Fix #20595 - automatically unittest destroy(T.init)

[thewilsonator]

No description provided.

[dlang-bot]

Thanks for your pull request and interest in making D better, @thewilsonator! We are looking forward to reviewing it, and you should be hearing from a maintainer soon.
Please verify that your PR follows this checklist:

• My PR is fully covered with tests (you can see the coverage diff by visiting the details link of the codecov check)
• My PR is as minimal as possible (smaller, focused PRs are easier to review than big ones)
• I have provided a detailed rationale explaining my changes
• New or modified functions have Ddoc comments (with Params: and Returns:)

Please see CONTRIBUTING.md for more information.

---

If you have addressed all reviews or aren't sure how to proceed, don't hesitate to ping us with a simple comment.

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.

⚠️⚠️⚠️ Warnings ⚠️⚠️⚠️

• In preparation for migrating from Bugzilla to GitHub Issues, the issue reference syntax has changed. Please add the word "Bugzilla" to issue references. For example, Fix Bugzilla Issue 12345 or Fix Bugzilla 12345.(Reminder: the edit needs to be done in the Git commit message, not the GitHub pull request.)

Testing this PR locally

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

dub run digger -- build "master + dmd#20598"

[thewilsonator]

cc @pbackus could you please provide a detailed explanation here?

[pbackus]

Short answer: I was wrong. Close this.

Long answer: the requirement I was thinking of is that destroying T.init must not cause undefined behavior, because T.init is always available in @safe code. But that does not mean it has to succeed—it could still throw an exception, enter an infinite loop, terminate the process, etc.

[Herringway]

Long answer: the requirement I was thinking of is that destroying T.init must not cause undefined behavior, because T.init is always available in @safe code. But that does not mean it has to succeed—it could still throw an exception, enter an infinite loop, terminate the process, etc.

I don't think you can even rely on T.init always being available in @safe code due to it still being possible to define an init() method, which may be @system.

[pbackus]

Even if a type defines an init() method, you can still obtain its .init value by using something like (T[1]).init.

There is currently no reliable way to prevent @safe code from accessing a type's .init value.

[Herringway]

There is currently no reliable way to prevent @safe code from accessing a type's .init value.

A @system init field is also legal - struct Foo { @system noreturn init; }, for example. The (T[1]).init trick doesn't work around that one. It's only a deprecation right now, so TECHNICALLY it's allowed... for now.

Whoops, testing error

[pbackus]

This works for me, even with -preview=systemVariables enabled:

struct S
{
    int n = 123;
    @disable this();
    @system noreturn init;
}

void main() @safe
{
    auto a = (S[1]).init;
    assert(a[0].n == 123); // passes
}

[thewilsonator]

Hmm:

struct S
{
    int n = 123;
    //@disable this();
    @system static immutable S init;
}

void main() @safe
{
    auto a = S.init;
    assert(a.n == 123); // passes
}

successfully compiles with -preview=systemVariables which seems like a bug to me

[pbackus]

It's definitely a bug, but it's also irrelevant to this PR which—again—should be closed.