Copy head@r302406 to stable/11 as part of the 11.0-RELEASE cycle.
Prune svn:mergeinfo from the new branch, as nothing has been merged
here.

Additional commits post-branch will follow.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

DIRDEPS_BUILD: Regenerate without local dependencies.

These are no longer needed after the recent 'beforebuild: depend' changes
and hooking DIRDEPS_BUILD into a subset of FAST_DEPEND which supports
skipping 'make depend'.

Sponsored by: EMC / Isilon Storage Division

META MODE: Don't create .meta files when symlinking sources into the obj directory.

Tracking these leads to situations where meta mode will consider the
file to be out of date if /bin/sh or /bin/ln are newer than the source
file. There's no reason for meta mode to do this as make is already
handling the rebuild dependency fine.

Sponsored by: EMC / Isilon Storage Division

META MODE: Prefer INSTALL=tools/install.sh to lessen the need for xinstall.host.

This both avoids some dependencies on xinstall.host and allows
bootstrapping on older releases to work due to lack of at least 'install -l'
support.

Sponsored by: EMC / Isilon Storage Division

Add META_MODE support.

Off by default, build behaves normally.
WITH_META_MODE we get auto objdir creation, the ability to
start build from anywhere in the tree.

Still need to add real targets under targets/ to build packages.

Differential Revision: D2796
Reviewed by: brooks imp

Remove gnu/ info pages to unbreak the build with MK_GCC != no, etc

Reported by: Ivan Klymenko <fidaj@ukr.net>
Reviewed by: bapt
Differential Revision: D1412

Convert to LIBADD
Reduce overlinking

Remove MK_ARM_EABI, the armeb issues have been fixed. The code to support
the oabi is still in the tree, but it is expected this will be removed
as developers work on surrounding code.

With this commit the ARM EABI is the only supported supported ABI by
FreeBSD on ARMa 32-bit processors.

X-MFC after: never
Relnotes: yes
Differential Revision: https://reviews.freebsd.org/D876

Allow libstdc++ and libsupc++ to compile with clang again, after the
bsd.*.mk infrastructure changes. Apparently, you must now modify
CXXFLAGS *before* including bsd.lib.mk, or your changes will be lost.

Use src.opts.mk in preference to bsd.own.mk except where we need stuff
from the latter.

To avoid having to explicitly test COMPILER_TYPE for setting
clang-specific or gcc-specific flags, introduce the following new
variables for use in Makefiles:

CFLAGS.clang
CFLAGS.gcc
CXXFLAGS.clang
CXXFLAGS.gcc

In bsd.sys.mk, these get appended to the regular CFLAGS or CXXFLAGS for
the right compiler.

MFC after: 1 week

After r255321, clang uses libc++ by default. This leads to a lot of
errors when you enable WITH_GNUCXX to build libstdc++, since it will
include C++ headers from the libc++ installation under ${WORLDTMP}, and
those are not compatible with libstdc++ at all.

To fix this, add -stdlib=libstdc++ to CXXFLAGS when building libstdc++
(and its companion libsupc++) with clang.

Approved by: re (delphij)

Teach libstdc++ about logl(3).

The logl(3) family of functions were implemented in r251292.
Define them in libstdc++'s configuration so they can be used.

Reviewed by: dim

Regenerate libstdc++'s config.h, synchronizing it with our current
almost-C99 headers.

MFC after: 1 week

Add compiler support for the ARM EABI.

ARM EABI support is disabled by default and can be enabled by setting
WITH_ARM_EABI when building, however only the kernel-toolchain target will
work with this flag until the rest of the support is added.

For building GNU libstdc++ and libsupc++, filter out libc++-specific and
-std= flags above c++98 or gnu++98 from CXXFLAGS *after* including
bsd.lib.mk, otherwise some additions to CXXFLAGS done in bsd.*.mk might
not be preserved.

MFC after: 1 week

For building libstdc++ and libsupc++, filter out any -stdlib=libc++
option from CXXFLAGS, otherwise these libraries will not build.
Similarly, filter out any -std=xxx options that aren't supported.

Submitted by: Yamaya Takashi <yamayan@kbh.biglobe.ne.jp>
MFC after: 2 weeks

Turn on TLS support for arm on here as it is supported since r231618/
r231619 and working since r233106.

Make libsupc++ build as a shared library and make libstdc++ a filter library
for it.

This allows people to swap out libsupc++ for libcxxrt easily, so we can begin
the libstdc++ -> libc++ migration.

Approved by: dim (mentor)

When one attempts to compile the tree with -march=i386, which also used
to be gcc's default before r198344, calls to atomic builtins will not be
expanded inline. Instead, they will be generated as calls to external
functions (e.g. __sync_fetch_and_add_N), leading to linking errors later
on.

