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

IRC log for #mojo, 2014-07-08

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

All times shown according to UTC.

Time Nick Message
00:06 Averna joined #mojo
00:20 jberger sri: my desire was for a non-mango backend, for ease of demonstration and possibility of inclusion in core (which seems like a long-shot now)
00:21 jberger I didn't finish my DBM::Deep backend because the ::File backend is comparable
00:21 jberger I expect that a DBM::Deep backend might be slightly more performant
00:21 jberger but not by much probably
00:22 jberger batman: that looks awesome
00:22 * jberger WANTS
00:23 jberger batman++
00:26 jberger rjbs++ # http://www.nntp.perl.org/group/perl.perl5.porters/2014/07/msg217530.html
00:31 sri Mikey: oh, i just remembered another reason pub/sub didn't work with mongodb... if you have 100 idle workers... and a job is announced via pub/sub... they will all try to grab it right away and DoS the database
00:35 bpmedley jberger: Something like that has been needed for a while.
00:39 sri jberger: a DBM::Deep backend would scale quite a bit better
00:40 sri the file backend has to read and write the entire database for everything
00:48 * jberger is still trying to find the fork bug :/
00:55 laouji joined #mojo
01:06 jberger I don't think its dying
01:07 punter joined #mojo
01:07 jberger I think it is calling the system abort
01:07 Adura Yeah, it calls abort with 3.
01:09 jberger Adura: do you know how to tell from where it calls it?
01:09 jberger me is searching the perl code base :o
01:09 Adura I uh... well, the warning... forget the name of it.
01:09 Adura In the IRC logs somewhere.
01:11 Adura Attempt to free unreferenced scalar I think it was.
01:11 Adura http://blog.booking.com/the-dreaded-attempt-to-free-unreferenced-scalar.html sri linked this at the time.
01:14 jberger I fixed the unreferenced scalar yesterday ;-)
01:14 jberger sadly that didn't fix everything
01:18 * jberger join
01:19 Mikey an inotify based event trigger should work with DBM::Deep, too.  Tomorrow, I'm gonna try and write one that works with the ::File backend and see how that performs.
01:47 klapperl_ joined #mojo
01:55 deserted joined #mojo
02:12 punter joined #mojo
02:29 jberger well, there's your answer: https://github.com/jberger/Mojo-IOLoop-ForkCall/issues/5
02:29 jberger not that it helps anything
02:37 preaction that sounds like a fun bit of yak shaving
02:37 jberger I'm just impressed that I shaved it to the point of knowing there is a yak underneath
02:37 preaction aw ffs cpants... you choose now to ruin my night?
02:38 jberger and I must say bulk88++ for having all of the perl core bugs in his/her head to be able to notice it
02:55 sri jberger++
02:59 KCL joined #mojo
03:00 mattastrophe joined #mojo
03:07 sri Mikey: a modernized patch for notifications http://pastie.org/9366134
03:12 basic6 joined #mojo
03:20 sri http://pastie.org/9366172 # this one is actually a little safer, since it doesn't have to reset the notifications collection
03:25 sri i'm on the edge about it...
03:26 sri the many idle worker problem is kinda annoying, but on the other hand it's more predictable than sleep 5 with arbitrary start times
03:26 sri don't have 1000 idle workers is kind of a simple rule to follow
03:27 * jberger assembles a herd for the sake of stampeding
03:28 jberger *thundering #dammitstevan
03:40 sri think i'll actually try it https://github.com/kraih/minion/commit/63fb0cb423891168c85331be2cd1a33c29d39f99
04:38 mattastrophe joined #mojo
04:52 sujithm joined #mojo
05:19 R0b0t1 I was thinking about using Google's code prettifier to highlight code, but it runs on the client as javascript. Anything anyone knows of that runs server-side?
05:21 preaction R0b0t1: i'm helping you in freenode #perl. don't cross post.
05:22 R0b0t1 Correction, you just started helping me.
05:23 sri R0b0t1: cross posting is considered very rude
05:24 R0b0t1 I don't agree.
05:24 sri then i'm afraid you might not last here very long
05:24 preaction you're not the one helping people
05:26 sri R0b0t1: to be more clear, consider yourself officially warned
05:27 R0b0t1 I don't get it. If someone sees cross posting, help the person where it seems most applicable. If you are not willing to help, ignore it. What is the point of being offended? How is it even offensive?
05:28 sri not everyone is on all channels, different people will give you the same advice, and you will have made them waste time
05:28 preaction because now the discussion is evolving on #perl. turns out you want more than just Perl syntax highlighting. teasing out all that information in more than one place is a waste of time
05:31 R0b0t1 So you did the first thing I suggested one would do in this situation: pick the channel that seems most appropriate, and continue there. Should a person not be in multiple channels, either I would 1) inform them I was helped to a greater extent elsewhere or 2) ignore them and the conversation would die before they invested any time (admittedly kind of rude). In any case, I bother asking more than one place because the advice usually *isn't* the
05:31 R0b0t1 same.
05:33 crab my potted basil has suddenly developed spots on its leaves. is there something i can do to restore it to good health?
05:33 crab (please ignore if not applicable)
05:33 preaction crab: you asked that in #blackthumb
05:34 crab how dare you question my motives
05:35 R0b0t1 crab, syntax highlighting seems at least reasonably related to mojo and/or its usage.
05:36 crab oh. i never thought that the spots might be because of syntax highlighting.
05:36 * R0b0t1 squints at crab
05:36 sri have you tried turning it off and on again?
05:38 denis_boyun joined #mojo
06:16 denis_boyun_ joined #mojo
06:18 denis_boyun joined #mojo
06:24 chankey joined #mojo
06:28 sujithm joined #mojo
06:33 cpan_mojo Minion 0.30 by Sebastian Riedel - http://metacpan.org/release/SRI/Minion-0.30 (depends on Mojolicious)
06:33 Jonis hooray
06:35 sri Mikey: you'll like this release :)
06:35 sri \o\
06:35 sri /o/
06:36 sri i'm not sure if notifications stay... but i want to try it in the wild at least
06:51 dod joined #mojo
07:01 eseyman joined #mojo
07:04 rawler joined #mojo
07:06 Eke- joined #mojo
07:11 denis_boyun joined #mojo
07:18 sujithm_ joined #mojo
07:23 neyasov_ joined #mojo
07:27 basiliscos joined #mojo
07:32 mire joined #mojo
07:36 mattastrophe joined #mojo
07:37 trone joined #mojo
07:46 igorsutton joined #mojo
07:58 chankeypathak joined #mojo
08:13 dod joined #mojo
08:16 denis_boyun joined #mojo
08:18 neyasov_ joined #mojo
08:19 Vandal joined #mojo
08:31 edestler joined #mojo
09:01 dod joined #mojo
09:05 dod joined #mojo
09:05 eseyman joined #mojo
09:13 basiliscos joined #mojo
09:36 basiliscos joined #mojo
09:38 dod joined #mojo
09:42 basiliscos joined #mojo
09:47 Eke|| joined #mojo
09:54 basiliscos joined #mojo
10:33 basiliscos joined #mojo
10:34 meshl joined #mojo
10:39 sujithm joined #mojo
10:40 basiliscos joined #mojo
10:42 basiliscos joined #mojo
10:52 basiliscos joined #mojo
11:14 dotandimet joined #mojo
11:19 bpmedley_ joined #mojo
11:21 mattastrophe joined #mojo
11:24 cpan_mojo MojoX-Renderer-IncludeLater 0.02 by CHGOVUK - http://metacpan.org/release/CHGOVUK/MojoX-Renderer-IncludeLater-0.02
11:29 dotan1 joined #mojo
11:43 jberger Hahaha "have you tried turning it off and on again?" hahahaha
11:45 * jberger starts writing 'the hackers guide to gardening'
11:46 jberger crab: if rebooting doesn't help, you might need to reinstall the os
11:46 neilhwatson joined #mojo
11:47 jberger But this time i recommend one of the free plant oses
11:59 dwierenga joined #mojo
12:09 aleksey joined #mojo
12:09 aleksey left #mojo
12:17 crab do i need root for that?
