Older GCCs don't have all the exciting options we have now.  let's just
turn them off.

	libcody/
	* Makefile.in: Disable some flags.

I missed adding these two configure files.

	contrib/
	* gcc_update: Add c++tools & libcody.

When the divisor is bool or has [0, 1] range, as division by
0 is UB, the only remaining option in valid programs is division by 1,
so we can optimize X / bool_range_Y into X.

2020-12-15  Jakub Jelinek  <jakub@redhat.com>

	PR tree-optimization/96094
	* match.pd (X / bool_range_Y -> X): New simplification.

	* gcc.dg/tree-ssa/pr96094.c: New test.

The DMD front-end shouldn't, but can sometimes leak manifest constants
in the AST passed to the code generator.  To prevent this being an
issue, the setting of DECL_INITIAL has been moved to the point where the
CONST_DECL is used, rather than in the declaration handler.

gcc/d/ChangeLog:

	PR d/98277
	* decl.cc (DeclVisitor::visit (VarDeclaration *)): Move setting of
	DECL_INITIAL for manifest constants to ...
	(get_symbol_decl): ... here.

gcc/testsuite/ChangeLog:

	PR d/98277
	* gdc.dg/pr98277.d: New test.

This matches changes in 788d2048.

For golang/go#43200

Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/278452

The test case for this is issue11614.go.

Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/278192

This is specifically for the test fixedbugs/issue23781.go, which
builds a large static array.  The code does compile and work without
this change, but it takes a long time and generates a large object
file.  Discarding the unnecessary static initializer makes this test
much faster.

Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/278172

A digit character is not invalid in general, but it's not permitted
at the start of an identifier.  Report a better error message.
The test case is issue11359.go in the source repo.

Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/278174

I missed some other places that used +=

	* config.m4: Avoid var+=...
	* configure: Rebuilt

Apparently 'var+=...' is not a dash thing.  Fixed thusly.

	* config.m4: Avoid non-dash idiom
	* configure: Rebuilt.

This reverts commit cf4ed3b4.

libstdc++-v3/ChangeLog:

    PR libstdc++/98108
    * include/std/iostream (__ioinit): Remove init_priority attribute.

The current libstdc++ basic_file_stdio.cc code assumes a POSIX API
underneath the stdio implementation provided by the host libc. This
means that the host must provide a fairly broad POSIX file API,
including read, write, open, close, lseek and ioctl.

This patch changes basic_file_stdio.cc to only use basic ANSI-C stdio
functions, allowing it to be used with libc implementations like
picolibc which may not have a POSIX operating system underneath.

This is enabled by a new --enable-cstdio=stdio_pure configure option.

Aided-by: Jonathan Wakely <jwakely@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>

libstdc++-v3/ChangeLog:

	* acinclude.m4 (GLIBCXX_ENABLE_CSTDIO): Allow "stdio_pure"
	option and define _GLIBCXX_USE_PURE_STDIO when it is used. Also
	add "stdio_posix" option as an alias for "stdio".
	* config/io/basic_file_stdio.cc [_GLIBCXX_USE_PURE_STDIO]: Only
	use defined stdio entry points for all I/O operations, without
	direct calls to underlying POSIX functions.
	* config.h.in: Regenerate.
	* configure: Regenerate.

During satisfaction that's entered through the three-parameter version
of satisfy_declaration_constraints, current_function_decl gets set to
the dependent DECL_TEMPLATE_RESULT for sake of access checking.  This
makes the predicate in_template_function return true during satisfaction
from this entrypoint, which in turn makes calls to mark_used exit early
before it does its full processing.  This leads to us accepting the
invalid testcase below due to mark_used never attempting to deduce the
return type of A::foo() and detecting failure thereof.

It seems wrong for in_template_function to be true during instantiation
or during satisfaction, so this patch makes in_template_function inspect
the less volatile cfun->decl instead of current_function_decl.

gcc/cp/ChangeLog:

	* pt.c (in_template_function): Inspect cfun->decl instead of
	current_function_decl.

gcc/testsuite/ChangeLog:

	* g++.dg/cpp2a/concepts-requires23.C: New test.

Add ChangeLog files, so that update_version_git can then fill those in.

gcc/fortran/ChangeLog:

	PR fortran/90207
	* invoke.texi: Document -fdebug-aux-vars.
	* lang.opt: Add -fdebug-aux-vars.
	* trans.c (MAX_PREFIX_LEN): New macro.
	(create_var_debug_raw): New function.
	(gfc_create_var_np): Call create_var_debug_raw if
	flag_debug_aux_vars is set.

libstdc++-v3/ChangeLog:

	* testsuite/27_io/rvalue_streams.cc: Run the extraction to a char*
	for C++17 and lower only.

And here is the user-facing documentation.

	gcc/
	* doc/cppopts.texi: Document new cpp opt.
	* doc/invoke.texi: Add C++20 module option & documentation.

This is the C++ modules implementation.  Dropping in the real
module.cc file into the module-shaped hole I carved out.  I include
some sanity tests for now.  I'll commit the remaining tests when the
initial round of failures seems abated.

	gcc/cp/
	* module.cc: Replace stubs with implementation.
	gcc/testsuite/
	* g++.dg/modules/mod-decl-0-2a.C: New.
	* g++.dg/modules/mod-decl-0.C: New.
	* g++.dg/modules/mod-decl-1.C: New.
	* g++.dg/modules/mod-decl-2_a.C: New.
	* g++.dg/modules/mod-decl-2_b.C: New.
	* g++.dg/modules/mod-decl-2_c.C: New.
	* g++.dg/modules/mod-decl-3.C: New.
	* g++.dg/modules/mod-decl-5_a.C: New.
	* g++.dg/modules/mod-decl-5_b.C: New.

To avoid always requiring an active mapper to connect to, we provide a
default in-process mapper with similar functionality to the sample
server.  This is that code.  Also included is the client-side
connection mechanism, which determines what server to use and how to
connect to it.

	gcc/cp/
	* Make-lang.in (CXX_AND_OBJCXX_OBJS): Add mapper-client &
	mapper-resolver.
	* mapper-client.h: New.
	* mapper-client.cc: New.
	* mapper-resolver.cc: New.

These are the toplevel directory changes to add c++tools

ChangeLog:

	* Makefile.def
	* Makefile.in
	* Makefile.tpl
	* configure
	* configure.ac

Part of our module implementation adds a sample mapper server, the
guts of which are used by the default in-process mapping of cc1plus.
Rather than add another executable to gcc/cp/, this creates a new
c++tools directory where this and any other c++ tools might live.
The toplevel changes are a subsequent commit, because ... git.

c++tools/ChangeLog:

	* Makefile.in: New.
	* config.h.in: New.
	* configure: New.
	* configure.ac: New.
	* resolver.cc: New.
	* resolver.h: New.
	* server.cc: New.

These tests FAIL when testing debug mode with a small tool_timeout
value. Give them a longer relative timeout.

libstdc++-v3/ChangeLog:

	* testsuite/25_algorithms/lexicographical_compare/deque_iterators/1.cc:
	Add dg-timeout-factor directive.
	* testsuite/ext/pb_ds/regression/tree_map_rand_debug.cc:
	Increase timeout factor from 2.0 to 3.0.
	* testsuite/ext/pb_ds/regression/tree_set_rand_debug.cc:
	Likewise.
	* testsuite/lib/libstdc++.exp (check_effective_target_debug-mode):
	Define "debug-mode" as an effective-target keyword.

There's no point even checking is_constant_evaluated() in C++11 mode,
because the 'if' statement used for the assertion wouldn't be valid in a
C++11 constexpr function anyway.

libstdc++-v3/ChangeLog:

	* include/bits/c++config (__glibcxx_assert_1): Define as empty
	for C++11.

In order to separate compiler from build system, C++ Modules, as
implemented in GCC introduces a communication channel between those
two entities.  This is implemented by libcody.  It is anticipated that
other implementations will also implement this protocol, or use
libcody to provide it.

