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

IRC log for #mojo, 2015-02-05

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

All times shown according to UTC.

Time Nick Message
00:08 disputin joined #mojo
00:12 jberger gosh darnit, I'm loving doing mojo every day
00:12 jberger sri++
00:12 jberger (mojo core team)++
00:12 jberger (mojo users)++
00:13 jberger https://www.youtube.com/watch?v=StTqXEQ2l-Y
00:20 nathanael joined #mojo
00:21 Nemix agreed...I have barely scratched the surface and have been impressed at how much it makes easy to accomplish....and I knew nothing about web sockets before starting to use it
00:22 Nemix spending all my time learning html and javascript :)
00:23 Gedge joined #mojo
00:26 jberger luckily I work with a front-end dev
00:28 Nemix I am hoping my friend will pick up the UI soon once he gets things running :)
00:34 Grinnz (luckily i work with 4 front-end devs)
00:35 disputin joined #mojo
00:48 marty joined #mojo
00:49 marty joined #mojo
00:52 Gedge joined #mojo
00:54 pink_mist reneeb++
00:54 asarch joined #mojo
00:56 Grinnz does anyone know what system he runs those tests on? cause Mojo-Reactor-POE failed all of them so it's probably BSD or win32 lol
00:57 Grinnz actually no, i'm thinking of POE-Loop-Mojo_IOLoop... i don't know what would cause Mojo-Reactor-POE to fail
00:59 Grinnz haven't gotten any fails for it on cpantesters :/
01:01 pink_mist my guess would be that it breaks compatibility somewhere between Mojolicious 5.whateverversionitclaimstoneed and 5.72
01:02 pink_mist if things changed /after/ it was released, cpan testers would likely take a long time to notice
01:02 pink_mist if they ever did
01:03 pink_mist that said, I guess I'll run it through my cpan testers setup
01:03 Grinnz it requires 5.70, lol
01:03 pink_mist oh :P
01:13 Gedge joined #mojo
01:13 pink_mist alright, first try with 5.71 = PASS
01:14 asarch joined #mojo
01:16 preaction are you using git bisect? it can automate finding which commit changed something
01:16 Grinnz i havent done anything, it installs for me
01:16 pink_mist let me just confirm that things actually do break on latest Mojolicious :P
01:16 Grinnz or at least, it has when i last tried, and there hasnt been any version bumps since that i know of
01:20 pink_mist I notice that POE asks me to run network tests, with the default being to run them -- I'm not sure how cpanm handles that, but it's plausible it might fail because it's not allowed to run network tests, and that would possibly cause the entire build to fail meaning we get nok as result
01:21 Grinnz well yes the default is to run them
01:21 Grinnz are you saying he downloads the tarball then turns off his network?
01:21 pink_mist yes
01:21 Grinnz ...
01:21 Grinnz why?
01:22 pink_mist seems more secure to me
01:22 pink_mist anyway it's just a theory
01:22 Grinnz not when everything is network related
01:22 Grinnz also that would make it kind of hard to use cpanm-reporter :P
01:23 pink_mist true :P
01:23 pink_mist but he could turn the network on again afterwards =)
01:27 pink_mist well, it works on 5.77 for me
01:30 pink_mist hmm, actually, let me just double-check something here...
01:34 pink_mist alright, I'm installing to a local::lib here ... and one of POE's prereqs doesn't seem to install cleanly into the local::lib .... it seems POE can't find it if you install it to a local::lib at least ... no clue why
01:34 pink_mist if I install that without local::lib, POE installs fine into a local::lib :/
01:35 pink_mist IO::Tty is the culprit
01:35 Grinnz what prereq?
01:35 purl prereq is just a hash or a test module that doesn't pass its own tests
01:35 Grinnz hmm.. doesnt ring a bell
01:36 pink_mist I can post build log if you want
01:38 Grinnz sure
01:38 pink_mist http://p.0au.de/1b15fb66/
01:39 pink_mist local::lib is set to /tmp/8rBoOOeydi
01:39 preaction is that what that is? i had trouble installing POE on Cygwin because of IO::Tty
01:40 preaction but last time i tried was a few years ago
01:40 Grinnz oh this is with cpan not cpanm?
