ORD and CHR a file in Bash

问题:

I build ord and chr functions and they work just fine. But if I take a file that contains \n, for example:

hello
CHECK THIS HIT
YES

when I ord everything I don't get any new line values. Why is that? I'm writing in Bash.

Here is the code that I am using:

function ord { 
    ordr="`printf "%d\n" \'$1`" 
}

TEXT="`cat $1`"
for (( i=0; i<${#TEXT}; i++ )) 
do 
    ord "${TEXT:$i:1}" 
    echo "$ordr" 
done

回答1:

Your ord function is really weird. Maybe it would be better to write it as:

function ord { 
    printf -v ordr "%d" "'$1"
}

Then you would use it as:

TEXT=$(cat "$1")
for (( i=0; i<${#TEXT}; i++ )); do 
    ord "${TEXT:$i:1}" 
    printf '%s\n' "$ordr"
done

This still leaves two problems: you won't be able to have null bytes and you won't see trailing newlines. For example (I called your script banana and chmod +x banana):

$ ./banana <(printf 'a\0b\n')
97
98

Two problems show here: the null byte is removed from Bash in the TEXT=$(cat "$1") part, as a Bash variable can't contain null bytes. Moreover, this step also trims trailing newlines.

A more robust approach would be to use read:

while IFS= read -r -n 1 -d '' char; do
    ord "$char"
    printf '%s\n' "$ordr"
done < "$1"

With this modification:

$ ./banana <(printf 'a\0b\n')
97
0
98
10

Note that this script will depend on your locale. With my locale (LANG="en_US.UTF-8):

$ ./banana <(printf 'a\0ℂ\n')
97
0
8450
10

whereas:

$ LANG= ./banana <(printf 'a\0ℂ\n')
97
0
226
132
130
10

That's to show you that Bash doesn't read bytes, but characters. So depending on how you want Bash to treat your data, set LANG accordingly.

---

If your script only does that, it's much simpler to not use an ord function at all:

#!/bin/bash

while IFS= read -r -n 1 -d '' char; do
    printf '%d\n' "'$char"
done < "$1"

It's that simple!