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   llvm::SpecificBumpPtrAllocator<FormatToken> Allocator;
68   IdentifierTable IdentTable(getFormattingLangOpts(Style));
69   FormatTokenLexer Lex(Env.getSourceManager(), Env.getFileID(),
70                        Env.getFirstStartColumn(), Style, Encoding, Allocator,
71 
72                        IdentTable);
73   ArrayRef<FormatToken *> Toks(Lex.lex());
74   SmallVector<FormatToken *, 10> Tokens(Toks.begin(), Toks.end());
75   UnwrappedLineParser Parser(Style, Lex.getKeywords(),
76                              Env.getFirstStartColumn(), Tokens, *this);
77   Parser.parse();
78   assert(UnwrappedLines.rbegin()->empty());
79   unsigned Penalty = 0;
80   for (unsigned Run = 0, RunE = UnwrappedLines.size(); Run + 1 != RunE; ++Run) {
81     LLVM_DEBUG(llvm::dbgs() << "Run " << Run << "...\n");
82     SmallVector<AnnotatedLine *, 16> AnnotatedLines;
83 
84     TokenAnnotator Annotator(Style, Lex.getKeywords());
85     for (unsigned i = 0, e = UnwrappedLines[Run].size(); i != e; ++i) {
86       AnnotatedLines.push_back(new AnnotatedLine(UnwrappedLines[Run][i]));
87       Annotator.annotate(*AnnotatedLines.back());
88     }
89 
90     std::pair<tooling::Replacements, unsigned> RunResult =
91         analyze(Annotator, AnnotatedLines, Lex);
92 
93     LLVM_DEBUG({
94       llvm::dbgs() << "Replacements for run " << Run << ":\n";
95       for (tooling::Replacements::const_iterator I = RunResult.first.begin(),
96                                                  E = RunResult.first.end();
97            I != E; ++I) {
98         llvm::dbgs() << I->toString() << "\n";
99       }
100     });
101     for (unsigned i = 0, e = AnnotatedLines.size(); i != e; ++i) {
102       delete AnnotatedLines[i];
103     }
104 
105     Penalty += RunResult.second;
106     for (const auto &R : RunResult.first) {
107       auto Err = Result.add(R);
108       // FIXME: better error handling here. For now, simply return an empty
109       // Replacements to indicate failure.
110       if (Err) {
111         llvm::errs() << llvm::toString(std::move(Err)) << "\n";
112         return {tooling::Replacements(), 0};
113       }
114     }
115   }
116   return {Result, Penalty};
117 }
118 
119 void TokenAnalyzer::consumeUnwrappedLine(const UnwrappedLine &TheLine) {
120   assert(!UnwrappedLines.empty());
121   UnwrappedLines.back().push_back(TheLine);
122 }
123 
124 void TokenAnalyzer::finishRun() {
125   UnwrappedLines.push_back(SmallVector<UnwrappedLine, 16>());
126 }
127 
128 } // end namespace format
129 } // end namespace clang
130