Advance Toolchain 14.0-2
Advance Toolchain new location
Since the end of August 2019, Advance Toolchain has been available from a new FTP server: public.dhe.ibm.com
Please be advised the FTP server at Unicamp won't be updated any longer.
New features in 14.0-2
- GCC provides:
- Fix for Preserve variable-use locations.
- Fix for [C++14] <unresolved overloaded function type> for function template with auto return.
- Fix for vtable not emitted for class with no key method and defaulted destructor with attribute((used)).
- Fix for relative line numbers only working if gcc_{error,warning}_prefix defined.
- Fix for [testsuite, libgomp] dg-message not supported.
- Fix for ICE: Floating point exception.
- Fix for Incorrect result from alignof in templates, if also using __alignof__..
- Fix for ice on aggregate initialization of unmovable base.
- Fix for constraints not checked on nested class template.
- Fix for Layout of istreambuf_iterator subobject depends on -std mode.
- Fix for Large increase in preprocessed file sizes in C++2a mode.
- Fix for Compiler is unable to generate finalization wrapper.
- Fix for Fortran Location Data for Diagnostic lacks the column number – when passing on to ME.
- Fix for [C++20] copy deduction rejected when doing CTAD for NTTP.
- Fix for No overflow check in __atomic_futex_unsigned_base::_M_futex_wait_until.
- Fix for ICE in gfc_constructor_append_expr, at fortran/constructor.c:135.
- Fix for ICE: [[no_unique_address] when constructing via template helper.
- Fix for Concept checked after auto return type deduction.
- Fix for ICE: verify_flow_info failed (error: wrong outgoing edge flags at end of bb 2).
- Fix for memcmp being wrongly stripped like strcmp.
- Fix for std::list | take | transform, expression does not work cbegin() == end().
- Fix for ICE in gfc_match_subroutine, at fortran/decl.c:7913.
- Fix for ICE in build_field, at fortran/trans-common.c:301.
- Fix for ICE in trunc_int_for_mode, at explow.c:59 since r9-7156-g33579b59aaf02eb7.
- Fix for std::ranges::construct_at's placement new not intercepted.
- Fix for suboptimal memcpy with embedded zero bytes.
- Fix for ICE in get_kind, at fortran/simplify.c:129.
- Fix for Segmentation fault compiling with static libraries and using jthread::request_stop.
- Fix for Reference type of std::ranges::iota is __int128 with -std=c++2a?!.
- Fix for bogus alignment for negative step grouped access.
- Fix for ICE in gimplify_expr, at gimplify.c:14638.
- Fix for ICE in fold_convert_loc, at fold-const.c:2398.
- Fix for internal compiler error: in tsubst_copy with CTAD for alias templates.
- Fix for Invalid specialization accepted when also constrained in base template template parameter.
- Fix for ICE in verify_ctor_sanity.
- Fix for optional comparison with nullopt fails.
- Fix for gfortran 8-10 shape mismatch in assumed-length dummy argument character array.
- Fix for ICE in add_new_edges_to_heap, at ipa-inline.c:1746 (-O3 PGO).
- Fix for gnat.dg/opt39.adb is not scalarized.
- Fix for ICE during GIMPLE pass:vect.
- Fix for ICE: Segmentation fault in instantiate_template / pop_nested_class().
- Fix for gcov TOPN streaming is incorrect for shared libraries.
- Fix for ICE segmentation fault in tree-vectorizer at -O3.
- Fix for Switch with case values derived from constexpr function takes unreasonable time to compile.
- Fix for Missing code from consteval constructor initializing const variable.
- Fix for C++20 ADL and function templates that are not visible (P0846R0) fails on call with templated type.
- Fix for Missing DW_AT_declaration=1 in dwarf data.
- Fix for [MATMUL intrinsic] The value of result is wrong when vector (step size is negative) * matrix.
- Fix for Zero valued #line directive results in excessively large blocks of memory being allocated.
- Fix for assume_aligned is not constexpr.
- Fix for Sanitizer pointer-subtract breaks constexpr functions subtracting pointers.
- Fix for Build error with -mcpu=power9 on ppc64.
- Fix for zstd build failure for gcc 10 on Ubuntu 16.04.
- Fix for .gcno files are not written to same directory as the object file.
- Fix for construct_at on a union member is not a constant expression.
- Fix for With -O2, Incorrect -Werror=maybe-uninitialized thrown, leads to 'target_mem_ref' and 'dump_expr<expression error>' in message.
- Fix for Vectorizer gives a boolean a value of 255.
- Fix for Wrong answer from MAXLOC with character arg.
- Fix for Segmentation fault in lto-wrapper.
- Fix for ASAN "dynamic-stack-buffer-overflow" false positive with OpenMP reduction to std::vector.
- Fix for ICE on firefox built with lto+pgo: dist/include/mozilla/Casting.h:64:1: internal compiler error: in to_frequency, at profile-count.c:273.
- Fix for ICE in verify_ctor_sanity, at cp/constexpr.c:3995 since r10-7313-gb599bf9d6d1e180d.
- Fix for Unable to coalesce ssa_names which are marked as MUST COALESCE..
- Fix for ICE in range_on_exit.
- Fix for `__deref` in <functional> in debug mode clashes with internal macro in Windows system header.
- Fix for wrong code with __builtin_bswap16() of rotated 64bit value.
- Fix for g++.dg/asan/asan_test.C compilation failure starting with r11-3827.
- Fix for [concepts] ICE with requires requires and parameter packs.
- Fix for AddressSanitizer CHECK failed: detect_stack_use_after_return and detect_invalid_pointer_pairs.
- Fix for [accepts-invalid] coroutines accepts prmomise type with both return_value() and return_void().
- Fix for Wrong min value generated for DFP numbers.
- Fix for An incorrect optimization causes a function to always return the same value when using -flto.
- Fix for Wrong restriction for VALUE arguments of pure procedures.
- Fix for ICE: decompressed stream: Destination buffer is too small.
- Fix for ICE verify_ssa failed since r10-4200-gb7ff7cef5005721e.
- Fix for ICE: during RTL pass: cprop /segfault in sbitmap.
- Fix for New gfortran.dg/pdt_14.f03 failure after g:617695cdc2b3d950f1e4deb5ea85d5cc302943f4.
- Fix for [c++17] Function with return type 'unsigned' in nested namespace misinterpreted as deduction guide.
- Fix for terminate called in std::experimental::filesystem::recursive_directory_iterator.
- Fix for Preincrement of _Complex gives bogus warning = "value computed is not used".
- Fix for GCC outputs wrong values when compiling the testcase with -O3 since r10-4200-gb7ff7cef5005721e.
- Fix for 32-bit f951 ICE on code from OpenMolcas.
- Fix for [Fortran] Confused location information for OpenACC compute constructs.
- Fix for constexpr evaluation reports false positive memory leak.
- Fix for internal compiler error: in build2, at tree.c:4814.
- Fix for std::ranges::search_n does not work with counted_iterator<_List_iterator<...>>.
- Fix for ice compiling dxml.
- Fix for Bad code gen when concatenating to array.
- Fix for <version> defines __cpp_lib_span even when <span> doesn't provide an implementation.
- Fix for stop_token header doesn't compile on clang-8 with -std=c++20.
- Fix for d: OutOfMemoryError thrown when appending to an array with a side effect.
- Fix for ICE near htab_hash_string when LTO, -O & -g.
- Fix for Missing subsumption in iterator category detection.
- Fix for ICE in build2, at tree.c:4868.
- Fix for ext/stdio_filebuf/char/79820.cc is broken.
- Fix for gcc.dg/strncmp-2.c frees mproptected memory.
- Fix for incorrect Type'Mod during a loop whose range is computed by a variable.
- Fix for Dependency check fails when using "allocatable" instead of "pointer" (forall_3.f90).
- GNU C Library provides:
- Fix for (CVE-2016-10228) iconv(1) with -c option hangs on illegal multi-byte sequences (CVE-2016-10228).
- Fix for (CVE-2019-25013) iconv encounters segmentation fault when converting 0x00 0xfe in EUC-KR to UTF-8 (CVE-2019-25013).
- Fix for undefined reference to `__warn_memset_zero_len' when changing gnuc version.
