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/SourceLocation.h"
18#include "clang/Basic/SourceManager.h"
19#include "clang/Lex/Lexer.h"
20#include "clang/Rewrite/Core/RewriteBuffer.h"
21#include "clang/Rewrite/Core/RewriteRope.h"
22#include "llvm/ADT/SmallVector.h"
23#include "llvm/ADT/StringRef.h"
24#include "llvm/Support/Error.h"
25#include "llvm/Support/raw_ostream.h"
26#include <cassert>
27#include <iterator>
28#include <map>
29#include <utility>
30
31using namespace clang;
32
33raw_ostream &RewriteBuffer::write(raw_ostream &os) const {
34 // Walk RewriteRope chunks efficiently using MoveToNextPiece() instead of the
35 // character iterator.
36 for (RopePieceBTreeIterator I = begin(), E = end(); I != E;
37 I.MoveToNextPiece())
38 os << I.piece();
39 return os;
40}
41
42/// Return true if this character is non-new-line whitespace:
43/// ' ', '\\t', '\\f', '\\v', '\\r'.
44static inline bool isWhitespaceExceptNL(unsigned char c) {
45 switch (c) {
46 case ' ':
47 case '\t':
48 case '\f':
49 case '\v':
50 case '\r':
51 return true;
52 default:
53 return false;
54 }
55}
56
57void RewriteBuffer::RemoveText(unsigned OrigOffset, unsigned Size,
58 bool removeLineIfEmpty) {
59 // Nothing to remove, exit early.
60 if (Size == 0) return;
61
62 unsigned RealOffset = getMappedOffset(OrigOffset, AfterInserts: true);
63 assert(RealOffset+Size <= Buffer.size() && "Invalid location");
64
65 // Remove the dead characters.
66 Buffer.erase(Offset: RealOffset, NumBytes: Size);
67
68 // Add a delta so that future changes are offset correctly.
69 AddReplaceDelta(OrigOffset, Change: -Size);
70
71 if (removeLineIfEmpty) {
72 // Find the line that the remove occurred and if it is completely empty
73 // remove the line as well.
74
75 iterator curLineStart = begin();
76 unsigned curLineStartOffs = 0;
77 iterator posI = begin();
78 for (unsigned i = 0; i != RealOffset; ++i) {
79 if (*posI == '\n') {
80 curLineStart = posI;
81 ++curLineStart;
82 curLineStartOffs = i + 1;
83 }
84 ++posI;
85 }
86
87 unsigned lineSize = 0;
88 posI = curLineStart;
89 while (posI != end() && isWhitespaceExceptNL(c: *posI)) {
90 ++posI;
91 ++lineSize;
92 }
93 if (posI != end() && *posI == '\n') {
94 Buffer.erase(Offset: curLineStartOffs, NumBytes: lineSize + 1/* + '\n'*/);
95 // FIXME: Here, the offset of the start of the line is supposed to be
96 // expressed in terms of the original input not the "real" rewrite
97 // buffer. How do we compute that reliably? It might be tempting to use
98 // curLineStartOffs + OrigOffset - RealOffset, but that assumes the
99 // difference between the original and real offset is the same at the
100 // removed text and at the start of the line, but that's not true if
101 // edits were previously made earlier on the line. This bug is also
102 // documented by a FIXME on the definition of
103 // clang::Rewriter::RewriteOptions::RemoveLineIfEmpty. A reproducer for
104 // the implementation below is the test RemoveLineIfEmpty in
105 // clang/unittests/Rewrite/RewriteBufferTest.cpp.
106 AddReplaceDelta(OrigOffset: curLineStartOffs, Change: -(lineSize + 1/* + '\n'*/));
107 }
108 }
109}
110
111void RewriteBuffer::InsertText(unsigned OrigOffset, StringRef Str,
112 bool InsertAfter) {
113 // Nothing to insert, exit early.
114 if (Str.empty()) return;
115
116 unsigned RealOffset = getMappedOffset(OrigOffset, AfterInserts: InsertAfter);
117 Buffer.insert(Offset: RealOffset, Start: Str.begin(), End: Str.end());
118
119 // Add a delta so that future changes are offset correctly.
120 AddInsertDelta(OrigOffset, Change: Str.size());
121}
122
123/// ReplaceText - This method replaces a range of characters in the input
124/// buffer with a new string. This is effectively a combined "remove+insert"
125/// operation.
