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

IRC log for #mojo, 2015-02-17

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

All times shown according to UTC.

Time Nick Message
00:01 jberger I would more likely say, if it isn't tested and thus not a feature, it probably shouldn't be documented either
00:02 jberger if it can just go away, and because of accessor method not existing, apps that worked last version die on this one
00:02 jberger nm, it has been a long day at work
00:03 jberger probably not my best time to comment on this
00:03 sri simple solution, write more tests
00:03 sri when was the last time any of you wrote a test for an untested feature?
00:03 jberger that doesn't fix your removing of things that weren't tested before
00:04 jberger lets list them
00:04 jberger I would do so, and maybe its another thing for GSoC
00:04 * sri is not volunteering
00:05 jberger all I am suggesting is leaving a stub method of the same name as the attribute with deprecated called
00:05 jberger I don't even care if it does what its supposed to do
00:05 disco joined #mojo
00:05 sri finding out which parts are not tested does seem like a sensible gsoc project
00:06 jberger that it does
00:07 jberger and -a, suddenly people are going to be having <1 connection per worker?
00:10 jberger -A -> -a isn't even in the Changes
00:14 jberger and for the record, some of that is tested, perhaps not well tested, but the tests exist
00:14 jberger things like lock_file
00:14 jberger so if a user call ->lock_file and it now dies, what do we say? it wasn't tested well enough to be considered a feature?
00:15 jberger I say if we want to remove a bunch of cruft fine, lets do it at the major release, which we can do at any time
00:21 good_news_everyon joined #mojo
00:21 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Ndm4
00:21 good_news_everyon mojo/master 7b70d90 Sebastian Riedel: make sure the event loop can be stopped in regular intervals
00:21 good_news_everyon left #mojo
00:21 sri that one is totally untested too
00:21 sri SIGINT and SIGTERM are not officially supported
00:24 sri it's funny, i think i might have made a mistake with removing all recurring timers :)
00:25 sri if there's only a 15 second timer in the reactor, the reactor will sleep for 15 seconds
00:27 zivester joined #mojo
00:29 sri some actual technical feedback would be really valuable
00:32 bpmedley sri: I’m somewhat ignorant.  Would the new $SIG{INT} and $SIG{TERM} stop the app or the IOLoop?  Is there a difference?
00:32 sri the ioloop
00:32 purl the ioloop is 100% not the bottleneck, see microhttpd ;p
00:33 sri the problem is safe signals can't interrupt the very long syscall that happens in the reactor
00:33 sri which is a usleep or poll
00:37 bpmedley While I’m looking at the code, would startup scripts and whatnot have to be changed for -a vs -A change?
00:37 sri that change has long been reversed
00:37 bpmedley My bad
00:38 asarch joined #mojo
00:38 Grinnz sri, were you using the EV or poll reactor?
00:39 Grinnz i had an issue with EV where i had to add a watcher that did nothing so that it could handle perl signals
00:39 sri both
00:39 Grinnz hmm
00:39 sri it's a funny problem
00:40 Grinnz https://metacpan.org/pod/EV#PERL-SIGNALS
00:40 Grinnz is where i got it from
00:40 Grinnz poll shouldnt have that specific problem though
00:40 sri Mojo::IOLoop used to have a recurring timer that fired every second and checked if an attribute had been set to 0, and depending on that did a graceful shutdown
00:42 bpmedley sri: Last question; does the recurring 1 event cause the ioloop to break out of the syscall every second to give the SIGTERM and SIGINT time to be processed?
00:42 sri yes
00:42 sri suppose i could use EV::check and make Mojo::Reactor::Poll limit the sleep time to one second or so
00:42 Grinnz ah, so with the poll reactor it gets stuck in the _poll syscall
00:42 davido_ltp joined #mojo
00:44 punter joined #mojo
00:44 cpan_mojo Mojolicious-Plugin-MimeTypes 0.01 by Stefan Adams - http://metacpan.org/release/SADAMS/Mojolicious-Plugin-MimeTypes-0.01
00:44 good_news_everyon joined #mojo
00:44 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NdG9
00:44 good_news_everyon mojo/master 93e6c27 Sebastian Riedel: mention the shortcut as well
00:44 good_news_everyon left #mojo
00:50 bpmedley sri: Perhaps some usage of sigprocmask to determine if there is a waiting signal after your call of _poll with some timeout logic.  The timeout logic might break out of _poll ever 100 ms, check for signals, then determine if the primary timeout has been reached.
