CODEWITHSUNDEEP
unixksh
sheep
sheep

Reputation: 3

Unix - condition within condition

I've been looking all over the place but couldn't find an answer. In ksh, how do you do something like this:

while [ [ ! [ [ -n $var1 ] || [ [ -n $var2 ] && [ -n $var3 ] ] ] ] && [ ! [ [ -n $var1 ] && [ -n $var2 ] ] ] ]; do
    ...etc etc
done

or in pseudo/a bit easier way to see

while ((! ((var1 != none) or 
         ((var2 != none) and (var3 != none)))
      and
      (!((var1 != none) and (var2 != none)))) {
....
}

...essentially any kind of conditions that are grouped

I've found a lot on simple conditions like if [ -z $var1 ] && [ -n $var2 ]; then

but not like the one above.

Any help would be appreciated.

Upvotes: 0

Views: 111

Answers (2)

pynexj
pynexj

Reputation: 20748

It's much easier to use ksh's [[ ]] syntax. ( and ) don't need to be quoted and you can use && and || just as in C language. For example:

[[ ! ( -n $var1 && ( -n $var2 || -n $var3 ) ) ]]

Also, you don't need to double quote $var within [[ ]] which would make the code much cleaner.

Upvotes: 1

mikyra
mikyra

Reputation: 10367

There are a number of issues with the test you want to perform.

  • fragments like -n $var will fail if $var is empty as it will be expanded to a single -n.

Using -n "$var" instead is the bulletproof version here.

  • and is expressed with the -a operator not with &&

  • or is expressed with the -o operator not with ||

  • the grouping operator is () not another nested [] construct

Applying this rules your expression would come closer to looking something like that:

[ ! ( -n "$var1" -o ( -n "$var2" -a -n "$var3" ) ) -a ! ( -n "$var1" -a -n "$var2" ) ]

Note that use of spaces is mandatory here.

Depending on escaping rules you might have to escape ( and ).

As I don't know how those are handled in ksh I can't tell for sure if really necessary there, but the all shell proof version would look something like that:

[ ! '(' -n "$var1" -o '(' -n "$var2" -a -n "$var3" ')' ')' -a ! '(' -n "$var1" -a -n "$var2" ')' ]

Note that [ is an executable usually located in /usr/bin/ so there aren't any syntactic differences across different shells.

It's only different meta characters that you have to be aware of, when using [ with a different shell.

Upvotes: 1

Related Questions