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

IRC log for #mojo, 2016-01-08

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

All times shown according to UTC.

Time Nick Message
00:20 sue joined #mojo
00:31 disputin joined #mojo
00:35 jontaylor joined #mojo
00:47 mad_hatter http://pastebin.com/XvQaWM3c
00:47 mad_hatter I keep getting the error 'template not found' when I try to do redirect to the confirmation page
00:47 mad_hatter any ideas?
00:56 genio is that a copy/paste error or are you missing an = here     my $confirmation  $r->under ...
01:00 mad_hatter copy paste error
01:00 mad_hatter there's an = there
01:57 miller joined #mojo
02:39 jberger Awwwww yea
02:39 jberger Another jberger module that triggers an abort on Windows
02:39 jberger http://www.cpantesters.org/cpan/report/dd9c0c6c-6c12-1014-8e56-f5042b7077ee
02:40 sri \o/
02:40 jberger Though this time I think I probably get to blame Mojo::SQLite and Minion::Backend::SQLite
02:41 jberger And maybe my stuff too
02:41 jberger Who knows! It's Windows!
02:46 kaare joined #mojo
02:46 Grinnz considering how many win32 cpantesters fails i get from Minion::Backend::SQLite, yeah you can probably blame one of those :P
02:46 Grinnz not sure if ive ever seen *that* error message though
02:47 jberger Oh I have
02:47 jberger It's an abort()
02:48 jberger ForkCall still might throw them de temps en temps
02:48 Grinnz was that french? cause it almost translates in spanish
02:50 jberger I think so, it's been a while
03:03 punter joined #mojo
03:03 noganex_ joined #mojo
06:41 kes joined #mojo
06:54 orev joined #mojo
06:56 dod joined #mojo
07:02 dod joined #mojo
07:07 sue joined #mojo
07:50 trone joined #mojo
08:14 Vandal joined #mojo
08:23 AndrewIsh joined #mojo
08:37 osfabibisi joined #mojo
08:49 ashimema joined #mojo
09:25 ashimema joined #mojo
10:52 CandyAngel Is everything in Mojolicious broken out into separately installable (from CPAN) modules?
10:53 CandyAngel I'm considering using EventEmitter and Collection in my OpenHMD modules..
10:54 henq morning y’all, I want to get to the template name from within a layout ; I tried % my $mt = shift;
10:54 henq % my $name = $mt->name;       but that does not seem to work; any hints? tia.
10:56 CandyAngel It might be in the stash as mojo.template?
10:56 CandyAngel I have the feeling this has been asked before
10:57 CandyAngel http://mojolicious.org/perldoc/Mojolicious/Controller#stash
11:00 henq @CandyAngel  thx!
11:02 CandyAngel My pleasure :)
11:03 nic henq: You're not supposed to access private fields (that's what the 'mojo' prefix denotes)
11:04 nic If you ignore that, it'll work until it suddenly doesn't
11:04 henq I have a footer in the layout that i want to suppress for some templates; other ways?
11:06 henq (and lazy, do not want to include   footer => 1 in many templates)
11:06 nic there are many ways, but it boils down to designing your templates in a more modular way
11:06 henq okay, i could do a suppress => 1 in some , of course
11:07 henq okay, thx, i can work it out.
11:20 jontaylor joined #mojo
11:34 AndrewIsh joined #mojo
11:39 punter joined #mojo
12:01 neilhwatson joined #mojo
12:08 CandyAngel Nevermind, I figured out the answer to my question. Guess I'll have to reimplement or copy them :P
12:08 * CandyAngel steals the modules
12:08 CandyAngel Arrrrrrr
12:10 sue joined #mojo
12:11 CandyAngel Oh no, please don't sue me :(
12:11 * CandyAngel gives the source code back
13:04 sh4 joined #mojo
13:21 sue joined #mojo
13:34 vytas joined #mojo
14:07 jontaylor joined #mojo
14:22 ashimema hmm.. I seem to be having issues install Mojolicious on Perl 5.20.2 at the moment.
