diff options
author | Egor Tensin <Egor.Tensin@gmail.com> | 2020-01-03 00:58:56 +0300 |
---|---|---|
committer | Egor Tensin <Egor.Tensin@gmail.com> | 2020-01-03 00:58:56 +0300 |
commit | ad748449fb2d6154f0d7677f2a933aa29919075b (patch) | |
tree | 11278ebbf3890823fb6bf85014ade1b62921ff6e /server/lexer | |
parent | update cmake-common (diff) | |
download | math-server-ad748449fb2d6154f0d7677f2a933aa29919075b.tar.gz math-server-ad748449fb2d6154f0d7677f2a933aa29919075b.zip |
lexer: split out testing-exposed functions
Diffstat (limited to 'server/lexer')
-rw-r--r-- | server/lexer/CMakeLists.txt | 2 | ||||
-rw-r--r-- | server/lexer/details/parse.cpp | 107 | ||||
-rw-r--r-- | server/lexer/details/parse.hpp | 22 | ||||
-rw-r--r-- | server/lexer/lexer.cpp | 104 | ||||
-rw-r--r-- | server/lexer/lexer.hpp | 8 |
5 files changed, 135 insertions, 108 deletions
diff --git a/server/lexer/CMakeLists.txt b/server/lexer/CMakeLists.txt index 65523a7..b62e47a 100644 --- a/server/lexer/CMakeLists.txt +++ b/server/lexer/CMakeLists.txt @@ -1,2 +1,2 @@ -add_library(lexer lexer.cpp token.cpp token_type.cpp) +add_library(lexer details/parse.cpp lexer.cpp token.cpp token_type.cpp) target_link_libraries(lexer PUBLIC common) diff --git a/server/lexer/details/parse.cpp b/server/lexer/details/parse.cpp new file mode 100644 index 0000000..464dfb2 --- /dev/null +++ b/server/lexer/details/parse.cpp @@ -0,0 +1,107 @@ +// Copyright (c) 2020 Egor Tensin <Egor.Tensin@gmail.com> +// This file is part of the "math-server" project. +// For details, see https://github.com/egor-tensin/math-server. +// Distributed under the MIT License. + +#include "../error.hpp" +#include "../token_type.hpp" + +#include <cstddef> + +#include <exception> +#include <optional> +#include <regex> +#include <string> +#include <string_view> + +namespace math::server::lexer::details { +namespace { + +std::string_view match_number(const std::string_view& input) { + static constexpr std::regex::flag_type flags = + std::regex_constants::ECMAScript | + std::regex_constants::icase; + // This is a hacky attempt to describe a C-like grammar for floating-point + // numbers using a regex (the tests seem to pass though). + // A proper NFA would be better, I guess. + static const std::regex number_regex{R"REGEX(^(?:\d+(?:\.\d*)?|\.\d+)(e[+-]?(\d*))?)REGEX", flags}; + + std::cmatch match; + { + const auto begin = input.data(); + const auto end = begin + input.length(); + if (!std::regex_search(begin, end, match, number_regex)) { + return {}; + } + } + { + // If we have the numeric part of a number followed by 'e' and no digits, + // 1) that 'e' definitely belongs to this number token, + // 2) the user forgot to type in the required digits. + const auto& exponent = match[1]; + const auto& abs_power = match[2]; + if (exponent.matched && abs_power.matched && abs_power.length() == 0) { + throw LexerError{"exponent has no digits: " + match[0].str()}; + } + } + return {match[0].first, static_cast<std::size_t>(match[0].length())}; +} + +bool starts_with(const std::string_view& a, const std::string_view& b) noexcept { + return a.length() >= b.length() + && a.compare(0, b.length(), b) == 0; +} + +} + +std::optional<double> parse_number(const std::string_view& input, std::string_view& token) { + const auto view = match_number(input); + if (!view.data()) { + return {}; + } + try { + const auto result = std::stod(std::string{view}); + token = view; + return result; + } catch (const std::exception&) { + throw LexerError{"internal: couldn't parse number from: " + std::string{view}}; + } + return {}; +} + +std::optional<double> parse_number(const std::string_view& input) { + std::string_view token; + return parse_number(input, token); +} + +std::optional<token::Type> parse_const_token(const std::string_view& input, std::string_view& token) { + for (const auto type : token::const_tokens()) { + const auto str = token::type_to_string(type); + if (starts_with(input, str)) { + token = std::string_view(input.data(), str.length()); + return {type}; + } + } + return {}; +} + +std::optional<token::Type> parse_const_token(const std::string_view& input) { + std::string_view token; + return parse_const_token(input, token); +} + +std::string_view parse_whitespace(const std::string_view& input) { + static const std::regex ws_regex{R"(^\s+)"}; + + std::cmatch match; + { + const auto begin = input.data(); + const auto end = begin + input.length(); + if (std::regex_search(begin, end, match, ws_regex)) { + return std::string_view(match[0].first, match[0].length()); + } + } + return {}; +} + +} diff --git a/server/lexer/details/parse.hpp b/server/lexer/details/parse.hpp new file mode 100644 index 0000000..72da234 --- /dev/null +++ b/server/lexer/details/parse.hpp @@ -0,0 +1,22 @@ +// Copyright (c) 2020 Egor Tensin <Egor.Tensin@gmail.com> +// This file is part of the "math-server" project. +// For details, see https://github.com/egor-tensin/math-server. +// Distributed under the MIT License. + +#pragma once + +#include "../token_type.hpp" + +#include <optional> +#include <string_view> + +namespace math::server::lexer::details { + +// Exposed for testing: +std::string_view parse_whitespace(const std::string_view&); +std::optional<double> parse_number(const std::string_view&, std::string_view&); +std::optional<double> parse_number(const std::string_view&); +std::optional<token::Type> parse_const_token(const std::string_view&, std::string_view&); +std::optional<token::Type> parse_const_token(const std::string_view&); + +} diff --git a/server/lexer/lexer.cpp b/server/lexer/lexer.cpp index 24e3f14..cf24189 100644 --- a/server/lexer/lexer.cpp +++ b/server/lexer/lexer.cpp @@ -3,113 +3,19 @@ // For details, see https://github.com/egor-tensin/math-server. // Distributed under the MIT License. +#include "details/parse.hpp" #include "error.hpp" #include "lexer.hpp" #include "token.hpp" #include "token_type.hpp" -#include <exception> #include <optional> -#include <regex> -#include <string_view> #include <string> +#include <string_view> +#include <utility> #include <vector> namespace math::server { -namespace lexer { -namespace { - -std::string_view match_number(const std::string_view& input) { - static constexpr std::regex::flag_type flags = - std::regex_constants::ECMAScript | - std::regex_constants::icase; - // This is a hacky attempt to describe a C-like grammar for floating-point - // numbers using a regex (the tests seem to pass though). - // A proper NFA would be better, I guess. - static const std::regex number_regex{R"REGEX(^(?:\d+(?:\.\d*)?|\.\d+)(e[+-]?(\d*))?)REGEX", flags}; - - std::cmatch match; - { - const auto begin = input.data(); - const auto end = begin + input.length(); - if (!std::regex_search(begin, end, match, number_regex)) { - return {}; - } - } - { - // If we have the numeric part of a number followed by 'e' and no digits, - // 1) that 'e' definitely belongs to this number token, - // 2) the user forgot to type in the required digits. - const auto& exponent = match[1]; - const auto& abs_power = match[2]; - if (exponent.matched && abs_power.matched && abs_power.length() == 0) { - throw LexerError{"exponent has no digits: " + match[0].str()}; - } - } - return {match[0].first, static_cast<std::size_t>(match[0].length())}; -} - -std::optional<double> parse_number(const std::string_view& input, std::string_view& token) { - const auto view = match_number(input); - if (!view.data()) { - return {}; - } - try { - const auto result = std::stod(std::string{view}); - token = view; - return result; - } catch (const std::exception&) { - throw LexerError{"internal: couldn't parse number from: " + std::string{view}}; - } - return {}; -} - -bool starts_with(const std::string_view& a, const std::string_view& b) noexcept { - return a.length() >= b.length() - && a.compare(0, b.length(), b) == 0; -} - -std::optional<token::Type> parse_const_token(const std::string_view& input, std::string_view& token) { - for (const auto type : token::const_tokens()) { - const auto str = token::type_to_string(type); - if (starts_with(input, str)) { - token = std::string_view(input.data(), str.length()); - return {type}; - } - } - return {}; -} - -} - -namespace details { - -std::optional<double> parse_number(const std::string_view& input) { - std::string_view token; - return lexer::parse_number(input, token); -} - -std::optional<token::Type> parse_const_token(const std::string_view& input) { - std::string_view token; - return lexer::parse_const_token(input, token); -} - -std::string_view parse_whitespace(const std::string_view& input) { - static const std::regex ws_regex{R"(^\s+)"}; - - std::cmatch match; - { - const auto begin = input.data(); - const auto end = begin + input.length(); - if (std::regex_search(begin, end, match, ws_regex)) { - return std::string_view(match[0].first, match[0].length()); - } - } - return {}; -} - -} -} Lexer::Lexer(const std::string_view& input) : Lexer{lexer::Input{input}} { @@ -190,7 +96,7 @@ std::optional<Lexer::ParsedToken> Lexer::parse_whitespace() const { std::optional<Lexer::ParsedToken> Lexer::parse_const_token() const { std::string_view token_view; - const auto type = lexer::parse_const_token(m_input.get_input(), token_view); + const auto type = lexer::details::parse_const_token(m_input.get_input(), token_view); if (!type.has_value()) { return {}; } @@ -199,7 +105,7 @@ std::optional<Lexer::ParsedToken> Lexer::parse_const_token() const { std::optional<Lexer::ParsedToken> Lexer::parse_number() const { std::string_view token_view; - const auto number = lexer::parse_number(m_input.get_input(), token_view); + const auto number = lexer::details::parse_number(m_input.get_input(), token_view); if (!number.has_value()) { return {}; } diff --git a/server/lexer/lexer.hpp b/server/lexer/lexer.hpp index 44831d8..68950cb 100644 --- a/server/lexer/lexer.hpp +++ b/server/lexer/lexer.hpp @@ -15,14 +15,6 @@ #include <vector> namespace math::server { -namespace lexer::details { - -// Exposed for testing: -std::string_view parse_whitespace(const std::string_view&); -std::optional<double> parse_number(const std::string_view&); -std::optional<token::Type> parse_const_token(const std::string_view&); - -} class Lexer { public: |