Remove lang/gcc45 in favor of lang/gcc46 and later.  Nothing in the
tree uses this version of GCC (GCC 4.5.x) any longer and GCC 4.6 has
been the default for USE_FORTRAN=yes and USE_GCC=4.X+ for a while
and has proven itself.  Plus, only GCC 4.6 and GCC 4.7 have been
patched to work with FreeBSD 10.x and later.

Feature safe:   yes

Update to the 20111117 snapshot of GCC 4.5.4.  This is for reference,
this port is going away in favor of GCC 4.6 soon.

Feature safe:   yes

Update to the 20111103 snapshot of GCC 4.5.4.

Update to the 20111013 snapshot of GCC 4.5.4.

Update to the 20110929 snapshot of GCC 4.5.4.

Update to the 20110922 snapshot of GCC 4.5.4.

Update to the 20110908 snapshot of GCC 4.5.4.

Update to the 20110901 snapshot of GCC 4.5.4.

Update to the 20110825 snapshot of GCC 4.5.4.

Update to the 20110811 snapshot of GCC 4.5.4.

Update to the 20110721 snapshot of GCC 4.5.4.

Update to the 20110630 snapshot of GCC 4.5.4.

Update to the 20110609 snapshot of GCC 4.5.4.

Enable c, c++, and fortran (and only these) explicitly by default, and
Java when/where available. [1]

Reported by:    Scott Allendorf <scott-allendorf@uiowa.edu> [1]

Change BUILD_DEPENDS from ${LOCALBASE}/bin/zip to plain zip since,
really, any zip does.  This bring gcc45 in line with later lang/gcc*
ports.

Update to the 20110602 snapshot of GCC 4.5.4.  Remove the LTO options and
force LTO to off.  Anyone interested in that should be using GCC 4.6 now,
where support is a lot more mature, stable and complete.

Update to the 20110526 snapshot of GCC 4.5.4.  Note that upstream now only
provides a single tarball, so this brings a number of changes around that.

Update to the 20110519 snapshot of GCC 4.5.4.

Update to the 20110505 snapshot of GCC 4.5.4.

Update to the 20110428 snapshot of GCC 4.5.4 which basically coincides
with the upstream GCC 4.5.3 release.

Update to the 20110421 snapshot of GCC 4.5.3.

Replace the, now dysfunctional, post-patch target with a configure
option that marks this build of GCC as "FreeBSD Ports Collection". [1]

Reported by:    Bjoern Koenig <bkoenig@alpha-tierchen.de> [1]

Update to the 20110414 snapshot of GCC 4.5.3.

Update to the 20110407 snapshot of  GCC 4.5.3.

Update to the 20110331 snapshot of GCC 4.5.3.

Update to the 20110324 snapshot of GCC 4.5.3.

Update to the 20110317 snapshot of GCC 4.5.3.

Make the full-regression-test target conditional on MAINTAINER_MODE. [1]

Reported by:    andreast, marius [1]

Update to the 20110310 snapshot of GCC 4.5.3.

Update to the 20110303 snapshot of GCC 4.5.3.  The change for code
generation on i386 to default to i486 on FreeBSD 6 and above is now
upstream, so we can shed our specific patch.

Update to the 20110224 snapshot of GCC 4.5.3.

Update to the 20110217 snapshot of GCC 4.5.3.

By means of an extra patch code generation on i386 now defaults to
i486 on FreeBSD 6 and above. [1]

Submitted by:   tijl [1]
Reported by:    Yuri Karaban <tech@askold.net> [1]
PR:             154364 [1]

Update to the 20110210 snapshot of GCC 4.5.3.

Update to the 20110203 snapshot of GCC 4.5.3.

Feature safe:   yes

Update to the 20110127 snapshot of GCC 4.5.3.

Feature safe:   yes

Update to the 20110120 snapshot of GCC 4.5.3.

Feature safe:   yes

Update to the 20110113 snapshot of 4.5.3.

Feature safe:   yes

Update to the 20110106 snapshot of GCC 4.5.3.

Update to the 20101230 snapshot of GCC 4.5.3.

Update to the 20101223 snapshot of GCC 4.5.3.  This is nearly the same
as the 4.5.2 release.

Update to the 20101125 snapshot of GCC 4.5.2.

Update to the 20101118 snapshot of GCC 4.5.2.

Update to the 20101111 snapshot of GCC 4.5.2.

Update to the 20101104 snapshot of GCC 4.5.2.

Update to the 20101028 snapshot of GCC 4.5.2.

Update to the 20101021 snapshot of GCC 4.5.2.

Update to the 20101014 snapshot of GCC 4.5.2.

Update to the 20101007 snapshot of GCC 4.5.2.

Replace the explicit build dependency on devel/binutils by USE_BINUTILS.

Submitted by:   bf

Update to the 20100923 snapshot of GCC 4.5.2.

Update to the 20100916 snapshot of GCC 4.5.2 which already includes one of
my two patches for ia64.

Backport the change to make Java support an OPTION (JAVA) so it is easier
to deselect from lang/gcc46.

Update to the 20100909 snapshot of GCC 4.5.2.  With two new local patches
(one a backport of a patch of mine from GCC 4.6 development, the other a
first hack) FreeBSD/ia64 is now back working and ia64 can become part of
ONLY_FOR_ARCHS.

Update to the 20100902 snapshot of GCC 4.5.2.  This brings first _float128
support for FreeBSD.

Update to the 20100819 snapshot of GCC 4.5.2.

Replace NOT_FOR_ARCHS by ONLY_FOR_ARCHS. [1]

Suggested by:   linimon

Update to the 20100812 snapshot of GCC 4.5.2.

Update to the 20100805 snapshot of GCC 4.5.2.

Update to the 20100729 snapshot of GCC 4.5.1.

Update to the 20100722 snapshot of GCC 4.5.1, which basically coincides
with the first release candidate of this release.

Update to the 20100715 snapshot of GCC 4.5.1.

Feature safe:   yes

Update to the 20100708 snapshot of GCC 4.5.1.

Feature safe:   yes

Update to the 20100701 snapshot of GCC 4.5.1.

Feature safe:   yes

Simplify full-regression-test.

Feature safe:   yes

Update to the 20100624 snapshot of GCC 4.5.1.

Feature safe:   yes

Update to the 20100617 snapshot of GCC 4.5.1.

Feature safe:   yes

Update to the 20100610 snapshot of GCC 4.5.1.

Chase mpfr library and bump PORTREVISION.

Update to the 20100603 snapshot of GCC 4.5.1.

Update to the 20100527 snapshot of GCC 4.5.1.

Update to the 20100520 snapshot of GCC 4.5.1.

Update to the 20100513 snapshot of GCC 4.5.1.

Update to the 20100506 snapshot of GCC 4.5.1.

- Ok, this should work. I toasted the BUILD_DEPENDS with a "=" instead
  of a "+=".

Reported by:    QAT

- Re-added BUILD_DEPENDS on zip that I didn't think needed to be there,
  but it actually does.

Reported by:    QAT
Pointyhat to:   glarkin

- Added support for gcj (Java compiler) to prepare the print/pdftk port
  to migrate from gcj42 to gcj45.
- Bumped PORTREVISION (packing list & port functionality changes)

Approved by:    gerald (via email)

Add a new option LTO that enables the new link-time optimization framework
and first optimizations that GCC 4.5 brings.  This adds libelf as a new
dependency (if used) and is off by default; we may be enabling it at a
later point, though.

Bumping PORTREVISION since someone who has installed libelf, previously
to this change would get LTO enabled automatically and then run into
problems if libelf is removed -- which is possible, since this port
does not lay down a dependency on that case.

PR:             146055
Submitted by:   mm

Update to the 20100429 snapshot of GCC 4.5.1.

Update to the 20100422 snapshot of GCC 4.5.1.

Switch to use newer GMP version.

PR:             ports/144487
Submitted by:   ale
Approved by:    portmgr (-exp run by erwin)

Update to the 20100415 snapshot of GCC 4.5.1.  This is basically the
official release of GCC 4.5.0, thought that one may still have some
kinks and I recommend staying with GCC 4.4 for production use.

Update to the 20100401 snapshot of GCC 4.5.0.

Update to the 20100318 snapshot of GCC 4.5.0.

Update to the 20100311 snapshot of GCC 4.5.0.

Update to the 20100304 snapshot of GCC 4.5.0.

Update to the 20100225 snapshot of GCC 4.5.0.

Feature safe:   yes

Update to the 20100218 snapshot of GCC 4.5.0.

Feature safe:   yes

Update to the 20100211 snapshot of GCC 4.5.0.

Update to the 20100204 snapshot of GCC 4.5.0.

Update to the 20100128 snapshot of GCC 4.5.0.

Also build Java on amd64.

Submitted by:   Andreas Tobler <andreast@fgznet.ch>

Update to the 20100114 snapshot of GCC 4.5.0.

Since we already require devel/binutils for the assembler, also enforce
use of the linker from there.

Update to the 20100107 snapshot of GCC 4.5.0.

Update to the 20091231 snapshot of GCC 4.5.0 which brings some key
improvements to libffi (Java), especially on amd64.

Remove the WITH_JAVA_AWT knob which really has not been used a lot,
if at all.

Update to the 20091224 snapshot of GCC 4.5.0.  On the way, sort pkg-plist
a bit differently.

Update to the 20091217 snapshot of GCC 4.5.0.

Use current GNU binutils from ports (devel/binutils) instead of the
system tools.  The latter are half a decade old and fail to properly
support modern CPUs and their standard features such as SSE or AVX in
the case of AMD64/Intel64 which means that they even break when using
a simple -march=native.

On the way, sort CONFIGURE_ARGS.

PR:             140995, 136607

Update to the 20091126 snapshot of GCC 4.5.0.  Among others this fails
to build emulators/wine due to problems with debuginfo, but then this
is more of an experimental port than anything, so always use with care.

With lang/gcc295 gone (and not building for most of the last year), we can
now remove the check for a gcc295 binary that might cause build failure in
libgcj.

Update to the 20091119 snapshot of GCC 4.5.0.

No longer download Objective-C related files which we don't actually use.
Minor style changes.

Adjust the ports description to the new style I am using for the various
gcc ports and remove an obsolete note on memory consumption when building
libgcj (the Java library).

Update to the 20091112 snapshot of GCC 4.5.0.

Update to the 20091105 snapshot of GCC 4.5.0.