14:22 ashimema via cpanm under a plenv environment..
14:22 ashimema anyone else running similar.. know how to narrow down the issue.. it just seems to stall midway through install
14:27 ashimema runinng cpanm -v it appears to get stuck at: t/mojolicious/ojo.t ........................ 1/?
14:28 jberger CandyAngel: http://mojolicious.org/perldoc/Mojolicious/Guides/FAQ#Why-not-split-up-Mojolicious-into-many-smaller-distributions
14:29 jberger ashimema: do you have any MOJO_ vars in your environment?
14:29 ashimema hmm
14:29 ashimema not that I've deliberately put there..
14:29 ashimema will check
14:30 ashimema nope.. not anything show from prinenv with  MOJO_
14:32 jberger ashimema: I don't remember the order of the tests, is that the first test or do others pass before it stalls there
14:32 jberger ?
14:33 ashimema lots of others pass first
14:33 jberger and it always stalls there?
14:33 ashimema seems to.. in three consecutive attempts
14:34 CandyAngel jberger: Ah okies, thankies for the link
14:34 nic ashimema: what happens when you run it in isolation?
14:34 * ashimema feels silly..
14:34 jberger that was going to be my next question too
14:35 ashimema but I have no idea how to before it's installed
14:35 ashimema sure it's easy.. just not something i've tried
14:35 ashimema hitns?
14:35 nic cpanm Mojolicious --look
14:35 nic perl Makefile.PL
14:35 nic make
14:36 ashimema ooh.. nice..
14:36 nic prove t/mojolicious/ojo.t
14:36 nic (there might be a 'make test' variant for individual test; I'm not sure)
14:36 ashimema interesting.. it claims Mojo/Base.pm is missing
14:36 Grinnz that... shouldn't happen
14:36 jberger borked environment?
14:37 ashimema https://ssl.thorsen.pm/paste/21c79e9da513
14:37 jberger try a different plenv/perlbrew?
14:37 nic this is why I hate complicated ('clever') environments
14:37 Grinnz ashimema, can you paste your perl -V ?
14:38 ashimema https://ssl.thorsen.pm/paste/f5395a82a253
14:38 pink_mist 15:36 <nic> prove t/mojolicious/ojo.t <-- that might need to be prove -l t/mojolicious/ojo.t
14:38 pink_mist or prove -b maybe?
14:39 CandyAngel Is it okay if I reimplement them based on Mojo's though?
14:39 ashimema aha.. the prove -l gives me the same 'hanging' that the cpanm -v attempt to install did
14:39 ashimema as does -b
14:40 pink_mist try prove -v -b
14:40 jberger CandyAngel: why not just use Mojo?
14:40 jberger that's always the question at this point
14:40 ashimema hangs after test 19 by the looks
14:40 CandyAngel I don't want people who use my OpenHMD modules to have to install Mojolicious in its entirety
14:41 CandyAngel I'm not like... *super* against doing so though
14:41 jberger Mojolicious is tiny
14:42 jberger by perl module dependency chain metrics
14:42 jb360 joined #mojo
14:44 ashimema there's nothing in that test that looks like it should hang to me..
14:44 ashimema though I don't understand the benchmark stuff at the end to be honest ;)
14:45 nic I didn't spot anything bad in your perl -V
14:45 CandyAngel I was going to say it also means my modules will be subject to minimum perl of Mojolicious, but that would be true if it was separated anyay :P
14:45 nic ashimema: Did all tests pass when you built perl?
14:45 CandyAngel anyway*
14:45 jberger ashimema: that's testing the benchmark function "n"
14:47 ashimema yup.. tests passed when install perl
14:57 melo joined #mojo
14:58 sri feels weird to leave the perl6 community after all those years
15:01 * ashimema goes the 'if in doubt reboot' approach for good measure
15:03 sri jberger: any new refactoring branches?
15:05 Jonis sri: [14:55] [  [Coke] ] .tell _sri not sure if you'll get this, but I wasn't asking you to leave; I was trying to ask what you got out of the experience.
15:06 jberger sri: if after all that time they still don't value production quality then maybe it's time
15:06 Jonis (just passing on what I saw)
15:06 sri yea, it's not like i believe perl6 has a chance
15:08 sri think i'm already happier, a clear cut now makes future decisions much easier, like what to invest in
15:11 CandyAngel Happier is good :)
15:11 genio indeed
15:21 Coke joined #mojo
15:21 ashimema wow.. wasn't expecting it to.. but the restart worked..
15:21 ashimema Mojo installed fine this time!
15:21 ashimema must have been some stale env or somthing hanigng around
15:21 jberger kinda what I expected
15:22 jberger glad it turned out to be nothing worse than that
15:22 ashimema :)
15:25 sri yay, postgres 9.5 is now in homebrew, time to play with the new features! \o/
15:29 Coke left #mojo
15:43 sri time for some final SKIP LOCKED benchmarks
15:44 sri hmm, the difference is not that big with minion_bench.pl
15:45 sri up to 30% sounds about right
15:48 sri so yea, this makes minion around 30% faster https://gist.github.com/anonymous/864058ce7b8b0dda0185
15:53 kivilahtio joined #mojo
15:56 bpmedley Wow.  Thirty percent with a one line change?
15:56 jberger and upgrading pg from 9.4 to 9.5
15:57 jberger the one line change uses a new feature
15:57 genio \o/
15:58 sri yea, SKIP LOCKED avoids lock contention caused by many concurrent queries on the same table
15:59 genio explained decently here http://www.depesz.com/2014/10/10/waiting-for-9-5-implement-skip-locked-for-row-level-locks/
15:59 sri it's rather task specific, very very good for job queue kinda tables
16:00 sugar__ joined #mojo
16:08 jontaylor joined #mojo
16:26 vanHoesel joined #mojo
16:38 vanHoesel joined #mojo
16:44 lluad joined #mojo
16:51 vanHoesel joined #mojo
16:57 Kogurr joined #mojo
17:13 bpmedley joined #mojo
17:44 dod joined #mojo
18:05 PryMar56 joined #mojo
18:09 sri batman: are you going to say more about what your plans for #883 are?
18:13 batman sri: i don't know what to do yet.
18:15 sri :/
18:15 sri what about use cases?
18:17 fkafka joined #mojo
18:17 jberger I thought we did use cases on #mojo-core yesterday
18:18 batman sri: sorry. been working.
18:22 sri jberger: i didn't see any real ones
18:51 sri i think it would actually add quite a bit of complexity to Mojo::IOLoop::Server
18:51 sri and that's only the easy part
18:52 sri then you have to deal with Mojo::Transaction::local_*/remote_* and find a way to make it work with Mojo::UserAgent
18:52 sri and keep it working in the future
18:53 sri i ws hoping batman had real plans that could convince me there is not too much technical debt
18:57 sri the use cases we talked about were all meh at best
18:57 ribasushi heh, something I ran across while searching for a tweet: https://twitter.com/ribasushi/status/625339764275867648
19:01 jberger ribasushi++
19:19 punter joined #mojo
19:24 c-- joined #mojo
19:30 sri batman: i know i said my vote is neutral, but i tend towards -1 now
19:31 batman yeah... me as well :(
19:31 batman i can't seem to find my reasons for implementing it
19:31 batman vacations--
19:32 zivester joined #mojo
19:33 batman the only argument i can seem to find online is "file permissions", but... i don't care about that :P
19:34 jberger I think it might be kinda nice to use with persistent proceses communicating over websocket
19:34 jberger but there's no real reason that couldn't just be ports
19:36 sri you mean like a desktop app?
