mirror of
https://github.com/Ninjdai1/pokeemerald.git
synced 2024-12-27 12:14:20 +01:00
3530 lines
104 KiB
C++
Executable File
3530 lines
104 KiB
C++
Executable File
// MIT License
|
|
|
|
// Copyright (c) 2018 lbersch
|
|
|
|
// Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
// of this software and associated documentation files (the "Software"), to deal
|
|
// in the Software without restriction, including without limitation the rights
|
|
// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
// copies of the Software, and to permit persons to whom the Software is
|
|
// furnished to do so, subject to the following conditions:
|
|
|
|
// The above copyright notice and this permission notice shall be included in all
|
|
// copies or substantial portions of the Software.
|
|
|
|
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
// SOFTWARE.
|
|
|
|
|
|
// ---
|
|
|
|
|
|
// Copyright (c) 2009-2018 FIRST
|
|
// All rights reserved.
|
|
|
|
// Redistribution and use in source and binary forms, with or without
|
|
// modification, are permitted provided that the following conditions are met:
|
|
// * Redistributions of source code must retain the above copyright
|
|
// notice, this list of conditions and the following disclaimer.
|
|
// * Redistributions in binary form must reproduce the above copyright
|
|
// notice, this list of conditions and the following disclaimer in the
|
|
// documentation and/or other materials provided with the distribution.
|
|
// * Neither the name of the FIRST nor the
|
|
// names of its contributors may be used to endorse or promote products
|
|
// derived from this software without specific prior written permission.
|
|
|
|
// THIS SOFTWARE IS PROVIDED BY FIRST AND CONTRIBUTORS``AS IS'' AND ANY
|
|
// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
|
|
// WARRANTIES OF MERCHANTABILITY NONINFRINGEMENT AND FITNESS FOR A PARTICULAR
|
|
// PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL FIRST OR CONTRIBUTORS BE LIABLE FOR
|
|
// ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
|
|
// (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
|
|
// LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
|
|
// ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
|
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
|
// SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
#ifndef PANTOR_INJA_HPP
|
|
#define PANTOR_INJA_HPP
|
|
|
|
#include <functional>
|
|
#include <iostream>
|
|
#include <map>
|
|
#include <memory>
|
|
#include <sstream>
|
|
#include <string>
|
|
#include <vector>
|
|
|
|
#include <nlohmann/json.hpp>
|
|
|
|
// #include "environment.hpp"
|
|
#ifndef PANTOR_INJA_ENVIRONMENT_HPP
|
|
#define PANTOR_INJA_ENVIRONMENT_HPP
|
|
|
|
#include <memory>
|
|
#include <fstream>
|
|
#include <sstream>
|
|
#include <string>
|
|
|
|
#include <nlohmann/json.hpp>
|
|
|
|
// #include "config.hpp"
|
|
#ifndef PANTOR_INJA_CONFIG_HPP
|
|
#define PANTOR_INJA_CONFIG_HPP
|
|
|
|
#include <functional>
|
|
#include <string>
|
|
|
|
// #include "string_view.hpp"
|
|
// Copyright 2017-2019 by Martin Moene
|
|
//
|
|
// string-view lite, a C++17-like string_view for C++98 and later.
|
|
// For more information see https://github.com/martinmoene/string-view-lite
|
|
//
|
|
// Distributed under the Boost Software License, Version 1.0.
|
|
// (See accompanying file LICENSE.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
|
|
|
|
|
|
|
|
#ifndef NONSTD_SV_LITE_H_INCLUDED
|
|
#define NONSTD_SV_LITE_H_INCLUDED
|
|
|
|
#define string_view_lite_MAJOR 1
|
|
#define string_view_lite_MINOR 1
|
|
#define string_view_lite_PATCH 0
|
|
|
|
#define string_view_lite_VERSION nssv_STRINGIFY(string_view_lite_MAJOR) "." nssv_STRINGIFY(string_view_lite_MINOR) "." nssv_STRINGIFY(string_view_lite_PATCH)
|
|
|
|
#define nssv_STRINGIFY( x ) nssv_STRINGIFY_( x )
|
|
#define nssv_STRINGIFY_( x ) #x
|
|
|
|
// string-view lite configuration:
|
|
|
|
#define nssv_STRING_VIEW_DEFAULT 0
|
|
#define nssv_STRING_VIEW_NONSTD 1
|
|
#define nssv_STRING_VIEW_STD 2
|
|
|
|
#if !defined( nssv_CONFIG_SELECT_STRING_VIEW )
|
|
# define nssv_CONFIG_SELECT_STRING_VIEW ( nssv_HAVE_STD_STRING_VIEW ? nssv_STRING_VIEW_STD : nssv_STRING_VIEW_NONSTD )
|
|
#endif
|
|
|
|
#if defined( nssv_CONFIG_SELECT_STD_STRING_VIEW ) || defined( nssv_CONFIG_SELECT_NONSTD_STRING_VIEW )
|
|
# error nssv_CONFIG_SELECT_STD_STRING_VIEW and nssv_CONFIG_SELECT_NONSTD_STRING_VIEW are deprecated and removed, please use nssv_CONFIG_SELECT_STRING_VIEW=nssv_STRING_VIEW_...
|
|
#endif
|
|
|
|
#ifndef nssv_CONFIG_STD_SV_OPERATOR
|
|
# define nssv_CONFIG_STD_SV_OPERATOR 0
|
|
#endif
|
|
|
|
#ifndef nssv_CONFIG_USR_SV_OPERATOR
|
|
# define nssv_CONFIG_USR_SV_OPERATOR 1
|
|
#endif
|
|
|
|
#ifdef nssv_CONFIG_CONVERSION_STD_STRING
|
|
# define nssv_CONFIG_CONVERSION_STD_STRING_CLASS_METHODS nssv_CONFIG_CONVERSION_STD_STRING
|
|
# define nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS nssv_CONFIG_CONVERSION_STD_STRING
|
|
#endif
|
|
|
|
#ifndef nssv_CONFIG_CONVERSION_STD_STRING_CLASS_METHODS
|
|
# define nssv_CONFIG_CONVERSION_STD_STRING_CLASS_METHODS 1
|
|
#endif
|
|
|
|
#ifndef nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS
|
|
# define nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS 1
|
|
#endif
|
|
|
|
// Control presence of exception handling (try and auto discover):
|
|
|
|
#ifndef nssv_CONFIG_NO_EXCEPTIONS
|
|
# if defined(__cpp_exceptions) || defined(__EXCEPTIONS) || defined(_CPPUNWIND)
|
|
# define nssv_CONFIG_NO_EXCEPTIONS 0
|
|
# else
|
|
# define nssv_CONFIG_NO_EXCEPTIONS 1
|
|
# endif
|
|
#endif
|
|
|
|
// C++ language version detection (C++20 is speculative):
|
|
// Note: VC14.0/1900 (VS2015) lacks too much from C++14.
|
|
|
|
#ifndef nssv_CPLUSPLUS
|
|
# if defined(_MSVC_LANG ) && !defined(__clang__)
|
|
# define nssv_CPLUSPLUS (_MSC_VER == 1900 ? 201103L : _MSVC_LANG )
|
|
# else
|
|
# define nssv_CPLUSPLUS __cplusplus
|
|
# endif
|
|
#endif
|
|
|
|
#define nssv_CPP98_OR_GREATER ( nssv_CPLUSPLUS >= 199711L )
|
|
#define nssv_CPP11_OR_GREATER ( nssv_CPLUSPLUS >= 201103L )
|
|
#define nssv_CPP11_OR_GREATER_ ( nssv_CPLUSPLUS >= 201103L )
|
|
#define nssv_CPP14_OR_GREATER ( nssv_CPLUSPLUS >= 201402L )
|
|
#define nssv_CPP17_OR_GREATER ( nssv_CPLUSPLUS >= 201703L )
|
|
#define nssv_CPP20_OR_GREATER ( nssv_CPLUSPLUS >= 202000L )
|
|
|
|
// use C++17 std::string_view if available and requested:
|
|
|
|
#if nssv_CPP17_OR_GREATER && defined(__has_include )
|
|
# if __has_include( <string_view> )
|
|
# define nssv_HAVE_STD_STRING_VIEW 1
|
|
# else
|
|
# define nssv_HAVE_STD_STRING_VIEW 0
|
|
# endif
|
|
#else
|
|
# define nssv_HAVE_STD_STRING_VIEW 0
|
|
#endif
|
|
|
|
#define nssv_USES_STD_STRING_VIEW ( (nssv_CONFIG_SELECT_STRING_VIEW == nssv_STRING_VIEW_STD) || ((nssv_CONFIG_SELECT_STRING_VIEW == nssv_STRING_VIEW_DEFAULT) && nssv_HAVE_STD_STRING_VIEW) )
|
|
|
|
#define nssv_HAVE_STARTS_WITH ( nssv_CPP20_OR_GREATER || !nssv_USES_STD_STRING_VIEW )
|
|
#define nssv_HAVE_ENDS_WITH nssv_HAVE_STARTS_WITH
|
|
|
|
//
|
|
// Use C++17 std::string_view:
|
|
//
|
|
|
|
#if nssv_USES_STD_STRING_VIEW
|
|
|
|
#include <string_view>
|
|
|
|
// Extensions for std::string:
|
|
|
|
#if nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS
|
|
|
|
namespace nonstd {
|
|
|
|
template< class CharT, class Traits, class Allocator = std::allocator<CharT> >
|
|
std::basic_string<CharT, Traits, Allocator>
|
|
to_string( std::basic_string_view<CharT, Traits> v, Allocator const & a = Allocator() )
|
|
{
|
|
return std::basic_string<CharT,Traits, Allocator>( v.begin(), v.end(), a );
|
|
}
|
|
|
|
template< class CharT, class Traits, class Allocator >
|
|
std::basic_string_view<CharT, Traits>
|
|
to_string_view( std::basic_string<CharT, Traits, Allocator> const & s )
|
|
{
|
|
return std::basic_string_view<CharT, Traits>( s.data(), s.size() );
|
|
}
|
|
|
|
// Literal operators sv and _sv:
|
|
|
|
#if nssv_CONFIG_STD_SV_OPERATOR
|
|
|
|
using namespace std::literals::string_view_literals;
|
|
|
|
#endif
|
|
|
|
#if nssv_CONFIG_USR_SV_OPERATOR
|
|
|
|
inline namespace literals {
|
|
inline namespace string_view_literals {
|
|
|
|
|
|
constexpr std::string_view operator "" _sv( const char* str, size_t len ) noexcept // (1)
|
|
{
|
|
return std::string_view{ str, len };
|
|
}
|
|
|
|
constexpr std::u16string_view operator "" _sv( const char16_t* str, size_t len ) noexcept // (2)
|
|
{
|
|
return std::u16string_view{ str, len };
|
|
}
|
|
|
|
constexpr std::u32string_view operator "" _sv( const char32_t* str, size_t len ) noexcept // (3)
|
|
{
|
|
return std::u32string_view{ str, len };
|
|
}
|
|
|
|
constexpr std::wstring_view operator "" _sv( const wchar_t* str, size_t len ) noexcept // (4)
|
|
{
|
|
return std::wstring_view{ str, len };
|
|
}
|
|
|
|
}} // namespace literals::string_view_literals
|
|
|
|
#endif // nssv_CONFIG_USR_SV_OPERATOR
|
|
|
|
} // namespace nonstd
|
|
|
|
#endif // nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS
|
|
|
|
namespace nonstd {
|
|
|
|
using std::string_view;
|
|
using std::wstring_view;
|
|
using std::u16string_view;
|
|
using std::u32string_view;
|
|
using std::basic_string_view;
|
|
|
|
// literal "sv" and "_sv", see above
|
|
|
|
using std::operator==;
|
|
using std::operator!=;
|
|
using std::operator<;
|
|
using std::operator<=;
|
|
using std::operator>;
|
|
using std::operator>=;
|
|
|
|
using std::operator<<;
|
|
|
|
} // namespace nonstd
|
|
|
|
#else // nssv_HAVE_STD_STRING_VIEW
|
|
|
|
//
|
|
// Before C++17: use string_view lite:
|
|
//
|
|
|
|
// Compiler versions:
|
|
//
|
|
// MSVC++ 6.0 _MSC_VER == 1200 (Visual Studio 6.0)
|
|
// MSVC++ 7.0 _MSC_VER == 1300 (Visual Studio .NET 2002)
|
|
// MSVC++ 7.1 _MSC_VER == 1310 (Visual Studio .NET 2003)
|
|
// MSVC++ 8.0 _MSC_VER == 1400 (Visual Studio 2005)
|
|
// MSVC++ 9.0 _MSC_VER == 1500 (Visual Studio 2008)
|
|
// MSVC++ 10.0 _MSC_VER == 1600 (Visual Studio 2010)
|
|
// MSVC++ 11.0 _MSC_VER == 1700 (Visual Studio 2012)
|
|
// MSVC++ 12.0 _MSC_VER == 1800 (Visual Studio 2013)
|
|
// MSVC++ 14.0 _MSC_VER == 1900 (Visual Studio 2015)
|
|
// MSVC++ 14.1 _MSC_VER >= 1910 (Visual Studio 2017)
|
|
|
|
#if defined(_MSC_VER ) && !defined(__clang__)
|
|
# define nssv_COMPILER_MSVC_VER (_MSC_VER )
|
|
# define nssv_COMPILER_MSVC_VERSION (_MSC_VER / 10 - 10 * ( 5 + (_MSC_VER < 1900 ) ) )
|
|
#else
|
|
# define nssv_COMPILER_MSVC_VER 0
|
|
# define nssv_COMPILER_MSVC_VERSION 0
|
|
#endif
|
|
|
|
#define nssv_COMPILER_VERSION( major, minor, patch ) (10 * ( 10 * major + minor) + patch)
|
|
|
|
#if defined(__clang__)
|
|
# define nssv_COMPILER_CLANG_VERSION nssv_COMPILER_VERSION(__clang_major__, __clang_minor__, __clang_patchlevel__)
|
|
#else
|
|
# define nssv_COMPILER_CLANG_VERSION 0
|
|
#endif
|
|
|
|
#if defined(__GNUC__) && !defined(__clang__)
|
|
# define nssv_COMPILER_GNUC_VERSION nssv_COMPILER_VERSION(__GNUC__, __GNUC_MINOR__, __GNUC_PATCHLEVEL__)
|
|
#else
|
|
# define nssv_COMPILER_GNUC_VERSION 0
|
|
#endif
|
|
|
|
// half-open range [lo..hi):
|
|
#define nssv_BETWEEN( v, lo, hi ) ( (lo) <= (v) && (v) < (hi) )
|
|
|
|
// Presence of language and library features:
|
|
|
|
#ifdef _HAS_CPP0X
|
|
# define nssv_HAS_CPP0X _HAS_CPP0X
|
|
#else
|
|
# define nssv_HAS_CPP0X 0
|
|
#endif
|
|
|
|
// Unless defined otherwise below, consider VC14 as C++11 for variant-lite:
|
|
|
|
#if nssv_COMPILER_MSVC_VER >= 1900
|
|
# undef nssv_CPP11_OR_GREATER
|
|
# define nssv_CPP11_OR_GREATER 1
|
|
#endif
|
|
|
|
#define nssv_CPP11_90 (nssv_CPP11_OR_GREATER_ || nssv_COMPILER_MSVC_VER >= 1500)
|
|
#define nssv_CPP11_100 (nssv_CPP11_OR_GREATER_ || nssv_COMPILER_MSVC_VER >= 1600)
|
|
#define nssv_CPP11_110 (nssv_CPP11_OR_GREATER_ || nssv_COMPILER_MSVC_VER >= 1700)
|
|
#define nssv_CPP11_120 (nssv_CPP11_OR_GREATER_ || nssv_COMPILER_MSVC_VER >= 1800)
|
|
#define nssv_CPP11_140 (nssv_CPP11_OR_GREATER_ || nssv_COMPILER_MSVC_VER >= 1900)
|
|
#define nssv_CPP11_141 (nssv_CPP11_OR_GREATER_ || nssv_COMPILER_MSVC_VER >= 1910)
|
|
|
|
#define nssv_CPP14_000 (nssv_CPP14_OR_GREATER)
|
|
#define nssv_CPP17_000 (nssv_CPP17_OR_GREATER)
|
|
|
|
// Presence of C++11 language features:
|
|
|
|
#define nssv_HAVE_CONSTEXPR_11 nssv_CPP11_140
|
|
#define nssv_HAVE_EXPLICIT_CONVERSION nssv_CPP11_140
|
|
#define nssv_HAVE_INLINE_NAMESPACE nssv_CPP11_140
|
|
#define nssv_HAVE_NOEXCEPT nssv_CPP11_140
|
|
#define nssv_HAVE_NULLPTR nssv_CPP11_100
|
|
#define nssv_HAVE_REF_QUALIFIER nssv_CPP11_140
|
|
#define nssv_HAVE_UNICODE_LITERALS nssv_CPP11_140
|
|
#define nssv_HAVE_USER_DEFINED_LITERALS nssv_CPP11_140
|
|
#define nssv_HAVE_WCHAR16_T nssv_CPP11_100
|
|
#define nssv_HAVE_WCHAR32_T nssv_CPP11_100
|
|
|
|
#if ! ( ( nssv_CPP11 && nssv_COMPILER_CLANG_VERSION ) || nssv_BETWEEN( nssv_COMPILER_CLANG_VERSION, 300, 400 ) )
|
|
# define nssv_HAVE_STD_DEFINED_LITERALS nssv_CPP11_140
|
|
#endif
|
|
|
|
// Presence of C++14 language features:
