mjhjmtu

Say I have this:

delete_lock () 
 if grep -q 'PATTERN'; then
 # some command here
 fi
 cat >/dev/null


node foo.js | delete_lock

say the node.js process writes "foodog" to delete_lock, or it sends "sandbag". (It could be any unique name...)

How can I read that in the delete_lock function? delete_lock is going to receive some unique name and then delete the lock with that name from the filesystem...how can I do that?

The only thing that I can guess, is something like this:

 delete_lock () 
 if grep -q 'lockname:xxx'; then
 release_lock $xxx
 fi
 cat >/dev/null
 

but how do I reference the variable $xxx lol?

I'm assuming your xxx is a stand-in for a pattern of some sort.

delete_lock () 
 LOCK="$(grep -m1 -o "lockname:xxx")"
 if [ -n "$LOCK#lockname:" ]; then
 release_lock "$LOCK#lockname:"
 fi


node foo.js | delete_lock

This will stop processing data once it observes the first match for lockname:xxx. It then uses parameter expansion to remove the lockname: prefix and, assuming there was a match (e.g. xxx, runs release_lock on it (e.g. release_lock "xxx").

If you don't want the node call to be cut short, or you want more than one match, remove the -m1 option to grep, though note the quotes may not work with your release_lock code (and be careful about removing those quotes, you don't want to allow rogue characters!).