	* Makefile.def: Add libcody.
	* configure.ac: Add libcody.
	* Makefile.in: Regenerated.
	* configure: Regenerated.
	gcc/
	* Makefile.in (CODYINC, CODYLIB, CODYLIB_H): New. Use them.
	libcody/
	* configure.ac: New.
	* CMakeLists.txt: New.
	* CODING.md: New.
	* CONTRIB.md: New.
	* LICENSE: New.
	* LICENSE.gcc: New.
	* Makefile.in: New.
	* Makesub.in: New.
	* README.md: New.
	* buffer.cc: New.
	* build-aux/config.guess: New.
	* build-aux/config.sub: New.
	* build-aux/install-sh: New.
	* client.cc: New.
	* cmake/libcody-config-ix.cmake
	* cody.hh: New.
	* config.h.in: New.
	* config.m4: New.
	* configure: New.
	* configure.ac: New.
	* dox.cfg.in: New.
	* fatal.cc: New.
	* gdbinit.in: New.
	* internal.hh: New.
	* netclient.cc: New.
	* netserver.cc: New.
	* packet.cc: New.
	* resolver.cc: New.
	* server.cc: New.
	* tests/01-serialize/connect.cc: New.
	* tests/01-serialize/decoder.cc: New.
	* tests/01-serialize/encoder.cc: New.
	* tests/02-comms/client-1.cc: New.
	* tests/02-comms/pivot-1.cc: New.
	* tests/02-comms/server-1.cc: New.
	* tests/Makesub.in: New.
	* tests/jouster: New.

	contrib/
	* gcc-changelog/git_commit.py (changelog_locations): Add
	libcody, c++tools.

In manually applying a name-lookup patch, I managed to transpose two
calls.  That caused a test failure.  Fixed thusly.

	gcc/cp/
	* name-lookup.c (do_namespace_alias): Set originating module
	before pushing.

The testsuite for libstdc++ aims to skips test cases for which not all
required locales are installed. This patch adds missing directives about
required locales to one test case to avoid false positive test failures
on systems that have a partial set of locales installed.

Verified by test suite runs that this patch changes the test case from
FAIL to UNSUPPORTED when not all required locales are available and that
the test case will run and PASS when the necessary locales have been
added.

libstdc++-v3/ChangeLog:

	* testsuite/22_locale/locale/cons/5.cc: Add missing directives
	for required locales.

Now that the G++ bug is fixed we no longer need to protect this partial
specialization from complaining about subtracting void pointers.

libstdc++-v3/ChangeLog:

	* include/bits/iterator_concepts.h (incrementable_traits<Tp>):
	Remove workaround for PR c++/78173.

The AC_LIBTOOL_DLOPEN checks were previously disabled for newlib targets.
The patch applies similar logic to avr-libc based builds.

libstdc++-v3/ChangeLog:

	* configure.ac: Skip AC_LIBTOOL_DLOPEN check if avr-libc is used.
	* configure: Regenerate.

This causes the global objects that run the <iostream> initialization
code to be constructed earlier, which avoids some bugs in user code due
to incorrectly relying on static initialization order.

libstdc++-v3/ChangeLog:

	PR libstdc++/98108
	* include/std/iostream (__ioinit): Add init_priority attribute.

gcc/ada/

	* libgnat/g-sercom__linux.adb (Set): Use cfsetospeed and
	cfsetispeed to set the baud rate. Clear non-blocking serial port
	status when blocking is requested.

gcc/ada/

	* libgnat/s-valrea.adb (Integer_to_Real): Always use Extra.

gcc/ada/

	* par-ch5.adb (P_Condition): Simplify condition for warning
	about extra parens and make it easier to understand.

gcc/ada/

	* errout.ads (Error_Msg_Ada_2005_Extension): New routine (spec).
	* errout.adb (Error_Msg_Ada_2005_Extension): New routine (body).
	* par-ch10.adb: Reuse new routine; correct casing for "LIMITED
	WITH".
	* par-ch11.adb: Likewise.
	* par-ch12.adb: Likewise.
	* par-ch3.adb: Likewise.
	* par-ch4.adb: Likewise; replace "box" with "<>".
	* par-ch6.adb: Likewise.
	* par-ch9.adb: Likewise; correct casing for "THEN ABORT".

