On Fri, 21 Apr 2000, Don Maslin wrote:
> True, it is finding the sector numbers it is looking for, but it also is
> finding E5H where it expects to find the user area number or there could
> be either an error message or perhaps a garbaged directory display if it
> found 00H by chance.
> But MS-DOS does not use E5H and does use 00H fill in directory areas. I
> suspect that this might confuse the CP/M-86 directory program a bit - a
> whole bunch of null filled directory entries, all at user 0!
Actually, MS-DOS will accept 00 OR E5 for directory fill. I've seen
bothIf the first byte of a directory entry is E5, then the space used by
that entry is available. If 00, then it is available, and no subsequent
entries have been used, so any directory search routines, including DIR
can stop. Any time a file is deleted, in addition to appropriate
manipulation of the FAT data, the first byte is changed to E5. There
used to be some silly "prevent this deleted file from being undeleted"
shareware programs that would e5 the rest of the entry.
A freshly formatted MS-DOS SHOULD be filled with 00. But E5 will still
work, and SOME companies used it, particularly in the early days.
Also, some used F6, with a 00 or an E5 every 32 bytes.
We are talking about MS-DOS, but NOT IBM compatible. There were some
weird ones.
ALSO, due to reserved tracks, the MS-DOS directory is earlier on the disk
than the CP/M one. Depending on size of the MS-DOS directory, number of
CP/M reserved tracks, side patterns, and interleave, they might or might
not overlap. The CP/M directory is more likely to start at about cluster
number 2.
--
Fred Cisin cisin_at_xenosoft.com
XenoSoft http://www.xenosoft.com
2210 Sixth St. (510) 644-9366
Berkeley, CA 94710-2219
Received on Fri Apr 21 2000 - 16:53:30 BST