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

IRC log for #mojo, 2015-12-03

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

All times shown according to UTC.

Time Nick Message
00:00 nic Not Halloween and not dinosaurs.  A name that resonates with young devs and dev managers
00:00 Adura Papillion Perl sounds appropriate for 6.
00:00 Grinnz_ pumpkin perl is p5, not 6
00:00 Grinnz_ (an idea for p5, anyway)
00:02 nicomen U+1F383 I mean, what are we waiting for? ;
00:02 lsm joined #mojo
00:02 nic A name that feels comfortable to the Netflix generation.   Nitflex
00:02 nicomen purl? ;D
00:03 nicomen sorry mean purr, kittens and all that
00:04 nic that's the best I've heard so far
00:05 nic there's a small advantage if it's not a dictionary word tho
00:05 nicomen Nacre (/ˈneɪkər/ NAY-kər also /ˈnækrə/ NAK-rə), also known as mother of pearl, is an organic-inorganic composite material produced by some molluscs as an inner shell layer; it also makes up the outer coating of pearls. It is strong, resilient, and iridescent.
00:05 nic When .Net came out, it was impossible to google
00:06 lluad "pets-dot-com". It's perfect. There's already a logo. And I'm sure petsdotcom.io is available.
00:06 nicomen just drop a letter to get the pearl => perl effect ;)
00:07 nic dropping a vowel always helps.  Nacr
00:07 nic Any chemists able to tell us if NaCr has any significance?
00:08 nicomen it's nice for selling one-liners: not any carraige returns
00:09 nic and there could be built-in recursion for the not-another family of unimaginative phrases
00:21 jberger sri: this is your "you don't always see your query" poke
00:37 nic bpmedley++  # Mojo::IOLoop::Tail
00:38 nic been having fun with a websocket
00:38 nic tomorrow I'll try it with EventSource
00:38 marty Maybe Perl6 will be like bell bottom jeans.  Keep wearing them long enough and hipness is guaranteed.
00:41 jberger bpmedley: ummm the tests on that are a bit lacking
00:41 jberger bpmedley probably gets enough of that from me at work now :s
00:46 jberger I wonder if that module would tail the app's own log file
00:46 jberger I had to play a trick to make that work in a personal project (on my github, not cpan)
01:00 jontaylor joined #mojo
01:03 berov joined #mojo
01:12 asarch joined #mojo
01:16 berov1 joined #mojo
01:20 voldemortensen joined #mojo
01:21 bpmedley nic : Thanks!!  I'd love to beef up the tests.
01:32 jontaylor joined #mojo
01:47 sri lol https://atom.io/packages/activate-power-mode
01:49 sri tempire: admit it, you started this, right? http://www.slate.com/blogs/the_eye/2015/12/01/glitter_beard_is_a_2015_holiday_trend_that_has_grown_men_glitter_bombing.html
02:03 berov joined #mojo
02:07 jberger hahahhaaha
02:41 kaare joined #mojo
03:08 Zoffix joined #mojo
03:08 noganex_ joined #mojo
04:39 jontaylor joined #mojo
05:07 melo joined #mojo
05:13 melo1 joined #mojo
05:40 jontaylor joined #mojo
05:49 ribasushi joined #mojo
06:04 nnutter joined #mojo
06:12 cpan_mojo Mojo-Webqq-1.6.4 by SJDY https://metacpan.org/release/SJDY/Mojo-Webqq-1.6.4
06:32 stephen joined #mojo
06:50 dod joined #mojo
07:07 Vandal joined #mojo
07:14 dod joined #mojo
07:21 voldemortensen joined #mojo
07:37 esh joined #mojo
07:42 jontaylor joined #mojo
07:44 cpan_mojo AI-MicroStructure-0.20 by SANTEX https://metacpan.org/release/SANTEX/AI-MicroStructure-0.20
08:01 sugar joined #mojo
08:04 osfabibisi joined #mojo
08:14 trone joined #mojo
08:22 eseyman joined #mojo
08:25 kes joined #mojo
08:37 marcusr sri: I'm totally doing this for christmas.
08:39 marcusr glitter everywhere!
08:45 abra joined #mojo
08:50 berov joined #mojo
08:58 CandyAngel I really need to get a good understanding of Mojolicious so I can stop pestering with stupid questions -.-
09:08 nic ohhhhhhhhh, is there still time to grow a beard?
09:11 batman i really want to add in-memory asset support back into assetpack, but i don't know how to do it :/
09:11 batman anyone who wants to have a go are more than welcome
09:11 CandyAngel Grow a beard?
09:12 nic CandyAngel: http://irclog.perlgeek.de/mojo/2015-12-03#i_11648186
09:17 CandyAngel ... why would you do that to yourself? You'd never be rid of it1
09:17 CandyAngel !
09:18 nic that's a point
09:19 nic You have to choose: leave a trail of glitter everywhere you go....  or superglue it on
09:20 CandyAngel Looking on the bright side, you wouldn't have to do it next year, just ruffle your beard and more glitter would appear
09:20 nic :)
09:21 nic That would be impressive in meetings.  Start with a normal looking beard.  Encounter some tricky decisions, ruffle through beard...  GLITTER TIME
09:22 CandyAngel Pretty good distraction technique
09:22 CandyAngel Either people would be like "oooo glitter"
09:23 CandyAngel Or "Ewwww glitter *scramble away*"
09:48 woz joined #mojo
09:50 CandyAngel I can't even tell how I would set up a connection to this system where just one Minion worker can access it
09:50 * CandyAngel steals sri's brain
09:57 nic CandyAngel: maybe 'solo' would help:  https://josephscott.org/archives/2011/09/solo/
09:59 stephan48 CandyAngel: from your worker you access a backend system which has a connection limit?
10:00 CandyAngel That's the problem, the backend I am connecting to has a connection limit of 1. I'm not sure where to put that connection so only the worker (./myappl.pl minion worker) connects to it
10:01 CandyAngel As far as I can tell, if I just make an IOLoop::Client and authenticate $client->on(connect => sub {}) (I think), it'll try to be authed twice (by the main app, then again by Minion)
10:02 woz joined #mojo
10:05 CandyAngel Unless.. I create the client and authenticate using $minion->on(worker => sub{}) ..?
10:13 salva joined #mojo
10:14 kes joined #mojo
10:17 nic CandyAngel: I'm probably not following your problem entirely, but is it solved by having that worker be the only one that pulls from queue X?
10:17 nic that way it is the only worker attempting those jobs
10:19 nic sri: http://mojolicio.us/perldoc/Mojolicious/Guides/Cookbook#WebSocket-web-service uses ->to_abs
10:19 nic but http://mojolicio.us/perldoc/Mojolicious/Guides/Cookbook#EventSource-web-service does not
10:19 nic Is there a reason?
10:20 stephan48 CandyAngel: limit a worker to only one job and set a concurrency limit(maximal jobs to be processed at one time) to 1
10:20 nic Is it because the former is ws and the latter is get?
10:21 jontaylor joined #mojo
10:21 stephan48 CandyAngel: can't you install a event handler in your startup function to react on the worker event?
10:28 CandyAngel Let's see if I can explain this clearly
10:28 CandyAngel We have a system here, lets call it LEGSYS, which you connect to over port 23
10:29 CandyAngel You then choose which site you want (SS BRHO), then it asks for username and password (which is the same across all sites)
10:30 HtbaaPi with a quad core server with 8GB RAM that's going to handle a total of 700 video uploads, any suggestions on how to configure Hypnotoad? E.g. amount of workers etc.?
10:30 CandyAngel I can connect to BRHO and WESM concurrently, but I can't connect twice to BRHO
10:30 CandyAngel The second attempt will be refused (user already connected)
10:31 CandyAngel I'm using Minion to serialise the potentially concurrent requests for data to each site
10:32 nic HtbaaPi: 700 video uploads... at a rate of 1/week?
10:32 nic CandyAngel: Does the same worker need to be able to connect to both sites?
