Thank you Dominik, you are of course right ... I somehow missed that double-usage yesterday night ... :|
Andi. On 10.11.2014 21:23, Dominik Stadler wrote: > Yes I did run it. The failures are a bit strange, but I think you get > them because members are null that are cleared during close()ing, e.g. > POIXMLDocument.pkg is nullified. > > and the Workbook is referenced again after the first > writeOutandReadBack(), e.g. in one of the failing unit test there is > the following code: > > st = > XSSFTestDataSamples.writeOutAndReadBack(workbook).getStylesSource(); > > ... > > assertNotNull(XSSFTestDataSamples.writeOutAndReadBack(workbook)); > > here the workbook is passed two times to writeOutAndReadBack(), but in > your case it is already closed in the first call, thus the second > invocation fails. > > The other two unit-test failures in TestStylesTable are similar, the > workbook is accessed in the second writeOutAndReadBack() after it was > closed before and so the second write fails. > > Dominik. > > On Mon, Nov 10, 2014 at 6:45 PM, kiwiwings <[email protected]> wrote: >> Have you tried applying the pastebin to your local XSSFTestDataSamples class >> and run the TestStylesTable test? >> >> The original source reference is not used again in these tests and I've got >> unrelated errors, which seem not to be connected to closed resources. >> Last time when I've tried the ubuntu live cd, this also happened there, so I >> think it's not specific to the op-system. >> >> Andi >> >> >> >> -- >> View this message in context: >> http://apache-poi.1045710.n5.nabble.com/Why-is-this-breaking-the-tests-tp5717111p5717133.html >> Sent from the POI - Dev mailing list archive at Nabble.com. >> >> --------------------------------------------------------------------- >> 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] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
