The AUX compiler supports several languages: Ada, C, C++, Fortran and
Objective-C. Since Ada support must be built by an Ada-capable compiler,
only platforms for which a bootstrap compiler is available can build it.
The AUX compiler is based on release versions of the Free Software
Foundation's GNU Compiler Collection. It uses the GCC Runtime Library
Exception, so the resulting binaries have no licensing requirements.
Binaries produced by the AUX compiler should be legally handled the same
as binaries produced by any FSF compiler.
This compiler implements the full Ada-83, Ada-95, Ada-2005 and Ada-2012
standards.
WWW: http://www.dragonlace.net/
===> The following configuration options are available for gcc5-aux-20160603:
====> Options available for the group Standard
FORTRAN=on: Fortran language support
OBJC=on: Objective-C language support
NLS=on: Native Language Support
TESTSUITE=off: Activate test support
ALLSTAGES=off: Build compiler with itself (twice, brings in libcc1)
STATIC=off: Link compilers and drivers statically
====> Options available for the group Bootstrap
BOOTSTRAP=off: Create bootstrap compiler (overrides other options)
===> Use 'make config' to modify these settings
Deprecate this port which is essentially unmaintained and based on
a 15 month old snapshot of GCC 5. Users can still go for gcc6-aux
(which is in a similar situation but used by a couple of other ports
and at least only 7 months old).
Return the ports mistakenly reset to ports@ in r433856 to John Marino.
The mistake was completely on my part, I somehow connected the dots the
wrong way in my head.
The only exceptions (for now) are archivers/zstd and ports-mgmt/synth
which were already picked up by new volunteers in the mean time.
Backport from lang/gcc6-aux:
Rename the FORT option to FORTRAN to be aligned with other ports.
(FORTRAN_DESC is predefined in Mk/bsd.options.desc.mk.)
Approved by: marino (maintainer)
lang/gcc5-aux: Upgrade version 20151204 => 20160603 (gcc 5.3 => 5.4)
Remove the recently added ARM patch. This port has always been limited to
amd64 and i386 so patching ARM code has no effect.
lang/gcc5-aux: Upgrade version 20150609 => 20150716 (release 5.2.0)
This contains some minor fixes for DragonFly's master branch. The
patch diffs are less due to freebsd unwind support backport to 5 branch.
lang/gcc5-aux: Fix build of next version of gnatcoll on DF
The issue only affects DragonFly -- the proper configuration for openpty
support is already defined for FreeBSD using a macro, but the macro
for DragonFly was missing.
lang/gcc5-aux: Upgrade from Release 5.1.0 to 5.1.1 snapshot (20150609)
This captures some gcc fixes in the 2 months since the release. The
version info with --version option has also been restored.
lang/gcc5-aux: Minor bug fixes
When submitting patches to fix GNAT, I realized the exetime changes
were wrong. Moveover, the time patches not only are no longer needed,
but possibly bitrotted to the point of being wrong.
The next release of GCC5 (5.2) should have the FreeBSD unwind support
from gcc-trunk backported to GCC5 branch, so that should reduce the
patches by a lot. There are some minor changes involving NetBSD support
that are not used.
Also, apparently disable-libada functionality was broken by my patches,
but it's fixed now. This isn't used though.
MASTER_SITES cleanup.
- Replace ${MASTER_SITE_FOO} with FOO.
- Merge MASTER_SITE_SUBDIR into MASTER_SITES when possible. (This means 99.9%
of the time.)
- Remove occurrences of MASTER_SITE_LOCAL when no subdirectory was present and
no hint of what it should be was present.
- Fix some logic.
- And generally, make things more simple and easy to understand.
While there, add magic values to the FESTIVAL, GENTOO, GIMP, GNUPG, QT and
SAMBA macros.
Also, replace some EXTRACT_SUFX occurences with USES=tar:*.
