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

IRC log for #mojo, 2014-08-30

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

All times shown according to UTC.

Time Nick Message
00:01 sri yea, it's great
00:48 woz joined #mojo
00:53 runester joined #mojo
00:54 runester Hello! I have a quick question if anyone is about.
00:58 batman Extend LinkEmbedder to embed ALL THE THINGS! :)
00:59 batman sri: no idea about the 404. Sorry :/
00:59 batman Maybe you left a channel..?
00:59 runester In case anyone is watching ...
00:59 runester I've been building small Mojoliciouis::Lite applications and noticed that though the documentation states I can place static files in "public/" I actually have to place them in "public/public/" for them to get served. Is this deliberate?
01:00 batman Will look at the translucent thing to
01:00 batman ..tomorrow
01:01 batman runester: probably because you refer to the as /public/ in you HTML
01:02 batman Like my-app/public/foo.jpg need to be included as <img src="foo.jpg">
01:02 * batman goes to bed
01:03 runester OK, thank's batman ... I'll check on that.
01:03 ryozi joined #mojo
01:06 runester Ha! Got it. I was using a relative URL, but I was specifying "public/jquery.js" and didn't realize that 'public' was assumed, it didn't need to be specified. Now the path is just "jquery.js" and it's served out of my public folder.
01:06 runester That's been bugging me. Thanks!
01:15 runester OK. Have a good night, everyone! :)
01:15 runester left #mojo
01:16 dvinciguerra joined #mojo
01:28 klapperl joined #mojo
01:46 good_news_everyon joined #mojo
01:46 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Ug95Hw
01:46 good_news_everyon mojo/master c93a5b9 Sebastian Riedel: added fun example for content negotiation with static files
01:46 good_news_everyon left #mojo
01:47 sri question is a bit too common, might as well add a fun example for it
01:48 sri argh, typo!
01:50 good_news_everyon joined #mojo
01:50 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/zdVmyw
01:50 good_news_everyon mojo/master 1556ba0 Sebastian Riedel: fixed typo in content negotiation example
01:50 good_news_everyon left #mojo
01:56 Zx3 joined #mojo
02:16 disputin joined #mojo
02:33 noganex_ joined #mojo
02:53 sujithm joined #mojo
02:54 woz joined #mojo
02:56 jeneric joined #mojo
03:00 jberger that is a cool example!
03:00 jberger sri++
03:03 preaction joined #mojo
03:08 jeneric left #mojo
03:19 r0b3rt joined #mojo
03:20 good_news_everyon joined #mojo
03:20 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Exdn_A
03:20 good_news_everyon mojo/master ea73cf7 Sebastian Riedel: a few more additions to the tutorial
03:20 good_news_everyon left #mojo
03:22 sri \o/
03:22 sri the formats section was also a little weak
03:53 d4rkie joined #mojo
03:57 woz joined #mojo
03:59 good_news_everyon joined #mojo
03:59 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/kliyQA
03:59 good_news_everyon mojo/master beab983 Sebastian Riedel: more diverse PSGI examples
03:59 good_news_everyon left #mojo
04:11 r0b3rt What do we think about being able to set a custom buffer size here:  https://github.com/kraih/mojo/blob/master/lib/Mojo/IOLoop/Stream.pm#L113
04:12 good_news_everyon joined #mojo
04:12 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/D6Urtw
04:12 good_news_everyon mojo/master 725c0de Sebastian Riedel: show how to customize applications in PSGI environments
04:12 good_news_everyon left #mojo
04:16 d4rkie joined #mojo
04:21 good_news_everyon joined #mojo
04:21 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NTEu_A
04:21 good_news_everyon mojo/master ed64d42 Sebastian Riedel: better middleware recipe
04:21 good_news_everyon left #mojo
04:22 preaction joined #mojo
05:00 woz joined #mojo
05:17 basiliscos joined #mojo
06:42 KCL joined #mojo
07:03 woz joined #mojo
07:08 trone joined #mojo
07:43 basiliscos joined #mojo
07:56 irq joined #mojo
08:08 woz joined #mojo
08:21 Vandal joined #mojo
08:39 basiliscos joined #mojo
09:04 denis_boyun joined #mojo
09:05 _eugen joined #mojo
09:13 woz joined #mojo
09:49 sh4 joined #mojo
09:50 sujithm joined #mojo
09:57 sujithm_ joined #mojo
10:18 woz joined #mojo
10:19 batman r0b3rt: why would you want that?