01:40 pink_mist Grinnz: yes
01:40 pink_mist Grinnz: since I want to send test reports automatically =)
01:41 Grinnz i don't see where the failure is
01:41 Grinnz Tests succeeded but one dependency not OK (IO::Tty)
01:41 Grinnz what does that even mean?
01:42 Grinnz IO::Tty installation looks fine
01:42 pink_mist it means it couldn't find IO::Tty :/
01:42 pink_mist or something
01:43 Grinnz i just checked and i have IO::Tty installed in local::lib on both my systems
01:43 pink_mist huh
01:43 Grinnz from cpanm, though
01:43 Grinnz oh i see it now, Warning: Prerequisite 'POE => 1.366' for 'DBOOK/Mojo-Reactor-POE-0.005.tar.gz' failed when processing 'RCAPUTO/POE-1.366.tar.gz' with 'make_test => NO one dependency not OK (IO::Tty)'.
01:44 Grinnz wait thats for the next module lol
01:44 Grinnz it doesnt say that for POE
01:44 pink_mist yeah
01:44 Grinnz so i'm just confused
01:45 pink_mist it never installs POE because of it though
01:46 Grinnz what kind of system, linux?
01:46 pink_mist yeah
01:46 Grinnz could you try with cpanm?
01:47 Grinnz it looks like its just a regular runtime dependency...
01:47 Grinnz hasnt been updated lately
01:49 Grinnz one difference, your perl isn't threaded
01:49 Grinnz but neither are half of cpantesters
01:49 pink_mist I had the same thing on a threaded perl
01:49 Grinnz i guess it might not make it to cpantesters if it can't install POE
01:50 pink_mist indeed
01:50 Grinnz and your log even sends a pass report for POE, lol
01:50 Grinnz what the fuck
01:52 pink_mist cpanm -l ~/local Mojo::Reactor::POE <-- yeah, that did work ... huh
01:53 Grinnz i don't even... like, what the hell is determining IO::Tty is not ok -after- the tests?
01:54 pink_mist anyway, it's 3am here ... I need to get up early tomorrow :/
01:54 pink_mist good night
01:54 Grinnz anyway he's using cpanm so who knows if its even related
01:54 pink_mist yeah
01:55 Grinnz later, thanks for testing
01:55 good_news_everyon joined #mojo
01:55 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/bBad
01:55 good_news_everyon mojo/master 664adb6 Sebastian Riedel: center menu
01:55 good_news_everyon left #mojo
01:57 good_news_everyon joined #mojo
01:57 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/bBVc
01:57 good_news_everyon mojo/master 06bc223 Sebastian Riedel: mention changes to templates
01:57 good_news_everyon left #mojo
02:04 disputin joined #mojo
02:06 klapperl_ joined #mojo
02:31 Gedge joined #mojo
02:40 sri tempire: btw. think i've fixed the typography, kerning as well as rendering
02:41 sri Sketch native png export does not do anti aliasing so well
02:43 sri making something that doesn't look like an ad is actually not easy
02:51 brainbuz joined #mojo
03:27 noganex joined #mojo
04:29 inokenty-w joined #mojo
04:51 hshong joined #mojo
04:53 rem_lex joined #mojo
05:41 dotandimet joined #mojo
06:23 Gedge_ joined #mojo
06:23 davido__ joined #mojo
06:52 dod joined #mojo
06:58 dod joined #mojo
07:03 marcusr good morning
07:03 purl For you maybe.
07:06 dotandimet joined #mojo
07:10 marcusr purl: COFFE OR DEATH!
07:10 purl marcusr: i'm not following you...
07:10 marcusr purl: COFFEE OR DEATH!
07:10 purl marcusr: huh?
07:11 Snelius_ purl: http://ikle.ru/irc/ugt.html
07:11 purl Snelius_: huh?