Put in a seatbelt that disables use of atomic builtins in libstdc++ and
llvm, when tuning specifically for the real i386 CPU. This does not
protect against all possible issues, but it is better than nothing.

Now that TLS is supported for sparc64 by both binutils 2.17.50 committed
in r218822 and rtld(1) committed in r219533 turn on TLS support in GCC.

GCC defines built-ins for atomic instructions found on i486 and higher.
Because FreeBSD no longer supports the 80386 cpu all code targeting
FreeBSD/i386 necessarily runs on i486 or higher so the compiler
built-ins can be used by default inside libstdc++ and in C++ headers.
This allows newly compiled C++ code to inline some atomic operations.
Old binaries continue to use libstdc++ functions.

PR: 148926
Tested by: Yuri Karaban <tech askold net>
Reviewed by: kan
Approved by: kib (mentor)
MFC after: 2 weeks

MFtbemd:

Prefer MACHNE_CPUARCH to MACHINE_ARCH in most contexts where you want
to test of all the CPUs of a given family conform.

Non-GCC gcc compatible compilers may provide the same multimedia intrinsic
headers as GCC, but of their own implementation. So put the GCC ones into
their own header "namespace".

Requested by: ed

Use the new PO_CXXFLAGS from r202807 to remove the '-ffunction-sections'
flag when compiling profiled objects as it's ignored by the compiler.
This removes the associated warning for each file compiled.

MFC after: 1 month

There's no need to redundantly redefine atomicity.h.

Install extra include files that were forgotten in original GCC 4.2.x
import changes.

PR: 124647
Submitted by: Vlad GALU
MFC after: 2 days

Add MIPS to the list of "no TLS" architectures.

Approved by: cognet (mentor)

Install etx/numeric_traits.h.

PR: gnu/115250
Approved by: re (blanket)

Fix the fallout from over-zealous obsolete files removal. Use correct
atomicity.h file for arm and powerpc.

Update bmake glue to build GCC 4.2.

Also:
Switch FreeBSD to use libgcc_s.so.1.

Use dl_iterate_phdr to locate shared objects' exception frame
info instead of depending on older register_frame_info machinery.
This allows us to avoid depending on libgcc_s.so.1 in binaries
that do not use exception handling directly. As an additional
benefit it breaks circular libc <=> libgcc_s.so.1 dependency too.

Build newly added libgomp.so.1 library, the runtime support
bits for OpenMP.

Build LGPLed libssp library. Our libc provides our own
BSD-licensed SSP callbacks implementation, so this library
is only built to benefit applications that have hadcoded
knowledge of libssp.so and libssp_nonshared.a. When linked
in from command line, these libraries override libc
implementation.

The ARM platform does the setjmp/longjmp type of exception handling.

Fix build due to the incomplete removal of demangle.h.

Reduce diffs with file generated by FSF configure.

Include <sys/defs.h> to get __ISO_C_VISIBLE definition is one is not
already available.

Do not expose vfwscanf, vswscanf, vwscanf and wcstof functions if
they are not visible from wchar.h.

Bump the shared library version number of all libraries that have not
been bumped since RELENG_5.

Reviewed by: ru
Approved by: re (not needed for commit check but in principle...)

Enforce style.Makefile(5).

Glanced at by: ru (some time ago).

Teach libstdc++ about frexpl() and ldexpl().

More math functions.

Prodded by: stefanf

We have floorl() now. Reflect this in the library config file.

Of all templates only lists need a special implmementation file compiled
to support __GLIBCXX_DEBUG. Add debug_list.cc to the list of files to
compile.

Add -frandom-seed=RepeatabilityConsideredGood to CFLAGS for building this
library. As the value suggests, this allows the library to be built repeatably;
without this flag, gcc uses a random value in its parsing.

Since the random seed is only used when handling files which do not have any
externally-visible symbols, this change is not needed for any other libraries
in the FreeBSD base system.

Discussed on: freebsd-arch (in early November)
Approved by: kan
MFC after: 1 week

Add atomicity.cc to CLEANFILES.

Bmake glue for GCC 3.4.2-prerelease.

We generate this file from config.h automatically.

Fixed style of DPADD and LDADD assignments as per style.Makefile(5).

Correct the atomicity.h fallback path on architectures which do not
define their own.

Do not install files which are not part of GCC 3.3.x anymore.

Update bmake bits to build GCC 3.3.1-prerelease.

Update for GCC 3.3.1-pre.

Oops, remove hack to stop libstdc++.so being built now that gcc uses
-lgcc_pic by default on -fpic stuff. I forgot about this here.

Do not bogusly set CSTD to the empty string.