|
|
|
|
#define nssv_HAVE_CONSTEXPR_14 nssv_CPP14_000
|
|
|
|
// Presence of C++17 language features:
|
|
|
|
#define nssv_HAVE_NODISCARD nssv_CPP17_000
|
|
|
|
// Presence of C++ library features:
|
|
|
|
#define nssv_HAVE_STD_HASH nssv_CPP11_120
|
|
|
|
// C++ feature usage:
|
|
|
|
#if nssv_HAVE_CONSTEXPR_11
|
|
# define nssv_constexpr constexpr
|
|
#else
|
|
# define nssv_constexpr /*constexpr*/
|
|
#endif
|
|
|
|
#if nssv_HAVE_CONSTEXPR_14
|
|
# define nssv_constexpr14 constexpr
|
|
#else
|
|
# define nssv_constexpr14 /*constexpr*/
|
|
#endif
|
|
|
|
#if nssv_HAVE_EXPLICIT_CONVERSION
|
|
# define nssv_explicit explicit
|
|
#else
|
|
# define nssv_explicit /*explicit*/
|
|
#endif
|
|
|
|
#if nssv_HAVE_INLINE_NAMESPACE
|
|
# define nssv_inline_ns inline
|
|
#else
|
|
# define nssv_inline_ns /*inline*/
|
|
#endif
|
|
|
|
#if nssv_HAVE_NOEXCEPT
|
|
# define nssv_noexcept noexcept
|
|
#else
|
|
# define nssv_noexcept /*noexcept*/
|
|
#endif
|
|
|
|
//#if nssv_HAVE_REF_QUALIFIER
|
|
//# define nssv_ref_qual &
|
|
//# define nssv_refref_qual &&
|
|
//#else
|
|
//# define nssv_ref_qual /*&*/
|
|
//# define nssv_refref_qual /*&&*/
|
|
//#endif
|
|
|
|
#if nssv_HAVE_NULLPTR
|
|
# define nssv_nullptr nullptr
|
|
#else
|
|
# define nssv_nullptr NULL
|
|
#endif
|
|
|
|
#if nssv_HAVE_NODISCARD
|
|
# define nssv_nodiscard [[nodiscard]]
|
|
#else
|
|
# define nssv_nodiscard /*[[nodiscard]]*/
|
|
#endif
|
|
|
|
// Additional includes:
|
|
|
|
#include <algorithm>
|
|
#include <cassert>
|
|
#include <iterator>
|
|
#include <limits>
|
|
#include <ostream>
|
|
#include <string> // std::char_traits<>
|
|
|
|
#if ! nssv_CONFIG_NO_EXCEPTIONS
|
|
# include <stdexcept>
|
|
#endif
|
|
|
|
#if nssv_CPP11_OR_GREATER
|
|
# include <type_traits>
|
|
#endif
|
|
|
|
// Clang, GNUC, MSVC warning suppression macros:
|
|
|
|
#if defined(__clang__)
|
|
# pragma clang diagnostic ignored "-Wreserved-user-defined-literal"
|
|
# pragma clang diagnostic push
|
|
# pragma clang diagnostic ignored "-Wuser-defined-literals"
|
|
#elif defined(__GNUC__)
|
|
# pragma GCC diagnostic push
|
|
# pragma GCC diagnostic ignored "-Wliteral-suffix"
|
|
#endif // __clang__
|
|
|
|
#if nssv_COMPILER_MSVC_VERSION >= 140
|
|
# define nssv_SUPPRESS_MSGSL_WARNING(expr) [[gsl::suppress(expr)]]
|
|
# define nssv_SUPPRESS_MSVC_WARNING(code, descr) __pragma(warning(suppress: code) )
|
|
# define nssv_DISABLE_MSVC_WARNINGS(codes) __pragma(warning(push)) __pragma(warning(disable: codes))
|
|
#else
|
|
# define nssv_SUPPRESS_MSGSL_WARNING(expr)
|
|
# define nssv_SUPPRESS_MSVC_WARNING(code, descr)
|
|
# define nssv_DISABLE_MSVC_WARNINGS(codes)
|
|
#endif
|
|
|
|
#if defined(__clang__)
|
|
# define nssv_RESTORE_WARNINGS() _Pragma("clang diagnostic pop")
|
|
#elif defined(__GNUC__)
|
|
# define nssv_RESTORE_WARNINGS() _Pragma("GCC diagnostic pop")
|
|
#elif nssv_COMPILER_MSVC_VERSION >= 140
|
|
# define nssv_RESTORE_WARNINGS() __pragma(warning(pop ))
|
|
#else
|
|
# define nssv_RESTORE_WARNINGS()
|
|
#endif
|
|
|
|
// Suppress the following MSVC (GSL) warnings:
|
|
// - C4455, non-gsl : 'operator ""sv': literal suffix identifiers that do not
|
|
// start with an underscore are reserved
|
|
// - C26472, gsl::t.1 : don't use a static_cast for arithmetic conversions;
|
|
// use brace initialization, gsl::narrow_cast or gsl::narow
|
|
// - C26481: gsl::b.1 : don't use pointer arithmetic. Use span instead
|
|
|
|
nssv_DISABLE_MSVC_WARNINGS( 4455 26481 26472 )
|
|
//nssv_DISABLE_CLANG_WARNINGS( "-Wuser-defined-literals" )
|
|
//nssv_DISABLE_GNUC_WARNINGS( -Wliteral-suffix )
|
|
|
|
namespace nonstd { namespace sv_lite {
|
|
|
|
template
|
|
<
|
|
class CharT,
|
|
class Traits = std::char_traits<CharT>
|
|
>
|
|
class basic_string_view;
|
|
|
|
//
|
|
// basic_string_view:
|
|
//
|
|
|
|
template
|
|
<
|
|
class CharT,
|
|
class Traits /* = std::char_traits<CharT> */
|
|
>
|
|
class basic_string_view
|
|
{
|
|
public:
|
|
// Member types:
|
|
|
|
typedef Traits traits_type;
|
|
typedef CharT value_type;
|
|
|
|
typedef CharT * pointer;
|
|
typedef CharT const * const_pointer;
|
|
typedef CharT & reference;
|
|
typedef CharT const & const_reference;
|
|
|
|
typedef const_pointer iterator;
|
|
typedef const_pointer const_iterator;
|
|
typedef std::reverse_iterator< const_iterator > reverse_iterator;
|
|
typedef std::reverse_iterator< const_iterator > const_reverse_iterator;
|
|
|
|
typedef std::size_t size_type;
|
|
typedef std::ptrdiff_t difference_type;
|
|
|
|
// 24.4.2.1 Construction and assignment:
|
|
|
|
nssv_constexpr basic_string_view() nssv_noexcept
|
|
: data_( nssv_nullptr )
|
|
, size_( 0 )
|
|
{}
|
|
|
|
#if nssv_CPP11_OR_GREATER
|
|
nssv_constexpr basic_string_view( basic_string_view const & other ) nssv_noexcept = default;
|
|
#else
|
|
nssv_constexpr basic_string_view( basic_string_view const & other ) nssv_noexcept
|
|
: data_( other.data_)
|
|
, size_( other.size_)
|
|
{}
|
|
#endif
|
|
|
|
nssv_constexpr basic_string_view( CharT const * s, size_type count )
|
|
: data_( s )
|
|
, size_( count )
|
|
{}
|
|
|
|
nssv_constexpr basic_string_view( CharT const * s)
|
|
: data_( s )
|
|
, size_( Traits::length(s) )
|
|
{}
|
|
|
|
// Assignment:
|
|
|
|
#if nssv_CPP11_OR_GREATER
|
|
nssv_constexpr14 basic_string_view & operator=( basic_string_view const & other ) nssv_noexcept = default;
|
|
#else
|
|
nssv_constexpr14 basic_string_view & operator=( basic_string_view const & other ) nssv_noexcept
|
|
{
|
|
data_ = other.data_;
|
|
size_ = other.size_;
|
|
return *this;
|
|
}
|
|
#endif
|
|
|
|
// 24.4.2.2 Iterator support:
|
|
|
|
nssv_constexpr const_iterator begin() const nssv_noexcept { return data_; }
|
|
nssv_constexpr const_iterator end() const nssv_noexcept { return data_ + size_; }
|
|
|
|
nssv_constexpr const_iterator cbegin() const nssv_noexcept { return begin(); }
|
|
nssv_constexpr const_iterator cend() const nssv_noexcept { return end(); }
|
|
|
|
nssv_constexpr const_reverse_iterator rbegin() const nssv_noexcept { return const_reverse_iterator( end() ); }
|
|
nssv_constexpr const_reverse_iterator rend() const nssv_noexcept { return const_reverse_iterator( begin() ); }
|
|
|
|
nssv_constexpr const_reverse_iterator crbegin() const nssv_noexcept { return rbegin(); }
|
|
nssv_constexpr const_reverse_iterator crend() const nssv_noexcept { return rend(); }
|
|
|
|
// 24.4.2.3 Capacity:
|
|
|
|
nssv_constexpr size_type size() const nssv_noexcept { return size_; }
|
|
nssv_constexpr size_type length() const nssv_noexcept { return size_; }
|
|
nssv_constexpr size_type max_size() const nssv_noexcept { return (std::numeric_limits< size_type >::max)(); }
|
|
|
|
// since C++20
|
|
nssv_nodiscard nssv_constexpr bool empty() const nssv_noexcept
|
|
{
|
|
return 0 == size_;
|
|
}
|
|
|
|
// 24.4.2.4 Element access:
|
|
|
|
nssv_constexpr const_reference operator[]( size_type pos ) const
|
|
{
|
|
return data_at( pos );
|
|
}
|
|
|
|
nssv_constexpr14 const_reference at( size_type pos ) const
|
|
{
|
|
#if nssv_CONFIG_NO_EXCEPTIONS
|
|
assert( pos < size() );
|
|
#else
|
|
if ( pos >= size() )
|
|
{
|
|
throw std::out_of_range("nonst::string_view::at()");
|
|
}
|
|
#endif
|
|
return data_at( pos );
|
|
}
|
|
|
|
nssv_constexpr const_reference front() const { return data_at( 0 ); }
|
|
nssv_constexpr const_reference back() const { return data_at( size() - 1 ); }
|
|
|
|
nssv_constexpr const_pointer data() const nssv_noexcept { return data_; }
|
|
|
|
// 24.4.2.5 Modifiers:
|
|
|
|
nssv_constexpr14 void remove_prefix( size_type n )
|
|
{
|
|
assert( n <= size() );
|
|
data_ += n;
|
|
size_ -= n;
|
|
}
|
|
|
|
nssv_constexpr14 void remove_suffix( size_type n )
|
|
{
|
|
assert( n <= size() );
|
|
size_ -= n;
|
|
}
|
|
|
|
nssv_constexpr14 void swap( basic_string_view & other ) nssv_noexcept
|
|
{
|
|
using std::swap;
|
|
swap( data_, other.data_ );
|
|
swap( size_, other.size_ );
|
|
}
|
|
|
|
// 24.4.2.6 String operations:
|
|
|
|
size_type copy( CharT * dest, size_type n, size_type pos = 0 ) const
|
|
{
|
|
#if nssv_CONFIG_NO_EXCEPTIONS
|
|
assert( pos <= size() );
|
|
#else
|
|
if ( pos > size() )
|
|
{
|
|
throw std::out_of_range("nonst::string_view::copy()");
|
|
}
|
|
#endif
|
|
const size_type rlen = (std::min)( n, size() - pos );
|
|
|
|
(void) Traits::copy( dest, data() + pos, rlen );
|
|
|
|
return rlen;
|
|
}
|
|
|
|
nssv_constexpr14 basic_string_view substr( size_type pos = 0, size_type n = npos ) const
|
|
{
|
|
#if nssv_CONFIG_NO_EXCEPTIONS
|
|
assert( pos <= size() );
|
|
#else
|
|
if ( pos > size() )
|
|
{
|
|
throw std::out_of_range("nonst::string_view::substr()");
|
|
}
|
|
#endif
|
|
return basic_string_view( data() + pos, (std::min)( n, size() - pos ) );
|
|
}
|
|
|
|
// compare(), 6x:
|
|
|
|
nssv_constexpr14 int compare( basic_string_view other ) const nssv_noexcept // (1)
|
|
{
|
|
if ( const int result = Traits::compare( data(), other.data(), (std::min)( size(), other.size() ) ) )
|
|
return result;
|
|
|
|
return size() == other.size() ? 0 : size() < other.size() ? -1 : 1;
|
|
}
|
|
|
|
nssv_constexpr int compare( size_type pos1, size_type n1, basic_string_view other ) const // (2)
|
|
{
|
|
return substr( pos1, n1 ).compare( other );
|
|
}
|
|
|
|
nssv_constexpr int compare( size_type pos1, size_type n1, basic_string_view other, size_type pos2, size_type n2 ) const // (3)
|
|
{
|
|
return substr( pos1, n1 ).compare( other.substr( pos2, n2 ) );
|
|
}
|
|
|
|
nssv_constexpr int compare( CharT const * s ) const // (4)
|
|
{
|
|
return compare( basic_string_view( s ) );
|
|
}
|
|
|
|
nssv_constexpr int compare( size_type pos1, size_type n1, CharT const * s ) const // (5)
|
|
{
|
|
return substr( pos1, n1 ).compare( basic_string_view( s ) );
|
|
}
|
|
|
|
nssv_constexpr int compare( size_type pos1, size_type n1, CharT const * s, size_type n2 ) const // (6)
|
|
{
|
|
return substr( pos1, n1 ).compare( basic_string_view( s, n2 ) );
|
|
}
|
|
|
|
// 24.4.2.7 Searching:
|
|
|
|
// starts_with(), 3x, since C++20:
|
|
|
|
nssv_constexpr bool starts_with( basic_string_view v ) const nssv_noexcept // (1)
|
|
{
|
|
return size() >= v.size() && compare( 0, v.size(), v ) == 0;
|
|
}
|
|
|
|
nssv_constexpr bool starts_with( CharT c ) const nssv_noexcept // (2)
|
|
{
|
|
return starts_with( basic_string_view( &c, 1 ) );
|
|
}
|
|
|
|
nssv_constexpr bool starts_with( CharT const * s ) const // (3)
|
|
{
|
|
return starts_with( basic_string_view( s ) );
|
|
}
|
|
|
|
// ends_with(), 3x, since C++20:
|
|
|
|
nssv_constexpr bool ends_with( basic_string_view v ) const nssv_noexcept // (1)
|
|
{
|
|
return size() >= v.size() && compare( size() - v.size(), npos, v ) == 0;
|
|
}
|
|
|
|
nssv_constexpr bool ends_with( CharT c ) const nssv_noexcept // (2)
|
|
{
|
|
return ends_with( basic_string_view( &c, 1 ) );
|
|
}
|
|
|
|
nssv_constexpr bool ends_with( CharT const * s ) const // (3)
|
|
{
|
|
return ends_with( basic_string_view( s ) );
|
|
}
|
|
|
|
// find(), 4x:
|
|
|
|
nssv_constexpr14 size_type find( basic_string_view v, size_type pos = 0 ) const nssv_noexcept // (1)
|
|
{
|
|
return assert( v.size() == 0 || v.data() != nssv_nullptr )
|
|
, pos >= size()
|
|
? npos
|
|
: to_pos( std::search( cbegin() + pos, cend(), v.cbegin(), v.cend(), Traits::eq ) );
|
|
}
|
|
|
|
nssv_constexpr14 size_type find( CharT c, size_type pos = 0 ) const nssv_noexcept // (2)
|
|
{
|
|
return find( basic_string_view( &c, 1 ), pos );
|
|
}
|
|
|
|
nssv_constexpr14 size_type find( CharT const * s, size_type pos, size_type n ) const // (3)
|
|
{
|
|
return find( basic_string_view( s, n ), pos );
|
|
}
|
|
|
|
nssv_constexpr14 size_type find( CharT const * s, size_type pos = 0 ) const // (4)
|
|
{
|
|
return find( basic_string_view( s ), pos );
|
|
}
|
|
|
|
// rfind(), 4x:
|
|
|
|
nssv_constexpr14 size_type rfind( basic_string_view v, size_type pos = npos ) const nssv_noexcept // (1)
|
|
{
|
|
if ( size() < v.size() )
|
|
return npos;
|
|
|
|
if ( v.empty() )
|
|
return (std::min)( size(), pos );
|
|
|
|
const_iterator last = cbegin() + (std::min)( size() - v.size(), pos ) + v.size();
|
|
const_iterator result = std::find_end( cbegin(), last, v.cbegin(), v.cend(), Traits::eq );
|
|
|
|
return result != last ? size_type( result - cbegin() ) : npos;
|
|
}
|
|
|
|
nssv_constexpr14 size_type rfind( CharT c, size_type pos = npos ) const nssv_noexcept // (2)
|
|
{
|
|
return rfind( basic_string_view( &c, 1 ), pos );
|
|
}
|
|
|
|
nssv_constexpr14 size_type rfind( CharT const * s, size_type pos, size_type n ) const // (3)
|
|
{
|
|
return rfind( basic_string_view( s, n ), pos );
|
|
}
|
|
|
|
nssv_constexpr14 size_type rfind( CharT const * s, size_type pos = npos ) const // (4)
|
|
{
|
|
return rfind( basic_string_view( s ), pos );
|
|
}
|
|
|
|
// find_first_of(), 4x:
|
|
|
|
nssv_constexpr size_type find_first_of( basic_string_view v, size_type pos = 0 ) const nssv_noexcept // (1)
|
|
{
|
|
return pos >= size()
|
|
