Klemens Nanni <[email protected]> wrote: > On Sat, Dec 18, 2021 at 04:37:17PM +0300, Andrew Krasavin wrote: > > On Fri, Dec 17, 2021 at 01:04:35PM +0000, Klemens Nanni wrote: > > > An upcoming port uses ElfW(). I first patched the port, but looking > > > around shows that both NetBSD and FreeBSD adopted the macro: > > > > > > https://mail-index.netbsd.org/source-changes-hg/2018/07/31/msg027523.html > > > +#define ElfW(x) CONCAT(Elf,CONCAT(ELFSIZE,CONCAT(_,x))) > > > > > > https://cgit.freebsd.org/src/commit/sys/sys/elf_generic.h?id=34e7e4b6a059eee5e4e3e34de5b9d5f0d6e589f9 > > > +/* Define ElfW for compatibility with Linux, prefer __ElfN() in FreeBSD > > > code */ > > > +#define ElfW(x) __ElfN(x) > > > > > > So I suggest providing it as well. > > > > > > I have yet to put this through a release build on amd64 and/or sparc64 > > > to check if there is any fallout, but I don't suspect there will be. > > > > > > Feedback? Objections? OK (given the build succeeds)? > > > > > > > > > Index: sys/sys/exec_elf.h > > > =================================================================== > > > RCS file: /cvs/src/sys/sys/exec_elf.h,v > > > retrieving revision 1.93 > > > diff -u -p -r1.93 exec_elf.h > > > --- sys/sys/exec_elf.h 7 Dec 2021 22:17:03 -0000 1.93 > > > +++ sys/sys/exec_elf.h 17 Dec 2021 12:36:39 -0000 > > > @@ -725,6 +725,7 @@ struct elf_args { > > > #define CONCAT(x,y) __CONCAT(x,y) > > > #define ELFNAME(x) CONCAT(elf,CONCAT(ELFSIZE,CONCAT(_,x))) > > > #define ELFDEFNNAME(x) CONCAT(ELF,CONCAT(ELFSIZE,CONCAT(_,x))) > > > +#define ElfW(x) CONCAT(Elf,CONCAT(ELFSIZE,CONCAT(_,x))) > > > #endif > > > > > > #if defined(ELFSIZE) && (ELFSIZE == 32) > > > > > > > The code that requires this macro in sys/sys/exec_elf.h was > > eventually removed from the upcoming port (devel/abseil-cpp). > > > > I have no opinion on what you should do with this patch now, but > > I feel it is necessary to report that this change is no longer > > needed for my port. > > Even better! > > Unless there are objections, I'll make sure that this doesn't break > anything and then commit it, so that future ports that do require it > will build -- there is no downside to providing it, imho.
Why would you commit an unneccessary line to a protected namespace which doesn't need it?