19:37 batman when it comes to security and firewall... why not just listen to localhost instead of 0.0.0.0..?
19:37 batman won't that be the same?
19:39 punter joined #mojo
19:51 jberger sri: Mercury lets you do websockets as message broker, so with unix domain sockets it could be easy IPC
19:51 jberger again, not that it couldn't be done over local HTTP port, but its the best use case I can think of
19:52 Grinnz_ to do it over a port, everyone involved needs to know the port, which is harder than everyone involved knowing the socket file
19:52 jberger exactly
19:52 mspo socket files are faster
19:52 jberger plus its easier to be sure that its not open to the outside world
19:52 mspo except they have some weird/different error states if they get too busy I think
19:53 c-- what's the recommended way to have a common cache or data structure that is accessible to all (forked) hypnotoad child servers? maybe ipc::sharedmem somehow?
19:54 mspo memcached? :)
19:54 jberger c--: I'd recommend a database
19:54 Grinnz_ memcached, redis, ...
19:54 jberger redis for simple stuff, pg for real stuff
19:54 jberger really simple can be DBM::Deep but it doesn't scale well
19:55 jberger but its good for prototyping
19:55 Grinnz_ i wouldnt use it for caching
19:55 jberger true
19:55 jberger TTL is useful there
19:55 Grinnz_ redis keeps everything in memory, as does memcached hence the name, so that's a benefit there
19:55 c-- jberger, ok one of the keys i'm keeping in my common data structure could easily be stored in a database, so thanks - but my main use case is caching so i was wondering if it's better to use shared memory rather than a permanent db?
19:56 jberger c--: I've never done that
19:56 jberger check out redis though
19:56 Grinnz_ redis is close enough to shared memory, and much less complex
19:56 c-- i'll check it out
19:56 mspo memcached also supposed a socket file :)
19:56 jberger plus redis keys have TTL
19:56 Grinnz_ yeah redis has some neat features for caching
19:57 bpmedley c-- : I know people use memcache for caching.  Just another data point.
19:57 Grinnz_ i havent really used memcached, i just prefer redis because Mojo::Redis2 is awesome
19:57 jberger also, you might look into CHI
19:58 sri mspo: faster by how much?
19:58 sri so far i've only ever seen anecdotal evidence, but nobody has numbers
19:58 mspo sri: tcp memcached vs unix socket memcached was about 30% for me
19:59 sri is it even relevant for http?
19:59 mspo sri: udp vs socket file was slightless less of a margin
19:59 mspo what http client would support a socket file?
19:59 c-- should be as simple as possible, i don't think i need an additional complex db backend for this. dbm::deep looks interesting at first glance
19:59 Grinnz_ Mojo::UserAgent if it was implemented :P
19:59 mspo so if you were talking mojo -> mojo on the same host, a socket file would be a nice option
20:00 mspo for a signals alternative (hot reload, etc) I like the nagios setup of a named pipe instead of a socket file
20:00 Grinnz_ c--: then you have disk I/O bottlenecks, less parallelism, slower general access
20:00 c-- hm it should be completely in-memory, not on disk, if possible. it should be fast. actually thats why i thought about sharedmem
20:00 Grinnz_ yes. that's what redis is, as i said :P
20:01 mspo make sure to configure or start redis on localhost; it listens to 0.0.0.0 by default
20:01 mspo and embeds lua, so is dangerous
20:01 Grinnz_ there's really not much to configure for redis. no database names, no permissions
20:01 Grinnz_ just where it listens, and connect to it
20:02 sri mspo: seems like a terrible alternative to signals
20:02 mspo sri: it's handy as an ipc for nagios
20:02 sri mspo: you're changing the subject
20:03 sri or were you talking to c--?
