clang/lib/Rewrite/Rewriter.cpp

   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
  37 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'.
  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
  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
 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.
 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.
 145 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
 175 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.
 184 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   assert(EndOff >= StartOff && "Invalid iteration distance");
 227   std::advance(End, EndOff-StartOff);
 228
 229   return std::string(Start, End);
 230 }
 231
 232 unsigned Rewriter::getLocationOffsetAndFileID(SourceLocation Loc,
 233                                               FileID &FID) const {
 234   assert(Loc.isValid() && "Invalid location");
 235   std::pair<FileID, unsigned> V = SourceMgr->getDecomposedLoc(Loc);
 236   FID = V.first;
 237   return V.second;
 238 }
 239
 240 /// getEditBuffer - Get or create a RewriteBuffer for the specified FileID.
 241 RewriteBuffer &Rewriter::getEditBuffer(FileID FID) {
 242   std::map<FileID, RewriteBuffer>::iterator I =
 243     RewriteBuffers.lower_bound(FID);
 244   if (I != RewriteBuffers.end() && I->first == FID)
 245     return I->second;
 246   I = RewriteBuffers.insert(I, std::make_pair(FID, RewriteBuffer()));
 247
 248   StringRef MB = SourceMgr->getBufferData(FID);
 249   I->second.Initialize(MB.begin(), MB.end());
 250
 251   return I->second;
 252 }
 253
 254 /// InsertText - Insert the specified string at the specified location in the
 255 /// original buffer.
 256 bool Rewriter::InsertText(SourceLocation Loc, StringRef Str,
 257                           bool InsertAfter, bool indentNewLines) {
 258   if (!isRewritable(Loc)) return true;
 259   FileID FID;
 260   unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
 261
 262   SmallString<128> indentedStr;
 263   if (indentNewLines && Str.contains('\n')) {
 264     StringRef MB = SourceMgr->getBufferData(FID);
 265
 266     unsigned lineNo = SourceMgr->getLineNumber(FID, StartOffs) - 1;
 267     const SrcMgr::ContentCache *Content =
 268         &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
 269     unsigned lineOffs = Content->SourceLineCache[lineNo];
 270
 271     // Find the whitespace at the start of the line.
 272     StringRef indentSpace;
 273     {
 274       unsigned i = lineOffs;
 275       while (isWhitespaceExceptNL(MB[i]))
 276         ++i;
 277       indentSpace = MB.substr(lineOffs, i-lineOffs);
 278     }
 279
 280     SmallVector<StringRef, 4> lines;
 281     Str.split(lines, "\n");
 282
 283     for (unsigned i = 0, e = lines.size(); i != e; ++i) {
 284       indentedStr += lines[i];
 285       if (i < e-1) {
 286         indentedStr += '\n';
 287         indentedStr += indentSpace;
 288       }
 289     }
 290     Str = indentedStr.str();
 291   }
 292
 293   getEditBuffer(FID).InsertText(StartOffs, Str, InsertAfter);
 294   return false;
 295 }
 296
 297 bool Rewriter::InsertTextAfterToken(SourceLocation Loc, StringRef Str) {
 298   if (!isRewritable(Loc)) return true;
 299   FileID FID;
 300   unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
 301   RewriteOptions rangeOpts;
 302   rangeOpts.IncludeInsertsAtBeginOfRange = false;
 303   StartOffs += getRangeSize(SourceRange(Loc, Loc), rangeOpts);
 304   getEditBuffer(FID).InsertText(StartOffs, Str, /*InsertAfter*/true);
 305   return false;
 306 }
 307
 308 /// RemoveText - Remove the specified text region.
 309 bool Rewriter::RemoveText(SourceLocation Start, unsigned Length,
 310                           RewriteOptions opts) {
 311   if (!isRewritable(Start)) return true;
 312   FileID FID;
 313   unsigned StartOffs = getLocationOffsetAndFileID(Start, FID);
 314   getEditBuffer(FID).RemoveText(StartOffs, Length, opts.RemoveLineIfEmpty);
 315   return false;
 316 }
 317
 318 /// ReplaceText - This method replaces a range of characters in the input
 319 /// buffer with a new string.  This is effectively a combined "remove/insert"
 320 /// operation.
 321 bool Rewriter::ReplaceText(SourceLocation Start, unsigned OrigLength,
 322                            StringRef NewStr) {
 323   if (!isRewritable(Start)) return true;
 324   FileID StartFileID;
 325   unsigned StartOffs = getLocationOffsetAndFileID(Start, StartFileID);
 326
 327   getEditBuffer(StartFileID).ReplaceText(StartOffs, OrigLength, NewStr);
 328   return false;
 329 }
 330
 331 bool Rewriter::ReplaceText(SourceRange range, SourceRange replacementRange) {
 332   if (!isRewritable(range.getBegin())) return true;
 333   if (!isRewritable(range.getEnd())) return true;
 334   if (replacementRange.isInvalid()) return true;
 335   SourceLocation start = range.getBegin();
 336   unsigned origLength = getRangeSize(range);
 337   unsigned newLength = getRangeSize(replacementRange);
 338   FileID FID;
 339   unsigned newOffs = getLocationOffsetAndFileID(replacementRange.getBegin(),
 340                                                 FID);
 341   StringRef MB = SourceMgr->getBufferData(FID);
 342   return ReplaceText(start, origLength, MB.substr(newOffs, newLength));
 343 }
 344
 345 bool Rewriter::IncreaseIndentation(CharSourceRange range,
 346                                    SourceLocation parentIndent) {
 347   if (range.isInvalid()) return true;
 348   if (!isRewritable(range.getBegin())) return true;
 349   if (!isRewritable(range.getEnd())) return true;
 350   if (!isRewritable(parentIndent)) return true;
 351
 352   FileID StartFileID, EndFileID, parentFileID;
 353   unsigned StartOff, EndOff, parentOff;
 354
 355   StartOff = getLocationOffsetAndFileID(range.getBegin(), StartFileID);
 356   EndOff   = getLocationOffsetAndFileID(range.getEnd(), EndFileID);
 357   parentOff = getLocationOffsetAndFileID(parentIndent, parentFileID);
 358
 359   if (StartFileID != EndFileID || StartFileID != parentFileID)
 360     return true;
 361   if (StartOff > EndOff)
 362     return true;
 363
 364   FileID FID = StartFileID;
 365   StringRef MB = SourceMgr->getBufferData(FID);
 366
 367   unsigned parentLineNo = SourceMgr->getLineNumber(FID, parentOff) - 1;
 368   unsigned startLineNo = SourceMgr->getLineNumber(FID, StartOff) - 1;
 369   unsigned endLineNo = SourceMgr->getLineNumber(FID, EndOff) - 1;
 370
 371   const SrcMgr::ContentCache *Content =
 372       &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
 373
 374   // Find where the lines start.
