grep1-5  Exercise 3: USD Without a Dollar Sign, Other Currencies, and an Error That Is Not 'No Match'
=====================================================================================================

cat > prices.txt <<'EOF'
Price 100 USD, 250 EUR, 75 USD, $30 USD
EOF

echo "USD amounts without a dollar sign in front:"
grep -oP '(?<!\$)\b\d+(?= USD)' prices.txt
echo
echo "amounts in another currency:"
grep -oP '\b\d+(?= (?!USD)[A-Z]{3})' prices.txt
echo
head -c 30 /dev/zero | tr '\0' 'a' > aaa.txt; echo b >> aaa.txt
echo "a repeat inside a repeat (Perl mode):"
grep -P '^(a+)+$' aaa.txt; echo "status $?"
echo "the same idea written simply:"
grep -P '^a+$' aaa.txt; echo "status $?"
echo "and in extended mode:"
grep -E '^(a+)+$' aaa.txt; echo "status $?"


Output:

USD amounts without a dollar sign in front:
100
75

amounts in another currency:
250

a repeat inside a repeat (Perl mode):
grep: exceeded PCRE's backtracking limit
status 2
the same idea written simply:
status 1
and in extended mode:
status 1

WHY THIS WORKS AS AN ANSWER
---------------------------
(?<!\$) says 'not preceded by a dollar sign' and (?= USD) says 'followed by USD', so 100 and 75 qualify and $30 does not. \b stops the match starting in the middle of a number.
The second pattern uses a lookahead that itself contains a negative lookahead: after the number, a space, then (?!USD) 'not USD' and three capital letters. That finds 250 (EUR). Lookarounds can be nested.
The last three commands: the repeat inside a repeat makes grep -P back up through a huge number of ways to split the a's, hit its limit and report an ERROR (status 2). The plain version a+ gets the same answer, 'no match' (status 1), at once; extended mode reaches it too. A script must treat 2 and 1 differently.