20:03 jberger there are two active subjects atm
20:03 sri yes, i'm not sure which one mspo is talkign about
20:04 mspo sri: unix domain sockets are faster than tcp pretty much always; by upto 30% in my testing with memcached
20:04 mspo sri: which makes sense because they skip all of the kernel networking overhead AND can use big read/write sizes
20:04 mspo sri: whereas network traffic is often limited to tiny MTUs
20:05 mspo the socket size is a big tunable for things like memcached
20:05 mspo sri: in terms of http, using a unix socket would be difficult because so few clients would support it
20:05 sri looks like it doesn't matter anymore anyway https://github.com/kraih/mojo/issues/883#issuecomment-170104526
20:06 sri without someone to implement it it's toast
20:07 c-- jberger, hey CHI looks like a great common interface and i could use it with memcached - thanks for the tip!
20:08 jberger yeah
20:08 jberger pay special mind to the calculate method
20:08 jberger its magic
20:08 jberger s/magic/really nice/
20:08 preaction its layered caches are pretty awesome too
20:09 sri if someone here really cares about unix domain sockets, please speak up https://github.com/kraih/mojo/issues/883#issuecomment-170110340
20:09 sri although, you'd have to volunteer for implementing the feature too
20:12 mspo talking to minion over a domain socket makes sense
20:12 sri minion uses postgresql by default
20:13 bpmedley mspo : Isn't that up to the backend?  Can't you just configure Pg (or whatnot) to use domain sockets?
20:13 sri and what bpmedley said
20:13 jberger it does it however DBD::Pg talks to postgres
20:14 sri unix domain socket support in mojolicious would mostly be releavant for Mojo::Server::Daemon based servers, and Mojo::IOLoop::Client based clients... like Mojo::Redis2
20:15 mspo I mean if you had many processes talking to the same queue
20:15 jberger (reading my comment, it could be confusing, I was saying that it isn't relevant for Mojo::Pg)
20:15 mspo like if I had a cron that checked minion for things to do?
20:15 sri still postgresql
20:16 mspo I haven't actually looked if that's a normal use case since I mostly use rabbitmq for that kind of stuff
20:18 asarch joined #mojo
20:18 sri this one seems also pretty much over, although i can't shake the feeling that we are missing something https://github.com/kraih/mojo/issues/891
20:20 bpmedley Any thoughts on my javascript notifications? https://github.com/brianmed/notify/blob/master/templates/index/slash.html.ep#L19-L77
20:24 bpmedley Might be useful as a plugin on CPAN.
20:25 jberger there's almost no reason that multiplexing couldn't be a cpan plugin
20:25 bpmedley Given that, what do y'all think of the current example code?
20:25 preaction i'd love to see some kind of JS library for Mercury's message patterns
20:26 bpmedley Perhaps me and preaction could work on that.. :-0
20:26 sri bpmedley: you're not using a database, so it doesn't really count
20:27 preaction why does that matter?
20:27 sri the point here is to not just have messaging between one browser and a server, it's to have message passing between all servers and all browsers
20:27 preaction why do you need a database for that?
20:28 sri database or some other message bus
20:28 bpmedley sri : https://github.com/brianmed/notify/blob/master/lib/Chat/Notify.pm#L55
20:28 sri postgres or redis
20:28 bpmedley https://github.com/brianmed/notify/blob/master/lib/Mojolicious/Plugin/Notify.pm#L72
20:28 jberger sri: have you looked at Mercury?
20:28 sri bpmedley: oh sorry, i missed that
20:29 bpmedley It's fine.  I hadn't pointed that out and it's kinda hidden w/o any documentation.
20:29 sri wait, the database is initialized in the app?
20:29 sri and the plugin uses it?
20:29 sri that's wrong
20:30 sri jberger: no
20:31 jberger a message broker app just mojo websockets
20:31 bpmedley It would be better to abstract that out.
