X3: Regression 1.70: float parser precision

[gzauhar]

The following code is very simplified version of our math expression parser but it reveals regression introduced in 1.71.

#include <boost/spirit/home/x3.hpp>
#include <boost/spirit/home/x3/support/ast/variant.hpp>
#include <boost/variant/apply_visitor.hpp>
#include <boost/variant/static_visitor.hpp>

#include <iostream>
#include <string>

namespace x3 = boost::spirit::x3;

using ExpressionType = x3::variant<double>;

struct Expression : ExpressionType {
    using base_type::base_type;
    using base_type::operator=;
};

class ASTPrinter : public boost::static_visitor<void> {
public:
    void operator()(const double ast) const { printf("%e\n", ast); }
};

namespace grammar {

x3::real_parser<double, x3::ureal_policies<double>> const number = {};
x3::rule<class expression, Expression> const expression = "expression";

auto const expression_def =
        number;

BOOST_SPIRIT_DEFINE(
    expression
);

}  // namespace grammar

int main() {
    const std::string expression = "1e308";  // BAD
    // const std::string expression = "1.0e308";  // GOOD
    // const std::string expression = "1e307";  // GOOD
    auto position = expression.cbegin();

    Expression ast;
    x3::phrase_parse(position, expression.cend(), grammar::expression, x3::space, ast);

    ASTPrinter printer;
    boost::apply_visitor(printer, ast);
}

$ g++ -I boost-1.70 test.cpp && ./a.out 
1.000000e+308
$ g++ -I boost-1.71 test.cpp && ./a.out 
6.953216e-310

Are we misusing boost.spirit or there is a regression? This issue is blocking us from upgrading to 1.71 or newer. Thank you in advance!

[Kojoley]

#include <boost/spirit/home/x3.hpp>

#include <iostream>
#include <string>

int main() {
    namespace x3 = boost::spirit::x3;

    const std::string expression = "1e308";  // BAD
    // const std::string expression = "1.0e308";  // GOOD
    // const std::string expression = "1e307";  // GOOD
    auto position = expression.cbegin();

    x3::real_parser<double, x3::ureal_policies<double>> const number = {};

    double v;
    x3::phrase_parse(position, expression.cend(), number, x3::space, v);
    printf("%e\n", v);
    return v == 1e308 ? 0 : 1;
}

>git bisect start boost-1.71.0 boost-1.70.0
Bisecting: 25 revisions left to test after this (roughly 5 steps)
[d93e2c143cb8c8e02dbca50cdaf39f8a28d5bfa0] Lit suport for 32-bit unicode code points

>git bisect run issue567.bat
running issue567.bat
1.000000e+308
Bisecting: 12 revisions left to test after this (roughly 4 steps)
[90d03d9ae9cb16e0e43ea7e05f1e416d60d3f3ae] Fix for https://github.com/boostorg/spirit/issues/505
running issue567.bat
6.535756e-312
Bisecting: 6 revisions left to test after this (roughly 3 steps)
[747f7ebc41638bf1f2ed380b435bff35ce06d09f] deal with long double compares
running issue567.bat
6.535754e-312
Bisecting: 2 revisions left to test after this (roughly 2 steps)
[6b6f67720563c09a6edac49608eb801bb1c51ffd] cosmetic tweaks
running issue567.bat
1.000000e+308
Bisecting: 0 revisions left to test after this (roughly 1 step)
[5a001edc45a844c5d3328e454cc07c98caa75dc0] Added long_double parser
running issue567.bat
6.535755e-312
Bisecting: 0 revisions left to test after this (roughly 0 steps)
[793c070d1f8755399bc4ad1df59635102b7e2b57] Take advantage of constexpr. No more big hard coded table.
running issue567.bat
6.535761e-312
793c070d1f8755399bc4ad1df59635102b7e2b57 is the first bad commit
commit 793c070d1f8755399bc4ad1df59635102b7e2b57
Author: djowel <djowel@gmail.com>
Date:   Sat May 4 17:45:58 2019 +0800

    Take advantage of constexpr. No more big hard coded table.

:040000 040000 5c0f22d605a58ad3637ebb9a68f498c52ea7dc04 6e88bdba50a0845fec9ff494dc4ca7f0d7525b7a M      include
bisect run success

bisects to 793c070d1f8755399bc4ad1df59635102b7e2b57

[djowel]

Thanks, @Kojoley. I'll look into it as soon as I can.

[djowel]

Ah imprecise computations! Of course. Mea culpa!

[djowel]

I supose we'll have to bring back the big table, unless there's a better suggestion.

[djowel]

OK, reverted in develop.

[Kojoley]

I supose we'll have to bring back the big table, unless there's a better suggestion.

Use pow or fiddle with memcpy, but neither of them is constexpr even in c++2a.

[Kojoley]

The fix will be in 1.73 release. Thanks for your report!