 375   unsigned parentLineOffs = Content->SourceLineCache[parentLineNo];
 376   unsigned startLineOffs = Content->SourceLineCache[startLineNo];
 377
 378   // Find the whitespace at the start of each line.
 379   StringRef parentSpace, startSpace;
 380   {
 381     unsigned i = parentLineOffs;
 382     while (isWhitespaceExceptNL(MB[i]))
 383       ++i;
 384     parentSpace = MB.substr(parentLineOffs, i-parentLineOffs);
 385
 386     i = startLineOffs;
 387     while (isWhitespaceExceptNL(MB[i]))
 388       ++i;
 389     startSpace = MB.substr(startLineOffs, i-startLineOffs);
 390   }
 391   if (parentSpace.size() >= startSpace.size())
 392     return true;
 393   if (!startSpace.startswith(parentSpace))
 394     return true;
 395
 396   StringRef indent = startSpace.substr(parentSpace.size());
 397
 398   // Indent the lines between start/end offsets.
 399   RewriteBuffer &RB = getEditBuffer(FID);
 400   for (unsigned lineNo = startLineNo; lineNo <= endLineNo; ++lineNo) {
 401     unsigned offs = Content->SourceLineCache[lineNo];
 402     unsigned i = offs;
 403     while (isWhitespaceExceptNL(MB[i]))
 404       ++i;
 405     StringRef origIndent = MB.substr(offs, i-offs);
 406     if (origIndent.startswith(startSpace))
 407       RB.InsertText(offs, indent, /*InsertAfter=*/false);
 408   }
 409
 410   return false;
 411 }
 412
 413 namespace {
 414
 415 // A wrapper for a file stream that atomically overwrites the target.
 416 //
 417 // Creates a file output stream for a temporary file in the constructor,
 418 // which is later accessible via getStream() if ok() return true.
 419 // Flushes the stream and moves the temporary file to the target location
 420 // in the destructor.
 421 class AtomicallyMovedFile {
 422 public:
 423   AtomicallyMovedFile(DiagnosticsEngine &Diagnostics, StringRef Filename,
 424                       bool &AllWritten)
 425       : Diagnostics(Diagnostics), Filename(Filename), AllWritten(AllWritten) {
 426     TempFilename = Filename;
 427     TempFilename += "-%%%%%%%%";
 428     int FD;
 429     if (llvm::sys::fs::createUniqueFile(TempFilename, FD, TempFilename)) {
 430       AllWritten = false;
 431       Diagnostics.Report(clang::diag::err_unable_to_make_temp)
 432         << TempFilename;
 433     } else {
 434       FileStream.reset(new llvm::raw_fd_ostream(FD, /*shouldClose=*/true));
 435     }
 436   }
 437
 438   ~AtomicallyMovedFile() {
 439     if (!ok()) return;
 440
 441     // Close (will also flush) theFileStream.
 442     FileStream->close();
 443     if (std::error_code ec = llvm::sys::fs::rename(TempFilename, Filename)) {
 444       AllWritten = false;
 445       Diagnostics.Report(clang::diag::err_unable_to_rename_temp)
 446         << TempFilename << Filename << ec.message();
 447       // If the remove fails, there's not a lot we can do - this is already an
 448       // error.
 449       llvm::sys::fs::remove(TempFilename);
 450     }
 451   }
 452
 453   bool ok() { return (bool)FileStream; }
 454   raw_ostream &getStream() { return *FileStream; }
 455
 456 private:
 457   DiagnosticsEngine &Diagnostics;
 458   StringRef Filename;
 459   SmallString<128> TempFilename;
 460   std::unique_ptr<llvm::raw_fd_ostream> FileStream;
 461   bool &AllWritten;
 462 };
 463
 464 } // namespace
 465
 466 bool Rewriter::overwriteChangedFiles() {
 467   bool AllWritten = true;
 468   for (buffer_iterator I = buffer_begin(), E = buffer_end(); I != E; ++I) {
 469     const FileEntry *Entry =
 470         getSourceMgr().getFileEntryForID(I->first);
 471     AtomicallyMovedFile File(getSourceMgr().getDiagnostics(), Entry->getName(),
 472                              AllWritten);
 473     if (File.ok()) {
 474       I->second.write(File.getStream());
 475     }
 476   }
 477   return !AllWritten;
 478 }