00:51 bpmedley Does that make sense?
00:51 sri bpmedley: sounds complicated, i don't like complicated :)
00:52 bpmedley Hrmm.   Is it the fancy timeout?
00:52 sri yes
00:56 bpmedley Move the application logic at [1] to [2]?
00:56 bpmedley [1] https://github.com/kraih/mojo/blob/7b70d90ff30ebdf1a66f0e68db8021aa324ccffb/lib/Mojo/Server/Daemon.pm#L32
00:56 bpmedley [2] https://github.com/kraih/mojo/blob/7b70d90ff30ebdf1a66f0e68db8021aa324ccffb/lib/Mojo/Reactor/Poll.pm#L60
00:57 bpmedley And at [2] you could use the sigprocmask dance to determine if a signal is pending…
00:58 bpmedley I’m not sure that’s any better, especially since what you have probably works.
00:58 sri it works pretty well
00:58 sri prefork workers already have a recurring timer for sending heartbeats
01:08 good_news_everyon joined #mojo
01:08 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Ndli
01:08 good_news_everyon mojo/master 7dfd839 Sebastian Riedel: get rid of a few more warnings
01:08 good_news_everyon left #mojo
01:40 disco joined #mojo
01:41 n2plato joined #mojo
01:44 good_news_everyon joined #mojo
01:44 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NdgI
01:44 good_news_everyon mojo/master f72aca6 Sebastian Riedel: do not pass around transactions
01:44 good_news_everyon left #mojo
01:59 good_news_everyon joined #mojo
01:59 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NdVk
01:59 good_news_everyon mojo/master 767fb65 Sebastian Riedel: slightly faster reset
01:59 good_news_everyon left #mojo
02:09 disco joined #mojo
02:24 hy1001 joined #mojo
02:28 hy1001 joined #mojo
02:42 sri hmm, so i just wanted to document the signals Mojo::Server::Daemon responds to... but there's no tests... so i suppose they better stay undocumented
02:47 sri guess almost all signal docs have to go
02:48 ilbot2 joined #mojo
02:48 Topic for #mojo is now ???? hear me roar | http://mojolicio.us | http://irclog.perlgeek.de/mojo/today
02:49 klapperl joined #mojo
02:50 good_news_everyon joined #mojo
02:50 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Ndyc
02:50 good_news_everyon mojo/master 33f4c6c Sebastian Riedel: mention signals too
02:50 good_news_everyon left #mojo
02:50 sri actually, this upsets me quite a bit
02:50 sri there's lots of talk here, but very rarely any action anymore
02:51 sri there's actual tricky technical problems to solve, and the problem of the day is procedure
02:51 * sri switches to angry linus mode
02:52 Grinnz i think the concern is how it could affect people who are currently running apps, rather than the procedure itself
02:52 sri if you're running "apps", there is literally no effect whatsoever
02:53 sri if you're running a customized prefork daemon...
02:53 sri GET THE FUCK HERE, I WANT TO GET TO KNOW YOU!
02:53 Grinnz hehe
02:54 Grinnz yeah can't say i'm doing anything more than running morbo and hypnotoad
02:54 firnsy same here
02:54 sri hypnotoad config setting and command line args will just be ignored and are totally safe
02:54 firnsy TIL there was an -a option
02:55 sri the only way you'll stumble over this is a customized prefork server
02:55 sri and seriously, if you do that, you should be here and talk architecture with me!
