This consolidates the three static archives for extensions into one, so
that -lstdc++exp can be used to provide the definitions of all unstable
library features.
The libstdc++_libbacktrace.a archive is now just a "noinst" convenience
library that is only used during the build, not installed. Its contents
are added to libstdc++exp.a, along with the new non-inline definitions
of std::stacktrace symbols.
The libstdc++fs.a archive is still installed, but its contents are
duplicated in libstdc++exp.a now. This means -lstdc++exp can be used
instead of -lstdc++fs. For targets using the GNU linker we should
consider replacing libstdc++fs.a with a linker script that does
INPUT(libstdc++exp.a).
The tests for <experimental/filesystem> could be changed to use
-lstdc++exp instead of -lstdc++fs, which would allow removing
src/filesystem/.libs from the LDFLAGS in scripts/testsuite_flags.in,
but that can be done at a later date.
libstdc++-v3/ChangeLog:
* acinclude.m4 (GLIBCXX_CONFIGURE): Add c++23 directory.
* configure: Regenerate.
* doc/html/manual/*: Regenerate.
* doc/xml/manual/using.xml: Update documentation on linking.
* include/std/stacktrace: Remove declarations of libbacktrace
APIs.
(stacktrace_entry::_S_err_handler, stacktrace_entry::_S_init):
Remove.
(stacktrace_entry::_Info): New class.
(stacktrace_entry::_M_get_info): Use _Info.
(__stacktrace_impl): New class.
(basic_stacktrace): Derive from __stacktrace_impl.
(basic_stacktrace::current): Use __stacktrace_impl::_S_current.
* scripts/testsuite_flags.in: Adjust LDFLAGS to find
libstdc++exp instead of libstdc++_libbacktrace.
* src/Makefile.am (SUBDIRS): Add c++23 directory.
* src/Makefile.in: Regenerate.
* src/c++20/Makefile.am: Fix comment.
* src/c++20/Makefile.in: Regenerate.
* src/c++23/Makefile.am: New file.
* src/c++23/Makefile.in: New file.
* src/c++23/stacktrace.cc: New file with definitions of
stacktrace_entry::_Info and __stacktrace_impl members.
* src/experimental/Makefile.am: Use LIBADD to include other
libraries.
* src/experimental/Makefile.in: Regenerate.
* src/libbacktrace/Makefile.am: Use noinst_LTLIBRARIES.
* src/libbacktrace/Makefile.in: Regenerate.
* testsuite/19_diagnostics/stacktrace/current.cc: Adjust
dg-options to use -lstdc++exp.
* testsuite/19_diagnostics/stacktrace/entry.cc: Likewise.
* testsuite/19_diagnostics/stacktrace/stacktrace.cc: Likewise.
* testsuite/23_containers/vector/debug/assign4_backtrace_neg.cc:
Likewise.
The -Werror flag should depend on the --enable-werror configure option.
libstdc++-v3/ChangeLog:
* src/libbacktrace/Makefile.am: Remove -Werror.
* src/libbacktrace/Makefile.in: Regenerate.
This removes the 39 lines of shell commands that get echoed when
starting the testsuite. The fact that near the end of that output it
prints `echo "WARNING: could not find \`runtest'" 1>&2; :;` makes it
look like that warning is actually being shown to the user.
Suppress echoing the recipe, so that users only see the actual output
from the testsuite, not the makefile recipe as well.
libstdc++-v3/ChangeLog:
* testsuite/Makefile.am (check-DEJAGNU): Use @ in recipe.
* testsuite/Makefile.in: Regenerate.
For LoongArch, the toplevel library build is always aliased to
one of the multilib variants. This patch installs it with the
actual MULTISUBDIR (instead of ".") so that the headers can be
reached by the compiler.
This patch is an update of
https://gcc.gnu.org/pipermail/gcc-patches/2023-September/629435.html
libstdc++-v3/ChangeLog:
* configure.host: Register t-loongarch in tmake_file.
* config/cpu/loongarch/t-loongarch: New file. Manually refresh
MULTISUBDIR with $(shell $(CXX) --print-multi-directory).
The URI for namespaced docbook-xsl was updated to reflect the current
DocBook upstream at <https://cdn.docbook.org/>.
libstdc++-v3/Changelog:
* acinclude.m4: Update docbook xsl URI.
* configure: Regenerate.
The info manual isn't moved to the expected location after
generation which causes the install rule for it to fail.
libstdc++-v3/Changelog:
* doc/Makefile.am: Fix 'doc-install-info' rule.
Fix typo in commment.
* doc/Makefile.in: Regenerate.
A target selector allows multiple target triplets, it's not necessary to
use the || operator in a logical expression.
libstdc++-v3/ChangeLog:
* testsuite/27_io/filesystem/path/concat/94063.cc: Simplify
dg-do target selector.
The filesystem code was using these functions without checking for their
existence, assuming that any UNIX-like libc with <unistd.h> would always
provide them. That's not true for some newlib targets like arm-eabi.
libstdc++-v3/ChangeLog:
* acinclude.m4 (GLIBCXX_CHECK_FILESYSTEM_DEPS): Check for mkdir,
chmod, chdir, and getcwd.
* config.h.in: Regenerate.
* configure: Regenerate.
* src/c++17/fs_ops.cc (create_dir): Use USE_MKDIR macro.
(fs::current_path): Use USE_GETCWD and USE_CHDIR macros.
(fs::permissions): Use USE_CHMOD macro.
* src/filesystem/ops-common.h [FILESYSTEM_IS_WINDOWS]
(chmod, mkdir, getcwd, chdir): Define new macros.
[FILESYSTEM_IS_WINDOWS] (chmod, mkdir, getcwd, chdir): Use
new macros.
* src/filesystem/ops.cc (create_dir): Use USE_MKDIR macro.
(fs::current_path): Use USE_GETCWD and USE_CHDIR macros.
(fs::permissions): Use USE_CHMOD macro.
These tests do not run by default, because the c++20 effective target
selector isn't matched.
libstdc++-v3/ChangeLog:
* testsuite/23_containers/unordered_map/operations/1.cc: Add
dg-options for C++20 mode.
* testsuite/23_containers/unordered_multimap/operations/1.cc:
Likewise.
* testsuite/23_containers/unordered_multiset/operations/1.cc:
Likewise.
* testsuite/23_containers/unordered_set/operations/1.cc:
Likewise.
* testsuite/std/time/parse.cc: Move dg-options before dg-do.
The tests for the std::ranges access CPOs (ranges::begin etc) use
pathological types with ridiculous overload sets for begin/end/data
members, to exercise all the corner cases in the specification.
Since P2278R4 "cbegin should always return a constant iterator" was
implemented for C++23 mode, some of the range access CPOs now require
the argument to satisfy the range concept, which was not previously
required. The behaviour of the CPO also changes for corner cases where
the type is a range R for which constant_range<R> is satisfied in
addition to constant_range<const R> (meaning there's no need to wrap its
iterators in const_iterator). Adjust the expected results for those
pathological types that changed meaning in C++23, and add some new types
to verify other corner cases.
Some other range adaptor tests fail for C++20 because they assert that
ranges::end and ranges::cend return different types, which is not true
when the type satisfies constant_range.
This fixes the tests to PASS for both C++20 and C++23 (and later).
libstdc++-v3/ChangeLog:
* testsuite/std/ranges/access/cbegin.cc: Adjust for C++23
compatibility.
* testsuite/std/ranges/access/cdata.cc: Likewise.
* testsuite/std/ranges/access/cend.cc: Likewise.
* testsuite/std/ranges/access/crbegin.cc: Likewise.
* testsuite/std/ranges/access/crend.cc: Likewise.
* testsuite/std/ranges/adaptors/take.cc: Likewise.
* testsuite/std/ranges/adaptors/take_while.cc: Likewise.
* testsuite/std/ranges/adaptors/transform.cc: Likewise.
This implements the C++23 change "Poison Pills are Too Toxic". This
makes sense to do unconditionally for C++20, as the corner cases that it
fixes are considered to be defects in the C++20 design (e.g. LWG3480 was
needed to fix directory iterators because of these pills being too
toxic).
libstdc++-v3/ChangeLog:
* include/bits/iterator_concepts.h (__imove::iter_move): Define
poison pill as deleted for consistency.
