29 Dec
2009
29 Dec
'09
11:48 p.m.
It's not yet clear to me yet how the crypt(3)
assembler implementation
translated the binary output of the (very similar looking) algorithm to
printable text. I'm still staring at the assembler code. Evidently I
don't speak PDP-11 very well today.
Ok, finally got my lame brain to work.
After doing the rotor-cage-link crap, the password crypter takes
modulus-62 of each byte of the cryptext and base-62 encodes that
remainder using the set 0-9A-Za-z.
If I'm ambitious (hah!) tonight, maybe I'll see about hacking that into
the crypt(1) implementation and testing it.
De