jds

Activity Graph

Page 1 of 2 | Next »

2018-06-07

[13:03:12] jds: has joined #ruby

2016-05-18

[16:55:08] jds: has joined #RubyOnRails
[16:55:23] jds: has joined #ruby
[17:12:17] jds: Disconnected by services

2016-05-17

[08:12:37] jds: has joined #RubyOnRails
[08:12:47] jds: has joined #ruby
[10:59:34] jds: Quit: Connection closed for inactivity

2016-05-13

[09:37:38] jds: has joined #RubyOnRails
[09:37:47] jds: has joined #ruby
[12:49:32] jds: Quit: Connection closed for inactivity

2016-05-11

[09:59:31] jds: Quit: Connection closed for inactivity

2016-05-10

[15:30:31] jds: has joined #RubyOnRails
[15:30:41] jds: has joined #ruby

2016-05-08

[11:46:31] jds: has joined #RubyOnRails
[11:46:40] jds: has joined #ruby
[16:49:30] jds: Quit: Connection closed for inactivity

2016-05-06

[14:38:57] jds: has joined #RubyOnRails
[14:39:09] jds: has joined #ruby
[16:39:29] jds: Quit: Connection closed for inactivity

2016-05-03

[11:17:29] jds: has joined #RubyOnRails
[11:17:39] jds: has joined #ruby
[11:32:20] jds: Does anyone have capistrano set up to compile assets on a single box, then copy the results to others? I'm surprised I can't find anything about it on Google, but maybe I'm searching for the wrong things
[19:29:27] jds: Quit: Connection closed for inactivity

2016-04-26

[18:49:23] jds: Quit: Connection closed for inactivity

2016-04-25

[14:29:22] jds: Quit: Connection closed for inactivity

2016-03-17

[10:07:47] jds: has joined #RubyOnRails
[10:07:51] jds: has joined #ruby
[14:37:06] jds: Quit: Connection closed for inactivity
[15:07:24] jds: has joined #RubyOnRails
[15:07:28] jds: has joined #ruby
[17:27:06] jds: Quit: Connection closed for inactivity

2016-03-15

[16:20:08] jds: has joined #RubyOnRails
[16:20:12] jds: has joined #ruby
[20:07:05] jds: Quit: Connection closed for inactivity

2016-02-22

[11:57:46] jds: has joined #RubyOnRails
[11:57:50] jds: has joined #ruby
[15:36:59] jds: Quit: Connection closed for inactivity

2016-02-18

[09:07:35] jds: has joined #RubyOnRails
[09:07:39] jds: has joined #ruby
[19:06:58] jds: Quit: Connection closed for inactivity

2016-02-12

[12:35:15] jds: has joined #RubyOnRails
[12:35:19] jds: has joined #ruby
[12:40:26] jds: Is it possible to tell sprockets that a file has a dependency on an arbitrary ruby value? Along the lines of `depends_on`, but not tied to a file path
[12:41:34] jds: More concretely, I have a .js.erb file with `app.environment = "<%= Rails.env %>"`, and my tests break if assets have been precompiled for production, because sprockets doesn't bother recompiling that file in the test environment
[12:53:43] jds: @universa1 the stuff compiled to public/assets isn't actually a problem - it's the cache in tmp/cache/sprockets that's biting me
[12:53:58] jds: I guess I could move that aside between test runs
[12:54:08] jds: or tweak assets.version in the test environment
[15:36:57] jds: Quit: Connection closed for inactivity

2016-02-10

[20:45:30] jds: has joined #RubyOnRails
[20:45:34] jds: has joined #ruby

2016-02-05

[15:12:28] jds: has joined #RubyOnRails
[15:12:32] jds: has joined #ruby
[17:26:55] jds: Quit: Connection closed for inactivity

2016-02-02

[09:31:10] jds: has joined #RubyOnRails
[09:31:14] jds: has joined #ruby
[11:56:54] jds: Quit: Connection closed for inactivity

2016-02-01

[17:27:46] jds: has joined #RubyOnRails
[17:27:50] jds: has joined #ruby
[21:16:54] jds: Quit: Connection closed for inactivity

