TokenAnalyzer.cpp revision 303233 
1//===--- TokenAnalyzer.cpp - Analyze Token Streams --------------*- C++ -*-===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9///
10/// \file
11/// \brief This file implements an abstract TokenAnalyzer and associated helper
12/// classes. TokenAnalyzer can be extended to generate replacements based on
13/// an annotated and pre-processed token stream.
14///
15//===----------------------------------------------------------------------===//
16
17#include "TokenAnalyzer.h"
18#include "AffectedRangeManager.h"
19#include "Encoding.h"
20#include "FormatToken.h"
21#include "FormatTokenLexer.h"
22#include "TokenAnnotator.h"
23#include "UnwrappedLineParser.h"
24#include "clang/Basic/Diagnostic.h"
25#include "clang/Basic/DiagnosticOptions.h"
26#include "clang/Basic/FileManager.h"
27#include "clang/Basic/SourceManager.h"
28#include "clang/Format/Format.h"
29#include "llvm/ADT/STLExtras.h"
30#include "llvm/Support/Debug.h"
31
32#define DEBUG_TYPE "format-formatter"
33
34namespace clang {
35namespace format {
36
37// This sets up an virtual file system with file \p FileName containing \p
38// Code.
39std::unique_ptr<Environment>
40Environment::CreateVirtualEnvironment(StringRef Code, StringRef FileName,
41                                      ArrayRef<tooling::Range> Ranges) {
42  // This is referenced by `FileMgr` and will be released by `FileMgr` when it
43  // is deleted.
44  IntrusiveRefCntPtr<vfs::InMemoryFileSystem> InMemoryFileSystem(
45      new vfs::InMemoryFileSystem);
46  // This is passed to `SM` as reference, so the pointer has to be referenced
47  // in `Environment` so that `FileMgr` can out-live this function scope.
48  std::unique_ptr<FileManager> FileMgr(
49      new FileManager(FileSystemOptions(), InMemoryFileSystem));
50  // This is passed to `SM` as reference, so the pointer has to be referenced
51  // by `Environment` due to the same reason above.
52  std::unique_ptr<DiagnosticsEngine> Diagnostics(new DiagnosticsEngine(
53      IntrusiveRefCntPtr<DiagnosticIDs>(new DiagnosticIDs),
54      new DiagnosticOptions));
55  // This will be stored as reference, so the pointer has to be stored in
56  // due to the same reason above.
57  std::unique_ptr<SourceManager> VirtualSM(
58      new SourceManager(*Diagnostics, *FileMgr));
59  InMemoryFileSystem->addFile(
60      FileName, 0, llvm::MemoryBuffer::getMemBuffer(
61                       Code, FileName, /*RequiresNullTerminator=*/false));
62  FileID ID = VirtualSM->createFileID(FileMgr->getFile(FileName),
63                                      SourceLocation(), clang::SrcMgr::C_User);
64  assert(ID.isValid());
65  SourceLocation StartOfFile = VirtualSM->getLocForStartOfFile(ID);
66  std::vector<CharSourceRange> CharRanges;
67  for (const tooling::Range &Range : Ranges) {
68    SourceLocation Start = StartOfFile.getLocWithOffset(Range.getOffset());
69    SourceLocation End = Start.getLocWithOffset(Range.getLength());
70    CharRanges.push_back(CharSourceRange::getCharRange(Start, End));
71  }
72  return llvm::make_unique<Environment>(ID, std::move(FileMgr),
73                                        std::move(VirtualSM),
74                                        std::move(Diagnostics), CharRanges);
75}
76
77TokenAnalyzer::TokenAnalyzer(const Environment &Env, const FormatStyle &Style)
78    : Style(Style), Env(Env),
79      AffectedRangeMgr(Env.getSourceManager(), Env.getCharRanges()),
80      UnwrappedLines(1),
81      Encoding(encoding::detectEncoding(
82          Env.getSourceManager().getBufferData(Env.getFileID()))) {
83  DEBUG(
84      llvm::dbgs() << "File encoding: "
85                   << (Encoding == encoding::Encoding_UTF8 ? "UTF8" : "unknown")
86                   << "\n");
87  DEBUG(llvm::dbgs() << "Language: " << getLanguageName(Style.Language)
88                     << "\n");
89}
90
91tooling::Replacements TokenAnalyzer::process() {
92  tooling::Replacements Result;
93  FormatTokenLexer Tokens(Env.getSourceManager(), Env.getFileID(), Style,
94                          Encoding);
95
96  UnwrappedLineParser Parser(Style, Tokens.getKeywords(), Tokens.lex(), *this);
97  Parser.parse();
98  assert(UnwrappedLines.rbegin()->empty());
99  for (unsigned Run = 0, RunE = UnwrappedLines.size(); Run + 1 != RunE; ++Run) {
100    DEBUG(llvm::dbgs() << "Run " << Run << "...\n");
101    SmallVector<AnnotatedLine *, 16> AnnotatedLines;
102
103    TokenAnnotator Annotator(Style, Tokens.getKeywords());
104    for (unsigned i = 0, e = UnwrappedLines[Run].size(); i != e; ++i) {
105      AnnotatedLines.push_back(new AnnotatedLine(UnwrappedLines[Run][i]));
106      Annotator.annotate(*AnnotatedLines.back());
107    }
108
109    tooling::Replacements RunResult =
110        analyze(Annotator, AnnotatedLines, Tokens, Result);
111
112    DEBUG({
113      llvm::dbgs() << "Replacements for run " << Run << ":\n";
114      for (tooling::Replacements::iterator I = RunResult.begin(),
115                                           E = RunResult.end();
116           I != E; ++I) {
117        llvm::dbgs() << I->toString() << "\n";
118      }
119    });
120    for (unsigned i = 0, e = AnnotatedLines.size(); i != e; ++i) {
121      delete AnnotatedLines[i];
122    }
123    Result.insert(RunResult.begin(), RunResult.end());
124  }
125  return Result;
126}
127
128void TokenAnalyzer::consumeUnwrappedLine(const UnwrappedLine &TheLine) {
129  assert(!UnwrappedLines.empty());
130  UnwrappedLines.back().push_back(TheLine);
131}
132
133void TokenAnalyzer::finishRun() {
134  UnwrappedLines.push_back(SmallVector<UnwrappedLine, 16>());
135}
136
137} // end namespace format
138} // end namespace clang
139