Thanks, I was aware that I'd checked in bad code, but hadn't been able to access a system since to be able to fix it. I might get to it tonight if someone doesn't beat me to it.
Stephen --- Stefan Bodewig <[EMAIL PROTECTED]> wrote: > Hi, > > this is the manual version of what Gump is trying to > tell you 8-) > > FilenameUtils.getPrefixLength fails on Unix if you > pass a string of > length one as argument since it unconditionally > tries to access the > second character (just to never use it). Therefore > the test case > fails since it tests "/". > > The trivial patch appended fixes this problem > > It will not fix the build since the tests still > fails, but I'm not > sure whether the code ot the test is wrong. The > test failure > remaining is > > [junit] Testcase: testNormalize took 0,02 sec > [junit] FAILED > [junit] Check if '/../' normalized to 'null', > was '' expected:<null> but was:<> > [junit] junit.framework.ComparisonFailure: Check > if '/../' normalized to 'null', was '' > expected:<null> but was:<> > > Cheers > > Stefan > > Index: > io/src/java/org/apache/commons/io/FilenameUtils.java > =================================================================== > RCS file: > /home/cvspublic/jakarta-commons/io/src/java/org/apache/commons/io/FilenameUtils.java,v > retrieving revision 1.27 > diff -u -r1.27 FilenameUtils.java > --- > io/src/java/org/apache/commons/io/FilenameUtils.java > 23 Nov 2004 00:04:29 -0000 1.27 > +++ > io/src/java/org/apache/commons/io/FilenameUtils.java > 26 Nov 2004 07:51:24 -0000 > @@ -416,7 +416,6 @@ > } > } else { > char ch0 = filename.charAt(0); > - char ch1 = filename.charAt(1); > if (ch0 == '~') { > if (len == 1) { > return -1; > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: > [EMAIL PROTECTED] > For additional commands, e-mail: > [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]