>> Today I ran into the same problem and after applying this fix, all seems to
>> work well again. Any reason this fix hasn't been applied in the CVS code ?
>
> This is not the correct fix. That field should be endian converted.
>
> I'm not sure what the correct fix would be at the moment. I'll be able to
> look into it further if someone could supply a core file.
>
> Thanks,
> Ivan
Actually, I was never convinced it was right, either. It just "made it
work", and I never got back to look at it. We're currently not using
xfsdump here, so I quit working on it.
My suspicion is that the little->big->little endian (or is it
big->little->big?) conversion doesn't work for the initial value of
(-1).
Steve
----------------------------------------------------------------------------
Steve Roseman
Lehigh University Information Resources
sgr0@xxxxxxxxxx
|