Let's say in a Rails app you have some gems that you use in your app (we'll call them "primary gems") and you have vendored them for portability.
Let's say that those "primary gems" also require gems of their own - we'll call these "secondary gems".
When you are setting up your environment.rb, you have to say:
config.gem 'primary-gem'
for any of the gems you are using directly.
But, do you also need to say . . .
config.gem 'secondary-gem'
even if you are not using that gem explicitly in your app?
Or is it simply enough to include the gem in your vendor/gems directory for it to get picked up by your app?
At deploy time rails knows about your dependencies, so if you want to freeze your gems then you can run
rake gems:unpack:dependencies
to freeze them into the vendor directory.
At runtime however it's the gems job to load it's dependencies, and usually the gems do this, so a config.gem 'primary' should work.
No, you don't or at least you shouldn't. Each GEM specification should include it's own list of dependencies. When primary gem is installed, RubyGems automatically will install each gem dependency on cascade.
In other words, if A requires B that requires C+D, you only need to write
config.gem 'A'
When the command
gem install A
is run, RubyGems will resolve all the dependencies and install them.
You can view all A dependencies running (from a Rails project)
rake gems
Sometimes, a GEM author may forget to include some GEM dependencies in the specification. In this case you should specify them in your environment.rb to force the application to install them. Off course, it's also a good idea to contact the GEM maintainer so that it can fix the problem.
Related
I'm creating a gem (let's call it mygem) that is essentially a Sinatra server intended to be mounted within Rack based apps.
Inside my gem's gemspec file, I have the following:
gem.add_dependency 'kss'
And Inside my gem's Gemfile, I have the following
source 'https://rubygems.org'
gemspec
gem "kss", :path => "/Users/me/code/kss"
Now when running the server from within mygem's folder, this works exactly as expected: instead of fetching out for the kss dependency, it will look on my local drive and load that version.
The problem comes in when I add mygem to a Rails test app Gemfile. In my Rails test app Gemfile, I have the following line:
gem "mygem", :path => "/Users/me/code/mygem"
I would expect, upon a bundle install, that Bundler would load mygem and its dependencies; but for the kss dependency, instead of loading the local dependency, Bundler actually does fetches out to rubygems to find and load it. I'm assuming because in this case, it's only reading from the gemspec line and not including my dependency override.
Is there anything I can do to fix this behavior? I'd very much like to be able to run and test this stuff locally, but Bundler doesn't seem to recognize dependency overrides from a higher level app.
I'm completely open to any suggestions or changes if I'm going about this the wrong way.
Dependencies listed in your gemspec will become dependencies of the implementing application, while dependencies defined in the Gemfile will not become dependencies of the implementing application. I think you should be able to simply adjust the Rails test app Gemfile to be:
gem "kss", :path => "/Users/me/code/kss"
gem "mygem", :path => "/Users/me/code/mygem"
I have strange old buggy project on Rails 2.
It have gem's dependencies in config/environment.rb like
config.gem "andand"
config.gem "json"
config.gem "chronic"
config.gem "mini_fb"
all those gems are located in vendor/gems/
andand-1.3.3/
chronic-0.6.7/
json-1.7.3/
mini_fb-1.1.7/
rbet-1.0.3/
redis-3.0.1/
responsys_client-0.0.1/
but when i start unicorn server with this app it always complain that it can't find this gems. Why?
UPDATE
After building and installing gem from vendor/gems rails still complain about it.
I have tweake mini_fb gem into custom mini_fb_custom gem. Changed all references in gemspec and other files from mini_fb to mini_fb_my, installed it and it is shown in gem list as mini_fb_my. But it fails to load from config/environment.rb and complains that
Missing these required gems:
mini_fb_my >= 0
maybe i should rename lib/mini_fb.rb to lib/mini_fb_my.rb
i'll check it.
UPDATE 2
Yes, renaming files rocks!
You still need to install them from those folders, or unicorn will not know where to look for them.
Just install the gems from that directory and unicorn should pick them up.
UPDATE
You can install your gems locally with this command
gem install --local vendor/gems/gem/gem-name.gem
On more recent versions of rails you just specify path on the Gemfile
gem "gem-name", path: "path/to/gem"
My advice: replace the obsolete gem configuration with bundler (it works fine with rails 2, there should be a tutorial for rails 2 available on their website).