2016-01-01

[07:56:49] jds: Quit: Connection closed for inactivity

2015-12-31

[13:09:46] jds: has joined #RubyOnRails
[13:09:50] jds: has joined #ruby

2015-12-22

[17:51:32] jds: has joined #RubyOnRails
[17:51:37] jds: has joined #ruby
[20:16:47] jds: Quit: Connection closed for inactivity

2015-12-21

[09:03:52] jds: has joined #RubyOnRails
[09:03:56] jds: has joined #ruby
[12:06:47] jds: Quit: Connection closed for inactivity

2015-12-20

[17:10:18] jds: has joined #RubyOnRails
[17:10:23] jds: has joined #ruby
[17:11:54] jds: Brakeman reckons `redirect_to params.slice(:view_filter, :user_id, :format).to_h, status: 301` is an unprotected redirect. There's no combination of input where that could be unprotected, right?
[17:21:22] jds: Hates_: `redirect_to foo` is only a security issue if foo might contain a host, right? So either an absolute string, or a hash containing a :host key?
[20:46:47] jds: Quit: Connection closed for inactivity

2015-12-18

[10:17:57] jds: has joined #RubyOnRails
[10:18:01] jds: has joined #ruby
[19:36:46] jds: Quit: Connection closed for inactivity

2015-12-11

[11:19:40] jds: has joined #RubyOnRails
[11:19:44] jds: has joined #ruby
[15:36:45] jds: Quit: Connection closed for inactivity

2015-12-10

[01:54:48] jds: Ping timeout: 264 seconds
[01:55:28] jds: has joined #RubyOnRails
[01:55:33] jds: has joined #ruby
[12:56:45] jds: Quit: Connection closed for inactivity

2015-12-09

[10:17:17] jds: has joined #RubyOnRails
[10:17:22] jds: has joined #ruby
[14:06:45] jds: Quit: Connection closed for inactivity
[15:46:17] jds: has joined #RubyOnRails
[15:46:21] jds: has joined #ruby

2015-12-04

[12:06:44] jds: Quit: Connection closed for inactivity

2015-12-03

[15:54:25] jds: has joined #ruby
[15:54:25] jds: has joined #RubyOnRails

2015-12-02

[14:12:32] jds: Quit: Connection closed for inactivity

2015-12-01

[16:21:54] jds: has joined #ruby
[16:22:30] jds: has joined #RubyOnRails
[16:23:18] jds: Anyone aware of anything that creates a table called 'encoding_test' ? It keeps randomly cropping up in my schema.rb - I delete it, then something occurs that recreates it a few days later
[16:38:33] jds: @jhass using DROP TABLE in mysql, and removing it from schema.rb
[16:42:51] jds: jhass: Yeah, it doesn't exist. Not in my gems either
[16:42:57] jds: so there's clearly something crazy occuring
[16:43:43] jds: I've changed my database privileges so that CREATE TABLE will be denied, hopefully it'll make it more obvious when it tries to recreate it
[16:43:51] jds: thanks.
[16:49:26] jds: IIRC '-' needs to be first in the character class
[16:50:10] jds: huh, ok
[16:54:57] jds: kitsuneyo: ChIJ2_UmUkxNekgRqmv-BDgUvtk has an underscore
[16:56:11] jds: Oh, sorry. I think I need more coffee.

2015-11-30

[08:01:17] jds: has joined #ruby
[08:14:36] jds: Ping timeout: 264 seconds
[08:16:22] jds: has joined #ruby
[08:30:57] jds: Changing host
[08:30:58] jds: has joined #ruby
[08:50:36] jds: Ping timeout: 264 seconds
[09:00:18] jds: has joined #ruby
[10:42:32] jds: Quit: Connection closed for inactivity

2015-11-29

[19:25:43] jds: has joined #ruby
[22:02:37] jds: Quit: Connection closed for inactivity

2015-10-23

[12:52:20] jds: Quit: Connection closed for inactivity

2015-10-22

[11:38:57] jds: has joined #ruby

2015-10-21

[08:38:41] jds: has joined #ruby
[20:42:19] jds: Quit: Connection closed for inactivity

2015-09-29

[16:30:53] jds: has joined #ruby
[20:52:13] jds: Quit: Connection closed for inactivity

2015-09-28

[08:47:54] jds: has joined #ruby
[14:42:12] jds: Quit: Connection closed for inactivity

2015-09-12

[16:59:11] jds: has joined #ruby
[19:52:09] jds: Quit: Connection closed for inactivity