gcc/ada/

	* doc/gnat_ugn/gnat_and_program_execution.rst: Minor fix.
	* gnat_ugn.texi: Regenerate.
	* libgnat/s-valuer.ads (Precision_Limit): New formal parameter.
	* libgnat/s-valuer.adb (Precision_Limit): Remove.
	(Scan_Decimal_Digits): Robustify overflow check.
	(Scan_Integral_Digits): Likewise.
	* libgnat/s-valrea.adb: Add assertion on the size of the unsigned
	type and instantiate System.Value_R with the mantissa limit.
	(Integer_to_Real): Add Extra parameter and take it into account.
	(Scan_Real): Pass Extra to Integer_to_Real.
	(Value_Real): Likewise.
	* libgnat/s-valued.adb: Add assertion on the size of the unsigned
	type and instantiate System.Value_R with the mantissa limit.
	* libgnat/s-valuef.adb: Likewise.

gcc/ada/

	* contracts.adb, contracts.ads (Build_Postconditions_Procedure):
	Add declarations for Postcond_Enabled,
	Result_Object_For_Postcondition, and
	Return_Success_For_Postcond, and place all postconditions within
	an if statement to control their execution for interactions when
	cleanup actions get generated.
	(Get_Postcond_Enabled): Created to fetch object declared to
	handle new expansion of postconditions.
	(Get_Result_Object_For_Postcond): Created to fetch object
	declared to handle new expansion of postconditions.
	(Get_Return_Success_For_Postcond): Created to fetch object
	declared to handle new expansion of postconditions.
	* einfo.adb, einfo.ads: Modify flag Stores_Attribute_Old_Prefix
	to apply to constants, variables, and types.
	* exp_ch6.adb (Add_Return): Add assignment to
	Return_Success_For_Postcond.
	(Expand_Non_Function_Return): Add assignment to
	Return_Success_For_Postcond
	(Expand_Simple_Function_Return): Add assignment to
	Result_Object_For_Postcond and Return_Success_For_Postcond.
	* exp_ch7.adb (Build_Finalization_Master): Mark finalization
	masters which finalize types created store 'Old objects as
	storing 'Old objects.
	(Build_Finalizer): Created to generated a unified and special
	expansion for finalization when postconditions are present.
	(Build_Finalizer_Helper): Renamed Build_Finalizer and added
	parameter to facilitate the creation of separate finalization
	routines for 'Old objects and general objects.
	(Create_Finalizer): Add condition for the insertion of the
	finalizer spec to avoid malformed trees.
	(Expand_Cleanup_Actions): Move _postconditions and related
	declarations to the new declarative section.  Fix the loop to
	properly stop at the subprogram declaration for the
	postconditions procedure and exclude its body from being moved
	to the new list of declarations to avoid freezing issues.
	* exp_prag.adb (Expand_Attributes): Mark temporary created to
	store 'Old objects as storing a 'Old attribute.
	* sem_ch6.adb (Find_What_Applies_To): Remove strange exception
	to postconditions when traversing the scope stack.
	* sem_prag.adb (Find_Related_Declaration_Or_Body): Use the newly
	created Enclosing_HSS function to find the HSS for a potentially
	nested statement.
	* sem_util.adb, sem_util.ads (Declare_Indirect_Temp): Mark types
	created to store 'Old objects as storing 'Old attributes.
	(Enclosing_HSS): Created to find the enclosing handled sequence
	of statements for a given statement.
	* snames.ads-tmpl: Add multiple names to aid in the expansion of
	finalization and to control the evaluation of postconditions.
	Including _finalization_controller, a new routine to centralize
	finalization actions and postcondition evaluation.

gcc/ada/

	* par-ch5.adb (P_Loop_Parameter_Specification): Complain about
	missing -gnat2020 switch.
	(P_Iterator_Specification): Likewise.

gcc/ada/

	* par-ch4.adb (P_Aggregate_Or_Paren_Expr): Simplify with
	Append_New.