11:20 woz joined #mojo
11:32 corgi joined #mojo
11:42 d4rkie_ joined #mojo
11:43 ryozi joined #mojo
12:05 woz joined #mojo
12:13 d4rkie joined #mojo
12:49 basiliscos joined #mojo
13:16 good_news_everyon joined #mojo
13:16 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/D552MA
13:16 good_news_everyon mojo/master feade2e Sebastian Riedel: improved monkey_patch to be able to name generated functions
13:16 good_news_everyon left #mojo
13:16 sri Sub::Util is out \o/
13:17 sri no more anonymous functions in nytprof
13:42 * jberger installs!
13:44 Kundun joined #mojo
13:44 Kundun joined #mojo
13:48 jberger someone asked about the mojolicious url in another channel, so I gave them the .us one
13:48 jberger then I gave them the .org one too, just for completeness
13:49 jberger then I wondered what the .com had, so I checked, and its basically just parked
13:49 jberger sri: have you inquired about the .com or is there any way to take it?
13:52 batman what is steady_time() ? i thought it was unix time, with floating point precision, but perl -Mojo -le'print join " ", Mojo::Util::steady_time, Time::HiRes::time' returns 15093.203765606 1409406686.75495
13:52 batman so what can i use steady_time for?
13:55 sri jberger: nope, i have not
13:56 sri batman: http://mojolicio.us/perldoc/Mojo/Util#steady_time
13:57 batman i looked at it, and i read "High resolution time" as in "C<time()> with floating point precision"
13:58 batman sri: ^
13:58 sri google for monotonic clock
13:58 batman ok
13:59 basiliscos joined #mojo
13:59 sri short version, a monotonic clock can't ever jump backwards
14:01 batman looks like it's relative to reboot time
14:01 batman s/ re/ /
14:01 marcus batman: it's always moving forward.
14:02 marcus never back.
14:02 jberger batman: imagine someone changes the system clock
14:02 sri say you have Mojo::IOLoop->timer(5 => sub {...}), now suddenly your system clock gets set back by a year, when would that timer trigger?
14:02 batman ok, so i guess i should never use steady_time to store something in a database
14:03 sri never
14:03 jberger sri: whois tells me that the .com registration expires on 9/17
14:03 batman right. i think i get it now
14:03 batman marcus: i'm fixing the unit tests for day_change now
14:03 jberger not sure what time zone that is
14:06 good_news_everyon joined #mojo
14:06 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/S8sRsA
14:06 good_news_everyon mojo/master 9e62f76 Sebastian Riedel: better description for steady_time
14:06 good_news_everyon left #mojo
14:08 jberger gah! I have forgotten what a mess it is trying to jump a domain name
14:08 jberger nm, I don't think I'm interested
14:08 marcus batman: I was reading diff, and wondering if there's not an easier way to do the day changed helper.
14:08 batman marcus: please tell me how.
14:08 marcus batman: Time::Piece is core? Use it, and ->ymd ?
14:09 batman just day you mean?
14:09 batman we already use Time::Piece
14:10 batman marcus: do you want to make a new object each time? i think it would use cpu that we don't need to use...
14:10 * batman is sad :(
14:10 batman when did i start thinking about cpu time?
14:10 jberger perl -MSub::Util=set_subname -MCarp::Always -E 'set_subname(demo => sub{ die "argh"})->()'
14:11 marcus batman: but if we already use it to format the epoch, maybe we can just do it a bit earlier?
14:12 batman marcus: go crazy. i don't want to change this now. i just thought i should get the feature complete, since you got it in the CHANGELOG ;)
14:13 marcus batman: yeah, I thought it had landed already :)
14:13 batman yeah.
14:13 marcus batman: seems we only use format_time in topic_by.html.ep
14:13 * batman need to eat. then i will continue on connection manger
14:14 batman marcus: i'm happy to give you feedback if you decide to change it, but i don't have the energy to work on it.
14:14 sri btw. i liked it when the timestamps next to the messages showed the full date for everything before today
14:14 batman i hate working on issues regarding dates + epoch + timezone + .....
14:15 batman sri: you want to change it, or you like it as it is?
