xref: /freebsd/contrib/llvm-project/clang/lib/Format/TokenAnalyzer.h (revision 349cc55c9796c4596a5b9904cd3281af295f878f)
10b57cec5SDimitry Andric //===--- TokenAnalyzer.h - Analyze Token Streams ----------------*- C++ -*-===//
20b57cec5SDimitry Andric //
30b57cec5SDimitry Andric // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
40b57cec5SDimitry Andric // See https://llvm.org/LICENSE.txt for license information.
50b57cec5SDimitry Andric // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
60b57cec5SDimitry Andric //
70b57cec5SDimitry Andric //===----------------------------------------------------------------------===//
80b57cec5SDimitry Andric ///
90b57cec5SDimitry Andric /// \file
100b57cec5SDimitry Andric /// This file declares an abstract TokenAnalyzer, and associated helper
110b57cec5SDimitry Andric /// classes. TokenAnalyzer can be extended to generate replacements based on
120b57cec5SDimitry Andric /// an annotated and pre-processed token stream.
130b57cec5SDimitry Andric ///
140b57cec5SDimitry Andric //===----------------------------------------------------------------------===//
150b57cec5SDimitry Andric 
160b57cec5SDimitry Andric #ifndef LLVM_CLANG_LIB_FORMAT_TOKENANALYZER_H
170b57cec5SDimitry Andric #define LLVM_CLANG_LIB_FORMAT_TOKENANALYZER_H
180b57cec5SDimitry Andric 
190b57cec5SDimitry Andric #include "AffectedRangeManager.h"
200b57cec5SDimitry Andric #include "Encoding.h"
210b57cec5SDimitry Andric #include "FormatToken.h"
220b57cec5SDimitry Andric #include "FormatTokenLexer.h"
230b57cec5SDimitry Andric #include "TokenAnnotator.h"
240b57cec5SDimitry Andric #include "UnwrappedLineParser.h"
250b57cec5SDimitry Andric #include "clang/Basic/Diagnostic.h"
260b57cec5SDimitry Andric #include "clang/Basic/DiagnosticOptions.h"
270b57cec5SDimitry Andric #include "clang/Basic/FileManager.h"
280b57cec5SDimitry Andric #include "clang/Basic/SourceManager.h"
290b57cec5SDimitry Andric #include "clang/Format/Format.h"
300b57cec5SDimitry Andric #include "llvm/ADT/STLExtras.h"
310b57cec5SDimitry Andric #include "llvm/Support/Debug.h"
32*349cc55cSDimitry Andric #include <memory>
330b57cec5SDimitry Andric 
340b57cec5SDimitry Andric namespace clang {
350b57cec5SDimitry Andric namespace format {
360b57cec5SDimitry Andric 
370b57cec5SDimitry Andric class Environment {
380b57cec5SDimitry Andric public:
390b57cec5SDimitry Andric   // This sets up an virtual file system with file \p FileName containing the
400b57cec5SDimitry Andric   // fragment \p Code. Assumes that \p Code starts at \p FirstStartColumn,
410b57cec5SDimitry Andric   // that the next lines of \p Code should start at \p NextStartColumn, and
420b57cec5SDimitry Andric   // that \p Code should end at \p LastStartColumn if it ends in newline.
430b57cec5SDimitry Andric   // See also the documentation of clang::format::internal::reformat.
44*349cc55cSDimitry Andric   Environment(StringRef Code, StringRef FileName, unsigned FirstStartColumn = 0,
450b57cec5SDimitry Andric               unsigned NextStartColumn = 0, unsigned LastStartColumn = 0);
460b57cec5SDimitry Andric 
470b57cec5SDimitry Andric   FileID getFileID() const { return ID; }
480b57cec5SDimitry Andric 
490b57cec5SDimitry Andric   const SourceManager &getSourceManager() const { return SM; }
500b57cec5SDimitry Andric 
510b57cec5SDimitry Andric   ArrayRef<CharSourceRange> getCharRanges() const { return CharRanges; }
520b57cec5SDimitry Andric 
530b57cec5SDimitry Andric   // Returns the column at which the fragment of code managed by this
540b57cec5SDimitry Andric   // environment starts.
550b57cec5SDimitry Andric   unsigned getFirstStartColumn() const { return FirstStartColumn; }
560b57cec5SDimitry Andric 
570b57cec5SDimitry Andric   // Returns the column at which subsequent lines of the fragment of code
580b57cec5SDimitry Andric   // managed by this environment should start.
