My question is similar to this , where they want to override rake spec's output format. The resolution to that question is to use the .rspec config file, which is limiting. I would like this to be a command line argument because I want this to vary on different machines.
The rspec executable has the -f option be defining format. rake spec has -f defining a rakefile. rake spec --format is invalid. Is this an oversight in rake spec? "Format" really isn't an option?
ANSWER: I'm self answering my question here. rake spec will take the SPEC_OPTS environment variable.
rake spec SPEC_OPTS="--format documentation"
A cleaner way todo this in Rakefile:
RSpec::Core::RakeTask.new(:spec) do |t|
t.rspec_opts = '--format documentation'
end
Related
The rspec command comes with several options you can use to customize
RSpec's behavior, including output formats, filtering examples, etc.
For example:
$ rspec path/to/spec_file.rb
$ rspec --example 'first example'
$ rspec --tag type:special
$ rspec -P "**/*_test.rb"
How can I do the same thing with rake spec (with full Rspec options).
My Rakefile:
require 'bundler/gem_tasks'
require 'rspec/core/rake_task'
task default: :spec
RSpec::Core::RakeTask.new(:spec)
I have been google but didn't find any complete answer for that. Thanks.
Command line arguments can be passed automatically to the ENV hash.
For example:
From command line: FOO=BAR rspec spec/*spec.rb
Inside RSpec: puts ENV["FOO"] # => "BAR"
In your Rakefile, you can use backticks to call the shell command.
You can do it but it requires some changes.
First of all you need to undefine already defined spec task if its present, then define it again.
Or use other name, like spec_with_opts. Though I went through renaming.
in Rakefile
Rake::Task["spec"].clear
RSpec::Core::RakeTask.new(:spec) do |task, args|
task.rspec_opts = ENV['RSPEC_OPTS'] if ENV['RSPEC_OPTS'].present?
task.pattern = ENV['RSPEC_PATTERN'] if ENV['RSPEC_PATTERN'].present?
task.exclude_pattern = ENV['RSPEC_EXCLUDE_PATTERN'] if ENV['RSPEC_EXCLUDE_PATTERN'].present?
end
task default: :spec
So it now can be run this way:
rake spec RSPEC_PATTERN=path/to/spec_file.rb
rake spec RSPEC_OPTS="--example 'first example'"
rake spec RSPEC_OPTS="--tag type:special"
This one wont work, you would need to use RSPEC_PATTERN
rake spec RSPEC_OPTS="-P '**/*_test.rb'"
You can find other options that can be defined in source file:
https://github.com/rspec/rspec-core/blob/master/lib/rspec/core/rake_task.rb
I have a suite of RSpec tests I want to group under the following hierarchy:
tests/
featA/
t1.rb
t2.rb
featB/
t3.rb
but when I run
$ rspec tests
I get the following:
rspec tests
No examples were matched. Perhaps {:unless=>#<Proc:0x00007f318919cc08#/usr/lib/ruby/gems/1.8/gems/rspec-core-2.5.1/lib/rspec/core/configuration.rb:51>, :if=>#<Proc:0x00007f318919cdc0#/usr/lib/ruby/gems/1.8/gems/rspec-core-2.5.1/lib/rspec/core/configuration.rb:50>} is excluding everything?
Finished in 0.00003 seconds
0 examples, 0 failures
I feel like I'm going mad, but there doesn't seem to be a way to get RSpec to recursively glob for test files? Does this functionality exist?
EDIT:
I have a workaround by doing this:
$ rspec `find tests -name "*.rb"`
but I suspect I shouldn't have to. Am I right?
You've exposed an oversight on my part! In rspec-1, you could say this:
spec test --pattern "**/*.rb"
But the --pattern option is missing in rspec-2. I've just added it (in development) and it will be included in the rspec-2.6.0 release.
