avoid using impl in error_printer example

[N-Dekker]

What does this change do?

Replaced access to toml::impl::control_char_escapes with manual handling of control characters, in print_string.

Would ease a possible future replacement of #include <toml++/toml.hpp> with import tomplplusplus (because toml::impl is not exported).

Is it related to an existing bug report or feature request?

• It's another follow-up to pull request let error_printer example print source line of parse error #259.

• I've read CONTRIBUTING.md
• I've rebased my changes against the current HEAD of origin/master (if necessary)
• I've added new test cases to verify my change
• I've regenerated toml.hpp (how-to)
• I've updated any affected documentation
• I've rebuilt and run the tests with at least one of:
  • Clang 8 or higher
  • GCC 8 or higher
  • MSVC 19.20 (Visual Studio 2019) or higher
• I've added my name to the list of contributors in README.md

[N-Dekker]

Note: Two simple escape sequences, \a (0x07) and \v (0x0B), are not in toml::impl::control_char_escapes:

tomlplusplus/include/toml++/impl/forward_declarations.hpp

Lines 127 to 142 in fea1d90

	inline constexpr std::string_view control_char_escapes[] =
	{
	"\\u0000"sv,
	"\\u0001"sv,
	"\\u0002"sv,
	"\\u0003"sv,
	"\\u0004"sv,
	"\\u0005"sv,
	"\\u0006"sv,
	"\\u0007"sv,
	"\\b"sv,
	"\\t"sv,
	"\\n"sv,
	"\\u000B"sv,
	"\\f"sv,
	"\\r"sv,

I found them at https://eel.is/c++draft/lex.ccon#nt:simple-escape-sequence-char saying:

simple-escape-sequence-char: one of
  '  "  ?  \ a  b  f  n  r  t  v

Would it be worth adding these two to toml::impl::control_char_escapes as well? Honestly it doesn't matter much to me. Just asking 😃

[marzer]

They are in there:

• \u0007
• \u000B

This list is not used for parsing, so does not need to exhaustively list all possible representations. It's only used during serialization. Each entry in the array is at an index that corresponds to the equivalent codepoint.

[N-Dekker]

Thanks @marzer Just for my understanding, do you mean it's better to use \u0007 and \u000B (rather than \a and \v). Or do you mean that it just doesn't matter? 😃

[N-Dekker]

The print_string function in error_printer example is meant to print a human-readable output, so I guess for print_string, the human-friendly simple-escape-sequences \a and \v may be preferred, whereas for impl::control_char_escapes, \u0007 and \u000B may be preferred. Right?

[marzer]

I mean it doesn't matter, either should work when writing TOML. But you can't add additional values to that impl array as you suggested because the index of the elements matches their content. I don't recall why I chose those representations, was a few years ago now 🤷🏼‍♂️

[N-Dekker]

OK, thanks! By the way, I noticed something weird (?) in the draft of the C++ Standard. Table 10 - Simple escape sequences [tab:lex.ccon.esc] mentions both \a and \v. But there's another table, Table 112 - Mapping of characters to escape sequences [tab:format.escape.sequences], that doesn't mention them at all! Could it be that they accidentally forgot to mention those two, in table 112? It is part of a description of how to construct an escaped string, in section [format.string.escaped]

[marzer]

Yeah, entirely possible. It's a huge document.