10:33 CandyAngel My problem is I don't know where to put the "connect to port 23 and authenticate to BRHO" so that it will only be done once for the (site specific worker) to use
10:33 HtbaaPi nic: yes pretty much
10:33 CandyAngel No, I can do some task dependency thing to "join" from different sites
10:33 CandyAngel Tasks will be like
10:33 HtbaaPi it's a one time thing only (some sort of nomination by video website)
10:34 CandyAngel BRHO_PERLIS
10:34 CandyAngel So one worker will get all BRHO_* tasks
10:34 CandyAngel (or process those queues, as they are now)
10:36 CandyAngel So each worker will only handle requests to one site
10:37 nic So one job type is in one queue attended by one worker, whose concurrency is set to one.  I'm not seeing the problem
10:37 HtbaaPi so looking at the docs I could set workers => 8 (because of 4 cores). Or if it's all handled non blocking I could up it and keep clients very low
10:37 CandyAngel nic: I don't know where to put the "connect and log into the site" bit so it is only done for the worker
10:37 nic HtbaaPi: check if you have hyperthreading (probably yes)
10:38 CandyAngel I don't want it to be done in the app and then again when the worker starts
10:38 CandyAngel Because the workers attempt will be rejected
10:38 HtbaaPi nic: server isn't up yet but I'll be able to configure it at will. Threading as in default perl threads?
10:38 nic why not only in the worker?  It sounds like that one worker is the only client to that one site
10:38 CandyAngel Only in the worker is what I want to do
10:39 CandyAngel But I don't see how I do it, unless it is in that $minion->on(worker => sub{}) event
10:39 HtbaaPi it's gonna be a Ubuntu 14.04 machine and it seems Perl is compiled with threads
10:39 nic HtbaaPi: no, don't use perl threads; in production you should always have a perl compiled without threads
10:39 nic hyperthreading usually shows up as 'ht' against the cpu info
10:39 HtbaaPi ah that kind of threading :)
10:40 nic HtbaaPi: you'll notice I've been saying 'hyperthreading' and you're the only one saying 'threading' :/
10:40 HtbaaPi nic: my bad, not much into that stuff
10:41 HtbaaPi on a similar server (just less memory) from the save provider cpuinfo does say ht is supported
10:41 HtbaaPi *same provider
10:42 nic So your 2 x cpuqty rule-of-thumb is looking good to go
10:42 CandyAngel What I am envisioning is: ./legsys.pl daemon (provides a web interface, enqueues BRHO, WESM jobs as needed), ./legsys.pl minion worker -j1 -q BRHO, ./legsys.pl minion worker -j1 -q WESM
10:43 CandyAngel And each worker connecting and authenticating when they start (not for every job)
10:44 nic Can't you put the connector in a worker attribute.  When it's time to use it, it's either connected or it's not.  If it's not....
10:45 HtbaaPi nic: that's good to know. Probably a dumb question, but with default settings how many concurrent connections does a worker handle?
10:46 nic that'll be in the docs
10:47 woz joined #mojo
10:50 CandyAngel I think so? As long as I can pass in which system it is for (queue name) to connect to, that would be fine
10:51 HtbaaPi nic: thanks. It seems that it's basically workers * clients
10:52 CandyAngel SO that would mean subclassing Minion::Worker..?
11:43 nic Has anyone used solo?  I'm wondering if it works for distributed processes.  That would be my only use-case (ie minion) but he doesn't mention it in the blurb.  If it only works localhost then I don't see the point
11:43 melo joined #mojo
11:44 woz joined #mojo
11:44 nic The EventSource example is quite a lot of fun on the commandline (fairly obvious but it took me by surprise)
11:46 woz joined #mojo
11:47 * sri wonders where his brain went
11:49 nic There were reports of it in N America this morning
11:50 nic At this rate we're going to need to customise one of those Santa Tracker websites to keep track of it
11:50 sri :o
11:55 nic btw, I know this isn't news, but hypnotoad hot-reloading is the awesome
11:56 cpan_mojo Swagger2-0.65 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Swagger2-0.65
11:56 nic our B2B dept are using their part of the app real-time, ajax-polling 4pages/sec
11:57 nic but I need to do reloads cos the Warehouse dept want rapid turnaround on their UI requests
11:57 nic by the magic of hypnotoad (and some help from morbo) those conflicting demands are all satisfied
11:58 nic (Each dept thinks I've written an app specifically for them, not realising their all sharing the same mega-app segregated into dept 'areas')
11:59 nic s/their/they're/
12:03 bpmedley joined #mojo
12:08 woz_ joined #mojo
12:11 neilhwatson joined #mojo
12:11 woz joined #mojo
12:14 kaare joined #mojo
12:21 theo joined #mojo
12:22 woz_ joined #mojo
12:58 CandyAngel nic: Reports of it in North America? Excellent :)
12:59 * CandyAngel isn't there :D
12:59 * sri wonders if the parse_frame and build_frame methods from Mojo::Transaction::WebSocket should go into Mojo::Channel::WebSocket or an entirely different package
13:00 sri we put the low level HTTP stuff in Mojo::Message/Mojo::Headers/Mojo::Content... perhaps a Mojo::WebSocket would make sense
13:00 CandyAngel He can have it back once I've transferred a complete understanding of Mojolicious from it to my brain
13:00 sri or even Mojo::WebSocketFrame
13:01 sri jberger/crab: trying to make your life easier little by little :)
13:03 sri those two methods can basically become functions
13:03 sri so it's a little odd
13:18 ajr_ joined #mojo
13:32 jontaylor joined #mojo
13:33 abra joined #mojo
13:38 vytas joined #mojo
13:42 ajr_ joined #mojo
13:49 jberger sri: it would certainly make it more analogous to the message classes
13:50 Zoffix left #mojo
13:50 jberger But that isn't the problem, it's event emination
13:50 sri yea, it's just a minor thing
13:51 * jberger spots sri's brain out the train window!
13:51 sri the change yesterday was more important though, that would have annoyed you when you discovered it https://github.com/kraih/mojo/commit/6ba2bcb7bbdfe44aadddb07415b00b999f19d729
13:52 jberger heh
13:56 ajr_ joined #mojo
14:00 sri suppose Mojo::WebSocket makes sense, considering that Mojo::HTTP2::* is most likely going to be a thing
14:17 CandyAngel SHould I be using Mojo::Transaction (or a subclass of) to talk with the remote server I am connecting to with Mojo::IOLoop::Client?
14:18 CandyAngel I'm making like.. a makeshift thing where I just use a command line to get specific requests for now
14:18 CandyAngel And then building into a Minion-backed thing later
14:30 CandyAngel Yay, progress! :D
14:35 CandyAngel Okay, so I have a worker that creates a client and starts Mojo::IOLoop. Then on 'connect' event, it sets up some callbacks on the stream (read/timeout) and writes the "hi, i'm connected" bytes and prints out a response from the server
14:35 CandyAngel Weeeeee
14:38 CandyAngel I think I need a transaction thing now for collecting the data until I have all the response..
14:41 sri hmm, looks like i'm not going to Mojo::WebSocket after all
14:41 CandyAngel Oh? How come?
14:41 sri it makes sense, but also a huge mess with our deprecation policy
14:42 sri first time our deprecation policy prevented a cleanup :/
14:45 sri if i went ahead it would make life much harder for jberger and crab, because of all the deprecation workarounds
14:46 bpmedley joined #mojo
14:46 CandyAngel Ah I see, that's a shame
14:46 sri looks like we are completely deadlocked with all new protocol work now
14:51 stephan48 mh. it sounds like it does make sense throu to make an exception to the policies for this? it needs to have enough warning time throu
14:51 irqq joined #mojo
14:52 sri already trashed my branch, don't have enough time to play by the rules :(
14:53 sri we really need fresh volunteers with more time
14:55 CandyAngel I'm trying! :P
14:55 stephan48 would it be feasible to feature freeze the current version(only allow bugfixes), do that for one for a year or two around(sort of LTS support) and in the meantime work on refactoring the internals towards your goals in the next major version?
