This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Current UI is way too minimalistic and misses: - simple way to rerun the same set of test (there are few RFEs about that) - simple way to rerun all and failed first or just failed tests (IMO must have if you think seriously about TDD) - support to run arbitrary test Sad to say that Eclipse has it right.
Assigning issues from Michael to Andrey.
1. Now there is a "Rerun" button on the results window. And it's enabled if the ant file and targets which were used to run tests were identified. 2. This feature is not implemented yet. 3. What do you mean by arbitrary test? You can run/debug any single test from the results tree (supported only for JavaSE projects). Why it was filed as DEFECT? If Eclipse has some feature not supported by Netbeans then it not means that this is DEFECT for NB. All the requests described in this issue looks like RFE or FEATURE. So transfer it to the FEATURE.
re 1.: nice. thanks re 3.: sorry but I do not see this (neither in j2se project nor in any other like apisupport). re why DEFECT: 1. I strongly believe it is defect because it blocks a common workflow for people doing test driven development. Go to source + run file is one more action and does not fix 'run one test case' problem and adds another editor window. I already expressed myself that it is a shame that junit plugin has big usability problems and this stems from a fact that it was designed by people who do not know what TDD means - 8 pages long UI spec where less than 10% is dedicated to 'how to run' tests topic and space is wasted on wizards for test creating is a proof. 2. RFEs are ignored 3. I can give you a list of good and bad things in Eclipse/NetBeans/IntelliJ. NetBeans is not in a position where it can ignore its competitors. Eclipse has done test execution support better than NetBeans.
I strongly support adding the missing features described by _rkubacki (no 2 and no3). Those have been a serious lack in NB for a long time. Regarding to no 3, I think the ability to run a single test should be provided at least in the contextual menu in the editor when the cursor is inside a test, in the navigator window, in the test results window. There are various reasons for needing no 3, but I will give you the one that has bothering me recently: running UI tests with JUnit+Jemmy. The UI I am testing is very complex, so its creation takes around 2/3 seconds (depending on the computer). Now, I have more than 200 tests just for this particular JPanel and if just one fails (say a malformed Tool Tip Text in just one of the components) I need to run all the tests (about 7/10 min). So this really blocks my progress. Another reason is when developing a single test. It would be very useful to run just the test under development and avoid waiting for the other tests to complete. Of course, after the new test passes, I would run all of them. I changed the version to 6.9 to see if we are lucky this time!
Another thought about this: I think that when re-running a failed test, the IDE should keep in the failing list all the others that failed, so we can run/debug them as well without having to run all the tests again.
No.2 feature was implemented. No.3 is available from the context menu of the test method node (result view) at least for the j2se applications. As for running single test methods from the editor there is another issues: 72080, 166252 Marking as resolved.