On Mon, Oct 20, 2025 at 09:33:00PM +0200, Rafael Sadowski wrote:
> On Mon Oct 20, 2025 at 07:57:52PM +0100, Stuart Henderson wrote:
> > not sure how best to fix or knock this out, but this is a 'nojunk' port
> > which causes problems for bulk builds if it's broken (if build of these
> > fails, dpb no longer removes surplus packages on the relevant node, so
> > it builds up too many installed packages and likely runs out of disk
> > space), so it would be good to either fix or mark as BROKEN on the
> > relevant arch(s)
> > 
> > 
> > -- Configuration summary has been written to /usr/obj/ports/qt6-qtbase-6.8.3/build-sparc64/config.summary
> > 
> > ERROR: Feature "no_direct_extern_access": Forcing to "ON" breaks its condition:
> >     NOT WIN32 AND TEST_no_direct_extern_access
> > Condition values dump:
> >     WIN32 = ""
> >     TEST_no_direct_extern_access = "FALSE"
> > 
> > 
> > CMake Error at cmake/QtBuildInformation.cmake:220 (message):
> >   Check the configuration messages for an error that has occurred.
> > Call Stack (most recent call first):
> >   cmake/QtBuildInformation.cmake:39 (qt_configure_print_summary)
> >   cmake/QtBuildRepoHelpers.cmake:392 (qt_print_feature_summary)
> >   cmake/QtBaseHelpers.cmake:226 (qt_build_repo_end)
> >   CMakeLists.txt:32 (qt_internal_qtbase_build_repo)
> > 
> > 
> This error indicates that on sparc64, the no_direct_extern_access
> feature is forcibly set to "ON", but the necessary condition
> TEST_no_direct_extern_access fails.
> 
> Simple solution: Remove FEATURE_no_direct_extern_access and leave
> the decision to QT. If TEST_no_direct_extern_access is successful,
> FEATURE_no_direct_extern_access is set.

make configure passes (on a machine using gcc/11 as default
ports-gcc).  Not sure you need a REVISION bump since that feature
should work on base-clang and killed the build on ports-gcc archs.

ok jca@ either way.

-- 
jca