+1 Anton Avtamonov wrote: > On 1/25/06, Mikhail Loenko <[EMAIL PROTECTED]> wrote: >> One more reason when logs are necessary: >> >> If testing is possible in some configurations only >> (like set of providers contains something or default encoding is ...), then >> 1) build failing in all different configs would be annoying >> 2) One has to be able to scan logs for warnings to verify that >> functionality is tested >> when the config is as expected >> >> A different exit status for the tests that can test in the given >> configuration would >> help. >> >> Thanks, >> Mikhail >> > > Well, I didn't say that using of logging in some particular case in a > test is absolutely terrible. What I'm talking about is that in general > (i.e. in the 99,9% cases) we should use standard JUnit features, try > to keep each test small and clear. Using only JUnit asserts is a > common approach and I beleive the major part of community does so... > I agree that if you have to 'try' several classes/providers/what_ever > in the test (since you absolutely CANNOT build the required > environment from the test case and have to find an approprite test > scenario instead) it is good chance to log which env. settings you met > and which scenario was tested. However I hope that the majority of the > tests can build the required environment for the tested scenario. > > -- > Anton Avtamonov, > Intel Middleware Products Division >
-- Tim Ellison ([EMAIL PROTECTED]) IBM Java technology centre, UK.