126void RewriteBuffer::ReplaceText(unsigned OrigOffset, unsigned OrigLength,
127 StringRef NewStr) {
128 unsigned RealOffset = getMappedOffset(OrigOffset, AfterInserts: true);
129 Buffer.erase(Offset: RealOffset, NumBytes: OrigLength);
130 Buffer.insert(Offset: RealOffset, Start: NewStr.begin(), End: NewStr.end());
131 if (OrigLength != NewStr.size())
132 AddReplaceDelta(OrigOffset, Change: NewStr.size() - OrigLength);
133}
134
135//===----------------------------------------------------------------------===//
136// Rewriter class
137//===----------------------------------------------------------------------===//
138
139/// getRangeSize - Return the size in bytes of the specified range if they
140/// are in the same file. If not, this returns -1.
141int Rewriter::getRangeSize(const CharSourceRange &Range,
142 RewriteOptions opts) const {
143 if (!isRewritable(Loc: Range.getBegin()) ||
144 !isRewritable(Loc: Range.getEnd())) return -1;
145
146 FileID StartFileID, EndFileID;
147 unsigned StartOff = getLocationOffsetAndFileID(Loc: Range.getBegin(), FID&: StartFileID);
148 unsigned EndOff = getLocationOffsetAndFileID(Loc: Range.getEnd(), FID&: EndFileID);
149
150 if (StartFileID != EndFileID)
151 return -1;
152
153 // If edits have been made to this buffer, the delta between the range may
154 // have changed.
155 std::map<FileID, RewriteBuffer>::const_iterator I =
156 RewriteBuffers.find(x: StartFileID);
157 if (I != RewriteBuffers.end()) {
158 const RewriteBuffer &RB = I->second;
159 EndOff = RB.getMappedOffset(OrigOffset: EndOff, AfterInserts: opts.IncludeInsertsAtEndOfRange);
160 StartOff = RB.getMappedOffset(OrigOffset: StartOff, AfterInserts: !opts.IncludeInsertsAtBeginOfRange);
161 }
162
163 // Adjust the end offset to the end of the last token, instead of being the
164 // start of the last token if this is a token range.
165 if (Range.isTokenRange())
166 EndOff += Lexer::MeasureTokenLength(Loc: Range.getEnd(), SM: *SourceMgr, LangOpts: *LangOpts);
167
168 return EndOff-StartOff;
169}
170
171int Rewriter::getRangeSize(SourceRange Range, RewriteOptions opts) const {
172 return getRangeSize(Range: CharSourceRange::getTokenRange(R: Range), opts);
173}
174
175/// getRewrittenText - Return the rewritten form of the text in the specified
176/// range. If the start or end of the range was unrewritable or if they are
177/// in different buffers, this returns an empty string.
178///
179/// Note that this method is not particularly efficient.
180std::string Rewriter::getRewrittenText(CharSourceRange Range) const {
181 if (!isRewritable(Loc: Range.getBegin()) ||
182 !isRewritable(Loc: Range.getEnd()))
183 return {};
184
185 FileID StartFileID, EndFileID;
186 unsigned StartOff, EndOff;
187 StartOff = getLocationOffsetAndFileID(Loc: Range.getBegin(), FID&: StartFileID);
188 EndOff = getLocationOffsetAndFileID(Loc: Range.getEnd(), FID&: EndFileID);
189
190 if (StartFileID != EndFileID)
191 return {}; // Start and end in different buffers.
192
193 // If edits have been made to this buffer, the delta between the range may
194 // have changed.
195 std::map<FileID, RewriteBuffer>::const_iterator I =
196 RewriteBuffers.find(x: StartFileID);
197 if (I == RewriteBuffers.end()) {
198 // If the buffer hasn't been rewritten, just return the text from the input.
199 const char *Ptr = SourceMgr->getCharacterData(SL: Range.getBegin());
200
201 // Adjust the end offset to the end of the last token, instead of being the
202 // start of the last token.
203 if (Range.isTokenRange())
204 EndOff +=
205 Lexer::MeasureTokenLength(Loc: Range.getEnd(), SM: *SourceMgr, LangOpts: *LangOpts);
206 return std::string(Ptr, Ptr+EndOff-StartOff);
207 }
208
209 const RewriteBuffer &RB = I->second;
210 EndOff = RB.getMappedOffset(OrigOffset: EndOff, AfterInserts: true);
211 StartOff = RB.getMappedOffset(OrigOffset: StartOff);
212
213 // Adjust the end offset to the end of the last token, instead of being the
214 // start of the last token.
215 if (Range.isTokenRange())
216 EndOff += Lexer::MeasureTokenLength(Loc: Range.getEnd(), SM: *SourceMgr, LangOpts: *LangOpts);
217
218 // Advance the iterators to the right spot, yay for linear time algorithms.