07:11 marcusr https://www.youtube.com/watch?v=To2gpt9sDO8
07:20 dod joined #mojo
07:24 Vandal joined #mojo
07:32 reneeb joined #mojo
07:57 eseyman joined #mojo
08:03 reneeb joined #mojo
08:26 trone joined #mojo
08:36 nathanael joined #mojo
09:17 amon joined #mojo
09:21 fhelmber_ joined #mojo
09:37 irq joined #mojo
10:13 dod joined #mojo
10:13 btyler_ discovered the joy of delays and Mojo::Pg last night. promptly ditched all the nested callback garbage I'd been doing...absolutely lovely
10:28 Dandre joined #mojo
10:42 meshl joined #mojo
11:37 sri hahahahaha https://groups.google.com/d/msg/mojolicious/FLKpY68ctaI/iQBhnfSLm7kJ
11:42 marcusr never putting autoload back in
11:42 nicomen btyler_: you wouldn't be in the mood for writing a small tutorial, I'd love to spread around to Pg lovers
11:48 btyler_ nicomen: there are actually great examples of using delay + Mojo::Pg right in the Mojo::Pg doc
11:55 btyler_ the one thing I'm not totally clear on is what happens when you fire off multiple things at once, as in the Mojo::IOLoop::Delay synopsis. does the third step always trigger after the longest-running callback/with that function's arguments to the callback?
11:55 btyler_ and the ones that run more quickly (such as the one-second timer) are basically called in async "void" context?
12:01 sri btyler_: http://mojolicio.us/perldoc/Mojolicious/Guides/Cookbook#Synchronizing-events
12:03 btyler_ ...that is awesome
12:04 btyler_ apologies for not searching the doc harder. that is so much more excellent than I had thought
12:26 sri there is actually also an example http://mojolicio.us/perldoc/Mojo/IOLoop/Delay#begin
12:35 Nemix so in the examples that show an method retruning an instance of the object it is part of (for example "$delay = $delay->data({foo => 'bar'})") that allows for chaining?
12:36 Nemix or maybe it is not doing what I think it is
12:37 sri that sentence sounds odd, but not wrong
12:38 Nemix heh
12:42 Nemix what throws me off is that as examples are written it is not clear to me (probably due to my limited knowledge) why one needs a return value that contains the thing you already have.
12:47 asarch joined #mojo
12:50 neilhwatson joined #mojo
12:52 michael the examples and docs in general are very terse, which seems at odds with the intended accessibility of M::Lite etc
13:08 jberger btyler_: have you read my blog series?
13:09 btyler_ jberger: I did, a while ago, then didn't make use of the information in sufficient time to turn it into learning
13:21 sri michael: what information is missing?
13:23 michael nothing's missing, but sometimes seems very short, even on key concepts. i understand you don't want duplicate documentation that can get out of sync, but sometimes i think DRY is taken to an extreme in documentation
13:24 jberger michael: I'm not sure why it is, but people still come here with ... critiques ... of the documentation
13:25 jberger I'm curious, which project does it better?
13:25 sri michael: "too terse" is not constructive criticism
13:25 sri if you want a book, say that you want a book
13:25 jberger we are always open to improvement, but I'm not sure what to improve
13:26 * jberger still would love to write a book
13:27 * jberger wishes he had more time
13:27 btyler_ I think the mojo + friends docs are pretty much the most perfect man pages I've ever encountered. it's just that man pages are a tricky way to learn things from scratch
13:27 btyler_ so, indeed, a book! :)
13:28 Jonis a book would be excellent
13:28 jberger the only other thing I can think of, since this seems to come up every now and again, is a community sourced book
13:29 jberger something of a compendium
13:29 michael not sure it needs to be a book, but just more tying of concepts together. Examples of real-world use-cases
13:29 jberger each chapter with a different author
13:30 michael often examples/tutorials are given for ::Lite but then not obvious how to translate into the full app.
