1 //===- Rewriter.cpp - Code rewriting interface ----------------------------===//
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 // This file defines the Rewriter class, which is used for code
10 // transformations.
11 //
12 //===----------------------------------------------------------------------===//
13
14 #include "clang/Rewrite/Core/Rewriter.h"
15 #include "clang/Basic/Diagnostic.h"
16 #include "clang/Basic/DiagnosticIDs.h"
17 #include "clang/Basic/FileManager.h"
18 #include "clang/Basic/SourceLocation.h"
19 #include "clang/Basic/SourceManager.h"
20 #include "clang/Lex/Lexer.h"
21 #include "clang/Rewrite/Core/RewriteBuffer.h"
22 #include "clang/Rewrite/Core/RewriteRope.h"
23 #include "llvm/ADT/SmallString.h"
24 #include "llvm/ADT/SmallVector.h"
25 #include "llvm/ADT/StringRef.h"
26 #include "llvm/Support/FileSystem.h"
27 #include "llvm/Support/raw_ostream.h"
28 #include <cassert>
29 #include <iterator>
30 #include <map>
31 #include <memory>
32 #include <system_error>
33 #include <utility>
34
35 using namespace clang;
36
write(raw_ostream & os) const37 raw_ostream &RewriteBuffer::write(raw_ostream &os) const {
38 // Walk RewriteRope chunks efficiently using MoveToNextPiece() instead of the
39 // character iterator.
40 for (RopePieceBTreeIterator I = begin(), E = end(); I != E;
41 I.MoveToNextPiece())
42 os << I.piece();
43 return os;
44 }
45
46 /// Return true if this character is non-new-line whitespace:
47 /// ' ', '\\t', '\\f', '\\v', '\\r'.
isWhitespaceExceptNL(unsigned char c)48 static inline bool isWhitespaceExceptNL(unsigned char c) {
49 switch (c) {
50 case ' ':
51 case '\t':
52 case '\f':
53 case '\v':
54 case '\r':
55 return true;
56 default:
57 return false;
58 }
59 }
60
RemoveText(unsigned OrigOffset,unsigned Size,bool removeLineIfEmpty)61 void RewriteBuffer::RemoveText(unsigned OrigOffset, unsigned Size,
62 bool removeLineIfEmpty) {
63 // Nothing to remove, exit early.
64 if (Size == 0) return;
65
66 unsigned RealOffset = getMappedOffset(OrigOffset, true);
67 assert(RealOffset+Size <= Buffer.size() && "Invalid location");
68
69 // Remove the dead characters.
70 Buffer.erase(RealOffset, Size);
71
72 // Add a delta so that future changes are offset correctly.
73 AddReplaceDelta(OrigOffset, -Size);
74
75 if (removeLineIfEmpty) {
76 // Find the line that the remove occurred and if it is completely empty
77 // remove the line as well.
78
79 iterator curLineStart = begin();
80 unsigned curLineStartOffs = 0;
81 iterator posI = begin();
82 for (unsigned i = 0; i != RealOffset; ++i) {
83 if (*posI == '\n') {
84 curLineStart = posI;
85 ++curLineStart;
86 curLineStartOffs = i + 1;
87 }
88 ++posI;
89 }
90
91 unsigned lineSize = 0;
92 posI = curLineStart;
93 while (posI != end() && isWhitespaceExceptNL(*posI)) {
94 ++posI;
95 ++lineSize;
96 }
97 if (posI != end() && *posI == '\n') {
98 Buffer.erase(curLineStartOffs, lineSize + 1/* + '\n'*/);
99 // FIXME: Here, the offset of the start of the line is supposed to be
100 // expressed in terms of the original input not the "real" rewrite
101 // buffer. How do we compute that reliably? It might be tempting to use
102 // curLineStartOffs + OrigOffset - RealOffset, but that assumes the
103 // difference between the original and real offset is the same at the
104 // removed text and at the start of the line, but that's not true if
105 // edits were previously made earlier on the line. This bug is also
106 // documented by a FIXME on the definition of
107 // clang::Rewriter::RewriteOptions::RemoveLineIfEmpty. A reproducer for
108 // the implementation below is the test RemoveLineIfEmpty in
109 // clang/unittests/Rewrite/RewriteBufferTest.cpp.
