clang  9.0.0svn
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  /// Replaces the whitespace in front of \p Tok. Only call once for
44  /// each \c AnnotatedToken.
45  ///
46  /// \p StartOfTokenColumn is the column at which the token will start after
47  /// this replacement. It is needed for determining how \p Spaces is turned
48  /// into tabs and spaces for some format styles.
49  void replaceWhitespace(FormatToken &Tok, unsigned Newlines, unsigned Spaces,
50  unsigned StartOfTokenColumn,
51  bool InPPDirective = false);
52 
53  /// Adds information about an unchangeable token's whitespace.
54  ///
55  /// Needs to be called for every token for which \c replaceWhitespace
56  /// was not called.
57  void addUntouchableToken(const FormatToken &Tok, bool InPPDirective);
58 
60 
61  /// Inserts or replaces whitespace in the middle of a token.
62  ///
63  /// Inserts \p PreviousPostfix, \p Newlines, \p Spaces and \p CurrentPrefix
64  /// (in this order) at \p Offset inside \p Tok, replacing \p ReplaceChars
65  /// characters.
66  ///
67  /// Note: \p Spaces can be negative to retain information about initial
68  /// relative column offset between a line of a block comment and the start of
69  /// the comment. This negative offset may be compensated by trailing comment
70  /// alignment here. In all other cases negative \p Spaces will be truncated to
71  /// 0.
72  ///
73  /// When \p InPPDirective is true, escaped newlines are inserted. \p Spaces is
74  /// used to align backslashes correctly.
75  void replaceWhitespaceInToken(const FormatToken &Tok, unsigned Offset,
76  unsigned ReplaceChars,
77  StringRef PreviousPostfix,
78  StringRef CurrentPrefix, bool InPPDirective,
79  unsigned Newlines, int Spaces);
80 
81  /// Returns all the \c Replacements created during formatting.
83 
84  /// Represents a change before a token, a break inside a token,
85  /// or the layout of an unchanged token (or whitespace within).
86  struct Change {
87  /// Functor to sort changes in original source order.
89  public:
90  IsBeforeInFile(const SourceManager &SourceMgr) : SourceMgr(SourceMgr) {}
91  bool operator()(const Change &C1, const Change &C2) const;
92 
93  private:
94  const SourceManager &SourceMgr;
95  };
96 
97  /// Creates a \c Change.
98  ///
99  /// The generated \c Change will replace the characters at
100  /// \p OriginalWhitespaceRange with a concatenation of
101  /// \p PreviousLinePostfix, \p NewlinesBefore line breaks, \p Spaces spaces
102  /// and \p CurrentLinePrefix.
103  ///
104  /// \p StartOfTokenColumn and \p InPPDirective will be used to lay out
105  /// trailing comments and escaped newlines.
106  Change(const FormatToken &Tok, bool CreateReplacement,
108  unsigned StartOfTokenColumn, unsigned NewlinesBefore,
109  StringRef PreviousLinePostfix, StringRef CurrentLinePrefix,
111 
112  // The kind of the token whose whitespace this change replaces, or in which
113  // this change inserts whitespace.
114  // FIXME: Currently this is not set correctly for breaks inside comments, as
115  // the \c BreakableToken is still doing its own alignment.
116  const FormatToken *Tok;
117 
119  // Changes might be in the middle of a token, so we cannot just keep the
120  // FormatToken around to query its information.
123  unsigned NewlinesBefore;
124  std::string PreviousLinePostfix;
125  std::string CurrentLinePrefix;
127 
128  // The number of spaces in front of the token or broken part of the token.
129  // This will be adapted when aligning tokens.
130  // Can be negative to retain information about the initial relative offset
131  // of the lines in a block comment. This is used when aligning trailing
132  // comments. Uncompensated negative offset is truncated to 0.
133  int Spaces;
134 
135  // If this change is inside of a token but not at the start of the token or
136  // directly after a newline.
138 
139  // \c IsTrailingComment, \c TokenLength, \c PreviousEndOfTokenColumn and
140  // \c EscapedNewlineColumn will be calculated in
141  // \c calculateLineBreakInformation.
143  unsigned TokenLength;
146 
147  // These fields are used to retain correct relative line indentation in a
148  // block comment when aligning trailing comments.
149  //
150  // If this Change represents a continuation of a block comment,
151  // \c StartOfBlockComment is pointer to the first Change in the block
152  // comment. \c IndentationOffset is a relative column offset to this
153  // change, so that the correct column can be reconstructed at the end of
154  // the alignment process.
157 
158  // A combination of indent level and nesting level, which are used in
159  // tandem to compute lexical scope, for the purposes of deciding
160  // when to stop consecutive alignment runs.
161  std::pair<unsigned, unsigned> indentAndNestingLevel() const {
162  return std::make_pair(Tok->IndentLevel, Tok->NestingLevel);
163  }
164  };
165 
166 private:
167  /// Calculate \c IsTrailingComment, \c TokenLength for the last tokens
168  /// or token parts in a line and \c PreviousEndOfTokenColumn and
169  /// \c EscapedNewlineColumn for the first tokens or token parts in a line.
170  void calculateLineBreakInformation();
171 
172  /// Align consecutive assignments over all \c Changes.
173  void alignConsecutiveAssignments();
174 
175  /// Align consecutive declarations over all \c Changes.
176  void alignConsecutiveDeclarations();
177 
178  /// Align trailing comments over all \c Changes.
179  void alignTrailingComments();
180 
181  /// Align trailing comments from change \p Start to change \p End at
182  /// the specified \p Column.
183  void alignTrailingComments(unsigned Start, unsigned End, unsigned Column);
184 
185  /// Align escaped newlines over all \c Changes.
186  void alignEscapedNewlines();
187 
188  /// Align escaped newlines from change \p Start to change \p End at
189  /// the specified \p Column.
190  void alignEscapedNewlines(unsigned Start, unsigned End, unsigned Column);
191 
192  /// Fill \c Replaces with the replacements for all effective changes.
193  void generateChanges();
194 
195  /// Stores \p Text as the replacement for the whitespace in \p Range.
196  void storeReplacement(SourceRange Range, StringRef Text);
197  void appendNewlineText(std::string &Text, unsigned Newlines);
198  void appendEscapedNewlineText(std::string &Text, unsigned Newlines,
199  unsigned PreviousEndOfTokenColumn,
200  unsigned EscapedNewlineColumn);
201  void appendIndentText(std::string &Text, unsigned IndentLevel,
202  unsigned Spaces, unsigned WhitespaceStartColumn);
203 
204  SmallVector<Change, 16> Changes;
205  const SourceManager &SourceMgr;
206  tooling::Replacements Replaces;
207  const FormatStyle &Style;
208  bool UseCRLF;
209 };
210 
211 } // namespace format
212 } // namespace clang
213 
214 #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:230
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:1676
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:126
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:233
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:1675
A trivial tuple used to represent a source range.
This class handles loading and caching of source files into memory.
const FormatStyle & Style