219 RewriteBuffer::iterator Start = RB.begin();
220 std::advance(i&: Start, n: StartOff);
221 RewriteBuffer::iterator End = Start;
222 assert(EndOff >= StartOff && "Invalid iteration distance");
223 std::advance(i&: End, n: EndOff-StartOff);
224
225 return std::string(Start, End);
226}
227
228unsigned Rewriter::getLocationOffsetAndFileID(SourceLocation Loc,
229 FileID &FID) const {
230 assert(Loc.isValid() && "Invalid location");
231 std::pair<FileID, unsigned> V = SourceMgr->getDecomposedLoc(Loc);
232 FID = V.first;
233 return V.second;
234}
235
236/// getEditBuffer - Get or create a RewriteBuffer for the specified FileID.
237RewriteBuffer &Rewriter::getEditBuffer(FileID FID) {
238 std::map<FileID, RewriteBuffer>::iterator I =
239 RewriteBuffers.lower_bound(x: FID);
240 if (I != RewriteBuffers.end() && I->first == FID)
241 return I->second;
242 I = RewriteBuffers.insert(position: I, x: std::make_pair(x&: FID, y: RewriteBuffer()));
243
244 StringRef MB = SourceMgr->getBufferData(FID);
245 I->second.Initialize(BufStart: MB.begin(), BufEnd: MB.end());
246
247 return I->second;
248}
249
250/// InsertText - Insert the specified string at the specified location in the
251/// original buffer.
252bool Rewriter::InsertText(SourceLocation Loc, StringRef Str,
253 bool InsertAfter, bool indentNewLines) {
254 if (!isRewritable(Loc)) return true;
255 FileID FID;
256 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
257
258 SmallString<128> indentedStr;
259 if (indentNewLines && Str.contains(C: '\n')) {
260 StringRef MB = SourceMgr->getBufferData(FID);
261
262 unsigned lineNo = SourceMgr->getLineNumber(FID, FilePos: StartOffs) - 1;
263 const SrcMgr::ContentCache *Content =
264 &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
265 unsigned lineOffs = Content->SourceLineCache[lineNo];
266
267 // Find the whitespace at the start of the line.
268 StringRef indentSpace;
269 {
270 unsigned i = lineOffs;
271 while (isWhitespaceExceptNL(c: MB[i]))
272 ++i;
273 indentSpace = MB.substr(Start: lineOffs, N: i-lineOffs);
274 }
275
276 SmallVector<StringRef, 4> lines;
277 Str.split(A&: lines, Separator: "\n");
278
279 for (unsigned i = 0, e = lines.size(); i != e; ++i) {
280 indentedStr += lines[i];
281 if (i < e-1) {
282 indentedStr += '\n';
283 indentedStr += indentSpace;
284 }
285 }
286 Str = indentedStr.str();
287 }
288
289 getEditBuffer(FID).InsertText(OrigOffset: StartOffs, Str, InsertAfter);
290 return false;
291}
292
293bool Rewriter::InsertTextAfterToken(SourceLocation Loc, StringRef Str) {
294 if (!isRewritable(Loc)) return true;
295 FileID FID;
296 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
297 RewriteOptions rangeOpts;
298 rangeOpts.IncludeInsertsAtBeginOfRange = false;
299 StartOffs += getRangeSize(Range: SourceRange(Loc, Loc), opts: rangeOpts);
300 getEditBuffer(FID).InsertText(OrigOffset: StartOffs, Str, /*InsertAfter*/true);
301 return false;
302}
303
304/// RemoveText - Remove the specified text region.
305bool Rewriter::RemoveText(SourceLocation Start, unsigned Length,
306 RewriteOptions opts) {
307 if (!isRewritable(Loc: Start)) return true;
308 FileID FID;
309 unsigned StartOffs = getLocationOffsetAndFileID(Loc: Start, FID);
310 getEditBuffer(FID).RemoveText(OrigOffset: StartOffs, Size: Length, removeLineIfEmpty: opts.RemoveLineIfEmpty);
311 return false;
312}
313
314/// ReplaceText - This method replaces a range of characters in the input
315/// buffer with a new string. This is effectively a combined "remove/insert"
316/// operation.
