Hash of an entire file?

A while back GravityScore made a program that computes the sha256 hash of a string. This is good for using to store a salted password but what if you want to use it for creating a checksum to verify data integrity, in other words see if a file has been modified by a malicious program or a person who knows what the edit command does :P/>. I thought this would make for a really good anti-virus and auto updater by taking the checksum of all executable files used in an operating system and comparing them to a list downloaded by maybe pastebin. The problem there is that the function only accepts strings and i believe there is a character limit in lua as with probably every language. This would be very easily exceeded by man programs. Is there a way of taking the hash of an ENTIRE FILE or is this the wrong way to go about what i'm trying to do.

well according to the Java — ComputerCraft uses LuaJ which is a Java implementation of Lua — spec the max size String is Integer.MAX_VALUE, or 2³¹-1 characters in length. If you manage to have a file read into a string and its larger than 2,147,483,647 then you may want to consider refactoring your code so its smaller!

So in summary, yes, you can use GravityScores SHA256 hashing function; although there's probably better things you could use.

So in theory I could just

local f = fs.open("/path","r")
local checksum = sha256(textutils.serialize(f.readAll()))

and it would give me a value? I will test it tomorrow, it's almost midnight where I live and I should be off to bed.

Enderbane, on 10 April 2014 - 07:57 AM said:

So in theory I could just

local f = fs.open("/path","r")
local checksum = sha256(textutils.serialize(f.readAll()))

and it would give me a value? I will test it tomorrow, it's almost midnight where I live and I should be off to bed.

no not quite.

local f = fs.open( '/path', 'r' )
local checksum = sha256( f.readAll() ) --# contents are a string, not a table, no need to serialise
f.close() --# don't forget to close

theoriginalbit, on 10 April 2014 - 08:09 AM said:

Enderbane, on 10 April 2014 - 07:57 AM said:

So in theory I could just

local f = fs.open("/path","r")
local checksum = sha256(textutils.serialize(f.readAll()))

and it would give me a value? I will test it tomorrow, it's almost midnight where I live and I should be off to bed.

no not quite.

local h = fs.open( '/path', 'r' )
local checksum = sha256( f.readAll() ) --# contents are a string, not a table, no need to serialise
f.close() --# don't forget to close

Nope, not quite…

local f = fs.open( '/path', 'r' ) --# you defined it as 'h' instead of 'f'
local checksum = sha256( f.readAll() ) --# contents are a string, not a table, no need to serialise
f.close() --# don't forget to close

;)/>

Apemanzilla, on 10 April 2014 - 01:28 PM said:

Nope, not quite…

Shhh… no one is to know, lets keep this our little secret. ;)/>
Thanks, I was clearly very distracted when typing this and went muscle memory; I use `h`.

theoriginalbit, on 10 April 2014 - 01:46 PM said:

Apemanzilla, on 10 April 2014 - 01:28 PM said:

Nope, not quite…

Shhh… no one is to know, lets keep this our little secret. ;)/>
Thanks, I was clearly very distracted when typing this and went muscle memory; I use `h`.

Hehe :P/>
If you want to make it a one-liner, you can use this:

local checksum = sha256(fs.open("path","r").readAll())

Apemanzilla, on 10 April 2014 - 01:51 PM said:

If you want to make it a one-liner, you can use this:

local checksum = sha256(fs.open("path","r").readAll())

Except for the lingering file handle. while in the case of ComputerCraft it can be debated acceptable, it is definitely not good practice.

I've exprimented with onelining the fs api in the past, you're kinda screwed since you can't close the file afterwards :P/>

Well, thanks for the help guys, I'm just glad that there isn't a 256 char limit to strings like SOME languages I know. Very glad.

Enderbane, on 10 April 2014 - 11:12 PM said:

Well, thanks for the help guys, I'm just glad that there isn't a 256 char limit to strings like SOME languages I know. Very glad.

Which languages are those? I know of none. I do know Pascal / Delphi, SQL, and a few others have a max string length of 255 chars, but I'm unaware of any language that is 256. All languages that I know of use null terminated strings, meaning you only get 255 chars due to the last char being null.

theoriginalbit, on 10 April 2014 - 11:49 PM said:

Enderbane, on 10 April 2014 - 11:12 PM said:

Well, thanks for the help guys, I'm just glad that there isn't a 256 char limit to strings like SOME languages I know. Very glad.

Which languages are those? I know of none. I do know Pascal / Delphi, SQL, and a few others have a max string length of 255 chars, but I'm unaware of any language that is 256. All languages that I know of use null terminated strings, meaning you only get 255 chars due to the last char being null.

Uh, Lua doesn't have null-terminated strings as part of the language specification. In fact:

Lua 5.1 Refernce Manual, Section 2.1 said:

Strings in Lua can contain any 8-bit value, including embedded zeros, which can be specified as '\0'.