2015-09-03

[09:01:09] jds: has joined #ruby
[09:06:53] jds: I've been trying to profile some slow rspec/rails specs to see where all the time is being spent, but the resulting call graph is pretty incomprehensible - I'm basically just seeing a lot of cyclic calls to instance_exec & Proc.call. Any tips?
[12:52:06] jds: Quit: Connection closed for inactivity

2015-09-01

[08:37:15] jds: has joined #ruby
[20:22:06] jds: Quit: Connection closed for inactivity

2015-08-30

[12:09:09] jds: has joined #ruby
[21:12:07] jds: Quit: Connection closed for inactivity

2015-08-22

[16:27:56] jds: has joined #ruby
[18:32:03] jds: Quit: Connection closed for inactivity

2015-08-17

[15:30:50] jds: has joined #ruby
[18:08:50] jds: Quit: Connection closed for inactivity

2015-08-14

[15:10:32] jds: has joined #ruby
[18:48:49] jds: Quit: Connection closed for inactivity

2015-08-07

[08:25:55] jds: has joined #ruby
[10:58:46] jds: Quit: Connection closed for inactivity

2015-08-06

[10:18:55] jds: has joined #ruby
[20:28:46] jds: Quit: Connection closed for inactivity

2015-07-17

[11:10:03] jds: has joined #ruby
[14:17:43] jds: Quit: Connection closed for inactivity

2015-06-25

[06:01:16] jds: has joined #ruby
[08:17:31] jds: Quit: Connection closed for inactivity

2015-06-17

[10:19:56] jds: has joined #ruby
[12:57:27] jds: Quit: Connection closed for inactivity

2015-06-16

[12:07:27] jds: Quit: Connection closed for inactivity

2015-06-15

[17:32:56] jds: has joined #ruby

2015-06-12

[11:19:36] jds: has joined #ruby
[17:17:26] jds: Quit: Connection closed for inactivity

2015-06-10

[09:29:58] jds: has joined #ruby
[12:07:25] jds: Quit: Connection closed for inactivity

2015-06-09

[08:47:13] jds: has joined #ruby
[10:57:25] jds: Quit: Connection closed for inactivity

2015-06-07

[15:57:33] jds: has joined #ruby
[20:21:28] jds: Quit: Connection closed for inactivity

2015-06-06

[08:50:33] jds: has joined #ruby
[14:55:51] jds: Quit: Connection closed for inactivity

2014-10-07

[14:08:15] jds: I'm using nokogiri, and want to run something along the lines of xml.xpath("//foo[@attr='#{user_input}']"). Is there anything in nokogiri to help me escape 'user_input' appropriately ?
[14:10:10] jds: ugh. http://stackoverflow.com/questions/14822153/escape-single-quote-in-xpath-with-nokogiri

2014-09-16

[12:11:56] jds: I'm converting an existing test suite over to fixtures, and finding the fixtures are actually slower than just repeatedly creating the models via Machinist. Struggling to explain why that might be - any guesses?
[12:19:59] jds: Hm. Let me see if I can benchmark the initial fixture creation
[12:40:56] jds: Ugh, looks like my fixture slowdown is due to rspec
[12:41:10] jds: It's re-loading them on each individual spec
[13:42:10] jds: Rails' view cache keys aren't scoped to SSL or the domain - ie the cache key looks something like "views/posts/_post.haml/123456". Is there an easy way of including the current request's domain & protocol in there?

2014-09-08

[17:54:37] jds: Is there any way of making translations in yml files reference other translation phrases?
[17:54:56] jds: eg greeting: "hello %{world}", where 'world' is another translated phrase
[17:55:20] jds: Hopefully without passing 'world' as a variable (t 'greeting', world: t('world'))

