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

IRC log for #mojo, 2015-06-05

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

All times shown according to UTC.

Time Nick Message
00:05 frederic_ joined #mojo
00:11 lluad joined #mojo
00:22 zivester joined #mojo
00:27 bwf joined #mojo
00:30 woz joined #mojo
00:44 zivester joined #mojo
00:48 bwf joined #mojo
00:57 zivester joined #mojo
01:05 bwf joined #mojo
01:14 jberger jabberwok: showusthecode
01:17 dave the classic geek memetic
01:21 jberger Soooo time for that libuv reactor: https://metacpan.org/pod/release/MLEHMANN/Canary-Stability-2001/Stability.pm
01:33 klapperl_ joined #mojo
02:09 mattastrophe joined #mojo
02:13 genio joined #mojo
02:20 woz joined #mojo
02:36 noganex_ joined #mojo
02:38 hshong joined #mojo
02:48 bwf joined #mojo
03:09 bpmedley http://bmedley.org/mojo.pdf <— It’s fugly and light on content; what do people think of the direction and content?
03:10 bpmedley https://github.com/brianmed/mojo_book <— Code for the book
03:11 bin_005 joined #mojo
03:13 bpmedley Probably not the best time to ask.. I’ll read the irc log for any responses..
03:23 basic6 joined #mojo
04:01 Flying_Squirrel joined #mojo
04:54 vmbrasseur joined #mojo
05:21 kaare joined #mojo
05:45 dod joined #mojo
05:52 dod joined #mojo
06:03 batman jberger: yeah, submodules are quite confusing to me as well. subtrees .... not so much :)
06:04 batman https://medium.com/@porteneuve/mastering-git-subtrees-943d29a798ec
06:05 batman i don't use subtrees (/modules) so often, but it's very nice when i want to include a css framework in vendor/foo
06:14 bakanov_da hi. Can you answer, please, guys, does IOLoop treats mojo::useragent connections as normal connections, when it performs graceful stop?
06:26 dod joined #mojo
06:30 batman bakanov_da: i'm not sure if i can answer, but i'm guessing that you have a problem that makes you ask that question.
06:31 batman if that's the case, then state your actual problem. your question might only be a symptom.
07:01 AndrewIsh joined #mojo
07:18 trone joined #mojo
07:34 eseyman joined #mojo
07:58 bakanov_da @batman it's not actual problem... Now I send messages by mojo::useragent from my mojolicious app. And I worry, if I will be relaunch workers with graceful stop, sometime I can lose message, that I send in the time of worker stop
08:00 bakanov_da @batman I have tested and I see, that process of sending message with mojo::useragent (in the async mode) is group of events. And, stop of the worker may happen in the middle of these events
08:03 bakanov_da @batman ... I mean between these events. Not in the event
08:06 batman bakanov_da: you will loose messages on graceful stop, unless you finish before graceful_timeout
08:06 batman i think the default is 15 or 20 seconds... can't remember
08:07 batman if you mean graceful stop in hypnotoad..?
08:07 bakanov_da @batman ok, I understand
08:09 bakanov_da @batman I forgot, that worker will not receive new connections in the time of gracefult timeout
08:10 batman no, it won't.
08:12 absolut_todd joined #mojo
08:12 damaya_ joined #mojo
08:12 damaya__ joined #mojo
08:19 Vandal joined #mojo
08:27 kaare joined #mojo
08:33 woz joined #mojo
08:36 fhelmber_ joined #mojo
08:55 dod joined #mojo
08:56 ans А фейсбук апи юзает кто нибудь?
08:56 ans wow sorry )
08:56 nicomen kak delak?
09:49 Snelius ans: vk only
09:57 ans nicomen:
09:57 ans Horosho
09:58 Snelius ans: what's problem fb api application about ?
10:00 ans The problem is about friends or user. Now you can`t get this list
10:00 ans )))
10:01 ans There are tons of tears all over internet about it.
10:02 ans So i have no questions now about it anymore ). You just can`t get the full friend list through fb Graph API Now
10:04 ans "The whole point of these changes( means the new api version) is to expose less data and give users more privacy. developers.facebook.com/docs/apps/changelog: “Friend list now only returns friends who also use your app: The list of friends returned via the /me/friends endpoint is now limited to the list of friends that have authorized your app.”
10:31 mattastrophe joined #mojo
10:35 illywhacker joined #mojo
10:38 illywhacker joined #mojo
11:02 batman i can see how that's annoying :)
11:02 batman but it's also very cool that fb made the change
11:04 ans Many applications are dead now because of that
11:05 ans But yes - facebook can do anything with it`s API. )
11:06 ans Tomorrow  they can said - "Ok, bye bye, we close our api now"
11:08 dod joined #mojo
11:54 ZoffixWork joined #mojo
