18 magarwal mst, please refer to the link i sent above
16:18 mst magarwal: the link you sent above doesn't tell me anything
16:19 mst answer the questions, please.
16:19 mibbit_sri joined #mojo
16:19 mst magarwal: again: what *are* those client objects?
16:20 mst surely you can queue the message to be sent to every client and let the ioloop take care of delivering them
16:20 magarwal mst, if there are 10000 clients connected and i am sending realtime updates in a loop, there will be a delay in the message dispatch
16:20 mst goddamnit
16:20 purl goddamnit are you ever online?
16:20 mibbit_sri looks like convos might have bigger problems :(
16:20 tempire magarwal: he's likely wanting to know what other solution you would propose
16:20 mst I don't even know what he's doing now
16:20 mst since he won't answer any of my questions
16:21 marcus joined #mojo
16:21 batman joined #mojo
16:21 mst magarwal: are you trying to avoid being helped so you can tell your boss it isn't your fault? because otherwise, start answering questions please
16:21 jberger_ joined #mojo
16:21 magarwal tempire, if i know the answer why would i post the question
16:21 mst magarwal: you're -assuming- the loop will be too slow
16:21 mst magarwal: and there's some magic trick to not do a loop
16:22 mst magarwal: there's going to have to be 10,000 message sends done *somewhere*
16:22 mst magarwal: that's the architecture you chose, and Mojo can't make that not true
16:22 mst magarwal: however you *can* make it relatively efficient to do so
16:22 mst magarwal: hence me asking - WHAT ARE THOSE CLIENT OBJECTS ?
16:22 magarwal i am asking something like ActivMQ broadcast to a topic does
16:22 * mst wonders why he even bothers
16:22 maybe_sri joined #mojo
16:22 mst magarwal: that would *still* send 10,000 messages
16:22 mst that's the point
16:23 mst what activeMQ broadcast to a topic does ... is pretty much a loop
16:23 magarwal client is a web app
16:23 mst now answer my question please!
16:23 maybe_sri batman: now i'm totally dead
16:24 sri joined #mojo
16:24 maybe_sri mst: no need to yell, he obviously doesn't know what he's doing or wants
16:24 mst I asked four times "what class is the object in your example?"
16:24 mst that's not hard
16:24 mst w 23
16:24 batman maybe_sri: you're resurrected
16:25 marcus \o/
16:25 mst marcus: *prod* -> /msg
16:25 sri \o/
16:25 r0b3rt left #mojo
16:25 batman sri: but the backend is rather messed up. no idea why :(
16:25 sri :S
16:26 mst magarwal: basically, what you'd do is jsonify once
16:26 mst magarwal: then you're just ending a frame to each client
16:26 mst magarwal: which you can do in an nb loop and then let the ioloop deliver them for you
16:26 mst magarwal: which is basically what a broadcast *is* when you've got to fan it out over individual TCP connections
16:26 mst https://metacpan.org/pod/Mojo::Transaction::WebSocket#send
16:27 r0b3rt joined #mojo
16:27 sri websocket sends are never blocking
16:27 batman sri: the issue is that the receiver haven't gotten the message either, right..?
16:27 mst try and avoid imagining extra problems that don't exist - programming is hard enough when you only have to deal with the real ones.
16:27 batman that's what you showed me in the log?
16:27 nicomen magarwal: using non blocking should at least save you the accumulated time of waiting for the clients to get all data (if send_json is synchronous)
16:27 sri batman: my messages arrived
16:27 sri i just didn't see any
16:27 magarwal mst, nicomen, thanks
16:28 batman sri: so let's say...mst see your messages, but you don't ?
16:28 nicomen magarwal: ah, if they are not blocking like sri says, then the old thing you could save time on is to_json the data upfront I guess
16:28 mst nicomen: which is what I said :)
16:28 mst magarwal is inventing problems and then imagining solutions that don't exist and don't need to exist :)
16:29 nicomen magarwal: depending on the data you want to send, maybe at some point you would need to partition the data, ie. send status updates only for a user's friends, and not the whole list of users or similar
16:29 mst also, something else will probably break catastrophically long before you get to 10,000 simultaneous clients
16:29 mst it almost always does
16:29 nicomen mst: yes, but it still seemed like you guys are talking past each other ;)
16:29 mgrimes joined #mojo
16:29 batman sri: could you do refresh in your browser?
