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

IRC log for #mojo, 2015-06-15

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

All times shown according to UTC.

Time Nick Message
00:04 woz joined #mojo
01:08 woz joined #mojo
01:10 jamesw joined #mojo
01:16 dvinciguerra joined #mojo
01:25 klapperl_ joined #mojo
01:32 hshong joined #mojo
01:54 absolut_todd joined #mojo
02:11 woz joined #mojo
02:24 noganex joined #mojo
04:20 batman btyler: Going to hand that one to marcus
04:21 batman I think it's a bit weird that we allowed *just* \n in parts of the regex before
04:21 batman At least that's how it looks to me
04:22 batman Will have a closer look later
04:23 batman btyler: what kind of IRC server is that? Sounds very non-standard
04:23 batman Almost like doing HTML headers without \r
04:39 lthemick joined #mojo
04:41 lthemick list
05:18 woz joined #mojo
05:24 mattastrophe joined #mojo
05:29 kaare joined #mojo
05:40 hshong joined #mojo
05:48 Onigiri joined #mojo
05:57 oetiker joined #mojo
06:03 oetiker joined #mojo
06:06 oetiker joined #mojo
06:08 McA joined #mojo
06:16 dod joined #mojo
06:20 dod joined #mojo
06:20 gatitskiy joined #mojo
06:21 dod joined #mojo
06:23 woz joined #mojo
06:24 gatitski_ joined #mojo
06:24 oetiker joined #mojo
06:24 btyler batman: that's why I linked the spec -- I was going to submit a PR to the IRC server until I saw that chunk of the RFC
06:26 batman btyler: ah. ok. just got time to look at it now.
06:26 btyler (in other words, yes, I agree that it is weird, but accepting either as a message terminator is part of the spec)
06:26 batman was on my phone earlier...
06:26 btyler as I am now :) no worries
06:26 batman btyler: change this https://github.com/jhthorsen/mojo-irc/blob/master/lib/Test/Mojo/IRC.pm to not have any "\r"
06:26 purl batman: that doesn't look right
06:27 batman that will be good enough for unit testing for me
06:28 Onigiri joined #mojo
06:28 btyler roger, will do
06:29 batman cool
06:30 dod joined #mojo
06:46 gatitskiy joined #mojo
06:48 gatitskiy joined #mojo
06:55 gatitski_ joined #mojo
06:56 oetiker joined #mojo
07:00 AndrewIsh joined #mojo
07:01 Vandal joined #mojo
07:04 trone joined #mojo
07:07 damaya_ joined #mojo
07:08 damaya_ Hey all, my session keeps expiring after some random interval (maybe 3 minutes). Is there an easy way to debug why it's happening?
07:13 jzawodn joined #mojo
07:17 woz joined #mojo
07:21 eseyman joined #mojo
07:21 Onigiri joined #mojo
07:21 gatitskiy joined #mojo
07:23 gatitskiy joined #mojo
07:31 gatitskiy joined #mojo
07:37 batman damaya_: look at the expire data set in your dev console
07:44 batman *date
07:44 batman there might be something else going on... i would also choose "persist history" or what it's called (not sure which browser you're using) to see what happens between requests
07:55 gatitski_ joined #mojo
07:56 odc joined #mojo
07:58 gatitskiy joined #mojo
08:02 gatitski_ joined #mojo
08:04 gatitskiy joined #mojo
08:09 damaya_ Figured it out. My cookie expires everytime I make a change and save it (using morbo).
08:10 btyler batman: hmm, my PR might be wrong after all. I was linking you to the server spec, not the client one; client one specifies only CRLF like you'd expect. but at least one popular client (irssi) only looks for \n to delimit messages, and just strips \r
08:10 btyler https://github.com/irssi/irssi/blob/c122a2a226363114e11f32072395e2f06ad6a450/src/core/line-split.c#L84
08:11 btyler going to be submitting a patch to this hacked up IRC server after all
08:13 batman btyler: well... if marcus thinks it's ok, then i don't mind the PR
08:14 btyler http://tools.ietf.org/html/rfc2812#section-2.3 is pretty clear. I suppose they figure it's easier to get server implemenations to be mindful about their output and just resign to clients sending all kinds of things
08:14 batman you can't have (just) \n in a message anyway so...