14:55 stephan48 no clue how to make CPAN handle that properly throu
14:56 jberger Actually, perhaps there is another way
14:56 jberger These are all going to be new classes, perhaps they can be implemented in parallel
14:56 jberger Not very mojoy I'll grant you
14:56 stephan48 ah. and then some logic to switch between the behaviours?
14:57 jberger But this is an unusual case
14:57 stephan48 yea
14:57 sri that only makes it harder for new people to join the effort
14:57 sri lots of dead code to get distracted
14:58 jberger But I actually mean implement channels add entirely new code
14:58 sri the transaction classes, user agent, and daemon are now essentially frozen until the refactoring task is done
14:58 jberger So no dead code at all
14:58 sri since deprecations would add too much complexity
15:00 crux joined #mojo
15:00 sri i wanted to add proper websocket protocol abstraction with Mojo::WebSocket, but that adds like a dozen deprecations and workarounds
15:00 bpmedley kes : Did you get an answer on running Mojo apps in the debugger?
15:00 sri sad thing is, it would make the refactoring task easier without the deprecations
15:01 sri very frustrating deadlock
15:01 jberger sri: I want to take another crack at a refactor
15:01 jberger I have another idea
15:03 CandyAngel You do that.. I've had my own idea: a custom useragent
15:03 asarch joined #mojo
15:04 sri we actually started with Mojo::Client, and then added a rewrite named Mojo::UserAgent
15:07 CandyAngel So I am going to go back? :P From what I've read, I need to set up a IOLoop::Client which connects/auths, then I can have a custom UA which builds the requests with custom transactor and just put the response in a Mojo::Message
15:07 CandyAngel Which I can then do my Expect stuff against..
15:07 * sri is so annoyed with himself for not abstracting out protocols better in the first place
15:08 CandyAngel This is my current plan anyway
15:09 sri we have the perfect architecture for http/2, but a few silly mistakes i made will prevent us from taking advantage of it
15:12 sri i better take a break, having a bad day
15:13 CandyAngel :(
15:13 * CandyAngel offers sri a big hug
15:13 voldemortensen joined #mojo
15:22 ajr_ joined #mojo
15:28 sri everything went wrong after someone stole my brain
15:30 jberger sri: don't worry about http/2 too much just yet
15:30 jberger we'll get there
15:33 CandyAngel I don't know what you mean, it's right there *sneakily puts it on sri's head* Don't worry, I do that with my glasses all the time :|
15:50 kes bpmedley: No
15:50 bpmedley kes : I'd be happy to show you what I do.
15:50 bpmedley I use the "daemon" server and $DB::single = 1 lines in the code base.
15:50 marty_ joined #mojo
15:51 kes You debug locally or remote?
15:52 bpmedley I run the perl debugger on the same host as the server, so I guess locally.
15:52 bpmedley *same host as the Mojo code
15:53 sri jberger: it's not just http/2, we should have been more on the ball with spdy too
15:53 kes I debug remotely and also want the current file is opened in editor with cursor positioned at next execution line
15:53 sri jberger: just like we did with websockets back in the days
15:54 bpmedley kes : I see
15:54 kes I have read some docs about internals. And now have idea how to kill 'print' when debug code.
15:55 jberger sri: yeah
15:55 * sri was much more happy when mojolicious was bleeding edge
15:55 kes I hope the people will not use the 'print $x, $y' anymore =)
15:59 kes the prints will be too hard to use than that module
16:02 PryMar56 joined #mojo
16:02 jberger sri: I think a useful starting point will be to have build_tx logic move to the Channel
16:03 jberger that way code can be migrated into the Channel slowly
16:03 jberger Transactions then belong to the Channel by virtue of their creation
16:03 sri and Mojo::UserAgent::Transactor?
16:03 jberger same
16:04 sri don't think i agree with that
16:04 sri what about $ua->start($tx)?
16:04 jberger maybe the transactor has a Channel that it requests a transaction from
16:05 jberger but you get that transaction from the transactor
16:05 jberger $ua->build_tx delegates to the transactor as it is
16:05 jberger so delegate it one deeper to the transactor's channel
16:06 CandyAngel Hm, it might be possible for me to tag the requests I send to this server with an ID
16:06 CandyAngel Which comes back with the response..
