1 //===--- Rewriter.cpp - Code rewriting interface --------------------------===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // This file defines the Rewriter class, which is used for code
13 //===----------------------------------------------------------------------===//
15 #include "clang/Rewrite/Core/Rewriter.h"
16 #include "clang/Basic/Diagnostic.h"
17 #include "clang/Basic/DiagnosticIDs.h"
18 #include "clang/Basic/FileManager.h"
19 #include "clang/Basic/SourceManager.h"
20 #include "clang/Lex/Lexer.h"
21 #include "llvm/ADT/SmallString.h"
22 #include "llvm/Config/llvm-config.h"
23 #include "llvm/Support/FileSystem.h"
24 #include "llvm/Support/raw_ostream.h"
25 using namespace clang;
27 raw_ostream &RewriteBuffer::write(raw_ostream &os) const {
28 // Walk RewriteRope chunks efficiently using MoveToNextPiece() instead of the
29 // character iterator.
30 for (RopePieceBTreeIterator I = begin(), E = end(); I != E;
36 /// \brief Return true if this character is non-new-line whitespace:
37 /// ' ', '\\t', '\\f', '\\v', '\\r'.
38 static inline bool isWhitespace(unsigned char c) {
51 void RewriteBuffer::RemoveText(unsigned OrigOffset, unsigned Size,
52 bool removeLineIfEmpty) {
53 // Nothing to remove, exit early.
54 if (Size == 0) return;
56 unsigned RealOffset = getMappedOffset(OrigOffset, true);
57 assert(RealOffset+Size < Buffer.size() && "Invalid location");
59 // Remove the dead characters.
60 Buffer.erase(RealOffset, Size);
62 // Add a delta so that future changes are offset correctly.
63 AddReplaceDelta(OrigOffset, -Size);
65 if (removeLineIfEmpty) {
66 // Find the line that the remove occurred and if it is completely empty
67 // remove the line as well.
69 iterator curLineStart = begin();
70 unsigned curLineStartOffs = 0;
71 iterator posI = begin();
72 for (unsigned i = 0; i != RealOffset; ++i) {
76 curLineStartOffs = i + 1;
81 unsigned lineSize = 0;
83 while (posI != end() && isWhitespace(*posI)) {
87 if (posI != end() && *posI == '\n') {
88 Buffer.erase(curLineStartOffs, lineSize + 1/* + '\n'*/);
89 AddReplaceDelta(curLineStartOffs, -(lineSize + 1/* + '\n'*/));
94 void RewriteBuffer::InsertText(unsigned OrigOffset, StringRef Str,
97 // Nothing to insert, exit early.
98 if (Str.empty()) return;
100 unsigned RealOffset = getMappedOffset(OrigOffset, InsertAfter);
101 Buffer.insert(RealOffset, Str.begin(), Str.end());
103 // Add a delta so that future changes are offset correctly.
104 AddInsertDelta(OrigOffset, Str.size());
107 /// ReplaceText - This method replaces a range of characters in the input
108 /// buffer with a new string. This is effectively a combined "remove+insert"
110 void RewriteBuffer::ReplaceText(unsigned OrigOffset, unsigned OrigLength,
112 unsigned RealOffset = getMappedOffset(OrigOffset, true);
113 Buffer.erase(RealOffset, OrigLength);
114 Buffer.insert(RealOffset, NewStr.begin(), NewStr.end());
115 if (OrigLength != NewStr.size())
116 AddReplaceDelta(OrigOffset, NewStr.size() - OrigLength);
120 //===----------------------------------------------------------------------===//
122 //===----------------------------------------------------------------------===//
124 /// getRangeSize - Return the size in bytes of the specified range if they
125 /// are in the same file. If not, this returns -1.
