Camelia, the Perl 6 bug

IRC log for #mojo, 2012-09-22

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

All times shown according to UTC.

Time Nick Message
00:07 rem_lex|pivo joined #mojo
00:14 rem_lex| joined #mojo
00:42 rem_lex|pivo joined #mojo
00:48 zpmorgan marcus, I put together a lite demo app to show how 'subscribe' doesn't work, but then it worked :/
00:48 zpmorgan https://gist.github.com/3764706
00:52 zpmorgan I don't know why it fails when it fails; example: you can try subscribing to an empty string, I don't think it tells you that it won't work.
00:53 zpmorgan oh, oops, an empty string is a valid channel name
00:58 noganex joined #mojo
01:09 kitt_vl joined #mojo
01:48 xxtjaxx_ sri: Oh great master of Mojolcious, I present to you: Polo - volo for Mojolicious https://github.com/andreas-marschke/Polo ;)
01:57 human39 joined #mojo
02:18 l3l1p joined #mojo
02:41 noganex joined #mojo
03:06 xaka joined #mojo
03:36 crab for someone who doesn't know what "volo" is, what does that thing do? download jquery?
03:37 crab ah. pretty much.
03:59 Foxcool joined #mojo
04:09 l3l1p joined #mojo
04:35 Foxcool joined #mojo
04:45 l3l1p joined #mojo
04:51 davido joined #mojo
04:52 davido Hey everyone.  I got a reply to my recent PerlMonks post: http://www.perlmonks.org/?node_id=995023
04:54 davido The reply briefly discusses XSRF attack vectors.  My sense is that all that would be needed would be to verify the referrer, but I'm no expert on the topic.  Nevertheless, if there's a wise response to this reply, I'd like to provide *something*.
04:56 crab davido: there are multiple implementations of CSRF protection plugins for mojo
04:57 crab searching for mojo CSRF should find them
04:57 davido I'll look into those.  Thanks.  That's probably all that's necessary.  I don't intend to convey expertise on the subject, but simply want to keep the record presented there at PM accurate.
04:58 crab fwiw, checking the referer is not considered sufficient. the usual technique is to include a random token in forms and also in your session cookie, and when a frm submission arrives, check to see that the two tokens match
04:59 crab i could go on about this for several minutes :-), but i won.t
05:00 davido Well if you wish to, I'm always interested.  The wikipedia article lists checking the referrer as one strategy, but it does seem too simple. :)
05:00 davido The token method is also mentioned therein.
05:01 davido I found Mojolicious::Plugin::CSRFDefender
05:01 crab who is it by?
05:02 crab iirc there's CSRFDefender and CSRFProtect. one is older and one is newer.
05:03 davido Yuki Shibazaki
05:03 davido This one was last updated June-11
05:03 crab right. that's the older one. the other is by someone named viktor(?)
05:04 davido The newer one: Viktor Turskyi, updated in 2012.
05:04 davido M::P::CSRFProtect
05:05 crab there you go.
05:05 crab i suppose i should have made my implementation into a plugin, but i never bothered.
05:06 davido Looks a little more thorough.  One difference is that Victor's overrides "form_for", whereas Yuki's attaches to any post form
05:21 davido There we have it: http://www.perlmonks.org/?node_id=995042
05:32 crab good work.
05:33 davido Thanks.  Just trying to spread the *good word*
06:32 vishwanathms joined #mojo
06:34 vishwanathms_ joined #mojo
06:42 vishwanathms joined #mojo
07:19 Britzel joined #mojo
07:39 davido Is a submit_button's id not available in the request?
07:43 davido Or more to the point, is there an example somewhere of allowing my controller action to detect which of two buttons were clicked?