16:29 mst that's why I gave up on him saying anything useful and decided to just explain from first principles :)
16:30 nicomen magarwal: I think the real question is, why are you optimizing now, is it something that is starting to get slow?
16:30 nicomen mst: at least I understoood you ;)
16:30 mst that's usually a start :D
16:30 batman marcus: i think we need to enable ipv6 source ip... [username:freenode-org] Closing Link: demo.convos.by (Too many host connections (local))
16:31 sri batman: yes, that's what happened
16:31 mst batman: if you're running a gateway you can apply for a gateway cloak with an increased limit
16:31 batman mst: ok. cool :)
16:31 mst just be aware that you'll want to have some way of dealing with abuse
16:31 batman marcus: i'll assign that task to you, since you're more of a people person than me.
16:32 batman sri: could you do a refresh?
16:32 sri i actually asked in #toolchain if anyone knew since when https://cpan.metacpan.org has been available... but i might have missed the answer due to the convos problem
16:32 mst generally freenode is pretty kind to people, provided you wield the banhammer when required
16:32 sri anyone here know?
16:32 mst sri: nope
16:32 mst and nobody replied
16:32 mst I'd ask ranguard directly
16:33 batman demo.convos.by got 310 users now...
16:33 batman i will kill most of them...
16:33 mgrimes joined #mojo
16:33 sri ouch
16:34 sri and thanks
16:35 batman sure. sorry for the issues :(
16:35 batman it's not the 310 users though.
16:35 batman the server says $nick is missing, so it can't render a template.
16:35 sri oh, android lollipop is out
16:35 batman that's why i wanted you to refresh (and scroll back) sri
16:36 batman but i need to know _when_ you will do that
16:36 sri now
16:36 batman go for it
16:36 sri done
16:37 batman have you scrolled back as well?
16:37 sri yes
16:37 batman and still the messages are missing?
16:38 sri yes
16:38 marcus batman: I ???? people
16:38 dod joined #mojo
16:38 batman then i don't know after all...
16:38 batman this sucks
16:38 sri that nexus 6 is huuuuge
16:39 batman i will make sure users are deleted now and then i will go back to taking it slow on the couch
16:39 sri get well soon!
16:39 mst it's plague season everywhere
16:39 batman hopefully before tomorrow. got work to do :P
16:39 batman sri: thanks
16:39 marcus sri: Arne got a Iphone 6 plus, it's pretty huge too.
16:39 mst bloody universities
16:40 batman sri: btw: you can bug marcus as well ;)
16:40 batman i will give him access to the server soon
16:40 sri ooooh, amoled display
16:41 marcus Btw, now that it's more or less official I guess I should mention that I will be taking a new job soon.
16:41 marcus At Oslo Uni.
16:41 sri :o
16:41 mst gotboordaker?
16:42 marcus Burned out on consulting, and we've not managed to make output.digital sustainable so far.
16:42 batman marcus: another bug: users have uppercase usernames :(
16:42 marcus batman: oh noes.
16:43 batman marcus: i will keep jhthorsen, marcus and kraih. anyone else?
16:44 mst I'm still pretty amazed I haven't (burned out on consulting), tbh
16:44 marcus batman: gggritso ?
16:44 batman marcus: don't ask me. i'm asking you :P
16:45 batman gggritso is not a user
16:45 marcus batman: that's her freenode nick, but I don't know if it's her username
16:45 batman could you do /whois gggritso ?
16:45 marcus And I can't connect to freenode atm
16:45 marcus batman: tried
16:45 batman can you do it with another client, just to figure out?
16:46 marcus yes
16:46 batman i have no idea how to find user from nick
16:46 mgrimes in Test::Mojo is there an equivalent to WWW::Mech's submit_form_ok?
16:46 mgrimes rather than recreate via $t->post_ok(),  I'd like to specify a form and submit to the action url,  incorporating hidden fields, etc.
16:46 sujithm joined #mojo
16:46 stephanj i am in search of a nice pure CSS UI framework(no JS for basic functions like (dropdown)menues or the likes; not open for discussion)... can you maybe recommend some?
16:47 tempire purecss
16:47 marcus batman: forget about it, flush'em all except us three.
