Activity Graph

Page 1 of 1


[02:19:25] wkoszek: Ping timeout: 260 seconds


[14:05:14] wkoszek: havenwood: Thanks.


[21:48:29] wkoszek: havenwood: During out last discussion you mentioned Ruby can be statically compiled?
[21:48:42] wkoszek: I'm looking around and I can't seem to see people doing that.
[21:52:43] wkoszek: ACTION found gem-compile to pre-compile ruby gems; trying it out on 'json' extension


[06:52:45] wkoszek: has joined #ruby


[14:51:38] wkoszek: has joined #ruby


[00:15:34] wkoszek: Write error: Broken pipe


[04:18:02] wkoszek: from reading the rails framework code and some docs, my understanding is that `add_foreign_key` is same as t.references.. Am I get it right?
[04:19:04] wkoszek: I'm attempting to do: t.references :posts, foreign_key: true, on_delete: :cascade but the cascading doesn't seem to work (not getting this constraint executed in the SQL)
[05:40:28] wkoszek: hm. you actually need to use: foreign_key: {on_delete: :cascade}


[06:29:39] wkoszek: ACTION thinks he got pkgr to work for his empty Rails app template. Can deploy his Rails app via apt-get now.
[06:43:06] wkoszek: sevenseacat: the other approaches to just having rails app come up and work in case something happened (out of disk? out of ram?) required a lot of work
[06:43:53] wkoszek: pkgr have me a way to just apt-get my app; I got 'service <myapp> start', and a helper utility called 'myapp ...' and it seems to work
[06:44:19] wkoszek: not sure for how long it'll be the case, but if it ends up working like that longer term, that was worth a 2 day investment.
[06:44:52] wkoszek: yeah, it's ain't pretty, but works.


[02:14:18] wkoszek: I'm looking for a nice Ruby framework for DNS configuration. Something like this: https://github.com/AnalogJ/lexicon but in Ruby
[02:14:21] wkoszek: Any hints?


[01:17:27] wkoszek: ACTION reading about --deployment
[02:12:40] wkoszek: I think you just need to lookup the bigest consumers of Rails and see what their workload is and compare it t yours.
[02:13:22] wkoszek: Quite likely your workload is smaller, so Rails should be just fine for you.
[02:30:19] wkoszek: Bad performance of your code is a good problem to have. Typically it's less of an issue for all of your 5 users :)
[04:49:55] wkoszek: Any pkgr users here?
[05:12:58] wkoszek: Are there any pkgr users here? I'm trying to package Rails app with it and have it use system Ruby. But it insists on compiling a new one. Each time.


[22:58:15] wkoszek: Is there a reason why people don't commit their bundler configuration for Rails apps? I see that the default .gitignore for Rails has also the config of Bundler ignored


[04:05:05] wkoszek: has joined #ruby
[04:05:08] wkoszek: has joined #RubyOnRails


[07:29:08] wkoszek: Remote host closed the connection


[05:36:53] wkoszek: So I'm still looking for testers for my reproducible Ruby builds: https://github.com/wkoszek/ruby_packages


[18:44:35] wkoszek: Remote host closed the connection
[19:10:52] wkoszek: has joined #ruby
[19:10:54] wkoszek: has joined #RubyOnRails
[19:23:59] wkoszek: has joined #ruby
[19:23:59] wkoszek: has joined #RubyOnRails
[19:23:59] wkoszek: Changing host


