CONFIGURE_ARGS. While it may have had some value in the past, nowadays it's
just another (essentially untested) deviation from the mainstream.
Also:
* Restore -fstrict-aliasing and -fstrict-overflow behaviour
* Support -rdynamic on hppa
* libatomic is actually supported on every POSIX system, so remove useless
PFRAGs.
Tested on amd64, hppa and sparc64; ok tobiasu@
Additionally, from tobiasu@:
Disable the optimized shared Ada.Strings.Unbounded implementation until
overlapping memcpy() is fixed. Requires fresh adastrap's.
Rename all binaries starting with "egnat" back to "gnat". Reduces the
amount of workarounds needed to build various Ada software - including
GNAT itself.
Stop running the fixincludes script - saves time and pain. If our
headers are broken, we want to know and fix them ourselves, thank you
very much.
Drop gnat debug patch for a linker issue that was fixed some time ago.
ok tobiasu@
the past 3 bulks, or forever.
Various reasons: no d^Hrecent compiler available, base gcc generating
nice internal compiler errors, no atomic ops, etc, etc.
While here, mark all webkits as BROKEN-hppa, since this also requires
clang or g++ 4.7. Our g++ 4.8 port only builds on
i386|amd64|powerpc|sparc64 anyway...
if you're interested^Wbored to death, full past logs at
http://build-failures.rhaalovely.net/alpha/...
Marked broken because it uses instructions the default binutils don't
understand. The port itself seems to work.
I will probably post the update instructions to arm@ for those that are
interested in playing with this.
ok pascal@ a long time ago
C functions have been modified to use same-size types on both sides.
There are no checks in gcc to make sure the types match, and it happily
trashes the stack or reads bogus data without this fix.
The C shim in gnatlib is still casting down to long (by design), thus
gnat is not 2038-proof on 32 bit architectures. Fixing this is an
ongoing project, but I'm running out of time for this release.
Getting the timezone offset is now supported; the unused usleep
procedure that snuck in from FreeBSD support got removed.
Tested on all supported architectures, plus new bootstraps due to the
stack trashing, just in case...
Conceptually fine with brad@
deraadt@ is unhappy with the partial solution, fixing this remains a
high priority todo
enable it through symlinks in ${WRKDIR}/bin/, which were effectively
overriden by gcc4.port.mk or clang.port.mk - thus lying that ccache was
used while it was not.
Still investigating for a good solution to actually enable ccache support
for non-base compilers.
okay juanfra@, sthen@