14:15 batman marcus: sorry, i'm not your DateTime guy. this makes me exhausted :(
14:15 sri it was like that 2 days ago i believe
14:15 batman sri: it's always been like that :)
14:15 sri just noticed it is gone
14:16 sri maybe it was a bug i liked? Oo
14:16 batman sri: wow! no, it was supposed to be like that
14:16 batman wonder when that went away :(
14:16 * batman checks the log
14:23 marcus they are there even in this channel if you scroll far enough back.
14:23 marcus Seems like they are only shown for two days back, not one?
14:25 batman thanks for the input
14:28 sri :o
14:28 batman sri: marcus got it: it only adds date/month if it's more than 24h ago.
14:28 batman not if it's yesterday
14:29 * batman forgets about the issue
14:29 batman marcus or sri: please fix it if you like :)
14:30 batman https://github.com/Nordaaker/convos/blob/master/lib/Convos/Plugin/Helpers.pm#L277 # it's here
14:30 berov joined #mojo
14:31 sri *crickets*
14:32 batman hehe
14:32 batman i wonder if there's any programmer that enjoys working with date and time...
14:32 basiliscos1 joined #mojo
14:33 sri it's not too bad in ruby
14:33 batman i think we could save ONE GAZILLION trucks with money each year, if you just used some world time or space time or something universal...
14:33 batman timezones-- # sri: it's bad anywhere.
14:34 batman it's not obvious that convos is able to show you the correct time.
14:35 batman we send your localtime to the server (and save it in your session), so we can make sure it's the correct time no matter where you are in the world.
14:35 batman fun, fun, fun :(
14:35 batman it will be a lot simpler when we move to jsx
14:41 good_news_everyon joined #mojo
14:41 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/8IAQ1Q
14:41 good_news_everyon mojo/master e6466ae Sebastian Riedel: test monkey_patch with names
14:41 good_news_everyon left #mojo
14:41 jberger joined #mojo
14:42 sri batman: btw. did you see the translucent menubar example i linked to?
14:42 sri that kinda yosemite style could look cool in convos
14:43 sri http://www.telerik.com/fiddler # the bar that appears when you scroll down
14:44 sri maybe a little less translucent
14:54 sri in a quick test it does look nice
14:54 sri needs a few more tweaks though
14:57 sri the basic idea http://i.imgur.com/6zrZpLj.png
14:57 sri didn't fix the spacing, sidebar, or gray boxes you use for active links
14:59 sri basically all i did was change background-color to rgba(255,255,255,.95)
14:59 batman sri: the navbar used to have alpha in it, but i remove it.
14:59 marcus And kill the drop shadow :)
14:59 sri remove the drop shadow and add a little bottom-border
15:00 sri bottom-border is 1px solid #e5e5e5
15:01 * batman gets distracted
15:01 batman $batmna->mode("ignore_input");
15:01 batman :(
15:02 sri Global symbol "$batmna" requires explicit package name at batman.pl line 1.
15:02 batman hehe
15:03 batman the think is that my primary goal now is to make a simpler connection manager
15:03 batman ux stuff takes way to much time and it doesn't move convos forward.
15:03 batman sorry, i need to focus on the important stuff
15:03 batman or at least -- i try my best to do that.
15:04 marcus sri: pull request for flat translucent navbar? ;)
15:04 batman also, it must look good with the slide in sidebars
15:05 batman it need to work as a whole, so actual human testing is required.
15:05 sri marcus: i don't have a convos running locally for testing
15:05 marcus sri: it ships with carton, all you need is redis afaik.
15:05 jberger sri: HAHAHAHA # Global symbol "$batmna"
15:05 sri and yes, the sidebar needs updating too
15:05 batman sri: it's extremely simple to set up. just use the free service i linked to
15:06 jberger I was just looking at vedis to see if this could be made self-contained (where redis is not available)
15:08 batman vedis?
15:10 jberger http://vedis.symisc.net/
15:10 jberger embeddable redis-like c library
15:11 * sri thinks we need to be better represented in london and amsterdam
15:12 marcus sri: I heard there's a company who's hiring in Amsterdam ;)
15:21 sri we should get jberger to amsterdam
15:21 jberger wat!
15:21 sri it's a lot safer than chicago
15:22 jberger you want me hit by a bicycle instead!?
15:22 sri or drown in a canal... hmmm
15:22 sri *and/or
15:23 jberger argh, hit by a bicycle, fall into a canal and drown, that's a terrible way to go!