gcc/ada/

	* Makefile.rtl (GNATRTL_NONTASKING_OBJS): Remove s-fatsfl$(objext)
	and add s-valflt$(objext), s-vallfl$(objext), s-valllf$(objext).
	* exp_attr.adb (Find_Fat_Info): Merge Short_Float and Float cases.
	* exp_imgv.adb (Expand_Value_Attribute): Replace RE_Value_Real with
	RE_Value_Long_Long_Float for fixed-point types and use appropriate
	base type for floating-point types.
	* rtsfind.ads (RTU_Id): Remove System_Fat_IEEE_Long_Float,
	System_Fat_IEEE_Short_Float and System_Val_Real, add System_Val_Flt,
	System_Val_LFlt and System_Val_LLF.
	(RE_Id): Remove RE_Attr_IEEE_Long, RE_Fat_IEEE_Long,
	RE_Attr_IEEE_Short, RE_Fat_IEEE_Short, RE_Attr_Short_Float, add
	RE_Value_Float, RE_Value_Long_Float, RE_Value_Long_Long_Float,
	(RE_Unit_Table): Likewise.
	* libgnat/a-ticoau.ads: Add with clause for Float_Aux and make the
	package generic.
	(Get): Change parameter types to Num.
	(Put): Likewise.
	(Gets): Likewise.
	(Puts): Likewise.
	* libgnat/a-ticoau.adb: Remove clause and renaming for Float_Aux.
	(Get): Change parameter types to Num.
	(Gets): Likewise.
	(Put): Likewise.
	(Puts): Likewise.  Add conversion to Long_Long_Float.
	* libgnat/a-ticoio.adb: Remove with clause for Ada.Text_IO, add with
	clause for Float_Aux, add with and use clauses for System.Val_Flt,
	System.Val_LFlt and System.Val_LLF. Instantiate Float_Aux and
	Complex_Aux on Float, Long_Float, and Long_Long_Float.
	(OK_Float): New boolean constant.
	(OK_Long_Float): Likewise.
	(Get): Call appropriate Get routine from auxiliary package.
	(Get): Call appropriate Gets routine from auxiliary package.
	(Put): Call appropriate Put routine from auxiliary package.
	(Put): Call appropriate Puts routine from auxiliary package.
	* libgnat/a-tideau.adb: Remove with and use clause for Float_Aux.
	* libgnat/a-tifiau.adb: Likewise.
	* libgnat/a-tifiio.adb: Add with and use clause for System.Val_LLF.
	Instantiate Float_Aux on Long_Long_Float.
	(Get): Adjust call to Get routine from auxiliary package.
	(Get): Adjust call to Gets routine from auxiliary package.
	(Put): Adjust call to Put routine from auxiliary package.
	(Put): Adjust call to Puts routine from auxiliary package.
	* libgnat/a-tifiio__128.adb: Likewise.
	(Get): Likewise.
	(Get): Likewise.
	(Put): Likewise.
	(Put): Likewise.
	* libgnat/a-tiflau.ads: Make the package generic.
	(Get): Change parameter type to Num.
	(Put): Likewise.
	(Gets): Likewise.
	(Puts): Likewise.
	* libgnat/a-tiflau.adb: Remove clauses for System.Val_Real.
	(Get): Change parameter type to Num and call Scan routine.
	(Gets): Likewise.
	(Load_Real): Move to...
	(Put): Change parameter type and add conversion to Long_Long_Float.
	(Puts): Likewise.
	* libgnat/a-tiflio.adb: Add with and use clauses for System.Val_Flt,
	System.Val_LFlt and System.Val_LLF. Instantiate Float_Aux on Float,
	Long_Float and Long_Long_Float.
	(OK_Float): New boolean constant.
	(OK_Long_Float): Likewise.
	(Get): Call appropriate Get routine from auxiliary package.
	(Get): Call previous variant.
	(Get): Call appropriate Gets routine from auxiliary package.
	(Put): Call appropriate Put routine from auxiliary package.
	(Put): Call previous variant.
	(Put): Call appropriate Puts routine from auxiliary package.
	* libgnat/a-tigeau.ads (Load_Real): New procedure.
