2012/6/28 Corinna Vinschen <vinsc...@redhat.com>: > On Jun 28 22:02, Kai Tietz wrote: >> 2012/6/28 Corinna Vinschen <vinsc...@redhat.com>: >> > On Jun 28 21:47, Kai Tietz wrote: >> >> 2012/6/28 Corinna Vinschen <vinsc...@redhat.com>: >> >> > Hi, >> >> > >> >> > the below patch skips the definition of struct in6_addr and the >> >> > definitions concerning struct sockaddr_storage when building Cygwin. >> >> > >> >> > >> >> > Thanks, >> >> > Corinna >> >> >> >> Hmm, type in6_addr is required by ras.h and by ipexport.h, ws2ipdef.h, >> >> and ws2tcpip.h header. >> >> I am not sure if it is wise to simply not define it at this place. >> > >> > It's only undefined whern building Cygwin itself, not for Cygwin >> > applications using the windows headers. The reason is that the >> > definition clashes with our own POSIX definitions. >> > >> > >> > Corinna >> >> Ok, so wouldn't it be a good point to include exactly here in this >> case cygwin/in6.h header? > > The problem is that Cygwin creates the POSIX API using the Winsock API. > That requires to include both headers in the same files in a couple of > cases, and the __INSIDE_CYGWIN__ are used to make sure we don't have > type collisions where Windows uses POSIX datatype names. > > So, are you proposing to include cygwin/in6.h from in6addr.h or > winsock2.h? What would be the advantage? You'd still need to guard > these includes with __INSIDE_CYGWIN__ and you'd still have to disable > the winsock2.h definitions the same way.
Right. I am just a fan of having dependencies resolved in place. Means winsock header defines normally in6_addr, so why not do this here too. By this you avoid implicit dependencies. >> Anyway that's you business ... so patch is ok. your* > Thanks. I'll apply it tomorrow, unless you'd like to discuss this more. > I admit that the usage of __INSIDE_CYGWIN__ in the winsock headers looks > a bit random sometimes, but please keep in mind that this has grown > historically. We only ever diabled what collided and left everything > else intact as far as possible. Well, the usage of __INSIDE_CYGWIN__ might be a bit random, but of more importance for now is, that things remain consistent. The chance to clean it up a bit - if possible - we have later on, too. So please go ahead with that patch. Regards, Kai ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
