From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id 69B473858C1F; Wed, 31 May 2023 13:51:05 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 69B473858C1F
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1685541065;
	bh=Om8qNkJY/89pHLs+rtWepjz7SDmXGDrqoTvYi6Bvuoo=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=E2FRkQtEDesuNl6WJ1/z2+79sH1FkGmKG1cx+1FSLO0FnNZIJGfI0yidD6ZhnHruN
	 U0FibdCa6y41Ke8l3YDuwfAsxbY/42aCXgNNsEnJVWvwfi4fCss41CEEutDysg4D/B
	 tRfAkK0rcynYDDKHmcyAX9JeXXWEjiZQ8OU0YKqE=
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/104772] std::numeric_limits<__float128> should be
 specialized
Date: Wed, 31 May 2023 13:51:01 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: libstdc++
X-Bugzilla-Version: 12.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: enhancement
X-Bugzilla-Who: jakub at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 14.0
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-104772-4-UPHMv0k1AB@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-104772-4@http.gcc.gnu.org/bugzilla/>
References: <bug-104772-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D104772

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Dunno.
I think you can add support even without any compiler changes, at least if
_GLIBCXX_DOUBLE_IS_IEEE_BINARY64:

#ifdef __STRICT_ANSI__
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0static=C2=A0_GLIBCXX_CONSTEXPR=C2=A0__f=
loat128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0min()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return=C2=A0__extension__=C2=A03=
.36210314311209350626267781732175260e-4932Q;=C2=A0}

=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0static=C2=A0_GLIBCXX_CONSTEXPR=C2=A0__f=
loat128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0max()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return=C2=A0__extension__=C2=A01=
.18973149535723176508575932662800702e+4932Q;=C2=A0}

=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0static=C2=A0_GLIBCXX_CONSTEXPR=C2=A0__f=
loat128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0epsilon()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return=C2=A0__extension__=C2=A01=
.92592994438723585305597794258492732e-34Q;=C2=A0}

=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0static=C2=A0_GLIBCXX_CONSTEXPR=C2=A0__f=
loat128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0denorm_min()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return=C2=A0__extension__=C2=A06=
.47517511943802511092443895822764655e-4966Q;=C2=A0}
#else
      static _GLIBCXX_CONSTEXPR __float128
      _S_1pm4088 () _GLIBCXX_USE_NOEXCEPT
      { return __float128 (0x1.0p-1022) * 0x1.0p-1022 * 0x1.0p-1022 *
0x1.0p-1022; }

      static _GLIBCXX_CONSTEXPR __float128
      _S_1pm16352 () _GLIBCXX_USE_NOEXCEPT
      { return _S_1pm4088 () * _S_1pm4088 () * _S_1pm4088 () * _S_1pm4088 (=
); }

      static _GLIBCXX_CONSTEXPR __float128
      _S_1p4064 () _GLIBCXX_USE_NOEXCEPT
      { return __float128 (0x1.0p+1016) * 0x1.0p-1016 * 0x1.0p-1016 *
0x1.0p-1016; }

      static _GLIBCXX_CONSTEXPR __float128
      _S_1p16256 () _GLIBCXX_USE_NOEXCEPT
      { return _S_1p4064 () * _S_1p4064 () * _S_1p4064 () * _S_1p4064 (); }

      static _GLIBCXX_CONSTEXPR __float128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0min()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return /* 0x1.0p-16382Q */ 0x1.0=
p-30 * _S_1pm16352 ();=C2=A0}

=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0static=C2=A0_GLIBCXX_CONSTEXPR=C2=A0__f=
loat128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0max()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return=C2=A0/* 0x1.fffffffffffff=
fffffffffffffffp+16383Q */
               (__float128 (0x1.fffffffffffffp+127) + 0x0.fffffffffffffp+75
                            + 0x0.ffp+23) * _S_1p16256 ();=C2=A0}

=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0static=C2=A0_GLIBCXX_CONSTEXPR=C2=A0__f=
loat128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0epsilon()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return=C2=A00x1.0p-112;=C2=A0}

=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0static=C2=A0_GLIBCXX_CONSTEXPR=C2=A0__f=
loat128
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0denorm_min()=C2=A0_GLIBCXX_USE_NOEXCEPT
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0{=C2=A0return=C2=A0/* 0x1.0p-16494Q */ =
0x1.0p-142 * _S_1pm16352 ();=C2=A0}
#endif

Except that the FE errors in C++98/11/14 modes on the hexadecimal constants.
I guess one could replace them with decimal constants (except those in comm=
ents
of course), but better use explicit double(whatever) to make sure they aren=
't
evaluated in higher precision.  Anyway, I think the above computations shou=
ld
(at least assuming the decimal constants map to those hexadecimal ones) alw=
ays
be exact and so should produce such results in all rounding modes.=