08:14 batman which server is this?
08:15 btyler https://github.com/Yaribz/SpringIrcBridge -- it translates IRC to a protocol used by a game server
08:16 batman cool :)
08:16 btyler in other words, wildly obscure, and I'll be trying to patch it to emit \r\n
08:16 batman :D
08:27 gatitskiy joined #mojo
08:33 gatitskiy joined #mojo
08:34 stephan48 batman, btyler: oh. the irssi-sends-only-\n-bug-again?
08:35 gatitskiy joined #mojo
08:37 btyler stephan48: well, in this case it was "obscure specialized IRC server sends only \n; Mojo::IRC expects everything to have \r\n"
08:37 btyler and I went looking at irssi to see what a popular client did
08:38 stephan48 yea
08:38 stephan48 irssi does that too
08:39 stephan48 in their proxy module
08:39 pink_mist yeah, \n is not allowed inside irc strings anyway; its only use is as part of the \r\n pair to delimit commands
08:40 pink_mist (should use \012 and \015 instead of \r and \n)
08:44 btyler pink_mist: mind elaborating for one less wise in the ways of protocol specs? (re \015 vs \r)
08:49 gatitski_ joined #mojo
08:49 marcusr So really the pragmatic thing for us to do would be to accept \n as well as \r\n?
08:50 marcusr or is there a downside to that?
08:52 gatitskiy joined #mojo
08:52 batman marcusr: i can't see the downside...
08:53 batman btyler: not going to happen unless you change two lines in Test::Mojo::IRC though...
08:53 btyler yes, I'll do that at lunch
08:54 gatitskiy joined #mojo
08:55 batman cool
08:56 gatitskiy joined #mojo
08:56 batman btyler: marcusr is very pragmatic, so i i think he'll hit the merge button :)
08:57 dexteruk joined #mojo
08:58 gatitskiy joined #mojo
08:58 btyler awesome. also, you guys were already doing the pragmatic thing for all but the trailing delimiter -- [^\r\n] matches them individually or together
08:58 btyler I think, at least
08:58 marcusr mm
09:00 gatitski_ joined #mojo
09:03 stephan48 cool. thanks for doing it btyler!
09:12 gatitskiy joined #mojo
09:16 gatitskiy joined #mojo
09:19 gatitski_ joined #mojo
09:38 batman btyler: yeah, that's how i read it as well... it's very strange :/
09:42 gatitskiy joined #mojo
09:59 gatitskiy joined #mojo
10:09 absolut_todd joined #mojo
10:15 gatitskiy joined #mojo
10:21 pink_mist btyler: \015 is always going to be that specific byte. \r might be some other bytes on some platforms.
10:21 pink_mist btyler: same with \012 and \n
10:28 gatitski_ joined #mojo
10:29 bin_005 joined #mojo
10:29 sri this is what we do for HTTP headers https://github.com/kraih/mojo/blob/master/lib/Mojo/Headers.pm#L92
10:30 sri still one of the hot spots in mojolicious, so if anyone can think of an optimization, that would be very welcome!