13:30 jberger michael: real world use cases take a ton of space, especially when you get in to non-blocking architecture
13:30 jberger michael: we devote more space to that than any other concept
13:31 jberger that's what the Growing guide is
13:31 sri if by the end of this guide you're not able to translate lite concepts to full apps, that is something that needs to be fixed http://mojolicio.us/perldoc/Mojolicious/Guides/Growing
13:31 michael i meant an example of "here's how to use Log4perl with Mojolicious", or "here's how to write a json api"
13:31 sri that is out of scope
13:32 michael of the docs, that was in response to jberger's suggestion
13:32 jberger michael: those are definitely useful things, but they aren't really Mojo doc
13:32 jberger that is where books or blog posts take over
13:32 michael that was in response to the book
13:40 sri it's frustrating that we can't find someone to write a book, even with the amount of interest
13:43 sri and so much potential for selling new editions with all the new features we add
13:44 meshl joined #mojo
13:46 ignacio_ joined #mojo
13:50 basiliscos joined #mojo
13:55 jberger for me its about time
13:55 jberger I would LOVE to
13:56 jberger who knows, maybe one of these days
13:56 jberger but I really want to get my phantom tester and this minion monitor out too (cause I want to use them)
13:56 sri and at the same time you keep blogging ;p
13:56 reneeb I would help to write a book...
13:57 sri you know, there are many ways to write a book ;p
13:57 jberger its been a while, have you noticed?
13:57 sri the best web dev book i've seen was actually written as a series of blog posts
13:57 jberger I would gladly donate my blog posts to a compendium book
13:57 sri http://flaskbook.com/
13:59 sri and flask is not even particularly good :)
14:00 Insane why don't you want to use github as book hosting and dev environment?)
14:00 jberger Insane: by start of a mojo book is on github
14:00 jberger its about 6 pages
14:00 Insane wow ;)
14:01 sri re community book, has that ever worked well?
14:01 bwf https://www.gitbook.com/ ?
14:01 sri all the books i like had professional editing
14:01 jberger sri: its more common in scientific communities
14:02 jberger and not for textbooks but for emerging research fields
14:02 jberger each author gets a topic/chapter
14:02 jberger and the editor puts it together
14:03 mst that's how we did it for catalyst
14:03 sri ah, so there's a real editor
14:03 mst but, yeah, you still need a real editor
14:03 jberger there are two
14:03 jberger technical editor and publishing editor
14:03 jberger you are thinking of publishing editor
14:03 jberger which there still should be one of those, yes
14:04 mst an outside technical editor is really helpful too
14:04 jberger sure, we'll take one of those too
14:04 mst we got given Jacinta
14:04 mst every time she sent me feedback on my work I read the email, decided I hated her, and didn't do anything else useful that day
14:04 mst then the following day I thought it through properly and made at least 75% of the changes she's suggested
14:05 jberger mst: that really is how it works
14:05 sri i could see that work for a mojolicious book
14:05 jberger I had that experience with every scientific publication I made, when the reviews would come back
14:06 jberger my first response was always "that guy didn't get it at all!" which it turns out is valid criticism, I didn't explain myself, try again
14:16 btyler mst/jberger: is there a place where more-or-less level headed people can talk in a reasonable way about how to avoid perl-world-burning naming drama? #mojo isn't the place
14:16 jberger I did start a little channel for it actually
14:17 mst I wonder if we need a generic channel for the various discussions though
14:17 jberger I wasn't sure if I wanted it public or not, but since you asked
14:17 btyler (context: jberger and I have been discussing this for a bit in private)
14:18 jberger I have just introduced a new topic in #pumpkin
14:18 btyler yeah, I dunno. just feels like wheel-spinning to lob cents into a discussion on #perl6 or here
14:18 btyler or wherever else
14:18 jberger "A place for level-headed discussion about the future of a multi-perl world"
14:19 * mst creates #versionshed
14:19 btyler hah
14:19 mst btyler: let's use there.
14:19 marty joined #mojo
14:20 ribasushi mst: sure #vershed or even #vshed is more perlish ;)
14:26 arthas joined #mojo
14:29 dp_ joined #mojo
14:31 marty joined #mojo
15:20 gryphon joined #mojo
15:22 reneeb joined #mojo
15:56 marty joined #mojo
15:59 ZadYree Hello! I wonder if there already is a way out to make a cpan module from a Mojolicious application. Is there any?
