169 lines · cpp
1//===-- DILLexer.cpp ------------------------------------------------------===//2//3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.4// See https://llvm.org/LICENSE.txt for license information.5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception6//7// This implements the recursive descent parser for the Data Inspection8// Language (DIL), and its helper functions, which will eventually underlie the9// 'frame variable' command. The language that this parser recognizes is10// described in lldb/docs/dil-expr-lang.ebnf11//12//===----------------------------------------------------------------------===//13 14#include "lldb/ValueObject/DILLexer.h"15#include "lldb/Utility/Status.h"16#include "lldb/ValueObject/DILParser.h"17#include "llvm/ADT/StringSwitch.h"18 19namespace lldb_private::dil {20 21llvm::StringRef Token::GetTokenName(Kind kind) {22 switch (kind) {23 case Kind::amp:24 return "amp";25 case Kind::arrow:26 return "arrow";27 case Kind::coloncolon:28 return "coloncolon";29 case Kind::eof:30 return "eof";31 case Kind::float_constant:32 return "float_constant";33 case Kind::identifier:34 return "identifier";35 case Kind::integer_constant:36 return "integer_constant";37 case Kind::kw_false:38 return "false";39 case Kind::kw_true:40 return "true";41 case Kind::l_paren:42 return "l_paren";43 case Kind::l_square:44 return "l_square";45 case Kind::minus:46 return "minus";47 case Kind::period:48 return "period";49 case Kind::plus:50 return "plus";51 case Kind::r_paren:52 return "r_paren";53 case Kind::r_square:54 return "r_square";55 case Token::star:56 return "star";57 }58 llvm_unreachable("Unknown token name");59}60 61static bool IsLetter(char c) {62 return ('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z');63}64 65static bool IsDigit(char c) { return '0' <= c && c <= '9'; }66 67// A word starts with a letter, underscore, or dollar sign, followed by68// letters ('a'..'z','A'..'Z'), digits ('0'..'9'), and/or underscores.69static std::optional<llvm::StringRef> IsWord(llvm::StringRef expr,70 llvm::StringRef &remainder) {71 // Find the longest prefix consisting of letters, digits, underscors and72 // '$'. If it doesn't start with a digit, then it's a word.73 llvm::StringRef candidate = remainder.take_while(74 [](char c) { return IsDigit(c) || IsLetter(c) || c == '_' || c == '$'; });75 if (candidate.empty() || IsDigit(candidate[0]))76 return std::nullopt;77 remainder = remainder.drop_front(candidate.size());78 return candidate;79}80 81static bool IsNumberBodyChar(char ch) {82 return IsDigit(ch) || IsLetter(ch) || ch == '.';83}84 85static std::optional<llvm::StringRef> IsNumber(llvm::StringRef &remainder,86 bool &isFloat) {87 llvm::StringRef tail = remainder;88 llvm::StringRef body = tail.take_while(IsNumberBodyChar);89 size_t dots = body.count('.');90 if (dots > 1 || dots == body.size())91 return std::nullopt;92 if (IsDigit(body.front()) || (body[0] == '.' && IsDigit(body[1]))) {93 isFloat = dots == 1;94 tail = tail.drop_front(body.size());95 bool isHex = body.contains_insensitive('x');96 bool hasExp = !isHex && body.contains_insensitive('e');97 bool hasHexExp = isHex && body.contains_insensitive('p');98 if (hasExp || hasHexExp) {99 isFloat = true; // This marks numbers like 0x1p1 and 1e1 as float100 if (body.ends_with_insensitive("e") || body.ends_with_insensitive("p"))101 if (tail.consume_front("+") || tail.consume_front("-"))102 tail = tail.drop_while(IsNumberBodyChar);103 }104 size_t number_length = remainder.size() - tail.size();105 llvm::StringRef number = remainder.take_front(number_length);106 remainder = remainder.drop_front(number_length);107 return number;108 }109 return std::nullopt;110}111 112llvm::Expected<DILLexer> DILLexer::Create(llvm::StringRef expr) {113 std::vector<Token> tokens;114 llvm::StringRef remainder = expr;115 do {116 if (llvm::Expected<Token> t = Lex(expr, remainder)) {117 tokens.push_back(std::move(*t));118 } else {119 return t.takeError();120 }121 } while (tokens.back().GetKind() != Token::eof);122 return DILLexer(expr, std::move(tokens));123}124 125llvm::Expected<Token> DILLexer::Lex(llvm::StringRef expr,126 llvm::StringRef &remainder) {127 // Skip over whitespace (spaces).128 remainder = remainder.ltrim();129 llvm::StringRef::iterator cur_pos = remainder.begin();130 131 // Check to see if we've reached the end of our input string.132 if (remainder.empty())133 return Token(Token::eof, "", (uint32_t)expr.size());134 135 uint32_t position = cur_pos - expr.begin();136 bool isFloat = false;137 std::optional<llvm::StringRef> maybe_number = IsNumber(remainder, isFloat);138 if (maybe_number) {139 auto kind = isFloat ? Token::float_constant : Token::integer_constant;140 return Token(kind, maybe_number->str(), position);141 }142 std::optional<llvm::StringRef> maybe_word = IsWord(expr, remainder);143 if (maybe_word) {144 llvm::StringRef word = *maybe_word;145 Token::Kind kind = llvm::StringSwitch<Token::Kind>(word)146 .Case("false", Token::kw_false)147 .Case("true", Token::kw_true)148 .Default(Token::identifier);149 return Token(kind, word.str(), position);150 }151 152 constexpr std::pair<Token::Kind, const char *> operators[] = {153 {Token::amp, "&"}, {Token::arrow, "->"}, {Token::coloncolon, "::"},154 {Token::l_paren, "("}, {Token::l_square, "["}, {Token::minus, "-"},155 {Token::period, "."}, {Token::plus, "+"}, {Token::r_paren, ")"},156 {Token::r_square, "]"}, {Token::star, "*"},157 };158 for (auto [kind, str] : operators) {159 if (remainder.consume_front(str))160 return Token(kind, str, position);161 }162 163 // Unrecognized character(s) in string; unable to lex it.164 return llvm::make_error<DILDiagnosticError>(expr, "unrecognized token",165 position);166}167 168} // namespace lldb_private::dil169