02:56 sri if anything, we have a problem with not enough people hacking on that stuff
02:56 Grinnz in the end, there will always be these untestable things especially related to prefork
03:03 noganex_ joined #mojo
03:12 sri for the record, i kinda like how node.js labels individual apis with a stability rating http://nodejs.org/api/crypto.html#crypto_crypto
03:17 sri of course i'm questioning how much jberger really cares, considering he didn't even open a pull request
03:17 sri so i'm gonna leave it now
03:17 hshong joined #mojo
03:18 sri (technical feedback: yay, discussing procedure: nay)
03:36 hy1001 joined #mojo
04:00 jnbek joined #mojo
04:20 inokenty-w joined #mojo
04:27 Oleg joined #mojo
04:33 jberger sorry, I had life going on
04:33 jberger and for the record I didn't disagree with anything
04:34 jberger I don't always have the time to make a pull request either (who did the wantarray fix in the middle of the night? )
04:35 jberger I was just noting that there was a lot of api change
04:35 jberger and no it probably won't affect many people
04:35 jberger but I'm not sure I like that precedent
04:37 jberger I've told you what my pr would be, roll to 6.0 and break whatever we want
04:37 jberger then there are no worries
04:39 jberger but hey if we want to have api breakage (If in a dark corner) then that is  up to the group
04:43 good_news_everyon joined #mojo
04:43 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NFOm
04:43 good_news_everyon mojo/master 1c1a662 Sebastian Riedel: fix bug in Mojo::IOLoop where connection limits were not properly enforced
04:43 good_news_everyon left #mojo
04:51 sri jberger: but 6.0 means i have to write a blog again and draw something
04:51 sri and if the changes don't work out, we're screwed
04:52 sri 7.0 two weeks later?
05:00 jberger I'm tired of arguing
05:00 jberger I think your changes make sense
05:01 jberger I also kinda like the stability rating
05:03 jberger but I think we could achieve a similar goal just by making out what is lacking coverage our where apis are shaky
05:04 jberger but would it have any force of effect? would we consider that when we would look to make a change? If not then I think it would be an illusion
05:04 jberger If so, it might be nice
05:05 jberger sometimes we do know where there might be future breakage, and that might be nice to convey
05:06 jberger and who knows, perhaps noting where test coverage is weak will help give volunteers something to hack on if they want to learn/help
05:16 good_news_everyon joined #mojo
05:16 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NFnt
05:16 good_news_everyon mojo/master 6d2ef7a Sebastian Riedel: generate multi_accept value only once
05:16 good_news_everyon left #mojo
05:18 sri kinda fun to look at MOJO_IOLOOP_DEBUG=1 now
05:18 sri to get a feel for the concurrency in an async app
05:21 dotandimet joined #mojo
05:24 jberger oh I never did rerun the chat
05:25 * jberger makes mental note for tomorrow
05:25 good_news_everyon joined #mojo
05:25 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NFc2
05:25 good_news_everyon mojo/master 7cf093a Sebastian Riedel: no need to assign multi_accept more than once
05:25 good_news_everyon left #mojo
05:26 sri jberger: you were talking about deprecating those attributes before, could have made a pull request for that
05:39 sri btw. my initial commit was full of flaws
05:39 sri in fact i'm still finding them
05:57 jberger I'll take a crack at it in the morning
05:58 jberger I haven't been at an actual computer sine I first read the patch
05:58 jberger sorry about that, yes my complaining and not patching is not a good m.o.
05:58 marmez joined #mojo
06:01 dotandimet1 joined #mojo
06:01 good_news_everyon joined #mojo
06:01 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NFBp
06:01 good_news_everyon mojo/master b563152 Sebastian Riedel: allow Mojo::IOLoop to signal when it is shutting down gracefully
06:01 good_news_everyon left #mojo
06:02 sri this might actually be the most controversial change
06:02 sri somehow the prefork server needs to know that the event loop wants to shut down
06:03 sri it used to be signaled by ->max_conenctions getting set to 0
06:03 sri which is a bit clumsy
06:03 sri making Mojo::IOLoop and eventemitter is a bigger change
06:04 sri but a finish event there means that apps can hook into that and clean up faster so hypnotoad can faster restart
06:05 sri websockets tend to just stay open until the manager kill it with a SIGKILL
06:06 sri downside, this is kind of server specific
06:06 sri it makes no sense for a single process server
06:06 sri so possibly confusion
06:16 Grinnz good to have a standard way to find out, probably
06:24 sri having it reactive should be good for shutting down persistent backend connections
06:24 sri those would delay the restart otherwise
06:25 sri lol
06:25 sri that's actually a reason to have $ua->reset now :)
06:25 sri Mojo::IOLoop->singleton->on(finish => sub { app->ua->reset });
06:26 Grinnz left #mojo
06:26 Grinnz joined #mojo
06:26 Grinnz whoops
06:33 good_news_everyon joined #mojo
06:33 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NF26
06:33 good_news_everyon mojo/master 035611a Sebastian Riedel: better description for finish event
06:33 good_news_everyon left #mojo
06:35 dotandimet joined #mojo
06:36 sri finish is kinda our standard name for that kind of event... maybe it's different enough to be called "graceful" or so
06:51 sujithm joined #mojo
06:52 sujithm joined #mojo
06:54 basiliscos joined #mojo
06:55 marcusr joined #mojo
06:59 melo joined #mojo
07:01 dotandimet joined #mojo
07:23 kaare_ joined #mojo
07:26 Insane joined #mojo
07:36 McA joined #mojo
07:41 Insane joined #mojo
08:01 reneeb joined #mojo
08:02 amon joined #mojo
08:03 eseyman joined #mojo
08:20 fhelmber_ joined #mojo
08:26 Vandal joined #mojo
08:27 trone joined #mojo
08:29 avkhozov Seems that "mojo version" command strange detect the version of Net::DNS::Native.
