opEquals() is not honoured with 0.14 alpha 8

[skruppy]

with Version 0.14.0.alpha8 opEquals() is not honoured anymore when asserting equality. It worked prior version 0.14. See the following MWE:

import std.stdio;
import fluent.asserts;

void main() {}

class Thing {
	int x;
	this(int x) { this.x = x; }
	override bool opEquals(Object o) {
		if(typeid(this) != typeid(o)) return false;
		alias a = this;
		auto b = cast(typeof(this)) o;
		return a.x == b.x;
	}
}

unittest {
	auto a1 = new Thing(1);
	auto b1 = new Thing(1);
	auto a2 = new Thing(2);
	
	assert(a1 == b1); // works
	assert(a1 != a2); // works
	
	a1.should.equal(b1); // a1 should equal Thing(132421593017436). Thing(132421593017402) is not equal to Thing(132421593017436)
	a1.should.not.equal(a2); // works
}

[gedaiu]

It's interesting that you have all of these edge cases... I guess you have some complex tests. Those asserts are fixed now on 0.14.0.alpha9

Please let me know how it works for you.

[skruppy]

Thanks for your quick response. I still have to test the new release in the "big" project, but for now, it seems I still have some edge cases up my sleeve https://gitlab.com/Skruppy/csg-fitter/-/jobs/1601647959 , sorry :smile:. Some of the assertion fails are from my project not fully ported to the newest dmd release, but some unfortunately are probably still fluent assers problems.

[skruppy]

I'm super sorry, but I found an other edge case. With alpha.9, the upper examples works, but by changing the variables to arrays it will fail again:

	assert([a1] == [b1]); // works
	assert([a1] != [a2]); // works
	
	[a1].should.equal([b1]); // [a1] should equal [Thing(132599446648072)]
	[a1].should.not.equal([a2]); // works

[gedaiu]

No worries! Unfortunately, I don't have much time for developing this library, so thanks for finding those use cases. Comparing object using opEqual is enabled just for the equal operation. I will reserve some time to do it for the others as well... maybe next weekend. I hope you don't need those right away.