jpayne@69: // Copyright (c) 2013-2014 Sandstorm Development Group, Inc. and contributors
jpayne@69: // Licensed under the MIT License:
jpayne@69: //
jpayne@69: // Permission is hereby granted, free of charge, to any person obtaining a copy
jpayne@69: // of this software and associated documentation files (the "Software"), to deal
jpayne@69: // in the Software without restriction, including without limitation the rights
jpayne@69: // to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
jpayne@69: // copies of the Software, and to permit persons to whom the Software is
jpayne@69: // furnished to do so, subject to the following conditions:
jpayne@69: //
jpayne@69: // The above copyright notice and this permission notice shall be included in
jpayne@69: // all copies or substantial portions of the Software.
jpayne@69: //
jpayne@69: // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
jpayne@69: // IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
jpayne@69: // FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
jpayne@69: // AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
jpayne@69: // LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
jpayne@69: // OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
jpayne@69: // THE SOFTWARE.
jpayne@69: 
jpayne@69: // This file contains parsers useful for character stream inputs, including parsers to parse
jpayne@69: // common kinds of tokens like identifiers, numbers, and quoted strings.
jpayne@69: 
jpayne@69: #pragma once
jpayne@69: 
jpayne@69: #include "common.h"
jpayne@69: #include "../string.h"
jpayne@69: #include <inttypes.h>
jpayne@69: 
jpayne@69: KJ_BEGIN_HEADER
jpayne@69: 
jpayne@69: namespace kj {
jpayne@69: namespace parse {
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: // Exact char/string.
jpayne@69: 
jpayne@69: class ExactString_ {
jpayne@69: public:
jpayne@69:   constexpr inline ExactString_(const char* str): str(str) {}
jpayne@69: 
jpayne@69:   template <typename Input>
jpayne@69:   Maybe<Tuple<>> operator()(Input& input) const {
jpayne@69:     const char* ptr = str;
jpayne@69: 
jpayne@69:     while (*ptr != '\0') {
jpayne@69:       if (input.atEnd() || input.current() != *ptr) return nullptr;
jpayne@69:       input.next();
jpayne@69:       ++ptr;
jpayne@69:     }
jpayne@69: 
jpayne@69:     return Tuple<>();
jpayne@69:   }
jpayne@69: 
jpayne@69: private:
jpayne@69:   const char* str;
jpayne@69: };
jpayne@69: 
jpayne@69: constexpr inline ExactString_ exactString(const char* str) {
jpayne@69:   return ExactString_(str);
jpayne@69: }
jpayne@69: 
jpayne@69: template <char c>
jpayne@69: constexpr ExactlyConst_<char, c> exactChar() {
jpayne@69:   // Returns a parser that matches exactly the character given by the template argument (returning
jpayne@69:   // no result).
jpayne@69:   return ExactlyConst_<char, c>();
jpayne@69: }
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: // Char ranges / sets
jpayne@69: 
jpayne@69: class CharGroup_ {
jpayne@69: public:
jpayne@69:   constexpr inline CharGroup_(): bits{0, 0, 0, 0} {}
jpayne@69: 
jpayne@69:   constexpr inline CharGroup_ orRange(unsigned char first, unsigned char last) const {
jpayne@69:     return CharGroup_(bits[0] | (oneBits(last +   1) & ~oneBits(first      )),
jpayne@69:                       bits[1] | (oneBits(last -  63) & ~oneBits(first -  64)),
jpayne@69:                       bits[2] | (oneBits(last - 127) & ~oneBits(first - 128)),
jpayne@69:                       bits[3] | (oneBits(last - 191) & ~oneBits(first - 192)));
jpayne@69:   }
jpayne@69: 
jpayne@69:   constexpr inline CharGroup_ orAny(const char* chars) const {
jpayne@69:     return *chars == 0 ? *this : orChar(*chars).orAny(chars + 1);
jpayne@69:   }
jpayne@69: 
jpayne@69:   constexpr inline CharGroup_ orChar(unsigned char c) const {
jpayne@69:     return CharGroup_(bits[0] | bit(c),
jpayne@69:                       bits[1] | bit(c - 64),
jpayne@69:                       bits[2] | bit(c - 128),
jpayne@69:                       bits[3] | bit(c - 256));
jpayne@69:   }
jpayne@69: 
jpayne@69:   constexpr inline CharGroup_ orGroup(CharGroup_ other) const {
jpayne@69:     return CharGroup_(bits[0] | other.bits[0],
