D051
- The two ways of writing a 1 near the end of the
superblock are two fields, not one
An image with a seed writes an index at 0x36C
and the seed after it. An image without one writes
1 at 0x368 and stops. Both had been measured
as "a 1 near the end of the header", and reading could never tell them
apart because nothing reads either.
The writer is what separated them, and it is the second time in this
area that writing has found something reading could not (D042 was the
first). Both offsets are now named. The inner filesystem is unseeded and
uses NO_SEED_INDEX; the outer one is seeded and uses
UNKNOWN_INDEX.
The same pass corrected N_BLOCK. It is not the block
count, despite the name: the source writes a literal 1 and
remarks that it always is. Writing the real count there would have been
more plausible and would have made every image this builds differ from
every image examined, in a field nothing checks.