From 9e7941b6ba32e39a5b6100be09abd7a091ff6ebf Mon Sep 17 00:00:00 2001
From: Niels <niels.lohmann@gmail.com>
Date: Tue, 17 Feb 2015 00:22:12 +0100
Subject: [PATCH] more type adjustments
---
src/json.hpp | 25 ++++++++++++++-----------
src/json.hpp.re2c | 25 ++++++++++++++-----------
test/unit.cpp | 1 +
3 files changed, 29 insertions(+), 22 deletions(-)
diff --git a/src/json.hpp b/src/json.hpp
index 186632f5..32b633fa 100644
--- a/src/json.hpp
+++ b/src/json.hpp
@@ -2529,11 +2529,11 @@ class basic_json
};
/// the char type to use in the lexer
- using lexer_char_t = typename string_t::value_type;
+ using lexer_char_t = unsigned char;
/// constructor with a given buffer
inline lexer(const string_t& s) noexcept
- : m_content(s.c_str())
+ : m_content(reinterpret_cast<const lexer_char_t*>(s.c_str()))
{
m_start = m_cursor = m_content;
m_limit = m_content + s.size();
@@ -2552,7 +2552,7 @@ class basic_json
@see <http://en.wikipedia.org/wiki/UTF-8#Sample_code>
*/
- inline static string_t to_unicode(const size_t codepoint1, size_t codepoint2 = 0)
+ inline static string_t to_unicode(const size_t codepoint1, const size_t codepoint2 = 0)
{
string_t result;
@@ -3380,7 +3380,8 @@ basic_json_parser_59:
/// return string representation of last read token
inline string_t get_token() const noexcept
{
- return string_t(m_start, static_cast<size_t>(m_cursor - m_start));
+ return string_t(reinterpret_cast<typename string_t::const_pointer>(m_start),
+ static_cast<size_t>(m_cursor - m_start));
}
/*!
@@ -3410,7 +3411,7 @@ basic_json_parser_59:
result.reserve(static_cast<size_t>(m_cursor - m_start - 2));
// iterate the result between the quotes
- for (const typename string_t::value_type* i = m_start + 1; i < m_cursor - 1; ++i)
+ for (const lexer_char_t* i = m_start + 1; i < m_cursor - 1; ++i)
{
// process escaped characters
if (*i == '\\')
@@ -3468,7 +3469,8 @@ basic_json_parser_59:
case 'u':
{
// get code xxxx from uxxxx
- auto codepoint = std::strtoul(std::string(i + 1, 4).c_str(), nullptr, 16);
+ auto codepoint = std::strtoul(std::string(reinterpret_cast<typename string_t::const_pointer>(i + 1),
+ 4).c_str(), nullptr, 16);
if (codepoint >= 0xD800 and codepoint <= 0xDBFF)
{
@@ -3479,7 +3481,8 @@ basic_json_parser_59:
}
// get code yyyy from uxxxx\uyyyy
- auto codepoint2 = std::strtoul(std::string(i + 7, 4).c_str(), nullptr, 16);
+ auto codepoint2 = std::strtoul(std::string(reinterpret_cast<typename string_t::const_pointer>
+ (i + 7), 4).c_str(), nullptr, 16);
result += to_unicode(codepoint, codepoint2);
// skip the next 11 characters (xxxx\uyyyy)
i += 11;
@@ -3499,7 +3502,7 @@ basic_json_parser_59:
{
// all other characters are just copied to the end of the
// string
- result.append(1, *i);
+ result.append(1, static_cast<typename string_t::value_type>(*i));
}
}
@@ -3526,13 +3529,13 @@ basic_json_parser_59:
inline number_float_t get_number() const
{
// conversion
- lexer_char_t* endptr;
- const auto float_val = std::strtod(reinterpret_cast<const lexer_char_t*>(m_start),
+ typename string_t::value_type* endptr;
+ const auto float_val = std::strtod(reinterpret_cast<typename string_t::const_pointer>(m_start),
&endptr);
// return float_val if the whole number was translated and NAN
// otherwise
- return (endptr == m_cursor) ? float_val : NAN;
+ return (reinterpret_cast<lexer_char_t*>(endptr) == m_cursor) ? float_val : NAN;
}
private:
diff --git a/src/json.hpp.re2c b/src/json.hpp.re2c
index 31d77c6b..4d1c88f2 100644
--- a/src/json.hpp.re2c
+++ b/src/json.hpp.re2c
@@ -2529,11 +2529,11 @@ class basic_json
};
/// the char type to use in the lexer
- using lexer_char_t = typename string_t::value_type;
+ using lexer_char_t = unsigned char;
/// constructor with a given buffer
inline lexer(const string_t& s) noexcept
- : m_content(s.c_str())
+ : m_content(reinterpret_cast<const lexer_char_t*>(s.c_str()))
{
m_start = m_cursor = m_content;
m_limit = m_content + s.size();
@@ -2552,7 +2552,7 @@ class basic_json
@see <http://en.wikipedia.org/wiki/UTF-8#Sample_code>
*/
- inline static string_t to_unicode(const size_t codepoint1, size_t codepoint2 = 0)
+ inline static string_t to_unicode(const size_t codepoint1, const size_t codepoint2 = 0)
{
string_t result;
@@ -2729,7 +2729,8 @@ class basic_json
/// return string representation of last read token
inline string_t get_token() const noexcept
{
- return string_t(m_start, static_cast<size_t>(m_cursor - m_start));
+ return string_t(reinterpret_cast<typename string_t::const_pointer>(m_start),
+ static_cast<size_t>(m_cursor - m_start));
}
/*!