jpayne@69:                       bits[1] | other.bits[1],
jpayne@69:                       bits[2] | other.bits[2],
jpayne@69:                       bits[3] | other.bits[3]);
jpayne@69:   }
jpayne@69: 
jpayne@69:   constexpr inline CharGroup_ invert() const {
jpayne@69:     return CharGroup_(~bits[0], ~bits[1], ~bits[2], ~bits[3]);
jpayne@69:   }
jpayne@69: 
jpayne@69:   constexpr inline bool contains(unsigned char c) const {
jpayne@69:     return (bits[c / 64] & (1ll << (c % 64))) != 0;
jpayne@69:   }
jpayne@69: 
jpayne@69:   inline bool containsAll(ArrayPtr<const char> text) const {
jpayne@69:     for (char c: text) {
jpayne@69:       if (!contains(c)) return false;
jpayne@69:     }
jpayne@69:     return true;
jpayne@69:   }
jpayne@69: 
jpayne@69:   template <typename Input>
jpayne@69:   Maybe<char> operator()(Input& input) const {
jpayne@69:     if (input.atEnd()) return nullptr;
jpayne@69:     unsigned char c = input.current();
jpayne@69:     if (contains(c)) {
jpayne@69:       input.next();
jpayne@69:       return c;
jpayne@69:     } else {
jpayne@69:       return nullptr;
jpayne@69:     }
jpayne@69:   }
jpayne@69: 
jpayne@69: private:
jpayne@69:   typedef unsigned long long Bits64;
jpayne@69: 
jpayne@69:   constexpr inline CharGroup_(Bits64 a, Bits64 b, Bits64 c, Bits64 d): bits{a, b, c, d} {}
jpayne@69:   Bits64 bits[4];
jpayne@69: 
jpayne@69:   static constexpr inline Bits64 oneBits(int count) {
jpayne@69:     return count <= 0 ? 0ll : count >= 64 ? -1ll : ((1ll << count) - 1);
jpayne@69:   }
jpayne@69:   static constexpr inline Bits64 bit(int index) {
jpayne@69:     return index < 0 ? 0 : index >= 64 ? 0 : (1ll << index);
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: constexpr inline CharGroup_ charRange(char first, char last) {
jpayne@69:   // Create a parser which accepts any character in the range from `first` to `last`, inclusive.
jpayne@69:   // For example: `charRange('a', 'z')` matches all lower-case letters.  The parser's result is the
jpayne@69:   // character matched.
jpayne@69:   //
jpayne@69:   // The returned object has methods which can be used to match more characters.  The following
jpayne@69:   // produces a parser which accepts any letter as well as '_', '+', '-', and '.'.
jpayne@69:   //
jpayne@69:   //     charRange('a', 'z').orRange('A', 'Z').orChar('_').orAny("+-.")
jpayne@69:   //
jpayne@69:   // You can also use `.invert()` to match the opposite set of characters.
jpayne@69: 
jpayne@69:   return CharGroup_().orRange(first, last);
jpayne@69: }
jpayne@69: 
jpayne@69: #if _MSC_VER && !defined(__clang__)
jpayne@69: #define anyOfChars(chars) CharGroup_().orAny(chars)
jpayne@69: // TODO(msvc): MSVC ICEs on the proper definition of `anyOfChars()`, which in turn prevents us from
jpayne@69: //   building the compiler or schema parser. We don't know why this happens, but Harris found that
jpayne@69: //   this horrible, horrible hack makes things work. This is awful, but it's better than nothing.
jpayne@69: //   Hopefully, MSVC will get fixed soon and we'll be able to remove this.
jpayne@69: #else
jpayne@69: constexpr inline CharGroup_ anyOfChars(const char* chars) {
jpayne@69:   // Returns a parser that accepts any of the characters in the given string (which should usually
jpayne@69:   // be a literal).  The returned parser is of the same type as returned by `charRange()` -- see
jpayne@69:   // that function for more info.
jpayne@69: 
jpayne@69:   return CharGroup_().orAny(chars);
jpayne@69: }
jpayne@69: #endif
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: 
jpayne@69: namespace _ {  // private
jpayne@69: 
jpayne@69: struct ArrayToString {
jpayne@69:   inline String operator()(const Array<char>& arr) const {
jpayne@69:     return heapString(arr);
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: }  // namespace _ (private)
jpayne@69: 
jpayne@69: template <typename SubParser>
jpayne@69: constexpr inline auto charsToString(SubParser&& subParser)
jpayne@69:     -> decltype(transform(kj::fwd<SubParser>(subParser), _::ArrayToString())) {
jpayne@69:   // Wraps a parser that returns Array<char> such that it returns String instead.