07:53 davido (I got it... use name instead)
07:56 Vandal joined #mojo
08:00 ryozi joined #mojo
08:02 crab form submission doesn't give you the id of anything
08:44 batman joined #mojo
09:07 dminus_ joined #mojo
09:07 ryozi_ joined #mojo
09:08 drift_ joined #mojo
09:09 korshak joined #mojo
09:09 Mad_Dud joined #mojo
09:10 marcus joined #mojo
09:10 nicomen joined #mojo
09:10 phillipadsmith joined #mojo
09:10 ask joined #mojo
09:10 Mikey joined #mojo
09:10 aptituz joined #mojo
09:10 phips joined #mojo
09:10 stephan48 joined #mojo
09:10 ryanc joined #mojo
09:10 HtbaaPi joined #mojo
09:10 hide joined #mojo
09:10 bobkare joined #mojo
09:10 nuba joined #mojo
09:10 Annihitek joined #mojo
09:11 xxtjaxx_ crab: crab Goal is: If you need extra assets from outside (such as jquery or another perl package pull in these dependencies and install them locally.
09:11 bjoernfan joined #mojo
09:11 davido joined #mojo
09:11 dwierenga joined #mojo
09:11 throughnothing joined #mojo
09:11 al joined #mojo
09:12 tokuhirom joined #mojo
09:12 kmx joined #mojo
09:15 xxtjaxx_ crab: https://github.com/volojs/volo
09:44 Foxcool joined #mojo
09:56 kitt_vl joined #mojo
10:04 nic I'm wondering whether there could be a single plugin, Mojolicious::Plugin::Delegate, which is flexible about its back-end
10:06 crab what does that even mean
10:10 nic (I was going to continue, the realised it's better to prepare a gist rather than start discussing code)
10:10 nic I'll prep example code before continuing my ramblings :)
10:15 crab well, i'd rather understand what you mean than read code out of the blue whose purpose is unclear
10:16 nic ok, just for you...
10:16 nic When you want to delegate some (blocking) code, you have a choice of back-ends
10:17 nic If you're happy to spawn ad hoc processes, then the successor of MojoX::Run (which I'm working on) will do the job
10:18 nic If you want to delegate to pre-spawned workers, then you might choose Resque or a non-redis scheduler I'm working on
10:19 nic I don't see a gain in writing the delegator against a different API tied to a particular back-end
10:19 nic I suppose it's a kind of AnyJobber
10:20 nic So if you want to switch your code to a different back-end, it would mainly be a task of changing the setup call to Mojolicious::Plugin::Delegate
10:22 nic Most of the use cases are fire-and-forget-till-its-complete -- you shouldn't be concerned/tied to a choice of back-end IMO
10:23 crab i see.
10:23 nic There is also the case where you want to communicate with the delegated job, which I'm not currently seeing how it would fit
10:24 nic eg your app continues to interact with a delegated 'bc' job, or incrementally feeds & checks some process like LaTeX processing
10:26 nic (I think I'll have to ignore that use case for now)
10:26 nic zpmorgan: Any thoughts on any of that?
10:27 crab it's not really delegation if you're doing that, though
10:27 nic true
10:28 nic "I want you to get on with this task, but I haven't finished giving you input, and I want to check your status each time you complete a step"
10:29 nic Mojolicious::Plugin:MicroManage :)
10:29 * sri doesn't think MojoX::Run is good for background jobs
10:29 nic what aspect of it?
10:30 sri hypnotoad workers are supposed to get restarted regularly, that might interfere with jobs that are not tied to an active connection
10:30 nic Obviously having a shared api means it would be trivial to compare performance, therefore helping people choose what's best for their app
10:31 nic Does IOLoop restart?
10:31 sri the whole worker process does
10:32 sri it exits and a new one is spawned
10:32 sri process watchers may trigger that as well with signals
10:32 nic ah, yes
10:32 sri to manage memory limits and the like
10:32 nic so anything registered against ->loop will be lost?