@@ -2759,7 +2760,7 @@ class basic_json
result.reserve(static_cast<size_t>(m_cursor - m_start - 2));
// iterate the result between the quotes
- for (const typename string_t::value_type* i = m_start + 1; i < m_cursor - 1; ++i)
+ for (const lexer_char_t* i = m_start + 1; i < m_cursor - 1; ++i)
{
// process escaped characters
if (*i == '\\')
@@ -2817,7 +2818,8 @@ class basic_json
case 'u':
{
// get code xxxx from uxxxx
- auto codepoint = std::strtoul(std::string(i + 1, 4).c_str(), nullptr, 16);
+ auto codepoint = std::strtoul(std::string(reinterpret_cast<typename string_t::const_pointer>(i + 1),
+ 4).c_str(), nullptr, 16);
if (codepoint >= 0xD800 and codepoint <= 0xDBFF)
{
@@ -2828,7 +2830,8 @@ class basic_json
}
// get code yyyy from uxxxx\uyyyy
- auto codepoint2 = std::strtoul(std::string(i + 7, 4).c_str(), nullptr, 16);
+ auto codepoint2 = std::strtoul(std::string(reinterpret_cast<typename string_t::const_pointer>
+ (i + 7), 4).c_str(), nullptr, 16);
result += to_unicode(codepoint, codepoint2);
// skip the next 11 characters (xxxx\uyyyy)
i += 11;
@@ -2848,7 +2851,7 @@ class basic_json
{
// all other characters are just copied to the end of the
// string
- result.append(1, *i);
+ result.append(1, static_cast<typename string_t::value_type>(*i));
}
}
@@ -2875,13 +2878,13 @@ class basic_json
inline number_float_t get_number() const
{
// conversion
- lexer_char_t* endptr;
- const auto float_val = std::strtod(reinterpret_cast<const lexer_char_t*>(m_start),
+ typename string_t::value_type* endptr;
+ const auto float_val = std::strtod(reinterpret_cast<typename string_t::const_pointer>(m_start),
&endptr);
// return float_val if the whole number was translated and NAN
// otherwise
- return (endptr == m_cursor) ? float_val : NAN;
+ return (reinterpret_cast<lexer_char_t*>(endptr) == m_cursor) ? float_val : NAN;
}
private:
diff --git a/test/unit.cpp b/test/unit.cpp
index e1a38065..5e2ff903 100644
--- a/test/unit.cpp
+++ b/test/unit.cpp
@@ -5652,6 +5652,7 @@ TEST_CASE("parser class")
CHECK(json::parser("\"\\u2000\"").parse().get<json::string_t>() == " ");
CHECK(json::parser("\"\\uFFFF\"").parse().get<json::string_t>() == "");
CHECK(json::parser("\"\\u20AC\"").parse().get<json::string_t>() == "€");
+ CHECK(json::parser("\"€\"").parse().get<json::string_t>() == "€");
CHECK(json::parse("\"\\ud80c\\udc60\"").get<json::string_t>() == u8"\U00013060");
CHECK(json::parse("\"\\ud83c\\udf1e\"").get<json::string_t>() == "🌞");