On 05/10/12, David Lutterkort wrote: > I get some errors on the modular branch; they need to be addressed > before we can commit this patch series. > > (1) Running 'rake test' gives me > > > /homes/lutter/.gem/ruby/1.8/gems/rake-0.9.2.2/lib/rake/ext/module.rb:36:in > `const_missing': uninitialized constant > Deltacloud::Helpers::Application::API_ROOT_URL (NameError)
Right, this task will trigger the Test::Unit tests we have. I can fix them to use Deltacloud::API. > (2) Trying to upload a blob through the HTML UI gives me > > > /homes/lutter/code/deltacloud/git/server/lib/deltacloud/./helpers/blob_stream_helper.rb:191:in > `is_put_blob': undefined local variable or method `settings' for > BlobStreamIO:Class (NameError) Fixed in 638886b3cd87a09e4b0ea244e4b6c0dd94463c25 > > (3) minitest isn't mentioned in the Gemfile Fixed in 314bba781b60e86f428ce30324e7b9d774954bf0 > > (4) running 'rake test:api' gives me > > test_0007_must_have_some_key_elements_inside_keys_(Deltacloud API > Keys) > [/homes/lutter/code/deltacloud/git/server/tests/drivers/api/keys_test.rb:43]: > Expected [] to not be empty. Fixed in 3e2e3d9857c668d3c466bfec69a50b53d12da08d > > (5) Running 'rake features' in tests/ gives me > > Error creating formatter: html (Errno::ENOENT) > > /homes/lutter/.gem/ruby/1.8/gems/cucumber-1.1.4/bin/../lib/cucumber/formatter/io.rb:7:in > `initialize' > The Cucumber tests should be launched from the 'server' directory: firefly ~/code/core/server $ rake cucumber:mock:test ........... 41 scenarios (41 passed) 292 steps (292 passed) 0m10.489s Note that I just fixed the Mock Cucumber suite, still working on EC2 one, but I assume since there were no changes to EC2 driver they will just work. > To commit these, we need to get to a state where all the tests pass; > that's the biggest blocker right now. Agree, I'll do that as high priority. Btw. I suggest to replace the 'mock' unit tests we have in server/tests/drivers/mock with the 'api' tests. They are doing the same thing now and the 'api' tests are more up-to-date. What do you think? > Also, to make the patch series a little less daunting, you might > consider splitting patches that aren't directly related to modularizing > DC out (e.g., 14/32 or 30-32/32) This will make it much easier to review > them and get them committed separately. Yeah, should I do that and send a new patch bomb? I can do it when rebasing against master when pushing but that will not be anyhow usefull for review ;-) -- michal -- Michal Fojtik Sr. Software Engineer, Deltacloud API (http://deltacloud.org)