10:32 sri it may
10:33 nic hmm, I should have thought about that; best test it out before I spend another day of coding :/
10:33 sri resque is a way better architecture
10:35 zpmorgan nic, I'm interested. It still seems handy to me, particularly for small, lite apps
10:35 zpmorgan although maybe it should be called Mojolicious::Plugin::DWIM
10:35 sri if the process is tied to a connection it shouldn't be a problem though
10:36 sri (active connection)
10:36 nic I don't disagree Resque is better -- I also have a work project which is going to use it -- but ad hoc spawning via IPC::Open3 is a quicker simpler solution that fits a number of other projects
10:39 nic I also have a project in the queue that will need both; it needs a persistent job queue and ad hoc processes that are guaranteed to die as soon as their task is complete
11:15 dpetrov_ joined #mojo
11:43 ZadYree joined #mojo
11:58 schizo joined #mojo
12:02 ZadYree joined #mojo
12:16 Lucas1 joined #mojo
12:26 vishwanathms joined #mojo
12:28 vishwanathms_ joined #mojo
12:36 vishwanathms joined #mojo
12:38 zpmorgan_ joined #mojo
13:06 bpmedley nic: Try gearman?
13:07 nic bpmedley: yep, I will (re)consider it as a back-end, but it's a bit too cumbersome so has been ruled out of my upcoming projects
13:08 bpmedley nic: Cumbersome?
13:08 nic another team are keen on it tho, cos they could share it for php job scheduling
13:09 bpmedley nic: Which part is cumbersome?
13:15 nic bpmedley: What are you proposing gearman as an alternative to?  Resque?
13:15 bpmedley nic: Well, honestly, I didn't read the entire history.  If you already have a solution, then I'd hate to step on your toes.
13:16 bpmedley nic: I've never used Resque.
13:17 nic IMO gearman wins big time with its multi-language api(s)
13:18 nic hence its likely use in a project we'll be sharing with a php team
13:18 bpmedley nic: It's also pretty simple to use; that's really why I was talking.  I don't understand which part is "cumbersome".
13:18 bpmedley nic: Or maybe my use-case is simple?
13:19 nic but resque is blindingly simple; compare that to the gearman overview and there's no contest when it comes to explaining it to mgrs and devs
13:19 bpmedley nic: How would you describe it to another dev?
13:19 nic and, thanks to diegok, resque has all the code ready to be forked and messed with/extended, unlike gearman which is in c
13:20 nic http://www.slideshare.net/andy.sh/gearman-and-perl is 101 slides
13:20 bpmedley nic: I do agree some of the documentation could be a bit.. toned down.  I found a presentation that made sense to me.
13:20 nic 101
13:21 bpmedley nic: I think I read that one.  Perhaps we need a 5 minute lightening slide for gearman?
13:21 nic resque + redis was set up in 10 mins, with 9.5 mins of that being for configuring options for redis
13:22 bpmedley nic: I see your point; there could be a "Quick intro to gearman" document.
13:23 nic bpmedley: right now (and tomorrow) i'm all about the test env, so if there's a quick route to having a gearman prototype installed and ready to throw stuff at, I can do that right now
13:24 bpmedley nic: Ok, I'll help.  What OS?  I installed in Centos via CPAN and some 3rd party repo.
13:24 nic what's my shortest route to setting it up on my test box?
13:24 nic this is linux
13:24 nic (debian 6.?)
13:25 bpmedley apt-get install gearman-job-server gearman-server  <-- from the web site
13:25 bpmedley Meaning, I haven't tested that route; hopefully it will work.
13:25 nic linux v3.0.0, debian wheezy, any version of perl > 10.0.1
13:26 nic k, trying that
13:28 bpmedley /usr/sbin/gearmand -l /tmp/gearman.log -L 127.0.0.1 -p 4730 -u nic -v -v
13:30 nic \o/ gearmand is running
13:30 bpmedley cool.. 1 minute\
13:31 mire_ joined #mojo
13:37 bpmedley client: http://paste.scsys.co.uk/208256
13:38 bpmedley worker: http://paste.scsys.co.uk/208257
13:39 bpmedley Start worker 1st pls then run the client.
13:51 bpmedley Also, please install the CPAN module Gearman::Client
13:55 vishwanathms joined #mojo
13:55 vishwanathms_ joined #mojo
13:55 bpmedley nic: Make sense?