08:29 avkhozov Sample http://pastie.org/9955901
08:29 avkhozov Any ideas why?
08:40 Snelius avkhozov: Order modules loading issue. DNS::Native is good working.
08:40 Snelius For datails: https://github.com/kraih/mojo/issues/716 https://groups.google.com/forum/#!topic/mojolicious/lGNpgW4cfow
08:42 Snelius avkhozov: U should load DNS::Native before Mojo
08:42 avkhozov Hmm, ok.
08:42 avkhozov Thanks!
09:06 vytas joined #mojo
09:19 teejay joined #mojo
09:19 jontaylor joined #mojo
09:47 teejay joined #mojo
10:17 kwa joined #mojo
10:18 csson joined #mojo
11:01 bzero joined #mojo
11:49 teejay joined #mojo
11:57 dod joined #mojo
12:00 teejay joined #mojo
12:11 bc547 joined #mojo
12:12 Insane joined #mojo
12:28 marcusr This is brilliant https://gist.github.com/piscisaureus/3342247
12:44 neilhwatson joined #mojo
12:47 teejay joined #mojo
12:55 dotandimet joined #mojo
13:00 marmez left #mojo
13:17 mishantil bu
13:17 mishantil Wrong hole, sorry.
13:17 jberger marcusr: cool
13:35 punter joined #mojo
14:08 jbob joined #mojo
14:09 punter long live perl
14:11 zivester joined #mojo
14:20 gryphon joined #mojo
14:21 jbob Hi, I have small issue with hypnotoad/mojo/mango (all the latest versions). Whenever a page is accessed that displays data from a MongoDB collection the hypnotoad worker that serverd the request stays at ~3-4% CPU usage even after the request is finished. This doesn't happen if morbo instead of hypnotoad is used or the page doesn't fetch data from Mongo.
14:22 jbob What is it possible to debug this, or what could be wrong? Is it maybe even expected behaviour? A sample controller method which results in this issue: http://paste42.de/7882/
14:23 jbob *How
14:23 bc547 joined #mojo
14:31 basiliscos joined #mojo
14:34 sri avkhozov: https://metacpan.org/pod/Net::DNS::Native#INSTALLATION-WARNING
14:35 sri jbob: lol, you might like the next release
14:36 sri that's literally what i've been working on yesterday
14:37 jbob Wow, what a coincidence :-)
14:37 gryphon_ joined #mojo
14:37 sri if you want to make sure it's released soon please test https://github.com/kraih/mojo
14:38 jbob Will do
15:34 kwa joined #mojo
15:41 davido_ltp joined #mojo
15:44 davido_ltp joined #mojo
15:45 davido___ joined #mojo
15:53 Ptolemarch joined #mojo
15:59 tencendur joined #mojo
16:01 good_news_everyon joined #mojo
16:01 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NAKn
16:01 good_news_everyon mojo/master 20c14d5 Sebastian Riedel: turn removals into deprecations
16:01 good_news_everyon left #mojo
16:11 good_news_everyon joined #mojo
16:11 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NA10
16:11 good_news_everyon mojo/master 3fd9ed3 Sebastian Riedel: allow Mojo::Reactor::Poll to stop a little faster
16:11 good_news_everyon left #mojo
16:15 cfedde How can I get the client IP/FQDN from a request?