Something funky is going on here... disabling CSTD makes it go away...

libstdc++.so breaks on amd64 due to bogons in our build, so prevent the
shared library being built for amd64. The problem is that libstdc++.so
is produced with 'cc -shared'. This has an internal -lgcc, which is
not PIC. libstdc++.so uses exceptions and the dwarf2 unwinder, which
are in libgcc.a. As a result, non-PIC code gets pulled into libstdc++.so.
This is fatal on amd64 when certain relocation types cannot be used in
PIC mode. The official FSF solution to this is to have libgcc.so with
internal ELF symbol versioning to solve the ABI problem, but I dont want
to fight that battle yet. I tried making libgcc_pic.a (which worked
fine), but thats not something for the 11th hour before a release.

Approved by: re (amd64 "safe" stuff)

Enable wchar_t support in libstdc++.

Enable ISO-C99 compatibility -- among other things, this tells libstdc++
not to blindly undef isnan() and other functions that became macros in C99.
Enable use of newly grown C99 functions: strtof(), strtold(), wcstof()

Submitted by: das

Update for GCC 3.2.2

Install one more header file which is new with GCC 3.2.2.

C++ exceptions with shared libraries are working on sparc64 now.
Build shared libstdc++so on sparc64 platform too.

Approved: re (rwatson)

Update for Gcc 3.2.1 release.
We also have more wide char support in libc now.

Remove files no longer part of the gcc_3_2_anoncvs_20021009 libstdc++.

Approved by: re(jhb)

Zap now-unused SHLIB_MINOR

Bmake bits for GCC 3.2.1-prerelease libstdc++.
Connect two new source and one header file to the build.

Approved by: obrien

Connect new files added with GCC 3.2.1 import to the build.

Approved by: obrien

Bandaid for sparc64. Do not build libstdc++.so for now on sparc64.
Binaries compiled with it segfault.

Disconnect the docs until we figure out if there are any with v3.

Turn on the math function stub stubs.

Submitted by: Martin Blapp <mb@imp.ch>
I believe OK with: peter

The new bsd.incs.mk is powerful enough to handle this wrinkle.

Tweak so that .o's are written into src/contrib/gcc

Build docs not out of libobc++.295

Add some build glue for libstdc++. I'm not a C++ person, but this seems
to work at least for the non-hairy stuff. The main wrinkle here is that
a whole mess of include files get installed and under different names.
An earlier version of this built a shadow include tree first in the obj
directory, but this depends on the 'make includes' functionality.

More tweaking is certainly going to be needed.

Added new bsd.incs.mk which handles installing of header files
via INCS. Implemented INCSLINKS (equivalent to SYMLINKS) to
handle symlinking include files. Allow for multiple groups of
include files to be installed, with the powerful INCSGROUPS knob.
Documentation to follow.

Added standard `includes' and `incsinstall' targets, use them
in Makefile.inc1. Headers from the following makefiles were
not installed before (during `includes' in Makefile.inc1):

kerberos5/lib/libtelnet/Makefile
lib/libbz2/Makefile
lib/libdevinfo/Makefile
lib/libform/Makefile
lib/libisc/Makefile
lib/libmenu/Makefile
lib/libmilter/Makefile
lib/libpanel/Makefile

Replaced all `beforeinstall' targets for installing includes
with the INCS stuff.

Renamed INCDIR to INCSDIR, for consistency with FILES and SCRIPTS,
and for compatibility with NetBSD. Similarly for INCOWN, INCGRP,
and INCMODE.

Consistently use INCLUDEDIR instead of /usr/include.

gnu/lib/libstdc++/Makefile and gnu/lib/libsupc++/Makefile changes
were only lightly tested due to the missing contrib/libstdc++-v3.
I fully tested the pre-WIP_GCC31 version of this patch with the
contrib/libstdc++.295 stuff.

These changes have been tested on i386 with the -DNO_WERROR "make
world" and "make release".

Bmake bits for Gcc 3.1.

Partially made possible by: Wilko.Bulte@compaq.com

libstdc++ version 3 (Gcc 3.1 is a consumer) uses the typical `config.h'.

Sigh. I should know better than to commit a one line change from a PR
w/o a full compile test, even when the PR seemed so authoritative on the
subject...

Install sstream, not the non-existent sstring.

Install sstring.

PR: 25927
Submitted by: Jeremy Norris <ishmael27@home.com>

Scoot things over to the temporary *.295 source while I do major construction
on the mainline sources.

Do not use "DEFAULT_VTABLE_THUNKS". The bugs that have existed since
EGCS 1.x have not been worked out. And with 4.0 RELEASE comming quickly
we need C++ to be stable and usable.

Stop my evil plot to stop the use of GNU info files (which I don't care for)
for people that don't have "NOINFO=yes" in /etc/make.conf.
[read: this still refered to contrib/egcs/]

Cut over the system compiler from from EGCS 1.1.2 to GCC 2.95.2.