15:59 Ptolemarch joined #mojo
16:12 Ptolemarch joined #mojo
16:19 disputin joined #mojo
16:20 Grinnz_ ZadYree: https://metacpan.org/pod/Mojolicious::Guides::Cookbook#Making-your-application-installable
16:21 ZadYree wOOt!
16:21 ZadYree Didn't see that! Thanks Grinnz_
16:21 Grinnz_ mojo even has some helpers for the makefile and uploading
16:21 Grinnz_ though personally i just use dist zilla ;)
16:25 dvinciguerra joined #mojo
16:29 crab "Since our HeavyThing library itself has hand-written assembler to accomplish all of TLS and gzip, and since our implementation is indeed faster than either of the stock-standard reference libraries, we ended up with a considerably higher performing webserver than not only nginx and lighttpd, but all that we could get our hands on."
16:29 crab crikey (https://2ton.com.au/rwasa/)
16:30 Grinnz_ reneeb: hi, on your page it says that Mojo-Reactor-POE failed all versions but i don't have any failing reports from cpantesters, do you happen to have a build log or could you generate one for me?
16:32 Grinnz_ crab: handwritten TLS and gzip... i hope they are ready for some security testing ;)
16:37 basiliscos joined #mojo
17:00 dod joined #mojo
17:16 disputin joined #mojo
17:40 sh4 joined #mojo
17:46 rem_lex joined #mojo
18:10 pink_mist Grinnz_: not just handwritten, handwritten /assembler/ implementations of those
18:12 Grinnz_ heh
18:23 stephan48 sounds fun
18:44 aramisf joined #mojo
18:52 kaare joined #mojo
18:55 Grinnz_ god dammit EV
18:55 Grinnz_ EV makes exceptions non-fatal so my timeout with "die" is ineffective at rendering an exception
18:56 Grinnz_ oh right i need to do that in the thing
18:57 jberger Grinnz_: fairly common for event loops to do that
18:57 Grinnz_ EV's the only one i've seen
18:57 jberger you don't want an application exception to bring down your loop (read: server)
18:57 Grinnz_ that shouldn't be up to the reactor
18:57 jberger Grinnz_: hmmmm, perhaps true
18:57 Grinnz_ anyway jberger any way i can kill a running ForkCall?
18:57 jberger remember though, EV (and AnyEvent) are really greedy
18:58 jberger kill the process
18:58 Grinnz_ how do i find the process id though
18:58 jberger attach to the spawn event
18:58 Grinnz_ ah, ok thanks
19:00 Grinnz_ hrm, i don't have access to the forkcall object outside the helper though
19:01 Grinnz_ i suppose the helper could just return it
19:04 sri Mojo::Reactor::Poll exceptions are generally non-fatal too, because Mojo::IOLoop subscribes to the error event
19:04 jberger Grinnz_: remember when I added the helper (reluctantly) I said there is no generic way to do what you were asking for?
19:04 jberger that is one of the things
19:04 Grinnz_ well right now, it's not returning anything significant
19:04 jberger how could it return the pid?
19:05 jberger it doesn't have one when it returns
19:05 Grinnz_ i meant return the forkcall object
19:05 jberger hmmmm
19:06 jberger perhaps
19:07 Grinnz_ $self->fork_call(...)->on(...) looks natural enough to me
19:08 jberger Grinnz_: except that by definition we have already set up an error handler, so it feels a little unclean
19:09 jberger you don't necessarily want the user to attach another error handler
19:09 Grinnz_ i don't think that's the module's problem
19:10 Grinnz_ can't think of any alternatives, anyway
19:11 jberger the alternative is to not use the helper, which I was always inclined to recommend
19:15 davido___ joined #mojo
19:19 jberger sri: http://blogs.perl.org/users/shadowcat_mdk/2015/02/gsoc-2015.html
19:19 jberger HTTP2?
19:19 purl it has been said that HTTP2 is now basically the shitty version of spdy
19:19 sri sure
19:19 sri all you need to know is in the issue https://github.com/kraih/mojo/issues/423
19:19 jberger anything else you can think of?