[05:34:30] wkoszek: has joined #ruby
[05:34:56] wkoszek: How can I open an issue at https://github.com/ruby/xmlrpc
[05:36:44] wkoszek: Radar: This is a part of the core Ruby library. I just wonder if the report should be made in ruby/ruby?
[05:38:12] wkoszek: Radar: I'll e-mail the hstb@ and ask him
[05:53:09] wkoszek: I'm attempting to understand if this module would let me do XML RPC request via SSL, but without cert validation. I could a piece of code: https://github.com/ruby/xmlrpc/blob/master/lib/xmlrpc/client.rb#L445
[05:58:26] wkoszek: ineb: Yeah, but from what I see, I have no way to pass it in this module
[06:04:43] wkoszek: ineb: I have a self-signed, which is pretty popular to just get encryption.
[06:28:07] wkoszek: By the way, I'd still be interested in getting some testers for my Ruby builds: https://github.com/wkoszek/ruby_packages
[07:31:31] wkoszek: ACTION needs to figure out how to make ruby be ruby-2.4 after ./configure && make
[07:31:45] wkoszek: ACTION wanted to take his ruby_packages repo and push it to a ppa
[21:23:28] wkoszek: has joined #RubyOnRails
[21:32:31] wkoszek: How many of you are installing > 1 Ruby version at one time in the production system? 2nd question would be: do you use RVM in production deployments, or only for your development flow?
[21:38:08] wkoszek: I've had a moment of frustration on my side, when I tried to marry Ansible to Rails and RVM caused me problems. Tried 2 ready-to-use playbooks, and they didn't work either, so I just went and did a reproducible Ruby 2.4.0 builds: https://github.com/wkoszek/ruby_packages And it seems to work for me, but to make it official, I'd like to push it to my own PPA.
[21:38:45] wkoszek: And just trying to think if what I have is enough (Ruby gets installed as 'ruby') or whether I should fuzzle with it a bit more and make it install as 'ruby-2.4.0' or something).
[21:39:20] wkoszek: But the idea is that I grab the .deb from https://github.com/wkoszek/ruby_packages/releases and I install it system-wide, and then just ask bundler to install dependencies locally.
[21:50:53] wkoszek: ACTION is looking for testers; got no success/failure stories so far. he's the only tester.
[22:17:24] wkoszek: lupine: I'm gravitating towards what you say, precisely. pkgr looks very interesting, so maybe I'll give it a shot too.
[22:36:19] wkoszek: I've also read many articles on: what can a Ruby on Rails hacker do to structure the migration in a way that DB and code don't blow up.
[22:37:07] wkoszek: But I can't seem to find good pieces on: how to seemlessly run the migration on a production database.
[22:37:48] wkoszek: I imagine people use Capistrano and just run rails db:migrate on the production system, but if there's an expensive operation, this basically ... locks your site for some time.
[22:44:54] wkoszek: lupine: yes, it is. btw, great link, thanks for that.
[22:47:20] wkoszek: I see it in my mind: I work in dev. and everything works OK, I test on my Vagrant closely mimicking the production, all tests pass, I'm ready to do migration in production. You do rails db:migrate remotely, and it gives you: "Sorry, can't do it. BTW, I made irreversible changes to your database, so your site may be broken now. Sorry"
[22:48:22] wkoszek: lupine: I use PostgresSQL, but I think it's same for any database I think.
[22:50:12] wkoszek: I'm thinking exactly that. To have a database setup figured out, and do the migration by setting up a slave, syncing data to it, running migrations on it, and once everything owrks, that's your new master.
[22:50:47] wkoszek: yeah. I'm not a unicorn yet, so I guess taking the site down for 10-20min isn't a big deal.
[22:51:43] wkoszek: ACTION dreams big: thousands of customers, millions of requests; but would like to get the dev. model figured out mostly becuase he doesn't trust himself
[22:54:31] wkoszek: i need to measure how much it takes to just dump the 1-2GB database to a file; probably it's not that bad
[22:59:25] wkoszek: https://medium.com/g-adventures-engineering/speeding-up-postgres-restores-de575149d17a <- this was a cool post about this very topic.


[05:14:33] wkoszek: Which open-source Rails projects have a nicely figured out deploy model, so that I could look at it?
[05:15:19] wkoszek: I'm interested in stealing a flow for running a migration on a remote machine and having staging/prod on a separate VPSes.
[05:21:13] wkoszek: has joined #ruby
[05:27:35] wkoszek: Before you do so, you may source ~/.profile && which gem <- whatever you get, should show you ~/.rvm/... bleh path.
[06:13:04] wkoszek: Read error: Connection reset by peer


[19:42:29] wkoszek: Ping timeout: 252 seconds


[05:04:50] wkoszek: has joined #RubyOnRails


[17:19:14] wkoszek: *.net *.split


[23:45:17] wkoszek: Read error: Connection reset by peer
[23:45:23] wkoszek: has joined #RubyOnRails


[16:08:15] wkoszek: has joined #RubyOnRails


[02:23:50] wkoszek: Ping timeout: 248 seconds
[02:23:54] wkoszek: has joined #RubyOnRails
[03:22:40] wkoszek: *.net *.split


[12:53:56] wkoszek: Ping timeout: 264 seconds
[13:25:26] wkoszek: has joined #RubyOnRails


[22:53:35] wkoszek: *.net *.split
[22:58:01] wkoszek: has joined #RubyOnRails


[13:17:43] wkoszek: Read error: Connection reset by peer
[13:17:49] wkoszek: has joined #RubyOnRails
[21:50:18] wkoszek: *.net *.split
[21:51:38] wkoszek: has joined #RubyOnRails


