On Sat, Aug 27, 2022 at 12:15 PM Shane <skullrat...@gmail.com> wrote: > > My understanding then is that cmd/dist first builds for the host here (in my > case, for Linux) and then builds for the target starting here. Since there is > always a build for the host first, then since my host OS is Linux, the > linux_syscall.c is always part of the compilation (since I do not disable > cgo). If anything is incorrect with my understanding, please correct me. > > I believe then I need to have both Linux system headers and Solaris system > headers available on my build machine for the cross-compile, if I want to > build Go for Solaris with cgo support. > > Is it possible for there to be two header files with the same name but > different OS have the same #include path in the cgo source code? If so, could > the C compiler get the correct header for the target of the cgo? For example, > how could the C compiler know to use the Linux unistd.h header for the Linux > build, and later use the Solaris unistd.h header for the Solaris build?
I'm not sure whether this answers your question, but normally a system will have a native compiler that will look for header files in /usr/include. It can also have a cross-compiler that looks for header files in some other location. A cross-compiler should never look in /usr/include, it should only look at the cross-compilation header files. Ian -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/CAOyqgcWgft7wLnAC-AOTePFve33zpz0%2Bugm%2Birk42XFhPpceDw%40mail.gmail.com.