16:16 sri http://mojolicio.us/perldoc/Mojolicious/Controller#tx
16:16 Grinnz_ i dunno about FQDN but the IP... yeah that
16:16 cfedde awesome!
16:16 * sri still wonders if finish is the right name for this event http://mojolicio.us/perldoc/Mojo/IOLoop#finish
16:25 sh4 joined #mojo
16:28 sri omg, a good pull request \o/
16:30 Grinnz_ sri: i can see reasons for calling it finish and reasons for calling it something different
16:30 sri Grinnz_: you're not exactly helping :)
16:30 Grinnz_ :P
16:31 Grinnz_ i think it's specific enough to warrant a different name
16:32 good_news_everyon joined #mojo
16:32 good_news_everyon [mojo] kraih pushed 2 new commits to master: http://git.io/NAFx
16:32 good_news_everyon mojo/master 010c68e David Webb: Fix regression - some web frameworks terminate cookies with semicolons
16:32 good_news_everyon mojo/master 160aef4 Sebastian Riedel: Merge pull request #741 from dmw397/master...
16:32 good_news_everyon left #mojo
16:36 good_news_everyon joined #mojo
16:36 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NAAI
16:36 good_news_everyon mojo/master fac9d9c Sebastian Riedel: update Changes
16:36 good_news_everyon left #mojo
16:38 Oleg joined #mojo
16:42 sri hmm, just stumbled over this http://mojolicio.us/perldoc/Mojo/Server/Prefork#finish
16:42 sri i suppose it's consistent
16:53 Ptolemarch joined #mojo
17:31 jberger sri: sent you an email
17:46 RenatoCRON joined #mojo
17:48 * sri reads
17:48 ua_ joined #mojo
17:49 sri :D
17:52 jberger I take that as a :+1:
17:52 jberger \o\
17:52 jberger /o/
17:53 sri looks pretty good
17:53 sri of course i'm no expert on new york
17:53 Grinnz_ i live 40 miles from new york <_<
17:54 sri then i suppose we'll see you in june :)
17:55 jberger assuming this holds, we would need to get the word out soon, so people could coordinate YAPC::NA travel plans
17:56 sri right
17:56 jberger which seems to be the point
17:56 sri it's a bit late in the game again
17:57 jberger yeah
17:57 jberger but I think next year will likely be copenhagen and from there we will have a proper bit process
17:57 jberger bid
17:58 * sri nods
17:58 jberger I think we should actually start a bid process for next year too, but with priority to them since we bumped them a bit
17:58 sri i'm cool with copenhagen next year
17:58 jberger but I would like to see bid deadlines especially, so that we can get on a better schedule
17:58 jberger me too
17:59 sujithm joined #mojo
17:59 sri any word on sponsors for new york?
17:59 jberger I'm assuming that is part of Ka Yi's mandate
18:00 sujithm joined #mojo
18:00 jberger I would also assume that riche has some contacts
18:00 jberger then we get to rope in NYC.pm
18:00 sri but you're right, most important is getting the date out
18:01 sri so peoples KEEP JUNE 4-6 FREE!!!
18:01 jberger other side of the country, but maybe jzawodn would still be interested in sponsoring, hint hint
18:01 * sri nudges jzawodn
18:10 punter joined #mojo
18:23 good_news_everyon joined #mojo
18:23 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NxQi
18:23 good_news_everyon mojo/master afa068e Sebastian Riedel: fix typo in Changes
18:23 good_news_everyon left #mojo
18:24 sri jbob: so, did it work?
18:28 nicomen neat-o
18:42 marcusr jberger: is tempire in the loop with regards to training?
18:43 * tempire is not
18:44 marcusr tempire: do you want to be? I think training day probably going to be an important aspect of financing this thing.
18:44 tempire I'm down with it.
18:44 marcusr good. :)
18:45 tempire sri, jberger: craigslist is open to sponsoring, I can forward the contact information of the people I was in contact with at HR
18:45 jberger tempire: emailed
18:45 jberger send that info to Ka Yi
18:51 tempire So is this all public info?
18:53 marcusr tempire: it certainly is now
18:53 marcusr given that this channel is logged in public
18:53 tempire Indeed.
18:54 tempire Ok, the craigslist contact and Ka Yi have been introduced.
18:54 marcusr <3
18:55 tempire Well cool. I didn't believe the rumors.
