From: Jonathan Wakely <jwakely@redhat.com>
To: Jakub Jelinek <jakub@redhat.com>
Cc: libstdc++@gcc.gnu.org, gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] libstdc++: Fix up some <cmath> templates [PR109883]
Date: Wed, 17 May 2023 20:15:07 +0100 [thread overview]
Message-ID: <CACb0b4=62XY00SCpf5FPSR=1H7HxQaHbadnCcbm9iLOZcZ6ObA@mail.gmail.com> (raw)
In-Reply-To: <ZGUaWQkrIkrtlbPa@tucnak>
[-- Attachment #1: Type: text/plain, Size: 18332 bytes --]
On Wed, 17 May 2023 at 19:18, Jakub Jelinek <jakub@redhat.com> wrote:
> Hi!
>
> As can be seen on the following testcase, for
>
> std::{atan2,fmod,pow,copysign,fdim,fmax,fmin,hypot,nextafter,remainder,remquo,fma}
> if one operand type is std::float{16,32,64,128}_t or std::bfloat16_t and
> another one some integral type or some other floating point type which
> promotes to the other operand's type, we can end up with endless recursion.
> This is because of a declaration ordering problem in <cmath>, where the
> float, double and long double overloads of those functions come before
> the templates which use __gnu_cxx::__promote_{2,3}, but the
> std::float{16,32,64,128}_t and std::bfloat16_t overloads come later in the
> file. If the result of those promotions is _Float{16,32,64,128} or
> __gnu_cxx::__bfloat16_t, say std::pow(_Float64, int) calls
> std::pow(_Float64, _Float64) and the latter calls itself.
>
> The following patch fixes that by moving those templates later in the file,
> so that the calls from those templates see also the other overloads.
>
I checked that each set of moved functions is still within the same #if
group, so there's no change in the logic for whether they are defined or
not, only the point of declaration within the file.
I think other templates in the file like e.g. isgreater etc. shouldn't be
> a problem, because those just use __builtin_isgreater etc. in their bodies.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk/13.2?
>
OK for both, thanks.
>
> 2023-05-17 Jakub Jelinek <jakub@redhat.com>
>
> PR libstdc++/109883
> * include/c_global/cmath (atan2, fmod, pow): Move
> __gnu_cxx::__promote_2 using templates after _Float{16,32,64,128}
> and
> __gnu_cxx::__bfloat16_t overloads.
> (copysign, fdim, fmax, fmin, hypot, nextafter, remainder, remquo):
> Likewise.
> (fma): Move __gnu_cxx::__promote_3 using template after
> _Float{16,32,64,128} and __gnu_cxx::__bfloat16_t overloads.
>
> * testsuite/26_numerics/headers/cmath/constexpr_std_c++23.cc: New
> test.
>
> --- libstdc++-v3/include/c_global/cmath.jj 2023-01-16
> 23:19:06.225717615 +0100
> +++ libstdc++-v3/include/c_global/cmath 2023-05-17 15:07:07.823657320 +0200
> @@ -151,15 +151,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_atan2l(__y, __x); }
> #endif
>
> - template<typename _Tp, typename _Up>
> - inline _GLIBCXX_CONSTEXPR
> - typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - atan2(_Tp __y, _Up __x)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return atan2(__type(__y), __type(__x));
> - }
> -
> using ::ceil;
>
> #ifndef __CORRECT_ISO_CPP_MATH_H_PROTO
> @@ -286,15 +277,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_fmodl(__x, __y); }
> #endif
>
> - template<typename _Tp, typename _Up>
> - inline _GLIBCXX_CONSTEXPR
> - typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - fmod(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return fmod(__type(__x), __type(__y));
> - }
> -
> using ::frexp;
>
> #ifndef __CORRECT_ISO_CPP_MATH_H_PROTO
> @@ -411,15 +393,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> #endif
> #endif
>
> - template<typename _Tp, typename _Up>
> - inline _GLIBCXX_CONSTEXPR
> - typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - pow(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return pow(__type(__x), __type(__y));
> - }
> -
> using ::sin;
>
> #ifndef __CORRECT_ISO_CPP_MATH_H_PROTO
> @@ -1073,6 +1046,33 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __gnu_cxx::__bfloat16_t(__builtin_tanhf(__x)); }
> #endif
>
> + template<typename _Tp, typename _Up>
> + inline _GLIBCXX_CONSTEXPR
> + typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + atan2(_Tp __y, _Up __x)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return atan2(__type(__y), __type(__x));
> + }
> +
> + template<typename _Tp, typename _Up>
> + inline _GLIBCXX_CONSTEXPR
> + typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + fmod(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return fmod(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + inline _GLIBCXX_CONSTEXPR
> + typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + pow(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return pow(__type(__x), __type(__y));
> + }
> +
> #if _GLIBCXX_USE_C99_MATH
> #if !_GLIBCXX_USE_C99_FP_MACROS_DYNAMIC
>
> @@ -2107,16 +2107,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_copysignl(__x, __y); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - copysign(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return copysign(__type(__x), __type(__y));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr float
> erf(float __x)
> @@ -2199,16 +2189,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_fdiml(__x, __y); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - fdim(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return fdim(__type(__x), __type(__y));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr float
> fma(float __x, float __y, float __z)
> @@ -2219,16 +2199,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_fmal(__x, __y, __z); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up, typename _Vp>
> - constexpr typename __gnu_cxx::__promote_3<_Tp, _Up, _Vp>::__type
> - fma(_Tp __x, _Up __y, _Vp __z)
> - {
> - typedef typename __gnu_cxx::__promote_3<_Tp, _Up, _Vp>::__type
> __type;
> - return fma(__type(__x), __type(__y), __type(__z));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr float
> fmax(float __x, float __y)
> @@ -2239,16 +2209,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_fmaxl(__x, __y); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - fmax(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return fmax(__type(__x), __type(__y));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr float
> fmin(float __x, float __y)
> @@ -2259,16 +2219,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_fminl(__x, __y); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - fmin(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return fmin(__type(__x), __type(__y));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr float
> hypot(float __x, float __y)
> @@ -2279,16 +2229,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_hypotl(__x, __y); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - hypot(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return hypot(__type(__x), __type(__y));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr int
> ilogb(float __x)
> @@ -2481,16 +2421,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_nextafterl(__x, __y); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - nextafter(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return nextafter(__type(__x), __type(__y));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr float
> nexttoward(float __x, long double __y)
> @@ -2519,16 +2449,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_remainderl(__x, __y); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - remainder(_Tp __x, _Up __y)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return remainder(__type(__x), __type(__y));
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> inline float
> remquo(float __x, float __y, int* __pquo)
> @@ -2539,16 +2459,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __builtin_remquol(__x, __y, __pquo); }
> #endif
>
> -#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> - template<typename _Tp, typename _Up>
> - inline typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> - remquo(_Tp __x, _Up __y, int* __pquo)
