How to check for inequality in bash for loop? - bash

Question

I have the following for loop in a bash script:
for (( j = i; ${1:j:3} != " "; j=j + 1 ))
do
sleep 0.1
done
printf '%s' "${letter[${1:i:j}]}"
i=$j
When run, it leads to an infinite loop of the following errors:
/home/com/morsecoder.sh: line 188: letter: bad array subscript
/home/com/morsecoder.sh: line 184: ((: ... != : syntax error: operand expected (error token is "... != ")
The problem is on the first line; the bad array subscript error is almost certainly a byproduct of that.
I can see the error is caused by my ${1:j:3} != " ". Basically, what I need is for the loop to run through the characters in a string until it finds three consecutive spaces. The string contains Morse code, and each letter is separated by 3 characters (because in American Morse, letters can contain 0, 1, or 2 spaces, so 3 is the minimum letter delimiter).
Afterward, I convert what I have detected to be a complete Morse letter to English and print it out, and then move on to the next Morse characters.
The printf part seems to be working fine, but the error here has me puzzled. I checked and I am using (( and )) properly as well as != to check for inequality. I also tried enclosing ${1:j:3} in quotes, but that did nothing. How can I rephrase the for loop so that I don't get an error about invalid syntax?

Answer 1

This form of the for loop is only for arithmetic operations. You need to use a while loop instead:
j=$i
while [[ ${1:j:3} != " " ]]; do
sleep 0.1
j=$((j+1))
done

Answer 2

Basically, what I need is for the loop to run through the characters
in a string until it finds three consecutive spaces. The string
contains Morse code, and each letter is separated by 3 characters
(because in American Morse, letters can contain 0, 1, or 2 spaces, so
3 is the minimum letter delimiter).
Well that seems an odd way to go about it. The shell has better mechanisms for this task than scanning the string by iterating over an index. For example,
# the value of $1, with the longest suffix matching glob pattern " *" removed.
letter=${1%% *}
The purpose of the sleep 0.1 in your example code is not clear to me, but if you're simply simulating the timing of receiving a signal via Morse code (complete with different timings for letters of different Morse length) then it can be addressed separately.
Afterward, I convert what I have detected to be a complete Morse
letter to English and print it out, and then move on to the next Morse
characters.
So I would approach it more like this:
morse=$1
while [[ -n "$morse" ]]; do
# extract the next letter
next_let=${morse%% *}
# sleep for a duration based on the (Morse) length of the letter
sleep "0.${#next_let}"
# print the corresponding decoded Latin letter
printf '%s' "${letter[${next_let}]}"
# remove the Morse letter and its delimiter, if any
morse=${morse:$((${#next_let}+3))}
done
That covers your loop over all the Morse letters, by the way, not just one.
Iterating over a string by index is not necessarily wrong in shell code, but it has bad code smell.