jpayne@69:   return parse::transform(kj::fwd<SubParser>(subParser), _::ArrayToString());
jpayne@69: }
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: // Basic character classes.
jpayne@69: 
jpayne@69: constexpr auto alpha = charRange('a', 'z').orRange('A', 'Z');
jpayne@69: constexpr auto digit = charRange('0', '9');
jpayne@69: constexpr auto alphaNumeric = alpha.orGroup(digit);
jpayne@69: constexpr auto nameStart = alpha.orChar('_');
jpayne@69: constexpr auto nameChar = alphaNumeric.orChar('_');
jpayne@69: constexpr auto hexDigit = charRange('0', '9').orRange('a', 'f').orRange('A', 'F');
jpayne@69: constexpr auto octDigit = charRange('0', '7');
jpayne@69: constexpr auto whitespaceChar = anyOfChars(" \f\n\r\t\v");
jpayne@69: constexpr auto controlChar = charRange(0, 0x1f).invert().orGroup(whitespaceChar).invert();
jpayne@69: 
jpayne@69: constexpr auto whitespace = many(anyOfChars(" \f\n\r\t\v"));
jpayne@69: 
jpayne@69: constexpr auto discardWhitespace = discard(many(discard(anyOfChars(" \f\n\r\t\v"))));
jpayne@69: // Like discard(whitespace) but avoids some memory allocation.
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: // Identifiers
jpayne@69: 
jpayne@69: namespace _ { // private
jpayne@69: 
jpayne@69: struct IdentifierToString {
jpayne@69:   inline String operator()(char first, const Array<char>& rest) const {
jpayne@69:     if (rest.size() == 0) return heapString(&first, 1);
jpayne@69:     String result = heapString(rest.size() + 1);
jpayne@69:     result[0] = first;
jpayne@69:     memcpy(result.begin() + 1, rest.begin(), rest.size());
jpayne@69:     return result;
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: }  // namespace _ (private)
jpayne@69: 
jpayne@69: constexpr auto identifier = transform(sequence(nameStart, many(nameChar)), _::IdentifierToString());
jpayne@69: // Parses an identifier (e.g. a C variable name).
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: // Integers
jpayne@69: 
jpayne@69: namespace _ {  // private
jpayne@69: 
jpayne@69: inline char parseDigit(char c) {
jpayne@69:   if (c < 'A') return c - '0';
jpayne@69:   if (c < 'a') return c - 'A' + 10;
jpayne@69:   return c - 'a' + 10;
jpayne@69: }
jpayne@69: 
jpayne@69: template <uint base>
jpayne@69: struct ParseInteger {
jpayne@69:   inline uint64_t operator()(const Array<char>& digits) const {
jpayne@69:     return operator()('0', digits);
jpayne@69:   }
jpayne@69:   uint64_t operator()(char first, const Array<char>& digits) const {
jpayne@69:     uint64_t result = parseDigit(first);
jpayne@69:     for (char digit: digits) {
jpayne@69:       result = result * base + parseDigit(digit);
jpayne@69:     }
jpayne@69:     return result;
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: 
jpayne@69: }  // namespace _ (private)
jpayne@69: 
jpayne@69: constexpr auto integer = sequence(
jpayne@69:     oneOf(
jpayne@69:       transform(sequence(exactChar<'0'>(), exactChar<'x'>(), oneOrMore(hexDigit)), _::ParseInteger<16>()),
jpayne@69:       transform(sequence(exactChar<'0'>(), many(octDigit)), _::ParseInteger<8>()),
jpayne@69:       transform(sequence(charRange('1', '9'), many(digit)), _::ParseInteger<10>())),
jpayne@69:     notLookingAt(alpha.orAny("_.")));
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: // Numbers (i.e. floats)
jpayne@69: 
jpayne@69: namespace _ {  // private
jpayne@69: 
jpayne@69: struct ParseFloat {
jpayne@69:   double operator()(const Array<char>& digits,
jpayne@69:                     const Maybe<Array<char>>& fraction,
jpayne@69:                     const Maybe<Tuple<Maybe<char>, Array<char>>>& exponent) const;
jpayne@69: };
jpayne@69: 
jpayne@69: }  // namespace _ (private)
