johnmcfarlane

template <int c, long s> struct _;
#if defined(__clang__)
_<__clang_major__, __cplusplus> clang_version_and_standard;
#elif defined(__GNUC__)
_<__GNUC__, __cplusplus> gcc_version_and_standard;
#elif defined(_MSC_VER)
_<_MSC_VER, __cplusplus> msvc_version_and_standard;
#else
#error "I don't know that compiler"
#endif

johnmcfarlane

# 1 "<built-in>"
# 1 "/home/john/ws/rounding-overflow/cnl/test/unit/scaled_integer/rounding/elastic/integer.cpp"

//          Copyright John McFarlane 2015 - 2016.
// Distributed under the Boost Software License, Version 1.0.
//  (See accompanying file ../../LICENSE_1_0.txt or copy at
//          http://www.boost.org/LICENSE_1_0.txt)

#if 0 /* expanded by -frewrite-includes */
#include <cnl/elastic_integer.h>

johnmcfarlane

set -exu pipeline

SCRIPT_DIR=$(cd "$(dirname "$0")"; pwd)
PROJECT_DIR="$SCRIPT_DIR"/llvm-project

mkdir -p build
cd build
cmake \
  -GNinja \
  -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \

johnmcfarlane

#include "wwf_lexicon.h"
namespace {
node nAAHED = {
  "",
  nullptr,
  true
};
node nAAHEe[] {nAAHED};
node nAAHE = {
  "D",

johnmcfarlane

Where I Think We Are With a Numeric TS

• This is in answer to Bryce's question about how we should be working towards a Numeric TS.
• It's my (John McF's) frank and personal opinion about the bits I care about.
• Written a day before the Cologne 2019 meeting.

Current situation

I interpreted from Kona LEWG-I minutes that we're still exploring the design space. Specific work done by me for Cologne:

johnmcfarlane

// Out of curiosity I took preprocessor_fun.h
// https://gist.github.com/aras-p/6224951
// and added it to a project at work to see what happens when random C
// preprocessor hacks are applied to a modern C++ project with CI.
// Unfortunately, about half were flagged as warnings/errors by the compiler.
// Almost all of the rest had no effect because the codebase uses too much
// of the Standard Library / just doesn't use those APIs.
// The only one which wasn't caught by static tools was __builtin_expect
// which showed up as a failed death test. It is used in <gsl/gsl_assert> so
// should cause a decent amount of consternation in GSL-dependent projects.

johnmcfarlane

//          Copyright John McFarlane 2018.
// Distributed under the Boost Software License, Version 1.0.
//    (See accompanying file ../LICENSE_1_0.txt or copy at
//          http://www.boost.org/LICENSE_1_0.txt)

// mechanically retrieved, single-header version of CNL library
// https://github.com/johnmcfarlane/cnl

#if ! defined(CNL_COMPLETE_H)

johnmcfarlane

Tested on Ubuntu 18.10

1. Cloned copperspice/doxypress

2. README has no instructions on how to build and instead sends me to DoxyPress docs: https://www.copperspice.com/docs/doxypress/build-doxypress.html

   • Six major dependencies are presented

     • They don't differentiate between DoxyPress and DoxyPressApp

• #5 claims that I need automake or CMake. This doesn't appear to be the case (*).

johnmcfarlane

#include <functional>
#include <type_traits>
#include <utility>

namespace cnl2 {
  ////////////////////////////////////////////////////////////////////////////////
  // Arithmetic - is a something that supports arithmetic operators
  //
  // As well as a fundamental arithmetic type (unsigned short int, __int128, etc.)
  // a class-based type (a pack<> or a number<> instantiation) although bool is

johnmcfarlane

//          Copyright John McFarlane 2018.
// Distributed under the Boost Software License, Version 1.0.
//    (See accompanying file ../LICENSE_1_0.txt or copy at
//          http://www.boost.org/LICENSE_1_0.txt)

// mechanically retrieved, single-header version of CNL library
// https://github.com/johnmcfarlane/cnl

#if ! defined(CNL_COMPLETE_H)