WINDOWS/XP:

Due to the update of Compaq Visual Fortran 6.1 -> 6.6 a lot of differences in
the testresults occured because of the different behaviour of the
      write(resLun,*) <var.s>
statement (addional spaces and/or less spaces).

Therefore, all testresults have been diffed with:
   diff -w resultsApproved/<dir>/w32/<filename> progsrc/<dir>/w32/<filename>
The -w option ignores all spaces.
See diff_CVF_6.1_6.6.sh for shellscript.

diff_CVF_6.1_6.6-res.txt contains the results:
- Diffs in DelftIO versions strings
- Diffs in the last printed decimal.

A few His/Map files also differ (see diff_CVF_6.1_6.6-binaries.txt).
TESTRewindTot.his
- differences in last printed decimal (see progsrc/tstAppend/w32/TESTRewind-res.txt); OK
  (binary diff on TESTRewindTot.his shows diff in least significant byte (e.g. B8->B9).

TESTHisLong.his
- TESTHisLong-res.txt contains no differences; however: prints only 6 decimals
- binary diff on TESTHisLong.his shows diff in least significant byte (e.g. 3D->3E); OK.

TESTHisData.his
- binary diff on TESTHisData.his: only version string differs; OK.
  (TESTHis-res.txt: no differences indeed)

TESTHisDouble.his
- binary diff on TESTHisDouble.his shows diff in least significant byte; OK.
- binary diff between TESTHisDouble.his and TESTHisData.his shows diff in least significant byte; OK.

Binary files resultsApproved/tstMap/w32/TESTMapLong.map and progsrc/tstMap/w32/TESTMapLong.map differ
- TESTMapLong-res.txt contains no differences; however: prints only 6 decimals
- binary diff on TESTMapLong.map shows diff in least significant byte; OK.

All tests OK.

Linux Intel Fortran Compiler (grep on DIFFERENCES in allTests-intel-res.txt):
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAll2D/unix/TEST2DFAutoFiles-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAll2D/unix/TEST2DFAutoShm-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAll2D/unix/TEST2DFSerial-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAll2D/unix/TEST2DFSynch-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAllF77/unix/TESTF77Serial-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAllF77/unix/TESTF77Synch-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAllPLT/unix/TESTPltAutoFiles-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAllPLT/unix/TESTPltAutoShm-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAllPLT/unix/TESTPltSerial-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAllPLT/unix/TESTPltSynch-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstAppend/unix/TESTRewind-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstDioConf/unix/TESTDioConf-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstShmPlt2DfF90/unix/TESTShmPlt2dfInMem-res.txt
DIFFERENCES in /f/sunfish.users/homes/hummel/tmp/dio/diof90/progsrc/tstShmPlt2DfF90/unix/TESTShmPlt2df-res.txt

All differences originate from the fact that the Intel Fortran compiler (ifc) was
used, while the stored resultsApproved are from the Lahey compiler (lf95).

After removing every block for timestep 11 to timestep 100 from the original
w32 results (needed for the test tstAll2D, tstAllF77 and tstAllPLT; these test
perform only 10 steps on linux), all testresults have been diffed with:
   diff -w resultsApproved/<dir>/w32/<filename> progsrc/<dir>/unix/<filename>
The -w option ignores all spaces.
See diff_lf95_ifc_linux.sh for shellscript.

The differences (see ) show only:
- Diffs in DelftIO versions strings
- Diffs in the last printed decimal.

All tests OK.