1 //===--- WhitespaceManager.cpp - Format C++ code --------------------------===//
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 WhitespaceManager class.
12 ///
13 //===----------------------------------------------------------------------===//
14
15 #include "WhitespaceManager.h"
16 #include "llvm/ADT/STLExtras.h"
17
18 namespace clang {
19 namespace format {
20
21 bool WhitespaceManager::Change::IsBeforeInFile::
operator ()(const Change & C1,const Change & C2) const22 operator()(const Change &C1, const Change &C2) const {
23 return SourceMgr.isBeforeInTranslationUnit(
24 C1.OriginalWhitespaceRange.getBegin(),
25 C2.OriginalWhitespaceRange.getBegin());
26 }
27
Change(bool CreateReplacement,const SourceRange & OriginalWhitespaceRange,unsigned IndentLevel,int Spaces,unsigned StartOfTokenColumn,unsigned NewlinesBefore,StringRef PreviousLinePostfix,StringRef CurrentLinePrefix,tok::TokenKind Kind,bool ContinuesPPDirective)28 WhitespaceManager::Change::Change(
29 bool CreateReplacement, const SourceRange &OriginalWhitespaceRange,
30 unsigned IndentLevel, int Spaces, unsigned StartOfTokenColumn,
31 unsigned NewlinesBefore, StringRef PreviousLinePostfix,
32 StringRef CurrentLinePrefix, tok::TokenKind Kind, bool ContinuesPPDirective)
33 : CreateReplacement(CreateReplacement),
34 OriginalWhitespaceRange(OriginalWhitespaceRange),
35 StartOfTokenColumn(StartOfTokenColumn), NewlinesBefore(NewlinesBefore),
36 PreviousLinePostfix(PreviousLinePostfix),
37 CurrentLinePrefix(CurrentLinePrefix), Kind(Kind),
38 ContinuesPPDirective(ContinuesPPDirective), IndentLevel(IndentLevel),
39 Spaces(Spaces) {}
40
reset()41 void WhitespaceManager::reset() {
42 Changes.clear();
43 Replaces.clear();
44 }
45
replaceWhitespace(FormatToken & Tok,unsigned Newlines,unsigned IndentLevel,unsigned Spaces,unsigned StartOfTokenColumn,bool InPPDirective)46 void WhitespaceManager::replaceWhitespace(FormatToken &Tok, unsigned Newlines,
47 unsigned IndentLevel, unsigned Spaces,
48 unsigned StartOfTokenColumn,
49 bool InPPDirective) {
50 if (Tok.Finalized)
51 return;
52 Tok.Decision = (Newlines > 0) ? FD_Break : FD_Continue;
53 Changes.push_back(Change(true, Tok.WhitespaceRange, IndentLevel, Spaces,
54 StartOfTokenColumn, Newlines, "", "",
55 Tok.Tok.getKind(), InPPDirective && !Tok.IsFirst));
56 }
57
addUntouchableToken(const FormatToken & Tok,bool InPPDirective)58 void WhitespaceManager::addUntouchableToken(const FormatToken &Tok,
59 bool InPPDirective) {
60 if (Tok.Finalized)
61 return;
62 Changes.push_back(Change(false, Tok.WhitespaceRange, /*IndentLevel=*/0,
63 /*Spaces=*/0, Tok.OriginalColumn, Tok.NewlinesBefore,
64 "", "", Tok.Tok.getKind(),
65 InPPDirective && !Tok.IsFirst));
66 }
67
replaceWhitespaceInToken(const FormatToken & Tok,unsigned Offset,unsigned ReplaceChars,StringRef PreviousPostfix,StringRef CurrentPrefix,bool InPPDirective,unsigned Newlines,unsigned IndentLevel,int Spaces)68 void WhitespaceManager::replaceWhitespaceInToken(
69 const FormatToken &Tok, unsigned Offset, unsigned ReplaceChars,
70 StringRef PreviousPostfix, StringRef CurrentPrefix, bool InPPDirective,
71 unsigned Newlines, unsigned IndentLevel, int Spaces) {
72 if (Tok.Finalized)
73 return;
74 SourceLocation Start = Tok.getStartOfNonWhitespace().getLocWithOffset(Offset);
75 Changes.push_back(Change(
76 true, SourceRange(Start, Start.getLocWithOffset(ReplaceChars)),
77 IndentLevel, Spaces, std::max(0, Spaces), Newlines, PreviousPostfix,
78 CurrentPrefix,
79 // If we don't add a newline this change doesn't start a comment. Thus,
80 // when we align line comments, we don't need to treat this change as one.
