My principle is to put such data in files in src/test/resources, which has a different inconvenience profile.
On Wed, Feb 2, 2011 at 7:57 AM, Andy Seaborne <[email protected]> wrote: > > > On 02/02/11 12:39, Benson Margulies wrote: >> >> The problem is that you've got real UTF-8 in some test files, and >> eclipse stupidly set itself up with non-UTF-8 for me (I am using the >> not-quite-there-yet new M2E plugins). >> >> So the tests pass, and some simple tests manual of my finish work on >> your command pass, so you might have a look the patches. > > So the tests are good tests then, well sort of - it's the Java source code > test data that's wrong here. ☺ ☻ Writing them all out in \u to keep within > ASCII would be ugly. > > I've added UTF-8 to the project settings (and Unix line endings). > > Andy > > > >> >> On Wed, Feb 2, 2011 at 3:52 AM, Andy Seaborne >> <[email protected]> wrote: >>> >>> Works here (there was a new SNAPSHOT build yesterday so it must have >>> worked >>> because test failures block deployment). >>> >>> What does the Surefire report say? >>> >>> Andy >>> >>> On 02/02/11 02:34, Benson Margulies wrote: >>>> >>>> https://issues.apache.org/jira/browse/JENA-37 has the first piece of >>>> the patch, for ARQ. >>>> >>>> Note that the ARQ build is failing for me in ways that seem unrelated >>>> to my change: >>>> >>>> Failed tests: >>>> nt_model_1(org.openjena.riot.lang.TestLangNTriples) >>>> >>>> Tests in error: >>>> codec6(org.openjena.atlas.lib.TestBytes) >>>> codec7(org.openjena.atlas.lib.TestBytes) >>> >
