Do you think (or AKira) that this is the problem for mips?
For sparc I am a bit stymed.
Yes, it's all about BigEndian architectures (and ICU and upmendex).
(And don't bother explaining that sparc is endian-neutral, the practical
reality is that it is still bigendian in our world.)
And the -DU_IS_BIG_ENDIAN=0 kludge I found is a big pain and I wish
someone, anyone, could find the real bug and the real fix. (For next
year.) I can't go there again.
Another approach, simpler and just as viable IMHO, would be (I assume)
--disable-upmendex :). --karl