? npos
|
|
: to_pos( std::find_first_of( cbegin() + pos, cend(), v.cbegin(), v.cend(), Traits::eq ) );
|
|
}
|
|
|
|
nssv_constexpr size_type find_first_of( CharT c, size_type pos = 0 ) const nssv_noexcept // (2)
|
|
{
|
|
return find_first_of( basic_string_view( &c, 1 ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_first_of( CharT const * s, size_type pos, size_type n ) const // (3)
|
|
{
|
|
return find_first_of( basic_string_view( s, n ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_first_of( CharT const * s, size_type pos = 0 ) const // (4)
|
|
{
|
|
return find_first_of( basic_string_view( s ), pos );
|
|
}
|
|
|
|
// find_last_of(), 4x:
|
|
|
|
nssv_constexpr size_type find_last_of( basic_string_view v, size_type pos = npos ) const nssv_noexcept // (1)
|
|
{
|
|
return empty()
|
|
? npos
|
|
: pos >= size()
|
|
? find_last_of( v, size() - 1 )
|
|
: to_pos( std::find_first_of( const_reverse_iterator( cbegin() + pos + 1 ), crend(), v.cbegin(), v.cend(), Traits::eq ) );
|
|
}
|
|
|
|
nssv_constexpr size_type find_last_of( CharT c, size_type pos = npos ) const nssv_noexcept // (2)
|
|
{
|
|
return find_last_of( basic_string_view( &c, 1 ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_last_of( CharT const * s, size_type pos, size_type count ) const // (3)
|
|
{
|
|
return find_last_of( basic_string_view( s, count ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_last_of( CharT const * s, size_type pos = npos ) const // (4)
|
|
{
|
|
return find_last_of( basic_string_view( s ), pos );
|
|
}
|
|
|
|
// find_first_not_of(), 4x:
|
|
|
|
nssv_constexpr size_type find_first_not_of( basic_string_view v, size_type pos = 0 ) const nssv_noexcept // (1)
|
|
{
|
|
return pos >= size()
|
|
? npos
|
|
: to_pos( std::find_if( cbegin() + pos, cend(), not_in_view( v ) ) );
|
|
}
|
|
|
|
nssv_constexpr size_type find_first_not_of( CharT c, size_type pos = 0 ) const nssv_noexcept // (2)
|
|
{
|
|
return find_first_not_of( basic_string_view( &c, 1 ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_first_not_of( CharT const * s, size_type pos, size_type count ) const // (3)
|
|
{
|
|
return find_first_not_of( basic_string_view( s, count ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_first_not_of( CharT const * s, size_type pos = 0 ) const // (4)
|
|
{
|
|
return find_first_not_of( basic_string_view( s ), pos );
|
|
}
|
|
|
|
// find_last_not_of(), 4x:
|
|
|
|
nssv_constexpr size_type find_last_not_of( basic_string_view v, size_type pos = npos ) const nssv_noexcept // (1)
|
|
{
|
|
return empty()
|
|
? npos
|
|
: pos >= size()
|
|
? find_last_not_of( v, size() - 1 )
|
|
: to_pos( std::find_if( const_reverse_iterator( cbegin() + pos + 1 ), crend(), not_in_view( v ) ) );
|
|
}
|
|
|
|
nssv_constexpr size_type find_last_not_of( CharT c, size_type pos = npos ) const nssv_noexcept // (2)
|
|
{
|
|
return find_last_not_of( basic_string_view( &c, 1 ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_last_not_of( CharT const * s, size_type pos, size_type count ) const // (3)
|
|
{
|
|
return find_last_not_of( basic_string_view( s, count ), pos );
|
|
}
|
|
|
|
nssv_constexpr size_type find_last_not_of( CharT const * s, size_type pos = npos ) const // (4)
|
|
{
|
|
return find_last_not_of( basic_string_view( s ), pos );
|
|
}
|
|
|
|
// Constants:
|
|
|
|
#if nssv_CPP17_OR_GREATER
|
|
static nssv_constexpr size_type npos = size_type(-1);
|
|
#elif nssv_CPP11_OR_GREATER
|
|
enum : size_type { npos = size_type(-1) };
|
|
#else
|
|
enum { npos = size_type(-1) };
|
|
#endif
|
|
|
|
private:
|
|
struct not_in_view
|
|
{
|
|
const basic_string_view v;
|
|
|
|
nssv_constexpr not_in_view( basic_string_view v ) : v( v ) {}
|
|
|
|
nssv_constexpr bool operator()( CharT c ) const
|
|
{
|
|
return npos == v.find_first_of( c );
|
|
}
|
|
};
|
|
|
|
nssv_constexpr size_type to_pos( const_iterator it ) const
|
|
{
|
|
return it == cend() ? npos : size_type( it - cbegin() );
|
|
}
|
|
|
|
nssv_constexpr size_type to_pos( const_reverse_iterator it ) const
|
|
{
|
|
return it == crend() ? npos : size_type( crend() - it - 1 );
|
|
}
|
|
|
|
nssv_constexpr const_reference data_at( size_type pos ) const
|
|
{
|
|
#if nssv_BETWEEN( nssv_COMPILER_GNUC_VERSION, 1, 500 )
|
|
return data_[pos];
|
|
#else
|
|
return assert( pos < size() ), data_[pos];
|
|
#endif
|
|
}
|
|
|
|
private:
|
|
const_pointer data_;
|
|
size_type size_;
|
|
|
|
public:
|
|
#if nssv_CONFIG_CONVERSION_STD_STRING_CLASS_METHODS
|
|
|
|
template< class Allocator >
|
|
basic_string_view( std::basic_string<CharT, Traits, Allocator> const & s ) nssv_noexcept
|
|
: data_( s.data() )
|
|
, size_( s.size() )
|
|
{}
|
|
|
|
#if nssv_HAVE_EXPLICIT_CONVERSION
|
|
|
|
template< class Allocator >
|
|
explicit operator std::basic_string<CharT, Traits, Allocator>() const
|
|
{
|
|
return to_string( Allocator() );
|
|
}
|
|
|
|
#endif // nssv_HAVE_EXPLICIT_CONVERSION
|
|
|
|
#if nssv_CPP11_OR_GREATER
|
|
|
|
template< class Allocator = std::allocator<CharT> >
|
|
std::basic_string<CharT, Traits, Allocator>
|
|
to_string( Allocator const & a = Allocator() ) const
|
|
{
|
|
return std::basic_string<CharT, Traits, Allocator>( begin(), end(), a );
|
|
}
|
|
|
|
#else
|
|
|
|
std::basic_string<CharT, Traits>
|
|
to_string() const
|
|
{
|
|
return std::basic_string<CharT, Traits>( begin(), end() );
|
|
}
|
|
|
|
template< class Allocator >
|
|
std::basic_string<CharT, Traits, Allocator>
|
|
to_string( Allocator const & a ) const
|
|
{
|
|
return std::basic_string<CharT, Traits, Allocator>( begin(), end(), a );
|
|
}
|
|
|
|
#endif // nssv_CPP11_OR_GREATER
|
|
|
|
#endif // nssv_CONFIG_CONVERSION_STD_STRING_CLASS_METHODS
|
|
};
|
|
|
|
//
|
|
// Non-member functions:
|
|
//
|
|
|
|
// 24.4.3 Non-member comparison functions:
|
|
// lexicographically compare two string views (function template):
|
|
|
|
template< class CharT, class Traits >
|
|
nssv_constexpr bool operator== (
|
|
basic_string_view <CharT, Traits> lhs,
|
|
basic_string_view <CharT, Traits> rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) == 0 ; }
|
|
|
|
template< class CharT, class Traits >
|
|
nssv_constexpr bool operator!= (
|
|
basic_string_view <CharT, Traits> lhs,
|
|
basic_string_view <CharT, Traits> rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) != 0 ; }
|
|
|
|
template< class CharT, class Traits >
|
|
nssv_constexpr bool operator< (
|
|
basic_string_view <CharT, Traits> lhs,
|
|
basic_string_view <CharT, Traits> rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) < 0 ; }
|
|
|
|
template< class CharT, class Traits >
|
|
nssv_constexpr bool operator<= (
|
|
basic_string_view <CharT, Traits> lhs,
|
|
basic_string_view <CharT, Traits> rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) <= 0 ; }
|
|
|
|
template< class CharT, class Traits >
|
|
nssv_constexpr bool operator> (
|
|
basic_string_view <CharT, Traits> lhs,
|
|
basic_string_view <CharT, Traits> rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) > 0 ; }
|
|
|
|
template< class CharT, class Traits >
|
|
nssv_constexpr bool operator>= (
|
|
basic_string_view <CharT, Traits> lhs,
|
|
basic_string_view <CharT, Traits> rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) >= 0 ; }
|
|
|
|
// Let S be basic_string_view<CharT, Traits>, and sv be an instance of S.
|
|
// Implementations shall provide sufficient additional overloads marked
|
|
// constexpr and noexcept so that an object t with an implicit conversion
|
|
// to S can be compared according to Table 67.
|
|
|
|
#if nssv_CPP11_OR_GREATER && ! nssv_BETWEEN( nssv_COMPILER_MSVC_VERSION, 100, 141 )
|
|
|
|
#define nssv_BASIC_STRING_VIEW_I(T,U) typename std::decay< basic_string_view<T,U> >::type
|
|
|
|
#if nssv_BETWEEN( nssv_COMPILER_MSVC_VERSION, 140, 150 )
|
|
# define nssv_MSVC_ORDER(x) , int=x
|
|
#else
|
|
# define nssv_MSVC_ORDER(x) /*, int=x*/
|
|
#endif
|
|
|
|
// ==
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(1) >
|
|
nssv_constexpr bool operator==(
|
|
basic_string_view <CharT, Traits> lhs,
|
|
nssv_BASIC_STRING_VIEW_I(CharT, Traits) rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) == 0; }
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(2) >
|
|
nssv_constexpr bool operator==(
|
|
nssv_BASIC_STRING_VIEW_I(CharT, Traits) lhs,
|
|
basic_string_view <CharT, Traits> rhs ) nssv_noexcept
|
|
{ return lhs.size() == rhs.size() && lhs.compare( rhs ) == 0; }
|
|
|
|
// !=
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(1) >
|
|
nssv_constexpr bool operator!= (
|
|
basic_string_view < CharT, Traits > lhs,
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) rhs ) nssv_noexcept
|
|
{ return lhs.size() != rhs.size() || lhs.compare( rhs ) != 0 ; }
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(2) >
|
|
nssv_constexpr bool operator!= (
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) lhs,
|
|
basic_string_view < CharT, Traits > rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) != 0 ; }
|
|
|
|
// <
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(1) >
|
|
nssv_constexpr bool operator< (
|
|
basic_string_view < CharT, Traits > lhs,
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) < 0 ; }
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(2) >
|
|
nssv_constexpr bool operator< (
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) lhs,
|
|
basic_string_view < CharT, Traits > rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) < 0 ; }
|
|
|
|
// <=
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(1) >
|
|
nssv_constexpr bool operator<= (
|
|
basic_string_view < CharT, Traits > lhs,
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) <= 0 ; }
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(2) >
|
|
nssv_constexpr bool operator<= (
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) lhs,
|
|
basic_string_view < CharT, Traits > rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) <= 0 ; }
|
|
|
|
// >
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(1) >
|
|
nssv_constexpr bool operator> (
|
|
basic_string_view < CharT, Traits > lhs,
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) > 0 ; }
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(2) >
|
|
nssv_constexpr bool operator> (
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) lhs,
|
|
basic_string_view < CharT, Traits > rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) > 0 ; }
|
|
|
|
// >=
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(1) >
|
|
nssv_constexpr bool operator>= (
|
|
basic_string_view < CharT, Traits > lhs,
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) >= 0 ; }
|
|
|
|
template< class CharT, class Traits nssv_MSVC_ORDER(2) >
|
|
nssv_constexpr bool operator>= (
|
|
nssv_BASIC_STRING_VIEW_I( CharT, Traits ) lhs,
|
|
basic_string_view < CharT, Traits > rhs ) nssv_noexcept
|
|
{ return lhs.compare( rhs ) >= 0 ; }
|
|
|
|
#undef nssv_MSVC_ORDER
|
|
#undef nssv_BASIC_STRING_VIEW_I
|
|
|
|
#endif // nssv_CPP11_OR_GREATER
|
|
|
|
// 24.4.4 Inserters and extractors:
|
|
|
|
namespace detail {
|
|
|
|
template< class Stream >
|
|
void write_padding( Stream & os, std::streamsize n )
|
|
{
|
|
for ( std::streamsize i = 0; i < n; ++i )
|
|
os.rdbuf()->sputc( os.fill() );
|
|
}
|
|
|
|
template< class Stream, class View >
|
|
Stream & write_to_stream( Stream & os, View const & sv )
|
|
{
|
|
typename Stream::sentry sentry( os );
|
|
|
|
if ( !os )
|
|
return os;
|
|
|
|
const std::streamsize length = static_cast<std::streamsize>( sv.length() );
|
|
|
|
// Whether, and how, to pad:
|
|
const bool pad = ( length < os.width() );
|
|
const bool left_pad = pad && ( os.flags() & std::ios_base::adjustfield ) == std::ios_base::right;
|
|
|
|
if ( left_pad )
|
|
write_padding( os, os.width() - length );
|
|
|
|
// Write span characters:
|
|
os.rdbuf()->sputn( sv.begin(), length );
|
|
|
|
if ( pad && !left_pad )
|
|
write_padding( os, os.width() - length );
|
|
|
|
// Reset output stream width:
|
|
os.width( 0 );
|
|
|
|
return os;
|
|
}
|
|
|
|
} // namespace detail
|
|
|
|
template< class CharT, class Traits >
|
|
std::basic_ostream<CharT, Traits> &
|
|
operator<<(
|
|
std::basic_ostream<CharT, Traits>& os,
|
|
basic_string_view <CharT, Traits> sv )
|
|
{
|
|
return detail::write_to_stream( os, sv );
|
|
}
|
|
|
|
// Several typedefs for common character types are provided:
|
|
|
|
typedef basic_string_view<char> string_view;
|
|
typedef basic_string_view<wchar_t> wstring_view;
|
|
#if nssv_HAVE_WCHAR16_T
|
|
typedef basic_string_view<char16_t> u16string_view;
|
|
typedef basic_string_view<char32_t> u32string_view;
|
|
#endif
|
|
|
|
}} // namespace nonstd::sv_lite
|
|
|
|
//
|
|
// 24.4.6 Suffix for basic_string_view literals:
|
|
//
|
|
|
|
#if nssv_HAVE_USER_DEFINED_LITERALS
|
|
|
|
namespace nonstd {
|
|
nssv_inline_ns namespace literals {
|
|
nssv_inline_ns namespace string_view_literals {
|
|
|
|