110 AddReplaceDelta(curLineStartOffs, -(lineSize + 1/* + '\n'*/));
111 }
112 }
113 }
114
InsertText(unsigned OrigOffset,StringRef Str,bool InsertAfter)115 void RewriteBuffer::InsertText(unsigned OrigOffset, StringRef Str,
116 bool InsertAfter) {
117 // Nothing to insert, exit early.
118 if (Str.empty()) return;
119
120 unsigned RealOffset = getMappedOffset(OrigOffset, InsertAfter);
121 Buffer.insert(RealOffset, Str.begin(), Str.end());
122
123 // Add a delta so that future changes are offset correctly.
124 AddInsertDelta(OrigOffset, Str.size());
125 }
126
127 /// ReplaceText - This method replaces a range of characters in the input
128 /// buffer with a new string. This is effectively a combined "remove+insert"
129 /// operation.
ReplaceText(unsigned OrigOffset,unsigned OrigLength,StringRef NewStr)130 void RewriteBuffer::ReplaceText(unsigned OrigOffset, unsigned OrigLength,
131 StringRef NewStr) {
132 unsigned RealOffset = getMappedOffset(OrigOffset, true);
133 Buffer.erase(RealOffset, OrigLength);
134 Buffer.insert(RealOffset, NewStr.begin(), NewStr.end());
135 if (OrigLength != NewStr.size())
136 AddReplaceDelta(OrigOffset, NewStr.size() - OrigLength);
137 }
138
139 //===----------------------------------------------------------------------===//
140 // Rewriter class
141 //===----------------------------------------------------------------------===//
142
143 /// getRangeSize - Return the size in bytes of the specified range if they
144 /// are in the same file. If not, this returns -1.
getRangeSize(const CharSourceRange & Range,RewriteOptions opts) const145 int Rewriter::getRangeSize(const CharSourceRange &Range,
146 RewriteOptions opts) const {
147 if (!isRewritable(Range.getBegin()) ||
148 !isRewritable(Range.getEnd())) return -1;
149
150 FileID StartFileID, EndFileID;
151 unsigned StartOff = getLocationOffsetAndFileID(Range.getBegin(), StartFileID);
152 unsigned EndOff = getLocationOffsetAndFileID(Range.getEnd(), EndFileID);
153
154 if (StartFileID != EndFileID)
155 return -1;
156
157 // If edits have been made to this buffer, the delta between the range may
158 // have changed.
159 std::map<FileID, RewriteBuffer>::const_iterator I =
160 RewriteBuffers.find(StartFileID);
161 if (I != RewriteBuffers.end()) {
162 const RewriteBuffer &RB = I->second;
163 EndOff = RB.getMappedOffset(EndOff, opts.IncludeInsertsAtEndOfRange);
164 StartOff = RB.getMappedOffset(StartOff, !opts.IncludeInsertsAtBeginOfRange);
165 }
166
167 // Adjust the end offset to the end of the last token, instead of being the
168 // start of the last token if this is a token range.
169 if (Range.isTokenRange())
170 EndOff += Lexer::MeasureTokenLength(Range.getEnd(), *SourceMgr, *LangOpts);
171
172 return EndOff-StartOff;
173 }
174
getRangeSize(SourceRange Range,RewriteOptions opts) const175 int Rewriter::getRangeSize(SourceRange Range, RewriteOptions opts) const {
176 return getRangeSize(CharSourceRange::getTokenRange(Range), opts);
177 }
178
179 /// getRewrittenText - Return the rewritten form of the text in the specified
180 /// range. If the start or end of the range was unrewritable or if they are
181 /// in different buffers, this returns an empty string.
182 ///
183 /// Note that this method is not particularly efficient.
getRewrittenText(CharSourceRange Range) const184 std::string Rewriter::getRewrittenText(CharSourceRange Range) const {
185 if (!isRewritable(Range.getBegin()) ||
186 !isRewritable(Range.getEnd()))
187 return {};
188
189 FileID StartFileID, EndFileID;
190 unsigned StartOff, EndOff;
191 StartOff = getLocationOffsetAndFileID(Range.getBegin(), StartFileID);
192 EndOff = getLocationOffsetAndFileID(Range.getEnd(), EndFileID);
193
194 if (StartFileID != EndFileID)
195 return {}; // Start and end in different buffers.
196
197 // If edits have been made to this buffer, the delta between the range may
198 // have changed.
199 std::map<FileID, RewriteBuffer>::const_iterator I =
200 RewriteBuffers.find(StartFileID);
201 if (I == RewriteBuffers.end()) {
202 // If the buffer hasn't been rewritten, just return the text from the input.
203 const char *Ptr = SourceMgr->getCharacterData(Range.getBegin());
204
205 // Adjust the end offset to the end of the last token, instead of being the
206 // start of the last token.
207 if (Range.isTokenRange())
208 EndOff +=
209 Lexer::MeasureTokenLength(Range.getEnd(), *SourceMgr, *LangOpts);
210 return std::string(Ptr, Ptr+EndOff-StartOff);
211 }
212
213 const RewriteBuffer &RB = I->second;
214 EndOff = RB.getMappedOffset(EndOff, true);
215 StartOff = RB.getMappedOffset(StartOff);
216
217 // Adjust the end offset to the end of the last token, instead of being the
218 // start of the last token.
219 if (Range.isTokenRange())
220 EndOff += Lexer::MeasureTokenLength(Range.getEnd(), *SourceMgr, *LangOpts);
221
222 // Advance the iterators to the right spot, yay for linear time algorithms.
223 RewriteBuffer::iterator Start = RB.begin();
224 std::advance(Start, StartOff);
225 RewriteBuffer::iterator End = Start;
226 std::advance(End, EndOff-StartOff);
227
228 return std::string(Start, End);
229 }
230
getLocationOffsetAndFileID(SourceLocation Loc,FileID & FID) const231 unsigned Rewriter::getLocationOffsetAndFileID(SourceLocation Loc,
232 FileID &FID) const {
233 assert(Loc.isValid() && "Invalid location");
234 std::pair<FileID, unsigned> V = SourceMgr->getDecomposedLoc(Loc);
235 FID = V.first;
236 return V.second;
237 }
238
239 /// getEditBuffer - Get or create a RewriteBuffer for the specified FileID.
getEditBuffer(FileID FID)240 RewriteBuffer &Rewriter::getEditBuffer(FileID FID) {
241 std::map<FileID, RewriteBuffer>::iterator I =
242 RewriteBuffers.lower_bound(FID);
243 if (I != RewriteBuffers.end() && I->first == FID)
244 return I->second;
245 I = RewriteBuffers.insert(I, std::make_pair(FID, RewriteBuffer()));
246
247 StringRef MB = SourceMgr->getBufferData(FID);
248 I->second.Initialize(MB.begin(), MB.end());
249
250 return I->second;
251 }
252
253 /// InsertText - Insert the specified string at the specified location in the
254 /// original buffer.
InsertText(SourceLocation Loc,StringRef Str,bool InsertAfter,bool indentNewLines)255 bool Rewriter::InsertText(SourceLocation Loc, StringRef Str,
256 bool InsertAfter, bool indentNewLines) {
257 if (!isRewritable(Loc)) return true;
258 FileID FID;
259 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
260
261 SmallString<128> indentedStr;
262 if (indentNewLines && Str.find('\n') != StringRef::npos) {
263 StringRef MB = SourceMgr->getBufferData(FID);
264
265 unsigned lineNo = SourceMgr->getLineNumber(FID, StartOffs) - 1;
266 const SrcMgr::ContentCache *Content =
267 &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
268 unsigned lineOffs = Content->SourceLineCache[lineNo];
269
270 // Find the whitespace at the start of the line.