[19:35:53] wkoszek: *.net *.split
[19:35:58] wkoszek: has joined #RubyOnRails


[22:43:09] wkoszek: Radar: happy to hear you come from php background. will make more wishlist entries for rails internals book
[22:45:17] wkoszek: More of a deployment comparison, since most of people are more familiar with mod_php than they are with how rails elements work
[22:52:12] wkoszek: Sometimes makes me wonder why Rails isn't just deployed over xinetd/inetd.
[22:52:31] wkoszek: Since my understanding is that in Rails ecosystem threads are non-existent.
[22:53:13] wkoszek: sevenseacat: if you have any good examples feel free to share some references.
[22:57:47] wkoszek: sevenseacat: I will check the last 2 resources out.
[22:59:37] wkoszek: sevenseacat: I found a lot of: here's the commands you type... bla bla, but few on how it actually works.
[23:00:50] wkoszek: It's super different from Php, where everything is handled by the mod_php and within Apache process space.
[23:03:01] wkoszek: Passanger is the best I've found so far, yes.
[23:03:34] wkoszek: CornishPasty: I don't understand.
[23:05:23] wkoszek: ACTION still needs to find Ubuntu packages for passanger, since he doesn't agree for installing software outside of apt repos.
[23:05:51] wkoszek: I liked passanger best since it seemed to give me a way to have more apps on 1 server.
[23:06:15] wkoszek: Papierkorb: unfortuantely I break the rule for Ruby since there's no other way. I use rvm.
[23:06:53] wkoszek: CornishPasty: i think you describe the world pre-fastcgi.
[23:08:48] wkoszek: Papierkorb: That's what I'm look at right now.
[23:08:58] wkoszek: Honestly, unicorn marketing is interesting.
[23:09:15] wkoszek: as if I knew if the clients connecting to me will be on low-latency links and fast.
[23:10:07] wkoszek: nginx is just a forwarder, so it'll hit unicorn anyway
[23:15:03] wkoszek: unicorn/puma work is similar way: they'll pre-fork N processes and basically start handling what nginx will give them.
[23:20:07] wkoszek: ACTION is trying to understand underlying reason why rails seems to be process based.
[23:20:41] wkoszek: ACTION notes he likes rails, but seems like performance in all sorts of benchmarks is so-so.
[23:20:46] wkoszek: ACTION tries to understand the reason
[23:22:26] wkoszek: tubbo: this is true point of view if your hosting is for free
[23:22:51] wkoszek: i get that, but i'm trying to understand how good/bad it is.
[23:23:12] wkoszek: and what people do when they hit performance problems.
[23:24:12] wkoszek: I just mentioned PHP since I've done deployments with it and I feel like I understand it better.
[23:25:14] wkoszek: rails is cool and i'm studying it, but my experience with ruby speed is so so.
[23:25:28] wkoszek: so i'm wondering how much other people notice the perf. issues.
[23:28:50] wkoszek: mordocai: how many concurrent users/requests do you guys have?
[23:29:29] wkoszek: ACTION notes benchmarks he looked at where for(10000 requests) { get("/") } smart, so he can't really comment
[23:31:35] wkoszek: tubbo: what's your site?
[23:32:40] wkoszek: tubbo: of course.
[23:33:48] wkoszek: tubbo: seems fast. so that's the whole setup you're hosting it on?
[23:41:14] wkoszek: ACTION knows he's more likely to fail with his projects than hit performance problems, but the cost of learning rails is worth to consider speed of rails
[23:44:24] wkoszek: tubbo: sorry i missed 1 part: was this puma / unicorn?
[23:44:40] wkoszek: from what i can tell from this channel, it's basically a matter of the preference.


[00:33:19] wkoszek: argoneus: Don't push to Git.
[00:33:45] wkoszek: argoneus: Make a backup and make sure this file isn't in Git and that you have .gitignore set properly.
[15:35:54] wkoszek: *.net *.split
[16:28:43] wkoszek: has joined #RubyOnRails


