On Fri, Nov 03 2017, Nigel Taylor <njtaylor0...@btinternet.com> wrote: > As you say it's not a FLAVOR what's this in the Makefile for
What I meant is that PSEUDO_FLAVORS are meant to disable subpackages, they are not proper flavors. > -CONFIGURE_STYLE= gnu > +MULTI_PACKAGES = -main -qt -py2 -py3 > +PSEUDO_FLAVORS = no_qt no_py2 no_py3 > +FLAVOR ?= > > As subpackage is even worse so even the simplest ports using gpgme needs > Qt5 built to build. And I agree with you that requiring Qt5 to build gpgme in bulk builds is not reasonable, as said in my previous mail. Availability of a C++11 compiler, portability problems on low-powered/"just different" architectures, build times (gpgme is on the path of important ports). Maybe even circular deps? (I did not check.). All those are hard problems. I also agree with you that forcing Qt5 on people who just need gpgme is not good. FLAVOR ?= no_qt bootstrap could be a way to avoid Qt5 except when explicitely needed. But qt5.port.mk changes the C (and ofc C++) compiler used for the whole build, if base doesn't provide clang; and this is not acceptable for a PSEUDO_FLAVOR. So we need something smarter. Thoughts? -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE