[PoC] In-place struct initialization

[wilzbach]

So this is my WIP hack to have in-place struct initialization into DMD.
As during the parsing stage we don't have much information about the struct and can only parse a StructInitializer, I created an intermediate StructLiteralExp2 that will be converted into a StructLiteralExp during expression semantic.
I would like to have a more elegant solution and this still needs a lot of work, but it already passes the test suite for the common cases of this DIP and I thought it's good to show that it's possible to do this in dmd for submitting the DIP. Also I get to see what the CIs say.

What's in-place struct initialization?

Full DIP: dlang/DIPs#71

tl;dr:

struct StructFoo
{
    int a;
}

void bar(StructFoo foo)
{
    assert(foo.a == 3);
}

StructFoo f = {a: 1};
f = StructFoo({a: 2});
assert(f.a == 2);
bar(StructFoo({a: 3}));

[dlang-bot]

Thanks for your pull request, @wilzbach!

Bugzilla references

Your PR doesn't reference any Bugzilla issue.

If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog.

Testing this PR locally

If you don't have a local development environment setup, you can use Digger to test this PR:

dub fetch digger
dub run digger -- build "master + dmd#8460"

[jacob-carlborg]

Why do you need a new expression type?

[RazvanN7]

Can't you just use the existing StructLiteralExp and add the fields fields you want + a bool field that tells you if StructLiteralExp is used for in-place initialization?

[andralex]

status?

[jacob-carlborg]

status?

@andralex there's a DIP that hasn't been approved. Personally I don't think a DIP is needed. But we seem to disagree on the syntax.

[wilzbach]

status?

This is a working PoC, but there are multiple ways to integrate this feature in the D grammar (see the DIP for more details) and as there was never a consensus on which syntax to use. @andralex what are your thoughts on the syntax for this?
Once that's decided, polishing & merging this would be easy...

[dkorpel]

This has been superseded by named arguments, which have been implemented for struct constructors.