$Id$ -> $FreeBSD$

{EGCSDIR}->{SRCDIR} + associated tweaks.
This is in prepreparation of the gcc-2.95 upgrade.

Purely cosmetic changes -- fix Id's

_G_USING_THUNKS probably applies to a.out too.
(well we define DEFAULT_VTABLE_THUNKS=1 for G++ in the a.out case, so it seems
likely)

Submitted by: jdp

Remove several explicit rules. The two for exceptioni.o and typeionfoi.o
aren't needed as those files aren't part of libstdc++ any longer. Another
isn't needed as Cygnus doesn't compile with -frtti.

Noticed by: bde & my initial mispelling of ".So".

Foo. jdp changed .so to .So during the time I did the EGCS work and I
missed bring in his changes into my work.

Missed a "\" in a continuation line.

Get stdargs right on the Alpha.

Submitted by: Hidetoshi Shimokawa <simokawa@sat.t.u-tokyo.ac.jp>

Cometic cleanup.

Pre-bdelinted by: bde

Don't generate _G_config.h, use fixed version.
I knew better... too dependant on the environment we generate in, but...

This fixes the
/usr/lib/libstdc++.so: undefined reference to `filebuf virtual table'
/usr/lib/libstdc++.so: undefined reference to `stdiobuf virtual table'
errors seen after the initial bootstrap from gcc 2.7.2 to EGCS.

Agreed with by: bde & jdp

Crap! I grabbed the wrong Makefile from the test machine.

Fix .for loop continuation lines style bugs.

Submitted by: bde

Add iostream Info docs.

_G_config.h is now dynamically created.

Minimum set of changes to switch from Gcc 2.7.2 (in contrib/gcc) to Egcs 1.1.2
(in contrib/egcs)

Switch to using ".So" as the extension for PIC object files rather
than ".so". The old extension conflicted with well-established
naming conventions for dynamically loadable modules.

The "clean" targets continue to remove ".so" files too, to deal with
old systems.

Avoid using `ld -O' (as in bsd.lib.mk). `ld -O' seems to have caused only
non-problems for elf -- `ld -elf -O foo.tmp -x -r foo' apparently puts
the result in a.out so it has essentially no effect.

libg++ and libstdc++ also have internal calls to libm. Programs using the
shared versions should not need to add -lm unless the program uses libm
itself. Strictly speaking, libg++ depends on libstdc++, but libstdc++
has dynamic dependencies on some exception tables and binutils doesn't
seem to like it when a secondary library has undefined references. It
doesn't seem to care when -lstdc++ is added on the command line to ld
though. Anyway, the c++ driver adds -lstdc++ explicitly, so that should
be OK. c++ also adds -lm explicitly too, even though it wouldn't need
to now. [except for statically linked binaries as .a files don't have an
equivalent of automatic internal shared object dependencies.]

Get definition of va_list from <stdarg.h>.

This was broken for ELF builds.

Don't use beforedepend to complicate and break things. The usual
fix of putting generated source files in SRCS breaks many special
cases (many dependencies had to be incomplete for ${.ALLSRC}
not to give .h files that would break compiling with c++ -c).
Reduce these special cases to the general case so that SRCS works
normally and bsd.lib.mk handles most of the complications. Now
we only have to duplicate rules from bsd.lib.mk to add special
CFLAGS in some cases.

Don't add streambuf.cc to SRCS twice.

_G_VTABLE_LABLE_PREFIX is different between ELF and a.out
('_vt.' vs '__vt$')

Add ${DESTDIR} in front of absolute paths.

Reviewed by: bde

Bye bye CPLUSPLUSLIB hack! It's not needed any more.

Don't merge this into -2.2 unless you understand the dependencies on
c++rt0, bsd.lib.mk, and gcc -shared. I.e., let me do it.

Oops. Missed the Complex number add case of code generation.. :-]
Fixes PR#3239.

Submitted by: Takeshi WATANABE <watanabe@crayon.earth.s.kobe-u.ac.jp>

Revert $FreeBSD$ to $Id$

Make the long-awaited change from $Id$ to $FreeBSD$

This will make a number of things easier in the future, as well as (finally!)
avoiding the Id-smashing problem which has plagued developers for so long.

Boy, I'm glad we're not using sup anymore. This update would have been
insane otherwise.

STL object files missing from libstdc++ - fix.
This closes PR#1975 and was on my TODO list, so I'm breaking my own "no more
commits before I fly!" rule and taking care of this, since it was at the
very top and Peter shamed me in to taking care of it. This is definitely
a 2.2 candidate.

Submitted-By: Mark Diekhans <markd@Grizzly.COM>

Initial shot at a bmakefile for libstdc++