12:28 irq joined #mojo
12:37 dotan joined #mojo
12:40 john joined #mojo
12:41 dotandimet joined #mojo
12:42 john I'm about to implement event source, as per the example in the docs, with a small change
12:42 john rather than subscribe to the log event, I want to write my own emitter.
12:43 john subscribing to the event seems straight forward
12:43 john however, how might I go about emitting from elsewhere in my app?
12:44 john I presume I need to put the event source object somewhere, but where?
12:45 jberger_ joined #mojo
12:53 fhelmber_ joined #mojo
12:55 igorsutton joined #mojo
12:57 bpmedley john: What if you try and send an event from a separate process?
12:57 bpmedley john: You could possible do pub/sub from a database?
13:01 john separate as in a different worker under hypnotoad? I'm guessing it just works. The example subscribes to the log. If it only gets events from that process it isn't much use.
13:02 bpmedley john: Perhaps I mis-understood.  If you emit from elsewhere in your app, then how do you know that the request doing the emitting and the reqeust doing the listening are in the same process?
13:02 john I've currently got the event source wating async on a pipe, but that can't be as good as using the mojo pub/sub.
13:02 john I don't know that they are in the same process. Nor should I care
13:02 bpmedley I see.  So, you’re using IPC with pipes?
13:03 john I am, but am trying a different method.
13:03 john I want to dupe the event source example in the docs, but use my own publisher, not log.
13:03 bpmedley john: http://www.postgresql.org/docs/9.3/static/sql-notify.html <— Would something like this help?
13:04 john Unfortunately, I'm using mysql.
13:05 john I could always poll memcached, but that's a bit kludgey also.
13:06 john I'll see if the log example works in a useful manner first, then come back here
13:06 Adura joined #mojo
13:06 bpmedley john: Is everything currently on the same server?  Is it Linux or BSD?
13:07 john mojo app is on a linux server. I want a process on another machine to issue a get/post to it to signal the process has completed, I then want that info to use SSE to tell the browser.
13:08 john I have that working, but I'm not happy with it.
13:08 bpmedley Hrmm.  I’m not sure if inotify works with NFS.
13:09 john inotify or async read of a pipe is similar in this case. Using the mojo pub/sub seems a way better way to go.
13:10 bpmedley http://asana.github.io/kraken/  <— Something like this?
13:11 john well that's a pub/sub. So yes. But actually using Mojo::EventEmitter
13:11 john is what I'm after
13:11 purl after is using the map which doesn't trigger a query at all because rel_interfaces is prefetched
13:12 john I'll have a play and see what happens. I'll gain some understanding of the mojo architecture at least.
13:21 d4rkie joined #mojo
13:31 Akron joined #mojo
13:43 lipizzan joined #mojo
13:45 DaniBunny joined #mojo
13:48 lipizzan is there a recommended way to ( from within mojo) discover\follow links that are normally dynamically created by javascript in the browser?
13:48 jberger__ joined #mojo
13:48 bpmedley lipizzan: Are you wanting the backend to communicate somehow with the frontend?  I’m rather confused.
13:49 d4rkie joined #mojo
13:49 moritz lipizzan: you mean, you want to automate websites that use JS with Mojo::UserAgent?
13:49 lipizzan moritz: yes.
13:50 lipizzan moritz: or whatever modules would be  the current convention.
13:51 tomboh lipizzan: using plain perl you're likely to get frustrated
13:51 tomboh you might consider talking to a web browser (or something else) using the WebDriver protocol
13:52 tomboh as lots of things support WebDriver, it means you can swap the user agent you want to use easily
13:52 bpmedley http://phantomjs.org/ <— Maybe?
13:52 tomboh yeah, that's one option
13:52 tomboh the Selenium::Remote::Driver provides a perl client interface to the WebDriver protocol
13:54 lipizzan Anyone used WWW:Mechanize::Firefox?  I was hoping to go with something lighter weight than Selenium.