2014-09-02

[10:14:03] jds: Anything I should know about with the assets pipeline when upgrading from 4.0 to 4.1?
[10:14:20] jds: For some reason it seems like assets aren't getting generated properly for my production environment, though they are for staging :/
[10:24:24] jds: Hmph. Does the asset digest now depend on the ruby version or something?
[10:34:06] jds: My balancer is missing assets. I had to revert it pretty quickly so I'm still not 100% sure, but it looks like my rails servers precompiled a different set of assets than the balancer machines did
[10:34:41] jds: So rails generates pages referencing "desktop-abc.css", and the balancer (where the assets are served from) only has "desktop-xyz.css"
[10:39:07] jds: I've been using the asset pipeline for years, it's definitely something different with rails 4.0 vs 4.1 (or the related sprockets version)
[10:40:06] jds: sorry - that came out arrogant-sounding. I mean the deployment workflow I've got is pretty well tested, it auto-restarts the servers after updating assets, etc, etc
[10:41:51] jds: Yeah, it does look that way from the source
[10:43:27] jds: Same version of rails, though our balancers are still ruby 1.9 and the servers are on 2.1
[11:33:11] jds: workmad3: in case you were curious : https://github.com/rails/sprockets-rails/issues/138
[11:33:24] jds: is what was causing my asset digests to be different
[14:06:26] jds: Anyone tried the Large Hadron Migrator?
[14:06:55] jds: I've got an index I need to add to a large table. It's taking over half an hour on my dev machine, I don't fancy locking our production database for that length of time
[14:24:58] jds: Yeah, sadly I'm on mysql

2014-09-01

[08:48:40] jds: In fixtures, are has_one-associations-via-labels completely broken or is it just me?
[08:48:54] jds: The documentation claims they work, but I'm really struggling
[08:53:11] jds: Say I have a 'post' fixture, with 'has_one :image'. In images.yml, I can reference the original post fine by name. If I try to do the reverse association - eg using "image: kitten" in posts.yml to reference a image fixture called 'kitten' - then I get "Unknown column 'image' in 'field list'" when Rails tries to instantiate the fixture

2014-08-11

[11:53:40] jds: Has rails-master had some major changes to url helpers? Anywhere I'm calling something like posts_path(foo: 'bar') I'm getting "wrong number of arguments (3 for 0..1)>", rather than "/posts?foo=bar"
[11:58:05] jds: Ohhh. ssl_requirement is fucking with url_for

2014-08-09

[13:12:48] jds: I'm looking into reducing the amount of stuff that ruby allocates on a given request in our app. Some pages alloc 400k objects, which seems excessive
[13:13:22] jds: I'm starting to suspect haml is a major factor in my problem, it allocs an awful lot of SafeBuffers

2014-07-25

[11:24:08] jds: What does ruby use to determine equality on hash keys?
[11:27:15] jds: Hmph. So I've got a hash with a number keys that are identical, both from the hash value and if I compare them via eql?
[11:27:29] jds: which seems pretty broken
[11:27:41] jds: Yeah, me neither
[11:28:32] jds: https://gist.github.com/jdelStrother/0618d1a430f20b8abeb4
[11:28:48] jds: @details_keys is a hash, where the keys are all hashes themselves.
[11:29:14] jds: Note the latter bunch of keys, which are all identical
[11:29:39] jds: different object_ids, but the hash values are the same, and @details_keys.keys[2].eql?(@details_keys.keys[3]) => true
[11:30:35] jds: Hmm, sounds plausible
[11:31:01] jds: thanks, will give it a try
[11:35:33] jds: apeiros: thanks for the help, I think I might finally be on my way to fixing this memory leak. It's only taken me about 40 hours so far...

2014-07-24

[13:56:29] jds: My 2.1.2 Rails app is leaking like crazy, hitting around 800MB per instance after 20 minutes / 800 requests. It's a pretty big app, and I'm struggling to figure out where the leak is
[13:57:09] jds: I've been using rack-mini-profiler a lot, especially with profile-gc-ruby-head, and it's not actually showing me anything like that much memory. I'm beginning to suspect leakage in C
[13:57:42] jds: centrx: Don't think so, no. Although I guess it might be worth my trying to monkeypath Symbol#initialize & double-check
[16:09:39] jds: centrx: you mentioned symbols wrt my memory leak earlier
[16:10:03] jds: Which I initially dismissed because _obviously_ I'm not dumb enough to go around calling user_input.to_sym all over the place
[16:10:37] jds: But thanks for the suggestion, because it turned out we had user_hash.deep_symbolize_keys buried deep somewhere
[16:10:58] jds: workmad3: Yeah, way back in 2009
[16:11:17] jds: It's gone unnoticed for a long time, but seems like our upgrade to ruby 2.1 made the problem way more obvious