16:47 purl marcus, I didn't have anything matching about it, flush'em all except us three
16:47 tempire oh, they use js for dropdowns, I think.
16:47 batman marcus: :)
16:47 stephanj tempire: ye
16:49 stephanj tempire: thanks for the tipp throu - will investigate
16:50 Mso150 joined #mojo
16:50 sujithm_ joined #mojo
16:54 marcus sri: getting a nexus 6?
16:55 * mst likes his nexus 4
16:55 mst WHY DO ALL THE NEW PHONES HAVE TO BE FUCKING MONSTROSITIES I ALREADY HAVE A TABLET
16:55 * sri likes his nexus 5
16:56 sri the nexus 6 is quite pretty, but i'm not sure about the size
16:56 marcus I think I want a iphone 6+ tbh
16:56 marcus and no more tablets
16:56 tempire I would get the 6+
16:57 marcus tempire: btw, autolayout has become so much more pleasant to work with in xcode6 \o/
16:57 tempire Yes.
16:57 sri hmm, same price for nexus 6 as iphone 6+ makes the decision harder
16:57 tempire Also, probably due to me understanding it more now.
16:58 tempire But it's too big to have in my pocket while dancing
16:58 * marcus has hacked together an open source ios8 app this weekend. https://github.com/meatspaces/meatspace-ios
16:59 mst it'd fit in the kilt fine, the nexus 7 does
16:59 batman crap. i have to restart the backend again :(
16:59 marcus I suspect tempire can't dance in a kilt.
16:59 mst I dunno if it'd fit in the pockets of any of my skirts, but basically fuck all else does either
16:59 batman 30 seconds marcus + sri
16:59 mst other than headbanging, I mostly don't dance without wearing a kilt or a skirt
17:00 tempire marcus: get on the swift bandwagon
17:00 tempire debugging with println is all the rage.
17:00 marcus joined #mojo
17:00 marcus tempire: didn't want to spend the extra time, think I'm getting out of ios dev :)
17:01 tempire understandable.
17:01 marcus But maybe I'll just backport it to clean it up once 1.0 is out.
17:01 marcus As a weekend project.
17:01 tempire I've been banging my head against it for a while, but I think I've got the hang of it.
17:01 tempire mostly
17:01 marcus The socket.io client I use is horrible btw :(
17:01 batman joined #mojo
17:01 marcus it's using a uiwebview and javascript under the hood.
17:02 marcus_ joined #mojo
17:02 tempire cheater
17:02 HtbaaPi when running hypnotoad behind a proxy (and proxy => 1), and running the proxyserver behind SSL, shouldn't url_for() detect https as the scheme to use?
17:02 marcus Which is why I have to    dispatch_async(dispatch_get_main_queue(), ^{ }) all the things
17:03 nicomen mgrimes: here:
17:03 sri joined #mojo
17:03 batman WTF?!
17:03 batman marcus: why don't we bundle IO::Socket::SSL??
17:03 * batman need to restart AGAIN.
17:03 stephanj ooops
17:03 wtfmarcus batman: it requires libssl?
17:03 batman hm...
17:04 * batman updates ansible
17:04 batman this is awful :(
17:04 batman why didn't i remember this
17:04 nicomen mgrims: https://gist.github.com/nicomen/dccb9c4843b649b91c0d#file-mojolicious-submit_form_ok
17:04 HtbaaPi when using to_abs with url_for() of course...
17:04 wtfmarcus batman: did you install EV?
17:04 wtfmarcus otherwise you will be sorry
17:04 batman marcus: i am sorry :(
17:05 wtfmarcus I'm sorry for not reviewing better :(
17:06 mgrimes nicomen: perfect. thanks!
17:06 sri joined #mojo
17:07 batman joined #mojo
17:07 marcus joined #mojo
17:07 batman are we here now? marcus, sri?
17:07 sri am i?
17:07 batman yes
17:08 mst /kill sri not anymore
17:08 wtfmarcus I'm not
17:08 nicomen HtbaaPi: not necessarily, the HTTP request is mainly GET /path, and Host: domain.com (so even if it went thru SSL, the app doesn't see it) SO there are mechanisms to tell the web app which protocol was used in the proxy, and one  is setting a header X-Forwarded-Proto. Wether mojolicious or your proxy are reading and setting it, I don't know
17:09 sri resurrections always get me hungy
17:09 batman marcus: sorry for screaming
17:09 batman sri: i don't have to resurrect you manually anymore. i've fixed you permanently :)
17:09 sri feel free to scream at me if you need to let out some steam :)
17:09 wtfmarcus batman: I just get logged out as soon as I log in now.
