1 //===--- TokenAnalyzer.cpp - Analyze Token Streams --------------*- 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 /// \file 10 /// This file implements an abstract TokenAnalyzer and associated helper 11 /// classes. TokenAnalyzer can be extended to generate replacements based on 12 /// an annotated and pre-processed token stream. 13 /// 14 //===----------------------------------------------------------------------===// 15 16 #include "TokenAnalyzer.h" 17 #include "AffectedRangeManager.h" 18 #include "Encoding.h" 19 #include "FormatToken.h" 20 #include "FormatTokenLexer.h" 21 #include "TokenAnnotator.h" 22 #include "UnwrappedLineParser.h" 23 #include "clang/Basic/Diagnostic.h" 24 #include "clang/Basic/DiagnosticOptions.h" 25 #include "clang/Basic/FileManager.h" 26 #include "clang/Basic/SourceManager.h" 27 #include "clang/Format/Format.h" 28 #include "llvm/ADT/STLExtras.h" 29 #include "llvm/Support/Debug.h" 30 31 #define DEBUG_TYPE "format-formatter" 32 33 namespace clang { 34 namespace format { 35 36 Environment::Environment(StringRef Code, StringRef FileName, 37 ArrayRef<tooling::Range> Ranges, 38 unsigned FirstStartColumn, unsigned NextStartColumn, 39 unsigned LastStartColumn) 40 : VirtualSM(new SourceManagerForFile(FileName, Code)), SM(VirtualSM->get()), 41 ID(VirtualSM->get().getMainFileID()), FirstStartColumn(FirstStartColumn), 42 NextStartColumn(NextStartColumn), LastStartColumn(LastStartColumn) { 43 SourceLocation StartOfFile = SM.getLocForStartOfFile(ID); 44 for (const tooling::Range &Range : Ranges) { 45 SourceLocation Start = StartOfFile.getLocWithOffset(Range.getOffset()); 46 SourceLocation End = Start.getLocWithOffset(Range.getLength()); 47 CharRanges.push_back(CharSourceRange::getCharRange(Start, End)); 48 } 49 } 50 51 TokenAnalyzer::TokenAnalyzer(const Environment &Env, const FormatStyle &Style) 52 : Style(Style), Env(Env), 53 AffectedRangeMgr(Env.getSourceManager(), Env.getCharRanges()), 54 UnwrappedLines(1), 55 Encoding(encoding::detectEncoding( 56 Env.getSourceManager().getBufferData(Env.getFileID()))) { 57 LLVM_DEBUG( 58 llvm::dbgs() << "File encoding: " 59 << (Encoding == encoding::Encoding_UTF8 ? "UTF8" : "unknown") 60 << "\n"); 61 LLVM_DEBUG(llvm::dbgs() << "Language: " << getLanguageName(Style.Language) 62 << "\n"); 63 } 64 65 std::pair<tooling::Replacements, unsigned> TokenAnalyzer::process() { 66 tooling::Replacements Result; 67 FormatTokenLexer Tokens(Env.getSourceManager(), Env.getFileID(), 68 Env.getFirstStartColumn(), Style, Encoding); 69 70 UnwrappedLineParser Parser(Style, Tokens.getKeywords(), 71 Env.getFirstStartColumn(), Tokens.lex(), *this); 72 Parser.parse(); 73 assert(UnwrappedLines.rbegin()->empty()); 74 unsigned Penalty = 0; 75 for (unsigned Run = 0, RunE = UnwrappedLines.size(); Run + 1 != RunE; ++Run) { 76 LLVM_DEBUG(llvm::dbgs() << "Run " << Run << "...\n"); 77 SmallVector<AnnotatedLine *, 16> AnnotatedLines; 78 79 TokenAnnotator Annotator(Style, Tokens.getKeywords()); 80 for (unsigned i = 0, e = UnwrappedLines[Run].size(); i != e; ++i) { 81 AnnotatedLines.push_back(new AnnotatedLine(UnwrappedLines[Run][i])); 82 Annotator.annotate(*AnnotatedLines.back()); 83 } 84 85 std::pair<tooling::Replacements, unsigned> RunResult = 86 analyze(Annotator, AnnotatedLines, Tokens); 87 88 LLVM_DEBUG({ 89 llvm::dbgs() << "Replacements for run " << Run << ":\n"; 90 for (tooling::Replacements::const_iterator I = RunResult.first.begin(), 91 E = RunResult.first.end(); 92 I != E; ++I) { 93 llvm::dbgs() << I->toString() << "\n"; 94 } 95 }); 96 for (unsigned i = 0, e = AnnotatedLines.size(); i != e; ++i) { 97 delete AnnotatedLines[i]; 98 } 99 100 Penalty += RunResult.second; 101 for (const auto &R : RunResult.first) { 102 auto Err = Result.add(R); 103 // FIXME: better error handling here. For now, simply return an empty 104 // Replacements to indicate failure. 105 if (Err) { 106 llvm::errs() << llvm::toString(std::move(Err)) << "\n"; 107 return {tooling::Replacements(), 0}; 108 } 109 } 110 } 111 return {Result, Penalty}; 112 } 113 114 void TokenAnalyzer::consumeUnwrappedLine(const UnwrappedLine &TheLine) { 115 assert(!UnwrappedLines.empty()); 116 UnwrappedLines.back().push_back(TheLine); 117 } 118 119 void TokenAnalyzer::finishRun() { 120 UnwrappedLines.push_back(SmallVector<UnwrappedLine, 16>()); 121 } 122 123 } // end namespace format 124 } // end namespace clang 125