2014-03-24

[09:43:42] jds: Anyone using compass-rails? I like Compass, but it seems like compass-rails breaks quite a lot of the time, and I can't quite figure out what benefits it's providing me over just importing the compass sass directly
[10:42:15] jds: provideal: I've been using oneskyapp.com which is pretty good

2014-03-03

[13:50:51] jds: Heya - anyone tried posting on jobs.github.com ?
[13:52:22] jds: It seems pretty US-centric

2014-02-26

[11:28:44] jds: Routing defaults are expect to work with querystring params, right? With a "get '/foo', defaults: {x:1}" route, if I get "/foo?x=2", then I see x=1 in the params
[11:29:09] jds: request.GET shows me the requested x=2 param, but that's not ideal
[11:29:14] jds: esp. for caching...

2014-02-11

[14:12:46] jds: The caveat with polymorphic associations + STI seems weird - http://api.rubyonrails.org/classes/ActiveRecord/Associations/ClassMethods.html#label-Polymorphic+Associations
[14:12:57] jds: Why doesn't rails just run that logic itself?

2014-02-10

[16:26:05] jds: this. And at this point I'm a bit troubled because I can't seem to find anyone else who's run into the same problem. What am I missing?
[16:26:05] jds: In order for a publicly-visible page to be cachable upstream without Must-Revalidate, it's going to have to be completely independent on cookies & logged-in state, right? So I'm starting to consider including ?logged_in=1 in my all my query strings, or prepending all my url paths with /l/. And so I'm considering monkey patching url_for to auto-include

2014-01-22

[12:16:59] jds: Anyone else struggling to install from rubygems at the moment?
[12:17:38] jds: "gem install blah" sits there for ages before dying with some variant of "no such name (https://api.rubygems.org/quick/Marshal.4.8/nokogiri-1.3.2-x86-mswin32.gemspec.rz)"

2014-01-07

[18:08:46] jds: I upgraded from 3.2.16 to 4.0.2 recently. It seems like less caching is occurring - at least, memcached is serving less data - but it's a fairly large site and I'm struggling to pin down exactly what caches might not be getting used. Any suggestions what I might have missed?
[18:08:58] jds: We're still using action caching, and have added the plugin
[18:16:34] jds: ...huh. Rack::Cache disappeared. Didn't see that in the release notes...
[18:17:12] jds: no, 4.0
[18:19:03] jds: tubbo: Or at least, config.action_dispatch.rack_cache defaulted to true in 3.2 and false in 4.0
[18:19:08] jds: tubbo: yeah, possible
[18:19:26] jds: Guess I ought to generate a fresh app and diff my environment files

2013-12-31

[12:47:39] jds: I've run into a problem with routes in Rails 4, caused by https://github.com/rails/rails/commit/f1d8f2af72e21d41efd02488f1c2dcf829e17783
[12:47:57] jds: If I have routes like https://gist.github.com/jdelStrother/8196261, then the default param is ignored in url_for
[12:48:32] jds: eg url_for(controller:'users', action:'edit', bar:2) returns '/foo/a', regardless of the 'bar' value
[12:48:46] jds: Is there a better way of writing those routes?

2013-12-30

[11:28:34] jds: Ran into some problems with STI scoping in Rails 4. Say you've got STI subclasses Car & Bus. If I do something like User.first.cars.scoping { Bus.create }, then it actually creates a Car. Anyone know if it's a bug or a feature?

2013-12-24

[16:57:25] jds: AR::Base#becomes doesn't work right when saving, ever since the partial-insert support arrived. Is it supposed to?

2013-12-21

[16:33:26] jds: I'm upgrading to Rails 4, and using the protected_attributes gem. How can I write Foo.find_or_create_by(key:'123'), if 'key' is a protected attribute?

2013-12-13

[09:40:34] jds: Duckily: if you want an actual reason not to do it, you'll be re-defining the method every time the view renders, which will blow away ruby's method cache, making your app slower
[09:40:45] jds: +1 for sticking it in a helper

2013-12-12

[18:21:52] jds: Is there a sane way of having an HTTP-cachable page that can be logged into, short of having different urls for logged in users?
[18:22:03] jds: I'm currently sending Must-Revalidate

2013-12-01

[11:57:47] jds: Is there a sane way of detecting CollectionProxy objects? eg if I'm counting classes in ObjectSpace.each_object{...}
[11:57:55] jds: calling .class on them will load the association