Re: [MiNT] mv does delete when fail of fat (was: 1-18-cur trunk stability issues)

[Paul Wratt <paul.wratt@gmail.com>]

I did some tests years ago at "hiding" files on FAT drives, and how
delete and undelete worked. If you only E5 the 1st char (manually) the
FAT entries in FAT1 would get trash by CHKDSK (set to free) and you
would loose the file (unless you kept another fat list elsewhere and
no new files over wrote the data)

maybe the fat pointer in the dir listing is zero'd (but that would
stop undelete from working), I meant that there could still be valid
dir entries in the dir list, even tho the filenames might be invalid
to TOS/fs for output, because fs check only for valid entries in the
dir list, and that the size recorded there matches the cluster list in
the fat. the FAT is only (normally) zero'd with special tool, not on
delete

I do remember FAT1==FAT2 is not always true, ie not always maintained
by OS (and so not usable for its intended purpose, to restore FAT1 -
circa 1980)

Anyways, it appears the Marks initial problem is mv version related,
not fs related

Paul

2011/11/30 Vincent Rivière <vincent.riviere@freesbee.fr>:
> Paul Wratt wrote:
>>
>> remember that a deleted file simply has the first char of&hE5
>
>
> No, the FAT entries are freed, too.
>
> --
> Vincent Rivière
>
>