15:23 * jberger hunkers down in Chicago
15:24 jberger sri: you might be right about the representative tho
15:24 jberger I see that sjn went to yapc::eu but talked more about the mongering than the mojoing
15:25 jberger the talk is in my watch later queue
15:25 jberger just watched su-shee
15:27 sri fun talk
15:27 jberger sri: using the get command, is there an easy way to get a certain row/column out of table?
15:28 sri 'tr:nth-of-type(3) td:nth-of-type(1)' maybe
15:28 jberger yeah, that's what I was coming up with too
15:29 jberger I might just keep my example to extracting a certain row
15:29 * jberger has a blog post in the works
15:46 basiliscos joined #mojo
16:32 marcus I liked the su-shee talk.
16:33 dod joined #mojo
16:34 good_news_everyon joined #mojo
16:34 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/5CjfFQ
16:34 good_news_everyon mojo/master 8b66c45 Sebastian Riedel: let assets handle their mtime
16:34 good_news_everyon left #mojo
16:35 irq joined #mojo
16:40 sri app->static->serve_asset becomes quite a bit more useful with that
16:41 sri you can render something, put it into a memory asset, set an mtime based on some stash data you passed to the template, and have serve_asset negotiate etag, last-modified and range
16:42 batman sri: why do you have "sub _check { $_[0] ? $_[1] ? (stat $_[0])[9] : -s $_[0] : 0 }" ?
16:42 batman can't you just do it inside the mtime() and size() methods?
16:44 sri of course you have to render first to do that, so maybe my way of thought is wrong
16:44 r0b3rt joined #mojo
16:46 r0b3rt left #mojo
16:46 sri yea, it's a bit counter intuitive now that we have is_fresh
16:52 good_news_everyon joined #mojo
16:52 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Y5YvrQ
16:52 good_news_everyon mojo/master 29d9bc2 Sebastian Riedel: removed mtime handling from assets again
16:52 good_news_everyon left #mojo
16:54 basiliscos joined #mojo
17:09 dvinciguerra joined #mojo
17:23 basiliscos joined #mojo
17:26 berov Hi guys
17:27 berov May be some of you will be interested to know how Ado works on top of Mojolisious http://www.slideshare.net/kberov/ado-yapceu2014
17:28 berov I see today the videos were uploaded on youtube too
17:29 berov my first talk at YAPC, so be merciful https://www.youtube.com/watch?v=8kBOxd6DQsM
17:31 sh4|2 joined #mojo
17:37 marcus berov: congrats with your talk, I plan to check out Ada some time soon.
17:38 marcus I'm trying to understand the actual meaning of https://twitter.com/vtivti/status/505744640327557120
17:42 berov Thanks :)
17:42 __eugen joined #mojo
17:45 berov marcus: will be very grateful to have some feedback
17:47 berov I wish it evolves to a level were a guy colling him self webdeveloper unpacks it on a cheap shared hosting and builds a site on top of it
17:48 berov Ado provides an htaccess generator
17:48 berov and some other like crud and adoplugin
17:49 berov Ado is not non-blocking, tough :D
17:50 berov still did not tried to find out how to make DBI calls non-blocking
18:00 marcus berov: there has been some work on mojo non-block dbi lately.
18:01 disputin joined #mojo
18:02 berov thanks, markus, will grep the docs.
18:08 berov hm... not finding any mentioning of DBI. just pulled the latest mojo
18:13 og01 joined #mojo
18:14 og01 Hi does anyone know how to db.dropDatabase with Mango?
18:15 marcus berov: didn't mean in core. jzawodn was experimenting yesterday, and batman has been doing something with mojo-sql on github.
18:19 sri og01: i would assume it's just a command
18:21 sri funny thing, i almost started a non-blocking postgres driver and orm instead of mango... but when crab said he was working on something too that settled my decision to focus on mongodb :)
18:21 og01 its a great driver, used it in serveral projects
18:21 batman og01: $connection->storage->db->command(dropDatabase => 1);
18:22 og01 batman: great thanks!
18:22 batman you're welcome :)
18:23 og01 sri: I've used it in a few EV projects aswell
18:24 og01 afaik its the only non-blocking database driver thats anygood
18:24 sri \o/
18:25 og01 are there any tools to help with writing tests under Mojo::IOLoop?
18:25 og01 I wrote some helpers for my projects using EV
18:25 sri like?
