fix issue 20138 - is expression not evaluating correctly?

[aG0aep6G]

When passing a shared const T into a shared U parameter, then no const
conversion is needed to infer U as const int. Same with inout int and
inout const int.

[dlang-bot]

Thanks for your pull request and interest in making D better, @aG0aep6G! 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
✓	20138	blocker	is expression not evaluating correctly?

⚠️⚠️⚠️ Warnings ⚠️⚠️⚠️

• Regression or critical bug fixes should always target the stable branch. Learn more about rebasing to stable or the D release process.

To target stable perform these two steps:

1. Rebase your branch to upstream/stable:

git rebase --onto upstream/stable upstream/master

2. Change the base branch of your PR to stable

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#10317"

[TurkeyMan]

I don't know how to review this...

[aG0aep6G]

PR to prepare Phobos for this fix: dlang/phobos#7147

[aG0aep6G]

For reference, dlang/druntime#2739 seems to be blocked by this.

[Biotronic]

It seems to me this fix only deals with shared T, and ignores other qualifiers - is(shared const int U == const U) should also be true (and U should be shared int), likewise for inout.

[aG0aep6G]

It seems to me this fix only deals with shared T, and ignores other qualifiers - is(shared const int U == const U) should also be true (and U should be shared int), likewise for inout.

Right. Added those cases.

[thewilsonator]

Phobos PR merged.

[TurkeyMan]

Force rebuild?

[Biotronic]

Suggest adding static assert(is(inout const S U == inout U) && is(U == const shared S));

[aG0aep6G]

Suggest adding static assert(is(inout const S U == inout U) && is(U == const shared S));

I'd rather leave that out for now, because it's less obvious how to handle it.

For the others, I just changed MATCH.constant to MATCH.exact. For this combination, deduceTypeHelper currently returns MATCH.nomatch. But it works in IFTI, so there must be something else going on. My guess is that it's handled by deduceTypeHelper, which isn't even called for is expressions.

[aG0aep6G]

My guess is that it's handled by deduceTypeHelper, which isn't even called for is expressions.

That was supposed to be: My guess is that it's handled by deduceWildHelper

[Biotronic]

I was gonna say we could do something like this:

    *at = t.mutableOf().unSharedOf();
    if ((t.mod & MODFlags.const_) && !(tParam.mod & MODFlags.const_))
        *at = (*at).constOf();
    if ((t.mod & MODFlags.shared_) && !(tParam.mod & MODFlags.shared_))
        *at = (*at).sharedOf();
    if ((t.mod & MODFlags.immutable_) && !(tParam.mod & MODFlags.immutable_))
        *at = (*at).immutableOf();
    if ((t.mod & MODFlags.wild) && !(tParam.mod & MODFlags.wild))
        *at = (*at).wildOf();

However, apparently calling sharedOf() on a const type doesn't return a const shared type, and likewise for the other *Of()s. This somewhat disappoints me.

[aG0aep6G]

I'm not sure if it's really necessary to mention this in the changelog, but @thewilsonator has requested an entry, so I've added one.

[TurkeyMan]

Is this good?

[thewilsonator]

so I've added one.

Thanks. I wanted it because this is a non-obvious change of behaviour that would be better documented than not. It probably won't break anyones code, but, OTOH it did require a change in phobos.