HP OpenVMS Systems Documentation

HP Pascal for OpenVMS
Language Reference Manual

A D_floating-point value is represented by eight contiguous bytes. The bits are numbered from the right, 0 through 63, as shown in Figure A-4.

Figure A-4 D_floating-Point Data Representation

A D_floating-point value is specified by its address A, the address of the byte containing bit 0. The form of this value is identical to that of a F_floating-point value except for an additional 32 low-significance fraction bits. Within the fraction, bits of increasing significance are numbered 48 through 63, 32 through 47, 16 through 31, and 0 through 6.

For More Information:

• On D_floating-point range and precision ( Section 2.2)

A G_floating-point value is represented by eight contiguous bytes. The bits are numbered from the right, 0 through 63, as shown in Figure A-5.

Figure A-5 G_floating-Point Data Representation

A G_floating-point value is specified by its address A, the address of the byte containing bit 0. The form of this value is sign magnitude as follows:

• Bit 15 is the sign bit.
• Bits 14 through 4 are an excess 1024 binary exponent.
• Bits 3 through 0 and 63 through 16 represent a normalized 53-bit fraction without the redundant most significant fraction bit. Within the fraction, bits of increasing significance go from 48 through 63, 32 through 47, 16 through 31, and 0 through 3.

For More Information:

• On G_floating-point range and precision ( Section 2.2)

A T_floating-point value is represented by 8 contiguous bytes. The bits are numbered from the right 0 through 63, as shown in Figure A-6.

Figure A-6 T_floating-Point Data Representation

A T_floating-point value is specified by its address A, the address of the byte containing bit 0. The form of this value is sign magnitude as follows:

• Bit 63 is the sign bit (0 for positive numbers, 1 for negative numbers).
• Bits 62 through 52 are an excess 127 exponent.
• Bits 51 through 0 are a normalized 53-bit fraction with the redundant most significant fraction bit not represented.

For More Information:

• On T_floating-point range and precision ( Section 2.2)

An H_floating-point value is represented by 16 contiguous bytes. The bits are numbered from the right 0 through 127, as shown in Figure A-7.

Figure A-7 H_floating-Point Data Representation

An H_floating-point value is specified by its address A, the address of the byte containing bit 0. The form of this value is sign magnitude as follows:

• Bit 15 is the sign bit.
• Bits 14 through 0 are an excess 16,384 binary exponent.
• Bits 127 through 16 are a normalized 113-bit fraction with the redundant most significant fraction bit not represented. Within the fraction, bits of increasing significance go from 112 through 127, 96 through 111, 80 through 95, 64 through 79, 48 through 63, 32 through 47, and 16 through 31.

For More Information:

• On H_floating-point range and precision ( Section 2.2)

An X_floating-point value is represented by 16 contiguous bytes. The bits are numbered from the right 0 through 127, as shown in Figure A-8.

Figure A-8 X_floating-Point Data Representation

An X_floating-point value is specified by its address A, the address of the byte containing bit 0. The form of this value is sign magnitude as follows:

• Bit 127 is the sign bit.
• Bits 112 through 126 are an excess 16,383 binary exponent.
• Bits 0 through 111 are a normalized 112-bit fraction with the redundant most significant fraction bit not represented.

For More Information:

• On X_floating-point range and precision ( Section 2.2)

A.3.3 Representation of Nonstatic Types and Variables

Each nonstatic data type has some storage associated with it, called the control part. Figure A-9 shows the layout of a control part of a nonstatic data type.

Figure A-9 Storage of Nonstatic Data Types

In the top portion of the control part, HP Pascal stores each actual discriminant of the schema type in a longword of storage. The additional information piece of the control part varies in content and size depending on the type specification, and can contain any of the following:

• No information (if the schema type is simple), as follows:

  TYPE A_Char( x,y : CHAR ) = x..y;

• Control parts of nested discriminated schema types, as follows:

  TYPE My_Record( a, b : INTEGER ) = RECORD f1 : STRING( a ); f2 : STRING( b ); END;

• Values for all expressions appearing in the type definition, as follows:

  TYPE My_Subrange( a, b : INTEGER ) = a..a+b;

  
  HP Pascal evaluates the expression a+b when the schema type is discriminated and saves the result in the control part.
• The total size of the data part, if it can vary based on actual discriminants, as follows:

  TYPE Arr( a, b : INTEGER ) = ARRAY[a..b] OF REAL;

If you declare more than one variable of a discriminated schema type, each variable shares the information in the control part for that type.

A.3.3.2 Representation of Variables of Nonstatic Types

When allocating storage for a variable of a nonstatic type, HP Pascal allocates a pointer part and a data part. HP Pascal allocates and initializes the pointer part (to point to the data part); you cannot access the pointer part in your program. HP Pascal associates each object with a control part according to its data type.

If the type of the object involves more than one nonstatic type, HP Pascal associates that object with all applicable control parts. Consider the following example:

TYPE
   Sub_Range( a, b : INTEGER ) = a..b;

VAR  {x requires information in two control parts:}
   x : ARRAY[Sub_Range( i, j ), Sub_Range( k, l )] OF INTEGER;

Figure A-10 shows the layout for an object of a nonstatic type.

Figure A-10 Storage of Variables of Nonstatic Types

In Figure A-10, the pointer part is directly accessible by your program. The data part is allocated in heap when you use the NEW procedure. For example:

TYPE
   dstr = STRING( I );

VAR
   ptr : ^dstr;
{In the executable section:}
NEW( ptr );

You cannot create variables of undiscriminated types, but you can also show the representation for pointers to nonstatic types (except undiscriminated schema).

Figure A-11 shows the layout for a pointer to an undiscriminated schema type.

Figure A-11 Storage of Pointer Variables to Undiscriminated Schema Types

In Figure A-11, the control part and data part are allocated together by a call to the NEW procedure. Each object allocated this way has its own control part since the base type of the pointer is undiscriminated and does not have a control part. Consider the following example:

VAR
   x, y, z : ^STRING;
{In the executable section:}
NEW( x, 10 );   {x has a control and data part}
y := x;         {y points to same control and data part as x}
NEW( z, 10 );   {z has separate control and data parts}

Each variable created by NEW contains a unique control part attached to the data part.

A.3.3.3 Representation of Nonstatic Record Fields

If a record object contains a field of a nonstatic type, HP Pascal stores the field in one piece of storage within the record's storage (HP Pascal does not create a pointer part and a data part). HP Pascal determines the offset of the object by accessing the information in the control part of the field's data type and information in the control part of the record.

If you need to write portable code, you should not use the language features that are HP Pascal extensions. The following sections provide information on HP Pascal extensions:

• Extensions to the unextended Pascal standards ( Section B.1)
• Extensions to the Extended Pascal standard ( Section B.2)

For More Information:

• On Pascal standards ( Section 1.1)

Table B-1 summarizes the language features provided in HP Pascal that are not part of the unextended Pascal language definitions.

Table B-2 summarizes the language features provided in HP Pascal that are not part of the Extended Pascal language definitions.