Re: path name limits

>> I wonder, how you implement
>>
>> for directory size >= 64
>
>Under DR DOS 6.0 (1991), the PROMPT $P$G will display
>
>C:\4\5\6\7\8\9\0\1\2\3\4\5\6\7\8\9\0\1\2\3\4\5\6\7\8\9\0\1\2\3\4\5>
>
>So, it is truncating the string at the beginning. INT 21h/AH=47h will
>return with CY set and AX=0003h "path not found", the 64-bytes buffer
>at DS:SI, however, will be filled with

now I wonder, what DRDOS will do for the *common* scenario

	...
	GetCurrentDirectory(OriginalDirectory);

	SetCurrentDirectory("\");
	DosSomeInterestingStuff();

	SetCurrentDirectory(OriginalDirectory);

It's probably not, what you the user wanted to do.
there are probably more problems around.

> a long discussion of encryption/access verification

a) if it's enough to boot from MSDOS from floppy to
  get around your user access rights, it's not very hard to break in.

b) if, additionally, I have to use NU 'find partitions', because you
  scrambled the boot record, it's not very hard, either.

c) if your user password has a 16-bit hash, as your mail suggests,
  it's easy to break in - a computer will try to open a file with 65000
  different passwords in practical no time.

d) if you encrypt everything using 8 byte keywords, you have an effective
  cypher strength of ~ 64^8 = 2^48, as virtually passwords are
  a..zA..Z0..9.
  2^48 isn't particular good.
  if however, you include a 16 bit hash value of the keyword in the
filesystem,
  you have an effective security of 2^32, which is broken in no time.

I conclude, that no DRDOS has a security, better then to prevent the very
casual user
from tampering with the data.

tom