OS=CentOS-5.2 Ruby=1.8.6 Rails=2.2.1_RC Gems all up to date I have been experimenting with autotest and I have a few questions. 1. given export AUTOFEATURES=TRUE if ./test exists (with tests) then these tests are run and the features are not. Is this intended behaviour? 2. given require ''autotest/redgreen'' in .autotest if /test tests are run then all I get colourized is a line of ''======'' but if /test is removed and features run then the Failed message and the backtrace are colourized. Is this intentional as well? -- Posted via http://www.ruby-forum.com/.
James Byrne wrote:> OS=CentOS-5.2 > Ruby=1.8.6 > Rails=2.2.1_RC > Gems all up to date > > I have been experimenting with autotest and I have a few questions. > > 1. given export AUTOFEATURES=TRUE if ./test exists (with tests) then > these tests are run and the features are not. Is this intended > behaviour? > >Try doing TRUE in all lowercase. -Ben
Ben Mabey wrote:> James Byrne wrote: >> >> > Try doing TRUE in all lowercase. > > -Ben$ set AUTOFEATURE=true BASH=/bin/bash ... -- Posted via http://www.ruby-forum.com/.
James Byrne wrote:> OS=CentOS-5.2 > Ruby=1.8.6 > Rails=2.2.1_RC > Gems all up to date > > I have been experimenting with autotest and I have a few questions. > > 1. given: export AUTOFEATURES=true, if ./test exists (with tests) then > these tests are run and the features are not. Is this intended > behaviour? >I have investigated further and the situation is this. If all the unit/functional tests under ~/test pass then, and only then, cucumber feature tests are run. If any unit/functional tests fail then cucumber is not invoked. It is not immediately clear to me whether this behaviour is intended or not. Can someone provide the answer? -- Posted via http://www.ruby-forum.com/.
James Byrne wrote:> James Byrne wrote: > >> OS=CentOS-5.2 >> Ruby=1.8.6 >> Rails=2.2.1_RC >> Gems all up to date >> >> I have been experimenting with autotest and I have a few questions. >> >> 1. given: export AUTOFEATURES=true, if ./test exists (with tests) then >> these tests are run and the features are not. Is this intended >> behaviour? >> >> > > I have investigated further and the situation is this. If all the > unit/functional tests under ~/test pass then, and only then, cucumber > feature tests are run. If any unit/functional tests fail then cucumber > is not invoked. It is not immediately clear to me whether this > behaviour is intended or not. Can someone provide the answer? >Ahh... I didn''t notice that you said you were using TestUnit. Sorry about not catching on sooner.... The cucumber autotest integration was developed with autospec and rspec in mind. I wouldn''t be surprised if it never was tested with testunit so it might have some rough edges. Peter Jaros, the one who wrote the plugin, explained how the plugin is suppose to work/behave in the wiki: http://github.com/aslakhellesoy/cucumber/wikis/autotest-integration (under ''What does it do?'') After reading the wiki I think what you are seeing is the expected behaviour. I don''t think it is suppose to run the features until all your tests are passing. -Ben
On Thu, Nov 20, 2008 at 11:55 AM, Ben Mabey <ben at benmabey.com> wrote:> James Byrne wrote: > >> It is not immediately clear to me whether this behaviour is >> intended or not. Can someone provide the answer? >> > > After reading the wiki I think what you are seeing is the expected > behaviour. I don''t think it is suppose to run the features until all your > tests are passing.Yes, that''s the intended behavior. According to that workflow, if you''ve got unit tests failing, that''s what you should be working on. In RSpec, if you have failing examples which you don''t actually want to deal with yet, you can make them pending, and Autotest will ignore the fact that they''re not working. I don''t think Test::Unit provides an equivalent feature, though. Peter
Peter Jaros wrote:> In RSpec, if you have failing examples which you don''t actually want > to deal with yet, you can make them pending, and Autotest will ignore > the fact that they''re not working. I don''t think Test::Unit provides > an equivalent feature, though. > > PeterYes, and no. You can emulate a pending test by creating an empty method: def test_should_show_me #TODO Pending end As you imply, this will not show in the test run output. It does show up with the rake notes task though: $ rake notes (in ~/Software/Development/Projects/proforma.git) test/functional/clients_controller_test.rb: * [ 33] [TODO] Pending -- Posted via http://www.ruby-forum.com/.