Camelia, the Perl 6 bug

IRC log for #mojo, 2011-12-22

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

All times shown according to UTC.

Time Nick Message
00:02 GitHub157 joined #mojo
00:02 GitHub157 [mojo] kraih pushed 1 new commit to master: http://git.io/FVpx0A
00:02 GitHub157 [mojo/master] improved event handling in user agent tests - Sebastian Riedel
00:02 GitHub157 left #mojo
00:03 sri i think i have everything covered now
00:06 sri a gold star for you if you actually understand the latest test case :)
00:08 sri https://github.com/kraih/mojo/blo​b/master/t/mojo/user_agent.t#L114
00:08 sri it's rather cool, testing a chain of events on the server side
00:12 ccushing I want a different config file for tests, how would I do that?
00:14 xaka ccushing: via application mode, env variable, ...
00:15 marcus ccushing: set $ENV{MOJO_MODE}='testing'; and use myapp.testing.conf
00:16 marcus ccushing: it will fall back to myapp.conf for keys you don't define.
00:16 marcus sri: can I find the originating line of an event?
00:16 marcus sri: was looking at your test example, could I find out where that event originated in the code?
00:17 sri marcus: not sure i understand
00:17 ispeak joined #mojo
00:18 marcus sri: $err contains the entire die message?
00:18 sri yes
00:18 marcus no question, then.
00:18 sri :)
00:19 * marcus goes to bed, later.
00:19 sri eventemitter even does pass by reference
00:19 sri since it doesn't touch @_ and just passes it through
00:20 sri so you could modify the variables passed to emit()
00:20 sri nn
00:20 * sri should sleep soonish too
00:23 sri btw. user agent errors are special
00:24 sri since most can be associated with a transaction, and therefore just get passed along with them
00:24 xaka i've never used config before. is it JSON or Perl or what the format?
00:25 xaka ah, got it, nvm
00:25 sri the event (and previously the logger) are only for truly exceptional exceptions, like dead callbacks
00:25 sri (in case someone is still listening) :)
00:26 sri it has historically been a gray area, since there was no good solution, events solve it really well now though
00:28 tempire EVENTS!
00:28 sri !
00:30 tempire Events are within / Leaf fluttering in the wind / It's mojolicious!
00:30 mikegrb TO THE CLOUD!
00:31 tempire GO GO GO!
00:33 xaka would it be possible change log level for messages like "GET <url> ..." and "200 OK" from debug to info? i really want to see them in log file, but i don't want to see "Dispatching ..."
00:42 amoore joined #mojo
00:47 amoore joined #mojo
01:23 mattastrophe joined #mojo
01:49 j3nnn1 joined #mojo
01:52 ccushing joined #mojo
02:05 xaka sri: transaction's "finish" event is emitted twice when inactivity timeout happens. is it normal?
02:21 gshank_ joined #mojo
03:04 rsimoes left #mojo
03:08 rsimoes joined #mojo
03:36 Psyche^ joined #mojo
03:56 d4rkie joined #mojo
03:57 xaka joined #mojo
04:24 dryman joined #mojo
05:07 arpadszasz joined #mojo
05:17 marty_ joined #mojo
05:34 sherr joined #mojo
05:48 tempire I think this may have been addressed before, but…why is Mojo::Util?
05:49 tempire oh, it's functional
06:02 Foxcool joined #mojo
06:08 D4RK-PH0ENiX joined #mojo
06:11 Therstrium joined #mojo
06:47 Vandal joined #mojo
06:47 xaka joined #mojo
07:03 koban joined #mojo
07:03 koban left #mojo
07:03 koban joined #mojo
07:05 sugar joined #mojo
07:16 espent joined #mojo
07:21 sugar joined #mojo
07:23 crab "why is Mojo::Util"? huh?
07:27 batman joined #mojo
08:02 Psyche^ joined #mojo
08:03 xaka joined #mojo
08:09 spleenjack joined #mojo
08:39 alnewkirk joined #mojo
08:39 MojoGuest56 joined #mojo
08:39 MojoGuest56 From: http://twitter.com/ (7 hits)
08:39 berov joined #mojo
08:41 marcus TO THE CLOUDMOBILE?
08:42 judofyr joined #mojo
08:54 d4rkie joined #mojo
09:06 sri moin
09:20 tempire marcus: what sort of interaction did you have for aiclass?
09:21 marcus tempire: interaction?
09:21 purl interaction is probably completely optional.
09:21 tempire I presume it was more than just watching lectures
09:21 marcus tempire: I watched video lectures, and did homework, and had a midterm and a finals
09:21 tempire ah, ok.  so it's completely viable to get the full knowledge by watching the lectures after the fact, then
09:22 tempire I'm 6 months out at least, but I want to take that class
09:22 marcus homework every week, and you could skip 2 of them or something. % on statement was based on homework, midterm and finals
09:22 marcus tempire: I think they will run it again.
09:22 marcus tempire: it's good to follow the program tho, to pressure yourself to do the work
09:23 marcus had some late nights with the homework on the deadline :)
09:25 sri think we could use a release soon, marcus will do this one?
09:37 judofyr sri: I was really confused when I looked at some docs at metacpan and did *not* see "Sebastian" as release author
09:39 abra_ joined #mojo
09:45 * tempire wonders if it's a good idea to shift the power of mojolicious to norway
10:29 Quai mohahaha
10:31 omega tempire: you act like you don't know that  Norway owns the world already?
10:31 omega did you not get the memo?
10:53 mmcleric joined #mojo
10:56 sri norway was the capital of sweden right?
11:00 netfeed sri: teh horror
11:25 spleenjack1 joined #mojo
11:27 al joined #mojo
11:31 memowe akhand?
11:32 GitHub184 joined #mojo
11:32 GitHub184 [mojo] kraih pushed 1 new commit to master: http://git.io/2CwyPA
11:32 GitHub184 [mojo/master] improved user agent tests slightly - Sebastian Riedel
11:32 GitHub184 left #mojo
11:49 mire_ joined #mojo
11:52 AmeliePoulain joined #mojo
12:04 marcus sri: I can do that.
12:04 sri YES YOU CAN!
12:14 sri marcus: think master is in a good state for a release atm
12:17 marcus sri: we keep the timestamp at 00:00:00 ?
12:18 sri yea
12:18 marcus I see version has been bumped as well, not much for me to do but tag it and bag it then.
12:18 sri indeed
12:19 sri i usually bump the version with the first commit after a release
12:20 marcus urgh, I got home detection failures. Stupid non-case sensitive fs
12:20 * marcus retests with right dirname
12:21 sri so you're not using zshell?
12:21 marcus bash
12:21 sri zshell prevents that
12:21 marcus nice
12:21 marcus maybe I should switch
12:21 sri refuses to cd into a directory with wrong case
12:22 sri <3 zshell
12:23 sri it also knows which files i need to git add and stuff and lists only the right files for completion
12:24 sri configuring it right is an art though :)
12:24 marcus bash autocomplete does some of that too
12:24 marcus like autocompleting branches
12:25 sri zshell is a lot smarter than bash autocomplete
12:26 marcus hmm, guess I should add .tm_properties to MANIFEST.skip
12:26 marcus sri: you ship the distro with the perltidyrc in it?
12:26 sri all files that are in git
12:27 sri http://api.metacpan.org/so​urce/AMS/Mojolicious-2.38/
12:27 sri the documentation referrs to the .perltidyrc
12:28 marcus makes sense
12:28 GitHub50 joined #mojo
12:28 GitHub50 [mojo] marcusramberg pushed 1 new commit to master: http://git.io/awve5Q
12:28 GitHub50 [mojo/master] Add textmate2 properties to skipfile - Marcus Ramberg
12:28 GitHub50 left #mojo
12:28 purl makes sense is http://xrl.us/bjq3kq
12:30 sri hmm
12:30 sri i don't like having editor specific stuff in the repo
12:31 sri neither emacs nor vim require it
12:31 sri if textmate does i consider it a textmate bug
12:31 marcus sri: I don't keep it in the repo
12:31 marcus sri: I have it in my gitignore.
12:31 sri then why did you add it?
12:31 marcus I just put it in manifest skip to avoid it going into the manifest
12:32 sri yea, that's wrong
12:32 sri we shouldn't have to do that for every random editor a core dev uses
12:32 marcus POSTing upload for Mojolicious-2.39.tar.gz
12:33 marcus I do remember seeing backup files for emacs in manifest skips in the past
12:33 sri i don't see it
12:33 judofyr I usually do "git clone . release && cd release" to get a clean directory for release
12:33 marcus sri: catalyst-runtime manifest.skip has a bunch of that stuff
12:34 marcus https://gist.github.com/1510160
12:34 sri marcus: yea, think that through, where do we draw the line?
12:34 sri do we add exceptions for every editor in existence?
12:34 marcus sri: how many core devs do you plan to have? :p
12:35 sri i don't know
12:38 netfeed 67 no more, no less
12:39 netfeed and then a fight to the death with a current core developer if you want to become one
12:40 marty_ moin
12:42 sri marcus: https://github.com/rafl/catalyst-​runtime/blob/master/MANIFEST.SKIP # great example ;p
12:46 marcus sri: I see it has been optimized a bit in trunk :)
12:46 marcus sri: looks like that one only lists the files they want to keep.