#if nssv_CONFIG_STD_SV_OPERATOR && nssv_HAVE_STD_DEFINED_LITERALS
|
|
|
|
nssv_constexpr nonstd::sv_lite::string_view operator "" sv( const char* str, size_t len ) nssv_noexcept // (1)
|
|
{
|
|
return nonstd::sv_lite::string_view{ str, len };
|
|
}
|
|
|
|
nssv_constexpr nonstd::sv_lite::u16string_view operator "" sv( const char16_t* str, size_t len ) nssv_noexcept // (2)
|
|
{
|
|
return nonstd::sv_lite::u16string_view{ str, len };
|
|
}
|
|
|
|
nssv_constexpr nonstd::sv_lite::u32string_view operator "" sv( const char32_t* str, size_t len ) nssv_noexcept // (3)
|
|
{
|
|
return nonstd::sv_lite::u32string_view{ str, len };
|
|
}
|
|
|
|
nssv_constexpr nonstd::sv_lite::wstring_view operator "" sv( const wchar_t* str, size_t len ) nssv_noexcept // (4)
|
|
{
|
|
return nonstd::sv_lite::wstring_view{ str, len };
|
|
}
|
|
|
|
#endif // nssv_CONFIG_STD_SV_OPERATOR && nssv_HAVE_STD_DEFINED_LITERALS
|
|
|
|
#if nssv_CONFIG_USR_SV_OPERATOR
|
|
|
|
nssv_constexpr nonstd::sv_lite::string_view operator "" _sv( const char* str, size_t len ) nssv_noexcept // (1)
|
|
{
|
|
return nonstd::sv_lite::string_view{ str, len };
|
|
}
|
|
|
|
nssv_constexpr nonstd::sv_lite::u16string_view operator "" _sv( const char16_t* str, size_t len ) nssv_noexcept // (2)
|
|
{
|
|
return nonstd::sv_lite::u16string_view{ str, len };
|
|
}
|
|
|
|
nssv_constexpr nonstd::sv_lite::u32string_view operator "" _sv( const char32_t* str, size_t len ) nssv_noexcept // (3)
|
|
{
|
|
return nonstd::sv_lite::u32string_view{ str, len };
|
|
}
|
|
|
|
nssv_constexpr nonstd::sv_lite::wstring_view operator "" _sv( const wchar_t* str, size_t len ) nssv_noexcept // (4)
|
|
{
|
|
return nonstd::sv_lite::wstring_view{ str, len };
|
|
}
|
|
|
|
#endif // nssv_CONFIG_USR_SV_OPERATOR
|
|
|
|
}}} // namespace nonstd::literals::string_view_literals
|
|
|
|
#endif
|
|
|
|
//
|
|
// Extensions for std::string:
|
|
//
|
|
|
|
#if nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS
|
|
|
|
namespace nonstd {
|
|
namespace sv_lite {
|
|
|
|
// Exclude MSVC 14 (19.00): it yields ambiguous to_string():
|
|
|
|
#if nssv_CPP11_OR_GREATER && nssv_COMPILER_MSVC_VERSION != 140
|
|
|
|
template< class CharT, class Traits, class Allocator = std::allocator<CharT> >
|
|
std::basic_string<CharT, Traits, Allocator>
|
|
to_string( basic_string_view<CharT, Traits> v, Allocator const & a = Allocator() )
|
|
{
|
|
return std::basic_string<CharT,Traits, Allocator>( v.begin(), v.end(), a );
|
|
}
|
|
|
|
#else
|
|
|
|
template< class CharT, class Traits >
|
|
std::basic_string<CharT, Traits>
|
|
to_string( basic_string_view<CharT, Traits> v )
|
|
{
|
|
return std::basic_string<CharT, Traits>( v.begin(), v.end() );
|
|
}
|
|
|
|
template< class CharT, class Traits, class Allocator >
|
|
std::basic_string<CharT, Traits, Allocator>
|
|
to_string( basic_string_view<CharT, Traits> v, Allocator const & a )
|
|
{
|
|
return std::basic_string<CharT, Traits, Allocator>( v.begin(), v.end(), a );
|
|
}
|
|
|
|
#endif // nssv_CPP11_OR_GREATER
|
|
|
|
template< class CharT, class Traits, class Allocator >
|
|
basic_string_view<CharT, Traits>
|
|
to_string_view( std::basic_string<CharT, Traits, Allocator> const & s )
|
|
{
|
|
return basic_string_view<CharT, Traits>( s.data(), s.size() );
|
|
}
|
|
|
|
}} // namespace nonstd::sv_lite
|
|
|
|
#endif // nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS
|
|
|
|
//
|
|
// make types and algorithms available in namespace nonstd:
|
|
//
|
|
|
|
namespace nonstd {
|
|
|
|
using sv_lite::basic_string_view;
|
|
using sv_lite::string_view;
|
|
using sv_lite::wstring_view;
|
|
|
|
#if nssv_HAVE_WCHAR16_T
|
|
using sv_lite::u16string_view;
|
|
#endif
|
|
#if nssv_HAVE_WCHAR32_T
|
|
using sv_lite::u32string_view;
|
|
#endif
|
|
|
|
// literal "sv"
|
|
|
|
using sv_lite::operator==;
|
|
using sv_lite::operator!=;
|
|
using sv_lite::operator<;
|
|
using sv_lite::operator<=;
|
|
using sv_lite::operator>;
|
|
using sv_lite::operator>=;
|
|
|
|
using sv_lite::operator<<;
|
|
|
|
#if nssv_CONFIG_CONVERSION_STD_STRING_FREE_FUNCTIONS
|
|
using sv_lite::to_string;
|
|
using sv_lite::to_string_view;
|
|
#endif
|
|
|
|
} // namespace nonstd
|
|
|
|
// 24.4.5 Hash support (C++11):
|
|
|
|
// Note: The hash value of a string view object is equal to the hash value of
|
|
// the corresponding string object.
|
|
|
|
#if nssv_HAVE_STD_HASH
|
|
|
|
#include <functional>
|
|
|
|
namespace std {
|
|
|
|
template<>
|
|
struct hash< nonstd::string_view >
|
|
{
|
|
public:
|
|
std::size_t operator()( nonstd::string_view v ) const nssv_noexcept
|
|
{
|
|
return std::hash<std::string>()( std::string( v.data(), v.size() ) );
|
|
}
|
|
};
|
|
|
|
template<>
|
|
struct hash< nonstd::wstring_view >
|
|
{
|
|
public:
|
|
std::size_t operator()( nonstd::wstring_view v ) const nssv_noexcept
|
|
{
|
|
return std::hash<std::wstring>()( std::wstring( v.data(), v.size() ) );
|
|
}
|
|
};
|
|
|
|
template<>
|
|
struct hash< nonstd::u16string_view >
|
|
{
|
|
public:
|
|
std::size_t operator()( nonstd::u16string_view v ) const nssv_noexcept
|
|
{
|
|
return std::hash<std::u16string>()( std::u16string( v.data(), v.size() ) );
|
|
}
|
|
};
|
|
|
|
template<>
|
|
struct hash< nonstd::u32string_view >
|
|
{
|
|
public:
|
|
std::size_t operator()( nonstd::u32string_view v ) const nssv_noexcept
|
|
{
|
|
return std::hash<std::u32string>()( std::u32string( v.data(), v.size() ) );
|
|
}
|
|
};
|
|
|
|
} // namespace std
|
|
|
|
#endif // nssv_HAVE_STD_HASH
|
|
|
|
nssv_RESTORE_WARNINGS()
|
|
|
|
#endif // nssv_HAVE_STD_STRING_VIEW
|
|
#endif // NONSTD_SV_LITE_H_INCLUDED
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
enum class ElementNotation {
|
|
Dot,
|
|
Pointer
|
|
};
|
|
|
|
/*!
|
|
* \brief Class for lexer configuration.
|
|
*/
|
|
struct LexerConfig {
|
|
std::string statement_open {"{%"};
|
|
std::string statement_close {"%}"};
|
|
std::string line_statement {"##"};
|
|
std::string expression_open {"{{"};
|
|
std::string expression_close {"}}"};
|
|
std::string comment_open {"{#"};
|
|
std::string comment_close {"#}"};
|
|
std::string open_chars {"#{"};
|
|
|
|
bool trim_blocks {false};
|
|
bool lstrip_blocks {false};
|
|
|
|
void update_open_chars() {
|
|
open_chars = "";
|
|
if (open_chars.find(line_statement[0]) == std::string::npos) {
|
|
open_chars += line_statement[0];
|
|
}
|
|
if (open_chars.find(statement_open[0]) == std::string::npos) {
|
|
open_chars += statement_open[0];
|
|
}
|
|
if (open_chars.find(expression_open[0]) == std::string::npos) {
|
|
open_chars += expression_open[0];
|
|
}
|
|
if (open_chars.find(comment_open[0]) == std::string::npos) {
|
|
open_chars += comment_open[0];
|
|
}
|
|
}
|
|
};
|
|
|
|
/*!
|
|
* \brief Class for parser configuration.
|
|
*/
|
|
struct ParserConfig {
|
|
ElementNotation notation {ElementNotation::Dot};
|
|
};
|
|
|
|
}
|
|
|
|
#endif // PANTOR_INJA_CONFIG_HPP
|
|
|
|
// #include "function_storage.hpp"
|
|
#ifndef PANTOR_INJA_FUNCTION_STORAGE_HPP
|
|
#define PANTOR_INJA_FUNCTION_STORAGE_HPP
|
|
|
|
#include <vector>
|
|
|
|
// #include "bytecode.hpp"
|
|
#ifndef PANTOR_INJA_BYTECODE_HPP
|
|
#define PANTOR_INJA_BYTECODE_HPP
|
|
|
|
#include <string>
|
|
#include <utility>
|
|
|
|
#include <nlohmann/json.hpp>
|
|
|
|
// #include "string_view.hpp"
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
using json = nlohmann::json;
|
|
|
|
|
|
struct Bytecode {
|
|
enum class Op : uint8_t {
|
|
Nop,
|
|
// print StringRef (always immediate)
|
|
PrintText,
|
|
// print value
|
|
PrintValue,
|
|
// push value onto stack (always immediate)
|
|
Push,
|
|
|
|
// builtin functions
|
|
// result is pushed to stack
|
|
// args specify number of arguments
|
|
// all functions can take their "last" argument either immediate
|
|
// or popped off stack (e.g. if immediate, it's like the immediate was
|
|
// just pushed to the stack)
|
|
Not,
|
|
And,
|
|
Or,
|
|
In,
|
|
Equal,
|
|
Greater,
|
|
GreaterEqual,
|
|
Less,
|
|
LessEqual,
|
|
At,
|
|
Different,
|
|
DivisibleBy,
|
|
Even,
|
|
First,
|
|
Float,
|
|
Int,
|
|
Last,
|
|
Length,
|
|
Lower,
|
|
Max,
|
|
Min,
|
|
Odd,
|
|
Range,
|
|
Result,
|
|
Round,
|
|
Sort,
|
|
Upper,
|
|
Exists,
|
|
ExistsInObject,
|
|
IsBoolean,
|
|
IsNumber,
|
|
IsInteger,
|
|
IsFloat,
|
|
IsObject,
|
|
IsArray,
|
|
IsString,
|
|
Default,
|
|
|
|
// include another template
|
|
// value is the template name
|
|
Include,
|
|
|
|
// callback function
|
|
// str is the function name (this means it cannot be a lookup)
|
|
// args specify number of arguments
|
|
// as with builtin functions, "last" argument can be immediate
|
|
Callback,
|
|
|
|
// unconditional jump
|
|
// args is the index of the bytecode to jump to.
|
|
Jump,
|
|
|
|
// conditional jump
|
|
// value popped off stack is checked for truthyness
|
|
// if false, args is the index of the bytecode to jump to.
|
|
// if true, no action is taken (falls through)
|
|
ConditionalJump,
|
|
|
|
// start loop
|
|
// value popped off stack is what is iterated over
|
|
// args is index of bytecode after end loop (jumped to if iterable is
|
|
// empty)
|
|
// immediate value is key name (for maps)
|
|
// str is value name
|
|
StartLoop,
|
|
|
|
// end a loop
|
|
// args is index of the first bytecode in the loop body
|
|
EndLoop,
|
|
};
|
|
|
|
enum Flag {
|
|
// location of value for value-taking ops (mask)
|
|
ValueMask = 0x03,
|
|
// pop value off stack
|
|
ValuePop = 0x00,
|
|
// value is immediate rather than on stack
|
|
ValueImmediate = 0x01,
|
|
// lookup immediate str (dot notation)
|
|
ValueLookupDot = 0x02,
|
|
// lookup immediate str (json pointer notation)
|
|
ValueLookupPointer = 0x03,
|
|
};
|
|
|
|
Op op {Op::Nop};
|
|
uint32_t args: 30;
|
|
uint32_t flags: 2;
|
|
|
|
json value;
|
|
std::string str;
|
|
|
|
Bytecode(): args(0), flags(0) {}
|
|
explicit Bytecode(Op op, unsigned int args = 0): op(op), args(args), flags(0) {}
|
|
explicit Bytecode(Op op, nonstd::string_view str, unsigned int flags): op(op), args(0), flags(flags), str(str) {}
|
|
explicit Bytecode(Op op, json&& value, unsigned int flags): op(op), args(0), flags(flags), value(std::move(value)) {}
|
|
};
|
|
|
|
} // namespace inja
|
|
|
|
#endif // PANTOR_INJA_BYTECODE_HPP
|
|
|
|
// #include "string_view.hpp"
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
using namespace nlohmann;
|
|
|
|
using Arguments = std::vector<const json*>;
|
|
using CallbackFunction = std::function<json(Arguments& args)>;
|
|
|
|
/*!
|
|
* \brief Class for builtin functions and user-defined callbacks.
|
|
*/
|
|
class FunctionStorage {
|
|
public:
|
|
void add_builtin(nonstd::string_view name, unsigned int num_args, Bytecode::Op op) {
|
|
auto& data = get_or_new(name, num_args);
|
|
data.op = op;
|
|
}
|
|
|
|
void add_callback(nonstd::string_view name, unsigned int num_args, const CallbackFunction& function) {
|
|
auto& data = get_or_new(name, num_args);
|
|
data.function = function;
|
|
}
|
|
|
|
Bytecode::Op find_builtin(nonstd::string_view name, unsigned int num_args) const {
|
|
if (auto ptr = get(name, num_args)) {
|
|
return ptr->op;
|
|
}
|
|
return Bytecode::Op::Nop;
|
|
}
|
|
|
|
CallbackFunction find_callback(nonstd::string_view name, unsigned int num_args) const {
|
|
if (auto ptr = get(name, num_args)) {
|
|
return ptr->function;
|
|
}
|
|
return nullptr;
|
|
}
|
|
|
|
private:
|
|
struct FunctionData {
|
|
unsigned int num_args {0};
|
|
Bytecode::Op op {Bytecode::Op::Nop}; // for builtins
|
|
CallbackFunction function; // for callbacks
|
|
};
|
|
|
|
FunctionData& get_or_new(nonstd::string_view name, unsigned int num_args) {
|
|
auto &vec = m_map[static_cast<std::string>(name)];
|
|
for (auto &i: vec) {
|
|
if (i.num_args == num_args) return i;
|
|
}
|
|
vec.emplace_back();
|
|
vec.back().num_args = num_args;
|
|
return vec.back();
|
|
}
|
|
|
|
const FunctionData* get(nonstd::string_view name, unsigned int num_args) const {
|
|
auto it = m_map.find(static_cast<std::string>(name));
|
|
if (it == m_map.end()) return nullptr;
|
|
for (auto &&i: it->second) {
|
|
if (i.num_args == num_args) return &i;
|
|
}
|
|
return nullptr;
|
|
}
|
|
|
|
std::map<std::string, std::vector<FunctionData>> m_map;
|
|
};
|
|
|
|
}
|
|
|
|
#endif // PANTOR_INJA_FUNCTION_STORAGE_HPP
|
|
|
|
// #include "parser.hpp"
|
|
#ifndef PANTOR_INJA_PARSER_HPP
|
|
#define PANTOR_INJA_PARSER_HPP
|
|
|
|
#include <limits>
|
|
#include <string>
|
|
#include <utility>
|
|
#include <vector>
|
|
|
|
// #include "bytecode.hpp"
|
|
|
|
// #include "config.hpp"
|
|
|
|
// #include "function_storage.hpp"
|
|
|
|
// #include "lexer.hpp"
|
|
#ifndef PANTOR_INJA_LEXER_HPP
|
|
#define PANTOR_INJA_LEXER_HPP
|
|
|
|
#include <cctype>
|
|
#include <locale>
|
|
|
|
// #include "config.hpp"
|
|
|
|
// #include "token.hpp"
|
|
#ifndef PANTOR_INJA_TOKEN_HPP
|
|
#define PANTOR_INJA_TOKEN_HPP
|
|
|
|
#include <string>
|
|
|
|
// #include "string_view.hpp"
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
/*!
