adblock: initscript seems to exit on any valid argument

[brianjmurrell]

I'm not sure I'm understanding the logic/intention of this line:

packages/net/adblock/files/adblock.init

Line 23 in d16b8ca

[ -s "${adb_pidfile}" ] && { [ "${action}" = "start" ] || [ "${action}" = "stop" ] || [ "${action}" = "restart" ] || [ "${action}" = "reload" ] || [ "${action}" = "report" ] || [ "${action}" = "suspend" ] || [ "${action}" = "resume" ] || [ "${action}" = "query" ] || { [ "${action}" = "list" ] && [ -n "${1}" ]; }; } && exit 1

To me, this reads as if the PIDFILE exists (and is not empty) AND if argument 1 of the script is any of the valid actions for the script then exit 1.

Am I reading that correctly? If so why would that initscript want to exit 1 when called with a valid action?

While this seems pretty basic to me, what's making me question my understanding here is that this logic seems to have been in place for 6 years now and it just seems unbelievable to me that such a fundamental aspect of the script could be broken for that long. So I feel I must be missing something, probably pretty obvious here.

Digging further, it seems that in bc299d0 the logic of the individual [ "${action}" != "…" ] && list was inverted but without inverting the the result of the entire group to maintain the overall logic as such:

[ -s "${adb_pidfile}" ] && ! { [ "${action}" = "start" ] ||
                               [ "${action}" = "stop" ] ||
                               [ "${action}" = "restart" ] ||
                               [ "${action}" = "reload" ] ||
                               [ "${action}" = "report" ] ||
                               [ "${action}" = "suspend" ] ||
                               [ "${action}" = "resume" ] ||
                               [ "${action}" = "query" ] || {
                                   [ "${action}" = "list" ] &&
                                   [ -n "${1}" ];
                               };
                             } && exit 1

Although, quite honestly, I find even that logic a bit awkward. Not sure why the state of the ${adb_pidfile} is even being included in that test. In any case the above change/fix does seem to fix the overall problem that /etc/init.d/adblock restart didn't do anything useful.