18:56 tempire And I can actually go given the dates.
18:56 marcusr \o/
18:57 marcusr We should probably contact our sponors from last year as well and see if they want a repeat show =]
18:58 sri <3
18:58 tempire I liked that startsiden guy.
18:58 tempire Seemed like a cool chap.
19:03 batman i really hope ka yi can pull it off :-)
19:03 batman i feel like shit not replying to the emails until today.
19:06 sri jberger: if there's anything that needs to be announced, just throw the information my way
19:06 jberger lets wait for Ka Yi to confirm, then we would announce
19:06 sri yea
19:07 sri no site to link to yet either, so just a tweet so people know about the date
19:08 jberger batman, the old site is still around? could it be reused?
19:09 jberger also, I'm going to send her an email saying the core team agrees, since it seems we do
19:09 jberger if someone doesn't say so now!
19:09 sri +1
19:09 purl 1
19:09 jberger crab: is it possible to get you to NY?
19:11 jberger sent
19:12 jberger batman: I see your email
19:13 sri apparently i'm out of the loop ;p
19:13 jberger sri: IIRC you asked to be
19:13 jberger I looped in marcus and batman because they had info from last year
19:14 sri ah, ok
19:14 tempire I have the artwork I created.
19:14 tempire It's pretty cool, I think. Once I have a date, I can put it up.
19:14 tempire A confirmed date, I mean.
19:14 batman jberger: +1 on nyc
19:15 batman jberger: "the old site is still around could it be reused": not any more than what i describe in the email
19:15 sri jberger: you can totally cc me though, then i won't miss the announcement stuff at least
19:16 jberger k
19:16 batman sri: i will cc you, when i reply with the budget details...
19:16 * batman is translating
19:17 sri jberger: btw. https://github.com/kraih/mojo/commit/20c14d5b7c921445e737ffa6c35367de4583b6cb
19:18 jberger sri++ thanks, that was on my docket for tonight
19:18 sri read the ioloop internals instead, they sre surprisingly readable now
19:19 sri like, literally no magic anymore
19:21 jberger yeah, theres nothing to it :-)
19:21 jberger and you think the reason is that kernels have improved?
19:22 jberger (the reason you can do this without mutexes/locks)
19:22 sri about that i'm only guessing
19:22 sri not sure we ever really needed an accept mutex
19:23 jberger ah
19:23 sri we are just not fast enough
19:23 sri even for nginx it's a sub 10% difference in performance
19:23 basiliscos joined #mojo
19:24 jberger I will read the changes more in-depth tonight
19:24 sri of course the reactors goa lot better, so all workers waking up at the same time is just meh now regarding cpu
19:24 sri s/goa/got/
19:24 jberger in a topic that is only related if you know what I'm working on, do people here use JIRA or something like it?
19:25 jberger we are looking for something, and so far JIRA seems the leader (as sad as that is)
19:25 sri my last interaction with jira was for mongodb :S
19:25 sri they use it religiously
19:26 dotandimet joined #mojo
19:29 sri those deprecations brough test coverage down to 91.9% again :(
19:32 sri suppose all deprecations bring us to 92.4%
19:34 disputin joined #mojo
19:34 batman where did tempire go?
19:35 sri if i had to guess i'd say dancing
19:35 batman :)
19:40 batman i don't get why there are two email threads... I might have cc'ed you on the "old" one sri :(
19:42 jberger the perils of planning over email
19:42 jberger should I try to get her on to irc?
19:42 jberger mibbit to the rescue?
19:42 batman sure.