12:46 Foxcool joined #mojo
12:46 marcus that's pretty weird for a skip file
12:47 sri marcus: don't forget to tweet the release
12:48 GitHub167 joined #mojo
12:48 GitHub167 [mojo] kraih pushed 1 new commit to master: http://git.io/Z7eCvw
12:48 GitHub167 [mojo/master] fixed repository to not favor specific editors - Sebastian Riedel
12:48 GitHub167 left #mojo
12:49 sri i wouldn't feel comfortable with vim modeline magic either
12:50 * marcus adds vim modeline magic
12:50 * sri removes it again
12:51 * marcus pushes a release with .tm_properties in it by accident
12:51 marcus guess I'll do judofyr's trick next time
12:52 judofyr alias release="mkdir release && git clone . release && cd release && rake"
12:53 judofyr just replace rake with prove -lv
12:53 sri isn't there a smarter way? like ignore all ^\.* files except for .perltidyrc?
12:57 marcus judofyr: or make manifest && make disttest
12:57 sri marcus: your tweet is missing the link to changes
12:57 judofyr marcus: I need to create a bash function which smartly runs the test suit
12:57 sri and #perl tag
12:57 sri perhaps we do need a release guide :)
12:57 sri or checklist
12:57 judofyr e.g. if Makefile.PL exists, use make. if Rakefile, use rake. if t/ use prove …
12:59 sri marcus++
12:59 marcus sri: deleted and retweeted.
13:00 sri gold star for the summary :)
13:01 judofyr btw, is there some standard Changelog format?
13:01 sri there are many standard ChangeLog formats
13:01 judofyr hehe
13:03 sri we use the one promoted by h2xs
13:04 sri https://github.com/rafl/cataly​st-runtime/blob/master/Changes # good example, went through multiple changes in style
13:09 crab hi
13:09 purl what's up, crab.
13:09 crab i'm going away for a few days.
13:09 * sri waves
13:09 marcus crab: did you finish jsonpointer first? :p
13:10 marcus crab: going birdwatching?
13:14 GitHub13 joined #mojo
13:14 GitHub13 [mojo] kraih pushed 1 new commit to master: http://git.io/nnUcBw
13:14 GitHub13 [mojo/master] made MANIFEST.SKIP a little smarter - Sebastian Riedel
13:14 GitHub13 left #mojo
13:14 sri marcus: that should solve it
13:14 marcus sri: yeah, nice
13:21 j3nnn1 joined #mojo
13:22 koban joined #mojo
13:25 sri we are waiting for jsonpointer! :)
13:26 marcus crab: if you have work in progress, maybe you could push it to a feature branch before leaving?
13:31 sri btw. i kinda like the kill a core dev to become one rule
13:31 d4rkie joined #mojo
13:32 memowe you don't like your core devs? ;)
13:35 sri well, "you keep what you kill" worked for the necromonger
13:36 sri …wait
13:40 * sri just saw rumors about another riddick movie
13:40 marcus I like my odds in this game
13:40 * marcus sharpens his viking axe
13:41 * sri hides
13:41 marcus more like sharpens my squash racket, I guess :p
13:42 sri nitpicking
13:42 purl nitpicking is a tool of oppression by The Man
13:44 * sri sharpens his iphone… just in case
13:50 marcus I think we should look up for crab and his deadly bird friends
13:50 marcus and tempire's unicorn horn
13:50 sri i might have not thought this through
13:58 elb0w this was awesome until I rebooted http://elbowrage.com/keyboard.jpg
14:05 elb0w LOL
14:10 crab marcus: (a) no, i haven't finished json-pointer. i wanted to, but there are lots of last-minute chores. (b) i've pushed to the json-pointer branch on github, there's currently a stub M::J::P class with the right API and some documentation, plus the changes to the other classes (Mojo::Message and Test::Mojo)
14:10 marcus crab++
14:11 marcus crab: what about c) birds? :)
14:11 crab marcus: (c) yes, sort of. my girlfriend and i are going to thailand to hang around at the best site in the world to see spoon-billed sandpipers. but we're really going there to eat seafood until we have lobsters coming out of our ears.
14:11 crab :-)
14:11 marcus excellent!
14:11 * purl plays air guitar
14:11 marcus enjoy thailand
14:11 * marcus has been there a couple of times, but not to watch birds.
14:12 crab sri/marcus: if i'm already a core developer, i don't have to kill you, right? unless you insist, of course.
14:13 marcus crab: it's fine for me if you don't
14:13 crab (what is necromonger? some kind of mongodb library)
14:13 crab s/.$/?)/
14:13 marcus just pointing out sri might be the weakest link if someone is looking to join the core.
14:14 sri :(
14:15 amoore joined #mojo
14:15 sri make sure to watch hangover2 before going to thailand, it's a great travel guide :)
14:15 marcus I was pretty dissapointed by that movie.
14:15 marcus not as much as bridesmaids, but still.
14:16 sri it wasn't that bad, maybe your expectations were too high
14:17 marcus well, I was measuring it against the first one
14:17 * sri nods
14:25 osprey joined #mojo
14:27 d4rkie joined #mojo
14:34 jmhgtx joined #mojo
14:39 jmhgtx Hi, I'm having some trouble getting my routing to work correctly. I'm setting $r->namespace('PathTo::Controller')  and then setting a route as  $r->get('/user')->to('user#userGet')  where User.pm is the name of the controller and userGet is the subroutine I'm trying to call.  When I run script/driver_file  routes, I get an output of '/user  GET user' . I'm not sure that that is correct.  In any case, I'm getting 404 errors when I try and access those routes.
14:39 jmhgtx Anyone see anything clearly wrong here?
14:42 cosimo joined #mojo
14:47 marty jmhgtx:   $r->get('/user')->to('user#userGet') does not look right.
14:47 jmhgtx I
14:48 jmhgtx I'm getting that syntax right from the docs.
14:48 marty did you mean  $r->route('/user')->via('get')->to('user#userGet')
14:51 jmhgtx Hmm. That seemed to work.  I could swear I tried that that yesterday but with a capital GET and it didn't work.
14:51 jmhgtx That syntax is in the docs though so I thought it would work.
14:52 jmhgtx Thanks
14:52 marty np. :)
14:54 sri marty: that route looks perfectly fine, except for userGet
14:54 judofyr no camels please
14:55 jmhgtx What's wrong with the .... oh.  Ok.
14:55 sri (which is not a bug, but ugly as hell and not officially supported)
14:55 marty Ahh ok.   makes sense.
14:55 jmhgtx I'm not crazy about it either.
14:56 sri $r->get()->to() is fine though
14:56 judofyr jmhgtx: also user#userGet is kinda redundant. why not user#show?
14:57 jmhgtx It's just a vestige of a Catalyst REST implementation. I was more concerned about getting it running. But yeah, you're right.
14:59 nightmare joined #mojo
15:00 marty jhmgtx:  can you paste a link to the docs where $r->get()->to() documented?
15:00 sri marty: it's part of the growing guide
15:01 * marty is reading
15:01 sri and Mojolicious::Routes reference and http://mojolicio.us/perldoc/Mojolicious/G​uides/Routing#Mojolicious3A3ALite_routes
15:01 jmhgtx Under the 'Application Class'
15:02 jmhgtx Also in https://metacpan.org/module/Mojolicious::Routes   and the bottom of the Synopsis.
15:02 sri it's what Mojolicious::Lite uses internally, so it's extremely well tested
15:02 sri as in thousands of tests :)
15:04 marty wouldn't you know it.  It's under Mojolicious::Lite routes.  Since I never use lite I skipped that section.   Guess I will not skip it any more.  :)
15:05 sri you should really know better by now ;p
15:05 * marty nods
15:06 judofyr sri: what do you think about Backbone/Knockout/Ember?
15:07 jmhgtx Is it correct to show the routes as /user GET /user ?  as the output from script/file routes?
15:08 jmhgtx That should be "/user GET user"
15:09 sri judofyr: not a fan of too heavy client side web apps
15:10 netfeed less server work though
15:10 sri i meant that as a developer *and* user
15:10 marcus I like it for writing mobile apps in phonegap :)
15:11 sri nothing is more irritating when browsing the web than those pseudo-apps
15:12 sri github does it right, facebook not
15:12 judofyr I don't think facebook is *that* bad
15:12 netfeed heh, there's a bit of NIH-syndrome at work on the UI-side so they don't use such fancy things like backbone, even though almost all in our 3 largest sections of the site is javascript :)
15:13 sri of course there are worse examples, facebook is just a good example for a website you don't just browse through
15:16 netfeed i wouldn't mind a desktop client or a simplified facebook ui for just browsing what people i know have posted and check events, i don't really need the rest
15:16 sri i'm one of those people who liked the old twitter a lot more than the new and new-new ones
15:17 judofyr I haven't got the new-new one yet :/
15:17 judofyr twitter's mobile web app is just crappy
15:17 sri WHAT KIND OF HIPSTER ARE YOU?
