Any

I don't know if there is much harm from mysql_config returning the options even if the server was built without embedded; but I think the result that MySQL achieved by their change is awfully confusing.

# Source tree was built with WITHOUT_SERVER=ON

$ scripts/mysql_config --libmysqld-libs

Usage: scripts/mysql_config [OPTIONS]

Options:

        --cflags         [-I/usr/local/mysql/include -fabi-version=2 -fno-omit-frame-pointer]

        --cxxflags       [-I/usr/local/mysql/include -fabi-version=2 -fno-omit-frame-pointer]

        --include        [-I/usr/local/mysql/include]

        --libs           [-L/usr/local/mysql/lib -lmysqlclient -lpthread -lm -lrt -ldl]

        --libs_r         [-L/usr/local/mysql/lib -lmysqlclient -lpthread -lm -lrt -ldl]

        --plugindir      [/usr/local/mysql/lib/plugin]

        --socket         [/tmp/mysql.sock]

        --port           [0]

        --version        [5.7.10]

        --variable=VAR   VAR is one of:

                pkgincludedir [/usr/local/mysql/include]

                pkglibdir     [/usr/local/mysql/lib]

                plugindir     [/usr/local/mysql/lib/plugin]

I could understand if it properly reported the absence of the [embedded] server, or even if it returned an empty result, but complaining about improper usage of an option is silly – how can a user understand from this what is going on? Do you really want it to work this way?

I could understand if it properly reported the absence of the [embedded] server, or even if it returned an empty result, but complaining about improper usage of an option is silly – how can a user understand from this what is going on? Do you really want it to work this way?

The point is to "exit 1" so that a "configure" script fails, which it hopefully would check for.
Setting it blank could have the effect of failing at the same point in the compile as a filled in version.