12:15 ZoffixWork bpmedley, RE: http://bmedley.org/mojo.pdf  The immediate concern is all the hyphens in shell commands in the book are actually UTF8 minus signs, so it makes it impossible to copy-paste them (I'm assuming the book *will* be available in digital format) and copy/pasting code examples leaves spaces after nearly every character. You also keep using full paths to morbo, et. al. and I don't think that's necessary and should be avoided f
12:15 ZoffixWork or clarity. <html>, <head>, and <body> are all optional tags. Perhaps, it will be clearer if they are omitted in the code examples (at least in Hello World example) to simplify overall snippet. In fact, I think the "Getting Started" section on homepage of http://mojolicio.us/ is the perfect example to use: it's just 3 lines. To the uninitiated, it'll likely be confusing why you first use Lite and the full apps. I think there shoul
12:15 ZoffixWork d be a section explaining the purpose of Lite apps. There should be a section (say, section 0.7.0) explaining the MVC concept. People unfamiliar with it with get confused right when you start talking about Controllers. The code example after section 0.7.2 looks really messy. I think it should be written with 70-col, or so, max line width so there would be no wraps. I also don't think this form for dereference is the common one: $$
12:15 ZoffixWork site_config{hypnotoad_ip}
12:16 ZoffixWork Sorry for the noise... Anyway, that's my two cents.
12:29 Tiger joined #mojo
12:29 ToApolytoXaos joined #mojo
12:33 neilhwatson joined #mojo
12:42 ajr_ joined #mojo
12:51 marcus Getting a lot of calendar notifications for a certain non-event today. :-/
12:57 mattastrophe joined #mojo
12:58 batman marcus: :(
13:01 zivester joined #mojo
13:22 dave joined #mojo
13:28 Adura Yeah, screw World Environment Day.
13:46 davido__ joined #mojo
13:47 fhelmber_ joined #mojo
13:55 punter joined #mojo
14:21 mishanti1 How long do you guys spend compiling perl? 1m3s seems a bit excessive. :-/ Are there any tricks to get compiletime down?
14:24 nicomen I never compile perl
14:24 nicomen just use binaries from the dists I use
14:25 batman mishanti1: is that a long time..?
14:25 pink_mist on my laptop, 1m3s doesn't seem excessive at all
14:25 pink_mist but if I were on a 32-core 512GB ram machine, I'd consider it slow too
14:27 batman mishanti1: do you do it manully or use something like plenv/perlbrew?
14:27 batman i was thinking you could skip tests maybe...
14:37 jberger mishanti1: how many build jobs are you using?
14:38 Grinnz_ 1m3s doesnt seem that long to me either
14:38 Grinnz_ i compiled it with both perlbrew and perl-build recently and it took longer than a minute i'm sure
14:38 pink_mist on my laptop it probably takes upwards of 10 minutes or so
14:39 jberger also, you can save a bunch of time by not running the tests
14:40 jberger if that is your way about things
14:40 pink_mist well yeah, I did run the tests, because that's how I roll
14:40 genio serial tester!
14:40 pink_mist and honestly, 10 minutes isn't too bad anyway
14:40 Grinnz_ it's not like you have to build a new perl that often
14:41 pink_mist well, I was building all the perls I could get my hands on :P so I have about 40 different perls on that laptop now
14:41 McA joined #mojo
14:41 pink_mist maybe 25
14:41 Grinnz_ lol
14:41 Grinnz_ i just have 1 for each major release on my server
14:41 Grinnz_ so thats like 8
14:45 genio That's too many Perls to keep up with.  System+latest
14:45 Grinnz_ well thats what perlbrew is good for :P
14:46 Grinnz_ i only ever use the latest one unless i need to test something
14:46 Grinnz_ perlbrew exec++
14:47 pink_mist (perlbrew exec)++ #itym
14:47 zackiv31 joined #mojo
14:48 Grinnz_ lol
14:56 punter joined #mojo
15:10 mishanti1 Thanks for the feedback. :) Didn't think about the --notest option. Helped quite a bit.
15:11 mishanti1 jberger: Running with 8 jobs
15:11 jberger how many cores do you have?
15:11 mishanti1 8
15:11 jberger then you are about at your fastest I think
15:11 mishanti1 Methinks so too.
15:12 mishanti1 I just remember back in the day when I did c++ w/ Qt how insanely fast those compiles could be done with the right make-flags.
15:12 jberger you could try with a few more jobs, but the time will start becoming constant with added jobs
15:12 batman mishanti1: how fast when you skipped tests?
15:12 pink_mist some people claim one should use cores+1 for number of jobs ... but it depends on what cpu scheduler you're using
15:13 mishanti1 batman: right around 23s
15:14 mishanti1 batman: Small disclaimer though: this was measured with `time perlbrew install yadiyadi..`, and not with any form of proper timing-setup.
