MinGW: Fix error: cannot convert 'size_t*' to 'int*'

[yadij]

Removing another addrinfo as we do.

[rousskov]

Thanks a lot for this cleanup!

P.S. We should avoid (poor) code duplication that this PR and recent 1800e25 are creating, but I think we should continue to duplicate that code for a little bit longer before there are enough use cases to suggest a proper deduplication solution.

[yadij]

Thanks a lot for this cleanup!

P.S. We should avoid (poor) code duplication that this PR and recent 1800e25 are creating, but I think we should continue to duplicate that code for a little bit longer before there are enough use cases to suggest a proper deduplication solution.

I assume you are referring to the pattern of 2 local variables with a reinterpret_cast (if not please correct me). FWIW, I hope long term that we only need to de-duplicate the set/get syscalls and the rest are uncommon enough to be okay without anything special.

[rousskov]

P.S. We should avoid (poor) code duplication that this PR and recent 1800e25 are creating, but I think we should continue to duplicate that code for a little bit longer before there are enough use cases to suggest a proper deduplication solution.

I assume you are referring to the pattern of 2 local variables with a reinterpret_cast (if not please correct me).

The pattern includes two local variable definition/initialization code, a reinterpret_cast, an assertion, and a conversion to an Ip::Address object (at least).

FWIW, I hope long term that we only need to de-duplicate the set/get syscalls and the rest are uncommon enough to be okay without anything special.

Yeah, let's see where these initial (and duplicated) changes lead us. I see no reason to discuss the details of that potential future duplication avoidance effort until we have more merged use cases. We are making good progress so far...