Deprecate and mark BROKEN the rtems ports.
There are many issues with the ports and these will require a dedicated
maintainer to fix them.
Reviewed by: swills, draco@marino.st
- Return my ports back to the pool. I was unable to make any fixes to
my ports in the past 3 weeks while ports were broken on any 10.x
machines, which means I'm unable to maintain them. So let people know
that there's no available support for them until things are back to
normal (which also means that anyone with spare time will be able
to fix them without getting approval).
- Update the rtems ports collection. Now these ports will be slaves of
more generic cross-* ports, that will simplify updating and bug fixes.
* Update binutils to 2.17
* Update gcc to 4.2.0
* Update gdb to 6.6
- Deprecate ports for i960 architecture abandoned by mainstream
- Deprecate *-g77 and *-objc ports, support for these languages can
be enabled via OPTIONS in relative *-gcc ports.
- Take maintainership.
Approved by: ade (maintainer)
Since there are a number of issues associated with the *-rtems-* ports
as a result of the new gettext, and given the jeh@ is currently on
sabbatical, with these ports (amongst others) being reset to ports@,
temporarily take over maintainership to cut through a little red tape
to get things back in order.
Various slave ports are broken with the new gcc 4.2 in HEAD.
For simplicities sake, mark the master port as broken until such
time as it can be fixed.
Submitted by: krismail
Approved by: maintainer on sabattical
- Fix build by explicitly tell configure how to correctly link gettext
- Add extra empty line between PATCH_SITES and DISTFILES to make
portlint(1) happy
- Use += for *_DEPENDS after pre.mk to per portlint(1)
Submitted By: Rong-En Fan
PR: 111404
Various kinds of BROKEN:
RTEMS_ARCH=sh broken everywhere
RTEMS_ARCH=m68k broken everywhere
RTEMS_ARCH=powerpc and arm broken on sparc64
Approved by: portmgr (implicit)
Revert back to previous method to remove the CPU related entries from
CFLAGS. At least on 4.9, the new changes to bsd.port.mk don't like
if _CPUCFLAGS is empty.
Since once mk/bsd.cpu.mk adds the -mcpu and/or -march flags they can't
be removed. And setting NO_CPU_CFLAGS in a ports Makefile does not help.
It might be nice if Mk/bsd.port.mk would remove it if NO_CPU_CFLAGS was
set, patch coming soon.
* UPdate the *-rtems-gcc ports to include gcj (Java). Since the gcj
depends so heavily on c++ it was getting silly to keep it seperate for
all the mangeling of the Makefiles that was required
* Remove the *-rtems-gcj ports (MOVED) to be commited next.
Add MACHINE_ARCH to MAKE_ENV to try and stop stupid breakage on 5.x
bento builds caused by it forcing it's maching arch onto the ports.
No PORTREVISION bump required, they don't build now.
Update to the latest version from RTEMS, gcc-3.2.3 with newlib-1.11.0
This does not include PR 58548 because I could not get it to work no
matter what I tried.
Update to use the new grouping capability. This allows the removal
of the rtems-gcc hack to fetch the files from a different subdir
then newlib lived in. Remove devel/rtems-gcc
No functionality change so no PORTREVSION bump.
Minor update in where the files are fetched from. Maybe now I will stop
getting "Unfetchable distfiles" message for files that do get fetched in the
real-world (tm). Suggested by sf, and changed.
