Reading/Writing Binary files

Question

I'm just trying to read/write from a binary file. I've been following this tutorial, and it works... except it seems to be writing things to a txt file. I named the file test.bin when testing it, but notepad can open it and display it properly, so I don't think it's actually a binary file. I've told it that it's a binary file with "wb" and "rb" right?

if arg[1] == "write" then
    local output = assert(io.open(arg[2], "wb"))

    output:write(arg[3]) --3rd argument is written to the file.

    assert(output:close())
elseif arg[1] == "read" then
    local input = assert(io.open(arg[2], "rb"))

    print(input:read(1)) --Should read one byte, not one char/int. Right?
end

Answer 1

If you only write ASCII characters to a file, it will be possible to open it in Notepad or any other text editor just fine:

local out = io.open("file.bin", "wb")
local str = string.char(72,101,108,108,111,10) -- "Hello\n"
out:write(str)
out:close()

The resulting file will contain:

Hello

On the other hand if you write real binary data (for instance random bytes) you will get garbage:

local out = io.open("file.bin", "wb")
local t = {}
for i=1,1000 do t[i] = math.random(0,255) end
local str = string.char(unpack(t))
out:write(str)
out:close()

This is similar to those video game save files you have seen.

If you still don't get it, try to write all possible octets to a file:

local out = io.open("file.bin", "wb")
local t = {}
for i=0,255 do t[i+1] = i end
local str = string.char(unpack(t))
out:write(str)
out:close()

and then open it with a hexadecimal editor (here I used Hex Fiend on Mac OS) to see the correspondences:

Here, on the left, you have the bytes in hexadecimal and on the right you have their text representation. I have selected uppercase H which, as you can see on the left, corresponds to 0x48. 0x48 is 4*16 + 8 = 72 in base 10 (look at the bottom bar of the screenshot which tells you that).

Now look at my first code example and guess what the code for lowercase e is...

And finally look at the 4 last rows of the screenshot (bytes 128 to 255). This is the garbage you were seeing.