In this code :
echo hello > hello.txt
read X <<< $(grep hello hello.txt)
echo $?
$? refers to the exit code of the read statement which is 0. Is there a way to know if the grep failed (e.g., if hello.txt has been deleted by another process) without splitting the read and grep in two statements (i.e., first grep then check $? then read).
Use process substitution instead of command substitution + here string:
read X < <(grep 'hello' hello.txt)
This will get you 1 when using echo $?.
PS: If grep fails it will write an error on your terminal.
If you want to suppress error then use:
read X < <(grep 'hello' hello.txt 2>/dev/null)
Just:
X=$(grep hello hello.txt)
echo $?
For general case where you want to use read to do word splitting, then use a temporary variable:
tmp=$(grep hello hello.txt)
echo $?
IFS=' ' read -r name something somethingelse <<<"$tmp"
read X <<< $()? Why not justX=$()? What part ofreaddo you need?command $(grep hello hello.txt)and then check if thegrepsucceeded.tmp=$(grep hello hello.txt); grep_exit_status=$?; command "$tmp".