# Rake test:functionals -\> Task not supported by '' **URL:** \ **Category:** Rails\ **Created:** [March 6, 2009, 9:37pm UTC](https://www.ruby-forum.com/t/rake-test-functionals-task-not-supported-by/162214 "2009-03-06T21:37:09Z")\ **Posts on this page:** 3\ **Page:** 1
**Author:** ![fritza](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/f/5f9b8f/32.png) [@fritza](https://www.ruby-forum.com/u/fritza)\ **Post date:** [March 6, 2009, 9:37pm UTC](https://www.ruby-forum.com/t/rake-test-functionals-task-not-supported-by/162214/1 "2009-03-06T21:37:09Z")
I have a Rails project in which I’ve been using “rake test” tasks with great success. At some point (there were a few changes in the interim), rake started aborting test tasks. Here is the dialog: === $ rake test:functionals (in /Users/fritza/myproject) rake aborted! Task not supported by ‘’ # (See full trace by running task with --trace) The separate suites (e.g., “ruby test/functional/mycontroller\_test.rb”) work, reporting failures or not as usual. rake test:units, test:functionals, and test:integration all fail the same way. “rake test” successfully runs the functional and integration tests, but aborts with the same message when it attempts the unit tests. Among my changes were some shifting-around between a local SQLite database and a MySQL server, but I’d think that wouldn’t be a problem if the individual suites (and two subtasks of test) run normally. The problem persists no matter which DB I put in database.yml. I think the database issue is a red herring. “rake RAILS\_ENV=test db:migrate” was run, with the expected progress output, before the rake test invocations. Re-running the migration outputs the “(in…)” line and nothing else, as expected. Google turns up little, and nothing with an empty supported-by string. Apparently I’ve bent something, but I don’t know what. Will someone please help me get back on track? ``` — F ``` ---
**Author:** ![fritza](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/f/5f9b8f/32.png) [@fritza](https://www.ruby-forum.com/u/fritza)\ **Post date:** [March 6, 2009, 9:52pm UTC](https://www.ruby-forum.com/t/rake-test-functionals-task-not-supported-by/162214/2 "2009-03-06T21:52:15Z")
Fritz A. wrote: > I have a Rails project in which I’ve been using “rake test” tasks with > great success. At some point (there were a few changes in the interim), > rake started aborting test tasks. > … > Among my changes were some shifting-around between a local SQLite > database and a MySQL server > … Okay, posting a problem publicly gets you an answer very quickly. My database.yml file looked like this: === test: test\_mysql test\_mysql: adapter: mysql #etc # test\_local: adapter: sqlite3 database: db/test.sqlite3 timeout: 5000 Rails itself knows that when it finds a String as the value of an environment key, it should use that String as the key for another lookup in the YAML, in search of a Hash it can use for configuration. The rake task, however, stops at the first lookup. It can’t pull an “adapter” value from the String, and therefore reports that the task is not supported by this empty-named adapter. I think there may be a better way to parameterize YAML values, but the server for the YAML in Five Minutes tutorial, where I remember seeing it, is down. So this question is answered, and I give thanks/apologies for your effort. ``` — F ``` ---
**Author:** ![fritza](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/f/5f9b8f/32.png) [@fritza](https://www.ruby-forum.com/u/fritza)\ **Post date:** [March 6, 2009, 9:53pm UTC](https://www.ruby-forum.com/t/rake-test-functionals-task-not-supported-by/162214/3 "2009-03-06T21:53:44Z")
On Mar 6, 8:37 pm, Fritz A. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > Google turns up little, and nothing with an empty supported-by string. > > Apparently I’ve bent something, but I don’t know what. Will someone > please help me get back on track? If the problem is a source code change rather than an environment change, have you tried revert to previous source to see if you can isolate the change that broke this? (if you are using git then git bisect is designed for this sort of stuff) Fred