Details
Description
This bug looks like a windows line-ending (CRLF) vs. unix line ending (LF) problem. I.e., real bug in test because it is line-ending sensitive. When I look at this in Emacs it shows me the CRs as ^M. Every line ends with a ^M except the ones I marked with ### below.
THere are several of this same failure. These are tests related to internal whitespace preservation, so it's a place where Windows vs. Linux sensitivity may come up.
[error] Test edu.illinois.ncsa.daffodil.section23.dfdl_expressions.TestDFDLExpressions.test_internal_space_preserved3b failed: java.lang.Exception:
[error] Comparison failed.
[error] Expected
[error] <expr_space3b><dummy>2</dummy><value>a
[error] b
[error] c</value></expr_space3b>
[error] Actual
[error] <expr_space3b><dummy>2</dummy><value>a ### no ^M here
[error] b ### no ^M here
[error] c</value></expr_space3b>
[error] Differences were (path, expected, actual):
[error] (expr_space3b/value,'a
[error] b
[error] c','a ### no ^M here
[error] b ### no ^M here
[error] c'), took 0.618 sec
We need to change these tests so they are portable somehow.