10:30 gatitskiy joined #mojo
10:39 bin_005 joined #mojo
10:40 mattastrophe joined #mojo
10:50 btyler pink_mist: cool, thank you
10:53 gatitski_ joined #mojo
10:53 mattastrophe joined #mojo
10:55 gatitskiy joined #mojo
11:04 trone joined #mojo
11:06 trone joined #mojo
11:07 gatitskiy joined #mojo
11:09 gatitski_ joined #mojo
11:11 gatitskiy joined #mojo
11:13 gatitski_ joined #mojo
11:24 punter joined #mojo
11:32 gatitskiy joined #mojo
11:34 gatitsk__ joined #mojo
11:40 gatitskiy joined #mojo
11:43 bakanov_da joined #mojo
11:46 bakanov_da hello. Am I right, that 'catch' method from mojo::eventemitter can't be used to catch exception from 'die' and other fatal exceptions? If it can't be, what should I use to handle exceptions? I store some information in redis and what to delete it, if something went wrong
11:50 gatitskiy joined #mojo
12:02 neilhwatson joined #mojo
12:06 pheleas_frog joined #mojo
12:17 btyler batman/marcusr: I updated the PR to hit Test::Mojo::IRC as well. also added a commit to s/\r/\015/g and s/\n/\012/g per pink_mist's suggestion above
12:18 marcusr btyler++
12:19 marcusr btyler: I'll merge it if Travis says pass
12:20 marcusr done
12:20 btyler awesome *opensource highfive*
12:20 marcusr o/
12:26 gatitski_ joined #mojo
12:32 cpan_mojo Mojolicious-Plugin-PetalTinyRenderer-0.05 by LBR https://metacpan.org/release/LBR/Mojolicious-Plugin-PetalTinyRenderer-0.05
12:58 ajr_ joined #mojo
13:19 cpan_mojo Mojolicious-Plugin-SQLiteViewerLite-0.13 by KIMOTO https://metacpan.org/release/KIMOTO/Mojolicious-Plugin-SQLiteViewerLite-0.13
13:19 gatitskiy joined #mojo
13:22 mattastrophe joined #mojo
13:22 ToApolytoXaos joined #mojo
13:24 batman btyler: how is \015 an \012 better than \r and \n?
13:25 btyler batman: http://irclog.perlgeek.de/mojo/2015-06-15#i_10751439
13:26 batman but how is that possible?
13:28 ribasushi batman: IO layer translation
13:28 ribasushi batman: when on win32 you print "foo\n"
13:28 ribasushi the destination ends up with actual \r\n bytes
13:29 batman crazy. i didn't know that.
13:29 ribasushi batman: this is the main reason your perl cowkrs on win32 - how often do you print a literal \r in your code? yet almost nothing windows deals with a lack of an \r
13:29 ribasushi errr s/cowrks/works/ (wtf?)
13:30 batman ribasushi: i never cowork on win32 :P
13:31 Grinnz joined #mojo
13:31 btyler ribasushi: thanks for clarifying, the example of \n --> \r\n on win32 makes it much more concrete
13:32 ribasushi ????
13:33 pink_mist also note that on old mac os, pre os x, \r and \n were reversed iirc ... then there's ebcdic systems which is a whole nother can of worms
13:33 cpan_mojo Mojo-IRC-0.25 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Mojo-IRC-0.25
13:33 ribasushi I doubt mojo works on ebcdic (though would be nice if it did ;)
13:34 batman btyler: thanks and enjoy ^ :)
13:34 btyler batman: woo! awesome :)
13:34 * batman tries to join the *opensource highfive*
13:34 batman :)
13:36 dotan joined #mojo
13:37 wingfold joined #mojo
13:39 gatitski_ joined #mojo
13:41 dvinciguerra joined #mojo
13:41 gatitskiy joined #mojo