Configuration through gem command, freezing gems, etc. is just pain in the a** and it seemed kinda buggy to me when I'd used it (long time ago).
for a Ruby standalone script what Rails like deployment features such as Gemfile / "bundle install" etc
that is assuming you are developing a Ruby script that you want to test and then deployment, and perhaps ship to others, what Rails like deployment approach would you use for say:
a) GEM - marking GEM requirements & having them installed as required - e.g. Rails "Gemfile" where you mark what gems you need and then "bundle install" to install them
b) File Require - automatically loading *.rb files if they are in your script directory (I'm thinking of in Rails where if you put a class file in the apps/model directory Rails automatically load/require's the file for you)
depends on whether it's a tool you expect people to use on every host they find it on or not. also depends on whether the tool can be shared with pubic repos.
if it just has to work, without worrying about whether you've installed the gems via bundler already or not, you can use something like the following from within your standalone script to install gems if not already present (be mindful of system vs. user ruby):
#!/usr/bin/env ruby
require 'rubygems'
def install_gem(name, version=Gem::Requirement.default)
begin
gem name, version
rescue LoadError
print "ruby gem '#{name}' not found, " <<
"would you like to install it (y/N)? : "
answer = gets
if answer[0].downcase.include? "y"
Gem.install name, version
else
exit(1)
end
end
end
# any of the following will work...
install_gem 'activesupport'
install_gem 'activesupport', '= 4.2.5'
install_gem 'activesupport', '~> 4.2.5'
require 'active_support/all'
...
In my humble opinion, a gem is the way to go. Bundler makes it easy to get started; it starts a skeleton for you when you run the command…
bundle gem <GEM_NAME>
Take a look. As long as you specify your dependencies in your gem's .gemspec file, and somebody installs your packaged gem (they won't need bundler, just RubyGems' gem command), the dependencies will be installed as gems along with it.
I've been told that doing:
config.gem 'tzinfo'
doesn't obviate the need to require 'tzinfo'. Is this true of all gems? If yes, what exactly does adding config.gem WHATEVER do?
config.gem should cause the gem to be automatically required. You should not need to do a manual 'require' call.
config.gem
Tells Rails to load this gem automatically
Tells Rails that this gem is needed for the application, so that rake gems:install will install it
The :source option can tell rails to get it from a nonstandard repository
The :lib option can tell rails to load a non-standard file from the gem (i.e. something not named after the gem itself)
If i'm correct, during the environment initialization 'config.gem' allows your app to setup and require GEM dependencies from within the app, without the need to have to install them manually. (As we did before) By calling "config.gem tzinfo" as you did above, it automagically requires the gem across the app. This helps when you deploy to an external server and need to prepare the app along with necessary gems, etc. You can then run RAKE GEMS:INSTALL and rails will pull in all your gems and require them.
A thing to note though is that if you DO NOT want a gem to be required across your app. Then add ":lib => false" after config.gem i.e (config.gem 'tzinfo' :lib => false).
In some cases, (I followed your link) if you're getting an uninitialized gem, and you've manually installed it. Make sure that the config.gem ":lib" directory matches with the correct :lib directory of the gem. I.E a gem may be packaged and installed as "nlewis-supergem", however I may need to point the lib at "supergem". i.e "config.gem "nlewis-supergem" :lib=>"supergem". It all depends on how some people package their gem and the corresponding libraries.
A quick tip is instead of installing manually always install the gem via "config.gem" and then rake GEMS:INSTALL to catch any wierd errors before deployment.
Hope this helps.
In my rails application I once used authlogic-oid and ruby-openid. Now I want to get rid of them and I removed both gems and also their config.gem lines from my environment.rb.
Although my application works, I can't do any database migrations because I get a "Missing these required gems" error. Also if I run rake gems:install these gems are re-installed.
Where are the references to the gems stored?
The standard way to define a gem dependency is in the environment configuration. It usually takes place in the environment.rb file for any environment, but some gems might be specified also per-environment. Check the environment files in config/environments.
Also make sure some file doesn't include the gem with the classic RubyGems gem command.
Finally, check these gems are not required by other gems or plugins used by your application.