13:55 basiliscos joined #mojo
13:57 john Would I be correct in thinking there is no in built IPC between toad workers?
14:05 _eugen joined #mojo
14:07 tomboh lipizzan: you don't need Selenium to use WebDriver
14:07 tomboh although many docs might lead you to believe that
14:07 igorsutton joined #mojo
14:08 tomboh it's an agnostic protocol that gives you the flexibility to swap out the user-agent you use
14:08 tomboh for example, phantonjs with ghostdriver doesn't need selenium
14:09 lipizzan There's a new module WWW::Mechanize::PhantomJS on CPAN. Might try that. I've already got the final scraping code done in Mojo. So if I can retrieve the js result url, I should be done.
14:09 tomboh oh, sorry, I think I might be wrong and you do need Selenium
14:10 lipizzan I haven't looked at the dependencies of  WWW::Mechanize::PhantomJS, maybe it requires Selenium too.
14:11 tomboh yeah, it seems I misremembered - it's a while since I worked with these things
14:12 bluescreen joined #mojo
14:14 Mikey wow sri, thanks.. 0.30 has replaced my hacked version, but it performs identically
14:16 btyler joined #mojo
14:21 sujithm joined #mojo
14:59 KCL_ joined #mojo
15:17 nishttal2 joined #mojo
15:18 nishttal2 hello ppl.. I am having trouble with templates.. I have a very simple template but keeps giving me an error.. please see http://fpaste.org/116394/48325341/ the error is copied there as well
15:18 nishttal2 does any one know what I am doing wrong
15:20 nishttal2 i dont get any error if I remove line #8
15:25 batman jberger: what looks awesome..?
15:27 sh4 joined #mojo
15:31 mire joined #mojo
15:31 disputin joined #mojo
15:47 sri everyone please test minion 0.30!
15:53 disputin joined #mojo
15:55 dvinciguerra joined #mojo
15:57 go|dfish nishttal2: well line 8 begins with %= so the whole line is a Perl expression, you don't need the other <%= %>, so something like: %= include "main/$app->{Name}/${p}_content"
16:00 nicomen I'm going to test minion right now
16:01 nicomen but I want to not have to start it separately, would it be possible to let the web server do that somehow?
16:01 denis_boyun joined #mojo
16:02 nishttal2 go|dfish, that was it .. thanks :)
16:04 meshl joined #mojo
16:05 jberger__ batman: the convos menu (screen usage) video you posted
16:09 denis_boyun_ joined #mojo
16:11 denis_boyun joined #mojo
16:21 denis_boyun joined #mojo
16:41 basiliscos joined #mojo
16:49 dod joined #mojo
16:50 disputin joined #mojo
16:54 howitdo joined #mojo
17:01 fhelmber_ joined #mojo
17:09 jberger__ joined #mojo
17:46 risugg joined #mojo
17:51 meshl joined #mojo
17:53 mire joined #mojo
17:58 meshl joined #mojo
18:15 sri at least it looks like mongodb cpu usage doesn't increase with the number of pub/sub clients
18:17 bobkare joined #mojo
18:19 sri 7% constant cpu usage is not great... but manageable i guess
18:20 sri gonna have to play with mongodb 2.7.3... first dev release with document level locking and pluggable storage engines
18:21 sri those concurrency improvements will be so great for mango
18:21 mire joined #mojo
18:21 davido_ joined #mojo
18:23 lb i don't understand nosql - if you want to gamble with data integrity for the sake of speed, why not just use mysql ;)
18:26 Mikey on my mac it looks like it bumps up about 2% cpu for every worker i start
18:27 Mikey i've decided tho that since i have a slave postgres db doing a bunch of nothing that mongodb has a place in my infra, so i'm going to use mongo so i can put workers anywhere
18:29 sri lb: and then you have to scale up and denormalize your schema...and basically build your own version of nosql on top of postgres ;p
18:30 Mikey nosql has its place.  the original nosql database was ldap tho ;)
18:30 sri sharding with sql... eeeeeep
18:31 Mikey once upon a time i badly badly misused ldap in very similar ways to how people use mongodb.  and i'll never be forgiven for those sins.