15:18 netfeed i'm one of those people who doesn't get the point of twitter, could be because most of my communication goes via irc, but... :)
15:18 sri well, new-new is better than new
15:18 judofyr new^2
15:18 sri they seem to get their act together eventually
15:18 elb0w I never got into twitter
15:18 elb0w I tried
15:18 elb0w I dont really see the allure of it
15:18 netfeed i'm not alone :)
15:19 elb0w I made it follow sri
15:19 elb0w then I was like ok now what
15:19 sri it's all about the 140 characters
15:19 judofyr READ EVERYTHING SRI SAYS
15:19 judofyr elb0w: that's step two
15:19 elb0w @hackernews @sri
15:19 elb0w thats it
15:19 elb0w rofl
15:20 sri i actually use twitter as my daily news feed these days
15:20 netfeed got google reader for that :)
15:20 netfeed give google all your data, let's assimilate in the google machine
15:20 sri easy to skim through
15:21 elb0w Meh, all that crap ends up on hn and reddit/programming
15:21 sri just using rss for blogs now, all news via twitter
15:21 marcus me too
15:21 marcus I actually read twitter through flipboard on the ipad.
15:21 marcus to pick up news from my stream
15:22 netfeed news-news or it-news?
15:23 sri both
15:23 metaperl joined #mojo
15:24 netfeed isn't it a bit limiting or is it just for the headlines?
15:24 sri headlines with links
15:25 sri https://twitter.com/#!/cnnbrk # example
15:26 sri https://twitter.com/#!/techcrunch # another one
15:26 netfeed alright
15:27 sri https://twitter.com/#!/newsycombinator # of course
15:28 sri oh, there's a new mongodb book from manning
15:28 sri bigger one this time
15:29 netfeed i rather go with the rss-feed for HN, links to the comments too and not just for the articles
15:29 inokenty joined #mojo
15:30 netfeed hmm, they're not the same though
15:32 crab judofyr: i'm cautiously optimistic about backbone/ender
15:35 crab but i might not actually use either of them.
15:37 sri marcus: got the riak book yet?
15:38 marcus sri: buying it tonight I guess
15:38 sri ah
15:40 judofyr sri: you've read it?
15:42 sri judofyr: nope, wanted to know from marcus if it was any good :)
15:43 marcus buying it as soon as I'm able to figure out what our company paypal account is. :)
15:44 marcus GOT IT
16:05 mattastrophe joined #mojo
16:07 sri let me know how it is
16:08 crab hm. it'll be next year soon, and then i can buy another book.
16:08 crab but, it won't be the riak book.
16:35 metaperl|2 joined #mojo
16:37 j3nnn1 joined #mojo
16:44 marcus crab: what's the limit?
16:44 purl the limit is 255
16:45 marcus O'RILLY?
16:45 noganex_ joined #mojo
16:46 kaare joined #mojo
16:48 inokenty joined #mojo
16:52 crab marcus: it's more like selecting a book and working towards seeing it on my bookshelf, not shopping to a limit.
16:54 marcus crab: oh. I buy all my books electronically. most of the work involves punching in credit card info.
16:56 jnap joined #mojo
16:56 marcus almost all my paper books are in the closet in paper boxes.
16:57 sri same
16:58 marcus I'm also glad to report that the CD is dead.
17:00 xaka joined #mojo
17:03 crab everyone's releasing a new js templating engine.
17:03 marcus any good ones?
17:04 marcus I wanted to test xslate
17:04 crab marcus: dust.js and handlebars.js both look promising. don't know how the new hogan.js (from twitter) uis yet.
17:04 marcus handlebars can't be that new?
17:05 crab no, it's hogan that i was thinking of just now.
17:05 marcus LICENSE file  is from April
17:05 arpadszasz joined #mojo
17:06 marcus crab: hogan looks a lot like mustache too :)
17:07 crab yes.
17:08 marcus didn't understand the difference, actually
17:09 marcus oh, it's faster
17:10 crab right.
17:18 MojoGuest837 joined #mojo
17:18 MojoGuest837 From: http://twitter.com/ (8 hits)
17:37 sri crab: there is no jsonpointer branch on github as far as i can see
17:54 mire_ joined #mojo
17:57 GitHub34 joined #mojo
17:57 GitHub34 [mojo] amenonsen created json-pointer (+3 new commits): http://git.io/tHWC0Q
17:57 GitHub34 [mojo/json-pointer] Add a stub Mojo::JSON::Pointer class - Abhijit Menon-Sen
17:57 GitHub34 [mojo/json-pointer] Make Message::json take an optional pointer argument - Abhijit Menon-Sen
17:57 GitHub34 [mojo/json-pointer] Add json_{is,has,has_not} tests to Test::Mojo - Abhijit Menon-Sen
17:57 GitHub34 left #mojo
17:58 crab sri: sorry, i'd pushed it to amenonsen/mojo.git by mistake
18:03 sri crab: maybe i'll finish it
18:03 crab sri: if you really want to, sure.