15:15 mishanti1 But yeah, it seems like these day one should really put some effort into choosing the right scheduler for the kind of operations you'll be doing.
15:20 batman mishanti1: cool
15:24 Ptolemarch joined #mojo
15:42 dod joined #mojo
15:56 Ptolemarch joined #mojo
16:48 Ptolemarch joined #mojo
16:48 ans joined #mojo
16:53 absolut__ joined #mojo
17:17 * tempire is going to have to look into ECONNRESET
17:19 jberger tempire++
17:19 jberger I've been meaning to do that too, only I have been putting all my $freetime towards YAPC::NA talk prep
17:19 jberger which has involved releasing a surprising number of modules too
17:19 batman tempire: that's cool. i wouldn't dare :(
17:25 lluad joined #mojo
17:28 punter joined #mojo
17:30 batman marcus: i'm so glad you told me about riot.js.
17:30 batman it's so smooth.
17:37 sh4 joined #mojo
17:48 Ptolemarch joined #mojo
18:04 amon joined #mojo
18:06 ajr_ joined #mojo
18:18 disputin joined #mojo
18:37 sri the cost of voting against the use of new perl features in the documentation https://groups.google.com/forum/#!topic/mojolicious/FG4E4Atdsf0
18:41 mst the cost of nobody writing the example with Sub::Current and a mention __SUB__ also being available :)
18:44 jberger I don't think I ever voted against using features of higher versions of perl, just not experimental features thereof
18:44 jberger I would definitely support examples that use __SUB__
18:45 jberger and if they can't get 5.16 (I think?) then either we could mention the possible fallbacks or encourage them to upgrade
18:46 jberger I just dont want to demonstrate things that might break that aren't under our own deprecation policy, which would include core experimental features
18:46 jberger __SUB__ is here to stay I hope
18:46 dod joined #mojo
19:00 sh4 joined #mojo
19:03 bin_005 joined #mojo
19:15 mst batman: seems to be an attempt at React::Tiny
19:15 mst I wonder how much you'll end up reinventing over time
19:30 Grinnz_ https://github.com/Grinnz/Mojo-Reactor-UV havent gotten the tests to pass yet, and UV is super outdated, but it's almost workable
19:31 Grinnz_ the libuv in UV, that is
19:35 bin_005 joined #mojo
19:37 berov joined #mojo
19:37 sri Grinnz_: i think UV still has segfault problems with threads
19:37 sri it needs the same fixes as Net::DNS::Native
19:38 sri you know, the linking stuff
19:38 Grinnz_ hmm
19:38 sri (i guess it needs a lot of updates)
19:38 Grinnz_ yeah
19:39 Grinnz_ ive opened a couple issues on the GH but i wouldnt have any idea how to fix that stuff myself
19:40 cpan_mojo Mojolicious-Plugin-JIPConf-0.01 by VLADIMIRZ https://metacpan.org/release/VLADIMIRZ/Mojolicious-Plugin-JIPConf-0.01
19:41 kaare_ joined #mojo
19:41 Grinnz_ its interesting all the functions that are directly implemented in libuv now (most of it not in the perl UV)
19:41 Grinnz_ https://github.com/thlorenz/libuv-dox/blob/master/methods.md
19:42 Grinnz_ ip address stuff, filesystem stuff
19:42 Grinnz_ i can see why it would be nice for node
19:51 Repaster joined #mojo
19:55 stephan48 Grinnz_: hey, which POE<->Mojo Loop thing should i use if i want to run them together?
19:55 Grinnz_ probably Mojo::Reactor::POE
19:55 ans joined #mojo
20:03 kaare__ joined #mojo
20:10 disputin joined #mojo
20:20 PryMar56 joined #mojo
20:37 Grinnz_ woo, tests passing
20:59 da5id joined #mojo
21:54 woz joined #mojo
21:56 Kogurr joined #mojo
21:57 woz joined #mojo
22:12 cpan_mojo Mojo-Reactor-UV-0.001 by DBOOK https://metacpan.org/release/DBOOK/Mojo-Reactor-UV-0.001
22:12 jberger :o
22:12 jberger Grinnz_ += 3
22:13 Grinnz_ it probably has some issues, but it seems to "work" :P
22:13 Grinnz_ from the dumb-benchmark it's oddly faster than EV on my centos box but slower than either EV or poll on my home server
22:16 woz joined #mojo
22:24 Grinnz_ sri: from all indications, libuv is claiming to be single-threaded still
22:24 Grinnz_ at least, without the thread methods that aren't implemented in the perl module
22:25 Grinnz_ or the filesystem methods
22:30 ajr_ joined #mojo
23:33 Grinnz stephan48, if you are planning to use Mojo::Reactor::POE, i'm about to push a change so it will be more consistent with other reactors; the reactor won't keep running if it doesn't have anything to watch regardless of other sessions, but it won't destroy existing sessions when it's stopped
23:34 stephan48 okey
23:38 Grinnz but to answer your question better, either Mojo::Reactor::POE or POE::Loop::Mojo_IOLoop depending which interface you want to use to start the loop
23:41 woz joined #mojo
23:42 stephan48 thanks

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