17:09 sri \o/
17:10 batman wtfmarcus: i guess that's a new *feature*
17:10 wtfmarcus \o/
17:10 batman wtfmarcus: can you clear cookies and try agiain?
17:10 wtfmarcus ya
17:11 marcus ok, I'm here now \o/
17:11 powerman left #mojo
17:12 batman sweet
17:16 sri hmm, i guess i'd rather get a moto x with bamboo back than a nexus 6 :o
17:16 HtbaaPi nicomen: thanks, I didn't set X-Forwarded-Proto correctly
17:17 HtbaaPi after setting it to https it works the way it should
17:18 sri oh well, lets see how lollipop looks like on the nexus 5 :)
17:19 r0b3rt_ joined #mojo
17:22 sri and they keep selling the nexus 5, website has been updated with lollipop too
17:23 sri_ joined #mojo
17:24 batman_ joined #mojo
17:24 marcus_ joined #mojo
17:27 sri batman: something is wrong again
17:27 sri_ batman: something is wrong again
17:27 sri marcus: HALP
17:27 sri_ marcus: HALP
17:27 tempire It's the internet.
17:27 ignacio_ joined #mojo
17:27 tempire It's all coming to an end.
17:27 sri there's two of me now :O
17:27 sri_ there's two of me now :O
17:28 tempire http://www.blogcdn.com/www.engadget.com/media/2010/07/4768582930d23c62f411.jpg
17:28 sri joined #mojo
17:28 batman joined #mojo
17:29 marcus joined #mojo
17:30 marcus wow mom wow
17:30 batman sri, marcus: no idea how, but the locking mechanism for the backend failed, so it spawned twice
17:31 jzawodn_ sri has forked
17:31 batman marcus: i don't havet the energy to figure that out now :(
17:31 sri aww... my clone died :(
17:31 marcus Now we will have twice as many releases
17:31 batman sorry about that
17:31 marcus oh shoot
17:31 * batman fold
17:31 marcus I have root on the box now
17:31 marcus so we're all doomed
17:31 sri so much for fixing twice as many bugs
17:32 batman i *hope* it happened because convos started before redis on boot, but i'm not sure
17:32 batman that's the best theory i have...
17:32 batman it's worth zero though
17:33 batman marcus: happy bug hunting
17:33 batman i will hopefully not be back until tomorrow
17:33 mst batman: Daemon::Control is your friend. or some sort of lock file or something.
17:33 batman mst: you have no idea
17:34 batman mst: don't try to read the forking code in convos.
17:34 batman you will hate me.
17:34 mst fork-and-exec-a-backend-on-demand is not an easy problem to solve without it getting gnarly
17:34 batman it gets even worse when you're only supposed to fork in some environments...
17:35 batman and the environments aren't documented :P
17:36 mst there's a directory on my dev box with about a dozen different ways to do this wrong, and half a dozen that weren't totally wrong but made me hate myself anyway
17:36 mst if I ever get round to producing a reasonably simple such thing, I'll tell you and you can either use it or steal the logic or something
17:39 batman that would be nice, but you would have to take into consideration that *it* should be started when mojo starts in either fork or daemon mode, and maybe fork in either one of the cases....
17:40 batman it would be so much easier if we just didn't design for heroku kind of environments (which doesn't work anyway)
17:40 mst oh, where there's multiple webheads and you want the first one up to start the backend?
17:41 batman yeah, that as well
17:41 mst bleh
17:41 batman i know.
17:41 * batman gives up for to night
17:41 batman <3
17:43 genio YAY.  Paris
17:43 genio \o/
17:44 marcus Texas?
17:44 purl Texas is one of the most conservative states in the union... Drubbly tingle with that though or the Australia of the U.S. or the reason that the president's dead or where tanya tucker wants to go when she's dead or full of shitholes (: dallas) or the only state you can legally drive 800 miles in a straight line in with a assault rifle on my gun rack or where you can shoot people
17:44 genio nope. the real one
17:45 marcus I still can't connect to freenode :'(
17:45 marcus Can't understand how it can have too many connections if batman killed everyone.