18:31 Mikey but my app was fast as hell ;)
18:31 * lb forgives Mikey for THAT one
18:32 * sri loves document databases in general though... lets me try more things and make faster changes
18:32 r0b3rt there is always the EAV model for small databases ....but it does not scale well.
18:33 lb pg as a document database looks nice, then you get to keep the parts of your model in whatever makes sense, without polluting the tool-pool too much
18:34 lb i just came up with tool-pool, rather proud of myself
18:34 Mikey the frenzy of 4 workers all trying to grab a job bumps up mongo cpu usage by about 4% for that second
18:35 Mikey i need to see this on the slower AWS cpus tho.. i'm prolly gonna have to cap at 8 workers or so for the env.
18:35 Mikey at least until they fix mongodb.  maybe i'll just install 2.7.3 as minion is experimental anyway.
18:36 Mikey i love postgres.  mongo's not bad.  the implementation is way worse than the idea, but i mean, the implementation will improve.  look at mysql.
18:36 lb i lost you
18:36 lb you lost me
18:37 Mikey sorry.. the last thing i said was a reply to you, the other three were asides about my minion testing
18:37 lb mysql is still majorly terrible
18:37 lb unless you don't care about your data
18:37 Mikey you must not be a long time mysql user ;)
18:37 lb no, i try to stay clear
18:38 sri mongodb 2.8 will be *a lot* less terrible
18:39 lb like ovid tweeted, http://dev.mysql.com/doc/refman/5.0/en/sql-mode.html#sqlmode_allow_invalid_dates http://dev.mysql.com/doc/refman/5.1/en/constraint-invalid-data.html
18:39 lb By default, MySQL is forgiving of invalid or improper data values and coerces them to valid values for data entry.
18:39 lb NO! JUST NO!
18:39 sri being able to use different storage engines for every box in your cluster is actually pretty cool... have an in-memory engine for fast access... heavy durability one for failover... and so on
18:40 Mikey i have a feeling that i'll lean more and more on mongo as i scale, especially if the implementation improves
18:40 genio mongo failovers have left a poor taste in my mouth.  Mongo's been just nothing but annoying for us in multi-region setups.  Mongo-- # genio has spoken
18:40 Mikey i started my current project on mongodb and backed it out in favor of postgres about 8 months in
18:40 lb genio: you get to do your own acid!
18:40 lb which will make you want to do acid!
18:40 Mikey but back then mango didn't exist, and i really wanted an orm
18:40 Mikey which still doesn't exist
18:41 sri mongodb is still very young, but they are really working hard on it... unlike oracle with mysql...
18:41 sri you notice it's a database that happened by accident if you look closer... but it improves constantly
18:42 lb i'm currently doing work for an absolute oracle-shop - just trying to slowly force their software stack in a direction, so they can save a lot of money by moving to pg
18:42 Mikey we're an oracle shop, so i didn't use any native database types / stored procedures or anything, so our app is completely rdbms agnostic
18:43 Mikey some of my devs use mysql, some use postgres, some use sqlite
18:43 Mikey i have it split up like that so we can know if we've broken something
18:43 Mikey i really should be writing tests and running them against multiple backends.. currently our test suite just uses sqlite-in-ram
18:44 lb running tests against a different db than prod is exactly the same as mock-objects
18:46 Mikey i know.  we'd need to create a test box tho with all the rdbms' we plan to support installed and updated and yuck
18:46 Mikey sqlite in ram == no setup, no drama, runs everywhere
18:47 sri don't have a ci setup?
18:47 Mikey sadly, no :(
18:48 sri with stuff like drone available... i don't think there's much of a reason not to have ci https://github.com/drone/drone
18:48 lb i have started moving run-time checks of configuration and templates into the test-suite
18:48 lb the CI will alert us immediately
18:49 Mikey checking drone out
18:50 lb we're just using jenkins
18:50 lb it helps if you have a guy who's used jenkins a lot
18:50 lipizzan joined #mojo
18:50 Mikey this'll help me test on everything
18:50 Mikey mac, linux amd64 and arm
18:51 Mikey are the three platforms we care about
18:52 Mikey my management would love if i started testing against oracle
18:53 genio I'm starting to sound like a negative nancy today, but... Oracle--   #eww
18:53 lb genio: you're not wrong
18:54 * genio thinks he must have worked on the worst projects in the world to hate so many things so much.