590b57cec5SDimitry Andric   unsigned getNextStartColumn() const { return NextStartColumn; }
600b57cec5SDimitry Andric 
610b57cec5SDimitry Andric   // Returns the column at which the fragment of code managed by this
620b57cec5SDimitry Andric   // environment should end if it ends in a newline.
630b57cec5SDimitry Andric   unsigned getLastStartColumn() const { return LastStartColumn; }
640b57cec5SDimitry Andric 
65*349cc55cSDimitry Andric   // Returns nullptr and prints a diagnostic to stderr if the environment
66*349cc55cSDimitry Andric   // can't be created.
67*349cc55cSDimitry Andric   static std::unique_ptr<Environment> make(StringRef Code, StringRef FileName,
68*349cc55cSDimitry Andric                                            ArrayRef<tooling::Range> Ranges,
69*349cc55cSDimitry Andric                                            unsigned FirstStartColumn = 0,
70*349cc55cSDimitry Andric                                            unsigned NextStartColumn = 0,
71*349cc55cSDimitry Andric                                            unsigned LastStartColumn = 0);
72*349cc55cSDimitry Andric 
730b57cec5SDimitry Andric private:
740b57cec5SDimitry Andric   // This is only set if constructed from string.
750b57cec5SDimitry Andric   std::unique_ptr<SourceManagerForFile> VirtualSM;
760b57cec5SDimitry Andric 
770b57cec5SDimitry Andric   // This refers to either a SourceManager provided by users or VirtualSM
780b57cec5SDimitry Andric   // created for a single file.
790b57cec5SDimitry Andric   SourceManager &SM;
800b57cec5SDimitry Andric   FileID ID;
810b57cec5SDimitry Andric 
820b57cec5SDimitry Andric   SmallVector<CharSourceRange, 8> CharRanges;
830b57cec5SDimitry Andric   unsigned FirstStartColumn;
840b57cec5SDimitry Andric   unsigned NextStartColumn;
850b57cec5SDimitry Andric   unsigned LastStartColumn;
860b57cec5SDimitry Andric };
870b57cec5SDimitry Andric 
880b57cec5SDimitry Andric class TokenAnalyzer : public UnwrappedLineConsumer {
890b57cec5SDimitry Andric public:
900b57cec5SDimitry Andric   TokenAnalyzer(const Environment &Env, const FormatStyle &Style);
910b57cec5SDimitry Andric 
920b57cec5SDimitry Andric   std::pair<tooling::Replacements, unsigned> process();
930b57cec5SDimitry Andric 
940b57cec5SDimitry Andric protected:
950b57cec5SDimitry Andric   virtual std::pair<tooling::Replacements, unsigned>
960b57cec5SDimitry Andric   analyze(TokenAnnotator &Annotator,
970b57cec5SDimitry Andric           SmallVectorImpl<AnnotatedLine *> &AnnotatedLines,
980b57cec5SDimitry Andric           FormatTokenLexer &Tokens) = 0;
990b57cec5SDimitry Andric 
1000b57cec5SDimitry Andric   void consumeUnwrappedLine(const UnwrappedLine &TheLine) override;
1010b57cec5SDimitry Andric 
1020b57cec5SDimitry Andric   void finishRun() override;
1030b57cec5SDimitry Andric 
1040b57cec5SDimitry Andric   FormatStyle Style;
1050b57cec5SDimitry Andric   // Stores Style, FileID and SourceManager etc.
1060b57cec5SDimitry Andric   const Environment &Env;
1070b57cec5SDimitry Andric   // AffectedRangeMgr stores ranges to be fixed.
1080b57cec5SDimitry Andric   AffectedRangeManager AffectedRangeMgr;
1090b57cec5SDimitry Andric   SmallVector<SmallVector<UnwrappedLine, 16>, 2> UnwrappedLines;
1100b57cec5SDimitry Andric   encoding::Encoding Encoding;
1110b57cec5SDimitry Andric };
1120b57cec5SDimitry Andric 
1130b57cec5SDimitry Andric } // end namespace format
1140b57cec5SDimitry Andric } // end namespace clang
1150b57cec5SDimitry Andric 
1160b57cec5SDimitry Andric #endif
117