17:46 marcus Predictive text on iOS 8 is pretty neat btw. ::)
17:46 genio neat idea, but it's too big and gets in the way of seeing anything at all above what you're typing
17:47 marcus You can swipe it down when it's in the way.
17:48 marcus Also, it's apple's way of telling you that you need a bigger phone. :)
17:50 genio I have one more year on my contract :(
17:50 disputin joined #mojo
17:52 stephanj can't you request a connection limit increase?
17:53 mst yes
17:53 sri btw. i don't think we need to start signing mojolicious releases, cpanm --verify also uses CHECKSUMS from the authors directory, which gets signed by PAUSE
17:53 mst but that's a separate question to "why are we at the limit with nobody connected?"
17:53 stephanj maybe caching?
17:53 sri so that cpanm one-liner is really pretty good regarding security
17:55 ignacio_ joined #mojo
18:07 cpan_mojo XML-Loy 0.27 by Nils Diewald - http://metacpan.org/release/AKRON/XML-Loy-0.27 (depends on Mojolicious)
18:07 jamesaxl joined #mojo
18:27 preaction joined #mojo
18:32 da5id joined #mojo
18:36 Mso150 joined #mojo
19:07 Mso150 joined #mojo
19:09 basic6_ joined #mojo
19:14 Mso150 joined #mojo
19:23 denis_boyun joined #mojo
19:34 d4rkie joined #mojo
19:35 irq joined #mojo
19:35 Mso150 joined #mojo
19:50 Mso150 joined #mojo
20:07 jb360 joined #mojo
20:25 D4RK-PH0ENiX joined #mojo
20:45 disputin joined #mojo
20:59 mib_jtp2ka joined #mojo
21:06 tbushell joined #mojo
21:12 good_news_everyon joined #mojo
21:12 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/X5oBZQ
21:12 good_news_everyon mojo/master 8219cce Sebastian Riedel: use more secure download addresses
21:12 good_news_everyon left #mojo
21:17 berov joined #mojo
21:19 good_news_everyon joined #mojo
21:19 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/a-bF9w
21:19 good_news_everyon mojo/master 0431628 Sebastian Riedel: more secure links
21:19 good_news_everyon left #mojo
21:26 good_news_everyon joined #mojo
21:26 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/pz_ofw
21:26 good_news_everyon mojo/master 35852ab Sebastian Riedel: fixed documentation search
21:26 good_news_everyon left #mojo
21:30 stryx` joined #mojo
21:33 good_news_everyon joined #mojo
21:33 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/jLwOlA
21:33 good_news_everyon mojo/master 33b9dcc Sebastian Riedel: do not surround error messages with parentheses
21:33 good_news_everyon left #mojo
21:36 jnbek awww..... I was just getting to like those!
21:37 * sri sets jnbek on fire
21:56 jnbek lol
21:58 good_news_everyon joined #mojo
21:58 good_news_everyon [mojo] kraih tagged v5.50 at 9d13f44: http://git.io/GcprLQ
21:58 good_news_everyon left #mojo
21:59 good_news_everyon joined #mojo
21:59 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Lve0HA
21:59 good_news_everyon mojo/master 665fcce Sebastian Riedel: bump version
21:59 good_news_everyon left #mojo
22:00 cpan_mojo Mojolicious 5.50 by Sebastian Riedel - http://metacpan.org/release/SRI/Mojolicious-5.50
22:01 davido_ joined #mojo
23:16 wim joined #mojo
23:48 chansen sri: Today I came up with a very elegant way to format RFC2616 date strings:
23:48 chansen $ perl -wle 'print sprintf "%s, %02d %s %s %s GMT", (split /\s+/, scalar gmtime)[0,2,1,4,3];'
23:50 jberger Hahahahaha, eating today's backlog was hilarious
23:51 jberger I wish my nick forked!
23:51 jberger but dammit I'm on a reliable home based convos instance
23:52 * jberger opens a feature request for user forking
23:53 chansen I should try convos, I'm still on IRCCloud since marcus's recommendation
23:53 chansen ... or invite
23:56 chansen_convos joined #mojo

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