	* libgnat/a-tigeau.adb (Load_Real): ...here.
	* libgnat/a-wtcoau.ads: Add with clause for Float_Aux and make the
	package generic.
	(Get): Change parameter types to Num.
	(Put): Likewise.
	(Gets): Likewise.
	(Puts): Likewise.
	* libgnat/a-wtcoau.adb: Remove clause and renaming for Float_Aux.
	(Get): Change parameter types to Num.
	(Gets): Likewise.
	(Put): Likewise.
	(Puts): Likewise.  Add conversion to Long_Long_Float.
	* libgnat/a-wtcoio.ads: Remove use clause for Complex_Types and use
	qualified names throughout accordingly.
	* libgnat/a-wtcoio.adb: Remove clause for Ada.Unchecked_Conversion,
	add with clause for Float_Aux, add clauses for System.Val_Flt,
	System.Val_LFlt and System.Val_LLF. Add clause for Complex_Types.
	Instantiate Float_Aux and Complex_Aux on Float, Long_Float, and
	Long_Long_Float.  Remove LLF subtype and TFT instantiation.
	(OK_Float): New boolean constant.
	(OK_Long_Float): Likewise.
	(Get): Call appropriate Get routine from auxiliary package.
	(Get): Call appropriate Gets routine from auxiliary package.
	(Put): Call appropriate Put routine from auxiliary package.
	(Put): Call appropriate Puts routine from auxiliary package.
	* libgnat/a-wtdeau.adb: Remove with and use clause for Float_Aux.
	* libgnat/a-wtfiau.adb: Likewise.
	* libgnat/a-wtfiio.adb: Add with and use clause for System.Val_LLF.
	Instantiate Float_Aux on Long_Long_Float.
	(Get): Adjust call to Get routine from auxiliary package.
	(Get): Adjust call to Gets routine from auxiliary package.
	(Put): Adjust call to Put routine from auxiliary package.
	(Put): Adjust call to Puts routine from auxiliary package.
	* libgnat/a-wtfiio__128.adb: Likewise.
	(Get): Likewise.
	(Get): Likewise.
	(Put): Likewise.
	(Put): Likewise.
	* libgnat/a-wtflau.ads: Make the package generic.
	(Get): Change parameter type to Num.
	(Put): Likewise.
	(Gets): Likewise.
	(Puts): Likewise.
	* libgnat/a-wtflau.adb: Remove clauses for System.Val_Real.
	(Get): Change parameter type to Num and call Scan routine. Set
	Ptr parameter lazily.
	(Gets): Likewise.
	(Load_Real): Move to...
	(Put): Change parameter type and add conversion to Long_Long_Float.
	Bump buffer length to Max_Real_Image_Length.
	(Puts): Likewise.
	* libgnat/a-wtflio.adb: Add with and use clauses for System.Val_Flt,
	System.Val_LFlt and System.Val_LLF. Instantiate Float_Aux on Float,
	Long_Float and Long_Long_Float.
	(OK_Float): New boolean constant.
	(OK_Long_Float): Likewise.
	(Get): Call appropriate Get routine from auxiliary package.  Add
	pragma Unsuppress (Range_Check) and manual validity check.
	(Get): Call appropriate Gets routine from auxiliary package. Add
	pragma Unsuppress (Range_Check) and manual validity check.
	(Put): Call appropriate Put routine from auxiliary package.
	(Put): Call appropriate Puts routine from auxiliary package.
	* libgnat/a-wtgeau.ads (Load_Real): New procedure.
	* libgnat/a-wtgeau.adb (Load_Real): ...here.
	* libgnat/a-ztcoau.ads: Add with clause for Float_Aux and make the
	package generic.
	(Get): Change parameter types to Num.
	(Put): Likewise.
	(Gets): Likewise.
	(Puts): Likewise.
	* libgnat/a-ztcoau.adb: Remove clause and renaming for Float_Aux.
	(Get): Change parameter types to Num.
	(Gets): Likewise.
	(Put): Likewise.
	(Puts): Likewise.  Add conversion to Long_Long_Float.