|
|
* \brief Helper-class for the inja Parser.
|
|
*/
|
|
struct Token {
|
|
enum class Kind {
|
|
Text,
|
|
ExpressionOpen, // {{
|
|
ExpressionClose, // }}
|
|
LineStatementOpen, // ##
|
|
LineStatementClose, // \n
|
|
StatementOpen, // {%
|
|
StatementClose, // %}
|
|
CommentOpen, // {#
|
|
CommentClose, // #}
|
|
Id, // this, this.foo
|
|
Number, // 1, 2, -1, 5.2, -5.3
|
|
String, // "this"
|
|
Comma, // ,
|
|
Colon, // :
|
|
LeftParen, // (
|
|
RightParen, // )
|
|
LeftBracket, // [
|
|
RightBracket, // ]
|
|
LeftBrace, // {
|
|
RightBrace, // }
|
|
Equal, // ==
|
|
GreaterThan, // >
|
|
GreaterEqual, // >=
|
|
LessThan, // <
|
|
LessEqual, // <=
|
|
NotEqual, // !=
|
|
Unknown,
|
|
Eof
|
|
} kind {Kind::Unknown};
|
|
|
|
nonstd::string_view text;
|
|
|
|
constexpr Token() = default;
|
|
constexpr Token(Kind kind, nonstd::string_view text): kind(kind), text(text) {}
|
|
|
|
std::string describe() const {
|
|
switch (kind) {
|
|
case Kind::Text:
|
|
return "<text>";
|
|
case Kind::LineStatementClose:
|
|
return "<eol>";
|
|
case Kind::Eof:
|
|
return "<eof>";
|
|
default:
|
|
return static_cast<std::string>(text);
|
|
}
|
|
}
|
|
};
|
|
|
|
}
|
|
|
|
#endif // PANTOR_INJA_TOKEN_HPP
|
|
|
|
// #include "utils.hpp"
|
|
#ifndef PANTOR_INJA_UTILS_HPP
|
|
#define PANTOR_INJA_UTILS_HPP
|
|
|
|
#include <algorithm>
|
|
#include <fstream>
|
|
#include <stdexcept>
|
|
#include <string>
|
|
#include <utility>
|
|
|
|
// #include "string_view.hpp"
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
inline void inja_throw(const std::string& type, const std::string& message) {
|
|
throw std::runtime_error("[inja.exception." + type + "] " + message);
|
|
}
|
|
|
|
inline std::ifstream open_file_or_throw(const std::string& path) {
|
|
std::ifstream file;
|
|
file.exceptions(std::ifstream::failbit | std::ifstream::badbit);
|
|
try {
|
|
file.open(path);
|
|
} catch(const std::ios_base::failure& e) {
|
|
inja_throw("file_error", "failed accessing file at '" + path + "'");
|
|
}
|
|
return file;
|
|
}
|
|
|
|
namespace string_view {
|
|
inline nonstd::string_view slice(nonstd::string_view view, size_t start, size_t end) {
|
|
start = std::min(start, view.size());
|
|
end = std::min(std::max(start, end), view.size());
|
|
return view.substr(start, end - start); // StringRef(Data + Start, End - Start);
|
|
}
|
|
|
|
inline std::pair<nonstd::string_view, nonstd::string_view> split(nonstd::string_view view, char Separator) {
|
|
size_t idx = view.find(Separator);
|
|
if (idx == nonstd::string_view::npos) {
|
|
return std::make_pair(view, nonstd::string_view());
|
|
}
|
|
return std::make_pair(slice(view, 0, idx), slice(view, idx + 1, nonstd::string_view::npos));
|
|
}
|
|
|
|
inline bool starts_with(nonstd::string_view view, nonstd::string_view prefix) {
|
|
return (view.size() >= prefix.size() && view.compare(0, prefix.size(), prefix) == 0);
|
|
}
|
|
} // namespace string
|
|
|
|
} // namespace inja
|
|
|
|
#endif // PANTOR_INJA_UTILS_HPP
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
/*!
|
|
* \brief Class for lexing an inja Template.
|
|
*/
|
|
class Lexer {
|
|
enum class State {
|
|
Text,
|
|
ExpressionStart,
|
|
ExpressionBody,
|
|
LineStart,
|
|
LineBody,
|
|
StatementStart,
|
|
StatementBody,
|
|
CommentStart,
|
|
CommentBody
|
|
} m_state;
|
|
|
|
const LexerConfig& m_config;
|
|
nonstd::string_view m_in;
|
|
size_t m_tok_start;
|
|
size_t m_pos;
|
|
|
|
public:
|
|
explicit Lexer(const LexerConfig& config) : m_config(config) {}
|
|
|
|
void start(nonstd::string_view in) {
|
|
m_in = in;
|
|
m_tok_start = 0;
|
|
m_pos = 0;
|
|
m_state = State::Text;
|
|
}
|
|
|
|
Token scan() {
|
|
m_tok_start = m_pos;
|
|
|
|
again:
|
|
if (m_tok_start >= m_in.size()) return make_token(Token::Kind::Eof);
|
|
|
|
switch (m_state) {
|
|
default:
|
|
case State::Text: {
|
|
// fast-scan to first open character
|
|
size_t open_start = m_in.substr(m_pos).find_first_of(m_config.open_chars);
|
|
if (open_start == nonstd::string_view::npos) {
|
|
// didn't find open, return remaining text as text token
|
|
m_pos = m_in.size();
|
|
return make_token(Token::Kind::Text);
|
|
}
|
|
m_pos += open_start;
|
|
|
|
// try to match one of the opening sequences, and get the close
|
|
nonstd::string_view open_str = m_in.substr(m_pos);
|
|
bool must_lstrip = false;
|
|
if (inja::string_view::starts_with(open_str, m_config.expression_open)) {
|
|
m_state = State::ExpressionStart;
|
|
} else if (inja::string_view::starts_with(open_str, m_config.statement_open)) {
|
|
m_state = State::StatementStart;
|
|
must_lstrip = m_config.lstrip_blocks;
|
|
} else if (inja::string_view::starts_with(open_str, m_config.comment_open)) {
|
|
m_state = State::CommentStart;
|
|
must_lstrip = m_config.lstrip_blocks;
|
|
} else if ((m_pos == 0 || m_in[m_pos - 1] == '\n') &&
|
|
inja::string_view::starts_with(open_str, m_config.line_statement)) {
|
|
m_state = State::LineStart;
|
|
} else {
|
|
m_pos += 1; // wasn't actually an opening sequence
|
|
goto again;
|
|
}
|
|
|
|
nonstd::string_view text = string_view::slice(m_in, m_tok_start, m_pos);
|
|
if (must_lstrip)
|
|
text = clear_final_line_if_whitespace(text);
|
|
|
|
if (text.empty()) goto again; // don't generate empty token
|
|
return Token(Token::Kind::Text, text);
|
|
}
|
|
case State::ExpressionStart: {
|
|
m_state = State::ExpressionBody;
|
|
m_pos += m_config.expression_open.size();
|
|
return make_token(Token::Kind::ExpressionOpen);
|
|
}
|
|
case State::LineStart: {
|
|
m_state = State::LineBody;
|
|
m_pos += m_config.line_statement.size();
|
|
return make_token(Token::Kind::LineStatementOpen);
|
|
}
|
|
case State::StatementStart: {
|
|
m_state = State::StatementBody;
|
|
m_pos += m_config.statement_open.size();
|
|
return make_token(Token::Kind::StatementOpen);
|
|
}
|
|
case State::CommentStart: {
|
|
m_state = State::CommentBody;
|
|
m_pos += m_config.comment_open.size();
|
|
return make_token(Token::Kind::CommentOpen);
|
|
}
|
|
case State::ExpressionBody:
|
|
return scan_body(m_config.expression_close, Token::Kind::ExpressionClose);
|
|
case State::LineBody:
|
|
return scan_body("\n", Token::Kind::LineStatementClose);
|
|
case State::StatementBody:
|
|
return scan_body(m_config.statement_close, Token::Kind::StatementClose, m_config.trim_blocks);
|
|
case State::CommentBody: {
|
|
// fast-scan to comment close
|
|
size_t end = m_in.substr(m_pos).find(m_config.comment_close);
|
|
if (end == nonstd::string_view::npos) {
|
|
m_pos = m_in.size();
|
|
return make_token(Token::Kind::Eof);
|
|
}
|
|
// return the entire comment in the close token
|
|
m_state = State::Text;
|
|
m_pos += end + m_config.comment_close.size();
|
|
Token tok = make_token(Token::Kind::CommentClose);
|
|
if (m_config.trim_blocks)
|
|
skip_newline();
|
|
return tok;
|
|
}
|
|
}
|
|
}
|
|
|
|
const LexerConfig& get_config() const { return m_config; }
|
|
|
|
private:
|
|
Token scan_body(nonstd::string_view close, Token::Kind closeKind, bool trim = false) {
|
|
again:
|
|
// skip whitespace (except for \n as it might be a close)
|
|
if (m_tok_start >= m_in.size()) return make_token(Token::Kind::Eof);
|
|
char ch = m_in[m_tok_start];
|
|
if (ch == ' ' || ch == '\t' || ch == '\r') {
|
|
m_tok_start += 1;
|
|
goto again;
|
|
}
|
|
|
|
// check for close
|
|
if (inja::string_view::starts_with(m_in.substr(m_tok_start), close)) {
|
|
m_state = State::Text;
|
|
m_pos = m_tok_start + close.size();
|
|
Token tok = make_token(closeKind);
|
|
if (trim)
|
|
skip_newline();
|
|
return tok;
|
|
}
|
|
|
|
// skip \n
|
|
if (ch == '\n') {
|
|
m_tok_start += 1;
|
|
goto again;
|
|
}
|
|
|
|
m_pos = m_tok_start + 1;
|
|
if (std::isalpha(ch)) return scan_id();
|
|
switch (ch) {
|
|
case ',':
|
|
return make_token(Token::Kind::Comma);
|
|
case ':':
|
|
return make_token(Token::Kind::Colon);
|
|
case '(':
|
|
return make_token(Token::Kind::LeftParen);
|
|
case ')':
|
|
return make_token(Token::Kind::RightParen);
|
|
case '[':
|
|
return make_token(Token::Kind::LeftBracket);
|
|
case ']':
|
|
return make_token(Token::Kind::RightBracket);
|
|
case '{':
|
|
return make_token(Token::Kind::LeftBrace);
|
|
case '}':
|
|
return make_token(Token::Kind::RightBrace);
|
|
case '>':
|
|
if (m_pos < m_in.size() && m_in[m_pos] == '=') {
|
|
m_pos += 1;
|
|
return make_token(Token::Kind::GreaterEqual);
|
|
}
|
|
return make_token(Token::Kind::GreaterThan);
|
|
case '<':
|
|
if (m_pos < m_in.size() && m_in[m_pos] == '=') {
|
|
m_pos += 1;
|
|
return make_token(Token::Kind::LessEqual);
|
|
}
|
|
return make_token(Token::Kind::LessThan);
|
|
case '=':
|
|
if (m_pos < m_in.size() && m_in[m_pos] == '=') {
|
|
m_pos += 1;
|
|
return make_token(Token::Kind::Equal);
|
|
}
|
|
return make_token(Token::Kind::Unknown);
|
|
case '!':
|
|
if (m_pos < m_in.size() && m_in[m_pos] == '=') {
|
|
m_pos += 1;
|
|
return make_token(Token::Kind::NotEqual);
|
|
}
|
|
return make_token(Token::Kind::Unknown);
|
|
case '\"':
|
|
return scan_string();
|
|
case '0':
|
|
case '1':
|
|
case '2':
|
|
case '3':
|
|
case '4':
|
|
case '5':
|
|
case '6':
|
|
case '7':
|
|
case '8':
|
|
case '9':
|
|
case '-':
|
|
return scan_number();
|
|
case '_':
|
|
return scan_id();
|
|
default:
|
|
return make_token(Token::Kind::Unknown);
|
|
}
|
|
}
|
|
|
|
Token scan_id() {
|
|
for (;;) {
|
|
if (m_pos >= m_in.size()) {
|
|
break;
|
|
}
|
|
char ch = m_in[m_pos];
|
|
if (!std::isalnum(ch) && ch != '.' && ch != '/' && ch != '_' && ch != '-') {
|
|
break;
|
|
}
|
|
m_pos += 1;
|
|
}
|
|
return make_token(Token::Kind::Id);
|
|
}
|
|
|
|
Token scan_number() {
|
|
for (;;) {
|
|
if (m_pos >= m_in.size()) {
|
|
break;
|
|
}
|
|
char ch = m_in[m_pos];
|
|
// be very permissive in lexer (we'll catch errors when conversion happens)
|
|
if (!std::isdigit(ch) && ch != '.' && ch != 'e' && ch != 'E' && ch != '+' && ch != '-') {
|
|
break;
|
|
}
|
|
m_pos += 1;
|
|
}
|
|
return make_token(Token::Kind::Number);
|
|
}
|
|
|
|
Token scan_string() {
|
|
bool escape {false};
|
|
for (;;) {
|
|
if (m_pos >= m_in.size()) break;
|
|
char ch = m_in[m_pos++];
|
|
if (ch == '\\') {
|
|
escape = true;
|
|
} else if (!escape && ch == m_in[m_tok_start]) {
|
|
break;
|
|
} else {
|
|
escape = false;
|
|
}
|
|
}
|
|
return make_token(Token::Kind::String);
|
|
}
|
|
|
|
Token make_token(Token::Kind kind) const {
|
|
return Token(kind, string_view::slice(m_in, m_tok_start, m_pos));
|
|
}
|
|
|
|
void skip_newline() {
|
|
if (m_pos < m_in.size()) {
|
|
char ch = m_in[m_pos];
|
|
if (ch == '\n')
|
|
m_pos += 1;
|
|
else if (ch == '\r') {
|
|
m_pos += 1;
|
|
if (m_pos < m_in.size() && m_in[m_pos] == '\n')
|
|
m_pos += 1;
|
|
}
|
|
}
|
|
}
|
|
|
|
static nonstd::string_view clear_final_line_if_whitespace(nonstd::string_view text)
|
|
{
|
|
nonstd::string_view result = text;
|
|
while (!result.empty()) {
|
|
char ch = result.back();
|
|
if (ch == ' ' || ch == '\t')
|
|
result.remove_suffix(1);
|
|
else if (ch == '\n' || ch == '\r')
|
|
break;
|
|
else
|
|
return text;
|
|
}
|
|
return result;
|
|
}
|
|
};
|
|
|
|
}
|
|
|
|
#endif // PANTOR_INJA_LEXER_HPP
|
|
|
|
// #include "template.hpp"
|
|
#ifndef PANTOR_INJA_TEMPLATE_HPP
|
|
#define PANTOR_INJA_TEMPLATE_HPP
|
|
|
|
#include <map>
|
|
#include <string>
|
|
#include <vector>
|
|
|
|
// #include "bytecode.hpp"
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
/*!
|
|
* \brief The main inja Template.