[22:00:14] wkoszek: Radar: that totally sucks
[22:00:25] wkoszek: ACTION sees Radar selling his soul to devil
[22:03:56] wkoszek: ACTION wonders if he prefers raspberry pi over rails 4 in action from santa
[22:07:11] wkoszek: Radar: Frugal miser here, but yes, both are good investments.
[22:54:08] wkoszek: Is there a good resource on Rails internals? Basically I'd like to understand some more about how Rails handled connections for example.
[22:55:18] wkoszek: While this thing is fairly well understood for Apache/nginx + .php stuff, Rails with all the Rack stuff and actual servers, is so-so understood I feel.
[22:58:06] wkoszek: Radar: That would be great.
[22:58:58] wkoszek: I know that having scalability problems nowadays is a good problem to have (e.g.: most of your projects are probably not going to be popular enough to hit issues), but I'd like to understand what I'm getting into wrt. Rails.
[22:59:57] wkoszek: Radar: You bet I will. :-)
[23:31:07] wkoszek: so what's the biggest rails deployment you've guys dealt with? (and how many req/s) ?
[23:31:13] wkoszek: or better: how many users/s
[23:33:21] wkoszek: How many machines are serving this?
[23:33:38] wkoszek: Radar: Ok. Sounds good.
[23:34:41] wkoszek: Ok, I guess I should get over my performance anxiety then.
[23:34:57] wkoszek: Radar: This is plain normal Ruby or JRuby or Rubinius?
[23:35:21] wkoszek: Radar: I see. That's not bad.
[23:36:31] wkoszek: I think there's no objections from Rails when I only want it to be the backend for REST stuff?
[23:36:48] wkoszek: e.g.: I susepct the biggest sense to write an app nowadsy is to do the frontend in something like Angular.
[23:37:11] wkoszek: so HAMLs and ERBs of this world IMO may not be a way to go.
[23:37:55] wkoszek: You need views, but they'd be rendering JSON instead of HTML.
[23:38:43] wkoszek: I think it's hard to write slick UI without Angular.
[23:39:02] wkoszek: or rather:
[23:39:14] wkoszek: Angular/whatever_your_ui_framework_is


[23:20:49] wkoszek: has joined #RubyOnRails
[23:24:52] wkoszek: http://www.isrubyfastyet.com/ <- pretty cool benchmark.
[23:40:21] wkoszek: Is there a web benchmarking tool where I could benchmark number of concurrent "sessions", where a session is defined as something real, e.g.: user loging in, selecting 10 items, shuffling them inside of the web app, saving and logging out?


[22:19:19] wkoszek: *.net *.split


[19:02:09] wkoszek: *.net *.split
[19:11:14] wkoszek: has joined #RubyOnRails


[00:31:17] wkoszek: Radar: Other than your book, I have Agile Web Development with Rails 4. Anything else I may want to order to get to the Rails intermediate level?
[00:32:59] wkoszek: Radar: oki doki.
[09:12:50] wkoszek: ACTION has a sad moment; instead of 'gci' typed 'gco', which in his vocabulary is "git checkout"
[09:13:51] wkoszek: rvanlieshout: git commit
[09:17:10] wkoszek: ACTION is having his 3hrs of coding saved by copy & pasting from iterm history
[19:21:30] wkoszek: When I type "rails g XXX", is there a way/plugin mechanism to write an extension to intercept all the arguments to "rails" ?
[20:28:45] wkoszek: Radar: You have a lot of good stuff on your website.


[01:10:07] wkoszek: Radar: I haven't tried vim-ruby. Will take a look at it.
[01:10:58] wkoszek: Radar: lag == work, yes :-)
[01:11:36] wkoszek: I've looked at RubyMine and I haven't ever managed to get that level of qualify of completion in Vim.
[01:11:46] wkoszek: But I tried couple of years ago, so wondering if things had changed.
[01:33:32] wkoszek: Aeyrix: Do you have method/variable completion? I apm install ruby-lanauges, restarted Atom, created sample.rb, and I have in it: s = "string". I type "s." and it doesn't pull up a menu with methods.
[01:34:55] wkoszek: Aeyrix: It only seems to be good for reserved keywords. So it knows "def" and "end" etc.
[01:40:10] wkoszek: Aeyrix: For me it only takes me using Ruby at home and Python at work to get confused about: string.length and len(string) and other silly stuff.
[01:40:28] wkoszek: But you're right, it's not an excuse to not keep remembering stuff.
[01:40:57] wkoszek: I just sometimes gets syntax mixed up.
[01:41:14] wkoszek: As we speak, I tried the same in RubyMine and it doesn't know string's methods either.
[01:41:40] wkoszek: I guess the nature of Ruby makes it hard for this tool to infer the type.
[01:42:30] wkoszek: Android Studio thanks to Java types is very smart.
[01:43:11] wkoszek: You want to pass a function into another function, and it'll only complete types that match the expected argument signature.
[01:44:06] wkoszek: Do you guys have a solution for saving the state of the Atom after I quit? I tried save-session, but it didn't work for me
[01:46:00] wkoszek: Aeyrix: I close it, since it's starts to suck the battery with 2-3 projects opened..
[01:46:51] wkoszek: Maybe I'll try sevenseacat's project-manager hint. Would be nice to basically be able to keep your settings preserved in case I need to reboot the mba
[23:30:34] wkoszek: Is "devise" the best thing for user management for Ruby? Looking for the lowest effort approach to get user/admin management solved, and preferably with FB/Twitter/G+ integration.
[23:32:44] wkoszek: Radar: Funny you metnioned it, since Rails 4 in Action is something I just saw on the bookshelf this weekend. I'll order it.
[23:33:08] wkoszek: Radar: Aha!
[23:33:20] wkoszek: I didn't know poets hang out on this channel.
[23:34:13] wkoszek: Radar: It looked very interesting. I spent at least 10min looking through it.
[23:35:00] wkoszek: Radar: I visited NYC this weekend and saw it there.
[23:35:39] wkoszek: Radar: Will ask my gf and let you know.
[23:36:04] wkoszek: Radar: Might have been Broadway street, but I don't remember the store's name.
[23:39:21] wkoszek: ACTION notes this bookstore had a pretty good ollections of great tech literature.
[23:40:18] wkoszek: Radar: The Strand, 828 Broadway, New York, NY 10003
[23:40:31] wkoszek: Radar: You're in NYC ?
[23:50:06] wkoszek: Radar: I see. Is it available in the e-form?
[23:50:25] wkoszek: Radar: I see your 1st book was via LeanPub
[23:51:11] wkoszek: Radar: I'm in the middle of Guy Kawasaki's APE book wrt. self-publishing.


