flang/lib/Parser/prescan.h

   1 //===-- lib/Parser/prescan.h ------------------------------------*- C++ -*-===//
   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-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #ifndef FORTRAN_PARSER_PRESCAN_H_
  10 #define FORTRAN_PARSER_PRESCAN_H_
  11
  12 // Defines a fast Fortran source prescanning phase that implements some
  13 // character-level features of the language that can be inefficient to
  14 // support directly in a backtracking parser.  This phase handles Fortran
  15 // line continuation, comment removal, card image margins, padding out
  16 // fixed form character literals on truncated card images, file
  17 // inclusion, and driving the Fortran source preprocessor.
  18
  19 #include "token-sequence.h"
  20 #include "flang/Common/Fortran-features.h"
  21 #include "flang/Parser/characters.h"
  22 #include "flang/Parser/message.h"
  23 #include "flang/Parser/provenance.h"
  24 #include <bitset>
  25 #include <optional>
  26 #include <string>
  27 #include <unordered_set>
  28
  29 namespace Fortran::parser {
  30
  31 class Messages;
  32 class Preprocessor;
  33
  34 class Prescanner {
  35 public:
  36   Prescanner(Messages &, CookedSource &, Preprocessor &,
  37       common::LanguageFeatureControl);
  38   Prescanner(const Prescanner &);
  39
  40   const AllSources &allSources() const { return allSources_; }
  41   AllSources &allSources() { return allSources_; }
  42   const Messages &messages() const { return messages_; }
  43   Messages &messages() { return messages_; }
  44   const Preprocessor &preprocessor() const { return preprocessor_; }
  45   Preprocessor &preprocessor() { return preprocessor_; }
  46
  47   Prescanner &set_fixedForm(bool yes) {
  48     inFixedForm_ = yes;
  49     return *this;
  50   }
  51   Prescanner &set_encoding(Encoding code) {
  52     encoding_ = code;
  53     return *this;
  54   }
  55   Prescanner &set_fixedFormColumnLimit(int limit) {
  56     fixedFormColumnLimit_ = limit;
  57     return *this;
  58   }
  59
  60   Prescanner &AddCompilerDirectiveSentinel(const std::string &);
  61
  62   void Prescan(ProvenanceRange);
  63   void Statement();
  64   void NextLine();
  65
  66   // Callbacks for use by Preprocessor.
  67   bool IsAtEnd() const { return nextLine_ >= limit_; }
  68   bool IsNextLinePreprocessorDirective() const;
  69   TokenSequence TokenizePreprocessorDirective();
  70   Provenance GetCurrentProvenance() const { return GetProvenance(at_); }
  71
  72   template <typename... A> Message &Say(A &&...a) {
  73     return messages_.Say(std::forward<A>(a)...);
  74   }
  75
  76 private:
  77   struct LineClassification {
  78     enum class Kind {
  79       Comment,
  80       ConditionalCompilationDirective,
  81       IncludeDirective, // #include
  82       DefinitionDirective, // #define & #undef
  83       PreprocessorDirective,
  84       IncludeLine, // Fortran INCLUDE
  85       CompilerDirective,
  86       Source
  87     };
  88     LineClassification(Kind k, std::size_t po = 0, const char *s = nullptr)
  89         : kind{k}, payloadOffset{po}, sentinel{s} {}
  90     LineClassification(LineClassification &&) = default;
  91     Kind kind;
  92     std::size_t payloadOffset; // byte offset of content
  93     const char *sentinel; // if it's a compiler directive
  94   };
  95
  96   void BeginSourceLine(const char *at) {
  97     at_ = at;
  98     column_ = 1;
  99     tabInCurrentLine_ = false;
 100   }
 101
 102   void BeginSourceLineAndAdvance() {
 103     BeginSourceLine(nextLine_);
 104     NextLine();
 105   }
 106
 107   void BeginStatementAndAdvance() {
 108     BeginSourceLineAndAdvance();
 109     slashInCurrentStatement_ = false;
 110     preventHollerith_ = false;
 111     delimiterNesting_ = 0;
 112     continuationLines_ = 0;
 113   }
 114
 115   Provenance GetProvenance(const char *sourceChar) const {
 116     return startProvenance_ + (sourceChar - start_);
 117   }
 118
 119   ProvenanceRange GetProvenanceRange(
 120       const char *first, const char *afterLast) const {
 121     std::size_t bytes = afterLast - first;
 122     return {startProvenance_ + (first - start_), bytes};
 123   }
 124
 125   void EmitChar(TokenSequence &tokens, char ch) {
 126     tokens.PutNextTokenChar(ch, GetCurrentProvenance());
 127   }
 128
 129   void EmitInsertedChar(TokenSequence &tokens, char ch) {
 130     Provenance provenance{allSources_.CompilerInsertionProvenance(ch)};
 131     tokens.PutNextTokenChar(ch, provenance);
 132   }
 133
 134   char EmitCharAndAdvance(TokenSequence &tokens, char ch) {
 135     EmitChar(tokens, ch);
 136     NextChar();
 137     return *at_;
 138   }
 139
 140   bool InCompilerDirective() const { return directiveSentinel_ != nullptr; }
 141   bool InFixedFormSource() const {