16:06 CandyAngel If I am understanding these docs correctly
16:07 sri i think we will always be starting with a transaction in the user agent, and pass that to a transport layer, which might be channels
16:10 sri jberger: it might even make sense to turn what the user agent keeps in $self->{connections} into objects and refactor from that direction
16:11 jberger sri: that's probably an easy win
16:11 jberger I almost thought those were objects and that there was encapsulation violation going on
16:12 jberger I think I actually commented here to that effect before I noticed
16:14 sri looks like there are 6 places where $self->{connections}{$id}{tx} is accessed
16:14 sri if those points were turned into a real api, it might be a reasoable place to move all the client_* stuff to
16:17 jberger sri++
16:17 * sri just ate donuts to lift his spirit :)
16:19 sri gotta go to them gym later to make up for it though :S
16:19 * pink_mist had a traditional swedish saffron bun
16:19 gryphon joined #mojo
16:21 Lee_ joined #mojo
16:32 * jberger realizes he never had his coffee
16:43 CandyAngel Aha!
16:44 CandyAngel sri was asked about a telnet client last year
16:56 disputin joined #mojo
16:56 * sri just had an idea for a perl6 shirt... a snail representing the 6, because it's so ridiculously slow :)
16:56 CandyAngel 6 (on its side) already looks like a snail shell
16:56 CandyAngel It's fate1
16:56 CandyAngel !
16:57 CandyAngel Or it could be a snail climbing up the wall of the p5 temple
16:57 sri i like where that is going
16:58 CandyAngel ANd because there are multiple interpreters.. basically, it's a bunch of snails climbing the pillar of p5
16:58 CandyAngel .. and if you don't draw it, I'm going to :P
17:09 sri i want to make the snail derpy too... but that's hard to do with only 3 colors :S
17:16 disputin joined #mojo
17:24 sh4 joined #mojo
17:45 ajr_ joined #mojo
17:51 sri hehe, not sure where i'm going with this https://imgur.com/W3K9gWI
17:51 sri at least the snail is derpy
17:52 marcus herp derp
17:53 sri it has to work with the "5 > 6" theme somehow
17:55 disputin joined #mojo
18:09 sri the basic theme https://imgur.com/T81PLkC
18:10 sri but i need a witty one-liner to work with
18:15 Akron joined #mojo
18:35 sri you gotta look on the bright side https://i.imgur.com/9sOYCh5.png
18:42 _dave_ joined #mojo
18:56 sri slowly getting somewhere https://i.imgur.com/KtxYJvi.png
18:56 disputin joined #mojo
18:57 disputin joined #mojo
18:57 _dave_ O.o
19:05 * sri likes derpy the perl6 snail
19:05 _dave_ I miss real science
19:06 _dave_ are you implying perl6 is slow?
19:07 trone joined #mojo
19:07 sri i didn't mean to imply it, i was trying to make that absolutely obvious
19:08 _dave_ I haven't been keeping up ... too much emotion out there over these issues
19:08 _dave_ I just want facts :)
19:09 sri in all comparisons so far perl6 has been orders of magnitude slower
19:09 _dave_ wow
19:09 _dave_ but...it's compiled! ;)
19:10 sri the perl5 hate in #perl6 annoyed me, so i'm thinking about making a shirt in response :)
19:12 _dave_ I find the hate between the two camps saddening
19:18 pink_mist I really want to like perl 6
19:18 pink_mist but I am not enamoured with the p6 community
19:19 _dave_ I've never been enamoured with any community myself .. groupthink can be evil
19:19 Zoffix joined #mojo
19:20 * sri attaches a brainslug to _dave_'s head
19:20 Zoffix sri, you don't mind if I post that on Twitter, do you?
