Select / case with "dynamic size"

[lilltroll]

Assume c is an array of size SIZE, e.g.

int main() {
chan c[Size];
....

Is it possible to write the following code in another way such as SIZE can be chosen at compile time ?
E.g.

#define SIZE 4

Code: Select all

do {
		t:>time;
		select {
			case slave {
				c[0]:>len;
				c[0]:>colour;
				for (int i = 0; i < len; i++)
				c[0]:>msg[i];
			}:
			{	cursorX,cursorY}=insert_str(cursorX, cursorY, msg,colour,cursorColour,table);
			break;
			case slave {
				c[1]:>len;
				c[1]:>colour;
				for (int i = 0; i < len; i++)
				c[1]:>msg[i];
			{	cursorX,cursorY}=insert_str(cursorX, cursorY, msg,colour,cursorColour,table);
			break;
			}
			case slave {
				c[2]:>len;
				c[2]:>colour;
				for (int i = 0; i < len; i++)
				c[2]:>msg[i];
			}:
			{	cursorX,cursorY}=insert_str(cursorX, cursorY, msg,colour,cursorColour,table);
			break;
			case slave {
				c[3]:>len;
				c[3]:>colour;
				for (int i = 0; i < len; i++)
				c[3]:>msg[i];
			}:
			{	cursorX,cursorY}=insert_str(cursorX, cursorY, msg,colour,cursorColour,table);
			break;
			case tc when timerafter(timec) :> timec:
...

[richard]

You can use a replicated case. A simple example is shown below.

Code: Select all

unsigned f(chanend c[], unsigned size)
{
  select {
  case (unsigned i = 0; i < size; i++)  c[i] :> int:
    return i;
  }
}

[lilltroll]

:idea: :!: If I havn't missed it, it should be added to chapter
3.6 Parallel Replication or similar.

I assumed there was an way to do it, like this.

[richard]

:idea: :!: If I havn't missed it, it should be added to chapter
3.6 Parallel Replication or similar.

I assumed there was an way to do it, like this.

Select replicators are mentioned in the 2.8 Parameterised Selection section on page 28 of Programming XC on XMOS Devices

"A more concise way to specify the top-level select statement is to use a replicator..."

[lilltroll]

... and you if you need to guarantee the timing in one case , you may reserve some time for it like...

Code: Select all

			case t when timerafter(time) :> int _: //Do not run another event
			t when timerafter(time+TIMERESERVE) :> int _; //if more than TIMERESERVE CPU cycles are available
...
...
...
t:>time;
time+=DELAY-TIMERESERVE;
break;

where the TIMERESERVE can be calculated from the worst XTA timing from the other cases.