18:04 sri i'll leave t alone if you want to of course
18:05 xaka is there any way to subscribe on "new transaction" event from application class perspective? if such event exists, sure
18:05 crab sri: let me see if i can finish it tomorrow. if not, i'll tell you so before i leave.
18:06 crab it's >2330 now, so i should sleep. i'll try to sneak it in tomorrow morning.
18:06 sri you need to be faster, just wrote a solution ;p
18:09 sri just reused the path code from Mojo::Path and it was half finished
18:09 crab ok.
18:09 sri :/
18:11 crab doesn't matter. if the end product works, i don't care how it got there.
18:11 sri just trashed it, it's your baby
18:14 sri reusing Mojo::Path is a bad idea anyway, since its security features are unnecessary
18:14 crab ok. i shall apply bathwater liberally tomorrow morning.
18:14 crab 'night
18:14 sri nn
18:25 sri json_has_not should still be json_hasnt though ;p
18:31 cfedde The unicode in the topic is scewing up irssi for me.
18:34 xaka cfedde: ++
18:35 cfedde oh well.
18:35 cfedde left #mojo
18:36 cfedde joined #mojo
18:36 cfedde yeah.  No good.  I'll have to wait till I get home to fix this.
18:36 cfedde left #mojo
18:52 rsimoes irssi--
18:56 xaka sri: can i use Mojo::IOLoop::Stream for my own purpose?
18:58 sri xaka: no you can't
18:58 jnap joined #mojo
18:58 xaka ok, then i just copy/paste it if you don't mind :) i really need its functionallity
18:59 rsimoes I thought the Mojo distribution is a toolkit
18:59 xaka i mean to not depend on your future changes because it's experimental
18:59 rsimoes oh
18:59 rsimoes experimental in all caps
18:59 rsimoes roger
19:01 xaka the only thing i don't like in Stream is that it uses timeout by default always and you cant even disable it. I have 3rd party streams which should live forever till the error on read/write operation
19:06 xaka rsimoes: i'd call it toolkit because it has a lot wonderful features not only for web, i can easly adopt it to any 3rd party library, writting backend clients for your web service is just fun with it
19:07 xaka you just need to keep your eyes on changelog and commits for insurance :)
19:22 tempire kill?  no killing!
19:22 * tempire detains the core devs
19:23 * tempire invades their countries under the auspice of installing democracy
19:23 * tempire sets up puppets
19:23 * tempire plans to kill them later
19:23 xaka tempire: hello? o_O
19:29 marcus tempire: we're already democracies.
19:29 marcus or should that be democrazies?
19:30 marcus tempire: you're the one living in a country that practices the death penalty. oh and maybe crab in theory.
19:31 marcus (they hanged one guy back in 2004)
19:45 cfedde joined #mojo
19:48 Therstrium joined #mojo
19:51 tempire marcus: clearly your motivated by an agenda.  don't try to confuse the issue with facts
19:51 tempire *you're
19:52 GitHub121 joined #mojo
19:52 GitHub121 [mojo] kraih pushed 1 new commit to master: http://git.io/mttuug
19:52 GitHub121 [mojo/master] improved inactivity timeouts by allowing them to be disabled - Sebastian Riedel
19:52 GitHub121 left #mojo
19:52 tempire what country is crab in?
19:52 sri since the topic comes up every other week
19:52 * xaka beeeeeeeerrr
19:52 sri who am i to keep people from shooting themselves
19:54 xaka author of mojo ;)
19:55 mire_ joined #mojo
19:57 marcus tempire: india
19:58 xaka anyway i'm still waiting for the day when we can separate client inacivity timeout from server inactivity timeout
19:59 sri if anyone thinks 0 timeouts have no actual value i'll gladly remove them again
20:01 sri because i'm not entirely convinced they have
20:02 xaka for mojo itself it doesn't, even i think this way, but for people who use mojo as the toolkit (carefully, knowing that experimental is under changes always) it's useful
20:03 sri that looks like an argument for removing it again
20:03 cfedde apache proxy seems to break for urls to static files.  Is that fixable?
20:03 xaka the bigger problem, imho, is that stream inactivity timeout also working as server inactivity timeout. this one really should be separated
20:05 sri marcus/tempite: opinion on 0 timeouts?
