https://bugs.documentfoundation.org/show_bug.cgi?id=152381
--- Comment #2 from Eike Rathke <er...@redhat.com> --- Sure, if code sets/uses a date with year 0 (or even day=0,month=0,year=0 in this case) for the proleptic Gregorian calendar then that code is wrong ;-) That so far didn't matter because the Base code had its own and wrong conversion. Problem here seems to be a NULL value of the database that is *not* a date to be converted to relative days, entering connectivity/source/commontools/dbconversion.cxx dbtools::DBTypeConversion::toDouble() with _rVal all 0 then via dbtools::DBTypeConversion::toDays() dbtools::implRelativeToAbsoluteNull() calls the now new comphelper::date::convertDateToDaysNormalizing() that via its calls asserts. The old dbtools::implRelativeToAbsoluteNull() implementation for day=0,month=0,year=0 returned -365 for that "date", which is completely off. Whatever the old now eliminated dbtools::implBuildFromRelative() did for such.. -- You are receiving this mail because: You are the assignee for the bug.