Niall Pemberton wrote at Freitag, 9. Oktober 2009 00:03: > Hi Jörg, > > Thanks for testing, comments inline... >
[snip] >> >> Issues with Sun JDK 1.3.1: >> ========================== >> >> Source compilable, but a lot of tests fail. > > All the tests pass for me - I was using Ant 1.7.1 and Sun's JDK > 1.3.1_18 on Windows XP - but I tried Ant 1.6.5 and JDK 1.3.1_20 (again > on WinXP) and also had no problems. Are the tests actually failing or > are you just seeing alot of exceptions in the output? Because > BeanUtils uses reflection to intialise the "cause" when re-throwing > exceptions for JDK 1.4+, but for JDK 1.3 it logs the stack trace of > the original exception. For example: > > http://commons.apache.org/beanutils/xref/org/apache/commons/beanutils/PropertyUtilsBean.html#2194 > http://commons.apache.org/beanutils/xref/org/apache/commons/beanutils/BeanUtilsBean.html#1041 I'll have a closer look this evening when I'm back from office, but, yes, it looks like a lot of those. >> Issues with IBM JDK 1.4.2 and 1.5: >> ================================== >> >> 1/ strange warnings on console for one test: > > The class thats being tested logs those warnings for null values: > > http://commons.apache.org/beanutils/xref/org/apache/commons/beanutils/BeanPropertyValueChangeClosure.html#170 [snip] I simply wondered, why these warnings appear for those two JDKs only. However, it might also related to the fact that I've build those with Maven 1 and the mechanism to catch the console output is different (or configured in a different way). >> >> 2/ org.apache.commons.beanutils.bugs.Jira347TestCase: 1 Error > > This test was copied from MemoryLeakTestCase and the same problem > occurred for BeanUtils 1.8.0 when you tested with maven1 and IBM JDKs. > Would be nice to get these tests working (wonder if they work with IBM > JDK and ant, rather than maven1) but I can't seem to find anywhere to > download the IBM JDK for windows. So since its a problem in the test > rather than actual BeanUtils functionality I can only suggest that we > ignore this again, unless someone can come up with a fix for the test: > > http://commons.markmail.org/message/6fhsjmrugz6c746y see my reply to sebb [snip] >> Issues with IBM JDK 1.6: >> ======================== >> >> 1/ org.apache.commons.beanutils.memoryleaktests.MemoryLeakTestCase >> ================ %< ==================== > > OK so the IBM JDK 1.6 got past the setup problems and fails the actual > test. This is disappointing because it means the fix we put in for the > memory leak that works on Sun's JDK doesn't seem to work for IBM JDK > 1.6. I've created a bug report for this: > > https://issues.apache.org/jira/browse/BEANUTILS-366 > > ...but its going to need someone to fix this or for me to get the IBM > JDK. Anyway those same tests fail in Sun's JDK 1.3 & JDK 1.4 - so > theres a check so they don't run on pre-1.5 JDKs - can't find anything > in the archives as to why they fail or the results of any > investigation. Anyway I don't believe this is something new in this > release and the attitude we took with the memory leak fixes is that if > they work in some JDKs (i.e Suns) then thats progress even if they > don't resolve it for all JDKs. Anyway I've logged the bug but I'm > hoping this won't stop this release since none of the changes made > since 1.8.0 go anywhere near the caching that causes/caused the memory > leaks. We might add something to release notes about known issues then. [snip] >> Minor nits: >> ========== >> xdocs/changes.xml: BEANUTILS-354 is reported twice > > Thanks I fixed this in trunk: Hehe, you additionally spotted the year of the release. I'm currently filling my tax forms for last year, so it did look quite too familiar for me :)) - Jörg --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
