Label

From Free Pascal wiki
Jump to navigationJump to search

Deutsch (de) English (en) français (fr) русский (ru)

The label keyword is used for declaration of labels (markers for unconditional jumps using goto keyword) used further in the unit/program. It is a reserved word.

Pascal

To declare a label (not defining it) label starts a section with (comma separated) identifiers:

program labelDemo(input, output, stderr);

{$goto on}
// familiarize the compiler with symbols
// that are meant as 'goto' targets
label
	// the symbol 'done' becomes of "type" label
	done;

You then define the label, that means associating a symbol with a statement and effectively in the compiled code with an address, by writing the previously declared label in a statement block followed by a colon:

var
	n: qword;
	x, sum, arithmeticMean: real;
begin
	n := 0;
	sum := 0;
	while true do
	begin
		readLn(x);
		if x < 0 then
		begin
			// leave loop and continue at 'done'
			goto done;
		end;
		sum := sum + x;
		inc(n);
	end;
done:
	arithmeticMean := sum / n;
	writeLn('arithemetic mean = ', arithmeticMean);
end.

Pascal imposes further restrictions on labels: Labels have to be associated with statements: Putting a label right before an end is not allowed. To bypass that, you can insert an empty instruction ; right after the label definition.

In addition to normal identfiers, labels may be non-negative integers.

Assembler

A label section is also required for jump targets in asm-blocks, unless the @-address-operator is used.

program sumExample(input, output, stderr);

{ iteratively calculates the sum over first n integers }
function iterativeSumFirstNIntegers(const n: longword): qword;
{$ifdef CPUX86_64} // ============= optimized implementation
// assembler modifier appended to routine declaration
assembler;
// you have to familiarize the compiler with symbols
// which are meant to be jump targets
{$goto on}
label
	isfni_iterate;
{$asmMode intel}
asm
	xor rax, rax // rax := 0
	// ecx is used as counter by loop instruction
	mov ecx, n   // ecx := n
isfni_iterate:
	add rax, qword(ecx) // rax := rax + ecx
	loop isfni_iterate  // dec(ecx)
	// if ecx <> 0 then goto isfni_iterate
	
	// the @result macro represents the functions return value
	mov @result, rax // result := rax
// note, a list of modified registers (here ['rax', 'ecx'])
//       is ignored for pure assembler routines
end;
{$else} // ========================== default implementation
var
	i: longword;
	x: qword;
begin
	x := 0; // mov rax, 0
	for i := n downto 1 do // mov ecx, n
	begin
		x := x + i; // add rax, ecx
	end; // loop isfni_iterate
	iterativeSumFirstNIntegers := x; // mov @result, rax
end;
{$endif}

// M A I N =================================================
var
	n: longword;
begin
	readLn(n);
	writeLn(iterativeSumFirstNIntegers(n));
end.

Of course in a production program, you would use an algorithm applying the formula sum := (n * (n + 1)) div 2 (“Gaussian sum formula”).

See also