20:05 sri ops
20:05 sri tempire
20:05 xaka there always should be an option, right? either you want it to be enabled or disabled, it's up to user. framework should have options
20:06 sri you couldn't be more wrong
20:06 sri frameworks are the opposite
20:08 sri and i think now i'm in fact removing 0 timeouts again
20:10 xaka if you don't like an idea of reusing mojo code by people, i don't mind. we still have an option - monkey patching :)
20:11 ki0 joined #mojo
20:15 marcus xaka: each feature/option you add increases the complexity of a framework. Adding unneeded ones makes the framework worse.
20:16 sri marcus: 0 timeout, yay or nay?
20:16 xaka marcus: i agree, but you forget to meausre such features/options. supporting disabled timeout doesn't cost anything
20:16 sri and that's completely wrong
20:16 sri it is maintenance overhead
20:17 sri nothing is free
20:17 xaka you saw your own commit, how many lines it is? ;)
20:19 tempire xaka: the problem is when that argument is used every time
20:20 tempire the reason he's so strict is because that's how catalyst got crazy
20:20 tempire has anyone ever presented a real use case?
20:20 xaka i like how mojo does, it teaches a lot, but sometimes it ties the hands of so hard when it could be less strict because it's possible and safe
20:21 tempire coding guidelines: "A feature is only needed when the majority of the userbase benefits from it."
20:21 tempire It's all about keeping consistent
20:21 GitHub84 joined #mojo
20:21 GitHub84 [mojo] kraih pushed 1 new commit to master: http://git.io/WQ2Sug
20:21 GitHub84 [mojo/master] removed ability to disable inactivity timeouts again - Sebastian Riedel
20:21 GitHub84 left #mojo
20:22 tempire I like the idea of having it available, but I also like the hard rule of having use cases
20:22 sri no sensible use case, no unit test, no go
20:22 xaka tempire: real case? my backend is dbus service and i use Net::DBus + signals. Net::DBus requires reactor and i adopted IOLoop to be the reactor. I use iowatcher instead of stream because timeout kills dbus connection, but i dont need it
20:23 xaka dbus is tcp/ip long-leaving connection saying in easy words, it has its own timeout also for some cases
20:23 sri funny how you talked me out of it after asking for it
20:23 tempire that's a good example of where to override or monkey patch.  if you're going that far to use ioloop separate from mojo (which is fine), it's perfectly reasonable to have a layer that customizes it to your needs
20:24 xaka tempire: using Stream could simplify my code so much, but this timeout doesn't allow
20:25 tempire you know what would be neat
20:25 tempire a way to inject events using an api
20:26 tempire right now it only does that in specific places.  but to be able to say, I want this event to happen right..here.
20:26 xaka anothen use case is async mongodb connection in another project. different projects doesn't share the same code base so i have to override again just because of timeout. jus because i can't disable it i have to do really lazy things
20:27 tempire wonder how xantus' esb thingy based on ioloop is coming long
20:28 marcus I'm actually pro disabling the timeout with 0, sri.
20:29 marcus for the record
20:30 marcus I just don't agree with xaka's sentiment in general.
20:30 xaka tempire: so, do you like my use cases? :) it's all about backends and adopting them to use ioloop (stream). i have another project where net::ssh2 is a backend and timeout again ties me
20:31 tempire I understand where you're coming from, but ioloop is not intended to be used that way.  you're free to do so, you just have to build a layer that customizes it to your liking.
20:31 sri marcus: why are you pro disabling?
20:32 marcus sri: I think you should be able to remove the safety equipment if you're a pro.
20:32 xaka tempire: if i won't use it, i have to somehow link together both reactors - mojo one and dbus one, worse stuff as you understand
20:32 sri marcus: sure, but why?
20:32 sri i think the right question is, what do we gain by allowing it?
20:33 xaka tempire: and if library allows me use different reactor, why not?
20:33 marcus sri: makes mojo ioloop more versatile in long poll situations.
20:33 sri new behavior a user wants to rely on is maintenance overhead
20:33 sri marcus: i disagree actually
20:34 marcus sri: but we already need to support different timeouts
20:34 xaka marcus: long polling is more about separation of client timeout from server timeout i believe
20:34 sri timeouts are tested, 0 timeout is a special case that is untested
20:34 marcus sri: I suggest that if xaka wants he, he should contribute an unit test.
20:34 xaka sri: just to change topic a bit. what do you think about such kind of separation?
20:35 tempire xaka: are you referring to the usage of ioloop in that manner?
20:35 xaka tempire: sorry?
20:35 purl It's okay, xaka.