126 int Rewriter::getRangeSize(const CharSourceRange &Range,
127 RewriteOptions opts) const {
128 if (!isRewritable(Range.getBegin()) ||
129 !isRewritable(Range.getEnd())) return -1;
131 FileID StartFileID, EndFileID;
132 unsigned StartOff, EndOff;
134 StartOff = getLocationOffsetAndFileID(Range.getBegin(), StartFileID);
135 EndOff = getLocationOffsetAndFileID(Range.getEnd(), EndFileID);
137 if (StartFileID != EndFileID)
140 // If edits have been made to this buffer, the delta between the range may
142 std::map<FileID, RewriteBuffer>::const_iterator I =
143 RewriteBuffers.find(StartFileID);
144 if (I != RewriteBuffers.end()) {
145 const RewriteBuffer &RB = I->second;
146 EndOff = RB.getMappedOffset(EndOff, opts.IncludeInsertsAtEndOfRange);
147 StartOff = RB.getMappedOffset(StartOff, !opts.IncludeInsertsAtBeginOfRange);
151 // Adjust the end offset to the end of the last token, instead of being the
152 // start of the last token if this is a token range.
153 if (Range.isTokenRange())
154 EndOff += Lexer::MeasureTokenLength(Range.getEnd(), *SourceMgr, *LangOpts);
156 return EndOff-StartOff;
159 int Rewriter::getRangeSize(SourceRange Range, RewriteOptions opts) const {
160 return getRangeSize(CharSourceRange::getTokenRange(Range), opts);
164 /// getRewrittenText - Return the rewritten form of the text in the specified
165 /// range. If the start or end of the range was unrewritable or if they are
166 /// in different buffers, this returns an empty string.
168 /// Note that this method is not particularly efficient.
170 std::string Rewriter::getRewrittenText(SourceRange Range) const {
171 if (!isRewritable(Range.getBegin()) ||
172 !isRewritable(Range.getEnd()))
175 FileID StartFileID, EndFileID;
176 unsigned StartOff, EndOff;
177 StartOff = getLocationOffsetAndFileID(Range.getBegin(), StartFileID);
178 EndOff = getLocationOffsetAndFileID(Range.getEnd(), EndFileID);
180 if (StartFileID != EndFileID)
181 return ""; // Start and end in different buffers.
183 // If edits have been made to this buffer, the delta between the range may
185 std::map<FileID, RewriteBuffer>::const_iterator I =
186 RewriteBuffers.find(StartFileID);
187 if (I == RewriteBuffers.end()) {
188 // If the buffer hasn't been rewritten, just return the text from the input.
189 const char *Ptr = SourceMgr->getCharacterData(Range.getBegin());
191 // Adjust the end offset to the end of the last token, instead of being the
192 // start of the last token.
193 EndOff += Lexer::MeasureTokenLength(Range.getEnd(), *SourceMgr, *LangOpts);
194 return std::string(Ptr, Ptr+EndOff-StartOff);
197 const RewriteBuffer &RB = I->second;
198 EndOff = RB.getMappedOffset(EndOff, true);
199 StartOff = RB.getMappedOffset(StartOff);
201 // Adjust the end offset to the end of the last token, instead of being the
202 // start of the last token.
203 EndOff += Lexer::MeasureTokenLength(Range.getEnd(), *SourceMgr, *LangOpts);
205 // Advance the iterators to the right spot, yay for linear time algorithms.
206 RewriteBuffer::iterator Start = RB.begin();
207 std::advance(Start, StartOff);
208 RewriteBuffer::iterator End = Start;
209 std::advance(End, EndOff-StartOff);
211 return std::string(Start, End);
214 unsigned Rewriter::getLocationOffsetAndFileID(SourceLocation Loc,
216 assert(Loc.isValid() && "Invalid location");
217 std::pair<FileID,unsigned> V = SourceMgr->getDecomposedLoc(Loc);
223 /// getEditBuffer - Get or create a RewriteBuffer for the specified FileID.
