makeGlobal: Use an enum instead of magic numbers

[MetaLang]

No description provided.

[dlang-bot]

Thanks for your pull request, @MetaLang! We are looking forward to reviewing it, and you should be hearing from a maintainer soon.

Some tips to help speed things up:

• smaller, focused PRs are easier to review than big ones

• try not to mix up refactoring or style changes with bug fixes or feature enhancements

• provide helpful commit messages explaining the rationale behind each change

Bear in mind that large or tricky changes may require multiple rounds of review and revision.

Please see CONTRIBUTING.md for more information.

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.

[schveiguy]

Would it be bad to make the enum an actual type, and then make the template parameter an instance of that enum? This would prevent for instance makeGlobal!5, which then also aliases to stderr without something like that. At the very least, this should say if(_iob >= stdin_handle && iob <= stderr_handle)

[MetaLang]

Yes, I'll do that instead. My goal was to be as unintrusive as possible.

[schveiguy]

I'm not an expert on whether this will trigger -betterC problems, but this looks much better to me.

[MetaLang]

I believe we should be okay now that betterC uses C's assert.

[schveiguy]

I'm just looking at the doc changes, I didn't realize the instantiation makes it into the docs. Something to think about for another PR I think.

[PetarKirov]

I'm planning on nuking most of makeGlobal while addressing some of the thread-safety issues with std.stdio so I'm not sure if your efforts are worth it, though I haven't had the chance to work lately on this, so I'm not 100% sure how this will play with Walter's druntime PR, but I think I'll still use core.stdc.stdio.std* directly (I don't think I need to use them as template arguments).

[schveiguy]

@ZombineDev while makeGlobal exists, we might as well make it work better than it does. Obviously Andrei thinks it's worthwhile, and the whole point is to enable support for using the libraries with -betterC. I am not sure exactly how this is going to work, or whether it will work. But if this is the push, we might as well not make it uglier than it has to be.

I'll warn you that std.stdio is a nest of hacks that is really difficult to untangle. I hope you can do it, but I definitely don't have much hope.

[MetaLang]

@ZombineDev I don't think it's a huge issue since this is such a simple PR. The code is working as-is so it doesn't matter if this gets merged or not; I just thought I'd fire a quick PR off based on Steven's comment here.

[PetarKirov]

Alright, in that case feel free to proceed with this PR, as I can't give an ETA on my work, and as Steven points out we should strive to improve the status quo with what we have.