The updated spec is still ambiguous. The phrase: "ignoring three spaces at the end of the string" can be interpreted several ways. Any of them can be handled using standard variable expansions or conditional expressions in bash or any other shell that handles basic POSIX-mandated variable expansions.

Does it mean that the test fails if there aren't at least three spaces at the end of the string? And, then the test passes only if three spaces appear anywhere in the remainder of the string? If so, try:

Code:

[ "${var%   *   }" != "${var}" ] && echo pass || echo fail

Does it mean that the test passes if three spaces appear anywhere in the string other than the last three characters in the string? Try:

Code:

[ "${var%   ?*}" != "${var}" ] && echo pass || echo fail

Does it mean that the test passes if there are three spaces followed by a non-space character anywhere in the string? If so, try:

Code:

[ "${var%   [! ]*}" != "${var}" ] && echo pass || echo fail

Does it mean that the test passes if there are three spaces followed by any character (including another space) anywhere in the string, but if the string ends with three spaces none of those three spaces can be used in the match? If so, try:

Code:

tmp="${var%   }"
[ "${tmp%   *}" != "${tmp}" ] && echo pass || echo fail

Obviously, any of the above can be changed from using single square brackets:

Code:

[ "${var%pattern}" != "${var}" ] && echo pass || echo fail

to:

Code:

[[ "${var%pattern}" != "${var}" ]] && echo pass || echo fail
                  or
test "${var%pattern}" != "${var}" && echo pass || echo fail
                  or
if [ "${var%pattern}" != "${var}" ]
then    echo pass
else    echo fail
fi
                  etc.