317bool Rewriter::ReplaceText(SourceLocation Start, unsigned OrigLength,
318 StringRef NewStr) {
319 if (!isRewritable(Loc: Start)) return true;
320 FileID StartFileID;
321 unsigned StartOffs = getLocationOffsetAndFileID(Loc: Start, FID&: StartFileID);
322
323 getEditBuffer(FID: StartFileID).ReplaceText(OrigOffset: StartOffs, OrigLength, NewStr);
324 return false;
325}
326
327bool Rewriter::ReplaceText(SourceRange range, SourceRange replacementRange) {
328 if (!isRewritable(Loc: range.getBegin())) return true;
329 if (!isRewritable(Loc: range.getEnd())) return true;
330 if (replacementRange.isInvalid()) return true;
331 SourceLocation start = range.getBegin();
332 unsigned origLength = getRangeSize(Range: range);
333 unsigned newLength = getRangeSize(Range: replacementRange);
334 FileID FID;
335 unsigned newOffs = getLocationOffsetAndFileID(Loc: replacementRange.getBegin(),
336 FID);
337 StringRef MB = SourceMgr->getBufferData(FID);
338 return ReplaceText(Start: start, OrigLength: origLength, NewStr: MB.substr(Start: newOffs, N: newLength));
339}
340
341bool Rewriter::IncreaseIndentation(CharSourceRange range,
342 SourceLocation parentIndent) {
343 if (range.isInvalid()) return true;
344 if (!isRewritable(Loc: range.getBegin())) return true;
345 if (!isRewritable(Loc: range.getEnd())) return true;
346 if (!isRewritable(Loc: parentIndent)) return true;
347
348 FileID StartFileID, EndFileID, parentFileID;
349 unsigned StartOff, EndOff, parentOff;
350
351 StartOff = getLocationOffsetAndFileID(Loc: range.getBegin(), FID&: StartFileID);
352 EndOff = getLocationOffsetAndFileID(Loc: range.getEnd(), FID&: EndFileID);
353 parentOff = getLocationOffsetAndFileID(Loc: parentIndent, FID&: parentFileID);
354
355 if (StartFileID != EndFileID || StartFileID != parentFileID)
356 return true;
357 if (StartOff > EndOff)
358 return true;
359
360 FileID FID = StartFileID;
361 StringRef MB = SourceMgr->getBufferData(FID);
362
363 unsigned parentLineNo = SourceMgr->getLineNumber(FID, FilePos: parentOff) - 1;
364 unsigned startLineNo = SourceMgr->getLineNumber(FID, FilePos: StartOff) - 1;
365 unsigned endLineNo = SourceMgr->getLineNumber(FID, FilePos: EndOff) - 1;
366
367 const SrcMgr::ContentCache *Content =
368 &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
369
370 // Find where the lines start.
371 unsigned parentLineOffs = Content->SourceLineCache[parentLineNo];
372 unsigned startLineOffs = Content->SourceLineCache[startLineNo];
373
374 // Find the whitespace at the start of each line.
375 StringRef parentSpace, startSpace;
376 {
377 unsigned i = parentLineOffs;
378 while (isWhitespaceExceptNL(c: MB[i]))
379 ++i;
380 parentSpace = MB.substr(Start: parentLineOffs, N: i-parentLineOffs);
381
382 i = startLineOffs;
383 while (isWhitespaceExceptNL(c: MB[i]))
384 ++i;
385 startSpace = MB.substr(Start: startLineOffs, N: i-startLineOffs);
386 }
387 if (parentSpace.size() >= startSpace.size())
388 return true;
389 if (!startSpace.starts_with(Prefix: parentSpace))
390 return true;
391
392 StringRef indent = startSpace.substr(Start: parentSpace.size());
393
394 // Indent the lines between start/end offsets.
395 RewriteBuffer &RB = getEditBuffer(FID);
396 for (unsigned lineNo = startLineNo; lineNo <= endLineNo; ++lineNo) {
397 unsigned offs = Content->SourceLineCache[lineNo];
398 unsigned i = offs;
399 while (isWhitespaceExceptNL(c: MB[i]))
400 ++i;
401 StringRef origIndent = MB.substr(Start: offs, N: i-offs);
402 if (origIndent.starts_with(Prefix: startSpace))
403 RB.InsertText(OrigOffset: offs, Str: indent, /*InsertAfter=*/false);
404 }
405
406 return false;
407}
408
409bool Rewriter::overwriteChangedFiles() {
410 bool AllWritten = true;
411 auto& Diag = getSourceMgr().getDiagnostics();
412 unsigned OverwriteFailure = Diag.getCustomDiagID(
413 L: DiagnosticsEngine::Error, FormatString: "unable to overwrite file %0: %1");
414 for (buffer_iterator I = buffer_begin(), E = buffer_end(); I != E; ++I) {
415 OptionalFileEntryRef Entry = getSourceMgr().getFileEntryRefForID(FID: I->first);
416 llvm::SmallString<128> Path(Entry->getName());
417 getSourceMgr().getFileManager().makeAbsolutePath(Path);
418 if (auto Error = llvm::writeToOutput(OutputFileName: Path, Write: [&](llvm::raw_ostream &OS) {
419 I->second.write(os&: OS);
420 return llvm::Error::success();
421 })) {
422 Diag.Report(DiagID: OverwriteFailure)
423 << Entry->getName() << llvm::toString(E: std::move(Error));
424 AllWritten = false;
425 }
426 }
427 return !AllWritten;
428}
429