Strange FileCore effects

By pure random I detected an odd FileCore behaviour a few days ago and I am meanwhile quite sure that it is a bug. I was able to reproduce the effect on different storage media.

I took HForm from Git. I formatted my media with:
20 Heads
32 Sectors
1310 Tracks

So format is not big (but you can reproduce it also using 40 heads).

On the empty disc I copied an arbitrary directory from my NAS. 919 files with 391MB. CheckMap and DiscKnight are saying that everthing is fine. Now I am wiping all the files (Wipe * R).

Fine. Disc is empty now. CheckMap
Map inconsitent with directory tree!?!? Sh..t
DiscKnight acknowledges this and gives more hints. A detailed examination showed the following:
There is a sub directory which contains 192 PDF files with together around 81MB.
This directory gets two Chunks inside Map. You can access all of the files. All fine but when deleting FileCore forgets to remove the second chunk entry …

Ok, location is overall a bit strange:
End of sector &661C3 (Map 131)
…
&75AE8000 <- &75AE is SIN of the directory
&00007781 (&00000001 after Wipe)
&80000000

ID length is 16 Bits. Either it has something to do with the location of the SIN exactly on upper word boundary in conjunction with 16 Bits ID length or it
is just a relict of the good old times where directories had only one chunk …

FileCore versions used for the tests were 3.77/3.78 but if looking into the changes I don’t think that this is playing a role. In worst case it is a bug in amber and someone forget this area as big directories where introduced. Perhaps it doesn’t have a practical meaning. The range is free and reuseable. I am not sure what will happen in case that the SIN will be used again. But it may be that the likelihood to win a lottery or being killed by a lightning is much higher than to get troubles by this effect.

Interesting.

DiscKnight should be able to delete the object if FileCore can’t. I’d need to refresh my memory of how it works first though!

DiscKnight is able to repair this of course. FileCore can’t and will remain believing that there is still this bug even you just repaired it with DiscKnight. You will have to perform a Dismount/Mount to make FileCore believing that the bug is gone.

DiscKnight is a great tool. Lucky I have to use it seldom but if I have to use it it is doing a great job!