271 StringRef indentSpace;
272 {
273 unsigned i = lineOffs;
274 while (isWhitespaceExceptNL(MB[i]))
275 ++i;
276 indentSpace = MB.substr(lineOffs, i-lineOffs);
277 }
278
279 SmallVector<StringRef, 4> lines;
280 Str.split(lines, "\n");
281
282 for (unsigned i = 0, e = lines.size(); i != e; ++i) {
283 indentedStr += lines[i];
284 if (i < e-1) {
285 indentedStr += '\n';
286 indentedStr += indentSpace;
287 }
288 }
289 Str = indentedStr.str();
290 }
291
292 getEditBuffer(FID).InsertText(StartOffs, Str, InsertAfter);
293 return false;
294 }
295
InsertTextAfterToken(SourceLocation Loc,StringRef Str)296 bool Rewriter::InsertTextAfterToken(SourceLocation Loc, StringRef Str) {
297 if (!isRewritable(Loc)) return true;
298 FileID FID;
299 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
300 RewriteOptions rangeOpts;
301 rangeOpts.IncludeInsertsAtBeginOfRange = false;
302 StartOffs += getRangeSize(SourceRange(Loc, Loc), rangeOpts);
303 getEditBuffer(FID).InsertText(StartOffs, Str, /*InsertAfter*/true);
304 return false;
305 }
306
307 /// RemoveText - Remove the specified text region.
RemoveText(SourceLocation Start,unsigned Length,RewriteOptions opts)308 bool Rewriter::RemoveText(SourceLocation Start, unsigned Length,
309 RewriteOptions opts) {
310 if (!isRewritable(Start)) return true;
311 FileID FID;
312 unsigned StartOffs = getLocationOffsetAndFileID(Start, FID);
313 getEditBuffer(FID).RemoveText(StartOffs, Length, opts.RemoveLineIfEmpty);
314 return false;
315 }
316
317 /// ReplaceText - This method replaces a range of characters in the input
318 /// buffer with a new string. This is effectively a combined "remove/insert"
319 /// operation.
ReplaceText(SourceLocation Start,unsigned OrigLength,StringRef NewStr)320 bool Rewriter::ReplaceText(SourceLocation Start, unsigned OrigLength,
321 StringRef NewStr) {
322 if (!isRewritable(Start)) return true;
323 FileID StartFileID;
324 unsigned StartOffs = getLocationOffsetAndFileID(Start, StartFileID);
325
326 getEditBuffer(StartFileID).ReplaceText(StartOffs, OrigLength, NewStr);
327 return false;
328 }
329
ReplaceText(SourceRange range,SourceRange replacementRange)330 bool Rewriter::ReplaceText(SourceRange range, SourceRange replacementRange) {
331 if (!isRewritable(range.getBegin())) return true;
332 if (!isRewritable(range.getEnd())) return true;
333 if (replacementRange.isInvalid()) return true;
334 SourceLocation start = range.getBegin();
335 unsigned origLength = getRangeSize(range);
336 unsigned newLength = getRangeSize(replacementRange);
337 FileID FID;
338 unsigned newOffs = getLocationOffsetAndFileID(replacementRange.getBegin(),
339 FID);
340 StringRef MB = SourceMgr->getBufferData(FID);
341 return ReplaceText(start, origLength, MB.substr(newOffs, newLength));
342 }
343
IncreaseIndentation(CharSourceRange range,SourceLocation parentIndent)344 bool Rewriter::IncreaseIndentation(CharSourceRange range,
345 SourceLocation parentIndent) {
346 if (range.isInvalid()) return true;
347 if (!isRewritable(range.getBegin())) return true;
348 if (!isRewritable(range.getEnd())) return true;
349 if (!isRewritable(parentIndent)) return true;
350
351 FileID StartFileID, EndFileID, parentFileID;
352 unsigned StartOff, EndOff, parentOff;
353
354 StartOff = getLocationOffsetAndFileID(range.getBegin(), StartFileID);
355 EndOff = getLocationOffsetAndFileID(range.getEnd(), EndFileID);
356 parentOff = getLocationOffsetAndFileID(parentIndent, parentFileID);
357
358 if (StartFileID != EndFileID || StartFileID != parentFileID)
359 return true;
360 if (StartOff > EndOff)
361 return true;
362
363 FileID FID = StartFileID;
364 StringRef MB = SourceMgr->getBufferData(FID);
365
366 unsigned parentLineNo = SourceMgr->getLineNumber(FID, parentOff) - 1;
367 unsigned startLineNo = SourceMgr->getLineNumber(FID, StartOff) - 1;
368 unsigned endLineNo = SourceMgr->getLineNumber(FID, EndOff) - 1;
369
370 const SrcMgr::ContentCache *Content =
371 &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
372
373 // Find where the lines start.