20:31 jberger because its another process it can broker
20:32 sri i don't really care about a specific broker
20:32 sri the problem is that mojolicious would have to a) choose a broker, or b) have a pluggable broker api
20:32 sri which is pretty heavy
20:32 jberger and why it makes for a better cpan module at the moment
20:33 sri sad thing about that is that we are less competitive that way
20:33 jberger well we can't have it both ways
20:33 jberger a pluggable broker api would be nice though
20:33 jberger I did a really simple one for Minion::Notifier
20:33 sri well, phoenix can have it both ways ;p
20:34 * sri waves fist angrily
20:34 jberger how does phoenix have it both ways?
20:34 sri it can use the erlangvm native message passing
20:35 sri language feature makes the whole thing a non-issue
20:35 jberger sri: https://github.com/jberger/Minion-Notifier/blob/master/lib/Minion/Notifier/Transport.pm
20:35 jberger https://github.com/jberger/Minion-Notifier/blob/master/lib/Minion/Notifier/Transport/Redis.pm
20:36 jberger and one for pg and one for mercury
20:37 sri that kind of abstraction seems more sensible than the one bpmedley uses
20:39 sri ok, bottom line, this needs to be a plugin
20:41 sri https://github.com/kraih/mojo/issues/891
20:41 sri so many issues closed
20:47 hesh joined #mojo
20:47 sri all the remaining issues depend on jberger :)
20:47 jberger :o
20:47 preaction no pressure
20:48 sri http://hope.houptlab.org/OnlyHopeSticker.png
20:50 sri hmm
20:50 sri i kinda want to make a perl raptor shirt in this style https://supporters.eff.org/files/unicorn-shirt-500b_0.png
20:50 sri or mojolicious cloud shirt
20:51 disputin joined #mojo
20:51 punter joined #mojo
21:00 hesh hi.... I have a simple login form with name and password fields. If authentication fails I want to re-render the login form with an error message, but without repopulating the errant name and password fields. I'm using ->render in my controller code, but the name field is populated via the input element on re-rendering. I hope someone knows how to null out text fields before re-rendering, or can point me to an example.
21:02 sri $c->param(name => undef) should do it
21:07 sugar joined #mojo
21:07 sri batman: talking about asset pipelines, just noticed that phoenix uses brunch :o http://www.phoenixframework.org/docs/static-assets
21:09 sri i must admit, i like their style a lot
21:11 sri even though, that tutorial is a mess
21:13 hesh sri: that did it. thanx
21:14 sugar_ joined #mojo
21:17 disputin joined #mojo
21:19 sri curious, the phoenix google group is about half the size of the mojolicious one
21:19 sri small community, much noise
21:24 mspo sri: the occasional hackernews story doesn't necessarily mean a lot of usage
21:24 sri they have a massive impact on rails
21:25 sri actioncable is a panic reaction, caused by phoenix
21:25 mspo well yeah
21:26 mspo elixer and phoenix were started by a rails guy, right?
21:26 sri yes
21:26 mspo also rails is as slow as it gets
21:26 sri bunch of rails guys now
21:26 mspo so no wonder
21:28 * sri is jealous that they already have a book
21:30 mspo sri: http://www.amazon.com/gp/seller-account/mm-summary-page.html?topic=200260520
21:30 mspo sri: plugin the mojo guides
21:30 mspo call it a day ;)
21:31 sri lets not -.-
21:32 mspo I have a connection with this guy: http://www.reedmedia.net/
21:34 jberger mspo: publisher isn't the problem, its content, no one has written a book yet
21:34 mspo jberger: that'll do it
21:34 mspo feels like you'd want something to go along with the book
21:34 mspo like a new major version
21:35 jberger its on my list of things to do but I'm having a hard time trying to find time to do what I can to refactor the transaction
21:35 jberger we do major versions often enough
21:35 mspo writing a book takes a while
21:35 sri if someone writes a book, i promise to schedule the next major release for it ;p
21:35 mspo learnyousomemojo
21:36 sri mojolicious developers are overemployed :(
21:36 jberger mojo the hard way
21:36 jberger actually, that title probably implies something else
21:36 sri jberger: get back to refactoring!