19:20 sri i hate seeing unfinished work published -.-
19:21 janus joined #mojo
19:21 Zoffix _dave_, yes, in some areas 100x slower. Not much has been optimized yet. A test parse page took 1.6 *minutes* to parse, while Mojo::DOM parsed it in like half a *second*
19:21 sri but who knows if i'll ever finish it.... so go for it
19:22 Zoffix _dave_, not to say it's impossible to make it faster, just that it hasn't been done yet :)
19:22 * sri still wants a perl6 critical shirt though
19:22 * _dave_ vanishes before sri can attach anything
19:23 _dave_ Zoffix: I see
19:23 Zoffix https://twitter.com/zoffix/status/672496350475460608
19:23 sri "gantlet"?
19:24 _dave_ that's not good for perl 6 .. but of course it hasn't been optimized fully yet (apparently)
19:24 _dave_ and people probably don't know how to write faster code in it yet
19:24 Zoffix sri, gauntlet... I typed it as guntlet and when spell checker didn't highlight gantlet, I figured I got it right :P
19:24 _dave_ so I grant a lot of slack personally ... I'm going to use both for a while and ignore the hate ;)
19:25 Zoffix _dave_, from what I heard, Larry considered this essay of a 100-Year Language... and if you read it, it's pretty critical of trying to express concept in terms of performance first and usability second: http://paulgraham.com/hundred.html
19:25 * _dave_ hands sri the red pill, which he clearly hasn't taken yet
19:27 Zoffix sri, BTW, I worked more with Bailador, and it's pretty unusable IMO at the moment. Pressing F5 rapidly actually makes it exit. I had to hack it to handle what I presume was an update to the PSGI module. And as far as features go, it's a half-done version of Mojolicious's templates and you have to define all routes yourself (even for static files), which is pretty annoying
19:27 Zoffix So as far as comparing P6 and P5 web frameworks... it's not even possible to do so yet, as P6 doesn't have an obvious contender.
19:28 Zoffix (Bailador == P6 slacky "port" of P5's Dancer)
19:28 disputin joined #mojo
19:29 _dave_ Zoffix: I remember reading that some time ago ... it's a sane concept to put usability first, but you can't ignore performance either. As usual, the balance point needs to be found.
19:30 _dave_ Zoffix: No one seems to write about why perl5 has lasted so long
19:31 Zoffix _dave_, sure. And current P6 FAQ entry reads that it uses a modern compiler, so optimization is certainly possible. We'll see how much of that statement is over-optimizm and how much is reality.
19:31 _dave_ I remember that the perl5 community was driving away people too as far back as 1996
19:32 _dave_ most people incorrectly left perl5 instead of just ignoring the community
19:32 Zoffix ehehe. I remember merlin foaming from the mouth in Freenode's #perl any time someone made a typo lol :P
19:33 * sri is pretty sure perl6 has no chance to reach critical mass
19:33 _dave_ heh I deliberately failed an interview with him
19:33 _dave_ sri: that's what I said about java back in the day
19:33 Zoffix heh
19:33 sri java has a niche though
19:34 _dave_ minecraft ;)
19:34 sri boring enterprise stuff
19:34 Grinnz_ :|
19:34 _dave_ what, you don't like minecraft?
19:34 sri all the languages that have been successful recently had a niche
19:34 Grinnz_ p6's only chance of a niche is probably somewhere in its concurrency or its grammar support
19:34 sri look at go and swift
19:35 Zoffix sri, are there languages that could produce a new "language" non-computer people could use? I'm wondering what's available in research areas where you could essentially write all the math the way you'd normally do and it'd be a computer program.
19:35 _dave_ think also about language dilution ...there's only so many developers
19:35 Grinnz_ Zoffix: thats sort of what matlab is i guess...
19:35 _dave_ too many languages = not enough critical mass for any of them
19:35 Zoffix Grinnz_, ah
19:35 sri Zoffix: swift has a huge chance to go more mainstream with its playground stuff
19:36 sri coincidentally, swift went open source today :)
19:36 Grinnz_ o rly
19:36 Zoffix Which means Apple doesn't care about it anymore :P
19:36 _dave_ See, that's why you need the red pill sri. I claim "mainstream" is irrelevent.