(__access::begin): Replace with a single declaration.
* include/bits/ranges_base.h (__access::end, __access::rbegin)
(__access::rend, __access::size): Likewise.
* include/bits/version.def (ranges): Update value for C++23.
* include/bits/version.h: Regenerate.
* libsupc++/compare (__compare): Add missing poison pill
overloads.
* testsuite/std/ranges/version_c++23.cc: Adjust expected value
of __cpp_lib_ranges.
* testsuite/std/ranges/access/p2602.cc: New test.
This makes the naming of the CPO types and namespaces simpler and more
consistent. With this change the string "cust" won't appear in
diagnostics related to these CPOs, which avoids confusion about what it
means (customization? customer?). Users don't really need to care that
these are called "customization point objects", so don't show them the
string "cust". Names like "__imove::_IterMove" are preferable to names
like "__cust_imove::_IMove" as the former is more obviously related to
the public API "ranges::iter_move".
Instead of a plethora of inline namespaces for all the different CPO
objects, define them all in an inline namespace called _Cpo (which isn't
shown to users anyway, unlike the types of those objects).
libstdc++-v3/ChangeLog:
* include/bits/iterator_concepts.h (ranges::__cust_imove):
Rename to ranges::__imove.
(_IMove): Rename to _IterMove.
(ranges::__cust_iswap): Rename to ranges::__iswap.
(ranges::__cust): Rename to ranges::_Cpo.
(ranges::__cust_access): Rename to ranges::__access.
* include/bits/ranges_base.h (ranges::__cust_access): Rename to
ranges::__access.
(ranges::__cust): Rename to ranges::_Cpo.
* include/std/concepts (ranges::__cust_swap): Rename to
ranges::__swap.
(ranges::__cust): Rename to ranges::_Cpo.
* libsupc++/compare (__cmp_cust): Rename to __compare.
(__cmp_algo): Rename to _Cpo.
This is a C++23 feature that relaxes the constraints on some range
adaptors, to support move-only types.
libstdc++-v3/ChangeLog:
* include/bits/version.def (ranges): Update value.
* include/bits/version.h: Regenerate.
* include/std/ranges (__detail::__boxable): Use
move_constructible instead of copy_constructible for C++23.
(__detail::__box<T>): Adjust constraints for partial
specialization.
(single_view, transform_view): Use __box_constructible instead
of copy_constructible in constraints.
(zip_transform_view, adjacent_transform_view, repeat_view): Use
move_constructible instead of copy_constructible in constraints.
* testsuite/std/ranges/adaptors/adjacent_transform/1.cc: Check
construction from move-only argument.
* testsuite/std/ranges/adaptors/transform.cc: Likewise.
* testsuite/std/ranges/repeat/1.cc: Likewise.
* testsuite/std/ranges/single_view.cc: Likewise.
* testsuite/std/ranges/zip_transform/1.cc: Likewise.
* testsuite/std/ranges/version_c++23.cc: Adjust expected value
of __cpp_lib_ranges.
As discussed in PR104167 (comments #8 and below), and PR111238, using
-Wl,-gc-sections in the libstdc++ testsuite for arm-eabi
(cross-toolchain) avoids link failures for a few tests:
27_io/filesystem/path/108636.cc
std/time/clock/gps/1.cc
std/time/clock/gps/io.cc
std/time/clock/tai/1.cc
std/time/clock/tai/io.cc
std/time/clock/utc/1.cc
std/time/clock/utc/io.cc
std/time/clock/utc/leap_second_info.cc
std/time/exceptions.cc
std/time/format.cc
std/time/time_zone/get_info_local.cc
std/time/time_zone/get_info_sys.cc
std/time/tzdb/1.cc
std/time/tzdb/leap_seconds.cc
std/time/tzdb_list/1.cc
std/time/zoned_time/1.cc
std/time/zoned_time/custom.cc
std/time/zoned_time/io.cc
std/time/zoned_traits.cc
This patch achieves this by calling GLIBCXX_CHECK_LINKER_FEATURES in
cross-build cases, like we already do for native builds. We keep not
doing so in Canadian-cross builds.
