Make Socket.receive @nogc.

[quickfur]

No description provided.

[dlang-bot]

Thanks for your pull request, @quickfur!

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 + phobos#7433"

[CyberShadow]

Hi & thanks for the PR,

As with past attempts to improve the Socket methods' attributes, I think this should be a breaking change, as making receive @nogc in Socket would mean that all classes that subclass Socket would now need to have their receive also @nogc. We have a test for this but for some reason it is not triggered by this change. What am I missing?

[schveiguy]

Why would that test trigger? asserts are @nogc.

[CyberShadow]

We have a test for this but for some reason it is not triggered by this change. What am I missing?

OK, I see the compiler is assuming that the method in the derived class is @nogc , which our stub method in the test is, but that may not be true for the real cases which the test is supposed to represent, which means the test is broken. I will submit a fix.

Why would that test trigger? asserts are @nogc.

Because the method in the derived class was not annotated @nogc. (I forgot that method annotations implicitly propagate to derived classes.)

[schveiguy]

And I agree, this should be considered a breaking change.

[CyberShadow]

RFC: #7434

[n8sh]

There is a previous non-accepted PR identical to this one, Add @nogc attribute to Socket receive methods (#6730).

[pinver]

It's not "not accepted", but pending for a decision: note the "label" and the discussion.

[n8sh]

Perhaps "open" would have been a better term than "non-accepted."

[quickfur]

I don't have the time / patience to work on this any more. Somebody else please pick this up.

[pinver]

@thewilsonator , as stated above, #7434 exactly the same issue, with a little more details: I think this can simply be closed.