 142     return inFixedForm_ && !inPreprocessorDirective_ && !InCompilerDirective();
 143   }
 144
 145   bool IsCComment(const char *p) const {
 146     return p[0] == '/' && p[1] == '*' &&
 147         (inPreprocessorDirective_ ||
 148             (!inCharLiteral_ &&
 149                 features_.IsEnabled(
 150                     common::LanguageFeature::ClassicCComments)));
 151   }
 152
 153   void LabelField(TokenSequence &);
 154   void EnforceStupidEndStatementRules(const TokenSequence &);
 155   void SkipToEndOfLine();
 156   bool MustSkipToEndOfLine() const;
 157   void NextChar();
 158   void SkipToNextSignificantCharacter();
 159   void SkipCComments();
 160   void SkipSpaces();
 161   static const char *SkipWhiteSpace(const char *);
 162   const char *SkipWhiteSpaceAndCComments(const char *) const;
 163   const char *SkipCComment(const char *) const;
 164   bool NextToken(TokenSequence &);
 165   bool ExponentAndKind(TokenSequence &);
 166   void QuotedCharacterLiteral(TokenSequence &, const char *start);
 167   void Hollerith(TokenSequence &, int count, const char *start);
 168   bool PadOutCharacterLiteral(TokenSequence &);
 169   bool SkipCommentLine(bool afterAmpersand);
 170   bool IsFixedFormCommentLine(const char *) const;
 171   const char *IsFreeFormComment(const char *) const;
 172   std::optional<std::size_t> IsIncludeLine(const char *) const;
 173   void FortranInclude(const char *quote);
 174   const char *IsPreprocessorDirectiveLine(const char *) const;
 175   const char *FixedFormContinuationLine(bool mightNeedSpace);
 176   const char *FreeFormContinuationLine(bool ampersand);
 177   bool IsImplicitContinuation() const;
 178   bool FixedFormContinuation(bool mightNeedSpace);
 179   bool FreeFormContinuation();
 180   bool Continuation(bool mightNeedFixedFormSpace);
 181   std::optional<LineClassification> IsFixedFormCompilerDirectiveLine(
 182       const char *) const;
 183   std::optional<LineClassification> IsFreeFormCompilerDirectiveLine(
 184       const char *) const;
 185   const char *IsCompilerDirectiveSentinel(const char *) const;
 186   LineClassification ClassifyLine(const char *) const;
 187   void SourceFormChange(std::string &&);
 188
 189   Messages &messages_;
 190   CookedSource &cooked_;
 191   Preprocessor &preprocessor_;
 192   AllSources &allSources_;
 193   common::LanguageFeatureControl features_;
 194   bool inFixedForm_{false};
 195   int fixedFormColumnLimit_{72};
 196   Encoding encoding_{Encoding::UTF_8};
 197   int delimiterNesting_{0};
 198   int prescannerNesting_{0};
 199   int continuationLines_{0};
 200
 201   Provenance startProvenance_;
 202   const char *start_{nullptr}; // beginning of current source file content
 203   const char *limit_{nullptr}; // first address after end of current source
 204   const char *nextLine_{nullptr}; // next line to process; <= limit_
 205   const char *directiveSentinel_{nullptr}; // current compiler directive
 206
 207   // This data members are state for processing the source line containing
 208   // "at_", which goes to up to the newline character before "nextLine_".
 209   const char *at_{nullptr}; // next character to process; < nextLine_
 210   int column_{1}; // card image column position of next character
 211   bool tabInCurrentLine_{false};
 212   bool slashInCurrentStatement_{false};
 213   bool preventHollerith_{false}; // CHARACTER*4HIMOM not Hollerith
 214   bool inCharLiteral_{false};
 215   bool inPreprocessorDirective_{false};
 216
 217   // In some edge cases of compiler directive continuation lines, it
 218   // is necessary to treat the line break as a space character by
 219   // setting this flag, which is cleared by EmitChar().
 220   bool insertASpace_{false};
 221
 222   // When a free form continuation marker (&) appears at the end of a line
 223   // before a INCLUDE or #include, we delete it and omit the newline, so
 224   // that the first line of the included file is truly a continuation of
 225   // the line before.  Also used when the & appears at the end of the last
 226   // line in an include file.
 227   bool omitNewline_{false};
 228   bool skipLeadingAmpersand_{false};
 229
 230   const Provenance spaceProvenance_{
 231       allSources_.CompilerInsertionProvenance(' ')};
 232   const Provenance backslashProvenance_{
 233       allSources_.CompilerInsertionProvenance('\\')};
 234
 235   // To avoid probing the set of active compiler directive sentinel strings
 236   // on every comment line, they're checked first with a cheap Bloom filter.
 237   static const int prime1{1019}, prime2{1021};
 238   std::bitset<prime2> compilerDirectiveBloomFilter_; // 128 bytes
 239   std::unordered_set<std::string> compilerDirectiveSentinels_;
 240 };
 241 } // namespace Fortran::parser
 242 #endif // FORTRAN_PARSER_PRESCAN_H_