18:54 sri you're just a negative nancy
18:54 * genio lowers his head and cries in the corner
18:54 lb i defined for our project that all talk with the db will ideally happen through DBIx::Class::Schema::Loader
18:54 lb so it's just perl
18:55 davido_ This month's Salt Lake Perl Mongers is going to demonstrate refurbished 1970 Teletypes, interfaced with the Internet via custom hardware, Perl, and Mojolicious.  Set up to retrieve email, drudge reports, wikipedia articles, etc.
18:55 sri hmm, i guess chrome is the new ie
18:55 genio DBIx::Class and SQL::Abstract are nice to deal with.
18:55 genio sri: in what sense?
18:56 sri google hangouts just went chrome exclusive
18:56 sri marketshare corrupts
18:56 genio Do they still offer the stand-alone client?
18:56 stephanj cool
18:57 sri http://www.google.com/hangouts/ # try the "on your computer" link with anything else than chrome
18:59 Mikey davido_: that sounds awesome
18:59 sri and even in chrome it's a play store extension :o
18:59 davido_ He gave a 50 minute version of this presentation at OpenWest, but we're giving him 90 so he has more time for the Perl content, as well as for allowing audience hands' on.
19:01 genio the hell?!?  I try to download hangouts from FF and get a page: "Get a fast, free web browser" telling me how to download Chrome
19:01 sri yea :D
19:01 Mikey i'm still installing flash and java updates
19:02 sri the open web according to google
19:03 stephanj fuck off google.
19:03 sri mozilla really needs to get some marketshare back
19:05 Mikey agreed.
19:06 stephanj a collegue in the company i stared working for on monday, frowned when i installed firefox and started to tell me of all the awesomeness of chrome... i just replied that i did a well informedchoice and took the lesser evil and that made him stop dead in his tracks
19:07 Mikey unfortunately, to many mainstream users, the more important thing is what a piece of software does for them, not what it doesn't do to them.
19:07 Mikey chrome is spyware done right.
19:08 Mikey and google knows this, that's why they made hangouts chrome only.  and this will only further marginalize firefox unfortunately.
19:09 lb no need for mainstream
19:09 Mikey maybe for us.  but we're perl developers.  so... we're not exactly mainstream.
19:09 lb no, but we're bitten by convenience too
19:10 stephanj my firefox is a ram hogging slow piece of shit especially with many tabs... but hey i got used to live with its quirks
19:11 lb chrome is a ram hogging slow piece of shit especially with many tabs - but it's been years since i've ever lost a tab to a crash
19:11 lb which was a frequent problem with firefox and opera before that
19:12 * lb often just kills the processes for tabs he doesn't want/need right now
19:13 bluescreen_ joined #mojo
19:20 Mikey meh 2.7.3 still uses a bunch of cpu on my mac
19:20 disputin joined #mojo
19:22 irq joined #mojo
19:24 Mikey sri: what do you think about a worker forking up to a configurable number of subworkers?  that way, you have only one mongodb listener that can get more work done?
19:24 irq_ joined #mojo
19:24 sri not portable
19:25 Mikey damn.  i figured you'd already thought of that ;)
19:25 sri at least with the proven hypnotoad architecture
19:26 sri you're welcome to give it a try, just remember to test on windows
19:26 sri or you make your own unix only worker command
19:27 Mikey i'd have to have my own Backend:: too
19:27 Mikey actually the fork is in Job.pm
19:27 sri the primitives are all there, a worker can already claim multiple jobs
19:27 Mikey hmm
19:28 sri no, you don't need to own anything
19:28 sri minion was designed with custom worker daemons in mind
19:28 Mikey if i do $job->perform it's gonna block on waitpid isn't it?