	* libgnat/a-ztcoio.ads: Remove use clause for Complex_Types and use
	qualified names throughout accordingly.
	* libgnat/a-ztcoio.adb: Remove clause for Ada.Unchecked_Conversion,
	add with clause for Float_Aux, add clauses for System.Val_Flt,
	System.Val_LFlt and System.Val_LLF. Add clause for Complex_Types.
	Instantiate Float_Aux and Complex_Aux on Float, Long_Float, and
	Long_Long_Float. Remove LLF subtype and TFT instantiation.
	(OK_Float): New boolean constant.
	(OK_Long_Float): Likewise.
	(Get): Call appropriate Get routine from auxiliary package.
	(Get): Call appropriate Gets routine from auxiliary package.
	(Put): Call appropriate Put routine from auxiliary package.
	(Put): Call appropriate Puts routine from auxiliary package.
	* libgnat/a-ztdeau.adb: Remove with and use clause for Float_Aux.
	* libgnat/a-ztfiau.adb: Likewise.
	* libgnat/a-ztfiio.adb: Add with and use clause for System.Val_LLF.
	Instantiate Float_Aux on Long_Long_Float.
	(Get): Adjust call to Get routine from auxiliary package.
	(Get): Adjust call to Gets routine from auxiliary package.
	(Put): Adjust call to Put routine from auxiliary package.
	(Put): Adjust call to Puts routine from auxiliary package.
	* libgnat/a-ztfiio__128.adb: Likewise.
	(Get): Likewise.
	(Get): Likewise.
	(Put): Likewise.
	(Put): Likewise.
	* libgnat/a-ztflau.ads: Make the package generic.
	(Get): Change parameter type to Num.
	(Put): Likewise.
	(Gets): Likewise.
	(Puts): Likewise.
	* libgnat/a-ztflau.adb: Remove clauses for System.Val_Real.
	(Get): Change parameter type to Num and call Scan routine. Set
	Ptr parameter lazily.
	(Gets): Likewise.
	(Load_Real): Move to...
	(Put): Change parameter type and add conversion to Long_Long_Float.
	Bump buffer length to Max_Real_Image_Length.
	(Puts): Likewise.
	* libgnat/a-ztflio.adb: Add with and use clauses for System.Val_Flt,
	System.Val_LFlt and System.Val_LLF. Instantiate Float_Aux on Float,
	Long_Float and Long_Long_Float.
	(OK_Float): New boolean constant.
	(OK_Long_Float): Likewise.
	(Get): Call appropriate Get routine from auxiliary package.  Add
	pragma Unsuppress (Range_Check) and manual validity check.
	(Get): Call appropriate Gets routine from auxiliary package. Add
	pragma Unsuppress (Range_Check) and manual validity check.
	(Put): Call appropriate Put routine from auxiliary package.
	(Put): Call appropriate Puts routine from auxiliary package.
	* libgnat/a-ztgeau.ads (Load_Real): New procedure.
	* libgnat/a-ztgeau.adb (Load_Real): ...here.
	* libgnat/s-fatsfl.ads: Delete.
	* libgnat/s-valflt.ads: New package.
	* libgnat/s-vallfl.ads: Likewise.
	* libgnat/s-valllf.ads: Likewise.
	* libgnat/s-valrea.ads: Make generic. Add assertions, defensive
	code and clarify intent.
	(Scan_Real): Change parameter type to Num.
	(Value_Real): Likewise.
	* libgnat/s-valrea.adb: Instantiate Value_R on Uns.
	(Integer_to_Real): Change parameter and result to Num.
	Call Float_Control.Reset only if the mantissa is 64 bits.  Use
	a divide to compute the final value if the scale is negative.
	(Scan_Real): Change result to Num.
	(Value_Real): Likewise.
	* libgnat/s-valuer.adb: Add assertions, defensive code and
	clarify intent.
	(F_Limit): Delete.
	(I_Limit): Likewise.
	(Precision_Limit): Always use the integer limit.
	* libgnat/s-fatgen.adb: Add pragma Annotate.