Checked by: make fetch-urlall-list
With hat: portmgr
Sponsored by: Absolight
lang/gcc5-aux: set build/host_configargs again
Discussing with tijl resulted in decision ot restore *_configargs
in order to fix libiconv linkage. This does build on DragonFly.
lang/gcc5-aux: Build in one stage by default, new bootstrap for DF
The original intention was to build update all the bootstrap compilers
to be based on GCC5, but right now I'm only doing that for DragonFly.
After June 30, I might rebuild the bootstrap compiler on FreeBSD 10.1,
but right now gcc5-aux is building fine on FreeBSD in one stage, even
though their bootstrap is based on GCC 4.9.
Right now, gcc5-aux is building with 3 full stages. I'm switching it to
a single stage build, but adding an off-by-default option to build it
in three stages again. With one stage, libcc1 is not built, but with
three stages it is. LibCC1 is meant to support the upcoming release of
GDB so it's currently not needed (it fails to build if gcc is not fully
bootstrapping which is why it's disabled)
This brings the build down to about 9-14 minutes versus 60-90 minutes
depending on the (average power) machine.
Also, I've explicitly limited the build to i386 and amd64 on FreeBSD and
I have dropped i386 support on DragonFly, which could only have been
build on DragonFly 3.8 or earlier.
Also, I removed the iconv dependency when a bootstrap compiler is being
created (the all-overriding BOOTSTRAP option for the maintainer)
lang/gcc5-aux: r384038 (iconv commit) broken DragonFly, Fix
Swapping LIBS with LDFLAGS and removing the *_configargs definitions
allows gcc5-aux to build on all supported platforms. Now that gcc5 has
been released, new bootstrap compilers should be created for both FreeBSD
and DragonFly in order remove bootstrapping requirement (previously
gcc5 built live but not in poudriere, reason still unknown).
converters/libiconv:
- Remove const qualifier from iconv(3) to match POSIX:
http://pubs.opengroup.org/onlinepubs/9699919799/functions/iconv.html
- Patch iconv.h to expose more GNU extensions when LIBICONV_PLUG is
defined because the base system iconv supports these extensions too.
Add/remove patches to/from ports to call iconv with non-const arguments.
This breaks some ports on FreeBSD 10 because base system iconv.h still has
the const qualifier. Fix this by letting USES=iconv add a build dependency
on converters/libiconv so ports can use its iconv.h (with LIBICONV_PLUG
defined) instead of the base system iconv.h.
This exposed some ports that link with libiconv when it is available instead
of using libc iconv. In these cases one of the following changes has been
(Only the first 15 lines of the commit message are shown above )
lang/gcc5-aux: Disable diagnostic checks
These diagnostic checks cause lang/gcc-aux to take nearly two hours
building as with 3 stages. They are pretty heavy; I'm going to disable
them (with the default checking for releases). FWIW, lang/gcc5 still
has these checks.
Technically this requires a revbump, but given the compiler is experiment
and that it takes two hours to build, I'll make the decision not to bump.
lang/gcc5-aux: Fix ada hardlink
The code to create a hardlink of gcc5-aux/bin/ada to gcc5-aux/bin/gcc
did not work. It appears that the "-" in front of the "if" condition
isn't working in this context. However, it's really not needed at all,
so just remove it so the hardlink gets created.
All Ada ports use gcc5-aux/bin/ada, so it was a critical problem.
lang/gcc5-aux: Make DragonFly build 3 stages too
This port also fails in poudriere (but not outside it) in DragonFly,
so just make all platforms build all 3 stages until a new bootstrap
compiler is available on both FreeBSD and DragonFly.
Add new port lang/gcc5-aux (next Ada compiler)
This is the initial version of gcc5-aux, which will eventually become the
default Ada compiler. It's not hooked into Mk/Uses/ada.mk yet, but it
does pass all Ada tests on both DragonFly64 and FreeBSD64.
For FreeBSD 10 amd64, the compiler built fine outside of poudriere but the
bootstrap compiler failed inside of it (seemingly as a result of using
base linker). Eventually a new bootstrap compiler needs to be made, but
for now gcc5-aux is built with a full bootstrap on FreeBSD. On DragonFly,
only a single stage is built as the bootstrap compiler still works fine.