225 RewriteBuffer &Rewriter::getEditBuffer(FileID FID) {
226 std::map<FileID, RewriteBuffer>::iterator I =
227 RewriteBuffers.lower_bound(FID);
228 if (I != RewriteBuffers.end() && I->first == FID)
230 I = RewriteBuffers.insert(I, std::make_pair(FID, RewriteBuffer()));
232 StringRef MB = SourceMgr->getBufferData(FID);
233 I->second.Initialize(MB.begin(), MB.end());
238 /// InsertText - Insert the specified string at the specified location in the
240 bool Rewriter::InsertText(SourceLocation Loc, StringRef Str,
241 bool InsertAfter, bool indentNewLines) {
242 if (!isRewritable(Loc)) return true;
244 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
246 SmallString<128> indentedStr;
247 if (indentNewLines && Str.find('\n') != StringRef::npos) {
248 StringRef MB = SourceMgr->getBufferData(FID);
250 unsigned lineNo = SourceMgr->getLineNumber(FID, StartOffs) - 1;
251 const SrcMgr::ContentCache *
252 Content = SourceMgr->getSLocEntry(FID).getFile().getContentCache();
253 unsigned lineOffs = Content->SourceLineCache[lineNo];
255 // Find the whitespace at the start of the line.
256 StringRef indentSpace;
258 unsigned i = lineOffs;
259 while (isWhitespace(MB[i]))
261 indentSpace = MB.substr(lineOffs, i-lineOffs);
264 SmallVector<StringRef, 4> lines;
265 Str.split(lines, "\n");
267 for (unsigned i = 0, e = lines.size(); i != e; ++i) {
268 indentedStr += lines[i];
271 indentedStr += indentSpace;
274 Str = indentedStr.str();
277 getEditBuffer(FID).InsertText(StartOffs, Str, InsertAfter);
281 bool Rewriter::InsertTextAfterToken(SourceLocation Loc, StringRef Str) {
282 if (!isRewritable(Loc)) return true;
284 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
285 RewriteOptions rangeOpts;
286 rangeOpts.IncludeInsertsAtBeginOfRange = false;
287 StartOffs += getRangeSize(SourceRange(Loc, Loc), rangeOpts);
288 getEditBuffer(FID).InsertText(StartOffs, Str, /*InsertAfter*/true);
292 /// RemoveText - Remove the specified text region.
293 bool Rewriter::RemoveText(SourceLocation Start, unsigned Length,
294 RewriteOptions opts) {
295 if (!isRewritable(Start)) return true;
297 unsigned StartOffs = getLocationOffsetAndFileID(Start, FID);
298 getEditBuffer(FID).RemoveText(StartOffs, Length, opts.RemoveLineIfEmpty);
302 /// ReplaceText - This method replaces a range of characters in the input
303 /// buffer with a new string. This is effectively a combined "remove/insert"
305 bool Rewriter::ReplaceText(SourceLocation Start, unsigned OrigLength,
307 if (!isRewritable(Start)) return true;
309 unsigned StartOffs = getLocationOffsetAndFileID(Start, StartFileID);
311 getEditBuffer(StartFileID).ReplaceText(StartOffs, OrigLength, NewStr);
315 bool Rewriter::ReplaceText(SourceRange range, SourceRange replacementRange) {
316 if (!isRewritable(range.getBegin())) return true;
317 if (!isRewritable(range.getEnd())) return true;
318 if (replacementRange.isInvalid()) return true;
319 SourceLocation start = range.getBegin();
320 unsigned origLength = getRangeSize(range);
321 unsigned newLength = getRangeSize(replacementRange);
323 unsigned newOffs = getLocationOffsetAndFileID(replacementRange.getBegin(),
325 StringRef MB = SourceMgr->getBufferData(FID);
326 return ReplaceText(start, origLength, MB.substr(newOffs, newLength));
329 bool Rewriter::IncreaseIndentation(CharSourceRange range,
330 SourceLocation parentIndent) {
331 if (range.isInvalid()) return true;
332 if (!isRewritable(range.getBegin())) return true;
333 if (!isRewritable(range.getEnd())) return true;
334 if (!isRewritable(parentIndent)) return true;
336 FileID StartFileID, EndFileID, parentFileID;
337 unsigned StartOff, EndOff, parentOff;
339 StartOff = getLocationOffsetAndFileID(range.getBegin(), StartFileID);
340 EndOff = getLocationOffsetAndFileID(range.getEnd(), EndFileID);
341 parentOff = getLocationOffsetAndFileID(parentIndent, parentFileID);
343 if (StartFileID != EndFileID || StartFileID != parentFileID)
345 if (StartOff > EndOff)
348 FileID FID = StartFileID;
349 StringRef MB = SourceMgr->getBufferData(FID);
351 unsigned parentLineNo = SourceMgr->getLineNumber(FID, parentOff) - 1;
352 unsigned startLineNo = SourceMgr->getLineNumber(FID, StartOff) - 1;
353 unsigned endLineNo = SourceMgr->getLineNumber(FID, EndOff) - 1;
355 const SrcMgr::ContentCache *
356 Content = SourceMgr->getSLocEntry(FID).getFile().getContentCache();
358 // Find where the lines start.