13:56 nic (sorry, got waylaid, back now)
13:59 bpmedley nic: Did you get the pastes?
14:00 nic yep, just getting things across to my test/demo box right now
14:24 nic bpmedley: is that code known working or could there be some ironing out for me to do?
14:24 bpmedley nic: It works for me.
14:24 nic k, that tells me which end the problem is at :)
14:24 bpmedley What is happening?
14:25 bpmedley You should get a random number back.  The "get" sub should probably be called "random".
14:25 nic I don't get any file at /tmp/gearman.log -- does that tell me the server has a problem?
14:25 bpmedley No, that's fine.
14:25 nic the client has been hanging for a while -- doesn't come back
14:26 bpmedley Is anything in syslog?
14:26 nic sorry, I bet I need to start the server as 'root'
14:26 bpmedley No, I do not.
14:26 bpmedley It uses a port above 1024.
14:27 bpmedley Does netstat show a listener on 4730?
14:28 nic ah, no
14:28 bpmedley Is "mic" your username in LInux?
14:28 bpmedley *nic
14:28 l3l1p joined #mojo
14:30 nic ah, was going to ask you about that :D
14:31 nic also, it said it didn't recognise "-l -l -v -v" (even tho I only gave -l once)
14:31 bpmedley Try just "/usr/sbin/gearmand" as your normal user.
14:32 nic Unknown option: l
14:32 nic then again for l, then for u, then for v, then again for v
14:33 bpmedley odd..  will you "less /usr/sbin/gearmand" ?
14:33 * nic tries plain "./bin/gearmand &"
14:34 nic yay, I think worker.pl is running now...
14:35 bpmedley Cool.
14:35 nic Can't call method "syswrite" on an undefined value at /home/nsandfield/proj/gear​man/lib/Gearman/Taskset.pm line 202.
14:35 nic ...when I run client.pl
14:35 bpmedley Moment.
14:36 bpmedley What port is the gearmnd listening on?
14:37 * nic checks
14:39 nic Is 'none' a possible answer?
14:39 nic the proc is running but none of the open ports look likely
14:39 bpmedley Let's try "gearmand -p 4730"
14:40 bpmedley Rather: gearmand -p 4730 -v -v
14:41 nic k
14:42 bpmedley Did the port get bound to?
14:42 nic no
14:42 nic and it really doesn't like "-v"
14:42 bpmedley Any error message?
14:43 bpmedley What version of gourmand is it?
14:43 nic I'm going to try as 'root' in case I have local security blocking me
14:43 bpmedley Ok
14:44 nic how do I get it to tell me version number?
14:44 bpmedley gourmand --help did for me.. also maybe -V
14:44 bpmedley *gearmand
14:46 nic There seems to be something v fishy about the gearmand I have installed (from cpan)
14:46 nic it doesn't accept "--help", and it turns out one of the complaints about "-l" is cos it doesn't like "-L"
14:47 bpmedley Hrmm.  Is it a script?
14:47 nic ah, wasn't expecting it to be ascii -- I can read the docn within it
14:48 bpmedley My gearmand is a binary.  May be a packing thing in debian?
14:48 bpmedley Can you tell what they installed the server as?
14:50 nic \o/
14:50 nic it works
14:50 bpmedley Sweet.
14:50 nic sorry, i just wasn't thinking that it could be an awk script
14:51 bpmedley Understood.  Usually it's not.
14:51 nic (invocation is "./bin/gearmand --daemon -d" defaulting to port 7003)
14:51 bpmedley Ok; now change the perl to use port 7003.
14:51 nic yep, done that, hence the \o/
14:52 nic and another \o/ too
14:52 bpmedley :-)
14:52 nic thank you for your time and patience
14:52 bpmedley Sure thing.
14:52 nic one of the delays earlier was that this has highlighted a problem with my test box, so it was a double gain
14:53 nic and you've demonstrated that everything I had thought about gearman was wrong
14:53 nic the footprint is small and the protocol is light and versatile
14:54 bpmedley Well, they need a quick-start guide.
