RE: About integers
RE: About integers
- Subject: RE: About integers
- From: Doug McNutt <email@hidden>
- Date: Wed, 2 Mar 2005 13:23:24 -0700
At 17:42 +0100 3/2/05, Emmanuel wrote:
>If I'm right, then it looks like Mission impossible, you only have 2^32 integers, and you have much more possibilities for admin_user, so one integer *has* to code for *several* admin_user values.
UNIX does not maintain any copy of actual passwords as text. Rather it uses an algorithm that hashes the password into something binary. It is "irreversible" so that there is no way to get the original text back. Well, "no way" is approximate in the world of modern crypto but surely multiple combinations of text can end up stored as the same binary hash.
I know the hash was once 32 bits but I'm not sure it still is.
If that's what the original poster has in mind I think it's not going to work. The password police will get upset even if the hashing were to be done correctly. One can obtain a user's hashed value with simple root access but it can't be used.
And while I'm at it. . .
At 11:42 -0800 3/1/05, Christopher Nebel wrote:
>Regardless, it should work with "write ... as [unsigned] integer", assuming the number is in the appropriate range for the type. You're relying on Apple Event Manager coercions at that point.
Does that mean a 32 bit unsigned integer that needs the most significant 2 bits, but happens to be in a floating point 64 bit byte, will get written correctly in 32 bit format? Or is the "appropriate range for the type" still 2^29-1?
And further. . .
If we're in the Unicode world, are password characters that use the most significant bit converted to UTF-x before UNIX does its hashing?
--
--> There are 10 kinds of people: those who understand binary, and those who don't <--
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Applescript-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden