Triangulation operator<< prints the infinite vertex

[mglisse]

(from https://sympa.inria.fr/sympa/arc/cgal-discuss/2024-03/msg00000.html)

When printing a triangulation with operator<<, the dD class has a surprising behavior: it prints the infinite vertex. In practice, this can give

8
3 -1 2 -3
3 1 -2 -3
3 -1 -2 -3
3 1 2 3
3 -1 2 -3
3 1 -2 3
3 -1 -2 3
3 -1 2 3
3 1 2 -3

i.e. it announces 8 points then prints 9, where the first one randomly happens to match one of the other points.

See

cgal/Triangulation/include/CGAL/Triangulation.h

Lines 1413 to 1418 in 8c65aff

	// infinite vertex has index 0 (among all the vertices)
	index_of_vertex[tr.infinite_vertex()] = i++;
	if(IO::is_ascii(os))
	os << *tr.infinite_vertex() <<"\n";
	else
	write(os, *tr.infinite_vertex());

which explicitly prints the infinite vertex and compare with the code in Triangulation_3

cgal/Triangulation_3/include/CGAL/Triangulation_3.h

Lines 2507 to 2510 in 8c65aff

	V[tr.infinite_vertex()] = 0;
	for(i=1; i <= n; i++)
	{
	os << *TV[i];

which skips it and starts with the first finite vertex.

Note that while printing the point of the infinite vertex is a bad idea, printing its data/info looks like it could be useful.
Unlike Triangulation, Triangulation_3 does not print the info field of Triangulation_vertex_base_with_info_3 though.

[afabri]

@mglisse do you agree that as a first step we just do not write the infinite vertex?

[mglisse]

@mglisse do you agree that as a first step we just do not write the infinite vertex?

Ok.
It will be a small regression in that the info of the infinite vertex will be lost, but hopefully that causes less trouble than the extra vertex?
I don't know if there are tests already that check that << and >> do something consistent.
An alternative would be to print a recognizable placeholder in place of the coordinates of the infinite vertex (so the line is still there in the output, with the data of the infinite vertex).
I never use those functions myself...