19:28 Mikey i'll read the code again
19:28 sri look how ->perform works...all public api
19:30 Dandre joined #mojo
19:39 basic6_ joined #mojo
19:42 bobkare joined #mojo
19:57 jberger_ joined #mojo
20:14 denis_boyun joined #mojo
20:29 Jonis wow, brazil have well and truly collapsed
20:30 genio the score looks bad for brasil
20:31 Jonis germany scored 4 in 6 minutes
20:31 Jonis it's not even a game, it's a slaughter
20:34 sri crazy
20:34 genio is there a skunk-rule in soccer?
20:35 lb if there's a skunk on the field, they almost always postpone play
20:36 genio but does the clock continue to run during that downtime?
20:36 lb no
20:37 lb at the ref's sole discretion
20:40 crab huh. no goals in the last five minutes. this game sure turned boring quickly
20:40 genio heh
20:46 sri argentina vs germany should be more interesting
20:50 genio The rest of this game is going to just be Germany sitting on the ball laughing...
20:55 denis_boyun joined #mojo
20:57 denis_boyun_ joined #mojo
21:01 denis_boyun__ joined #mojo
21:02 lb i'm not staying up for the penalty kicks tonight
21:05 Jonis that's easily sorted. no way there's going to be penalties
21:05 Jonis unless brazil gets one for filming
21:14 meshl joined #mojo
21:30 davido_ joined #mojo
21:31 davido_ Is it correct that Mojolicious::Lite exhausts the DATA filehandle, and if so, is there a way to request it not do so?
21:32 davido_ I don't want to use the handle for templates, but did want to put some data there just for a demonstration.
22:25 sri hmmmm
22:26 sri i wanted to make Minion::Backend::File check mtime and file size before slurping the file... but it seems those values are unreliable :( https://github.com/kraih/minion/commit/82b3e60ef2e573bfb1e7c71627ba37fb31d9df67
22:26 jberger_ joined #mojo
22:33 disputin joined #mojo
22:39 nicomen sri: why unreliable?
22:40 sri because it doesn't work
22:40 nicomen modified time is not correct?
22:40 sri no
22:43 sri if anyone wants to hack on it, be my guest
22:45 neilhwatson joined #mojo
22:52 nicomen weird, when running a test here, one of the executions showed a wrong time, the rest seem correct and I can't reproduce
22:53 nicomen $ date; touch foo2; perl -wle 'print scalar localtime "" . (stat "foo2")[9];'
22:53 nicomen on. 09. juli 00:50:49 +0200 2014
22:53 nicomen Wed Jul  9 00:50:48 2014
22:53 nicomen must be different rounding?
23:04 nicomen sri: do you have a failing test to work with?
23:04 sri the minion tests fail randomly
23:11 deserted joined #mojo
23:13 disputin joined #mojo
23:16 d4rkie joined #mojo
23:31 d4rkie joined #mojo
23:40 deserted hey all, got a question in that may fit more in #httpd but wanted to ask here as I've had no luck with them as of yet, I've got a mojolicious app I'm trying to implement single sign on with, using apache with mod_auth_kerb acting as a reverse proxy to hypnotoad
23:40 deserted problem is that the remote_user header is showing as (null) when I dump $c->req->headers
23:42 deserted apache config can be viewed at http://pastie.org/9366042 and was wondering if anyone had encountered similar issues passing REMOTE_USER to mojolicious?
23:42 nicomen sri: which tests?
23:42 purl which tests are failing?
23:42 nicomen ah nevermind
23:52 firnsy deserted: does mod_auth_kerb actually provide the REMOTE_USER variable?
23:56 deserted supposedly yes
23:57 deserted and it's filled using phpinfo on a straight vhost
23:57 deserted but when trying to set it with RequestHeader set X-Forwarded-REMOTE-USER %{REMOTE_USER}s I just get a null value
23:59 firnsy deserted: i'm wracking my brain trying to remember my config, i wrote a mojo 2.x app in a gov department with the exact thing you're after

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