clang  10.0.0svn
WhitespaceManager.h
Go to the documentation of this file.
1 //===--- WhitespaceManager.h - Format C++ code ------------------*- 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 /// WhitespaceManager class manages whitespace around tokens and their
11 /// replacements.
12 ///
13 //===----------------------------------------------------------------------===//
14 
15 #ifndef LLVM_CLANG_LIB_FORMAT_WHITESPACEMANAGER_H
16 #define LLVM_CLANG_LIB_FORMAT_WHITESPACEMANAGER_H
17 
18 #include "TokenAnnotator.h"
20 #include "clang/Format/Format.h"
21 #include <string>
22 
23 namespace clang {
24 namespace format {
25 
26 /// Manages the whitespaces around tokens and their replacements.
27 ///
28 /// This includes special handling for certain constructs, e.g. the alignment of
29 /// trailing line comments.
30 ///
31 /// To guarantee correctness of alignment operations, the \c WhitespaceManager
32 /// must be informed about every token in the source file; for each token, there
33 /// must be exactly one call to either \c replaceWhitespace or
34 /// \c addUntouchableToken.
35 ///
36 /// There may be multiple calls to \c breakToken for a given token.
38 public:
39  WhitespaceManager(const SourceManager &SourceMgr, const FormatStyle &Style,
40  bool UseCRLF)
41  : SourceMgr(SourceMgr), Style(Style), UseCRLF(UseCRLF) {}
42 
43  bool useCRLF() const { return UseCRLF; }
44 
45  /// Replaces the whitespace in front of \p Tok. Only call once for
46  /// each \c AnnotatedToken.
47  ///
48  /// \p StartOfTokenColumn is the column at which the token will start after
49  /// this replacement. It is needed for determining how \p Spaces is turned
50  /// into tabs and spaces for some format styles.
51  void replaceWhitespace(FormatToken &Tok, unsigned Newlines, unsigned Spaces,
52  unsigned StartOfTokenColumn,
53  bool InPPDirective = false);
54 
55  /// Adds information about an unchangeable token's whitespace.
56  ///
57  /// Needs to be called for every token for which \c replaceWhitespace
58  /// was not called.
59  void addUntouchableToken(const FormatToken &Tok, bool InPPDirective);
60 
62 
63  /// Inserts or replaces whitespace in the middle of a token.
64  ///
65  /// Inserts \p PreviousPostfix, \p Newlines, \p Spaces and \p CurrentPrefix
66  /// (in this order) at \p Offset inside \p Tok, replacing \p ReplaceChars
67  /// characters.
68  ///
69  /// Note: \p Spaces can be negative to retain information about initial
70  /// relative column offset between a line of a block comment and the start of
71  /// the comment. This negative offset may be compensated by trailing comment
72  /// alignment here. In all other cases negative \p Spaces will be truncated to
73  /// 0.
74  ///
75  /// When \p InPPDirective is true, escaped newlines are inserted. \p Spaces is
76  /// used to align backslashes correctly.
77  void replaceWhitespaceInToken(const FormatToken &Tok, unsigned Offset,
78  unsigned ReplaceChars,
79  StringRef PreviousPostfix,
80  StringRef CurrentPrefix, bool InPPDirective,
81  unsigned Newlines, int Spaces);
82 
83  /// Returns all the \c Replacements created during formatting.
85 
86  /// Represents a change before a token, a break inside a token,
87  /// or the layout of an unchanged token (or whitespace within).
88  struct Change {
89  /// Functor to sort changes in original source order.
91  public:
92  IsBeforeInFile(const SourceManager &SourceMgr) : SourceMgr(SourceMgr) {}
93  bool operator()(const Change &C1, const Change &C2) const;
94 
95  private:
96  const SourceManager &SourceMgr;
97  };
98 
99  /// Creates a \c Change.
100  ///
101  /// The generated \c Change will replace the characters at
102  /// \p OriginalWhitespaceRange with a concatenation of
103  /// \p PreviousLinePostfix, \p NewlinesBefore line breaks, \p Spaces spaces
104  /// and \p CurrentLinePrefix.
105  ///
106  /// \p StartOfTokenColumn and \p InPPDirective will be used to lay out
107  /// trailing comments and escaped newlines.
108  Change(const FormatToken &Tok, bool CreateReplacement,
110  unsigned StartOfTokenColumn, unsigned NewlinesBefore,
111  StringRef PreviousLinePostfix, StringRef CurrentLinePrefix,
113 
114  // The kind of the token whose whitespace this change replaces, or in which
115  // this change inserts whitespace.
116  // FIXME: Currently this is not set correctly for breaks inside comments, as
117  // the \c BreakableToken is still doing its own alignment.
118  const FormatToken *Tok;
119 
121  // Changes might be in the middle of a token, so we cannot just keep the
122  // FormatToken around to query its information.
125  unsigned NewlinesBefore;
126  std::string PreviousLinePostfix;
127  std::string CurrentLinePrefix;
129 
130  // The number of spaces in front of the token or broken part of the token.
131  // This will be adapted when aligning tokens.
132  // Can be negative to retain information about the initial relative offset
133  // of the lines in a block comment. This is used when aligning trailing
134  // comments. Uncompensated negative offset is truncated to 0.
135  int Spaces;
136 
137  // If this change is inside of a token but not at the start of the token or
138  // directly after a newline.
140 
141  // \c IsTrailingComment, \c TokenLength, \c PreviousEndOfTokenColumn and
142  // \c EscapedNewlineColumn will be calculated in
143  // \c calculateLineBreakInformation.
145  unsigned TokenLength;
148 
149  // These fields are used to retain correct relative line indentation in a
150  // block comment when aligning trailing comments.
151  //
152  // If this Change represents a continuation of a block comment,
153  // \c StartOfBlockComment is pointer to the first Change in the block
154  // comment. \c IndentationOffset is a relative column offset to this
155  // change, so that the correct column can be reconstructed at the end of
156  // the alignment process.
159 
160  // A combination of indent level and nesting level, which are used in
161  // tandem to compute lexical scope, for the purposes of deciding
162  // when to stop consecutive alignment runs.
163  std::pair<unsigned, unsigned> indentAndNestingLevel() const {
164  return std::make_pair(Tok->IndentLevel, Tok->NestingLevel);
165  }
166  };
167 
168 private:
169  /// Calculate \c IsTrailingComment, \c TokenLength for the last tokens
170  /// or token parts in a line and \c PreviousEndOfTokenColumn and
171  /// \c EscapedNewlineColumn for the first tokens or token parts in a line.
172  void calculateLineBreakInformation();
173 
174  /// \brief Align consecutive C/C++ preprocessor macros over all \c Changes.
175  void alignConsecutiveMacros();
176 
177  /// Align consecutive assignments over all \c Changes.
178  void alignConsecutiveAssignments();
179 
180  /// Align consecutive declarations over all \c Changes.
181  void alignConsecutiveDeclarations();
182 
183  /// Align trailing comments over all \c Changes.
184  void alignTrailingComments();
185 
186  /// Align trailing comments from change \p Start to change \p End at
187  /// the specified \p Column.
188  void alignTrailingComments(unsigned Start, unsigned End, unsigned Column);
189 
190  /// Align escaped newlines over all \c Changes.
191  void alignEscapedNewlines();
192 
193  /// Align escaped newlines from change \p Start to change \p End at
194  /// the specified \p Column.
195  void alignEscapedNewlines(unsigned Start, unsigned End, unsigned Column);
196 
197  /// Fill \c Replaces with the replacements for all effective changes.
198  void generateChanges();
199 
200  /// Stores \p Text as the replacement for the whitespace in \p Range.
201  void storeReplacement(SourceRange Range, StringRef Text);
202  void appendNewlineText(std::string &Text, unsigned Newlines);
203  void appendEscapedNewlineText(std::string &Text, unsigned Newlines,
204  unsigned PreviousEndOfTokenColumn,
205  unsigned EscapedNewlineColumn);
206  void appendIndentText(std::string &Text, unsigned IndentLevel,
207  unsigned Spaces, unsigned WhitespaceStartColumn);
208 
209  SmallVector<Change, 16> Changes;
210  const SourceManager &SourceMgr;
211  tooling::Replacements Replaces;
212  const FormatStyle &Style;
213  bool UseCRLF;
214 };
215 
216 } // namespace format
217 } // namespace clang
218 
219 #endif
Change(const FormatToken &Tok, bool CreateReplacement, SourceRange OriginalWhitespaceRange, int Spaces, unsigned StartOfTokenColumn, unsigned NewlinesBefore, StringRef PreviousLinePostfix, StringRef CurrentLinePrefix, bool ContinuesPPDirective, bool IsInsideToken)
Creates a Change.
unsigned NestingLevel
The nesting level of this token, i.e.
Definition: FormatToken.h:233
Defines the SourceManager interface.
std::pair< unsigned, unsigned > indentAndNestingLevel() const
Maintains a set of replacements that are conflict-free.
Definition: Replacement.h:209
void replaceWhitespaceInToken(const FormatToken &Tok, unsigned Offset, unsigned ReplaceChars, StringRef PreviousPostfix, StringRef CurrentPrefix, bool InPPDirective, unsigned Newlines, int Spaces)
Inserts or replaces whitespace in the middle of a token.
This file implements a token annotator, i.e.
Manages the whitespaces around tokens and their replacements.
const FormatToken & Tok
WhitespaceManager(const SourceManager &SourceMgr, const FormatStyle &Style, bool UseCRLF)
A text replacement.
Definition: Replacement.h:83
unsigned Offset
Definition: Format.cpp:1728
SourceLocation End
llvm::Error addReplacement(const tooling::Replacement &Replacement)
A wrapper around a Token storing information about the whitespace characters preceding it...
Definition: FormatToken.h:129
const tooling::Replacements & generateReplacements()
Returns all the Replacements created during formatting.
Represents a change before a token, a break inside a token, or the layout of an unchanged token (or w...
void replaceWhitespace(FormatToken &Tok, unsigned Newlines, unsigned Spaces, unsigned StartOfTokenColumn, bool InPPDirective=false)
Replaces the whitespace in front of Tok.
const bool InPPDirective
Various functions to configurably format source code.
The FormatStyle is used to configure the formatting to follow specific guidelines.
Definition: Format.h:49
unsigned IndentLevel
The indent level of this token. Copied from the surrounding line.
Definition: FormatToken.h:236
Dataflow Directional Tag Classes.
Functor to sort changes in original source order.
void addUntouchableToken(const FormatToken &Tok, bool InPPDirective)
Adds information about an unchangeable token&#39;s whitespace.
bool operator()(const Change &C1, const Change &C2) const
StringRef Text
Definition: Format.cpp:1727
A trivial tuple used to represent a source range.
This class handles loading and caching of source files into memory.
const FormatStyle & Style