If the /usr/ports/base/gcc didn't create gcc5.4.0, I wouldn't raise this bug. But as I've built gcc5 from lang/gcc5 and build most ports this way, its worth mentioning that gnupg1 fails to build. The logic of the conditional probably needs refining, but I just wanted to jump over this "bug". --- /usr/ports/security/gnupg1/Makefile (revision 424968) +++ /usr/ports/security/gnupg1/Makefile (working copy) @@ -56,7 +56,7 @@ .include <bsd.port.options.mk> -.if ${ARCH} == "i386" && (${CC} == "clang" || ${OSVERSION} >= 1000024) +.if ${ARCH} == "i386" && ${CC} == "clang" CFLAGS+= -fheinous-gnu-extensions .endif FYI: In my case, # make -VCC -Vcc gcc5 /usr/local/libexec/ccache/gcc5
Hrmmn, what about: .if ${ARCH} == "i386" && empty(CC:Mgcc*)
A commit references this bug: Author: adamw Date: Mon Oct 31 16:25:00 UTC 2016 New revision: 425012 URL: https://svnweb.freebsd.org/changeset/ports/425012 Log: Improve the compiler check by using USES=compiler and not applying it blindly to 10+. PR: 213932 Reported by: dewayne heuristicsystems com au Changes: head/security/gnupg1/Makefile
Committed a slight variation on your patch. Please reopen this PR if the problem persists.
If I try to make, it says: /!\ gnupg1-1.4.21_1: Makefile errors /!\ you cannot include bsd.port[.pre].mk twice *** Error code 1 Stop. make[1]: stopped in /usr/ports/security/gnupg1
Thanks, Kurt. Fixed in r425027.