As well as the "it shall work, even if it must be magic" fix for the
sockaddr aliasing problems, has any consideration been given to a
named function for accessing the sa_family_t fields?
inline sa_family_t getsockfam(const struct sockaddr* sa) {
sa_family_t fam;
memcpy(&fam, (char*)sa + offse
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1650
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1651
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1652
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1653
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1654
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1655
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1656
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1657
==
Reported By:rillig
Assigned To:
==
The following issue has been SUBMITTED.
==
https://austingroupbugs.net/view.php?id=1658
==
Reported By:rillig
Assigned To:
==
< said:
> I don't know if guaranteeing that would break existing implementations
> though. It would be fine for Linux and Solaris, because the
> sa_family_t field is the first one, so as long as every sockaddr
> structure uses the same type for that first field, it's in the common
> initial sequen
11 matches
Mail list logo