| Message ID | ydda6so567m.fsf@CeBiTec.Uni-Bielefeld.DE |
|---|---|
| State | New |
| Headers | show |
| Series | libstdc++: Don't use reserved identifiers in simd headers | expand |
On Montag, 1. Februar 2021 13:21:33 CET Rainer Orth wrote: > Two simd tests FAIL on Solaris, both SPARC and x86: > > FAIL: experimental/simd/standard_abi_usable.cc -msse2 -O2 -Wno-psabi (test > for excess errors) FAIL: experimental/simd/standard_abi_usable_2.cc -msse2 > -O2 -Wno-psabi (test for excess errors) > > This happens because the simd headers use identifiers documented in the > libstdc++ manual as reserved by system headers. Sorry, this code was originally written as non-stdlib code, i.e. without any reserved identifiers. I had hoped I found all issues... > Fixed as follows, tested on i386-pc-solaris2.11, sparc-sun-solaris2.11, > and x86_64-pc-linux-gnu. > > Ok for master? Looks good to me. > As an aside, the use of vim: markers initially confused the hell out of > me. As an Emacs user, I rarely use vi for much more than a pager, but > when I wanted to check the lines mentioned in the g++ errors, I had no > idea what was going on or how to disable the folding enabled there: > > // vim: foldmethod=marker sw=2 noet ts=8 sts=2 tw=80 > > I can't help but feel that this is just a personal preference and > doesn't belong into the upstream code. Yes. I guess it's better to remove at least foldmethod. The rest isn't personal preference, but coding style requirements. However, I don't need any of it anymore: by now my vim config autodetects GCC / libstdc++ code. If the rest of libstdc++ doesn't have it, the simd headers probably shouldn't have it either. Best, Matthias
On 01/02/21 13:21 +0100, Rainer Orth wrote: >Two simd tests FAIL on Solaris, both SPARC and x86: > >FAIL: experimental/simd/standard_abi_usable.cc -msse2 -O2 -Wno-psabi (test for excess errors) >FAIL: experimental/simd/standard_abi_usable_2.cc -msse2 -O2 -Wno-psabi (test for excess errors) > >This happens because the simd headers use identifiers documented in the >libstdc++ manual as reserved by system headers. > >Fixed as follows, tested on i386-pc-solaris2.11, sparc-sun-solaris2.11, >and x86_64-pc-linux-gnu. > >Ok for master? OK, thanks. >As an aside, the use of vim: markers initially confused the hell out of >me. As an Emacs user, I rarely use vi for much more than a pager, but >when I wanted to check the lines mentioned in the g++ errors, I had no >idea what was going on or how to disable the folding enabled there: > >// vim: foldmethod=marker sw=2 noet ts=8 sts=2 tw=80 > >I can't help but feel that this is just a personal preference and >doesn't belong into the upstream code. > >For avoidance of doubt, I'd consider equivalent Emacs local variables >equally inappropriate. We do have Emacs mode lines in all the libstdc++ headers FWIW. I use a modified copy of https://www.vim.org/scripts/script.php?script_id=3381 to parse those modelines and set Vim config like sw=2 noet ts=2 etc. when it sees them. So I don't need vim: lines, because the Emacs ones work for me. I agree with removing the vim: lines if Matthias is happy to. The foldmethod is definitely awkward if that's not your preference.
# HG changeset patch # Parent 585afa0bd4a3567358ed9efd210d1a1dab740005 libstdc++: Don't use reserved identifiers in simd headers diff --git a/libstdc++-v3/include/experimental/bits/simd.h b/libstdc++-v3/include/experimental/bits/simd.h --- a/libstdc++-v3/include/experimental/bits/simd.h +++ b/libstdc++-v3/include/experimental/bits/simd.h @@ -201,8 +201,8 @@ template <size_t _Np> inline constexpr overaligned_tag<_Np> overaligned = {}; // }}} -template <size_t _X> - using _SizeConstant = integral_constant<size_t, _X>; +template <size_t _Xp> + using _SizeConstant = integral_constant<size_t, _Xp>; // unrolled/pack execution helpers // __execute_n_times{{{ @@ -4060,11 +4060,11 @@ template <template <int> class _A0, temp return typename __decay_abi<_A0<_Bytes>>::type{}; else { - using _B = + using _Bp = typename __find_next_valid_abi<_A0, _Bytes, _Tp>::type; - if constexpr (_B::template _S_is_valid_v< - _Tp> && _B::template _S_size<_Tp> <= _Np) - return _B{}; + if constexpr (_Bp::template _S_is_valid_v< + _Tp> && _Bp::template _S_size<_Tp> <= _Np) + return _Bp{}; else return typename _AbiList<_Rest...>::template _BestAbi<_Tp, _Np>{}; diff --git a/libstdc++-v3/include/experimental/bits/simd_builtin.h b/libstdc++-v3/include/experimental/bits/simd_builtin.h --- a/libstdc++-v3/include/experimental/bits/simd_builtin.h +++ b/libstdc++-v3/include/experimental/bits/simd_builtin.h @@ -894,12 +894,12 @@ template <typename _Tp, typename _Mp, ty class _SimdCastType2 { using _Ap = __intrinsic_type_t<_Tp, _Np>; - using _B = __vector_type_t<_Tp, _Np>; + using _Bp = __vector_type_t<_Tp, _Np>; _SimdMember _M_data; public: _SimdCastType2(_Ap __a) : _M_data(__vector_bitcast<_Tp>(__a)) {} - _SimdCastType2(_B __b) : _M_data(__b) {} + _SimdCastType2(_Bp __b) : _M_data(__b) {} operator _SimdMember() const { return _M_data; } }; diff --git a/libstdc++-v3/include/experimental/bits/simd_x86.h b/libstdc++-v3/include/experimental/bits/simd_x86.h --- a/libstdc++-v3/include/experimental/bits/simd_x86.h +++ b/libstdc++-v3/include/experimental/bits/simd_x86.h @@ -55,10 +55,10 @@ template <typename _TV, // }}} // __interleave128_lo {{{ -template <typename _Ap, typename _B, typename _Tp = common_type_t<_Ap, _B>, +template <typename _Ap, typename _Bp, typename _Tp = common_type_t<_Ap, _Bp>, typename _Trait = _VectorTraits<_Tp>> _GLIBCXX_SIMD_INTRINSIC constexpr _Tp - __interleave128_lo(const _Ap& __av, const _B& __bv) + __interleave128_lo(const _Ap& __av, const _Bp& __bv) { const _Tp __a(__av); const _Tp __b(__bv);