NetBeans 6.9 provides a custom Runner class for RSpec to be integrated into the IDE. I'm trying to get my Rails 3 applications specs to be correctly displayed inside NetBeans, but RSpec 2 seems no longer to support custom Runner classes in general.
Any ideas how to get the specs into the IDE anyway?
Just in. Oracle has just announced they are withdrawing support for Rails in future version of NetBeans. Time to start looking at other IDE options.
http://news.ycombinator.com/item?id=2148161
So far (NB 6.9.1) the only way I know to run Rspec2 tests from inside NetBeans is by using rake tasks. But I was not able to make it work with UI Test Runner, because of this and few other problems. So the best way is to avoid invoking UI runner, this can be done in many ways:
Disable it via tools -> options -> miscellaneous -> Ruby
modify project.properties file
give other name to task than 'spec', so naming task as 'rspec' will avoid invoking UI runner
This way you will have just test results in output pan, but it is still usable, because you can click anywhere on stack trace, and NB will take you immediately to that file:line.
There is one thing left, auto generated by NB Rakefile has not valid task (for Rails projects, there is NO such problem), to make it work one needs at least:
require 'rspec/core/rake_task'
Rspec::Core::RakeTask.new(:rspec)
I know this is not what you are expecting but you might want to check RubyMine3 out it comes out of the box, you do need to buy a licence but at least you can check it out in the 30 day trial
I am using RVM.
And at the minimum I wanted to be able to run my Ruby 1.9.2 / Rails 3 / RSpec 2 specs from inside the IDE and be able to click on stack traces for Netbeans to open the right files and lines.
I found a work-around for that:
Put somewhere in the project a ruby file that shells out to run the spec suite.
E. g. my ruby file has the following content:
system <<EOF
time ~/.rvm/wrappers/ruby-1.9.2-p290#default/rspec --drb spec
EOF
Change the ruby version and gemset as you need it.
The major limitation:
I cannot just run only a single spec. For that I have to change the "spec" parameter to the target spec file (which isn't such a big deal though).
Netbeans is fairly sluggish running specs (using rspec1 here), would recommend running specs from command line.
Related
I like to create rspec test from working rails 5 app.
It' could be template to work on.
For example:
-scaffold will create test files template which is nice.
-Or simplecov to help increase test ratio (if I understand correctly)
-So if any tools that could run through each line of controller method and re-create each expected/result put in the test file.
Are there any gem or solution ?
Why ask this question?.
I understand this is not a purpose of test. However test is to save time in future and now as well. Many rails app don't have test, and to go back each line of running code will cost again, assuming the app is good at this stage. If we can have all test at this point and use it to control / run for future development that would be good
I found another gem that answer my question
rspec-kickstarter
But to use with Rails 5 it need to edit the path that created.
I'm trying to create a test folder for my rails project using cucumber-rails generator (the gem is here: https://github.com/cucumber/cucumber-rails ).
The problem is that when I run "rails generate cucumber:install" it creates a folder in my project named "features" with the cucumber files inside that folder, which is - in my perspective - not very organized.
I would like all the cucumber files to be inside a folder named "test". Is there a way that allows me to do that? I tried to run the generator inside the "test" folder but it didn't seem to work.
Anyone knows how to do this? Thanks a lot!
There are essentially 3 test libraries in Ruby and corresponding common assumptions:
test is expected to contain Test::Unit or MiniTest tests
spec is expected to contain RSpec tests
features is expected to contain cucumber tests
You should follow these assumptions or you may end up having several troubles in the long run.
In any case, to answer your question, what you want to do is not possible. The features folder is hard-code in the generator and in the cucumber task file. Therefore, unless you want to introduce several hacks and workaround, you'll have to follow the conventions.
I think there is a pretty straight forward solution for this:
In an issue it is pointed out that we can point to different paths for our features in the cucumber.yml
default: features
// becomes
default: test/features
But, then it can't find the step_defintions if you also add your custom path as a required flag then it can pick up the step_defintions as well.
default: features
// becomes
default: test/features -r test/features/step_definitions
if you don't want to use cucumber.yml you can use the cucumber cli to take care of it:
bundle exec cucumber test/features -r test/features/step_definitions
NOTE: I am using this in a docker environment and this works fine for both local and docker implementations.
I am able to right-click on any of my 3 spec/models but when I right click the spec/models folder and select 'run all tests in models' I get 'Unable to attach test reporter to test framework'.
