Cisco IOS.sh supports an environment variable called PATH. It contains a percent character (%) separated list of names and directories, which tell the Cisco IOS.sh where to look when executing commands.

Elements that can be included in the PATH variable are:

• CLI
• User-defined functions
• Built-in functions
• User-defined file system directories

The order of the elements on the PATH variable indicates the order of places that the Cisco IOS.sh looks for commands. For example:

    "PATH=%CLI%;%Userfunctions%;%Builtins%;tmpsys:/;disk0:". 

In this case, the order of the command lookup is in the following order:

1. If the entered command matches an IOS CLI, then the matched IOS CLI will be invoked.
2. If the entered command matches a user-defined Cisco IOS.sh function, then this user-defined Cisco IOS.sh function will be invoked.
3. If the entered command matches a built-in function, then this built-in function will be invoked.
4. If the entered command matches a script in "tmpsys:lib/tcl" directory (is a virtual file system mapped to a part of the IOS image), the script will be invoked.
5. If the entered command is found in disk0:, the script will be executed.

There are five built-in variables that are set on command completion. These variables indicate the status of the command that was just executed.

• prc_change_mode
• prc_change_type
• prc_error_code
• prc_failure_type
• prc_ha_sync

The PRC variables contain the value of the PRC, as documented by IOS CLI documentation. In addition, the "$?" variable contains zero (0) on successful execution, and non-zero on failure. The use of the question mark is a historical precedent. Unfortunately, a question mark cannot be typed at the IOS prompt unless it is preceded by the IOS escape character 'Ctrl-V'.

An "if" statement is used to conditionally execute sets of statements. The "if" loop syntax is as follows:

if
 <test>
  statements
elif <test>
  statements
else 
  statements
fi

The "elif" and "else" commands are optional, and there can be multiple "elif" statements. The "end" for "if" statements is "fi" which is "if" backwards. This is done for historical reasons.

There are a set of conditional tests available for tests, which are prefixed by either "[[" or "((". (See conditional testing below).

In addition, built-in functions or CLI commands can be used as the test, and the PRC of the command will be used to determine TRUE or FALSE.

For loop syntax:

A "for" loop is a control construct that loops over a given set of values, and executes a set of statements with a variable set to each of the values in turn. Here is a simple example.

for x in 1 2 3 4 5
do
    echo $x
done

The loop will print out each of the values 1,2,3,4,5 in turn.

A 'while" statement is used the same test syntax as an 'if' statement, and executes the statements in the following block until the test returns FALSE. An example is

let x=0
while (( x++ < 10 )); do
  echo $x
done

This will print out the numbers from 1 to 10

Users can exit a 'while' or 'for' loop early by using a 'break' statement. Here is an example.

n=0
while true; do
    let n++
     if [[ $n -le 10 ]]; then
        echo $n
    else
break;
    fi
done

In addition to 'break', the 'continue' statement can be used to stop executing statements in the loop, and restart the loop. A 'while' statement that is interrupted by a 'continue' will simply stop executing statements in the body of the loop and start the 'while' statement over again. A 'for' statement that is interrupted by a continue will also start over, except that it will assign the next element in the list of values before doing so.

for i in 0 1 2 3 4 5 6
do
    statements1      #Executed for all values of ''I'', up to a disaster-condition if any.
    statements2
    if interface ethernet$i/0
    then
      statements
    else
      continue
    fi
    statements3
done

Note	"Continue" and 'break' statements have no effect outside of a loop.

Logical testing is invoked in 'if' or 'while' statements using the

'[[' operator. Here is an example.

if [[ "abc" = $foo ]]; then
   echo '$foo is equal to abc'
fi

If the variable '$foo' contains the string 'abc', then it will print the message.

There are quite a few different logical tests that can be used within

the '[[' to ']]'.

The following tests can be used on files in the file system:

-a <filename> - return TRUE if file exists

-e <filename> - same as -a

-d <dirname> - TRUE if arg is a directory

-f <filename> - TRUE if file is a regular file

-x <filename> - TRUE if file is executable

-r <filename> - TRUE if file is readable by user

-w <filename> - TRUE if file is writable by user

The following tests can be used on strings:

-z <string> - TRUE if string is 0 length

-n <string> - TRUE if string is not 0 length

<string> <= <string> - TRUE if first is lexographically less than second

<string> == <regex> - TRUE if regex matches ALL of string

<string> != <regex> - TRUE if regex does not match string

<string> =~ <regex> - TRUE if regex matches anything in string vars $v0-$v9 are set to contain matches that are enclosed in parenthesis.

<expr> && <expr> - TRUE if both expressions are non-null

<expr> || <expr> - TRUE if either expression is non-null

Additionally, there are some numerical tests that can be used here. These are intended to be used with variables that contain integers (floating point numbers are not supported), and will give unexpected results if given strings instead of numbers.

<varname>++ - postincrement a variable

<varname>-- - postdecrement a variable

<expr> * <expr> - TRUE if multiplication result is non-zero

<expr> / <expr> - TRUE if division is non-zero, and denominator is non-zero

<expr> + <expr> - TRUE if sum is non-zero

<expr> - <expr> - TRUE if difference is non-zero

<nexpr> -lt <nexpr> - TRUE if first numeric expr is < second

<nexpr> -le <nexpr> - TRUE if first numeric expr is <= second

<nexpr> -eq <nexpr> - TRUE if numeric expressions are equal

<nexpr> -ge <nexpr> - TRUE if first numeric expr is >= second

<nexpr> -gt <nexpr> - TRUE if first numeric expr is > second

The '((' can be used in tests. It has a different set of possibilities that can be performed than the '[[ ]]' syntax.

In the '(( ))' syntax, variables can be referenced without the '$' prefix. Here is the list of tests that are available. All of the following expressions expect numeric values. If the result is non-zero, it will evaluate to TRUE.

<varname> - evaluates to the variable value

<varname>++ - postincrement

<varname>-- - postdecrement

!<expr> - logical NOT

~<expr> - bitwise negation

<expr> * <expr> - multiplication

<expr> / <expr> - division

<expr> + <expr> - addition

<expr> - <expr> - subtraction

<expr> < <expr> - comparison. evaluates to 0 or 1

<expr> <= <expr> - comparison. evaluates to 0 or 1

<expr> == <expr> - comparison. evaluates to 0 or 1

<expr> >= <expr> - comparison. evaluates to 0 or 1

<expr> > <expr> - comparison. evaluates to 0 or 1

<expr> & <expr> - bitwise AND

<expr> | <expr> - bitwise OR

<expr> ^ <expr> - bitwise XOR

<expr> << <expr> - left shift

<expr> >> <expr> - right shift

<expr> && <expr> - logical AND

<expr> || <expr> - logical OR

<varname> = <expr> - assignment

<varname> += <expr> - assign var + expr to var

<varname> -= <expr> - assign var - expr to var

<varname> *= <expr> - assign var * expr to var

<varname> /= <expr> - assign var / expr to var

<varname> <<= <expr> - left shift value of var by expr bits, assign to var

<varname> >>= <expr> - right shift value of var by expr bits, assign to var

( <expr> ) - value of expression

Parenthesis can be used to group expressions. However, parenthesis must be quoted or escaped in order to be used