In CAST, the dimensions were declared between the type-identifier and the instance-identifier. Arrays in HAC are syntactically C-style, where the dimensions of the array follow the array's identifier.
Sparse arrays are also supported but with different syntax. In HAC, sparse arrays are always declared using a range expression in the indices. e.g. ‘inv x[1..1];’ declares a sparse 1D array populated at index 1. (Yes, some of you may find this inconvenient.)
The following table shows some examples of equivalent declarations in CAST and HAC, where inv is defined as a type.
Examples of dense and sparse array declarations:
| CAST | HAC | meaning
|
|---|---|---|
inv[2] x;
| inv x[2];
| 1D array with indices 0..1
|
inv[2] x, y;
| inv x[2], y[2];
| 2 1D arrays with indices 0..1
|
inv[2][3] x;
| inv x[2][3];
| 2D array with indices 0..1,0..2
|
inv x[2];
| inv x[2..2];
| sparse 1D array indexed 2 only
|
inv x[2..4];
| inv x[2..4];
| sparse 1D array indexed 2..4
|
inv x[2], x[4];
| inv x[2..2], x[4..4];
| sparse 1D array indexed 2, 4 only
|
A common pitfall is to declare sparse index of an array and pass port connections in the same statement, such as: ‘inv w[2](x, y, z);’. This illegal statement tries to declare an array indexed 0..1, and connect both instances with the same port parameters. In HAC, one cannot declare a collection and connect its ports in the same statement as in CAST, however, one may declare a scalar instance and connect its ports in the same statement. The proper way to instantiate and connect a sparse instance is to use a sparse range, just like in a sparse declaration: ‘inv w[2..2](x, y, z);’.