The first line of my model tests is: require 'spec_helper.rb'
I posted this answer for another question. It may help with this one as well:
On my system, the problem was the "redgreen" gem. It automatically colors the progress marks (dots) and the Test::Unit summary messages based on success or failure. There must be something in RubyMine that's trying to parse the Test::Unit results (the output of "rake test", I imagine), and it's choking on the ANSI sequences.
I commented out "require 'redgreen'" in my test/test_helper.rb, and the problem went away. I really like redgreen for executing "rake test" from the shell, though, so I put this in test_helper to make it work for both "rake test" and within RubyMine:
require 'redgreen' if $stdin.tty?
It may not be redgreen that's causing your problem, but be suspicious of anything which might create unconventional Test::Unit output.
Good luck!
I had the same issue and had to add the following to /etc/launcd.conf (I had to create this file as well):
setenv DYLD_LIBRARY_PATH /usr/local/mysql/lib/
and reboot. There are other ways to do it as well (a plist file, adding this environmental variable to RubyMine, etc… but this was most reliable. Of course, this assumes that you use MySQL.
The answer in the end for my setup was to fiddle around with the IDE settings including the ruby version, making sure it was 1.9.2 and the directories referenced in the config. screens were correct. This plus some restarts resolved the issue.
About 6 months ago I switched from TextMate to MacVim for all of my development work, which primarily consists of coding in Ruby, Ruby on Rails and JavaScript.
With TextMate, whenever I needed to run a spec or a test, I could just command+R on the test or spec file and another window would open and the results would be displayed with the 'pretty' format applied. If the spec or test was a lengthy one, I could just continue working with the codebase since the test/spec was running in a separate process/window. After the test ran, I could click through the results directly to the corresponding line in the spec file.
Tim Pope's excellent rails.vim plugin comes very close to emulating this behavior within the MacVim environment. Running :Rake when the current buffer is a test or spec runs the file then splits the buffer to display the results. You can navigate through the results and key through to the corresponding spot in the file.
The problem with the rails.vim approach is that it locks up the MacVim window while the test runs. This can be an issue with big apps that might have a lot of setup/teardown built into the tests. Also, the visual red/green html results that TextMate displays (via --format pretty, I'm assuming) is a bit easier to scan than the split window.
This guy came close about 18 mos ago: http://cassiomarques.wordpress.com/2009/01/09/running-rspec-files-from-vim-showing-the-results-in-firefox/ The script he has worked with a bit of hacking, but the tests still ran within MacVim and locked up the current window.
Any ideas on how to fully replicate the TextMate behavior described above in MacVim?
Thanks!
There is a plugin called vim-addon-background-cmd that can allow you to run tasks in the background instead of locking up the vim interface. You would have to create the call to run through the background command. See the docs for more information on how to do that.
A few months back I was looking for this same exact thing. Then I discovered autotest with rspec. Now I keep a separate terminal window open which shows my last run tests. If I change any relavent code files my tests are automatically run for me (the files are watched and if they change the tests run).
If you want the same autotest type behavior in a non-rails project you can look at the watchr gem. It's functionality is similar to autotest but you can use it in ANY framework.
What tools do you use for automated code sanity checks and adhering to the coding conventions in your Ruby apps? How do you incorporate them into your process? (I mean tools like roodi, reek, heckle, rcov, dcov, etc.)
I'd suggest taking a look at RuboCop. It is a Ruby code style checker based on the Ruby Style Guide. It's maintained pretty actively and it's based on standard Ruby tooling (like the ripper library). It works well with Ruby 1.9 and 2.0 and has great Emacs integration.
The metric_fu gem might be perfect for what you need. From it's README: "Metric-fu is a set of rake tasks that make it easy to generate metrics reports. It uses Saikuro, Flog, Rcov, and Rails'
built-in stats task to create a series of reports. It's designed to integrate easily with CruiseControl.rb by placing files in the Custom Build Artifacts folder." Since they converted it to a gem, it works with non-Rails applications as well. I'll bet you could add hooks for other tools as well.
There was some good discussion on this topic on the On-Ruby blog recently. For my personal development process I build quality tools into my tests, but only after all other tests have run. So I have a top-level rake task that looks something like this:
desc 'Runs all unit tests, acceptance tests and quality checks'
task 'test' => ['test:spec', 'test:features', 'test:quality']
I allow myself to commit if the last suite "fails", but I do try to get them to zero at least once each day.