Issue 12652 - Non-constant hash initializers should have a special-case diagnostic

[adelavais]

Hello! I am unsure if this is the best error message, but from what I can see the problem is not only for global [0] declaration (like in the issue), I discovered it in a struct [1], and it extends to classes [2] too. There is no error in main or in a function called by main [3].
[0]: https://run.dlang.io/is/RRJOBT
[1]: https://run.dlang.io/is/E53JkC
[2]: https://run.dlang.io/is/2jOMg6
[3]: https://run.dlang.io/is/Q2I4mN

[dlang-bot]

Thanks for your pull request and interest in making D better, @adelavais! 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

Auto-close	Bugzilla	Severity	Description
✓	12652	enhancement	Non-constant hash initializers should have a special-case diagnostic

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#11598"

[MoonlightSentinel]

This line is rather long, the sugession should be a second line using errorSuplemental.

[MoonlightSentinel]

The error message should rather mention that it's impossible to initialize AA at compile time instead of runtime (the difference between [1],[2] and [3],[4]).

(I assume that this distinction should be understandable for non-compiler developers but other suggestions are welcome]

[MoonlightSentinel]

No phobos in the test suite (it's also unused?)

[MoonlightSentinel]

Unnecessary?

[MoonlightSentinel]

Might as well add another global AA at module scope.

[Geod24]

A few suggestion:

• Use errorSupplemental: If you search for it in the source code, you'll see other usages. Usually we use error to print the error message, and errorSupplemental to suggest a resolution. One exception is printing instantiation stack trace.
• enum is not going to cut it, because it's still compile time;
• Try to make the message relevant to the code. In your example code, it's in a struct, so it should move to the struct constructor. Module constructor should only be suggested if the parent is a global scope. Check toParent for this.

[adelavais]

Thank you both for your feedback! I will update the PR as soon as possible.

[RazvanN7]

@adelavais are you still pursuing this?

[adelavais]

@adelavais are you still pursuing this?

Yes, I am. I will try to send a PR over the weekend.

[WalterBright]

Error messages are not capitalized

[WalterBright]

The error message does not quite make sense. The one from the bug report is better:

Associative array literals currently cannot be used to initialize globals. Try using a module constructor instead.

[WalterBright]

And remove "currently".

[WalterBright]

They aren't actually globals, either. They are statically allocated.

[RazvanN7]

@MoonlightSentinel @WalterBright @Geod24 I rebased this and addressed the concerns. Is this ready to go?

[RazvanN7]

cc @dkorpel

Addressed