Understanding CLI Scripting Errors

When an error is generated, an exception is thrown. The exception is generally an object that contains the following members:

code - a numeric code associated with the error
message - a human-readable message associated with the error

Exceptions can be caught and handled, or they may be thrown out of the script environment. If a script environment has an uncaught exception, the CLI will display the details. For example:

hostname:> script run('not a cmd')
error: uncaught error exception (code EAKSH_BADCMD) in script: invalid command
       "not a cmd" (encountered while attempting to run command "not a cmd")

You could see more details about the exception by catching it and dumping it out:

hostname:> script try { run('not a cmd') } catch (err) { dump(err); }
{
   toString: <function>,
   code: 10004,
   message: 'invalid command "not a cmd" (encountered while attempting to
                      run command "not a cmd")'
}

This also allows you to have rich error handling; for example:

#!/usr/bin/ksh -p

ssh -T root@hostname <<EOF
script
       try {
               run('shares select default select $1');
       } catch (err) {
               if (err.code == EAKSH_ENTITY_BADSELECT) {
                       printf('error: "$1" is not a share in the ' +
                           'default project\n');
                       exit(1);
               }

               throw (err);
       }

       printf('"default/$1": compression is %s\n', get('compression'));
       exit(0);
EOF

If this script is named share.ksh and run with an invalid share name, a rich error message will be generated:

$ ksh ./share.ksh bogus
error: "bogus" is not a share in the default project