Translate _d_array{,set}ctor to templates

[teodutu]

This PR is part of replacing runtime hooks in DMD with templates, as described in this project.
I also provided a short description of this project in this forum post.

This PR concludes Dan Printzell's previous work:

• [WIP] Translate _d_array{,set}ctor to templates #10102
• Translate _d_array{,set}ctor to templates druntime#2655

[dlang-bot]

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

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#13116"

[WalterBright]

Missing is why do this. What problem is being solved. It adds a bunch of code.

[PetarKirov]

@WalterBright You can find more information here: https://forum.dlang.org/thread/bbhmgddyabnuxajlxgqp@forum.dlang.org

@teodutu can you please update the commit message and pull request description with a brief information about the project goals and why the change is necessary?

[teodutu]

@teodutu can you please update the commit message and pull request description with a brief information about the project goals and why the change is necessary?

I mentioned the initial proposal: dlang/project-ideas#25
I hope this provides enough context.

[teodutu]

The code should pass (more) tests when this PR is merged: dlang/druntime#3582

[RazvanN7]

@teodutu Rebase this and lets see what happens.

[RazvanN7]

src/dmd/expressionsem.d' contains trailing whitespace at line 9763

[teodutu]

This is currently blocked by this issue: https://issues.dlang.org/show_bug.cgi?id=22372

[teodutu]

In this post, I highlighted that adding a third parameter to _d_arrayctor yielded better performance than declaring the returned array inside the hook. For this reason, I chose to go back to the former approach, which is implemented in druntime here: dlang/druntime#3587.

[RazvanN7]

I'm so happy this finally got in. Congrats for this awesome work @teodutu !

[andralex]

Awesome work, thanks! I just took a look at https://github.com/dlang/druntime/pull/2655/files and there are a few opportunities for improvements:

• Instead of doing the memcpy element + postblit loop, we can do just one big memcpy for the whole thing for certain (many!) types. Introspection will figure which is which.
• Types that don't throw from ctors (information we ca n get by introspection) don't need the try/catch. So we're looking at routines tailored for categories of types.
• Could some or all of the work of enforceRawArraysConformable be done during compilation?
• auto element_size = T.sizeof; is at best redundant and at worse a pessimization. Just use T.sizeof in the call to memcpy to make sure the compiler sees it's a compile-time constant.

[radcapricorn]

To add to Andrei's notes, if these functions are going to be instantiated by user code, the try/catches also need a version(D_BetterC) version.

[ibuclaw]

This PR caused a regression. https://issues.dlang.org/show_bug.cgi?id=23181