13:43 gatitski_ joined #mojo
13:46 gatitskiy joined #mojo
13:48 mattastrophe joined #mojo
13:48 gatitskiy joined #mojo
13:58 * sri starts a new wishlist thread https://groups.google.com/forum/#!topic/mojolicious/hZu18SdEmbc
14:04 absolut_todd joined #mojo
14:08 gatitski_ joined #mojo
14:11 gatitskiy joined #mojo
14:12 asarch joined #mojo
14:13 gatitski_ joined #mojo
14:15 gatitskiy joined #mojo
14:17 gatitski_ joined #mojo
14:21 bakanov_da joined #mojo
14:22 jberger wishlist \o/
14:34 cpan_mojo Mojolicious-Plugin-ValidateTiny-0.15 by KOORCHIK https://metacpan.org/release/KOORCHIK/Mojolicious-Plugin-ValidateTiny-0.15
14:39 ashimema joined #mojo
14:41 gatitskiy joined #mojo
14:49 idegen joined #mojo
14:51 wingfold joined #mojo
14:52 disputin joined #mojo
14:55 gatitski_ joined #mojo
14:55 ajr_ joined #mojo
14:58 gatitskiy joined #mojo
15:02 gatitskiy joined #mojo
15:04 gatitskiy joined #mojo
15:06 gatitskiy joined #mojo
15:08 gatitski_ joined #mojo
15:10 gatitskiy joined #mojo
15:14 sh4 joined #mojo
15:18 zackiv31 joined #mojo
15:19 amon joined #mojo
15:24 Kogurr joined #mojo
15:31 PryMar56 joined #mojo
15:40 lluad joined #mojo
15:48 punter joined #mojo
15:55 hernan604 joined #mojo
16:00 asarch joined #mojo
16:08 wingfold joined #mojo
16:20 wingfold_ joined #mojo
16:24 ajr_ joined #mojo
16:25 gtodd joined #mojo
16:35 cpan_mojo Mojolicious-Command-snoodoc-0.03 by CURTIS https://metacpan.org/release/CURTIS/Mojolicious-Command-snoodoc-0.03
16:37 lb so clearly Mojolicious::Command was a mistake
16:39 jberger lb: ?
16:39 * sri can't wait for postgres 9.5
16:39 cfedde joined #mojo
16:42 lb jberger: donuts and snoodoc
16:43 jberger lb: have a little fun :-)
16:43 lb i have lots of fun
16:44 jberger in concept I would have expected "doc" to be a subcommand of a snoo command
16:45 mst hey, it's not like he's using any namespace anybody else wants
16:48 lb My::Silly::API::whatevs could easily include Mojolicious::Command::whatevs without pooping on cpan
16:50 Mikey what's your fav. feature in pg 9.5, sri?
16:52 bif joined #mojo
16:54 sri Mikey: SKIP LOCKED
16:54 jberger lb: I don't see how that is any different in the long run
16:54 sri upsert is a close second, but i need SKIP LOCKED more
16:55 lb jberger: non-indexed
16:55 lb Mojolicious::Command:: is an implementation detail, so package\n
16:55 jberger lb: that would still be indexed
16:55 lb it would not
16:55 jberger you are right, not by hiding it from the indexer
16:55 disputin joined #mojo
16:56 jberger but then I come back to, who cares if the indexer sees it
16:56 lb well, apparently I do :)
16:56 jberger the only thing I would want to hide from the indexer is something that I would tolerate being provided by another package
16:56 lb this is exactly the same argument as classes using Moose shouldn't be in the Moose:: namespace
16:56 jberger s/package/dist/
16:57 lb it's just slly
16:57 lb +i
16:57 Grinnz in that sense, it should really be My::Silly::API::Command::whatever
16:57 jberger lb: but the way the command plugin system works, you need it to be in that namespace
16:57 Grinnz do you?