20:35 tempire regarding the question you asked sri
20:36 sri marcus: in long polling you need to support reconnects anyway, i don't see how unlimited timeout is any better than a one week timeout for example
20:36 xaka tempire: ah, no. it's long time ago topic, i don't an idea that stream's timeout working as server inactivity timeout. i'd like be able configure it separatly
20:37 marcus sri: I actually meant the other end of the long poll, talking to some backend service for a long time.@
20:38 sri marcus: doesn't that also require reconnect?
20:38 xaka browsers support disabling connection/activity timeout
20:38 xaka so you don't neet to reconnect
20:39 xaka just wait till the end of your days for response :)
20:40 marcus sri: I'm pretty happy that I can keep a ssh connection open without need for reconnecting, even if you can do it.
20:40 sri marcus: btw. sane long poll apis like twitter send whitespace characters every now and then to keep the connection alive
20:41 * marcus has been swearing in the past about firewalls killing idle ssh connections, and actually creating keepalive crap to avoid it.
20:41 xaka so, to refresh the air. 2 topics of the day: 1. allow to disable stream's timeout 2. separate client and server timeouts
20:41 sri tempire: where do you stand on timeout 0, yay or nay?
20:42 xaka purl: do you support polls?
20:42 purl xaka: wish i knew
20:42 sri whatever tempire says will be done
20:43 tempire just another tuesday :)
20:43 tempire you both have valid points.  let me think about it for a couple of hours.
20:43 sri yay or nay? :)
20:43 sri :S
20:44 sri no easy decision is a no in my book
20:45 xaka tempire: you have to save the world! sri is going to put no mark in the book
20:46 xaka what about 2nd topic, core team?
20:47 chansen joined #mojo
20:49 xaka marcus: btw, ssh has keepalive packets, how it could be idle?
20:50 marcus xaka: keepalive for ssh is a config setting, and I believe it didn't have it 10 years ago
20:51 marcus xaka: I really don't care about separating server and client timeouts
20:57 GitHub178 joined #mojo
20:57 GitHub178 [mojo] kraih pushed 1 new commit to master: http://git.io/MKSWIQ
20:57 GitHub178 [mojo/master] brought back ability to disable inactivity timeouts with test - Sebastian Riedel
20:57 GitHub178 left #mojo
20:57 sri for marcus
20:58 marcus xaka: merry christmas :p
20:58 xaka marcus: teach me how you did it!
20:58 marcus xaka: you could have done it as well, by shutting up :D
20:58 sri i can confirm that
20:59 * xaka confused
20:59 purl You won't be after this episode of Soap!
21:00 marcus xaka: you're the one who talked sri into removing it again after he added it an hour ago :)
21:00 marcus ne ponimasish?
21:01 marcus -s
21:14 mattastrophe joined #mojo
21:48 batman joined #mojo
21:49 batman joined #mojo
21:50 xaka marcus: da vse ya ponimayu, ti cho :)
21:51 marcus xaka molodetz!
22:12 MojoGuest746 joined #mojo
22:12 MojoGuest746 From: http://www.google.com/url?sa=t&amp;rct=j&amp​;q=irc%20websocket&amp;source=web&amp;cd=3&a​mp;ved=0CC4QFjAC&amp;url=http%3A%2F%2Fdev.xa​ntus.org%2F&amp;ei=B6vzTozlC5OAsgL56OyuAQ&am​p;usg=AFQjCNGw6nZv0wO42uzUnHhWcD0vcKNULg (1 hits)
22:12 MojoGuest746 hello?
22:12 purl hello is this thing on?
22:12 Success oh cool i didn't think it really works
22:12 Success gui is weird though but w.e
22:32 jnap joined #mojo
22:46 ccushing can I change where the config file is at runtime?
22:47 ccushing I thought maybe MOJO_HOME had something to to do with where it looks for them
22:47 ccushing but I set that
22:47 ccushing and it's still looking in my perlbrew directory
22:49 marcus ccushing: at runtime?
22:49 purl i think at runtime is better than not at all.. not sure setup time is possible..
22:50 marcus ccushing: it's read at startup time, so...
22:50 ccushing marcus well at startup time then
22:50 marcus you mean without changing the code?
22:50 ccushing yeah
22:51 mercutioviz joined #mojo
22:52 marcus ccushing: MOJO_CONFIG env variable
22:52 ccushing marcus thanks
22:53 marcus ccushing: it's documented in Mojolicious::Plugin::Config
22:53 marcus and the Cheatsheet guide
23:31 tempire it's amazing to me that sql injection is still a problem
23:32 metaperl joined #mojo
23:36 xaka hacked something? )
23:39 mattastrophe joined #mojo
23:40 j3nnn1 joined #mojo
23:45 Kallisti joined #mojo

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