- Fix for (CVE-2020-27618) iconv hangs when converting some invalid inputs from several IBM character sets (CVE-2020-27618).
- Fix for bind_textdomain_codeset doesn't accept //TRANSLIT anymore.
- Fix for Transaction ID collisions cause slow DNS lookups in getaddrinfo.
- Fix for 32-bit shmctl(IPC_INFO) crashes when shminfo struct is at the end of a memory mapping.
- Fix for semctl SEM_STAT_ANY fails to pass the buffer specified by the caller to the kernel.
- Fix for msgctl IPC_INFO and MSG_INFO return garbage.
- Fix for glibc 2.32 has default_cflags="-g -O" instead of "-g -O2".
- Binutils provides:
- Fix for Error: file table slot 1 is already occupied by a different file.
- Fix for nm -D -u now by default displays symbol versions, adding --with-symbol-versions duplicates them.
- Fix for Inconsistent undefined versioned symbols.
- Fix for -z ibt and -z shstk are mishandled.
- Fix for Extra filename generated by --gdwarf-4.
- Fix for Error: file table slot 1 is already occupied by a different file.
- Fix for [2.36 Regression] readelf: Warning: DIE at offset 0x232 refers to abbreviation number 77 which does not exist.
- Fix for ld: assertion fail ../../bfd/elflink.c:14806 on i686-linux-gnu since 382aae06322799a25ea52fe61b243cbca4db8d66.
- Fix for gold powerpc inserts jump to middle of long branch stub.
- Fix for [readelf] crash with ASAN in print_dynamic_symbol.
- Fix for nm -P portable output format regression.
- Fix for ppc32 segmentation fault in make_stub.
- OpenSSL provides:
- Version 1.1.1i, see the release notes for details.
- Fix for OPENSSL_cpuid_setup FreeBSD PowerPC.
- Fix for [crypto/dh] side channel hardening for computing DH shared keys .
- Fix for NULL pointer access caused by X509_ATTRIBUTE_create().
- Fix for Broken on 32 bit big endian.
- Fix for SSL_CTX_sess_accept gives negative value.
- Fix for nginx 1.19.4 feature "ssl_reject_handshake" does not work as intended with openssl.
- Fix for Low Memory SegFault in dtls.
- Fix for X509_cmp fails since 1.1.1f.
- Fix for Regression in `X509_verify_cert`.
- Fix for UTF8 BOM should be ignored when reading the configuration file.
- Expat provides:
- Version 2.2.10, see the release notes for details.
- Fix for Make testrunner output more helpful (related to #382) by hartwork.
- Fix for Detect unsupported VS at configure time (and not at compile time) by tbeu.
- Fix for Obsolescent macro AC_HEADER_STDC .
- Fix for Protect installed expat_config.h from regression + drop obsolescent macro AC_HEADER_STDC (#436) by hartwork.
- Python provides:
- Version 3.8.7, see the release notes for details.
- Fix for strange interaction between __slots__ and class-level attributes.
- Fix for fixcid.py is using instead of !=.
- Fix for Change the text signature format (again) to be more robust.
- Fix for statistics._decimal_to_ratio() produces non-integer ratio.
- Fix for argparse: Provide equivalent of optparse.OptionParser.{option_groups,option_list,get_option}.
- Fix for Access violation - pyc file.
- Fix for Use PyMem_New instead of PyMem_Malloc.
- Fix for Improve byte formatting compatibility between Py2 and Py3.
- Fix for Add tests for wsgiref.validate.
- Fix for multiprocessing: string arg to SystemExit.
- Fix for timeit accuracy could be better.
- Fix for expandvars removes single quotes ( ' ).
- Fix for Reloading tokenize module leads to error.
- Fix for sum() function docstring lists arguments incorrectly.
- Fix for Make stdprinter use DebugOutputString when no stdout/stderr available.
- Fix for Encrypted MSI fails to install with code 2755.
- Fix for Argument Clinic: wrap impl's declaration if it's too long.
- Fix for Add python.ini file for embedded/applocal installs.
- Fix for Incorrect TimeoutError referenced in concurrent.futures documentation.
- Fix for test_ssl failure on Fedora 22.
- Fix for More clearly expose/explain native and cross-build target information.
