Hi, I have uploaded a new (!) version of FORMAT 0.91j to
http://www.coli.uni-sb.de/~eric/stuff/soft/by-others/

This format-0.91j.zip version uses 0x0fffffff instead of 0x0ffffff8 as
end-of-chain for the root directory, marks the free_count in the filesystem
info sector backup copy as invalid, can mark bad clusters even in FAT32
(NOTE: will only SEARCH for bad clusters if you do FORMAT /U ... even MS
does not do wipe / surface scan on FAT32, I think. It would take hours on
big drives), and fixes a bug where the DMA overflow protection would spoil
proper clearing of the root directory or maybe even other areas in rare cases.

Open bugs and TODOs:
- should use Joes code to detect CD-ROM and network drives
- should allow the user to retry again even after 5 attempts of sector access
  (using some prompt like "retry/abort/infiniteretry-dontaskagain/ignore" ?)
- formatting to DD sizes in HD floppy drives still does not work well...
- you cannot unformat after FAT32 format (no "SafeFormat" yet)
- IF you have a broken sector in the FAT32 root dir, the root dir could be moved...
- 1,234,567 ASCII_CD display should use NLS settings
- 40:[9x] values should be displayed (baud rate, double stepping) and maybe
  even modified (only needed for XT BIOSes?)
- clusters taken by system files in /S ("call SYS") mode should be counted and
  displayed in the summary
- existing bad cluster marks in the FAT should not be overwritten (this would
  be controlled by a to-be-added /C switch). Maybe even surface scan / wipe
  should avoid them. The latter makes not much sense as surface scan is off by
  default, I would say.
- FAT32 FAT sizes should be optimized, but this is a KERNEL TODO.
- FAT32 boot sector backup copy should be updated after SYS. This is a SYS TODO.


Happy testing!

Eric.