/* Copyright 2015 Google Inc. All rights reserved. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. */ #ifndef JSONNET_LEXER_H #define JSONNET_LEXER_H #include #include #include #include #include #include #include #include "static_error.h" #include "unicode.h" namespace jsonnet::internal { /** Whitespace and comments. * * "Fodder" (as in cannon fodder) implies this data is expendable. */ struct FodderElement { enum Kind { /** The next token, paragraph, or interstitial should be on a new line. * * A single comment string is allowed, which flows before the new line. * * The LINE_END fodder specifies the indentation level and vertical spacing before whatever * comes next. */ LINE_END, /** A C-style comment that begins and ends on the same line. * * If it follows a token (i.e., it is the first fodder element) then it appears after the * token on the same line. If it follows another interstitial, it will also flow after it * on the same line. If it follows a new line or a paragraph, it is the first thing on the * following line, after the blank lines and indentation specified by the previous fodder. * * There is exactly one comment string. */ INTERSTITIAL, /** A comment consisting of at least one line. * * // and # style commes have exactly one line. C-style comments can have more than one * line. * * All lines of the comment are indented according to the indentation level of the previous * new line / paragraph fodder. * * The PARAGRAPH fodder specifies the indentation level and vertical spacing before whatever * comes next. */ PARAGRAPH, }; Kind kind; /** How many blank lines (vertical space) before the next fodder / token. */ unsigned blanks; /** How far the next fodder / token should be indented. */ unsigned indent; /** Whatever comments are part of this fodder. * * Constraints apply. See Kind, above. * * The strings include any delimiting characters, e.g. // # and C-style comment delimiters but * not newline characters or indentation. */ std::vector comment; FodderElement(Kind kind, unsigned blanks, unsigned indent, const std::vector &comment) : kind(kind), blanks(blanks), indent(indent), comment(comment) { assert(kind != LINE_END || comment.size() <= 1); assert(kind != INTERSTITIAL || (blanks == 0 && indent == 0 && comment.size() == 1)); assert(kind != PARAGRAPH || comment.size() >= 1); } }; static inline std::ostream &operator<<(std::ostream &o, const FodderElement &f) { switch (f.kind) { case FodderElement::LINE_END: o << "END(" << f.blanks << ", " << f.indent; if (!f.comment.empty()) { o << ", " << f.comment[0]; } o << ")"; break; case FodderElement::INTERSTITIAL: o << "INT(" << f.blanks << ", " << f.indent << ", " << f.comment[0] << ")"; break; case FodderElement::PARAGRAPH: o << "PAR(" << f.blanks << ", " << f.indent << ", " << f.comment[0] << "...)"; break; } return o; } /** A sequence of fodder elements, typically between two tokens. * * A LINE_END is not allowed to follow a PARAGRAPH or a LINE_END. This can be represented by * replacing the indent of the prior fodder and increasing the number of blank lines if necessary. * If there was a comment, it can be represented by changing the LINE_END to a paragraph containing * the same single comment string. * * There must be a LINE_END or a PARAGRAPH before a PARAGRAPH. * * TODO(sbarzowski) Make it a proper class */ typedef std::vector Fodder; static inline bool fodder_has_clean_endline(const Fodder &fodder) { return !fodder.empty() && fodder.back().kind != FodderElement::INTERSTITIAL; } /** As a.push_back(elem) but preserves constraints. * * See concat_fodder below. */ static inline void fodder_push_back(Fodder &a, const FodderElement &elem) { if (fodder_has_clean_endline(a) && elem.kind == FodderElement::LINE_END) { if (elem.comment.size() > 0) { // The line end had a comment, so create a single line paragraph for it. a.emplace_back(FodderElement::PARAGRAPH, elem.blanks, elem.indent, elem.comment); } else { // Merge it into the previous line end. a.back().indent = elem.indent; a.back().blanks += elem.blanks; } } else { if (!fodder_has_clean_endline(a) && elem.kind == FodderElement::PARAGRAPH) { a.emplace_back(FodderElement::LINE_END, 0, elem.indent, std::vector()); } a.push_back(elem); } } /** As a + b but preserves constraints. * * Namely, a LINE_END is not allowed to follow a PARAGRAPH or a LINE_END. */ static inline Fodder concat_fodder(const Fodder &a, const Fodder &b) { if (a.size() == 0) return b; if (b.size() == 0) return a; Fodder r = a; // Carefully add the first element of b. fodder_push_back(r, b[0]); // Add the rest of b. for (unsigned i = 1; i < b.size(); ++i) { r.push_back(b[i]); } return r; } /** Move b to the front of a. */ static inline void fodder_move_front(Fodder &a, Fodder &b) { a = concat_fodder(b, a); b.clear(); } static inline Fodder make_fodder(const FodderElement &elem) { Fodder fodder; fodder_push_back(fodder, elem); return fodder; } static inline void ensureCleanNewline(Fodder &fodder) { if (!fodder_has_clean_endline(fodder)) { fodder_push_back(fodder, FodderElement(FodderElement::Kind::LINE_END, 0, 0, {})); } } static inline int countNewlines(const FodderElement &elem) { switch (elem.kind) { case FodderElement::INTERSTITIAL: return 0; case FodderElement::LINE_END: return 1; case FodderElement::PARAGRAPH: return elem.comment.size() + elem.blanks; } std::cerr << "Unknown FodderElement kind" << std::endl; abort(); } static inline int countNewlines(const Fodder &fodder) { int sum = 0; for (const auto &elem : fodder) { sum += countNewlines(elem); } return sum; } static inline std::ostream &operator<<(std::ostream &o, const Fodder &fodder) { bool first = true; for (const auto &f : fodder) { o << (first ? "[" : ", "); first = false; o << f; } o << (first ? "[]" : "]"); return o; } struct Token { enum Kind { // Symbols BRACE_L, BRACE_R, BRACKET_L, BRACKET_R, COMMA, DOLLAR, DOT, PAREN_L, PAREN_R, SEMICOLON, // Arbitrary length lexemes IDENTIFIER, NUMBER, OPERATOR, STRING_DOUBLE, STRING_SINGLE, STRING_BLOCK, VERBATIM_STRING_SINGLE, VERBATIM_STRING_DOUBLE, // Keywords ASSERT, ELSE, ERROR, FALSE, FOR, FUNCTION, IF, IMPORT, IMPORTSTR, IMPORTBIN, IN, LOCAL, NULL_LIT, TAILSTRICT, THEN, SELF, SUPER, TRUE, // A special token that holds line/column information about the end of the file. END_OF_FILE } kind; /** Fodder before this token. */ Fodder fodder; /** Content of the token if it wasn't a keyword. */ std::string data; /** If kind == STRING_BLOCK then stores the sequence of whitespace that indented the block. */ std::string stringBlockIndent; /** If kind == STRING_BLOCK then stores the sequence of whitespace that indented the end of * the block. * * This is always fewer whitespace characters than in stringBlockIndent. */ std::string stringBlockTermIndent; UString data32(void) const { return decode_utf8(data); } LocationRange location; Token(Kind kind, const Fodder &fodder, const std::string &data, const std::string &string_block_indent, const std::string &string_block_term_indent, const LocationRange &location) : kind(kind), fodder(fodder), data(data), stringBlockIndent(string_block_indent), stringBlockTermIndent(string_block_term_indent), location(location) { } Token(Kind kind, const std::string &data = "") : kind(kind), data(data) {} static const char *toString(Kind v) { switch (v) { case BRACE_L: return "\"{\""; case BRACE_R: return "\"}\""; case BRACKET_L: return "\"[\""; case BRACKET_R: return "\"]\""; case COMMA: return "\",\""; case DOLLAR: return "\"$\""; case DOT: return "\".\""; case PAREN_L: return "\"(\""; case PAREN_R: return "\")\""; case SEMICOLON: return "\";\""; case IDENTIFIER: return "IDENTIFIER"; case NUMBER: return "NUMBER"; case OPERATOR: return "OPERATOR"; case STRING_SINGLE: return "STRING_SINGLE"; case STRING_DOUBLE: return "STRING_DOUBLE"; case VERBATIM_STRING_SINGLE: return "VERBATIM_STRING_SINGLE"; case VERBATIM_STRING_DOUBLE: return "VERBATIM_STRING_DOUBLE"; case STRING_BLOCK: return "STRING_BLOCK"; case ASSERT: return "assert"; case ELSE: return "else"; case ERROR: return "error"; case FALSE: return "false"; case FOR: return "for"; case FUNCTION: return "function"; case IF: return "if"; case IMPORT: return "import"; case IMPORTSTR: return "importstr"; case IMPORTBIN: return "importbin"; case IN: return "in"; case LOCAL: return "local"; case NULL_LIT: return "null"; case SELF: return "self"; case SUPER: return "super"; case TAILSTRICT: return "tailstrict"; case THEN: return "then"; case TRUE: return "true"; case END_OF_FILE: return "end of file"; default: std::cerr << "INTERNAL ERROR: Unknown token kind: " << v << std::endl; std::abort(); } } }; /** The result of lexing. * * Because of the EOF token, this will always contain at least one token. So element 0 can be used * to get the filename. */ typedef std::list Tokens; static inline bool operator==(const Token &a, const Token &b) { if (a.kind != b.kind) return false; if (a.data != b.data) return false; return true; } static inline std::ostream &operator<<(std::ostream &o, Token::Kind v) { o << Token::toString(v); return o; } static inline std::ostream &operator<<(std::ostream &o, const Token &v) { if (v.data == "") { o << Token::toString(v.kind); } else if (v.kind == Token::OPERATOR) { o << "\"" << v.data << "\""; } else { o << "(" << Token::toString(v.kind) << ", \"" << v.data << "\")"; } return o; } /** IF the given identifier is a keyword, return its kind, otherwise return IDENTIFIER. */ Token::Kind lex_get_keyword_kind(const std::string &identifier); Tokens jsonnet_lex(const std::string &filename, const char *input); std::string jsonnet_unlex(const Tokens &tokens); } // namespace jsonnet::internal #endif // JSONNET_LEXER_H