> - {
> - typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> - return remquo(__type(__x), __type(__y), __pquo);
> - }
> -#endif
> -
> #ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_FP
> constexpr float
> rint(float __x)
> @@ -3555,6 +3465,79 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> { return __gnu_cxx::__bfloat16_t(__builtin_truncf(__x)); }
> #endif
>
> +#ifndef __CORRECT_ISO_CPP11_MATH_H_PROTO_INT
> + template<typename _Tp, typename _Up>
> + constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + copysign(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return copysign(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + fdim(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return fdim(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + fmax(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return fmax(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + fmin(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return fmin(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + hypot(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return hypot(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + nextafter(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return nextafter(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + constexpr typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + remainder(_Tp __x, _Up __y)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return remainder(__type(__x), __type(__y));
> + }
> +
> + template<typename _Tp, typename _Up>
> + inline typename __gnu_cxx::__promote_2<_Tp, _Up>::__type
> + remquo(_Tp __x, _Up __y, int* __pquo)
> + {
> + typedef typename __gnu_cxx::__promote_2<_Tp, _Up>::__type __type;
> + return remquo(__type(__x), __type(__y), __pquo);
> + }
> +
> + template<typename _Tp, typename _Up, typename _Vp>
> + constexpr typename __gnu_cxx::__promote_3<_Tp, _Up, _Vp>::__type
> + fma(_Tp __x, _Up __y, _Vp __z)
> + {
> + typedef typename __gnu_cxx::__promote_3<_Tp, _Up, _Vp>::__type
> __type;
> + return fma(__type(__x), __type(__y), __type(__z));
> + }
> +#endif
>
> #endif // _GLIBCXX_USE_C99_MATH_TR1
> #endif // C++11
> ---
> libstdc++-v3/testsuite/26_numerics/headers/cmath/constexpr_std_c++23.cc.jj
> 2023-05-17 15:23:55.120327327 +0200
> +++
> libstdc++-v3/testsuite/26_numerics/headers/cmath/constexpr_std_c++23.cc
> 2023-05-17 16:19:27.386931508 +0200
> @@ -0,0 +1,129 @@
> +// Copyright (C) 2023 Free Software Foundation, Inc.
> +//
> +// This file is part of the GNU ISO C++ Library. This library is free
> +// software; you can redistribute it and/or modify it under the
> +// terms of the GNU General Public License as published by the
> +// Free Software Foundation; either version 3, or (at your option)
> +// any later version.
> +
> +// This library is distributed in the hope that it will be useful,
> +// but WITHOUT ANY WARRANTY; without even the implied warranty of
> +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> +// GNU General Public License for more details.
> +
> +// You should have received a copy of the GNU General Public License along
> +// with this library; see the file COPYING3. If not see
> +// <http://www.gnu.org/licenses/>.
> +
> +// { dg-options "-std=gnu++2b" }
> +// { dg-do link { target c++23 } }
> +
> +#include <stdfloat>
> +#include <cmath>
> +
> +template <typename T>
> +void
> +test_functions()
> +{
> + constexpr T zero = 0;
> + constexpr T one = 1;
> + constexpr auto a = std::acos(one);
> + constexpr auto b = std::asin(zero);
> + constexpr auto c = std::atan(zero);
> + constexpr auto d = std::cos(zero);
> + constexpr auto e = std::sin(zero);
> + constexpr auto f = std::tan(zero);
> + constexpr auto g = std::acosh(one);
> + constexpr auto h = std::asinh(zero);
> + constexpr auto i = std::atanh(zero);
> + constexpr auto j = std::cosh(zero);
> + constexpr auto k = std::sinh(zero);
> + constexpr auto l = std::tanh(zero);
> + constexpr auto m = std::exp(zero);
> + constexpr auto n = std::exp2(zero);
> + constexpr auto o = std::expm1(one);
> + constexpr auto p = std::log(one);
> + constexpr auto q = std::log10(one);
> + constexpr auto r = std::log1p(zero);
> + constexpr auto s = std::log2(one);
> + constexpr auto t = std::logb(one);
> + constexpr auto u = std::cbrt(zero);
> + constexpr auto v = std::fabs(zero);
> + constexpr auto w = std::sqrt(one);
> + constexpr auto x = std::erf(zero);
> + constexpr auto y = std::erfc(zero);
> +// constexpr auto z = std::lgamma(one);
> + constexpr auto A = std::tgamma(one);
> + constexpr auto B = std::ceil(zero);
> + constexpr auto C = std::floor(zero);
> +// constexpr auto D = std::nearbyint(zero);
> +// constexpr auto E = std::rint(zero);
> + constexpr auto F = std::round(zero);
> + constexpr auto G = std::trunc(zero);
> + constexpr auto H = std::atan2(zero, one);
> + constexpr auto I = std::hypot(one, zero);
> + constexpr auto J = std::pow(one, zero);
> + constexpr auto K = std::fmod(zero, one);
> + constexpr auto L = std::remainder(one, one);
> + constexpr auto M = std::copysign(zero, zero);
> + constexpr auto N = std::nextafter(zero, zero);
> + constexpr auto O = std::fdim(zero, zero);
> + constexpr auto P = std::fmax(zero, one);
> + constexpr auto Q = std::fmin(zero, one);
> + constexpr auto R = std::ilogb(one);
> + constexpr auto S = std::ldexp(one, 0);
> + constexpr auto U = std::scalbn(one, 1);
> + constexpr auto V = std::scalbln(one, 1);
> +// constexpr auto W = std::lrint(one);
> +// constexpr auto X = std::llrint(one);
> + constexpr auto Y = std::lround(one);
> + constexpr auto Z = std::llround(one);
> + constexpr auto a1 = std::fma(one, one, one);
> + constexpr auto b1 = std::atan2(zero, 1);
> + constexpr auto c1 = std::hypot(one, 0);
> + constexpr auto d1 = std::pow(one, 0);
> + constexpr auto e1 = std::fmod(zero, 1);
> + constexpr auto f1 = std::remainder(one, 1);
> + constexpr auto g1 = std::copysign(zero, 0);
> + constexpr auto h1 = std::nextafter(zero, 0);
> + constexpr auto i1 = std::fdim(zero, 0);
> + constexpr auto j1 = std::fmax(zero, 1);
> + constexpr auto k1 = std::fmin(zero, 1);
> + constexpr auto l1 = std::fma(one, one, 1);
> + constexpr auto n1 = std::atan2(0, one);
> + constexpr auto o1 = std::hypot(1, zero);
> + constexpr auto p1 = std::pow(1, zero);
> + constexpr auto q1 = std::fmod(0, one);
> + constexpr auto r1 = std::remainder(1, one);
> + constexpr auto s1 = std::copysign(0, zero);
> + constexpr auto t1 = std::nextafter(0, zero);
> + constexpr auto u1 = std::fdim(0, zero);
> + constexpr auto v1 = std::fmax(0, one);
> + constexpr auto w1 = std::fmin(0, one);
> + constexpr auto x1 = std::fma(one, 1, one);
> + constexpr auto y1 = std::fma(1, one, one);
> + constexpr auto z1 = std::fma(1, 1, one);
> + constexpr auto A1 = std::fma(1, one, 1);
> +}
> +
> +int
> +main()
> +{
> +#if defined(__STDCPP_FLOAT16_T__) &&
> defined(_GLIBCXX_FLOAT_IS_IEEE_BINARY32)
> + test_functions<std::float16_t>();
> +#endif
> +#if defined(__STDCPP_FLOAT32_T__) &&
> defined(_GLIBCXX_FLOAT_IS_IEEE_BINARY32)
> + test_functions<std::float32_t>();
> +#endif
> +#if defined(__STDCPP_FLOAT64_T__) &&
> defined(_GLIBCXX_DOUBLE_IS_IEEE_BINARY64)
> + test_functions<std::float64_t>();
> +#endif
> +#if defined(__STDCPP_FLOAT128_T__) \
> + && (defined(_GLIBCXX_DOUBLE_IS_IEEE_BINARY128) \
> + || defined(_GLIBCXX_HAVE_FLOAT128_MATH))
> + test_functions<std::float128_t>();
> +#endif
> +#if defined(__STDCPP_BFLOAT16_T__) &&
> defined(_GLIBCXX_FLOAT_IS_IEEE_BINARY32)
> + test_functions<std::bfloat16_t>();
> +#endif
> +}
>
> Jakub
>
>
prev parent reply other threads:[~2023-05-17 19:15 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-17 18:18 Jakub Jelinek
2023-05-17 19:15 ` Jonathan Wakely [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CACb0b4=62XY00SCpf5FPSR=1H7HxQaHbadnCcbm9iLOZcZ6ObA@mail.gmail.com' \
--to=jwakely@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jakub@redhat.com \
--cc=libstdc++@gcc.gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).