14:54 nic If it could be made easier and faster to get beginners trying it out (like wot you did today) that would be a big win
14:55 nic I had tried to get into Gearman at the same time as looking at TheSchwartz and POE::Wheel; the former is old fashioned and the latter is too heavyweight (for my purposes)
14:56 nic and in the rush I forgot that I had failed to get into gearman at all and just tarred it with the same brush
14:57 bpmedley Gotcha.  I had a similar situation a few weeks ago.  Glad I could help.
15:08 batman joined #mojo
15:49 Lucas1 joined #mojo
15:57 berov joined #mojo
16:01 marcus hmm, I don't get it
16:01 berov Hi all!
16:01 berov I am making a site managemnet application which I will publish on CPAN.
16:01 berov Having in mind that it will ofer a default configuration file - appname.pl I was wandering how could I make so users installing the application put their configuration in a local.appname.pl config file
16:01 berov is there an easy way similar to $ENV{MOJO_CONFIG}
16:01 berov ?
16:02 berov or I should implement it my self
16:02 berov ?
16:02 batman berov: do you have some sort of backend to store data in?
16:03 berov yes SQLite or mysql
16:03 berov that was that I was thinnking
16:03 berov to store the local config in a table
16:03 berov which will be merged with the default config
16:04 batman yes. store it in the backend
16:04 berov ah, thanks , batman for the encouragement :) , will do it this way
16:04 batman :)
16:06 marcus sri: we're having some trouble with web socket handling. for some reason, it seems not to handle unicode. Mojo::JSON fails horribly if I have non-ascii things in the ws frame
16:06 marcus Could of course all be batman's fault :)
16:07 batman what? it's never my fault. even when it is...
16:07 batman :)
16:08 marcus {"cid":"2","cname":"#wirc","cmd":"hørggh"}
16:08 marcus that's in chrome debugger, batman.
16:08 berov batman: BTW, since you recomended metacpan.org I look only there :), I did not realised before that how better than cna it is really
16:09 marcus <3 metacpan
16:11 batman berov: that is actually awesome news!
16:12 batman good to hear that you like it
16:12 berov :)
16:14 abra joined #mojo
16:16 batman one thing i don't get about metacpan is why the index is just one level
16:16 batman it used to contain methods ++, but it's gone :(
16:18 berov not sure I understand, but anyway, will get to it too
16:21 batman https://metacpan.org/module/Mojo <-- look at the index. it only contains NAME, SYNOPSIS, .... not "home", "log", ...
16:26 * sri shoos marcus away
16:27 jzawodn joined #mojo
16:29 * sri hates encoding problems
16:29 Jason_ joined #mojo
16:29 batman joined #mojo
16:31 marcus sri: me too. :(
16:31 marcus sri: I see you have tests for unicode
16:31 batman sri: i totally agree. good thing marcus volunteered
16:31 batman :D
16:31 sri yay for volunteers
16:32 batman marcus++
16:32 * marcus cooks sri and batman
16:32 * marcus & # dinner
16:32 * sri is not edible
16:33 batman marcus: could it be double encoding?
16:33 sri btw. next few weeks i won't be around as much
16:33 batman what'up?
16:33 sri some vacationing
16:34 batman that sounds good :)
16:35 sri indeed \o/
16:36 batman i'm starting to like "...@{[...]}..." more and more
16:36 batman not sure that i like that i like it though:P
16:38 sri batman: haha, i know exactly what you mean :)
16:38 batman :)
16:41 sri before using it i now always make sure the alternatives don't look better
16:42 batman sounds like a lot of extra work...
16:51 sri http://hastebin.com/raw/sideduware # which style is best? :)
16:56 * sri kinda hates looking at the first two
16:57 dpetrov_ joined #mojo
16:59 dpetrov_ berov: you might want take a look at https://metacpan.org/module/Galileo
16:59 berov I saw it wonce
16:59 berov once
17:00 dpetrov_ I think there was command config
17:00 dpetrov_ which writes a config file
17:00 berov aha just looking at it
17:01 berov looks very similar to my  MYDLjE project
17:02 berov I now try to do something with no requirements and minimum additional programming aand not monolitic as before
17:05 sri hmm, i guess it would be nice to have Daemon::Control support on CPAN
17:05 sri so if anyone feels like uploading a release, knock yourself out!