[21:53:56] wkoszek: Changing host
[21:53:57] wkoszek: has joined #RubyOnRails
[21:54:07] wkoszek: What are the best options to get a smart Ruby code completion in Vim ? And by smart I mean: so that it'd be contextual, could see methods etc..


[17:05:06] wkoszek: Ping timeout: 240 seconds
[17:29:07] wkoszek: has joined #RubyOnRails
[17:45:33] wkoszek: Ping timeout: 250 seconds
[18:32:34] wkoszek: has joined #RubyOnRails
[18:43:46] wkoszek: Ping timeout: 240 seconds
[18:45:18] wkoszek: has joined #RubyOnRails
[19:20:10] wkoszek: Ping timeout: 246 seconds
[20:04:14] wkoszek: has joined #RubyOnRails


[02:21:24] wkoszek: Ping timeout: 255 seconds
[03:31:09] wkoszek: has joined #RubyOnRails
[03:35:28] wkoszek: Client Quit
[20:56:33] wkoszek: has joined #RubyOnRails
[20:56:56] wkoszek: Changing host
[20:56:57] wkoszek: has joined #RubyOnRails


[22:12:31] wkoszek: has joined #RubyOnRails
[22:19:36] wkoszek: Is there an API which would "explain" my Ruby script the structure of the Rails app: how many models, and what they are, how many controllers, and what they are etc.. ?
[22:20:30] wkoszek: rhizome: these are just stats.
[22:22:31] wkoszek: I want to make myself a CLI program which based on the Rails models structure, would give me a uniform argument parsing for the REST API I created.
[22:24:13] wkoszek: Well, basically for each model M added with variables A:string B:string C:number, I'd like to get a command line interface to my Rails app: myapp M --new --A value1 --B value2 --C 123
[22:24:30] wkoszek: And it'd call curl or wget for me.
[22:25:16] wkoszek: So they way to think about it is that either I'd have to remember the command line calls made to "rails" command, or (better), I could write a script which could pull out all this stuff from the existing Rails app.
[22:27:25] wkoszek: Yeah. Since I believe there's no real database backing each of the performed commands (e.g.: I think "rails g scaffold bleh" isn't registered anywhere), I believe one of the ways to approach this would be to "overload" names used in models syntax and basically execute all models' .rb files.
[22:28:48] wkoszek: rake routes at least gives me the routes, but I don't think I can easily pull out a knowledge of what I can send to each API endpoint.


[00:06:50] wkoszek: shevy: Thanks man. Will check #rack. Wrt. middleman: I use it for my website https://circleci.com/gh/wkoszek/me/70, and I added 1 article the same way as always, and it is not picked up.
[00:08:45] wkoszek: So I did simple debugging: copied the content from YYYYMMDD-new-article-here.md to YYYYMMDD-existing-....md and it rendered OK.
[00:09:33] wkoszek: (this eliminated my suspicionnthat I got frontmatter messed up). Then I copied existing article to a new file, changed the date in the frontmatter and it still doesn't render.
[02:20:37] wkoszek: Quit: leaving


[23:40:23] wkoszek: has joined #ruby
[23:48:22] wkoszek: Do you guys know any channel where Middleman is discussed? Tried #thoughbox, #middleman and #middlemanapp, but with no success.