|
|
*/
|
|
struct Template {
|
|
std::vector<Bytecode> bytecodes;
|
|
std::string content;
|
|
};
|
|
|
|
using TemplateStorage = std::map<std::string, Template>;
|
|
|
|
}
|
|
|
|
#endif // PANTOR_INJA_TEMPLATE_HPP
|
|
|
|
// #include "token.hpp"
|
|
|
|
// #include "utils.hpp"
|
|
|
|
|
|
#include <nlohmann/json.hpp>
|
|
|
|
|
|
namespace inja {
|
|
|
|
class ParserStatic {
|
|
ParserStatic() {
|
|
functions.add_builtin("at", 2, Bytecode::Op::At);
|
|
functions.add_builtin("default", 2, Bytecode::Op::Default);
|
|
functions.add_builtin("divisibleBy", 2, Bytecode::Op::DivisibleBy);
|
|
functions.add_builtin("even", 1, Bytecode::Op::Even);
|
|
functions.add_builtin("first", 1, Bytecode::Op::First);
|
|
functions.add_builtin("float", 1, Bytecode::Op::Float);
|
|
functions.add_builtin("int", 1, Bytecode::Op::Int);
|
|
functions.add_builtin("last", 1, Bytecode::Op::Last);
|
|
functions.add_builtin("length", 1, Bytecode::Op::Length);
|
|
functions.add_builtin("lower", 1, Bytecode::Op::Lower);
|
|
functions.add_builtin("max", 1, Bytecode::Op::Max);
|
|
functions.add_builtin("min", 1, Bytecode::Op::Min);
|
|
functions.add_builtin("odd", 1, Bytecode::Op::Odd);
|
|
functions.add_builtin("range", 1, Bytecode::Op::Range);
|
|
functions.add_builtin("round", 2, Bytecode::Op::Round);
|
|
functions.add_builtin("sort", 1, Bytecode::Op::Sort);
|
|
functions.add_builtin("upper", 1, Bytecode::Op::Upper);
|
|
functions.add_builtin("exists", 1, Bytecode::Op::Exists);
|
|
functions.add_builtin("existsIn", 2, Bytecode::Op::ExistsInObject);
|
|
functions.add_builtin("isBoolean", 1, Bytecode::Op::IsBoolean);
|
|
functions.add_builtin("isNumber", 1, Bytecode::Op::IsNumber);
|
|
functions.add_builtin("isInteger", 1, Bytecode::Op::IsInteger);
|
|
functions.add_builtin("isFloat", 1, Bytecode::Op::IsFloat);
|
|
functions.add_builtin("isObject", 1, Bytecode::Op::IsObject);
|
|
functions.add_builtin("isArray", 1, Bytecode::Op::IsArray);
|
|
functions.add_builtin("isString", 1, Bytecode::Op::IsString);
|
|
}
|
|
|
|
public:
|
|
ParserStatic(const ParserStatic&) = delete;
|
|
ParserStatic& operator=(const ParserStatic&) = delete;
|
|
|
|
static const ParserStatic& get_instance() {
|
|
static ParserStatic inst;
|
|
return inst;
|
|
}
|
|
|
|
FunctionStorage functions;
|
|
};
|
|
|
|
/*!
|
|
* \brief Class for parsing an inja Template.
|
|
*/
|
|
class Parser {
|
|
public:
|
|
explicit Parser(const ParserConfig& parser_config, const LexerConfig& lexer_config, TemplateStorage& included_templates): m_config(parser_config), m_lexer(lexer_config), m_included_templates(included_templates), m_static(ParserStatic::get_instance()) { }
|
|
|
|
bool parse_expression(Template& tmpl) {
|
|
if (!parse_expression_and(tmpl)) return false;
|
|
if (m_tok.kind != Token::Kind::Id || m_tok.text != static_cast<decltype(m_tok.text)>("or")) return true;
|
|
get_next_token();
|
|
if (!parse_expression_and(tmpl)) return false;
|
|
append_function(tmpl, Bytecode::Op::Or, 2);
|
|
return true;
|
|
}
|
|
|
|
bool parse_expression_and(Template& tmpl) {
|
|
if (!parse_expression_not(tmpl)) return false;
|
|
if (m_tok.kind != Token::Kind::Id || m_tok.text != static_cast<decltype(m_tok.text)>("and")) return true;
|
|
get_next_token();
|
|
if (!parse_expression_not(tmpl)) return false;
|
|
append_function(tmpl, Bytecode::Op::And, 2);
|
|
return true;
|
|
}
|
|
|
|
bool parse_expression_not(Template& tmpl) {
|
|
if (m_tok.kind == Token::Kind::Id && m_tok.text == static_cast<decltype(m_tok.text)>("not")) {
|
|
get_next_token();
|
|
if (!parse_expression_not(tmpl)) return false;
|
|
append_function(tmpl, Bytecode::Op::Not, 1);
|
|
return true;
|
|
} else {
|
|
return parse_expression_comparison(tmpl);
|
|
}
|
|
}
|
|
|
|
bool parse_expression_comparison(Template& tmpl) {
|
|
if (!parse_expression_datum(tmpl)) return false;
|
|
Bytecode::Op op;
|
|
switch (m_tok.kind) {
|
|
case Token::Kind::Id:
|
|
if (m_tok.text == static_cast<decltype(m_tok.text)>("in"))
|
|
op = Bytecode::Op::In;
|
|
else
|
|
return true;
|
|
break;
|
|
case Token::Kind::Equal:
|
|
op = Bytecode::Op::Equal;
|
|
break;
|
|
case Token::Kind::GreaterThan:
|
|
op = Bytecode::Op::Greater;
|
|
break;
|
|
case Token::Kind::LessThan:
|
|
op = Bytecode::Op::Less;
|
|
break;
|
|
case Token::Kind::LessEqual:
|
|
op = Bytecode::Op::LessEqual;
|
|
break;
|
|
case Token::Kind::GreaterEqual:
|
|
op = Bytecode::Op::GreaterEqual;
|
|
break;
|
|
case Token::Kind::NotEqual:
|
|
op = Bytecode::Op::Different;
|
|
break;
|
|
default:
|
|
return true;
|
|
}
|
|
get_next_token();
|
|
if (!parse_expression_datum(tmpl)) return false;
|
|
append_function(tmpl, op, 2);
|
|
return true;
|
|
}
|
|
|
|
bool parse_expression_datum(Template& tmpl) {
|
|
nonstd::string_view json_first;
|
|
size_t bracket_level = 0;
|
|
size_t brace_level = 0;
|
|
|
|
for (;;) {
|
|
switch (m_tok.kind) {
|
|
case Token::Kind::LeftParen: {
|
|
get_next_token();
|
|
if (!parse_expression(tmpl)) return false;
|
|
if (m_tok.kind != Token::Kind::RightParen) {
|
|
inja_throw("parser_error", "unmatched '('");
|
|
}
|
|
get_next_token();
|
|
return true;
|
|
}
|
|
case Token::Kind::Id:
|
|
get_peek_token();
|
|
if (m_peek_tok.kind == Token::Kind::LeftParen) {
|
|
// function call, parse arguments
|
|
Token func_token = m_tok;
|
|
get_next_token(); // id
|
|
get_next_token(); // leftParen
|
|
unsigned int num_args = 0;
|
|
if (m_tok.kind == Token::Kind::RightParen) {
|
|
// no args
|
|
get_next_token();
|
|
} else {
|
|
for (;;) {
|
|
if (!parse_expression(tmpl)) {
|
|
inja_throw("parser_error", "expected expression, got '" + m_tok.describe() + "'");
|
|
}
|
|
num_args += 1;
|
|
if (m_tok.kind == Token::Kind::RightParen) {
|
|
get_next_token();
|
|
break;
|
|
}
|
|
if (m_tok.kind != Token::Kind::Comma) {
|
|
inja_throw("parser_error", "expected ')' or ',', got '" + m_tok.describe() + "'");
|
|
}
|
|
get_next_token();
|
|
}
|
|
}
|
|
|
|
auto op = m_static.functions.find_builtin(func_token.text, num_args);
|
|
|
|
if (op != Bytecode::Op::Nop) {
|
|
// swap arguments for default(); see comment in RenderTo()
|
|
if (op == Bytecode::Op::Default)
|
|
std::swap(tmpl.bytecodes.back(), *(tmpl.bytecodes.rbegin() + 1));
|
|
append_function(tmpl, op, num_args);
|
|
return true;
|
|
} else {
|
|
append_callback(tmpl, func_token.text, num_args);
|
|
return true;
|
|
}
|
|
} else if (m_tok.text == static_cast<decltype(m_tok.text)>("true") ||
|
|
m_tok.text == static_cast<decltype(m_tok.text)>("false") ||
|
|
m_tok.text == static_cast<decltype(m_tok.text)>("null")) {
|
|
// true, false, null are json literals
|
|
if (brace_level == 0 && bracket_level == 0) {
|
|
json_first = m_tok.text;
|
|
goto returnJson;
|
|
}
|
|
break;
|
|
} else {
|
|
// normal literal (json read)
|
|
tmpl.bytecodes.emplace_back(
|
|
Bytecode::Op::Push, m_tok.text,
|
|
m_config.notation == ElementNotation::Pointer ? Bytecode::Flag::ValueLookupPointer : Bytecode::Flag::ValueLookupDot);
|
|
get_next_token();
|
|
return true;
|
|
}
|
|
// json passthrough
|
|
case Token::Kind::Number:
|
|
case Token::Kind::String:
|
|
if (brace_level == 0 && bracket_level == 0) {
|
|
json_first = m_tok.text;
|
|
goto returnJson;
|
|
}
|
|
break;
|
|
case Token::Kind::Comma:
|
|
case Token::Kind::Colon:
|
|
if (brace_level == 0 && bracket_level == 0) {
|
|
inja_throw("parser_error", "unexpected token '" + m_tok.describe() + "'");
|
|
}
|
|
break;
|
|
case Token::Kind::LeftBracket:
|
|
if (brace_level == 0 && bracket_level == 0) {
|
|
json_first = m_tok.text;
|
|
}
|
|
bracket_level += 1;
|
|
break;
|
|
case Token::Kind::LeftBrace:
|
|
if (brace_level == 0 && bracket_level == 0) {
|
|
json_first = m_tok.text;
|
|
}
|
|
brace_level += 1;
|
|
break;
|
|
case Token::Kind::RightBracket:
|
|
if (bracket_level == 0) {
|
|
inja_throw("parser_error", "unexpected ']'");
|
|
}
|
|
--bracket_level;
|
|
if (brace_level == 0 && bracket_level == 0) goto returnJson;
|
|
break;
|
|
case Token::Kind::RightBrace:
|
|
if (brace_level == 0) {
|
|
inja_throw("parser_error", "unexpected '}'");
|
|
}
|
|
--brace_level;
|
|
if (brace_level == 0 && bracket_level == 0) goto returnJson;
|
|
break;
|
|
default:
|
|
if (brace_level != 0) {
|
|
inja_throw("parser_error", "unmatched '{'");
|
|
}
|
|
if (bracket_level != 0) {
|
|
inja_throw("parser_error", "unmatched '['");
|
|
}
|
|
return false;
|
|
}
|
|
|
|
get_next_token();
|
|
}
|
|
|
|
returnJson:
|
|
// bridge across all intermediate tokens
|
|
nonstd::string_view json_text(json_first.data(), m_tok.text.data() - json_first.data() + m_tok.text.size());
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::Push, json::parse(json_text), Bytecode::Flag::ValueImmediate);
|
|
get_next_token();
|
|
return true;
|
|
}
|
|
|
|
bool parse_statement(Template& tmpl, nonstd::string_view path) {
|
|
if (m_tok.kind != Token::Kind::Id) return false;
|
|
|
|
if (m_tok.text == static_cast<decltype(m_tok.text)>("if")) {
|
|
get_next_token();
|
|
|
|
// evaluate expression
|
|
if (!parse_expression(tmpl)) return false;
|
|
|
|
// start a new if block on if stack
|
|
m_if_stack.emplace_back(tmpl.bytecodes.size());
|
|
|
|
// conditional jump; destination will be filled in by else or endif
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::ConditionalJump);
|
|
} else if (m_tok.text == static_cast<decltype(m_tok.text)>("endif")) {
|
|
if (m_if_stack.empty()) {
|
|
inja_throw("parser_error", "endif without matching if");
|
|
}
|
|
auto& if_data = m_if_stack.back();
|
|
get_next_token();
|
|
|
|
// previous conditional jump jumps here
|
|
if (if_data.prev_cond_jump != std::numeric_limits<unsigned int>::max()) {
|
|
tmpl.bytecodes[if_data.prev_cond_jump].args = tmpl.bytecodes.size();
|
|
}
|
|
|
|
// update all previous unconditional jumps to here
|
|
for (unsigned int i: if_data.uncond_jumps) {
|
|
tmpl.bytecodes[i].args = tmpl.bytecodes.size();
|
|
}
|
|
|
|
// pop if stack
|
|
m_if_stack.pop_back();
|
|
} else if (m_tok.text == static_cast<decltype(m_tok.text)>("else")) {
|
|
if (m_if_stack.empty())
|
|
inja_throw("parser_error", "else without matching if");
|
|
auto& if_data = m_if_stack.back();
|
|
get_next_token();
|
|
|
|
// end previous block with unconditional jump to endif; destination will be
|
|
// filled in by endif
|
|
if_data.uncond_jumps.push_back(tmpl.bytecodes.size());
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::Jump);
|
|
|
|
// previous conditional jump jumps here
|
|
tmpl.bytecodes[if_data.prev_cond_jump].args = tmpl.bytecodes.size();
|
|
if_data.prev_cond_jump = std::numeric_limits<unsigned int>::max();
|
|
|
|
// chained else if
|
|
if (m_tok.kind == Token::Kind::Id && m_tok.text == static_cast<decltype(m_tok.text)>("if")) {
|
|
get_next_token();
|
|
|
|
// evaluate expression
|
|
if (!parse_expression(tmpl)) return false;
|
|
|
|
// update "previous jump"
|
|
if_data.prev_cond_jump = tmpl.bytecodes.size();
|
|
|
|
// conditional jump; destination will be filled in by else or endif
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::ConditionalJump);
|
|
}
|
|
} else if (m_tok.text == static_cast<decltype(m_tok.text)>("for")) {
|
|
get_next_token();
|
|
|
|
// options: for a in arr; for a, b in obj
|
|
if (m_tok.kind != Token::Kind::Id)
|
|
inja_throw("parser_error", "expected id, got '" + m_tok.describe() + "'");
|
|
Token value_token = m_tok;
|
|
get_next_token();
|
|
|
|
Token key_token;
|
|
if (m_tok.kind == Token::Kind::Comma) {
|
|
get_next_token();
|
|
if (m_tok.kind != Token::Kind::Id)
|
|
inja_throw("parser_error", "expected id, got '" + m_tok.describe() + "'");
|
|
key_token = std::move(value_token);
|
|
value_token = m_tok;
|
|
get_next_token();
|
|
}
|
|
|
|
if (m_tok.kind != Token::Kind::Id || m_tok.text != static_cast<decltype(m_tok.text)>("in"))
|
|
inja_throw("parser_error",
|
|
"expected 'in', got '" + m_tok.describe() + "'");
|
|
get_next_token();
|
|
|
|
if (!parse_expression(tmpl)) return false;
|
|
|
|
m_loop_stack.push_back(tmpl.bytecodes.size());
|
|
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::StartLoop);
|
|
if (!key_token.text.empty()) {
|
|
tmpl.bytecodes.back().value = key_token.text;
|
|
}
|
|
tmpl.bytecodes.back().str = static_cast<std::string>(value_token.text);
|
|
} else if (m_tok.text == static_cast<decltype(m_tok.text)>("endfor")) {
|
|
get_next_token();
|
|
if (m_loop_stack.empty()) {
|
|
inja_throw("parser_error", "endfor without matching for");
|
|
}
|
|
|
|
// update loop with EndLoop index (for empty case)
|
|
tmpl.bytecodes[m_loop_stack.back()].args = tmpl.bytecodes.size();
|
|
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::EndLoop);
|
|
tmpl.bytecodes.back().args = m_loop_stack.back() + 1; // loop body
|
|
m_loop_stack.pop_back();
|
|
} else if (m_tok.text == static_cast<decltype(m_tok.text)>("include")) {
|
|
get_next_token();
|
|
|
|
if (m_tok.kind != Token::Kind::String) {
|
|
inja_throw("parser_error", "expected string, got '" + m_tok.describe() + "'");
|
|
}
|
|
|
|
// build the relative path
|
|
json json_name = json::parse(m_tok.text);
|
|
std::string pathname = static_cast<std::string>(path);
|
|
pathname += json_name.get_ref<const std::string&>();
|
|
if (pathname.compare(0, 2, "./") == 0) {
|
|
pathname.erase(0, 2);
|
|
}
|
|
// sys::path::remove_dots(pathname, true, sys::path::Style::posix);
|
|
|
|
if (m_included_templates.find(pathname) == m_included_templates.end()) {
|
|
Template include_template = parse_template(pathname);
|
|
m_included_templates.emplace(pathname, include_template);
|
|
}
|
|
|
|
// generate a reference bytecode
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::Include, json(pathname), Bytecode::Flag::ValueImmediate);
|
|
|
|
get_next_token();
|
|
} else {
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
void append_function(Template& tmpl, Bytecode::Op op, unsigned int num_args) {
|
|
// we can merge with back-to-back push
|
|
if (!tmpl.bytecodes.empty()) {
|
|
