C and C++ to JSON schema mapping

The DFHLS2JS utility program supports mappings between C and C++ data types and JSON schema definitions.

How C and C++ names are converted to JSON

C and C++ names are converted to JSON names according to the following rules:

  1. Characters that are not valid in JSON property names are replaced with ' X '.

    For example, monthly-total becomes monthlyXtotal.

  2. Duplicate names are made unique by the addition of one or more numeric digits.

    For example, two instances of year become year and year1.

How C and C++ data types are mapped to JSON

DFHLS2JS maps C and C++ data types to schema elements according to Table 1.

The _Packed qualifier is supported for structures.

Restriction:
  • C and C++ types that are not shown in Table 1 are not supported by DFHLS2JS.
  • Header files must contain a top level struct instance.
  • You cannot declare a structure type that contains itself as a member.
  • The following C and C++ data types are not supported:
    • decimal
    • long double
    • wchar_t (C++ only)
  • The following are ignored if they are present in the header file.
    Storage class specifiers:
    • auto
    • register
    • static
    • extern
    • mutable
    Qualifiers
    • const
    • volatile
    • _Export (C++ only)
    Function specifiers
    • inline (C++ only)
    • virtual (C++ only)
  • The header file must not contain these items:
    • Unions
    • Class declarations
    • Enumeration data types
    • Pointer type variables
    • Template declarations
    • Predefined macros; that is, macros with names that start and end with two underscore characters ( __ )
    • The line continuation sequence (a \ symbol that is immediately followed by a newline character)
    • Prototype function declarators
    • Preprocessor directives
    • Bit fields
    • The __cdecl (or _cdecl) keyword (C++ only)
  • The application programmer must use a 32-bit compiler to ensure that an int maps to 4 bytes.
  • The following C++ reserved keywords are not supported:
    • explicit
    • using
    • namespace
    • typename
    • typeid
  • If the MAPPING-LEVEL parameter is set to 1.2 or higher and the CHAR-VARYING parameter is set to NULL, character arrays are mapped to an string and are processed as null-terminated strings.
  • If the MAPPING-LEVEL parameter is set to 1.2 or higher and the CHAR-VARYING parameter is set to BINARY, character arrays are mapped to xsd:base64Binary and are processed as binary data.
  • If the MAPPING-LEVEL parameter is set to 1.2 or higher and the CHAR-VARYING parameter is set to COLLAPSE, <xsd:whiteSpace value="collapse"/> is generated for strings.
Table 1. Mapping reference for C and C++ data types
C and C++ data type Schema simpleType
char[ z ] 
"type":"string"
"maxlength":z
char16_t[ n ]

At mapping level 4.0 and higher:

"type":"string"
"maxlength":n

At run time, CICS® populates the application data structure field with UTF-16 data.

char[8]

Supported at mapping level 3.0 and higher when DATETIME=PACKED15 

"type":"string"
"format":"date-time"
The format of the timestamp is defined by RFC3339.

char
short
int
long
long long
 
"type":"integer",
"minimum":- (n+1),
"maximum":n
where  n  is the maximum value that can be represented by the primitive.

unsigned char
unsigned short
unsigned int
unsigned long
unsigned long long
 
"type":"integer",
"minimum":0,
"maximum":n
where  n  is the maximum value that can be represented by the primitive. 
bool
(C++ only) 
"type":"boolean"
 
float
Supported at mapping level 1.2 and higher. 
"type":"number",
"format":"float"
Note: The IBM® Hexadecimal Floating Point (HFP) data representation is not exactly the same as the IEEE-754-1985 representation used for JSON. Some values may not convert exactly from one representation to the other. Some extremely large or small values might not be valid for float data types. Some values may lose precision when converted to or from HFP representation. If precise conversions are important, consider replacing use of float data types with fixed precision alternatives. 
double
Supported at mapping level 1.2 and higher. 
"type":"number",
"format":"double"
Note: The IBM Hexadecimal Floating Point (HFP) data representation is not exactly the same as the IEEE-754-1985 representation used for JSON. Some values may not convert exactly from one representation to the other. Some extremely large or small values might not be valid for double data types. Some values may lose precision when converted to or from HFP representation. If precise conversions are important, consider replacing use of double data types with fixed precision alternatives.
type name [ n ] For primitives: 
"type":"array"
"maxItems":n
"minItems":n
"items":{
    "type":"object",
    "properties":{name:{type JSON}}
    "required":[name]
         }
For structs: 
"type":"array"
"maxItems":n
"minItems":n
"items":{type JSON}
Where type JSON is the JSON schema representation of the C or C++ type.