I need to run a bunch of ruby scripts that I have written on a server that I don't have sudo access to.
On my own machine, I have installed a bunch of gems using 'sudo gem install ..' and used them in my code..
Is there any mechanism which would let me use these gems without formally installing them on a remote machine?
You can, but it's tricky.
First, install them using the --install-dir option, i.e.:
gem install gem_name --install-dir /some/directory/you/can/write/to
Second, make sure you have a .gemrc file in your home directory that looks something like this:
gemhome: /some/directory/you/can/write/to
gempath:
- /some/directory/you/can/write/to
- /usr/local/lib/ruby/gems/1.8
gemhome is where gems should look first when seeking a gem. gempath is all the paths it should check in when seeking a gem. So in the .gemrc above, I'm telling my code to look first in the local directory, and if not found, check the system gem directory.
Third, be aware that some code - even code within gems - can make assumptions about where gems are located. Some code may programmatically alter gempath or gemhome. You may need to "alter it back" in your own code.
There's not a lot (read: no) documentation on how to do that - the best way to figure it out is to read the tests that are included with the RubyGems source. Here's how I hack the gem paths in a rake task to point to my frozen version of capistrano:
Gem.use_paths(Gem.dir, ["#{RAILS_ROOT}/vendor/gems"])
Gem.refresh # picks up path changes
Related
How do I configure bundler so that when I run bundle install it looks for gems under /my/custom/path first and if it doesn't find there then try to fetch them from a ruby gem remote repository hosted under https://a.nice.host and downloads those into ./local/relative/path (relative to cwd for example). I would like to avoid the bundler looking at default gem installation system path or rubygems.org
The syntax for sourcing a gem from a local folder is:
gem 'some-gem-name', path: '/my/custom/path'
And the syntax for specifying a custom source is:
gem 'another-gem-name', source: 'https://a.nice.host'
And to install gems into a specific local folder, you can run:
bundle install --path ./local/relative/path
Now, that's probably all the tools you need, in truth... (And in fact, especially for that last requirement, you may instead wish to look into rvm gemsets, or using bundle install --deployment.)
But you did also ask about "looking in a local folder first, and only falling back to a remote source if it doesn't exist". That's quite an odd requirement (usually you'd only want to explicitly opt-in to fetching gems from a local path?!), but to answer this question as you've asked it...
A Gemfile is literally just ruby code! So you can define this logic using... You guessed it, ruby! For example:
if File.exists?('/my/custom/path')
gem 'some-gem-name', path: '/my/custom/path'
else
gem 'some-gem-name', source: 'https://a.nice.host'
end
If this (unusual) pattern needs to be repeated in multiple places, you could wrap it into some helper method.
For more information on the configuration options of bundler, please see the documentation.
I have a project that depends on Ruby to do something. I need to tell these people to install bundler, create a Gemfile (or update an existing one) and then run bundler install.
To be very clear, these people do not care about Ruby, they don't know what Ruby is and they don't need to know what Ruby is.
Currently my documentation is:
Run this command in terminal:
gem install bundler
Create a new file name Gemfile and add these contents:
source 'https://rubygems.org'
gem 'lightning_sites'
Or if there is already a Gemfile then edit that file and add the line gem 'lightning_sites' at the bottom.
Go back to the terminal and run:
bundle install --path vendor/bundle
I would like to replace the documentation for step 2 and preferable replace it with a command line. Is there a tool that ships by default with Ruby or bundler that accomplishes this?
If you want to avoid bundler you need to force-bundle all your dependencies inside your application. This is only really practical if none of your dependencies have compiled extensions, so if they're all pure Ruby you'll be able to do it.
What you end up doing is a bundle install --path gems/ for example, then package up everything including that directory as a deployable application. You may want to make a script that performs this step and creates a .zip file of the final result for distribution purposes.
This is a heavy-handed approach, so it's best to do this only if absolutely necessary.
You don't have to use Bundler to install gems, Ruby provides the gem command for installing gems individually.