19:20 jberger I can think of Minion::Monitor work perhaps or Mojo::Pg, each of which are related to Mojolicious
19:21 sri Mojo::Pg is done
19:21 jberger I didn't know if you had anything more to add, so scratch that
19:22 sri Minion problems might be too hard
19:22 jberger Minion::Monitor is more of an app than Minion
19:23 jberger but I think it makes it a lot more approachable
19:23 sri oh, i meant problems i would like to see addressed ;p
19:23 jberger oh ok
19:23 sri like better monitoring of workers
19:23 jberger yeah, I don
19:23 sri i don't even know how to do it myself
19:23 jberger don't know what I would add there
19:23 jberger right
19:24 sri a prefork worker command might be nice (as a 3rd party cpan module)
19:24 jberger the next win for Mojo::Pg will be if/when Postgres adds a RESTful interface
19:25 sri that's not going to happen in the next few releases i think
19:25 jberger too bad
19:25 sri can't say i care very much
19:25 mst not everything has to be HTTP
19:25 sri multi-master replication is much cooler
19:26 jberger mst: its only about the non-blocking transport without having to implement the wire protocol
19:26 sri jberger: i'm ok with blocking i/o so far
19:26 jberger and as sri says, not a high priority in a geographically well laid out architecture
19:35 human39 joined #mojo
19:48 Ptolemarch joined #mojo
19:51 Grinnz_ so another oddity... morbo was reaching the 15 second default inactivity timeout, and apache would *re-send* the POST
19:52 Grinnz_ thats uhhh not good
19:52 Grinnz_ i set the timeout higher but idk if theres something else wrong...
19:55 jberger Grinnz_: you are using apache as a reverse proxy?
20:05 dotandimet joined #mojo
20:14 meshl joined #mojo
20:14 jberger so I have an API that has some HTML representations
20:15 jberger and I'm trying to make it do the right thing when someone attempts to access the resource when not authenticated
20:15 jberger do you think its sane to do 401 if XHR header is present, and 302 to the login page otherwise?
20:20 human39 joined #mojo
20:20 rem_lex joined #mojo
20:22 Grinnz_ jberger: yeah
20:23 jberger should be right most of the time, and our own site is almost always the consumer (via XHR)
20:23 meshl joined #mojo
20:23 jberger oh, Grinnz_ you were replying to my question to you, not my question at large
20:24 Grinnz_ lol
20:38 jberger ok, I chatted with mdk and he likes the idea
20:38 jberger I said that I would be primary contact, with sri available for more specific questions (seemed true as that is normally the case anyway)
20:42 Grinnz_ re: blocking I/O in Pg, i don't think that matters too much since it's always going to send you all the data at once, right?
20:43 jberger Grinnz_: yeah
20:44 jberger or at least, that is my understanding
20:56 meshl joined #mojo
20:59 sri it does
20:59 sri and for performance it doesn't even use prepared statements
21:01 Ptolemarch joined #mojo
21:08 sri (so i/o is pretty straight forward, just request/response every time)
21:22 marmez joined #mojo
21:28 alnewkirk joined #mojo
21:29 irq joined #mojo
21:50 pink_mist Grinnz_: so did you ever find out why reneeb's test service fails that module?
21:52 Grinnz_ not yet
21:52 Grinnz_ btw if you want to help figure out what was wrong with your cpan, i mentioned your build log in #toolchain
21:53 Grinnz_ <xdg> I'd want to see the CPAN config file and a perl -V dump and any details about how it was run
22:07 pink_mist I'll provide those tomorrow; am already in bed, and don't have /that/ laptop with me :P
22:18 disputin joined #mojo
22:23 punter joined #mojo
22:26 Gedge joined #mojo
22:40 tgt joined #mojo
22:51 xdg pink_mist, and I'm lurking here, too.  :-)
23:03 jberger NCC-1701. No bloody A, B, C, or D
23:24 marty joined #mojo
23:25 asarch joined #mojo
23:34 Grinnz_ btw, around_dispatch actually did work out for the warning i wanted to do on requests that took a certain amount of time
23:35 Grinnz_ because any deferred rendering isn't inside that hook
23:35 Grinnz_ so it doesnt count
23:56 Averna joined #mojo

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