17:06 ZadYree joined #mojo
17:11 batman sri: about the paste: i'm doing a lot of sprintf(), but @{[]} works nicely on simple stuff
17:13 sri @{[]} is often shorter than sprintf for simple stuff
17:13 Britzel joined #mojo
17:13 sri too bad we don't have the python version :)
17:14 batman which is?
17:15 sri "foo: %s bar: %s" % $a, $b
17:15 batman isn't that sprintf?
17:16 sri it has a few more features, but basically yes
17:16 sri "foo: %(foo) bar: %(bar)" % {foo => 'lalala', bar => 'yada'}
17:17 sri ops
17:17 sri "foo: %(foo)s bar: %(bar)s" % {foo => 'lalala', bar => 'yada'}
17:17 sri forgot the s
17:19 sri bit out of character for python to be so conveninent :)
17:20 batman looks easy enough to implement...
17:22 sri batman: btw. did you give up on getting use Scalar::Util 'weaken' into Mojo::Base?
17:22 sri Scalar::Util::blessed would make sense too
17:24 dross joined #mojo
17:24 batman i think i don't care enough
17:24 batman it's not that difficult to do use s::u qw/.../; so....
17:25 sri i just remembered that there may also be potential for exploits :)
17:25 sri perl -Mojo -E 'package Mojolicious::Controller; use Scalar::Util "weaken"; package main; Mojolicious::Controller->new->weaken'
17:28 * tempire loves the turtle operator
17:28 * tempire also likes "${$thing->methodcall}"
17:29 tempire but more & more, I think sprintf is the better option
17:30 nordmarcus joined #mojo
17:30 nordmarcus Møøse!
17:30 purl møøse is better (unicode) or 牛牛頭犬 or 箆 or 駝鹿 or 大鹿 or bit my sister once
17:31 nordmarcus hømpf
17:31 nordmarcus tempire: moi.tempi.re doesn't use websockets?
17:31 tempire no, wanted it to work in all browsers
17:31 tempire websockets only worked in one at the time
17:32 marcus mkay, so much for seeing if you'd solved the unicode issue :)
17:32 batman sri: the exploit is not a reason to put it into Mojo::Base imo
17:32 batman there's no way to prevent such function leakage unless the programmer takes care of it
17:33 sri marcus: you sure you didn't just mix up bytes and chars? :)
17:33 marcus sri: it's quite possible.
17:33 batman marcus: have you tried to do Dump($input) ?
17:33 marcus sri: I'm not sure what javascript does here.
17:33 sri yea, i have no clue about that either
17:33 marcus sri: I just know that Mojo::JSON hates the json I get back in on_message
17:34 sri http://mojolicio.us/perldoc/​Mojolicious/Controller#send
17:34 sri the examples have a list of basically everything you can do
17:35 marcus sri: this is the content I get back from the browser tho, not the messages I send in.
17:35 marcus (actually, it's both)
17:36 sri i suppose, theoretically this should be your best bet $c->send({text => Mojo::JSON->new->encode({hello => 'world'})});
17:38 sri since there's no charset change required by the json decoder
17:41 sri from browser to server is actually a bit more tricky, since we decode incoming text messages, you may have to encode them again before feeding them to a json parser :)
17:41 batman marcus: i think that is the problem ^
17:42 marcus yeah, that makes sense
17:42 sri correct way to handle json would be to use binary frames
17:42 sri but i bet that would cause a lot of pain with browsers
17:42 sri json in text frames is a total hack basically
17:42 marcus encode works!
17:43 batman :)
17:43 sri since text frames on the wire are required to be utf-8 encoded
17:44 sri use binary frames, save a tree!
