Mark, These errors will be fixed by HARMONY-6590 which is marked as a dependency on the HARMONY-6594 jira. Our logic of OutputStream.close() is wrong according to java spec (wrongly use CharsetEncoder.flush()), but previously our CharsetEncoder also doesn't comply with RI so that 2 errors by chance let OutputStream.close() pass luni unit test before . Please see my comments on https://issues.apache.org/jira/browse/HARMONY-6590. Thanks a lot!
2010/7/27 Mark Hindess <mark.hind...@googlemail.com> > > In message <201007270818.o6r8ir9m021...@d12av02.megacenter.de.ibm.com>, > "Mark Hindess" writes: > > > > I've just fixed a break (r979569) that was stopping the hudson builds > > running all the tests. However even with the aborted tests run, there > > were a lot of regressions. See: > > > > > http://hudson.zones.apache.org/hudson/view/Harmony/job/Harmony-1.5-head-lin > > ux-x86_64-full-tests/272/ > > > > Unfortunately, hudson hadn't done a build in a while so quite a few > > changes went into that build which makes it harder to see what might > > have caused the failures. > > > > The latest test run: > > > > > http://hudson.zones.apache.org/hudson/view/Harmony/job/Harmony-1.5-head-lin > > ux-x86_64-full-tests/275/ > > > > should at least run all the tests. > > > > Please can we focus on fixing these regressions? > > Looks like: > > r967069 | tellison | 2010-07-23 13:24:45 +0100 (Fri, 23 Jul 2010) | 2 > lines > Changed paths: > M > /harmony/enhanced/java/trunk/classlib/modules/nio_char/src/main/java/java/nio/charset/CharsetEncoder.java > M > /harmony/enhanced/java/trunk/classlib/modules/nio_char/src/test/java/org/apache/harmony/nio_char/tests/java/nio/charset/ASCIICharsetEncoderTest.java > M > /harmony/enhanced/java/trunk/classlib/modules/nio_char/src/test/java/tests/api/java/nio/charset/CharsetEncoderTest.java > > Apply patch for HARMONY-6594 ([classlib][nio_char] > CharsetEncoder.flush(Bytebuffer) doesn't follow the spec and RI's > behavior) > > is responsible for most of the test failures. > > Deven/Tim: can you take a look please? > > Regards, > Mark. > > > >