16:57 Grinnz my app doesn't
16:57 jberger if you want it to show up for the mojo command
16:58 Grinnz oh, for mojo yeah
16:58 lb hence my point, Mojolicious::Command:: is an implementation detail,
16:58 jberger commands that do not come with their own app
16:58 Grinnz no, if you want to attach it to the mojo binary, it's not an implementation detail
16:58 Mikey can push onto namespaces, which is what i do for My::Silly::API
16:58 Grinnz "binary"
16:58 lb it most certainly is
16:58 sri minion does that for its commands too
16:58 lb it needs to put stuff in the Mojolicious::Command namespace on installation, but it doesn't need to index as Mojolicious::Command on cpan
16:59 sri you use the plugin, it registers the minion command namespace
16:59 lb and it certainly doesn't need its package to be in the Mojolicious::Command namespace
16:59 Grinnz yeah to do it properly it should set up its own executable and namespace
16:59 damaya_ joined #mojo
16:59 lb You can question the importance of the issue, but I am right :)
17:00 Grinnz i still disagree with not-indexing it as it is
17:00 sri i'd argue that finding a donut shop is kind of a global issue, not app specific
17:00 Grinnz lol
17:00 jberger and a very important one at that
17:00 Grinnz right but in this case, it would be "an app to find a donut shop" instead of "a mojo command to find a donut shop"
17:00 Grinnz and then using mojo would just be an "implementation detail"
17:01 sri clearly we need Mojolicious::Command::acme
17:01 jberger Grinnz/lb: I have Mojolicious::Command::nopaste
17:01 lb jberger: you monster
17:01 Grinnz lol
17:01 jberger it's quite handy
17:01 lb so is sri's phone
17:01 jberger git diff | mojo nopaste gist
17:02 Grinnz too bad App::mojopaste is already taken :P
17:02 sri can confirm, phone is quite handy
17:02 jberger different thing
17:02 Grinnz i can't think of a better name to make it a standalone script
17:02 sri for some reason i'm getting worse at hearthstone :/
17:02 * lb likes the packagen trick, as the other classes are just noise, https://metacpan.org/source/LBR/Mojolicious-Plugin-PetalTinyRenderer-0.05/lib/Mojolicious/Plugin/PetalTinyRenderer.pm
17:03 mst no_index ftw
17:03 * sri doesn't like the trick
17:03 lb sri hates everything i like
17:03 * Grinnz thinks if it's a legit module, people should be able to find it unless there's a real reason not to index like dist conflicts
17:04 mst Grinnz: but then if you remove it, you end up with the pause index detritus problem
17:04 jberger lb: as mst says, if you really want to hide it, I would use no_index
17:04 Grinnz eh
17:04 Grinnz i guess that's a question of at what point in the process you discover the conflicts
17:04 sri lb: you had to attack donuts...
17:04 jberger but I still wouldn't want some other module to stop on my packages, just because I was over zealous about no_index-ing them
17:04 lb sri: i'm anti-sugar
17:04 sri :O
17:04 Grinnz lol
17:05 lb if you remove carbs and vegetable oil from donuts, what's left? ;)
17:06 * jberger grabs DBIx::Class::Candy before lb takes it away!
17:06 lb jberger: I used to use Class::DBI::Sweet ;)
17:06 Grinnz ::Bro
17:06 lb someone needs to say that once in a while, to keep mst humble
17:07 mst I remember that code
17:07 mst somebody claimed it was impossible to implement prefetch for ::Sweet
17:07 mst sometimes I wish I hadn't proved them wrong
17:09 lb it is ever thus
17:09 ribasushi mst: without the test suite to show what people want, my iteration would not happen, so all worked out for the best ;)
17:10 bif joined #mojo
17:16 * sri wonders how sharding in postgres 9.5 will be working out
17:19 lb that's a word you have to be careful with
17:19 lb I sharted :(
17:20 disputin joined #mojo
17:21 wingfold_ joined #mojo
17:27 batman Grinnz: what do you want to do with the App::mojopaste namespace?
17:28 Grinnz nothing, i was just trying to think of an executable name for a standalone version of mojo nopaste
17:28 batman oh. well mojopaste is the server counter part
17:29 Grinnz yeah, i saw
17:29 mst Grinnz: App::mo'paste
17:29 Grinnz hah
17:31 * lb deprecates mst before any more damage
17:35 pink_mist Grinnz: could use App::mojopaster
17:37 lb the server should be App::mojopastee then
17:37 Grinnz lol
17:37 genio App::mojopaste-it :/  I suck at naming things
17:50 Adura Sometimes I wish Mojo::DOM had a debug env var.
17:51 cfedde what whould the debug env var do?
17:51 Grinnz remove the bugs, of course
17:52 cfedde duh!
17:52 cfedde that would be cool!
17:52 lb it would not
17:52 * jabberwok hands round DWIM keycaps
17:52 lb terrible idea
17:53 lb as a consultant i do not approve of people solving the halting problem
17:53 cfedde I for one will be happy when I can finaly automate myself out of a job.
17:53 lb must be nice being independently wealthy ;)
17:53 jabberwok cfedde: if you're a cashier at McDonald's, you could have that opportunity today
17:54 cfedde lb: you can have the job pumping gas in New Jersey.
