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.
Sometimes tests fail with error "The button [Finish] is not found or disabled in the dialog [NetBeans IDE Installer]. Need to increase delay time before selecting of radio button to be sure that problem occurs not because dialog was finally initialized after radio button was selected (it probably can happen if test host is slow).
It works for me in all cases on all (SunOS sparc/x86, Linux, WinXP) platforms (see the attached WinXP_results.zip). Of course, if you need to increase delay somewhere in automated tests or gravy library for investigation of test failing, you can do it. If test host is slow, it should be changed to another one, which is faster.
Created attachment 54145 [details] test results
See results: http://test4u.czech.sun.com/test4u-users/rave-qe-spb/results/20071210100004/rave-blade1/ All other tests on the same host at the same time passed. Probably the reason is that UC wasn't available, but it was in the rest of tests. So I think the reason is just that Plugins dialog was initialized after method of PluginsOperator was invoked. It is very easy to fix. Your results are excellent, but sometimes tests fail (not always and not often) because of PluginsOperator, so it should be fixed to be sure that it works perfectly. Probably you have computer better than test host and therefore you don't have problem. What about faster test host, so we don't have it. If you want to buy us one, you are welcome. But anyway, tests should work on different hosts, on fast and on slow. Not all customer have fastest computers.
When I run automated tests on all (SunOS sparc/x86, Linux, WinXP) platforms, they works for me. I don't know, which delay value is better for your machines/tests. So, it's really easy to change delay to desired value. Please do it.
You has developed this operator and you knows better where delay should be increased. Tests fail not on my host, they fail on test hosts and you have access to them. It will be better to double delay. And please, don't reassign this issue to me. You has developed this operator, so please fix it.
Please, don't reassign this issue to me. This operator is a part of common gravy library, so, any engineer can change delay to desired value if it's needed.
Every engineer can, but you has developed this operator and you knows it better than anybody. And therefore you are the best person to fix this. If you don't want to do this, just tell this.
Fixed in release60 and in trunk.
Verified.