Thanks for this find! Indeed, Dates are written as local timestamp with explicit time zone offset. The specification does not prescribe that, but I think that having this information is valuable. Unfortunately, this means that the PDF representation of Dates cannot be tested in a time zone independent way L. Therefore, I removed this from the tests. Christian Von: pdf4st-bounce@xxxxxxxxxxxxx [mailto:pdf4st-bounce@xxxxxxxxxxxxx] Im Auftrag von bobn@xxxxxxxxxx Gesendet: Sonntag, 20. Mai 2012 10:31 An: pdf4st@xxxxxxxxxxxxx Betreff: [pdf4st] SUnit test timezones The SUnit test SimpleObjectTests>>testDate has a timezone dependency, such that... self assert: (File readTesterOn: '(D:199812231952-08''00)') readObject content = (Timestamp d: 24 m: 12 y: 1998 h: 4 m: 52). ...fails. The test would work in GMT+1 but fails in GMT-4. (File readTesterOn: '(D:199812231952-08''00)') readObject content = December 23, 1998 10:52:00 PM (Timestamp d: 24 m: 12 y: 1998 h: 4 m: 52) = December 24, 1998 4:52:00 AM (Timestamp d: 24 m: 12 y: 1998 h: 4 m: 52) asPDF pdfSource = '(D:19981224045200--5')' vs. '(D:19981224045200+01'')' Same problem for TrailerTests>>testCreationDate. I'm running the tests while renaming PDF4Smalltalk classes with a PDF4 prefix. These are the only two tests that fail, but in an understood way. I'll skip them for now. Bob