The latest amd64 bulk build tested a number of things and suffered quite a bit of fallout: (1) update to CMake-3.6.1 (2) mmap returning ENOTSUP for W^X mappings (3) printf %s NULL no longer being valid
databases/hs-postgresql-simple ghc mmap W^X violation devel/darcs ghc mmap W^X violation devel/hs-fgl ghc mmap W^X violation mail/mozilla-thunderbird xpcshell mmap W^X violation devel/jdk/1.7 java mprotect W^X violation devel/libdwarf ? devel/libtalloc printf %s NULL devel/xulrunner/24 xpcshell mmap W^X violation graphics/ctl cmake? alloca.h: No such file or directory lang/libv8 ? segfault lang/mono mono-boehm mmap W^X violation databases/tdb printf %s NULL lang/node ? segfault lang/pypy pypy mmap W^X violation lang/rakudo ? segfault lang/sbcl sbcl mmap W^X violation net/gupnp/av ? segfault net/libaccounts-glib xsltproc.core www/chromium display_source.h file not found www/seamonkey xpcshell mmap W^X violation x11/freerdp ? segfault Notes: * libaccounts-glib has sporadically shown this failure for some time. * mono doesn't build on a loaded machine anyway, the W^X violation is probably just incidental. * The dmesgs of the build machines show some further W^X violations that apparently didn't cause any build errors. Recommendably, pypy doesn't just die randomly but catches the error and provides an explicit error message: Got an unexpected error trying to allocate some memory for the JIT (tried to do mmap() with PROT_EXEC|PROT_READ|PROT_WRITE). This can be caused by a system policy like PAX. You need to find how to work around the policy on your system. Abort trap (core dumped) -- Christian "naddy" Weisgerber [email protected]