19:43 batman make a github project where we can share files and comment on issues
19:46 disputin joined #mojo
19:48 jberger hmmmmm, just trying mibbit again, their ads are ... a little ... less than professional
19:48 jberger :/
19:48 * batman doesn't know what mibbit is :P
19:48 jberger online irc client
19:49 batman aha. i've heard of another online client........... ;)
19:49 jberger indeed
19:49 jberger :-)
19:49 batman i can make an account
19:49 jberger sure
19:50 mst the nice thing about mibbit is you can just use a bounce link
19:51 mst e.g. https://chat.mibbit.com/#dbix-class@irc.perl.org
19:51 mst they originally had a hard coded list of allowed networks. I asked nicely and they changed it :D
19:51 hahainternet the nice thing about mibbit is it encodes the source IP address in the user's username
19:52 hahainternet so you can easily block
19:52 Ptolemarch joined #mojo
19:57 dotandimet joined #mojo
20:00 cpan_mojo Toadfarm 0.47 by Jan Henning Thorsen - http://metacpan.org/release/JHTHORSEN/Toadfarm-0.47 (depends on Mojolicious)
20:02 Grinnz_ btw batman, new mojo-irc release is great. now i just need to actually sit down and write tests and docs for my bot
20:02 batman Grinnz_: nice to hear! :)
20:05 jberger ooooh, I should update my convos
20:14 sri coincidentally mojolicious 5.80 should be very very good for convos :)
20:16 batman yeah. i can imagine :)
20:18 sri is convis usually deplyed with daemon or prefork?
20:18 sri s/i/o/
20:19 sri (i actually tried fixing "convis" and typoed it twice...)
20:20 Grinnz_ cinvis'
20:21 batman sri: prefork
20:21 batman sri: but not next version... (whenever it will be done)
20:23 sri wait, you're going minimalistic? :o
20:24 batman trying my best.
20:24 batman but time is not on my side
20:24 batman time--
20:25 batman sri: s/typeface/font/ (ref: email) :)
20:27 TitanOfOld joined #mojo
20:27 marmez joined #mojo
20:28 TitanOfOld How would I mock up single sign on for testing?
20:32 jberger TitanOfOld: http://trout.me.uk/vague.jpg
20:41 sri btw. in case anyone missed this little workaround yesterday https://github.com/kraih/mojo/blob/master/lib/Mojo/Server/Daemon.pm#L31
20:41 jberger HA
20:42 sri as it turns out, once i removed all recurring timers from the ioloop internals, our reactors wouls spend *a lot* of time in poll()/select()
20:42 sri which in turn prevents safe signals from having any effect for that time
20:42 Grinnz_ which is the same problem i had with a POE test for POE::Loop::EV where it starts the kernel without adding anything
20:43 sri it's a funny problem
20:43 Grinnz_ and then tries to stop it with SIGALRM
20:43 Grinnz_ lol
20:44 sri prefork workers actually have a 5s recurring timer for sending heartbeats to the manager, so they are immune to the problem
20:44 sri but daemon needs a workaround :)
20:45 sri especially since you want to stop daemon all the time with ctrl+c
20:46 Grinnz_ i actually run morbo as a daemonized process as well heh
20:46 Grinnz_ but it still needs to get signals
20:46 sri i hope you mean daemon, not morbo
20:47 Grinnz_ no, morbo, its a dev env
20:47 sri oh
20:47 Anon021 joined #mojo
20:47 Grinnz_ (it works ok as long as i don't chdir to /)
20:50 Grinnz_ if i had to tell other developers how to keep their morbo server running in a terminal i probably would not have sold them on mojo :P
20:54 TitanOfOld Okay, so here's the single sign on I'm going to use: https://developers.eveonline.com/resource/single-sign-on
20:55 TitanOfOld When I do the tests, I want the to remain non-interactive.
20:56 TitanOfOld Rather than having the only way to test the sign in by actually using the site, I was hoping to keep the tests offline by mocking up responses according to the documentation.
20:57 TitanOfOld From what I can tell, I'd have to write another (mojo) webapp to reply to this webapp I'm working on.
20:58 csson joined #mojo
21:00 batman sri: Building and testing Mojolicious-5.80 # Successfully installed
21:00 batman going to use it while working on convos...
21:00 batman anything in particular i should watch out for?
21:00 batman seems like morbo even starts faster..?
21:01 batman :)
21:03 jberger TitanOfOld: can you mock your login/out methods
21:05 batman mojolicious-plugin-cgi, mojo-redis2, mojo-ioloop-readwritefork, mojo-irc and mojo-mysql are all successful with the 5.80
21:06 batman crap. i didn't mean that the t-shirt should have the same design as last year :/
21:06 batman sri, jberger ^
21:07 TitanOfOld jberger, there's nothing preventing me from doing anything really. It's just me, and I'm at the start of my project.
21:07 TitanOfOld Or middle depending on how you look at it.
21:07 jberger then I would recommend making your sso login and logout methods something that you can easily mock for testing
21:08 TitanOfOld jberger, within the webapp, the test, or as a separate webapp?
21:08 jberger best to do it at the test level if possible
21:08 jberger you can make a test utility that does it for you so you can share it with all your tests
21:12 TitanOfOld jberger, so rather than testing the route itself ($t->get('/login')), I should duplicate the steps in the test that /login would perform?
21:18 dod joined #mojo
21:21 punter joined #mojo
21:22 jberger presumably that controller has some logic in it right?
21:23 jberger that should be abstracted into another method or helper which could then be mocked out
21:23 jberger the layout and behavior of the /login route is still testable
21:24 * Grinnz_ wonders if "non-blocking sendmail" is a thing
21:25 preaction open an smtp socket?
21:26 preaction open a pipe to sendmail?
21:26 sri let a minion deal with it ;p
21:26 preaction also that
21:26 purl well, also that is just backwards ...
21:26 Grinnz_ well, it's already in another process, i'm just thinking if theres some way without further forking :P
21:26 Grinnz_ but i supose it wouldnt be that bad
21:27 Grinnz_ an smtp socket would require knowing what smtp server we're using, heh
21:28 jberger <3 minion
21:29 sri purl: jberger?
21:29 purl i guess jberger is on the core team, but I will leave it up to the others if they want me to pursue this further or https://s-media-cache-ak0.pinimg.com/236x/7f/75/6b/7f756bcb3481db56650768cc5fc0cf50.jpg
21:29 preaction run a local sendmail daemon
21:29 Grinnz_ lol
21:29 preaction i seriously think that local smtpd should be on every machine, even just to handle outgoing mail
21:29 Grinnz_ preaction: there is one running apparently
21:29 preaction exactly. so open an smtp socket to localhost
21:30 Grinnz_ you can do that?
21:30 Grinnz_ hehe
21:30 cpan_mojo Mojo-HelloMojo 0.02 by Stefan Adams - http://metacpan.org/release/SADAMS/Mojo-HelloMojo-0.02
21:30 TitanOfOld jberger, the /login would generate a session token, generate a token for the state parameter, store the tokens in the DB, then redirect the user to the 3rd party sign on page. ($c->redirect_to('https://.../?state=token'))
21:30 Grinnz_ the extent of what i know about smtp is the 5 hours i spent trying to get ssmtp to work from my home server (mostly because ISPs block smtp so i had to go through my VPS, naturally)
21:30 preaction it's the best thing to do. then the smtpd handles queuing and errors and if it goes undeliverable sends an e-mail to the account that sent the original. if you've got a .forward file, it'll send to your own e-mail address
21:31 jberger TitanOfOld: your controller shouldn't have all that logic in its own method, each of those tasks should be other methods that the controller calls, that way you can mock the external call in a case like this
21:31 Grinnz_ yeah but with my home server, there's nothing there for it to deliver to
21:32 Grinnz_ either it sends it off to my VPS or it's gone
21:32 sri it's 2015... and sending email is still hard
21:32 Grinnz_ indeed
21:33 preaction blame spammers
21:33 sri all the time
21:35 Grinnz_ well its partly because they made email a "shoot first, ask later" protocol
21:36 Grinnz_ nobody needs to authenticate in order to put stuff into your inbox
21:36 Grinnz_ thats a tricky thing to work around
21:37 sri hahaha http://wentin.github.io/underlineJS/
21:37 Grinnz_ (but necessary, of course)
21:37 Grinnz_ lol
21:51 cfedde that's pretty cute.
22:04 sri reminds me of what happens when you click on the little button in the bottom right here https://events.google.com/io2015/
22:04 * sri loves material design
22:05 Grinnz_ when it says "you might want to grab headphones" it should also you "you might wnat to grab a cpu made after 2005"
22:05 Grinnz_ because this one is dying
22:05 sri lol
22:10 ua_ joined #mojo
22:22 Newbie__ joined #mojo
22:23 punter joined #mojo
22:36 basiliscos joined #mojo
22:56 tardisx joined #mojo
23:16 melo1 joined #mojo
23:44 melo joined #mojo
23:58 good_news_everyon joined #mojo
23:58 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NjBj
23:58 good_news_everyon mojo/master 7043493 Sebastian Riedel: the correct term is setting
23:58 good_news_everyon left #mojo

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