I usually manage running RSpec on my specs via rake. The relevant portion of my Rakefile looks something like this:
require 'rspec/core/rake_task'
RSpec::Core::RakeTask.new(:spec) do |t|
t.rspec_opts = ['--color', '-f progress', '-r ./spec/spec_helper.rb']
t.pattern = 'spec/**/*_spec.rb'
t.fail_on_error = false
end
Now rake spec runs RSpec with the appropriate options; you'll need to change t.pattern to match the specs you want to run.
Be sure to check out the RSpec2 site for more information.
How do I set global configuration for RSpec in Ubuntu.
Specifically so, --color and --format specdoc stay turned on, across all my projects (ie every time I run rspec anywhere).
As you can see in the docs here, the intended use is creating ~/.rspec and in it putting your options, such as --color.
To quickly create an ~/.rspec file with the --color option, just run:
echo '--color' >> ~/.rspec
One can also use a spec_helper.rb file in all projects. The file should include the following:
RSpec.configure do |config|
# Use color in STDOUT
config.color = true
# Use color not only in STDOUT but also in pagers and files
config.tty = true
# Use the specified formatter
config.formatter = :documentation # :progress, :html,
# :json, CustomFormatterClass
end
Any example file must require the helper to be able to use that options.
In your spec_helper.rb file, include the following option:
RSpec.configure do |config|
config.color_enabled = true
end
You then must require in each *_spec.rb file that should use that option.
If you use rake to run rspec tests then you can edit spec/spec.opts
http://rspec.info/rails/runners.html
Or simply add alias spec=spec --color --format specdoc to your ~/.bashrc file like me.
One thing to be aware of is the impact of the different ways of running RSpec.
I was trying to turn on the option with the following code in spec/spec_helper.rb -
Rspec.configure do |config|
config.tty = $stdout.tty?
end
calling the 'rspec' binary directly - or as 'bundle exec rspec' and checking $stdout.tty? will return true.
invoking the 'rake spec' task - or as 'bundle exec rake spec' - Rake will invoke rspec in a separate process, and $stdout.tty? will return false.
In the end I used the ~/.rspec option, with just --tty as its contents. Works well for me and keeps our CI server output clean.
With Test::Unit, I can run:
ruby path/to/test.rb --name=test_name_that_i_want_to_run
Thus far, I have not been able to figure out how to do this with test/spec specifications. I am wondering if the way that specifications are automatically named does not allow me to do something like this.
Take the following spec for example:
require 'rubygems'
require 'spec'
describe 'tests' do
it 'should be true' do
1.should == 1
end
it 'should be false' do
1.should_not == 2
end
end
You can execute a single spec by using the -e flag and providing the portion specified by the it block. e.g. ruby my_spec.rb -e 'should be false'
After contacting the gem maintainer, Christian Neukirchen, I found out how to do this, so I am documenting it here for future reference.
specrb path/to/test.rb --name ".*should behave this way.*"
I needed to use the specrb test runner, an extended version Test::Unit's test runner, rather than just the ruby command.
You can also do this with the ruby command:
ruby path/to/test.rb -n "/should behave this way/"
I want to be able to run a single spec file's tests — for the one file I'm editing, for example. rake spec executes all the specs. My project is not a Rails project, so rake spec:doc doesn't work.
Don't know if this matters, but here is my directory structure.
./Rakefile
./lib
./lib/cushion.rb
./lib/cushion
./lib/cushion/doc.rb
./lib/cushion/db.rb
./spec
./spec/spec.opts
./spec/spec_helper.rb
./spec/db_spec.rb
Or you can skip rake and use the 'rspec' command:
bundle exec rspec path/to/spec/file.rb
In your case I think as long as your ./spec/db_spec.rb file includes the appropriate helpers, it should work fine.
If you're using an older version of rspec it is:
bundle exec spec path/to/spec/file.rb
The raw invocation:
rake spec SPEC=spec/controllers/sessions_controller_spec.rb \
SPEC_OPTS="-e \"should log in with cookie\""
Now figure out how to embed this into your editor.