However, this would hide the fact that libstdc++ somehow forces the
user to use -Wl,-gc-sections to avoid undefined references to chdir,
mkdir, chmod, pathconf, ... so maybe it's better to keep the status
quo and not apply this patch?
2023-08-31 Christophe Lyon <christophe.lyon@linaro.org>
libstdc++-v3/ChangeLog:
PR libstdc++/111238
* configure: Regenerate.
* configure.ac: Call GLIBCXX_CHECK_LINKER_FEATURES in cross,
non-Canadian builds.
It's no longer true that 1.0if has type float _Complex when GNU
extensions are enabled, so remove the hardcoded -std option.
libstdc++-v3/ChangeLog:
* testsuite/26_numerics/complex/literals/types.cc: Remove
dg-options and add target selector instead.
These tests need slight adjustments to be valid in C++11 and later, but
there's no reason that can't be done, so that we test them in more
modes.
libstdc++-v3/ChangeLog:
* testsuite/tr1/6_containers/utility/pair.cc: Remove dg-options
and qualify ambiguous calls to get.
* testsuite/tr1/8_c_compatibility/cmath/pow_cmath.cc: Adjust
expected result for std::pow(float, int) as per DR 550.
There is no reason to only test std::auto_ptr with -std=c++03, we just
need to handle the deprecated warnings for C++11 and later.
libstdc++-v3/ChangeLog:
* testsuite/20_util/auto_ptr/1.cc: Remove dg-options -std=c++03
and add dg-warning for deprecation warnings.
* testsuite/20_util/auto_ptr/2.cc: Likewise.
* testsuite/20_util/auto_ptr/3.cc: Likewise.
* testsuite/20_util/auto_ptr/3946.cc: Likewise.
* testsuite/20_util/auto_ptr/4.cc: Likewise.
* testsuite/20_util/auto_ptr/5.cc: Likewise.
* testsuite/20_util/auto_ptr/6.cc: Likewise.
* testsuite/20_util/auto_ptr/7.cc: Likewise.
* testsuite/20_util/auto_ptr/assign_neg.cc: Likewise.
* testsuite/20_util/auto_ptr/requirements/explicit_instantiation/1.cc:
Likewise.
* testsuite/tr1/2_general_utilities/shared_ptr/assign/auto_ptr.cc:
Likewise.
* testsuite/tr1/2_general_utilities/shared_ptr/assign/auto_ptr_neg.cc:
Likewise.
* testsuite/tr1/2_general_utilities/shared_ptr/assign/auto_ptr_rvalue_neg.cc:
Likewise.
* testsuite/tr1/2_general_utilities/shared_ptr/cons/43820_neg.cc:
Likewise.
* testsuite/tr1/2_general_utilities/shared_ptr/cons/auto_ptr.cc:
Likewise.
* testsuite/tr1/2_general_utilities/shared_ptr/cons/auto_ptr_neg.cc:
Likewise.
These tests have transposed digits in the filenames and comments.
libstdc++-v3/ChangeLog:
PR libstdc++/26142
* testsuite/23_containers/vector/26412-1.cc: Moved to...
* testsuite/23_containers/vector/26142-1.cc: ...here.
* testsuite/23_containers/vector/26412-2.cc: Moved to...
* testsuite/23_containers/vector/26142-2.cc: ...here.
These test behaviour only seen with -std=c++03 so the target selector
should match.
libstdc++-v3/ChangeLog:
* testsuite/20_util/bitset/107037.cc: Add c++98_only selector.
* testsuite/26_numerics/complex/56111.cc: Likewise.
libstdc++-v3/ChangeLog:
* testsuite/23_containers/deque/requirements/explicit_instantiation/2.cc:
Add dg-options to restrict the test to C++98 mode.
* testsuite/23_containers/list/requirements/explicit_instantiation/2.cc:
Likewise.
This test should use a target selector of c++23 so that the explicit
-std=gnu++23 option can be removed, to allow testing with later
standards too.
libstdc++-v3/ChangeLog:
* testsuite/20_util/expected/bad.cc: Add missing target
selector.