374 unsigned parentLineOffs = Content->SourceLineCache[parentLineNo];
375 unsigned startLineOffs = Content->SourceLineCache[startLineNo];
376
377 // Find the whitespace at the start of each line.
378 StringRef parentSpace, startSpace;
379 {
380 unsigned i = parentLineOffs;
381 while (isWhitespaceExceptNL(MB[i]))
382 ++i;
383 parentSpace = MB.substr(parentLineOffs, i-parentLineOffs);
384
385 i = startLineOffs;
386 while (isWhitespaceExceptNL(MB[i]))
387 ++i;
388 startSpace = MB.substr(startLineOffs, i-startLineOffs);
389 }
390 if (parentSpace.size() >= startSpace.size())
391 return true;
392 if (!startSpace.startswith(parentSpace))
393 return true;
394
395 StringRef indent = startSpace.substr(parentSpace.size());
396
397 // Indent the lines between start/end offsets.
398 RewriteBuffer &RB = getEditBuffer(FID);
399 for (unsigned lineNo = startLineNo; lineNo <= endLineNo; ++lineNo) {
400 unsigned offs = Content->SourceLineCache[lineNo];
401 unsigned i = offs;
402 while (isWhitespaceExceptNL(MB[i]))
403 ++i;
404 StringRef origIndent = MB.substr(offs, i-offs);
405 if (origIndent.startswith(startSpace))
406 RB.InsertText(offs, indent, /*InsertAfter=*/false);
407 }
408
409 return false;
410 }
411
412 namespace {
413
414 // A wrapper for a file stream that atomically overwrites the target.
415 //
416 // Creates a file output stream for a temporary file in the constructor,
417 // which is later accessible via getStream() if ok() return true.
418 // Flushes the stream and moves the temporary file to the target location
419 // in the destructor.
420 class AtomicallyMovedFile {
421 public:
AtomicallyMovedFile(DiagnosticsEngine & Diagnostics,StringRef Filename,bool & AllWritten)422 AtomicallyMovedFile(DiagnosticsEngine &Diagnostics, StringRef Filename,
423 bool &AllWritten)
424 : Diagnostics(Diagnostics), Filename(Filename), AllWritten(AllWritten) {
425 TempFilename = Filename;
426 TempFilename += "-%%%%%%%%";
427 int FD;
428 if (llvm::sys::fs::createUniqueFile(TempFilename, FD, TempFilename)) {
429 AllWritten = false;
430 Diagnostics.Report(clang::diag::err_unable_to_make_temp)
431 << TempFilename;
432 } else {
433 FileStream.reset(new llvm::raw_fd_ostream(FD, /*shouldClose=*/true));
434 }
435 }
436
~AtomicallyMovedFile()437 ~AtomicallyMovedFile() {
438 if (!ok()) return;
439
440 // Close (will also flush) theFileStream.
441 FileStream->close();
442 if (std::error_code ec = llvm::sys::fs::rename(TempFilename, Filename)) {
443 AllWritten = false;
444 Diagnostics.Report(clang::diag::err_unable_to_rename_temp)
445 << TempFilename << Filename << ec.message();
446 // If the remove fails, there's not a lot we can do - this is already an
447 // error.
448 llvm::sys::fs::remove(TempFilename);
449 }
450 }
451
ok()452 bool ok() { return (bool)FileStream; }
getStream()453 raw_ostream &getStream() { return *FileStream; }
454
455 private:
456 DiagnosticsEngine &Diagnostics;
457 StringRef Filename;
458 SmallString<128> TempFilename;
459 std::unique_ptr<llvm::raw_fd_ostream> FileStream;
460 bool &AllWritten;
461 };
462
463 } // namespace
464
overwriteChangedFiles()465 bool Rewriter::overwriteChangedFiles() {
466 bool AllWritten = true;
467 for (buffer_iterator I = buffer_begin(), E = buffer_end(); I != E; ++I) {
468 const FileEntry *Entry =
469 getSourceMgr().getFileEntryForID(I->first);
470 AtomicallyMovedFile File(getSourceMgr().getDiagnostics(), Entry->getName(),
471 AllWritten);
472 if (File.ok()) {
473 I->second.write(File.getStream());
474 }
475 }
476 return !AllWritten;
477 }
478