Hi all,

The dtest pattern test is still not working for me. I thought I'd post a
few more observations about the problem, and see if the gurus are
listening this time :)

The mingw patch is also attached, if it helps.

To reiterate, scan_pattern_test is producing messages of the form:
"pattern check failed on file /flash/yaffs2/f1, size 10000 at position
4608. Got 0 instead of ffffedff
pattern check failed on file /flash/yaffs2/f1, size 10000 at position
4864. Got 0 instead of ffffecff"

All tests were started from blank 'flash' (emfiles deleted)

- Yaffs1 (512b pages), short op cache enabled: First failure is in
iteration 0!
- Yaffs1 (512b pages), short op cache disabled: First failure is in
iteration 0.
I'm not sure that any iterations actually *pass*.

- Yaffs2 (2k pages), short op cache enabled: First failure is in
iteration 22.
- Yaffs2 (2k pages), short op cache disabled: First failure is in
iteration 12.
The common element for these seems to be that the first error happens
just after the allocation of block 26.

There's also a rather odd glitch in iteration 18 of the yaffs2
without-cache case, where it finds 'f1' to be 18179904 bytes big instead
of 10000... ow.

James