These tests started failing at some point:
FAIL: 25_algorithms/copy/debug/constexpr_neg.cc (test for errors, line 49)
FAIL: 25_algorithms/copy/debug/constexpr_neg.cc (test for excess errors)
FAIL: 25_algorithms/equal/debug/constexpr_neg.cc (test for errors, line 47)
FAIL: 25_algorithms/equal/debug/constexpr_neg.cc (test for excess errors)
They only run with -D_GLIBCXX_DEBUG or make check-debug so seem to have
gone unnoticed until now.
libstdc++-v3/ChangeLog:
* testsuite/25_algorithms/copy/debug/constexpr_neg.cc: Adjust
expected errors.
* testsuite/25_algorithms/equal/debug/constexpr_neg.cc:
Likewise.
Fix ordering to prevent clobbering of class members by a call to deallocate
in _M_realloc_insert and _M_default_append.
Because of recent changes in _M_realloc_insert and _M_default_append,
calls to deallocate were ordered after assignment to class members of
std::vector (in the guard destructor), which is causing said members to
be call-clobbered. This is preventing further optimization, the
compiler is unable to move memory read out of a hot loop in this case.
This patch reorders the call to before assignments by putting guard in
its own block. Plus a new testsuite for this case. I'm not very happy
with the new testsuite, but I don't know how to properly test this.
PR libstdc++/110879
libstdc++-v3/ChangeLog:
* include/bits/vector.tcc (_M_realloc_insert): End guard
lifetime just before assignment to class members.
(_M_default_append): Likewise.
gcc/testsuite/ChangeLog:
* g++.dg/pr110879.C: New test.
Signed-off-by: Vladimir Palevich <palevichva@gmail.com>
There are a few places in the std::filesystem code that use a string as
a buffer for OS APIs to write to. We can use the new extension
__resize_and_overwrite to avoid redundant initialization of those
buffers.
libstdc++-v3/ChangeLog:
* src/c++17/fs_ops.cc (fs::absolute) [FILESYSTEM_IS_WINDOWS]:
Use __resize_and_overwrite to fill buffer.
(fs::read_symlink) [HAVE_READLINK]: Likewise.
* src/filesystem/ops-common.h (get_temp_directory_from_env)
[FILESYSTEM_IS_WINDOWS]: Likewise.
We need to use a loop in std::atomic_ref::compare_exchange_strong in
order to properly implement the C++20 requirement that padding bits do
not participate when checking the value for equality. The variable being
modified by a std::atomic_ref might have an initial value with non-zero
padding bits, so when the __atomic_compare_exchange built-in returns
false we need to check whether that was only because of non-equal
padding bits that are not part of the value representation. If the value
bits differ, it's just a failed compare-exchange. If the value bits are
the same, we need to retry the __atomic_compare_exchange using the value
that was just read by the previous failed call. As noted in the
comments, it's possible for that second try to also fail due to another
thread storing the same value but with differences in padding.
Because it's undefined to access a variable directly while it's held by
a std::atomic_ref, and because std::atomic_ref will only ever store
values with zeroed padding, we know that padding bits will never go from
zero to non-zero during the lifetime of a std::atomic_ref. They can only
go from an initial non-zero state to zero. This means the loop will
terminate, rather than looping indefinitely as padding bits flicker on
and off. In theory users could call __atomic_store etc. directly and
write a value with non-zero padding bits, but we don't need to support
that. Users doing that should ensure they do not write non-zero padding,
to be compatibile with our std::atomic_ref's invariants.
This isn't a problem for std::atomic<T>::compare_exchange_strong because
the initial value (and all later stores to the variable) are performed
by the library, so we ensure that stored values always have padding bits
cleared. That means we can simply clear the padding bits of the
'expected' value and we will be comparing two values with equal padding
bits. This means we don't need the loop for std::atomic, so update the
__atomic_impl::__compare_exchange function to take a bool parameter that
says whether it's being used by std::atomic_ref. If not, we can use a
simpler, non-looping implementation.
libstdc++-v3/ChangeLog:
PR libstdc++/111077
* include/bits/atomic_base.h (__atomic_impl::__compare_exchange):
Add _AtomicRef non-type template parameter and use a loop if it
is true.
(__atomic_impl::compare_exchange_weak): Add _AtomicRef NTTP.
