On amd64, when "make install" is run by non-root user, installation of x11-toolkits/linux-gtk finishes with success but nothing is really installed. % make install ...(snip)... ===> Installing for linux-gtk-1.2_4 ===> linux-gtk-1.2_4 depends on file: /compat/linux/usr/X11R6/lib/libX11.so.6 - found ===> Generating temporary packing list ===> Checking if x11-toolkits/linux-gtk already installed ===> Switching to root credentials for 'install' target gtk+-1.2.9-4.i386.rpm package gtk+-1.2.9-4 is for a different architecture ===> Registering installation for linux-gtk-1.2_4 ===> Returning to user credentials % pkg_info -qL linux-gtk-1.2_4 | xargs ls -l ls: /compat/linux/etc/gtk/gtkrc: No such file or directory ls: /compat/linux/etc/gtk/gtkrc.az: No such file or directory ...(snip)... When installed by root, everything is successful. This is because a separate make process is run (with ${SU_CMD}) when "install" target is executed by non-root user, and ${ARCH}, which is overwritten to "i386" by the 1st make process, is passed to the 2nd make, and the following part of the Makefile is not interpreted. .if (${ARCH} == "amd64") ARCH= i386 RPMFLAGS+= --ignorearch .endif This is the same problem as ports/80679 (so portupgrade also fails), and the same fix should be applied.
Responsible Changed From-To: freebsd-ports-bugs->freebsd-emulation Over to maintainer
Responsible Changed From-To: freebsd-emulation->emulation Use canonical address of the emulation mailing list.
State Changed From-To: open->feedback Can you please check if this is still an issue? I'm unable to repeat this, albeit with a hand-crafted test-setup. Also, using MACHINE_ARCH is usually the wrong solution, please read bsd.port.mk for the rationale.
State Changed From-To: feedback->closed Feedback timeout (1 month)