18:27 og01 i tend to use a partern of starting a timeout timer, setting up a test, and fetching the result of the test from a callback/promise and/or timing out
18:27 og01 i can write a helper function, but I wondered if anyone had already thoughtup/already exists testing tools
18:28 sri not that i know of
18:28 og01 no problem, just wondered
18:30 berov marcus: thanks. will look at mojo-sql.
18:32 meshl joined #mojo
18:34 berov batman++ ... and why not using DBI for mojo-sql ?
18:35 berov is it too hard to get the connection file descriptor? or how it is called exactly - not sure
18:41 marcus berov: jzawodn is doing it with mysql and getting the file descriptor. Don't think he has published anything yet.
18:42 KCL_ joined #mojo
18:42 marcus <jzawodn> yay, Mojo::IOLoop based sphinx query server handles 350 req/sec with lots of headroom using async MySQL interface
18:45 sri didn't jberger finish his non-blocking DBD::Pg module?
18:45 batman berov: not much to do ++ for, i haven't really looked into my project for some while :(
18:45 batman berov: i don't want to use DBI, since it doesn't seem to really care about non-blocking
18:45 sri you know, the one from mojoconf
18:49 berov so this is solved on DBD level by the drivers
18:50 berov I read somwhere sqlite will provide async natively whatever that means
18:51 berov I will be happy if I can make those 3 (mysql,sqlite,postgre) work
18:53 berov may be DBI can provide some API et least and let the drivers handle this?
18:54 sri good luck with that
18:54 purl SOL loser!
19:03 sri hmm... async DBD::Pg seems terrible after taking a closer look
19:05 sri you connect blocking, send the query blocking, then wait non-blocking for the number of results, and fetch the results blocking again
19:05 berov DBD::mysql seems to provide async queries
19:05 jberger sri: I stopped working on it when I found Mojolicious::Plugin::Pg
19:06 jberger *PgAsync
19:06 jberger https://metacpan.org/pod/Mojolicious::Plugin::PgAsync
19:06 berov https://metacpan.org/pod/DBD::mysql#ASYNCHRONOUS-QUERIES
19:06 sri same for DBD::mysql, also bad at non-blocking
19:08 sri i suppose the drivers have to work around limitations in DBI
19:10 berov sri: what do you mean by "good luck with that"
19:11 berov ?
19:17 berov ...although the setup with hypnotoad or fastcgi/mod_perl works for now for me. may be I still do not understand the real advantage of asynchronous stuff
19:25 Eke- joined #mojo
19:33 jberger I don't know why it is seemingly so hard to get a non-blocking driver for SQL dbs, but it seems to be
19:33 sri jberger: are you sure the connection pool is necessary?
19:34 sri as far as i can see waiting for multiple async queries is no problem, and most operations are blocking anyway
19:34 sri (on one dbh)
19:34 jberger if that is the case, then no, but I thought there was some state being kept in the dbh IIRC
19:35 jberger I guess it would really need to be the wire-protocol route to be truly "better" than that, right?
19:35 sri hmm, the plugin seems to be getting stuff wrong actually
19:35 sri it disconnects instead of canceling operations
19:36 sri you know... i'm soooo close to releasing a Mojo::Pg just to show how it's done ;p
19:39 sri the whole timeout thing is not good imo
19:40 sri i would not recommend the plugin
19:41 jberger would your Mojo::Pg use one of the existing underlying modules or would it implement the wire protocol (ala Mango)?
19:42 sri it would just use DBD::Pg properly
19:42 berov sri: please do release it even ifit just for the example
19:43 sri postgres wire protocol is crazy, not getting anywhere near that
19:44 sri only thing keeping from doing it is that i would have to spend some time reading up on postgres to make sure everything works... and i can't really afford to atm.
19:45 sri a decent version with tests and docs should be doable in a week otherwise
19:48 preaction joined #mojo
19:54 jberger are either of the branches made at or after MojoConf worth saving? https://github.com/marcusramberg/mojo-pg
19:55 sri not really, the one sth limit seems pointless
19:56 sri wait a minute
19:56 jberger it was a reaction to the belief that it was the only way it could work, I truly don't remember why that was at the time tho
19:56 jberger I'm rather sure we had a reason
19:56 sri why the hell is pg_result() called on the dbh?
19:57 sri can there actually only be one active async sth?
19:57 sri that makes no sense at all
19:58 * jberger seems to recall going through that thought process at the hackathon at mojoconf
20:00 sri yea, looks like there actually is such a limit
20:00 sri allright... make that two weeks...
