File Formats                                          compound_var(4)



NAME
     compound_var - format of serialised compound variable

DESCRIPTION
     An serialised compound variable file is text representing
     the value of a compound variable. A compound variable has the form
     name=(assignment ...) where "assignment" is a sequence of zero
     or more shell variable assignments in the form
     [<datatype><blank>]<name>=<value> (if <datatype><blank> is
     omitted the default shell datatype "string" will be used)
     seperated by newline or ';'.
     
     * DataTypes:
     <needs to be written.
       * Material:
         string
         integer (and unsigned integer)
         float (conformant to IEEE 754-2008)
         hexfloat (conformant to C99 and IEEE 754-2008)
         Array, indexed+associative
         binary (mime base64)
         compound variable
     >
     
     
     * Value encodings:
     <needs to be written.
       * Material:
         1. plain strings
         2. plain strings in double-quotes
         3. ANSI-C strings $'...'     
     >


     * Indexed arrays, sets, associative arrays:
     <needs to be written>
     
     
SEE ALSO
     ksh93(1), XML(4), libcompoundvar(3c)

NOTES
     It is recommended that applications implement compund variable
     output using the option "-C compound" or - if "-C" is already
     used - use option "-# compound".
     
     The format supports multibyte charatcers for both variable names
     and values but the users but ensure that the creator and consumers
     use a compatible character encoding.
     
     The recommended file extension for serialised compound variables
     is *.cpv

EXAMPLES
     1. Empty compound variable:
     -- snip --
     (
     )
     -- snip --

     2. Compound variable with a string member variable "foo"
     -- snip --
     (
             foo="bar"
     )
     -- snip --
     
     3. Compound variable with an indexed array with values "1", "2" and "3":
     -- snip --
     (
             typeset -a a=(
                     1
                     2
                     3
             )
     ) 
     -- snip --

     4. Compound variable with an indexed array with string values "1",
     "2" and "3":
     -- snip --
     (
             typeset -a a=(
                     1
                     2
                     3
             )
     ) 
     -- snip --

     5. Compound variable with a sparse indexed array containing strings:
     -- snip --
     (
             typeset -a a=(
                     [0]=1
                     [3]=3
                     [5]=2
             )
     )
     -- snip --

     6. Compound variable with a sparse indexed array containing
     compound variables:
     -- snip --
     (
             typeset -C -a a=(
                     [4]=(
                             msg="hello world"
			     food="roasted chicken"
                     ) 
                     [7]=(
                             msg="thank you for the fish"
			     food="tuna"
                     )

                     [9]=(
                             typeset -l -E temp=0.1
			     typeset -C nested_node=(
			          name="text node"
				  description="a text node"
				  value="hello fish"
                             )
                     )
             )
     ) 
     -- snip --

     7. Compound variable with an associative array containing
     compound variables:
     -- snip --
     (
             typeset -C -a a=(
                     ['a meal node']=(
                             msg="hello world"
			     food="roasted chicken"
                     ) 
                     ['another meal node']=(
                             msg="thank you for the fish"
			     food="tuna"
                     )

                     ['temperature difference']=(
                             typeset -l -E temp=0.1
                     )
             )
     ) 
     -- snip --


#### Material: ####
* Use sccsfile(4) as template

* String literals:
  1. plain strings
  2. Plain strings in double-quotes
  3. ANSI-C strings $'...'

* Datatypes:
  string
  integer (and unsigned integer)
  float (conformant to IEEE 754-2008)
  hexfloat (conformant to C99 and IEEE 754-2008)
  Array, indexed+associative
  binary (mime base64)
  compound variable

* ToDo:
-- snip --
<plocher> define the stability and specs for the following: Name, Grammer/Syntax,
Producer(s) and Consumer(s), as well as a Description of what the file is used for,
how its content can evolve (i.e., presume the parser stays same, but the grammer
grows, how can Consumers deal with change...)
-- snip --