Bytecode& last = tmpl.bytecodes.back();
|
|
if (last.op == Bytecode::Op::Push) {
|
|
last.op = op;
|
|
last.args = num_args;
|
|
return;
|
|
}
|
|
}
|
|
|
|
// otherwise just add it to the end
|
|
tmpl.bytecodes.emplace_back(op, num_args);
|
|
}
|
|
|
|
void append_callback(Template& tmpl, nonstd::string_view name, unsigned int num_args) {
|
|
// we can merge with back-to-back push value (not lookup)
|
|
if (!tmpl.bytecodes.empty()) {
|
|
Bytecode& last = tmpl.bytecodes.back();
|
|
if (last.op == Bytecode::Op::Push &&
|
|
(last.flags & Bytecode::Flag::ValueMask) == Bytecode::Flag::ValueImmediate) {
|
|
last.op = Bytecode::Op::Callback;
|
|
last.args = num_args;
|
|
last.str = static_cast<std::string>(name);
|
|
return;
|
|
}
|
|
}
|
|
|
|
// otherwise just add it to the end
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::Callback, num_args);
|
|
tmpl.bytecodes.back().str = static_cast<std::string>(name);
|
|
}
|
|
|
|
void parse_into(Template& tmpl, nonstd::string_view path) {
|
|
m_lexer.start(tmpl.content);
|
|
|
|
for (;;) {
|
|
get_next_token();
|
|
switch (m_tok.kind) {
|
|
case Token::Kind::Eof:
|
|
if (!m_if_stack.empty()) inja_throw("parser_error", "unmatched if");
|
|
if (!m_loop_stack.empty()) inja_throw("parser_error", "unmatched for");
|
|
return;
|
|
case Token::Kind::Text:
|
|
tmpl.bytecodes.emplace_back(Bytecode::Op::PrintText, m_tok.text, 0u);
|
|
break;
|
|
case Token::Kind::StatementOpen:
|
|
get_next_token();
|
|
if (!parse_statement(tmpl, path)) {
|
|
inja_throw("parser_error", "expected statement, got '" + m_tok.describe() + "'");
|
|
}
|
|
if (m_tok.kind != Token::Kind::StatementClose) {
|
|
inja_throw("parser_error", "expected statement close, got '" + m_tok.describe() + "'");
|
|
}
|
|
break;
|
|
case Token::Kind::LineStatementOpen:
|
|
get_next_token();
|
|
parse_statement(tmpl, path);
|
|
if (m_tok.kind != Token::Kind::LineStatementClose &&
|
|
m_tok.kind != Token::Kind::Eof) {
|
|
inja_throw("parser_error", "expected line statement close, got '" + m_tok.describe() + "'");
|
|
}
|
|
break;
|
|
case Token::Kind::ExpressionOpen:
|
|
get_next_token();
|
|
if (!parse_expression(tmpl)) {
|
|
inja_throw("parser_error", "expected expression, got '" + m_tok.describe() + "'");
|
|
}
|
|
append_function(tmpl, Bytecode::Op::PrintValue, 1);
|
|
if (m_tok.kind != Token::Kind::ExpressionClose) {
|
|
inja_throw("parser_error", "expected expression close, got '" + m_tok.describe() + "'");
|
|
}
|
|
break;
|
|
case Token::Kind::CommentOpen:
|
|
get_next_token();
|
|
if (m_tok.kind != Token::Kind::CommentClose) {
|
|
inja_throw("parser_error", "expected comment close, got '" + m_tok.describe() + "'");
|
|
}
|
|
break;
|
|
default:
|
|
inja_throw("parser_error", "unexpected token '" + m_tok.describe() + "'");
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
Template parse(nonstd::string_view input, nonstd::string_view path) {
|
|
Template result;
|
|
result.content = static_cast<std::string>(input);
|
|
parse_into(result, path);
|
|
return result;
|
|
}
|
|
|
|
Template parse(nonstd::string_view input) {
|
|
return parse(input, "./");
|
|
}
|
|
|
|
Template parse_template(nonstd::string_view filename) {
|
|
Template result;
|
|
result.content = load_file(filename);
|
|
|
|
nonstd::string_view path = filename.substr(0, filename.find_last_of("/\\") + 1);
|
|
// StringRef path = sys::path::parent_path(filename);
|
|
Parser(m_config, m_lexer.get_config(), m_included_templates).parse_into(result, path);
|
|
return result;
|
|
}
|
|
|
|
std::string load_file(nonstd::string_view filename) {
|
|
std::ifstream file = open_file_or_throw(static_cast<std::string>(filename));
|
|
std::string text((std::istreambuf_iterator<char>(file)), std::istreambuf_iterator<char>());
|
|
return text;
|
|
}
|
|
|
|
private:
|
|
const ParserConfig& m_config;
|
|
Lexer m_lexer;
|
|
Token m_tok;
|
|
Token m_peek_tok;
|
|
bool m_have_peek_tok {false};
|
|
TemplateStorage& m_included_templates;
|
|
const ParserStatic& m_static;
|
|
|
|
struct IfData {
|
|
unsigned int prev_cond_jump;
|
|
std::vector<unsigned int> uncond_jumps;
|
|
|
|
explicit IfData(unsigned int condJump): prev_cond_jump(condJump) {}
|
|
};
|
|
|
|
std::vector<IfData> m_if_stack;
|
|
std::vector<unsigned int> m_loop_stack;
|
|
|
|
void get_next_token() {
|
|
if (m_have_peek_tok) {
|
|
m_tok = m_peek_tok;
|
|
m_have_peek_tok = false;
|
|
} else {
|
|
m_tok = m_lexer.scan();
|
|
}
|
|
}
|
|
|
|
void get_peek_token() {
|
|
if (!m_have_peek_tok) {
|
|
m_peek_tok = m_lexer.scan();
|
|
m_have_peek_tok = true;
|
|
}
|
|
}
|
|
};
|
|
|
|
} // namespace inja
|
|
|
|
#endif // PANTOR_INJA_PARSER_HPP
|
|
|
|
// #include "polyfill.hpp"
|
|
#ifndef PANTOR_INJA_POLYFILL_HPP
|
|
#define PANTOR_INJA_POLYFILL_HPP
|
|
|
|
|
|
#if __cplusplus < 201402L
|
|
|
|
#include <cstddef>
|
|
#include <memory>
|
|
#include <type_traits>
|
|
#include <utility>
|
|
|
|
|
|
namespace stdinja {
|
|
template<class T> struct _Unique_if {
|
|
typedef std::unique_ptr<T> _Single_object;
|
|
};
|
|
|
|
template<class T> struct _Unique_if<T[]> {
|
|
typedef std::unique_ptr<T[]> _Unknown_bound;
|
|
};
|
|
|
|
template<class T, size_t N> struct _Unique_if<T[N]> {
|
|
typedef void _Known_bound;
|
|
};
|
|
|
|
template<class T, class... Args>
|
|
typename _Unique_if<T>::_Single_object
|
|
make_unique(Args&&... args) {
|
|
return std::unique_ptr<T>(new T(std::forward<Args>(args)...));
|
|
}
|
|
|
|
template<class T>
|
|
typename _Unique_if<T>::_Unknown_bound
|
|
make_unique(size_t n) {
|
|
typedef typename std::remove_extent<T>::type U;
|
|
return std::unique_ptr<T>(new U[n]());
|
|
}
|
|
|
|
template<class T, class... Args>
|
|
typename _Unique_if<T>::_Known_bound
|
|
make_unique(Args&&...) = delete;
|
|
}
|
|
|
|
#else
|
|
|
|
namespace stdinja = std;
|
|
|
|
#endif // memory */
|
|
|
|
|
|
#endif // PANTOR_INJA_POLYFILL_HPP
|
|
|
|
// #include "renderer.hpp"
|
|
#ifndef PANTOR_INJA_RENDERER_HPP
|
|
#define PANTOR_INJA_RENDERER_HPP
|
|
|
|
#include <algorithm>
|
|
#include <numeric>
|
|
#include <string>
|
|
#include <utility>
|
|
#include <vector>
|
|
|
|
#include <nlohmann/json.hpp>
|
|
|
|
// #include "bytecode.hpp"
|
|
|
|
// #include "template.hpp"
|
|
|
|
// #include "utils.hpp"
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
inline nonstd::string_view convert_dot_to_json_pointer(nonstd::string_view dot, std::string& out) {
|
|
out.clear();
|
|
do {
|
|
nonstd::string_view part;
|
|
std::tie(part, dot) = string_view::split(dot, '.');
|
|
out.push_back('/');
|
|
out.append(part.begin(), part.end());
|
|
} while (!dot.empty());
|
|
return nonstd::string_view(out.data(), out.size());
|
|
}
|
|
|
|
/*!
|
|
* \brief Class for rendering a Template with data.
|
|
*/
|
|
class Renderer {
|
|
std::vector<const json*>& get_args(const Bytecode& bc) {
|
|
m_tmp_args.clear();
|
|
|
|
bool has_imm = ((bc.flags & Bytecode::Flag::ValueMask) != Bytecode::Flag::ValuePop);
|
|
|
|
// get args from stack
|
|
unsigned int pop_args = bc.args;
|
|
if (has_imm) {
|
|
pop_args -= 1;
|
|
}
|
|
|
|
for (auto i = std::prev(m_stack.end(), pop_args); i != m_stack.end(); i++) {
|
|
m_tmp_args.push_back(&(*i));
|
|
}
|
|
|
|
// get immediate arg
|
|
if (has_imm) {
|
|
m_tmp_args.push_back(get_imm(bc));
|
|
}
|
|
|
|
return m_tmp_args;
|
|
}
|
|
|
|
void pop_args(const Bytecode& bc) {
|
|
unsigned int popArgs = bc.args;
|
|
if ((bc.flags & Bytecode::Flag::ValueMask) != Bytecode::Flag::ValuePop) {
|
|
popArgs -= 1;
|
|
}
|
|
for (unsigned int i = 0; i < popArgs; ++i) {
|
|
m_stack.pop_back();
|
|
}
|
|
}
|
|
|
|
const json* get_imm(const Bytecode& bc) {
|
|
std::string ptr_buffer;
|
|
nonstd::string_view ptr;
|
|
switch (bc.flags & Bytecode::Flag::ValueMask) {
|
|
case Bytecode::Flag::ValuePop:
|
|
return nullptr;
|
|
case Bytecode::Flag::ValueImmediate:
|
|
return &bc.value;
|
|
case Bytecode::Flag::ValueLookupDot:
|
|
ptr = convert_dot_to_json_pointer(bc.str, ptr_buffer);
|
|
break;
|
|
case Bytecode::Flag::ValueLookupPointer:
|
|
ptr_buffer += '/';
|
|
ptr_buffer += bc.str;
|
|
ptr = ptr_buffer;
|
|
break;
|
|
}
|
|
try {
|
|
return &m_data->at(json::json_pointer(ptr.data()));
|
|
} catch (std::exception&) {
|
|
// try to evaluate as a no-argument callback
|
|
if (auto callback = m_callbacks.find_callback(bc.str, 0)) {
|
|
std::vector<const json*> arguments {};
|
|
m_tmp_val = callback(arguments);
|
|
return &m_tmp_val;
|
|
}
|
|
inja_throw("render_error", "variable '" + static_cast<std::string>(bc.str) + "' not found");
|
|
return nullptr;
|
|
}
|
|
}
|
|
|
|
bool truthy(const json& var) const {
|
|
if (var.empty()) {
|
|
return false;
|
|
} else if (var.is_number()) {
|
|
return (var != 0);
|
|
} else if (var.is_string()) {
|
|
return !var.empty();
|
|
}
|
|
|
|
try {
|
|
return var.get<bool>();
|
|
} catch (json::type_error& e) {
|
|
inja_throw("json_error", e.what());
|
|
throw;
|
|
}
|
|
}
|
|
|
|
void update_loop_data() {
|
|
LoopLevel& level = m_loop_stack.back();
|
|
|
|
if (level.loop_type == LoopLevel::Type::Array) {
|
|
level.data[static_cast<std::string>(level.value_name)] = level.values.at(level.index); // *level.it;
|
|
auto& loopData = level.data["loop"];
|
|
loopData["index"] = level.index;
|
|
loopData["index1"] = level.index + 1;
|
|
loopData["is_first"] = (level.index == 0);
|
|
loopData["is_last"] = (level.index == level.size - 1);
|
|
} else {
|
|
level.data[static_cast<std::string>(level.key_name)] = level.map_it->first;
|
|
level.data[static_cast<std::string>(level.value_name)] = *level.map_it->second;
|
|
}
|
|
}
|
|
|
|
const TemplateStorage& m_included_templates;
|
|
const FunctionStorage& m_callbacks;
|
|
|
|
std::vector<json> m_stack;
|
|
|
|
|
|
struct LoopLevel {
|
|
enum class Type { Map, Array };
|
|
|
|
Type loop_type;
|
|
nonstd::string_view key_name; // variable name for keys
|
|
nonstd::string_view value_name; // variable name for values
|
|
json data; // data with loop info added
|
|
|
|
json values; // values to iterate over
|
|
|
|
// loop over list
|
|
size_t index; // current list index
|
|
size_t size; // length of list
|
|
|
|
// loop over map
|
|
using KeyValue = std::pair<nonstd::string_view, json*>;
|
|
using MapValues = std::vector<KeyValue>;
|
|
MapValues map_values; // values to iterate over
|
|
MapValues::iterator map_it; // iterator over values
|
|
|
|
};
|
|
|
|
std::vector<LoopLevel> m_loop_stack;
|
|
const json* m_data;
|
|
|
|
std::vector<const json*> m_tmp_args;
|
|
json m_tmp_val;
|
|
|
|
|
|
public:
|
|
Renderer(const TemplateStorage& included_templates, const FunctionStorage& callbacks): m_included_templates(included_templates), m_callbacks(callbacks) {
|
|
m_stack.reserve(16);
|
|
m_tmp_args.reserve(4);
|
|
m_loop_stack.reserve(16);
|
|
}
|
|
|
|
void render_to(std::ostream& os, const Template& tmpl, const json& data) {
|
|
m_data = &data;
|
|
|
|
for (size_t i = 0; i < tmpl.bytecodes.size(); ++i) {
|
|
const auto& bc = tmpl.bytecodes[i];
|
|
|
|
switch (bc.op) {
|
|
case Bytecode::Op::Nop: {
|
|
break;
|
|
}
|
|
case Bytecode::Op::PrintText: {
|
|
os << bc.str;
|
|
break;
|
|
}
|
|
case Bytecode::Op::PrintValue: {
|
|
const json& val = *get_args(bc)[0];
|
|
if (val.is_string()) {
|
|
os << val.get_ref<const std::string&>();
|
|
} else {
|
|
os << val.dump();
|
|
}
|
|
pop_args(bc);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Push: {
|
|
m_stack.emplace_back(*get_imm(bc));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Upper: {
|
|
auto result = get_args(bc)[0]->get<std::string>();
|
|
std::transform(result.begin(), result.end(), result.begin(), ::toupper);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::move(result));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Lower: {
|
|
auto result = get_args(bc)[0]->get<std::string>();
|
|
std::transform(result.begin(), result.end(), result.begin(), ::tolower);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::move(result));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Range: {
|
|
int number = get_args(bc)[0]->get<int>();
|
|
std::vector<int> result(number);
|
|
std::iota(std::begin(result), std::end(result), 0);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::move(result));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Length: {
|
|
const json& val = *get_args(bc)[0];
|
|
|
|
int result;
|
|
if (val.is_string()) {
|
|
result = val.get_ref<const std::string&>().length();
|
|
} else {
|
|
result = val.size();
|
|
}
|
|
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Sort: {
|
|
auto result = get_args(bc)[0]->get<std::vector<json>>();
|
|
std::sort(result.begin(), result.end());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::move(result));
|
|
break;
|
|
}
|
|
case Bytecode::Op::At: {
|
|
auto args = get_args(bc);
|
|
auto result = args[0]->at(args[1]->get<int>());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::First: {
|
|
auto result = get_args(bc)[0]->front();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Last: {
|
|
auto result = get_args(bc)[0]->back();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Round: {
|
|
auto args = get_args(bc);
|
|
double number = args[0]->get<double>();
|
|
int precision = args[1]->get<int>();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::round(number * std::pow(10.0, precision)) / std::pow(10.0, precision));
|
|
break;
|
|
}
|
|
case Bytecode::Op::DivisibleBy: {
|
|
auto args = get_args(bc);
|
|
int number = args[0]->get<int>();
|
|
int divisor = args[1]->get<int>();
|
|
pop_args(bc);
|
|
m_stack.emplace_back((divisor != 0) && (number % divisor == 0));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Odd: {
|
|
int number = get_args(bc)[0]->get<int>();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(number % 2 != 0);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Even: {
|
|
int number = get_args(bc)[0]->get<int>();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(number % 2 == 0);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Max: {
