Introduce AssignExpression to support multiple assignments

[oxisto]

This adds a new statement class, called AssignExpression as well as TupleType.

The AssignExpression holds a list of lhs and rhs for the respective values on the left and right side. Furthermore, it contains a list of declarations, since some languages can declare variables in an assignment-like statement (either implicitly or explicitly). The recommended way to fill these declarations are in an extra pass that runs after all language frontends have been executed.

[KuechA]

I'm surprised that there's no change necessary to the ControlFlowSensitiveDFGPass.

[sonarcloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
1 Code Smell

53.8% Coverage
0.0% Duplication

[oxisto]

I'm surprised that there's no change necessary to the ControlFlowSensitiveDFGPass.

To be honest, I have just not touched that yet, because for me this pass is black voodoo magic. There are some checks in the pass for = assignments in binary operators. Those probably need to be adjusted to the AssignExpression.

[KuechA]

I'm surprised that there's no change necessary to the ControlFlowSensitiveDFGPass.

To be honest, I have just not touched that yet, because for me this pass is black voodoo magic. There are some checks in the pass for = assignments in binary operators. Those probably need to be adjusted to the AssignExpression.

Ok, I guess I should look at it but we can do that in a separate PR. I just realized that we should probably also change the assign functions (or offer a "BinaryOperator" and a "AssignmentExpression" one) in the Fluent API and offer a tuple type there as well, right?

[oxisto]

I'm surprised that there's no change necessary to the ControlFlowSensitiveDFGPass.

To be honest, I have just not touched that yet, because for me this pass is black voodoo magic. There are some checks in the pass for = assignments in binary operators. Those probably need to be adjusted to the AssignExpression.

Ok, I guess I should look at it but we can do that in a separate PR. I just realized that we should probably also change the assign functions (or offer a "BinaryOperator" and a "AssignmentExpression" one) in the Fluent API and offer a tuple type there as well, right?

Yes that would be good; I would only offer the AssignExpression one. We want to get rid of the binary op assignment at some point.