81 // FIXME: We still need to take this change in account to properly
82 // calculate the new length of the comment and to calculate the changes
83 // for which to do the alignment when aligning comments.
84 Tok.Type == TT_LineComment && Newlines > 0 ? tok::comment : tok::unknown,
85 InPPDirective && !Tok.IsFirst));
86 }
87
generateReplacements()88 const tooling::Replacements &WhitespaceManager::generateReplacements() {
89 if (Changes.empty())
90 return Replaces;
91
92 std::sort(Changes.begin(), Changes.end(), Change::IsBeforeInFile(SourceMgr));
93 calculateLineBreakInformation();
94 alignTrailingComments();
95 alignEscapedNewlines();
96 generateChanges();
97
98 return Replaces;
99 }
100
calculateLineBreakInformation()101 void WhitespaceManager::calculateLineBreakInformation() {
102 Changes[0].PreviousEndOfTokenColumn = 0;
103 for (unsigned i = 1, e = Changes.size(); i != e; ++i) {
104 unsigned OriginalWhitespaceStart =
105 SourceMgr.getFileOffset(Changes[i].OriginalWhitespaceRange.getBegin());
106 unsigned PreviousOriginalWhitespaceEnd = SourceMgr.getFileOffset(
107 Changes[i - 1].OriginalWhitespaceRange.getEnd());
108 Changes[i - 1].TokenLength = OriginalWhitespaceStart -
109 PreviousOriginalWhitespaceEnd +
110 Changes[i].PreviousLinePostfix.size() +
111 Changes[i - 1].CurrentLinePrefix.size();
112
113 Changes[i].PreviousEndOfTokenColumn =
114 Changes[i - 1].StartOfTokenColumn + Changes[i - 1].TokenLength;
115
116 Changes[i - 1].IsTrailingComment =
117 (Changes[i].NewlinesBefore > 0 || Changes[i].Kind == tok::eof) &&
118 Changes[i - 1].Kind == tok::comment;
119 }
120 // FIXME: The last token is currently not always an eof token; in those
121 // cases, setting TokenLength of the last token to 0 is wrong.
122 Changes.back().TokenLength = 0;
123 Changes.back().IsTrailingComment = Changes.back().Kind == tok::comment;
124
125 const WhitespaceManager::Change *LastBlockComment = nullptr;
126 for (auto &Change : Changes) {
127 Change.StartOfBlockComment = nullptr;
128 Change.IndentationOffset = 0;
129 if (Change.Kind == tok::comment) {
130 LastBlockComment = &Change;
131 } else if (Change.Kind == tok::unknown) {
132 if ((Change.StartOfBlockComment = LastBlockComment))
133 Change.IndentationOffset =
134 Change.StartOfTokenColumn -
135 Change.StartOfBlockComment->StartOfTokenColumn;
136 } else {
137 LastBlockComment = nullptr;
138 }
139 }
140 }
141
alignTrailingComments()142 void WhitespaceManager::alignTrailingComments() {
143 unsigned MinColumn = 0;
144 unsigned MaxColumn = UINT_MAX;
145 unsigned StartOfSequence = 0;
146 bool BreakBeforeNext = false;
147 unsigned Newlines = 0;
148 for (unsigned i = 0, e = Changes.size(); i != e; ++i) {
149 if (Changes[i].StartOfBlockComment)
150 continue;
151 Newlines += Changes[i].NewlinesBefore;
152 if (!Changes[i].IsTrailingComment)
153 continue;
154
155 unsigned ChangeMinColumn = Changes[i].StartOfTokenColumn;
156 unsigned ChangeMaxColumn = Style.ColumnLimit - Changes[i].TokenLength;
157 if (i + 1 != e && Changes[i + 1].ContinuesPPDirective)
158 ChangeMaxColumn -= 2;
159 // If this comment follows an } in column 0, it probably documents the
160 // closing of a namespace and we don't want to align it.
161 bool FollowsRBraceInColumn0 = i > 0 && Changes[i].NewlinesBefore == 0 &&
162 Changes[i - 1].Kind == tok::r_brace &&
163 Changes[i - 1].StartOfTokenColumn == 0;
164 bool WasAlignedWithStartOfNextLine = false;
165 if (Changes[i].NewlinesBefore == 1) { // A comment on its own line.
166 for (unsigned j = i + 1; j != e; ++j) {
167 if (Changes[j].Kind != tok::comment) { // Skip over comments.
168 // The start of the next token was previously aligned with the
169 // start of this comment.
170 WasAlignedWithStartOfNextLine =
171 (SourceMgr.getSpellingColumnNumber(
172 Changes[i].OriginalWhitespaceRange.getEnd()) ==
173 SourceMgr.getSpellingColumnNumber(
174 Changes[j].OriginalWhitespaceRange.getEnd()));
175 break;
176 }
177 }
178 }
179 if (!Style.AlignTrailingComments || FollowsRBraceInColumn0) {
180 alignTrailingComments(StartOfSequence, i, MinColumn);
181 MinColumn = ChangeMinColumn;
182 MaxColumn = ChangeMinColumn;
183 StartOfSequence = i;
184 } else if (BreakBeforeNext || Newlines > 1 ||
185 (ChangeMinColumn > MaxColumn || ChangeMaxColumn < MinColumn) ||
186 // Break the comment sequence if the previous line did not end
187 // in a trailing comment.
188 (Changes[i].NewlinesBefore == 1 && i > 0 &&
189 !Changes[i - 1].IsTrailingComment) ||
190 WasAlignedWithStartOfNextLine) {
191 alignTrailingComments(StartOfSequence, i, MinColumn);
192 MinColumn = ChangeMinColumn;
193 MaxColumn = ChangeMaxColumn;
194 StartOfSequence = i;
195 } else {
196 MinColumn = std::max(MinColumn, ChangeMinColumn);
197 MaxColumn = std::min(MaxColumn, ChangeMaxColumn);
198 }
199 BreakBeforeNext =
200 (i == 0) || (Changes[i].NewlinesBefore > 1) ||
201 // Never start a sequence with a comment at the beginning of
202 // the line.
203 (Changes[i].NewlinesBefore == 1 && StartOfSequence == i);
204 Newlines = 0;
205 }
206 alignTrailingComments(StartOfSequence, Changes.size(), MinColumn);
207 }
208
alignTrailingComments(unsigned Start,unsigned End,unsigned Column)209 void WhitespaceManager::alignTrailingComments(unsigned Start, unsigned End,
210 unsigned Column) {
211 for (unsigned i = Start; i != End; ++i) {
212 int Shift = 0;
213 if (Changes[i].IsTrailingComment) {
214 Shift = Column - Changes[i].StartOfTokenColumn;
215 }
216 if (Changes[i].StartOfBlockComment) {
217 Shift = Changes[i].IndentationOffset +
218 Changes[i].StartOfBlockComment->StartOfTokenColumn -
219 Changes[i].StartOfTokenColumn;
220 }
221 assert(Shift >= 0);
222 Changes[i].Spaces += Shift;
223 if (i + 1 != End)
224 Changes[i + 1].PreviousEndOfTokenColumn += Shift;
225 Changes[i].StartOfTokenColumn += Shift;
226 }
227 }
228
alignEscapedNewlines()229 void WhitespaceManager::alignEscapedNewlines() {
230 unsigned MaxEndOfLine =
231 Style.AlignEscapedNewlinesLeft ? 0 : Style.ColumnLimit;
232 unsigned StartOfMacro = 0;
233 for (unsigned i = 1, e = Changes.size(); i < e; ++i) {
234 Change &C = Changes[i];
235 if (C.NewlinesBefore > 0) {
236 if (C.ContinuesPPDirective) {
237 MaxEndOfLine = std::max(C.PreviousEndOfTokenColumn + 2, MaxEndOfLine);
238 } else {
239 alignEscapedNewlines(StartOfMacro + 1, i, MaxEndOfLine);
240 MaxEndOfLine = Style.AlignEscapedNewlinesLeft ? 0 : Style.ColumnLimit;
241 StartOfMacro = i;
242 }
243 }
244 }
245 alignEscapedNewlines(StartOfMacro + 1, Changes.size(), MaxEndOfLine);
246 }
247
alignEscapedNewlines(unsigned Start,unsigned End,unsigned Column)248 void WhitespaceManager::alignEscapedNewlines(unsigned Start, unsigned End,
249 unsigned Column) {
250 for (unsigned i = Start; i < End; ++i) {
251 Change &C = Changes[i];
252 if (C.NewlinesBefore > 0) {
253 assert(C.ContinuesPPDirective);
254 if (C.PreviousEndOfTokenColumn + 1 > Column)
255 C.EscapedNewlineColumn = 0;
256 else
257 C.EscapedNewlineColumn = Column;
258 }
259 }
260 }
261
generateChanges()262 void WhitespaceManager::generateChanges() {
263 for (unsigned i = 0, e = Changes.size(); i != e; ++i) {
264 const Change &C = Changes[i];
265 if (C.CreateReplacement) {
266 std::string ReplacementText = C.PreviousLinePostfix;
267 if (C.ContinuesPPDirective)
268 appendNewlineText(ReplacementText, C.NewlinesBefore,
269 C.PreviousEndOfTokenColumn, C.EscapedNewlineColumn);
270 else
271 appendNewlineText(ReplacementText, C.NewlinesBefore);
272 appendIndentText(ReplacementText, C.IndentLevel, std::max(0, C.Spaces),
273 C.StartOfTokenColumn - std::max(0, C.Spaces));
274 ReplacementText.append(C.CurrentLinePrefix);
275 storeReplacement(C.OriginalWhitespaceRange, ReplacementText);
276 }
277 }
278 }
279
storeReplacement(const SourceRange & Range,StringRef Text)280 void WhitespaceManager::storeReplacement(const SourceRange &Range,
281 StringRef Text) {
282 unsigned WhitespaceLength = SourceMgr.getFileOffset(Range.getEnd()) -
283 SourceMgr.getFileOffset(Range.getBegin());
284 // Don't create a replacement, if it does not change anything.
285 if (StringRef(SourceMgr.getCharacterData(Range.getBegin()),
286 WhitespaceLength) == Text)
287 return;
288 Replaces.insert(tooling::Replacement(
289 SourceMgr, CharSourceRange::getCharRange(Range), Text));
290 }
291
appendNewlineText(std::string & Text,unsigned Newlines)292 void WhitespaceManager::appendNewlineText(std::string &Text,
293 unsigned Newlines) {
294 for (unsigned i = 0; i < Newlines; ++i)
295 Text.append(UseCRLF ? "\r\n" : "\n");
296 }
297
appendNewlineText(std::string & Text,unsigned Newlines,unsigned PreviousEndOfTokenColumn,unsigned EscapedNewlineColumn)298 void WhitespaceManager::appendNewlineText(std::string &Text, unsigned Newlines,
299 unsigned PreviousEndOfTokenColumn,
300 unsigned EscapedNewlineColumn) {
301 if (Newlines > 0) {
302 unsigned Offset =
303 std::min<int>(EscapedNewlineColumn - 1, PreviousEndOfTokenColumn);
304 for (unsigned i = 0; i < Newlines; ++i) {
305 Text.append(std::string(EscapedNewlineColumn - Offset - 1, ' '));
306 Text.append(UseCRLF ? "\\\r\n" : "\\\n");
307 Offset = 0;
308 }
309 }
310 }
311
appendIndentText(std::string & Text,unsigned IndentLevel,unsigned Spaces,unsigned WhitespaceStartColumn)312 void WhitespaceManager::appendIndentText(std::string &Text,
313 unsigned IndentLevel, unsigned Spaces,
314 unsigned WhitespaceStartColumn) {
315 switch (Style.UseTab) {
316 case FormatStyle::UT_Never:
317 Text.append(std::string(Spaces, ' '));
318 break;
319 case FormatStyle::UT_Always: {
320 unsigned FirstTabWidth =
321 Style.TabWidth - WhitespaceStartColumn % Style.TabWidth;
322 // Indent with tabs only when there's at least one full tab.
323 if (FirstTabWidth + Style.TabWidth <= Spaces) {
324 Spaces -= FirstTabWidth;
325 Text.append("\t");
326 }
327 Text.append(std::string(Spaces / Style.TabWidth, '\t'));
328 Text.append(std::string(Spaces % Style.TabWidth, ' '));
329 break;
330 }
331 case FormatStyle::UT_ForIndentation:
332 if (WhitespaceStartColumn == 0) {
333 unsigned Indentation = IndentLevel * Style.IndentWidth;
334 // This happens, e.g. when a line in a block comment is indented less than
335 // the first one.
336 if (Indentation > Spaces)
337 Indentation = Spaces;
338 unsigned Tabs = Indentation / Style.TabWidth;
339 Text.append(std::string(Tabs, '\t'));
340 Spaces -= Tabs * Style.TabWidth;
341 }
342 Text.append(std::string(Spaces, ' '));
343 break;
344 }
345 }
346
347 } // namespace format
348 } // namespace clang
349