19:36 _dave_ er irrelevant
19:36 sri https://github.com/apple/swift
19:36 _dave_ I wouldn't care if Mojo lost all it's mindshare for example, I'd still use it.
19:37 sri they really went all out with this one
19:38 sri i generally don't trust apple with open source one bit, but swift is from the clang team after all
19:38 _dave_ Apple may realize that swift won't go mainstream as long as Apple has control of it ;)
19:39 Zoffix left #mojo
19:41 sri Grinnz_: grammars is not much of a niche, and for concurrency there's better alternatives
19:42 Grinnz_ thats my assessment as well
19:42 sri like, there is just no way perl6 could ever compete with elixir when it comes to concurrency
19:42 Grinnz_ without performance, concurrency is not very inviting
19:46 disputin joined #mojo
19:58 kes joined #mojo
19:59 berov1 joined #mojo
20:09 sugar_ joined #mojo
20:20 marty joined #mojo
20:28 irqq joined #mojo
20:28 disputin joined #mojo
20:39 ajr_ joined #mojo
20:42 Trelane awesome
20:43 Trelane I just got to pr removal of a bunch of code that worked around the noisy re-registation of plugins
20:43 Trelane since mateu upgraded our mojolicious yesterday
20:44 Trelane s/plugins/helpers in plugins/
20:44 jberger I still personally worry that that is going to bite users :s
20:48 Trelane jberger: users that aren't me ;-)
21:01 disputin joined #mojo
21:07 bpmedley_ joined #mojo
21:24 ajr_ joined #mojo
21:39 sri public beta https://letsencrypt.org/2015/12/03/entering-public-beta.html
21:45 sri the 90 day cert lifetime is going to be fun, we absolutely need a plugin that can handle auto renewal
21:46 pink_mist should that really be the work of a plugin? shouldn't the letsencrypt clients do that?
21:46 sri you need to script a cron job then anyway
21:47 jberger that isn't going to be long-term is it?
21:47 pink_mist I think that's part of their plan
21:47 pink_mist yes it is jberger
21:47 pink_mist jberger: they specifically want to force you to haveto automate it
21:47 sri 90 day expiry is long-term
21:47 jberger hunh
21:47 sri they want to encourage automation with that
21:47 pink_mist (and they're meant to make automating it very easy)
21:48 pink_mist (but it's still beta, so dunno if it is easy yet)
21:48 jb360 joined #mojo
21:51 * sri is going to miss the times when tls used to mean something
21:51 sri of course we don't know how compromised the ca system already is
21:52 * sri bets it's only a matter of days before letsencrypt will get their first national security letter
21:57 sri people are starting to realize that https everywhere might have been a bad idea https://lists.w3.org/Archives/Public/ietf-http-wg/2015OctDec/0314.html
22:12 ajr_ joined #mojo
22:16 sugar_ joined #mojo
22:19 _dave_ the system crackers will attack the renewal cron job ;)
22:19 disputin joined #mojo
22:20 ajr_ joined #mojo
22:24 sri jberger: will you try the $self->{connections} path for refactoring?
22:24 jberger sure
22:24 sri \o/
22:24 jberger when I get a chance
22:25 jberger which hopefully is soon
22:25 sri i'm pretty sure now that it can work
22:25 jberger the thanksgiving break was anything but relaxing
22:25 jberger I was actually lucky to get as much time as I did
22:26 jberger but now without all the family in town all I have to deal with is a leaking shower :s
22:28 neilhwatson joined #mojo
22:39 asarch joined #mojo
22:57 ajr_ joined #mojo
22:57 cpan_mojo Paws-0.20 by JLMARTIN https://metacpan.org/release/JLMARTIN/Paws-0.20
23:00 orev curious if anyone has a good example of how best to use database connections?  anything like connection pooling, etc...?
23:11 disputin joined #mojo
23:14 ajr_ joined #mojo
23:14 jantore joined #mojo
23:16 sri http://mojolicio.us/perldoc#SPIN-OFFS
23:17 sri https://github.com/kraih/mojo-pg/tree/master/examples
23:23 disputin joined #mojo

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