17:55 cfedde is it worth paying someone to do if your customers are willing to do it for themselves?
17:56 cfedde every time I deploy a new website that makes some process easier I've cut the number of hours that someone gets to work.
17:57 Grinnz indeed, my job is automating things half our company does
17:57 Grinnz but it's so we can scale our business up 10x, not to put them out of a job :P
17:57 lb oh, sure, my job could easily be replaced by an army of drones
17:58 ajr_ joined #mojo
17:59 lb cfedde: my job is putting other people out of work, not myself! I mean, sheesh :)
17:59 cfedde it does turn out to be a two way street.
18:00 Grinnz people are just not as fast as computers at doing predictable things
18:00 Grinnz (or as reliable)
18:00 * lb is currently doing self service for a nic
18:00 cfedde at least my code usualy makes the same mistake each time.
18:01 Grinnz hehe
18:01 lb it's really just cost control
18:02 cfedde heh: _just_
18:02 lb of course
18:03 lb growing number of customers leading to growing number of supporters is bad business ;)
18:12 Britzel|2 joined #mojo
18:16 cfedde it has always been a ratio thing. Some industries have been more successful than others in pushing the ratio closer to zero.
18:16 sh4 joined #mojo
18:22 wingfold_ joined #mojo
18:33 Mikey devs should get a percentage cut of all savings realized by their creations.  also we need to make a strong tech worker's union and start hoarding knowledge, so we can gain the leverage necessary to negotiate deals like this.
18:39 cpan_mojo Mojolicious-Plugin-AccessLog-0.007 by GRAF https://metacpan.org/release/GRAF/Mojolicious-Plugin-AccessLog-0.007
18:41 gatitskiy joined #mojo
18:42 berov joined #mojo
18:43 bin_005 joined #mojo
18:46 Grinnz_ > Increase minimum Mojolicious version to 3.28.
18:46 Grinnz_ LIVING ON THE EDGE
18:49 mattastrophe joined #mojo
19:04 ajr_ joined #mojo
19:06 mst Grinnz++
19:12 trone joined #mojo
19:14 doby el8sch24
19:14 doby oops wrong window
19:14 purl the next time we have sex.... can it be to the tune of star trek??
19:14 doby purl you so naughty
19:14 purl doby: sorry...
19:17 Grinnz_ hahaha
19:20 wingfold_ joined #mojo
19:21 dod joined #mojo
19:21 mst purl++
19:32 wingfold_ joined #mojo
19:41 pink_mist botsnack
19:41 purl thanks pink_mist :)
19:44 disputin joined #mojo
19:50 doby joined #mojo
20:03 bin_005 joined #mojo
20:10 bwf joined #mojo
20:16 bjakubski joined #mojo
20:23 aggrolite joined #mojo
20:33 mattastrophe joined #mojo
20:46 thowe joined #mojo
20:46 marty joined #mojo
21:12 bin_005 joined #mojo
21:17 stephan48 what would be the easiest way to kickoff an reccouring minion job? i want to run a background task once a minute.. is there a nicer way then just having a cronjob inject the job?
21:18 stephan48 currently i am still using morbo but i will migrate to hypnotoad soon
21:18 jberger stephan48: the job could re-add itself at completion, but that sounds a tiny bit insane
21:18 stephan48 yea
21:19 stephan48 i thought about that but i decided that i would need a failsafe anyway when somehow the job or the worker crashed before completion
21:19 stephan48 (or add it at the beginning)
21:21 stephan48 i think i am going to settle for the cronjob method
21:22 stephan48 thanks!
21:51 jberger cron does that one thing really well
21:51 jberger stephan48: build a command and have cron execute that command
21:51 jberger simple and effective
22:10 dvinciguerra joined #mojo
22:14 mattastrophe joined #mojo
22:29 damaya_ joined #mojo
22:45 punter joined #mojo
23:04 DaniBunny joined #mojo
23:16 aggrolite joined #mojo
23:44 aggrolite joined #mojo

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