20:02 sri still wouldn't recommend the plugin
20:05 sri if someone puts a little more effort into it, you might even be able to have pluggable backends for pg and mysql
20:05 sri and fake it with sqlite (next_tick)
20:08 sri that's how i would do Mojo::DBI
20:15 lipizzan joined #mojo
20:16 sri the world would be so much better if all databases used http or the memcached/redis protocol
20:18 jberger http://xkcd.com/927/
20:18 jberger oh man, no xkcd integration?!
20:19 jberger http://imgs.xkcd.com/comics/standards.png
20:19 sri convos sucks!
20:19 sri this is unacceptable
20:20 preaction no good metadata in there. it'll have to be custom :(
20:21 * jberger starts working!
20:22 jberger btw, marcus/batman, any thoughts on my og fork?
20:22 * jberger opens pull request
20:23 batman sri: could you make an issue for LinkEmbedder for xkcd support?
20:23 batman jberger: I thought I had expressed myself clearly about your branch
20:23 jberger true
20:23 jberger marcus, thoughts?
20:23 batman I think you're making a super complex solution
20:24 batman I might have to hand the project over to you if you move it in that direction
20:24 batman You and Marcus can decide.
20:25 marcus jberger: I think it's batman's project, so his opinions are more important.
20:25 marcus (the embedder module is that is)
20:25 jberger true, I see you two working together so often, I sometimes forget to ask
20:25 jberger batman: your project, your decision
20:26 jberger my "complexity" was only to be able to support multiple types of meta data, but if you think that it can be handled together, then, ok
20:26 sri i can't even tell them apart anymore
20:26 * jberger works on xkcd support
20:27 * marcus pushes sri into the alligator pit.
20:27 sri the fact that marcus's avatar still wears glasses sure doesn't help ;p
20:27 sri AAAAAAAAAAAH
20:27 marcus Guess I should update my gravatar. Or start wearing glasses again.
20:34 batman jberger: marcus often saves me from
20:34 batman Crap.
20:35 batman jberger: marcus often saves me from my stubbornness, so I'm not sure if I would mind.
20:35 marcus I'd like to see the alternative to jberger's branch.
20:35 batman What I do mind is having to jump between modules that I consider solving the same thing.
20:36 batman marcus: I'm on my phone...
20:36 batman But the branch is on GitHub
20:36 batman Tok bad CodeHub (iOS app) doesn't allow copying links :(
20:36 marcus batman: I've seen his branch, I'm wondering how the alternative solution looks.
20:37 batman Yeah, and I'm saying my (incomplete) branch is also on GitHub
20:37 marcus feature/opengraph?
20:39 batman sounds right.
20:39 batman There should only be two og branches there
20:42 marcus I think I'm fine with batman's solution.
20:43 batman Mine is super basic. Does nothing more than I need it to...
20:45 basiliscos joined #mojo
20:47 marcus batman: why would you do things you don't need to do?
20:47 marcus :p
20:48 batman I rarely do. That's why I'm excused from most meetings at work ;)
20:51 cpan_mojo Mojolicious-Plugin-BootstrapHelpers 0.002 by CSSON - http://metacpan.org/release/CSSON/Mojolicious-Plugin-BootstrapHelpers-0.002
21:18 jberger so this was a fun one
21:18 good_news_everyon joined #mojo
21:18 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/2Dzf9w
21:18 good_news_everyon mojo/master c77d69b Sebastian Riedel: slightly faster range handling
21:18 good_news_everyon left #mojo
21:18 jberger I was having a brain fart and I wanted to check that an empty hashref evaluated to true
21:19 jberger so I did this: perl -E 'say {} ? "yes" : "no"'
21:19 jberger guess what error THAT gives :-P
21:21 sri Oo
21:22 sri this one works perl -E 'say [] ? "yes" : "no"'
21:26 sri go home perl, you're drunk!
21:26 jberger hehe
21:26 marcus perl -E'say \{} : "yes" : "no"' -> yes ;)
21:27 jberger if I understand correctly, the {} is being interpreted as a block evaluating to a filehandle, so then the next instruction is ? which is a ?PATTERN? I guess?