(__atomic_impl::compare_exchange_strong): Likewise.
(atomic_ref::compare_exchange_weak): Use true for NTTP.
(atomic_ref::compare_exchange_strong): Use true for NTTP.
* testsuite/29_atomics/atomic_ref/compare_exchange_padding.cc:
Fix test to not rely on atomic_ref::load() to return an object
with padding preserved.
This test expects to be able to link, which fails if there are undefined
references to chdir, mkdir etc. in fs_ops.o in the libstdc++.a archive.
libstdc++-v3/ChangeLog:
* testsuite/27_io/filesystem/path/108636.cc: Add dg-require for
filesystem support.
chrono::tzdb::current_zone uses filesystem::read_symlink, which creates
a dependency on the fs_ops.o object in libstdc++.a, which then creates
dependencies on several OS functions if --gc-sections isn't used. For
more details see PR libstdc++/104167 comment 8 and comment 11.
In the cases where that causes linker failures, we probably don't have
readlink anyway, so the filesystem::read_symlink call will always fail.
Repeat the preprocessor conditions for filesystem::read_symlink in the
body of chrono::tzdb::current_zone so that we don't create a
dependency on fs_ops.o for a function that will always fail.
libstdc++-v3/ChangeLog:
* src/c++20/tzdb.cc (tzdb::current_zone): Check configure macros
for POSIX readlink before using filesystem::read_symlink.
This causes libstdc++_libbacktrace.a to be built by default. This might
fail on some targets, in which case we can make the 'auto' choice expand
to either 'yes' or 'no' depending on the target.
libstdc++-v3/ChangeLog:
* acinclude.m4 (GLIBCXX_ENABLE_BACKTRACE): Default to yes.
* configure: Regenerate.
We fail to diagnose an error and extract an incorrect time for cases
like "25:59" >> parse("%H:%M", mins). The bad "25" hour value gets
ignored (on the basis that we might not care about it if trying to
extract something like a weekday or a month name), but then when we get
to the end of the function we think we have a valid time from "59" and
so the result is 00:59.
The problem is that the '__bad_h' value is used for "no hour value read
yet" as well as "bad hour value read". If we just set __h = __bad_h and
continue, we can't tell later that we read an invalid hour.
The fix is to set failbit early when we're trying to extract a
time-of-day (e.g. duration or time_point) and we encounter an invalid
hour, minute, or second value. We can still delay other error checking
to the end.
libstdc++-v3/ChangeLog:
* include/bits/chrono_io.h (_Parser::operator()): Set failbit
early if invalid values are read when _M_need & _TimeOfDay is
non-zero.
* testsuite/std/time/parse.cc: Check that "25:59" cannot be
parsed for "%H:%M".
libstdc++-v3/ChangeLog:
PR libstdc++/111162
* include/bits/chrono_io.h (_Parser::Operator()): Check %C
values are in range of year::min() to year::max().
* testsuite/std/time/parse.cc: Check out of range centuries.
Using an offset as the second argument instead of an iterator makes it
easier for callers, as they don't need to create an lvalue span in order
to get an iterator from it for the _M_reset call.
libstdc++-v3/ChangeLog:
* include/std/format (__format::_Sink::_M_reset): Change second
argument from iterator to offset.
When parsing a format string, the width is parsed into an unsigned short
but the result is not checked in the case the format string is not a
char string (such as a wide string). In case the parse fails, a null
pointer is returned which is used for pointer arithmetic which is
undefined behaviour.
Signed-off-by: Paul Dreik <gccpatches@pauldreik.se>
libstdc++-v3/ChangeLog:
PR libstdc++/111102
* include/std/format (__format::__parse_integer): Check for
non-null pointer.
Update a preprocessor condition using __cplusplus and _GLIBCXX_HOSTED
to use the relevant feature test macro for <syncstream>.
Also add comments to some conditions saying which C++ standard revision
the check corresponds to.
libstdc++-v3/ChangeLog:
* include/std/atomic: Add comment to #ifdef and fix indentation.
* include/std/ostream: Check __glibcxx_syncbuf instead of
__cplusplus and _GLIBCXX_HOSTED.
* include/std/thread: Add comment to #ifdef.