21:37 * sri cracks the whip
21:37 mspo the little liciouser
21:45 mad_hatter joined #mojo
21:48 patrickz joined #mojo
21:57 PryMar56 joined #mojo
22:21 PryMar56 joined #mojo
22:25 sri more postgres 9.5 hype \o/ http://www.enterprisedb.com/postgres-plus-edb-blog/jason-davis/what-s-under-radar-postgresql-95
22:28 Grinnz_ <Spock> Grinnz_: Google Fight! "postgres 9.5": 4.6 thousand; "perl 6": 63.6 thousand. Winner: "perl 6"!
22:28 Grinnz_ a somewhat surprising result :P
22:29 jberger 15 years, vs like 20 minutes
22:29 jberger normalize that out and see what you get
22:30 sri buzzkill
22:30 Grinnz_ this parallelism infrastructure seems very interesting
22:31 sri yea
22:32 * Grinnz_ will be continuing to push for migration from mysql... as long as it takes...
22:32 sri fighting the good fight
22:34 jberger Grinnz_: I introduce you to Phil21, go
22:46 sri you hear almost nothing about mongodb anymore these days
22:46 sri aside from postgres being part of mongodb now ;p
22:46 jberger that was hilarious
22:59 Phil21 lol
22:59 Phil21 make it not suck to admin please
22:59 Grinnz_ they did that in 9.0 as far as i could tell :P
23:00 Grinnz_ setting permissions in 8.3 was a nightmare
23:04 sri row level security is kinda cool
23:04 sri making all your app users actual db users is an interesting idea
23:05 lluad Not a good idea, though, unless you've got a very constrained set of users.
23:05 lluad (for that case it's nice, though)
23:12 good_news_everyon joined #mojo
23:12 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vuHfr
23:12 good_news_everyon mojo/master 3916e24 Sebastian Riedel: update jQuery to version 2.2.0
23:12 good_news_everyon left #mojo
23:29 sri jberger: i think you forgot to add Mojo::Headers examples
23:30 jberger I still have the tab open in my browser
23:30 jberger I'm like 3 deep in priority stack
23:30 sri stack overflow
23:31 jberger I'm like this close to getting my vnc over websocket thingy going at $work
23:33 ribasushi jberger: isn't this called webrtc?
23:34 jberger probably
23:34 jberger but I have to do it using novnc
23:34 jberger and I've replaced forking a websockify process per-host to a mojo tcp-over-websocket bridge I rolled myself
23:35 sugar_ joined #mojo
23:35 ribasushi jberger: I meant this: https://www.webrtc-experiment.com/screen-sharing/
23:36 sri doesn't sound like he wants p2p
23:37 jberger certainly not
23:39 sri do you remember if we found a solution for $c->{cb} in Mojo::UserAgent?
23:42 * sri feels like he wants to join a refactoring hackathon :o
23:46 jberger the general idea was that we were going to have some kind of $c->start($tx => $cb)
23:47 jberger which come to think of it is when it should subscribe to resume too
23:47 sri oh, i guess that works
23:48 sri since the callback runs before the channel gets upgraded to a websocket
23:48 sri somehow i was thinking about how a $cb could be passed around... but that's totally unnecessary
23:48 jberger its basically the same code, just in a different place
23:49 jberger which is how I've been operating this whole time
23:49 sri hmm, i guess there shouldn't be any more problems with the user agent refactoring
23:50 sri everything else seems pretty straightforward
23:50 jberger its mostly about time
23:50 jberger and me not understanding where the member data is
23:52 sri maybe we should try this a little differently
23:52 sri with a shared branch, where we all can commit to
23:58 sri naah, that would only work once the basics are done
23:59 jberger VNC!
23:59 jberger baby!

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