Saving a process ID when detaching a command in Bash

: is a shell builtin that is basically equivalent to the true command. It is often used as a no-op eg after an if statement. You can read more about it in this question from stack overflow.

The ${varname=value} basically means set the value of $varname to value if $varname is not already set, and then return the value of $varname. Though if you try to run that at the command line it will try to run the value returned. Putting the : in front as a no-op prevents bash from trying to run the value.

Note there are two slightly different forms:

${varname:=value}

sets varname to value if varname is either unset or null.

${varname=value}

only sets the value of varname if varname is currently unset (i.e., it will not change varname from "" to value)

(Thank you to chepner for clarifying that in a comment).

Someone else referencing this method

you could wrap this into a simple shell that waits for some input at the end of your command. For example, this one achieves this:

#!/bin/bash
cmd=$*
tmpcmd=/tmp/cygwr.$$
echo "#!/bin/bash" > $tmpcmd
echo "$cmd" >>  $tmpcmd
echo "read line" >>  $tmpcmd

chmod +x $tmpcmd
cygstart /bin/bash $tmpcmd
rm -f $tmpcmd

• copy this to a file, say ~/cygwrap.sh

• chmod +x ~/cygwrap.sh and then invoke it, example:

  ~/cygwrap.sh pwd