onsdag den 2 maj 2012 klockan 14:56 skrev Sergey Poznyakoff detta: > Mats Erik Andersson <g...@gisladisker.se> ha escrit: > > > How can you tell that <netinet/ip.h> be standardised, > > It is not standardised, either. FWIW, here is a list of *standard* > header files: http://pubs.opengroup.org/onlinepubs/9699919799/idx/head.html
Very good reading. Thank you! I notice that <netinet/tcp.h> is mentioned, but not <netinet/udp.h>, which seems awkward. > > Well, OpenSolaris, NetBSD, OpenBSD, FreeBSD, and DragonFly BSD are > > These are all very good, but how about Linux? Hurd? There is no operating system named Linux, nor Hurd. They are tied to Glibc, hence falls under my comment on duplicate macro sets, and they are thus rightly to be spoken of as GNU/Linux and GNU/Hurd as soon as the operating system is in focus. What deviations do SysV derived systems display concerning <netinet/icmp.h>? Systems we care about? (Remember that it is only since my appearance that the contemporary *BSD and OpenSolaris are getting compilable support inside GNU Inetutils. Their stability is also better than that of GNU/Linux.) Admittedly, I now appreciate Alfred's point of view, but I would like to understand better what impacts are circumvented by nourishing a private header like <libicmp/icmp.h>. Best regards, Mats E A
