DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 006D03858C56
Message-ID: <c079ef99-e047-4a9b-8179-eca7bbbbe334@gmail.com>
Date: Wed, 17 Jul 2024 01:27:41 +0200
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: Urgent gcc update to GCC 13.3
To: cygwin@cygwin.com
References: <CAN0SSYxqC85dMNXzdogMXkEOEZiU-YPCsp=oDq5oYfz3OH6NXg@mail.gmail.com>
 <CALXu0Udho3sY6zCojnEipZjn4pwbQe43nWv3k1n-p3oxRFTFJQ@mail.gmail.com>
 <82c9bc8e-4048-43d8-a9ad-8ed5485a24a8@gmail.com>
 <CAN0SSYwEG-EtPPhCfx5mW_X2RUA-8euKTDNep1ZcChEtbGB-Uw@mail.gmail.com>
Content-Language: en-GB, hu
In-Reply-To: <CAN0SSYwEG-EtPPhCfx5mW_X2RUA-8euKTDNep1ZcChEtbGB-Uw@mail.gmail.com>
Precedence: list
From: =?utf-8?q?Csaba_R=C3=A1duly_via_Cygwin?= <cygwin@cygwin.com>
Reply-To: =?UTF-8?Q?Csaba_R=C3=A1duly?= <rcsaba@gmail.com>
Content-Type: text/plain; charset="utf-8"; Format="flowed"
Errors-To: cygwin-bounces~archive-cygwin=delorie.com@cygwin.com
Sender: "Cygwin" <cygwin-bounces~archive-cygwin=delorie.com@cygwin.com>
Content-Transfer-Encoding: 8bit

On 15/07/2024 22:44, Mark Liam Brown via Cygwin wrote:
> On Mon, Jul 15, 2024 at 9:35 PM Csaba Ráduly via Cygwin
> <cygwin@cygwin.com> wrote:
>> On 09/07/2024 08:17, Cedric Blancher via Cygwin wrote:
>>> I'd agree that newer gcc 13.x to compile Cygwin and as /usr/bin/gcc
>>> would be a very good idea, not only for performance+STL fixes, but
>>> also since by default the Cygwin distro lacks C++17 support,
>> Huh? GCC supports C++17 since GCC 7
>>
>> https://en.cppreference.com/w/cpp/17
> No, it does not. *Some* features work, but gcc 13.1 was the first
> release where everything mandated by C++17 is working, or at least
> compiling.
> So far g++ in Cygwin 3.5.3 cannot compile anything in std::pmr:*, e.g.
> std::pmr::monotonic_buffer_resource, std::pmr::polymorphic_allocator,
> std::pmr::string, ruling out any realistic C++17 apps (of course,
> HelloWorld will work).
>
> Mark

A few examples of non-realistic, HelloWorld class projects:

LLVM:

~/wk/llvm-project $ ag -s --cpp -w pmr bolt/ clang* compiler-rt/ 
cross-project-tests/ flang/ libclc libunwind/ ll* mlir/ offload/ openmp/ 
polly/ pstl/ runtimes/ third-party/ utils/
clang/test/SemaCXX/overloaded-builtin-operators.cpp
116:  pmf &pmr = (pmf_ref = &X::f); // expected-error{{no viable 
overloaded '='}}
~/wk/llvm-project $

Abseil:

~/wk/abseil-cpp$ ag -s --cpp -w pmr absl/
absl/container/internal/raw_hash_set_allocator_test.cc
450:// This allocator is similar to std::pmr::polymorphic_allocator.
~/wk/abseil-cpp$

GCC:

~/wk/gcc-project$ ag -l -c -s --cpp -w --ignore libstdc++-v3 pmr
gcc/cp/std-name-hint.h:10
gcc/testsuite/g++.dg/rtti/typeid7.C:2
~/wk/gcc-project$

Electron

~/wk/electron$ ag -l -c -s --cpp -w pmr
~/wk/electron$

LibreOffice

~/wk/libreoffice$ ag -l -c -s --cpp -w pmr
~/wk/libreoffice$

TensorFlow

~/wk/tensorflow$ ag -l -c -s --cpp -w pmr
~/wk/tensorflow$

Polymorphic memory resources were supported by GCC 9, according to 
https://en.cppreference.com/w/cpp/compiler_support/17

Anyway, nobody cares about std::pmr::anything , except maybe John Lakos 
and the Intel TBB project. /hyperbole

Search results on github (thousands)

pmr::                8
set_symmetric_difference    8.4
lexicographical_compare    32
nth_element    25
std::rotate        17.5
vector            27000
std::vector    12100
std::string    16300


Csaba

-- 
Life is complex, with real and imaginary parts.


-- 
Problem reports:      https://cygwin.com/problems.html
FAQ:                  https://cygwin.com/faq/
Documentation:        https://cygwin.com/docs.html
Unsubscribe info:     https://cygwin.com/ml/#unsubscribe-simple