21:33 sri how odd, just found this patch lying around here https://gist.github.com/anonymous/eb7b8ea9ff0e23719dc3
21:33 jberger batman: https://github.com/jhthorsen/mojolicious-plugin-linkembedder/pull/4
21:34 jberger sri: I seem to recall that one
21:35 sri i just vaguely remember
21:35 jberger you didn't like the method dispatcher hacks (and probably tight coupling too)
21:38 irq_ joined #mojo
21:51 bobkare joined #mojo
21:57 good_news_everyon joined #mojo
21:57 good_news_everyon [mojo] kraih tagged v5.35 at 2b4f779: http://git.io/NKxa8A
21:57 good_news_everyon left #mojo
21:58 cpan_mojo Mojolicious 5.35 by Sebastian Riedel - http://metacpan.org/release/SRI/Mojolicious-5.35
21:58 good_news_everyon joined #mojo
21:58 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/4GdTiQ
21:58 good_news_everyon mojo/master cbce73c Sebastian Riedel: bump version
21:58 good_news_everyon left #mojo
22:02 basiliscos joined #mojo
22:10 jberger sri: the new __DATA__ handling broke Galileo!
22:10 sri yay?
22:11 sri how odd
22:11 jberger in the setup command I replace the router with a completely new router which renders templates in its __DATA__ section
22:12 jberger https://github.com/jberger/Galileo/blob/master/lib/Galileo/Command/setup.pm
22:12 sri umm...i...can't...even...
22:14 sri ah, because warmup runs on ->start, which is before the command
22:16 jberger I can move those templates to real files
22:17 jberger it was so cute as a one file solution though
22:17 sri may or may not work
22:17 sri template list is also pregenerated on warmup
22:17 sri for certain detection jobs
22:18 sri running app->renderer->_warmup again would work ;p
22:18 jberger so is there no way to change (or even add) templates at runtime (even if before the first request)
22:18 sri no
22:18 jberger :(
22:19 jberger ok, well, I don't much like that
22:19 jberger and calling _warmup seems like something that sri would scold jberger for in a few months
22:20 sri you know what to do, but there's a lot of stuff to consider... like if you want to formalize a complete reset there are more caches
22:20 sri oh, i would scold you today!
22:20 marcus jberger: you have to do it at startup time.
22:21 sri commands happen later
22:21 jberger but I can't know that the setup command was called until runtime
22:21 marcus Can't you just always load the templates regardless of setup command? Or too much overhead?
22:22 sri why it has to be the main app i'm not sure though
22:22 sri right
22:22 sri merge or split them completely
22:22 sri this feels very much like a hack
22:22 jberger security, I don't want any chance that an clever client can see how to change my configuration file
22:23 jberger ^^ that was @marcus
22:24 sri anyway, your feature proposal would be new templates at runtime
22:24 sri and i can already tell you it will be a tough sell
22:24 jberger aahhh, but it worked already!
22:25 sri haha, allright, lets revert the __DATA__ fix
22:26 sri ;p
22:26 * jberger reminds sri that he worked very hard to try to find any other fix for that
22:26 marcus https://twitter.com/TokenScandi/status/505796884871802880
22:26 meshl joined #mojo
22:27 * sri reminds jberger that he has not found one
22:27 sri marcus: i know, i ade the original complaints ;p
22:27 sri *made
22:28 marcus Sri: are you involved in the new plans?
22:28 sri 2 or 3 years ago actually... and the solutions being talked about are the ones i recommended too xD
22:28 sri nope, i'm still an outlaw
22:29 marcus Would be nice if they got it right.
22:30 sri i think that ship has sailed... they can only bolt on some fixes now
22:30 sri namely a close event and a drain event handler
22:31 sri and it's less async, more fix the current streaming api
22:33 sri best case endresult for us would be that we can do streaming responses on EV based psgi servers
22:33 preaction joined #mojo
22:33 sri that's it
22:34 marcus Guess we'd be hard pressed to run something like convos on it then.
22:35 jberger I guess I can load the templates at start time
22:35 bobkare joined #mojo
22:35 sri honestly, imo the age of rack, psgi, wsgi is over
22:35 jberger they don't do anything security related
22:36 jberger the logical is all in the command/new router
22:37 marcus jberger: \m/
22:38 jberger I kinda kicked over the hornets nest in Alien::Base world today too, so I'm managing all kinds of things atm
22:41 sri well, you can always flee the country and go to amsterdam *hint hint*
23:10 jzawodn .wi37
23:10 jzawodn oops
23:54 Claudio_ joined #mojo

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