|
|
auto args = get_args(bc);
|
|
auto result = *std::max_element(args[0]->begin(), args[0]->end());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::move(result));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Min: {
|
|
auto args = get_args(bc);
|
|
auto result = *std::min_element(args[0]->begin(), args[0]->end());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::move(result));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Not: {
|
|
bool result = !truthy(*get_args(bc)[0]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::And: {
|
|
auto args = get_args(bc);
|
|
bool result = truthy(*args[0]) && truthy(*args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Or: {
|
|
auto args = get_args(bc);
|
|
bool result = truthy(*args[0]) || truthy(*args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::In: {
|
|
auto args = get_args(bc);
|
|
bool result = std::find(args[1]->begin(), args[1]->end(), *args[0]) !=
|
|
args[1]->end();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Equal: {
|
|
auto args = get_args(bc);
|
|
bool result = (*args[0] == *args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Greater: {
|
|
auto args = get_args(bc);
|
|
bool result = (*args[0] > *args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Less: {
|
|
auto args = get_args(bc);
|
|
bool result = (*args[0] < *args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::GreaterEqual: {
|
|
auto args = get_args(bc);
|
|
bool result = (*args[0] >= *args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::LessEqual: {
|
|
auto args = get_args(bc);
|
|
bool result = (*args[0] <= *args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Different: {
|
|
auto args = get_args(bc);
|
|
bool result = (*args[0] != *args[1]);
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Float: {
|
|
double result =
|
|
std::stod(get_args(bc)[0]->get_ref<const std::string&>());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Int: {
|
|
int result = std::stoi(get_args(bc)[0]->get_ref<const std::string&>());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Exists: {
|
|
auto&& name = get_args(bc)[0]->get_ref<const std::string&>();
|
|
bool result = (data.find(name) != data.end());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::ExistsInObject: {
|
|
auto args = get_args(bc);
|
|
auto&& name = args[1]->get_ref<const std::string&>();
|
|
bool result = (args[0]->find(name) != args[0]->end());
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::IsBoolean: {
|
|
bool result = get_args(bc)[0]->is_boolean();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::IsNumber: {
|
|
bool result = get_args(bc)[0]->is_number();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::IsInteger: {
|
|
bool result = get_args(bc)[0]->is_number_integer();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::IsFloat: {
|
|
bool result = get_args(bc)[0]->is_number_float();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::IsObject: {
|
|
bool result = get_args(bc)[0]->is_object();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::IsArray: {
|
|
bool result = get_args(bc)[0]->is_array();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::IsString: {
|
|
bool result = get_args(bc)[0]->is_string();
|
|
pop_args(bc);
|
|
m_stack.emplace_back(result);
|
|
break;
|
|
}
|
|
case Bytecode::Op::Default: {
|
|
// default needs to be a bit "magic"; we can't evaluate the first
|
|
// argument during the push operation, so we swap the arguments during
|
|
// the parse phase so the second argument is pushed on the stack and
|
|
// the first argument is in the immediate
|
|
try {
|
|
const json* imm = get_imm(bc);
|
|
// if no exception was raised, replace the stack value with it
|
|
m_stack.back() = *imm;
|
|
} catch (std::exception&) {
|
|
// couldn't read immediate, just leave the stack as is
|
|
}
|
|
break;
|
|
}
|
|
case Bytecode::Op::Include:
|
|
Renderer(m_included_templates, m_callbacks).render_to(os, m_included_templates.find(get_imm(bc)->get_ref<const std::string&>())->second, *m_data);
|
|
break;
|
|
case Bytecode::Op::Callback: {
|
|
auto callback = m_callbacks.find_callback(bc.str, bc.args);
|
|
if (!callback) {
|
|
inja_throw("render_error", "function '" + static_cast<std::string>(bc.str) + "' (" + std::to_string(static_cast<unsigned int>(bc.args)) + ") not found");
|
|
}
|
|
json result = callback(get_args(bc));
|
|
pop_args(bc);
|
|
m_stack.emplace_back(std::move(result));
|
|
break;
|
|
}
|
|
case Bytecode::Op::Jump: {
|
|
i = bc.args - 1; // -1 due to ++i in loop
|
|
break;
|
|
}
|
|
case Bytecode::Op::ConditionalJump: {
|
|
if (!truthy(m_stack.back())) {
|
|
i = bc.args - 1; // -1 due to ++i in loop
|
|
}
|
|
m_stack.pop_back();
|
|
break;
|
|
}
|
|
case Bytecode::Op::StartLoop: {
|
|
// jump past loop body if empty
|
|
if (m_stack.back().empty()) {
|
|
m_stack.pop_back();
|
|
i = bc.args; // ++i in loop will take it past EndLoop
|
|
break;
|
|
}
|
|
|
|
m_loop_stack.emplace_back();
|
|
LoopLevel& level = m_loop_stack.back();
|
|
level.value_name = bc.str;
|
|
level.values = std::move(m_stack.back());
|
|
level.data = (*m_data);
|
|
m_stack.pop_back();
|
|
|
|
if (bc.value.is_string()) {
|
|
// map iterator
|
|
if (!level.values.is_object()) {
|
|
m_loop_stack.pop_back();
|
|
inja_throw("render_error", "for key, value requires object");
|
|
}
|
|
level.loop_type = LoopLevel::Type::Map;
|
|
level.key_name = bc.value.get_ref<const std::string&>();
|
|
|
|
// sort by key
|
|
for (auto it = level.values.begin(), end = level.values.end(); it != end; ++it) {
|
|
level.map_values.emplace_back(it.key(), &it.value());
|
|
}
|
|
std::sort(level.map_values.begin(), level.map_values.end(), [](const LoopLevel::KeyValue& a, const LoopLevel::KeyValue& b) { return a.first < b.first; });
|
|
level.map_it = level.map_values.begin();
|
|
} else {
|
|
if (!level.values.is_array()) {
|
|
m_loop_stack.pop_back();
|
|
inja_throw("render_error", "type must be array");
|
|
}
|
|
|
|
// list iterator
|
|
level.loop_type = LoopLevel::Type::Array;
|
|
level.index = 0;
|
|
level.size = level.values.size();
|
|
}
|
|
|
|
// provide parent access in nested loop
|
|
auto parent_loop_it = level.data.find("loop");
|
|
if (parent_loop_it != level.data.end()) {
|
|
json loop_copy = *parent_loop_it;
|
|
(*parent_loop_it)["parent"] = std::move(loop_copy);
|
|
}
|
|
|
|
// set "current" data to loop data
|
|
m_data = &level.data;
|
|
update_loop_data();
|
|
break;
|
|
}
|
|
case Bytecode::Op::EndLoop: {
|
|
if (m_loop_stack.empty()) {
|
|
inja_throw("render_error", "unexpected state in renderer");
|
|
}
|
|
LoopLevel& level = m_loop_stack.back();
|
|
|
|
bool done;
|
|
if (level.loop_type == LoopLevel::Type::Array) {
|
|
level.index += 1;
|
|
done = (level.index == level.values.size());
|
|
} else {
|
|
level.map_it += 1;
|
|
done = (level.map_it == level.map_values.end());
|
|
}
|
|
|
|
if (done) {
|
|
m_loop_stack.pop_back();
|
|
// set "current" data to outer loop data or main data as appropriate
|
|
if (!m_loop_stack.empty()) {
|
|
m_data = &m_loop_stack.back().data;
|
|
} else {
|
|
m_data = &data;
|
|
}
|
|
break;
|
|
}
|
|
|
|
update_loop_data();
|
|
|
|
// jump back to start of loop
|
|
i = bc.args - 1; // -1 due to ++i in loop
|
|
break;
|
|
}
|
|
default: {
|
|
inja_throw("render_error", "unknown op in renderer: " + std::to_string(static_cast<unsigned int>(bc.op)));
|
|
}
|
|
}
|
|
}
|
|
}
|
|
};
|
|
|
|
} // namespace inja
|
|
|
|
#endif // PANTOR_INJA_RENDERER_HPP
|
|
|
|
// #include "string_view.hpp"
|
|
|
|
// #include "template.hpp"
|
|
|
|
// #include "utils.hpp"
|
|
|
|
|
|
|
|
namespace inja {
|
|
|
|
using namespace nlohmann;
|
|
|
|
/*!
|
|
* \brief Class for changing the configuration.
|
|
*/
|
|
class Environment {
|
|
class Impl {
|
|
public:
|
|
std::string input_path;
|
|
std::string output_path;
|
|
|
|
LexerConfig lexer_config;
|
|
ParserConfig parser_config;
|
|
|
|
FunctionStorage callbacks;
|
|
TemplateStorage included_templates;
|
|
};
|
|
|
|
std::unique_ptr<Impl> m_impl;
|
|
|
|
public:
|
|
Environment(): Environment("") { }
|
|
|
|
explicit Environment(const std::string& global_path): m_impl(stdinja::make_unique<Impl>()) {
|
|
m_impl->input_path = global_path;
|
|
m_impl->output_path = global_path;
|
|
}
|
|
|
|
explicit Environment(const std::string& input_path, const std::string& output_path): m_impl(stdinja::make_unique<Impl>()) {
|
|
m_impl->input_path = input_path;
|
|
m_impl->output_path = output_path;
|
|
}
|
|
|
|
/// Sets the opener and closer for template statements
|
|
void set_statement(const std::string& open, const std::string& close) {
|
|
m_impl->lexer_config.statement_open = open;
|
|
m_impl->lexer_config.statement_close = close;
|
|
m_impl->lexer_config.update_open_chars();
|
|
}
|
|
|
|
/// Sets the opener for template line statements
|
|
void set_line_statement(const std::string& open) {
|
|
m_impl->lexer_config.line_statement = open;
|
|
m_impl->lexer_config.update_open_chars();
|
|
}
|
|
|
|
/// Sets the opener and closer for template expressions
|
|
void set_expression(const std::string& open, const std::string& close) {
|
|
m_impl->lexer_config.expression_open = open;
|
|
m_impl->lexer_config.expression_close = close;
|
|
m_impl->lexer_config.update_open_chars();
|
|
}
|
|
|
|
/// Sets the opener and closer for template comments
|
|
void set_comment(const std::string& open, const std::string& close) {
|
|
m_impl->lexer_config.comment_open = open;
|
|
m_impl->lexer_config.comment_close = close;
|
|
m_impl->lexer_config.update_open_chars();
|
|
}
|
|
|
|
/// Sets whether to remove the first newline after a block
|
|
void set_trim_blocks(bool trim_blocks) {
|
|
m_impl->lexer_config.trim_blocks = trim_blocks;
|
|
}
|
|
|
|
/// Sets whether to strip the spaces and tabs from the start of a line to a block
|
|
void set_lstrip_blocks(bool lstrip_blocks) {
|
|
m_impl->lexer_config.lstrip_blocks = lstrip_blocks;
|
|
}
|
|
|
|
/// Sets the element notation syntax
|
|
void set_element_notation(ElementNotation notation) {
|
|
m_impl->parser_config.notation = notation;
|
|
}
|
|
|
|
|
|
Template parse(nonstd::string_view input) {
|
|
Parser parser(m_impl->parser_config, m_impl->lexer_config, m_impl->included_templates);
|
|
return parser.parse(input);
|
|
}
|
|
|
|
Template parse_template(const std::string& filename) {
|
|
Parser parser(m_impl->parser_config, m_impl->lexer_config, m_impl->included_templates);
|
|
return parser.parse_template(m_impl->input_path + static_cast<std::string>(filename));
|
|
}
|
|
|
|
std::string render(nonstd::string_view input, const json& data) {
|
|
return render(parse(input), data);
|
|
}
|
|
|
|
std::string render(const Template& tmpl, const json& data) {
|
|
std::stringstream os;
|
|
render_to(os, tmpl, data);
|
|
return os.str();
|
|
}
|
|
|
|
std::string render_file(const std::string& filename, const json& data) {
|
|
return render(parse_template(filename), data);
|
|
}
|
|
|
|
std::string render_file_with_json_file(const std::string& filename, const std::string& filename_data) {
|
|
const json data = load_json(filename_data);
|
|
return render_file(filename, data);
|
|
}
|
|
|
|
void write(const std::string& filename, const json& data, const std::string& filename_out) {
|
|
std::ofstream file(m_impl->output_path + filename_out);
|
|
file << render_file(filename, data);
|
|
file.close();
|
|
}
|
|
|
|
void write(const Template& temp, const json& data, const std::string& filename_out) {
|
|
std::ofstream file(m_impl->output_path + filename_out);
|
|
file << render(temp, data);
|
|
file.close();
|
|
}
|
|
|
|
void write_with_json_file(const std::string& filename, const std::string& filename_data, const std::string& filename_out) {
|
|
const json data = load_json(filename_data);
|
|
write(filename, data, filename_out);
|
|
}
|
|
|
|
void write_with_json_file(const Template& temp, const std::string& filename_data, const std::string& filename_out) {
|
|
const json data = load_json(filename_data);
|
|
write(temp, data, filename_out);
|
|
}
|
|
|
|
std::ostream& render_to(std::ostream& os, const Template& tmpl, const json& data) {
|
|
Renderer(m_impl->included_templates, m_impl->callbacks).render_to(os, tmpl, data);
|
|
return os;
|
|
}
|
|
|
|
std::string load_file(const std::string& filename) {
|
|
Parser parser(m_impl->parser_config, m_impl->lexer_config, m_impl->included_templates);
|
|
return parser.load_file(m_impl->input_path + filename);
|
|
}
|
|
|
|
json load_json(const std::string& filename) {
|
|
std::ifstream file = open_file_or_throw(m_impl->input_path + filename);
|
|
json j;
|
|
file >> j;
|
|
return j;
|
|
}
|
|
|
|
void add_callback(const std::string& name, unsigned int numArgs, const CallbackFunction& callback) {
|
|
m_impl->callbacks.add_callback(name, numArgs, callback);
|
|
}
|
|
|
|
/** Includes a template with a given name into the environment.
|
|
* Then, a template can be rendered in another template using the
|
|
* include "<name>" syntax.
|
|
*/
|
|
void include_template(const std::string& name, const Template& tmpl) {
|
|
m_impl->included_templates[name] = tmpl;
|
|
}
|
|
};
|
|
|
|
/*!
|
|
@brief render with default settings to a string
|
|
*/
|
|
inline std::string render(nonstd::string_view input, const json& data) {
|
|
return Environment().render(input, data);
|
|
}
|
|
|
|
/*!
|
|
@brief render with default settings to the given output stream
|
|
*/
|
|
inline void render_to(std::ostream& os, nonstd::string_view input, const json& data) {
|
|
Environment env;
|
|
env.render_to(os, env.parse(input), data);
|
|
}
|
|
|
|
}
|
|
|
|
#endif // PANTOR_INJA_ENVIRONMENT_HPP
|
|
|
|
// #include "string_view.hpp"
|
|
|
|
// #include "template.hpp"
|
|
|
|
// #include "parser.hpp"
|
|
|
|
// #include "renderer.hpp"
|
|
|
|
|
|
|
|
#endif // PANTOR_INJA_HPP
|