- Fix for rename the platform directory from plat-$(MACHDEP) to plat-$(PLATFORM_TRIPLET).
- Fix for PEP 484 implementation.
- Fix for Undefined behavior in dtoa.c (rshift 32 of 32bit data type).
- Fix for Get rid of rare format units in PyArg_Parse*.
- Fix for Add a Sprints organization/preparation section to devguide.
- Fix for Add docstring to urllib.urlretrieve.
- Fix for Python heap corruption issue.
- Fix for str(bytes_obj) should raise an error.
- Fix for Python hangs forever in wait() of threading.py.
- Fix for GB2312 codec is using a wrong covert table.
- Fix for Argument Clinic: add the boolint converter.
- Fix for Clearify necessities for logging with timestamps.
- Fix for sum() several times slower on Python 3 64-bit.
- Fix for man page says -I implies -S. code says -s..
- Fix for inspect.getsourcelines ignores context and returns wrong line #.
- Fix for asyncio.Event.wait() Task was destroyed but it is pending.
- Fix for Obsolete note in argument parsing (c-api/arg.rst).
- Fix for yield expression confusion.
- Fix for argparse crashes with AssertionError.
- Fix for Use after free in Element.remove.
- Fix for Use after free in siftdown (2).
- Fix for Use after free in siftup.
- Fix for zipfile.ZipFile.write() does not accept bytes arcname.
- Fix for shlex constructor unreachable code.
- Fix for --with-pydebug has no effect when the final python binary is compiled.
- Fix for pathlib.(r)glob stops on PermissionDenied exception.
- Fix for Incorrect (misleading) statement in the execution model documentation.
- Fix for ABCMeta classes do not support the **kwargs standard class interface.
- Userspace RCU provides:
- Fix for configure: support Autoconf 2.70.
- Fix for bump tests thread limit to 4096.
- Fix for explicitly include urcu/config.h in files using CONFIG_RCU_ defines.
- Fix for add -lurcu-common to pkg-config libs for each flavor.
- Fix for race between rcu_barrier() and call_rcu_data_free().
Packages in 14.0-2
- The Advance Toolchain is a self contained toolchain which isn't reliant on the base system toolchain.
- Base toolchain packages:
- GNU Binutils - 2.35.1-4100e6999664 online docs
- Expat XML Parser - 2.2.10-aafd462fbdb9 online docs
- GCC (GNU Compiler Collection) - 10.2.1-2093e873bb6c [C, C++ (g++), fortran, Go] online docs
- GNU Standard C++ Library v3 (Libstdc++-v3) - 10.2.1-2093e873bb6c online docs
- GNU Libgomp - 10.2.1-2093e873bb6c online docs
- GNU C Library - 2.32-c542253f49f7 online docs
- GNU Multiple Precision Library - 6.1.2 online docs
- GNU MPC Library - 1.1.0-69ad1bcb8e9e online docs
- GNU MPFR Library - 4.0.2-13499 online docs
- Additional support libraries, including some additional floating point support libraries:
- Decimal Floating Point C Library - 1.0.15-1031321fa9d2 online docs
- Libhugetlbfs - 2.23-6b126a4d7da9 online docs
- Libvecpf - 1.1.0 online docs
- Open Secure Sockets Layer - 1.1.1i-76ed0c0ad119 online docs
- IBM Power Architecture Facilities Library - 0.3.0-cbb41aa0538a online docs
- Power Vector Library - 1.0.3-76ca105829a3 online docs
- Python - 3.8.7-187f76def8a5 online docs
- TCMalloc - 2.8-bda3c82e1161 online docs
- ZLIB Compression Library - 1.2.11-ba9df2111e9c online docs
- Development tools:
- GDB (GNU Debugger) - 9.2-9234a6aba1b6 online docs
- Additional performance and debugging tools:
- Valgrind - 3.16.1-24f63fd435e7 online docs
- Additional multi core support libraries:
- Boost - 1.73.0 online docs
- SPHDE - 1.4.0-6a630fad0fbe online docs
- Userspace RCU - 0.12.1-a2fd3e026bb9 online docs
- Thread Building Blocks - 2020.3-eca91f16d749 online docs