17:55 l3l1p joined #mojo
18:07 Foxcool joined #mojo
18:37 rem_lex joined #mojo
18:40 naquad joined #mojo
18:40 naquad hi
18:40 naquad is there any recommended ORM for Mojolicious? I need quiet simple DB support, but I don't want to mess with SQL
18:41 marcus naquad: I recommend DBIx::Class
18:49 l3l1p joined #mojo
19:09 memowe joined #mojo
19:09 rem_lex|pivo joined #mojo
19:12 marcus zrevrangebyscore <3
19:14 vishwanathms joined #mojo
19:14 vishwanathms_ joined #mojo
19:39 batman joined #mojo
19:46 nic bpmedley: You still around?
19:46 nic just wondering if you know anything about using mysql clients to gearman?
19:48 nic if that means a mysql daemon sending async messages to arbitrary workers, that would be awesome
19:49 nic (as long as it isn't a cludge with something polling the daemon)
19:49 nic tbh I got entirely bogged down again
19:50 nic apparently my gearman server was non-binary and on a different port because it is Old(TM)
19:51 nic the api has since changed, v1.1.1 has recently been released, but it's completely unclear where I get perl classes to code against that, if they exist
20:15 perlite joined #mojo
20:15 sri marcus: i accidentally a gravatar ;p https://github.com/mojolicious
20:16 marcus sri++ # looks good on my profile.
20:16 marcus hmm, I need a gravar without glasses, I suppose.
20:16 marcus since I don't need glasses anymore...
20:17 sri i should have added text
20:17 sri "we're sorry, something went very wrong with this avatar" or so ;p
20:18 sri it's already hard to spot the raptor on the tiny version though :S
20:19 marcus maybe just crop it to be the head puking the rainbow?
20:22 dross joined #mojo
20:23 sri marcus: like that?
20:24 sri it looks awkward when the arms are visible
20:24 marcus sri: yeah, better.
20:24 marcus "awkward dino"
20:25 marcus http://www.urbandictionary.com/de​fine.php?term=Awkward%20Dinosaur
20:26 ZadYree joined #mojo
20:29 sri \o/
20:30 xaka joined #mojo
20:35 l3l1p joined #mojo
20:44 ZadYree joined #mojo
20:58 tempire whoa
20:59 tempire I like the old github format
20:59 * tempire adjusts
21:00 tempire new format puts more focus on the dev, I guess.
21:01 tempire sri: the mojolicious avatar doesn't show up without being logged in
21:01 tempire shows up on the main page, though.
21:01 tempire strange.
21:01 purl But true.
21:02 sri prolly some caching quirk
21:03 tempire nice search features
21:06 stephan48 joined #mojo
21:13 sri oh, bind 10 is written in python :o
21:14 dross python <3
21:14 dross google's chosen langauge :3
21:14 sri not really
21:15 sri they do c++, java, python, go...
21:16 dross I'm aware they use other languages, however much is in Python :)
21:17 dross >.< now I'm anxious about friday again
21:17 dross interview
21:17 purl hmmm... interview is usually done on the "casting" couch
21:17 dross if you have devop in your descriptions or resume, they are instantly interested in you :3
21:20 * marcus adds devops borat to his linkedin
21:21 marcus hmm, I wonder if it would be useful to write a generator to make controllers and empty template files for all the defined routes.
21:21 dross marcus: you'll get spammed by google interviewers until you accept :3
21:51 Jason_ joined #mojo
22:06 memowe joined #mojo
22:17 sri nic: just in case you've not seen this awesome module yet https://metacpan.org/module/Child
22:18 sri that's the best fork abstraction i've seen so far
22:19 sri it actually works on windows :o
22:28 sri the only real disadvantage seems to be that you can't select/poll on the handle, since that's not portable
22:33 berov left #mojo
22:42 xaka wonder why people don't use socketpair for IPC instead of pipes. it's cross-platform in perl and even works in windows using simple accept/connect trick
22:49 * dross sighs
23:01 * dross ponders why his mongodb install isn't working >.<
23:15 xaka joined #mojo

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