This question is an old one, but it shows up at the top of Google when searching for how to run a single test. I don't know if it's a recent addition, but to run a single test out of a spec you can do the following:
rspec path/to/spec:<line number>
where -line number- is a line number that contains part of your test. For example, if you had a spec like:
1:
2: it "should be awesome" do
3: foo = 3
4: foo.should eq(3)
5: end
6:
Let's say it's saved in spec/models/foo_spec.rb. Then you would run:
rspec spec/models/foo_spec.rb:2
and it would just run that one spec. In fact, that number could be anything from 2 to 5.
You can also use the actual text of the *e*xample test case with -e !
So for:
it "shows the plane arrival time"
you can use
rspec path/to/spec/file.rb -e 'shows the plane arrival time'
./scripts/spec path/to/spec/file.rb -e 'shows the plane arrival time'
no need for rake here.
from help (spec -h):
-l, --line LINE_NUMBER Execute example group or example at given line.
(does not work for dynamically generated examples)
Example: spec spec/runner_spec.rb -l 162
To run all of your rspec files: rspec
note: you must be in the root of your project
To run one rspec file: rspec 'path_to/spec.rb'
note: replace 'path_to/spec.rb' with your path. Quotation marks optional.
To run one rspec test from one file: rspec 'path_to/spec.rb:7'
note: :7 is the line number where the test starts
If you installed rspec as a plugin rather than as a gem, then you won't have the spec executable.
At any rate, All you need to do is run the file using ruby. The rspec code is clever enough to run the tests for you.
eg:
ruby myclass_spec.rb
http://github.com/grosser/single_test lets you do stuff like..
rake spec:user #run spec/model/user_spec.rb (searches for user*_spec.rb)
rake test:users_c #run test/functional/users_controller_test.rb
rake spec:user:token #run the first spec in user_spec.rb that matches /token/
rake test:user:token #run all tests in user_test.rb that match /token/
rake test:last
rake spec:last
Ruby 1.9.2 and Rails 3 have an easy way to run one spec file:
ruby -I spec spec/models/user_spec.rb
Explanation:
ruby command tends to be faster than the rake command
-I spec means "include the 'spec' directory when looking for files"
spec/models/user_spec.rb is the file we want to run.
Although many great answers were written to this question, none of them uses the Rspec tags approach.
I use tags to run one or more specs in different files -- only those related to my current development task.
For example, I add the tag "dev" with the value "current":
it "creates an user", dev: :current do
user = create(:user)
expect(user.persisted?).to be_truthy
end
then I run
bundle exec rspec . --tag dev:current
Different tags/values can be set in individual specs or groups.
I was having trouble getting any of these examples to work, maybe because the post is old and the commands have changed?
After some poking around I found this works:
rspec spec/models/user_spec.rb
That will run just the single file and provides useful output in the terminal.
specky.vim
Alternatively, have a look at autotest.
Running autotest in a command window will mean that the spec file will be executed whenever you save it. Also, it will be run whenever the file you are speccing is run.
For instance, if you have a model spec file called person_spec.rb, and a model file that it is speccing called person.rb, then whenever you save either of these files from your editor, the spec file will be executed.
Lets say, you're running test for creating todo. You can always run that specific todo spec code using the file crete_spec.rb file as below.
rspec/spec/features/controller/spec_file_name.rb
Example:
Creating rspec spec/features/todos/create_spec.rb
Editing rspec spec/features/todos/edit_spec.rb
Deleting rspec spec/features/todos/destroy_spec.rb
If you want to run all the specs in one single short.
rspec
If you want to run all the specs in a specific controller user this.
rspec/spec/feaures/controller_name
Example: rspec/spec/features/todos
Hope it gives you more understanding!
And you can run specific line into your test file
rspec spec/models/model_spec.rb:47