This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Mary, It turns out I was wrong, the bug you reported was actually another symptom of the same bug that 1.8.4-patch1 fixed. Here's the explanation from the HDF5 developer: On 05/14/2010 02:55 PM, Russ Rew wrote: > We have reports of a possibly different bug with the symptoms that > > - a file with more than 8 attributes of an object is written on a > little-endian platform using version 1.8.4-patch1. > > - an attempt to read that file on a big-endian platform using version > before 1.8.4-patch1 results in a seg fault. > > Is this just another symptom of the same bug (the contrapositive of the > other bug :-), or is it a different bug? Yes, that is the same bug. Basically, there are two populations as far as the endianness of these values in the file goes: 1. All little-endian machines, and big-endian machines using 1.8.4-p1 or later 2. Big-endian machines using 1.8.4 or earlier Files written by one group cannot be read by the other. Files written by group 2 do not conform to the file format spec and are technically corrupt (though they can still be read just fine by group 2). Russ Rew UCAR Unidata Program address@hidden http://www.unidata.ucar.edu Ticket Details =================== Ticket ID: PEN-204681 Department: Support netCDF Priority: Urgent Status: Closed