You could simply run: gem install lightning_sites --install-dir lightning_sites and in whatever Ruby script is using the gem, programatically modify your GEM_PATH using Gem.paths before the require statement to include that install directory.
I am trying to contribute to a Ruby gem and I don't understand how to test a local gem without using a globally installed gem.
The gem I want to contribute to is a command line interface gem. I clone the gem into a directory then cd into that directory. However, when I run commands in the terminal when I'm in the cloned project directory it still uses the global gem. I've even run
gem uninstall gemname
then while inside the newly cloned gem directory I redo
gem install gemname.
No matter what changes I make to the gem, I can't see the results or what my contributions are doing because it's always running the global gem.
When I do try to type a command line command that is supposed to interact with the gem while in the cloned gem directory I get:
-bash: ~/.gem/ruby/2.1.0/bin/githubrepo: No such file or directory
I've done a ton of research but I'm just not getting it. Help?
gem install gemname will look for a .gem file in the current directory. If not found it will look for it on the web.
gem install --local /path/to/your/gemname.gem will allow you to target a particular directory. You may need to gem build gemname.gemspec first, so it has your changes.
Instead of doing this, I would write tests in the gem directory itself. It's likely that when running code in there, you can simply require 'gemname' in Ruby to get the gem functionality.
If it's a well-written gem, it should have tests already. They will most likely be in a directory called test or spec. Have a look at these tests and try to carry on in that style to test your changes. This will make your code changes far far more likely to be accepted as a pull request.
Is it possibile to have something like a 'local' gem repository for Ruby?
I'm working on a custom Linux distribution without admin rights. Ruby is installed on the machine (v.1.8.7) but apparently no 'gem' or 'bundle' or whatever are installed. I need to use some Ruby gems like Nokogiri.
Is it possible to use Ruby gems without installing them through gem install?
Yes. Any gem can be used standalone. You just have to either download the source from github, or download the gem and extract its contents manually.
After you've done that you have to add the lib folder of the gem into the load path ($:) of Ruby. For example:
$:.unshift(File.expand_path("nokogiri-1.6.1/lib"))
require 'nokogiri'
Assuming you are running Ruby in the current directory and the Nokogiri source is in the folder nokogiri-1.6.1.
But remember that first you have to do the same with all Nokogiri prerequisites. I.e. all the libraries Nokogiri depends on.
Another option, which is what I would do, is to install RubyGems in your home directory. That can get a little bit tricky though, but it's possible by downloading the RubyGems package and running something like:
ruby setup.rb --prefix=/home/my_user_name/rubygems
Then you need to set up the ENV variables GEM_HOME and GEM_PATH to point to a location in your home directory where you want all your gems to be installed. See "What's the difference between GEM_HOME and GEM_PATH?" for a description of what they do.
You will also need to add something like /home/my_user_name/rubygems/bin to your PATH so that the gem command can be found by the shell.
Ruby gem names are really throwing me off. For example if install a gem called jade, a wrapper script is placed at /usr/bin/jade1.9 on some systems and at /usr/bin/jade on others. I'm curious if anyone has a recommendation on how I can use jade in scripts without writing some hackish code that guesses the correct gem name. The naming also makes it difficult to write documentation for less savvy users. Is there a simple way of making gems install with same name on all systems?
To elaborate a bit:
gem install jade #jade is something I made up
ls /usr/bin/|grep "jade"
> jade1.9
When I want to execute:
jade --dosomething
I actually have to run
jade1.9 --dosomething
I've noticed certain gems such as rake are installed at both /usr/bin/rake and /usr/bin/rake1.9 with neither being a symlink and both files having the same md5sum.
I am trying to understand why gems get the ruby version appended on some systems and how I can make install without the trailing version number.
A list of some example of gems that install an executable:
ls /usr/bin/|grep "1.9"
amalgalite-pack1.9
bundle1.9
crate1.9
erb1.9
gem1.9
irb1.9
minitar1.9
rake1.9
rdoc1.9
ri1.9
rlock1.9
ruby1.9
rubyscript2exe1.9
testrb1.9