Incorrect result?

[bertram-gil]

Hi guys,

Consider the following program:

use ascent::ascent;
ascent!{
	relation a__(i32, i32);
	relation c__(i32, i32, i32);
	relation e__(i32);
	relation h__(i32, i32, i32);

	e__(a) <-- a__(b, a);
	h__(e, e, e) <-- a__(d, e), c__(e, f, e), e__(e);

}

fn main() {
	let mut prog = AscentProgram::default();
	prog.a__ = vec![(88,5), (37,24), (11,91)];
	prog.c__ = vec![(32,83,88), (2,8,5)];
	prog.e__ = vec![(44,), (83,)];
	prog.h__ = vec![(38,88,18), (76,18,65), (86,73,91), (98,26,91), (76,10,14)];

	prog.run();
	println!("{:?}", prog.h__);
}

If I run the program, I get the following for the relation h__:
[(38, 88, 18), (76, 18, 65), (86, 73, 91), (98, 26, 91), (76, 10, 14), (5, 5, 5)]

Now, if I add a subgoal c__(e, a, e) in the rule for h__ to get the following new program:

use ascent::ascent;
ascent!{
	relation a__(i32, i32);
	relation c__(i32, i32, i32);
	relation e__(i32);
	relation h__(i32, i32, i32);


	e__(a) <-- a__(b, a);
	h__(e, e, e) <-- a__(d, e), c__(e, f, e), e__(e), c__(e, a, e);
}

fn main() {
	let mut prog = AscentProgram::default();
	prog.a__ = vec![(88,5), (37,24), (11,91)];
	prog.c__ = vec![(32,83,88), (2,8,5)];
	prog.e__ = vec![(44,), (83,)];
	prog.h__ = vec![(38,88,18), (76,18,65), (86,73,91), (98,26,91), (76,10,14)];

	prog.run();
	println!("{:?}", prog.h__);
}

The result for h__ changes and I get:
[(38, 88, 18), (76, 18, 65), (86, 73, 91), (98, 26, 91), (76, 10, 14)]
Adding the subgoal c__(e, a, e) should not change the result for relation h__.
The tuple (5, 5, 5) disappears for the second program which should not have appeared for the first program in the first place.

My Cargo.toml file:

[package]
name = "ascent_project"
version = "0.1.0"
edition = "2021"

# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html

[dependencies]
ascent = "0.2"

Please let me know if you cannot reproduce this or if I am doing anything wrong.

[s-arash]

Thanks for finding it. Indeed this is a bug. It has to do with the interaction of repeated vars in c__(e, f, e) and reordering of clauses. I will push a fix soon. Will you be able to move to 0.3 when the fix is pushed?

And if you need a quick workaround, change it to c__(e, f, e1), if e == e1

[bertram-gil]

Thanks for confirming it.
Yes absolutely I can move to 0.3 when it is ready :)