Andrey,
It seems the rmf tests are failing on epimethieus which I gather is a bigendian (powerpc) MacOS X machine.
http://buildbot.osgeo.org:8500/builders/epimetheus-full/builds/3/steps/test/logs/stdio
------------ Failures ------------
Script: gdrivers/gsg.py
TEST: gsg_3 ... fail
Failed to open dataset: data/gsg_7binary.grd
Script: gdrivers/rmf.py
TEST: rmf_1 ... fail
Checksum for band 1 in "byte.rsw" is 0, but expected 4672.
TEST: rmf_2 ... fail
Checksum for band 1 in "byte-lzw.rsw" is 0, but expected 4672.
TEST: rmf_3 ... fail
Checksum for band 1 in "float64.mtw" is 0, but expected 4672.
TEST: rmf_4 ... fail
Checksum for band 3 in "rgbsmall.rsw" is 0, but expected 21349.
TEST: rmf_5 ... fail
Checksum for band 3 in "rgbsmall-lzw.rsw" is 0, but expected 21349.
TEST: rmf_6 ... fail
Did not get expected checksum on still-open file.
Got 0 instead of 4672.
TEST: rmf_7 ... fail
Did not get expected checksum on still-open file.
Got 0 instead of 21053.
----------------------------------
Is this something you can dig into? If not, would you like me to look into it on my bigendian mac? Actually, I'm just jumping to conclusions that it is an endian problem. It could be something else too.
