On 2016/03/16 15:26, Vadim Zhukov wrote:
> 2016-03-16 13:34 GMT+03:00 Stuart Henderson <[email protected]>:
> > As found by Peter Kay on misc, qt5 detects CPU features to enable
> > at build time, so produces binaries that don't work everywhere.
> > Does this seem a reasonable set to disable?
>
> I think so. But you'd better bump all subpackages except -qch and -html.
Ah I can do that. I thought I might as well just bump them all as
it was easier (hence REVISION rather than something like REVISION-main)
but I can exclude qch/html from that.
> > Index: Makefile
> > ===================================================================
> > RCS file: /cvs/ports/x11/qt5/Makefile,v
> > retrieving revision 1.48
> > diff -u -p -r1.48 Makefile
> > --- Makefile 6 Mar 2016 11:56:41 -0000 1.48
> > +++ Makefile 16 Mar 2016 09:33:26 -0000
> > @@ -4,6 +4,7 @@ SHARED_ONLY = Yes
> > ONLY_FOR_ARCHS = ${GCC4_ARCHS}
> >
> > VERSION = 5.5.1
> > +REVISION = 0
> > ENGINIO_VERSION = 1.2.1
> > DISTNAME = qt-everywhere-opensource-src-${VERSION}
> >
> > @@ -299,6 +300,15 @@ CONFIGURE_ENV = MAKE=make \
> > .if ${MACHINE_ARCH} == "powerpc"
> > CONFIGURE_ENV += LDFLAGS="-Wl,--relax"
> > .endif
> > +
> > +.if ${MACHINE_ARCH} == "amd64" || ${MACHINE_ARCH} == "i386"
> > +CONFIGURE_ARGS += -no-ssse3 -no-sse4.1 -no-sse4.2 -no-avx -no-avx2
> > +.endif
> > +
> > +.if ${MACHINE_ARCH} == "i386"
> > +CONFIGURE_ARGS += -no-sse2 -no-sse3
> > +.endif
> > +
> > .include <bsd.port.arch.mk>
> >
> > # QtWebkit (at least, maybe others) plays dirty games with inter-target
>
> --
> WBR,
> Vadim Zhukov
