Fix 14611 - socket.localAddress fails on Unix sockets with longer path (> 13 characters)

[dcarp]

See https://issues.dlang.org/show_bug.cgi?id=14611

[CyberShadow]

The previous implementation supported paths of any size. What is the reason for restricting the path length to the one defined in system headers? I don't think the length is important as far as the ABI is concerned, and your change may result in regressions.

On reviewing the documentation, it looks like this change is correct.

[CyberShadow]

Don't forget to add the test case from the bug report as a new unit test.

[dcarp]

assert(listener.localAddress.toString == name);

... in the existing unittest will trigger the bug, thus no additional test case.

[CyberShadow]

OK, thanks. LGTM.

[schveiguy]

This doesn't make sense. sun_path is a byte[], uninitializedFill is only needed for structs with copy ctors, and for cases where the data is uninitialized (in this case, it will be initialized).

I don't think we need a specialized function here, just:

sun.sun_path[] = '?';

And remove the import for uninitializedFill

[dcarp]

You are right! I didn't know that you can fill an array like this.

[schveiguy]

I updated the title. Please in the future include the bug title in the PR title so it's easy to know what the PR is about. Thanks.

LGTM except for the uninitializedFill thing

[schveiguy]

Auto-merge toggled on