jpayne@69: 
jpayne@69: constexpr auto number = transform(
jpayne@69:     sequence(
jpayne@69:         oneOrMore(digit),
jpayne@69:         optional(sequence(exactChar<'.'>(), many(digit))),
jpayne@69:         optional(sequence(discard(anyOfChars("eE")), optional(anyOfChars("+-")), many(digit))),
jpayne@69:         notLookingAt(alpha.orAny("_."))),
jpayne@69:     _::ParseFloat());
jpayne@69: 
jpayne@69: // =======================================================================================
jpayne@69: // Quoted strings
jpayne@69: 
jpayne@69: namespace _ {  // private
jpayne@69: 
jpayne@69: struct InterpretEscape {
jpayne@69:   char operator()(char c) const {
jpayne@69:     switch (c) {
jpayne@69:       case 'a': return '\a';
jpayne@69:       case 'b': return '\b';
jpayne@69:       case 'f': return '\f';
jpayne@69:       case 'n': return '\n';
jpayne@69:       case 'r': return '\r';
jpayne@69:       case 't': return '\t';
jpayne@69:       case 'v': return '\v';
jpayne@69:       default: return c;
jpayne@69:     }
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: struct ParseHexEscape {
jpayne@69:   inline char operator()(char first, char second) const {
jpayne@69:     return (parseDigit(first) << 4) | parseDigit(second);
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: struct ParseHexByte {
jpayne@69:   inline byte operator()(char first, char second) const {
jpayne@69:     return (parseDigit(first) << 4) | parseDigit(second);
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: struct ParseOctEscape {
jpayne@69:   inline char operator()(char first, Maybe<char> second, Maybe<char> third) const {
jpayne@69:     char result = first - '0';
jpayne@69:     KJ_IF_MAYBE(digit1, second) {
jpayne@69:       result = (result << 3) | (*digit1 - '0');
jpayne@69:       KJ_IF_MAYBE(digit2, third) {
jpayne@69:         result = (result << 3) | (*digit2 - '0');
jpayne@69:       }
jpayne@69:     }
jpayne@69:     return result;
jpayne@69:   }
jpayne@69: };
jpayne@69: 
jpayne@69: }  // namespace _ (private)
jpayne@69: 
jpayne@69: constexpr auto escapeSequence =
jpayne@69:     sequence(exactChar<'\\'>(), oneOf(
jpayne@69:         transform(anyOfChars("abfnrtv'\"\\\?"), _::InterpretEscape()),
jpayne@69:         transform(sequence(exactChar<'x'>(), hexDigit, hexDigit), _::ParseHexEscape()),
jpayne@69:         transform(sequence(octDigit, optional(octDigit), optional(octDigit)),
jpayne@69:                   _::ParseOctEscape())));
jpayne@69: // A parser that parses a C-string-style escape sequence (starting with a backslash).  Returns
jpayne@69: // a char.
jpayne@69: 
jpayne@69: constexpr auto doubleQuotedString = charsToString(sequence(
jpayne@69:     exactChar<'\"'>(),
jpayne@69:     many(oneOf(anyOfChars("\\\n\"").invert(), escapeSequence)),
jpayne@69:     exactChar<'\"'>()));
jpayne@69: // Parses a C-style double-quoted string.
jpayne@69: 
jpayne@69: constexpr auto singleQuotedString = charsToString(sequence(
jpayne@69:     exactChar<'\''>(),
jpayne@69:     many(oneOf(anyOfChars("\\\n\'").invert(), escapeSequence)),
jpayne@69:     exactChar<'\''>()));
jpayne@69: // Parses a C-style single-quoted string.
jpayne@69: 
jpayne@69: constexpr auto doubleQuotedHexBinary = sequence(
jpayne@69:     exactChar<'0'>(), exactChar<'x'>(), exactChar<'\"'>(),
jpayne@69:     oneOrMore(transform(sequence(discardWhitespace, hexDigit, hexDigit), _::ParseHexByte())),
jpayne@69:     discardWhitespace,
jpayne@69:     exactChar<'\"'>());
jpayne@69: // Parses a double-quoted hex binary literal. Returns Array<byte>.
jpayne@69: 
jpayne@69: }  // namespace parse
jpayne@69: }  // namespace kj
jpayne@69: 
jpayne@69: KJ_END_HEADER