359 unsigned parentLineOffs = Content->SourceLineCache[parentLineNo];
360 unsigned startLineOffs = Content->SourceLineCache[startLineNo];
362 // Find the whitespace at the start of each line.
363 StringRef parentSpace, startSpace;
365 unsigned i = parentLineOffs;
366 while (isWhitespace(MB[i]))
368 parentSpace = MB.substr(parentLineOffs, i-parentLineOffs);
371 while (isWhitespace(MB[i]))
373 startSpace = MB.substr(startLineOffs, i-startLineOffs);
375 if (parentSpace.size() >= startSpace.size())
377 if (!startSpace.startswith(parentSpace))
380 StringRef indent = startSpace.substr(parentSpace.size());
382 // Indent the lines between start/end offsets.
383 RewriteBuffer &RB = getEditBuffer(FID);
384 for (unsigned lineNo = startLineNo; lineNo <= endLineNo; ++lineNo) {
385 unsigned offs = Content->SourceLineCache[lineNo];
387 while (isWhitespace(MB[i]))
389 StringRef origIndent = MB.substr(offs, i-offs);
390 if (origIndent.startswith(startSpace))
391 RB.InsertText(offs, indent, /*InsertAfter=*/false);
398 // A wrapper for a file stream that atomically overwrites the target.
400 // Creates a file output stream for a temporary file in the constructor,
401 // which is later accessible via getStream() if ok() return true.
402 // Flushes the stream and moves the temporary file to the target location
403 // in the destructor.
404 class AtomicallyMovedFile {
406 AtomicallyMovedFile(DiagnosticsEngine &Diagnostics, StringRef Filename,
408 : Diagnostics(Diagnostics), Filename(Filename), AllWritten(AllWritten) {
409 TempFilename = Filename;
410 TempFilename += "-%%%%%%%%";
412 if (llvm::sys::fs::createUniqueFile(TempFilename.str(), FD, TempFilename)) {
414 Diagnostics.Report(clang::diag::err_unable_to_make_temp)
417 FileStream.reset(new llvm::raw_fd_ostream(FD, /*shouldClose=*/true));
421 ~AtomicallyMovedFile() {
426 // Win32 does not allow rename/removing opened files.
429 if (std::error_code ec =
430 llvm::sys::fs::rename(TempFilename.str(), Filename)) {
432 Diagnostics.Report(clang::diag::err_unable_to_rename_temp)
433 << TempFilename << Filename << ec.message();
434 // If the remove fails, there's not a lot we can do - this is already an
436 llvm::sys::fs::remove(TempFilename.str());
440 bool ok() { return (bool)FileStream; }
441 raw_ostream &getStream() { return *FileStream; }
444 DiagnosticsEngine &Diagnostics;
446 SmallString<128> TempFilename;
447 std::unique_ptr<llvm::raw_fd_ostream> FileStream;
450 } // end anonymous namespace
452 bool Rewriter::overwriteChangedFiles() {
453 bool AllWritten = true;
454 for (buffer_iterator I = buffer_begin(), E = buffer_end(); I != E; ++I) {
455 const FileEntry *Entry =
456 getSourceMgr().getFileEntryForID(I->first);
457 AtomicallyMovedFile File(getSourceMgr().getDiagnostics(), Entry->getName(),
460 I->second.write(File.getStream());