Return to Answer

edit Further review on url_decode() and the actual 'program'...

The other weird things about url_decode() are that you are repeating your decoding across two different loops, and carefully interpreting and handling for $num when $2 = auto (which isn't 'documented'). Also, since you mentioned that you prefer a more bash-like solution, perhaps you can consider the following too so that you can even not depend on grep and sed:

url_decode() { 
    local s="$1"; local n=${2:-0}; 
    while [[ $((n--)) -gt 0 || (-z $2 && $s =~ %[[:xdigit:]]{2}) ]]; do 
        s=$(echo -e ${s//%/\\\x}); 
    done; echo $s
}

Rather than defaulting to 2, you might as well incrementally apply the decoding until you don't see %XX, where X is a hexadecimal character. This is (better) represented using the regex character class [[:xdigit:]]. =~ replaces the egrep and ${s//%/\\\x} replaces the sed. The loop condition simply says:

• countdown $n until it reaches 0, i.e. [$n - 1, 0], or
• $2 is not specified and we still have 'leftover' values to decode.

If you still prefer to stick to a safe default such as 2, replace n=${2:-0} with n=${2:-2}, and then you can drop the || condition.

A quick test:

$ url_decode 'You%2BWon%2527t%2BBe%252BMissed'
You+Won't+Be+Missed

$ url_decode 'You%2BWon%2527t%2BBe%252BMissed' 1
You+Won%27t+Be%2BMissed

$ url_decode 'You%2BWon%2527t%2BBe%252BMissed' 2
You+Won't+Be+Missed

As for your actual 'program', I don't see many major problems with it...

To check for yes (or equivalent) inputs, I think a better approach is to normalize the casing first, then apply the egrep:

$ for i in y Y Yes YES no; do tr '[[:upper:]]' '[[:lower:]]' <<< $i | egrep '^y(es|)$' -q \
    && echo $i - OK; done
y - OK
Y - OK
Yes - OK
YES - OK

To check for yes (or equivalent) inputs, I think a better approach is to normalize the casing first, then apply a case-insensitive egrep:

$ for i in y Y Yes YES no; do egrep -i '^y(es|)$' -q <<< $i \
    && echo $i - OK; done
y - OK
Y - OK
Yes - OK
YES - OK

That can be simplified too...

That can be simplified too, by putting them all within one [[ ... ]]: