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 "llvm/ADT/RewriteBuffer.h"
21#include "llvm/ADT/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;
32using llvm::RewriteBuffer;
33
34//===----------------------------------------------------------------------===//
35// Rewriter class
36//===----------------------------------------------------------------------===//
37
38/// Return true if this character is non-new-line whitespace:
39/// ' ', '\\t', '\\f', '\\v', '\\r'.
40static inline bool isWhitespaceExceptNL(unsigned char c) {
41 return c == ' ' || c == '\t' || c == '\f' || c == '\v' || c == '\r';
42}
43
44/// getRangeSize - Return the size in bytes of the specified range if they
45/// are in the same file. If not, this returns -1.
46int Rewriter::getRangeSize(const CharSourceRange &Range,
47 RewriteOptions opts) const {
48 if (!isRewritable(Loc: Range.getBegin()) ||
49 !isRewritable(Loc: Range.getEnd())) return -1;
50
51 FileID StartFileID, EndFileID;
52 unsigned StartOff = getLocationOffsetAndFileID(Loc: Range.getBegin(), FID&: StartFileID);
53 unsigned EndOff = getLocationOffsetAndFileID(Loc: Range.getEnd(), FID&: EndFileID);
54
55 if (StartFileID != EndFileID)
56 return -1;
57
58 // If edits have been made to this buffer, the delta between the range may
59 // have changed.
60 std::map<FileID, RewriteBuffer>::const_iterator I =
61 RewriteBuffers.find(x: StartFileID);
62 if (I != RewriteBuffers.end()) {
63 const RewriteBuffer &RB = I->second;
64 EndOff = RB.getMappedOffset(OrigOffset: EndOff, AfterInserts: opts.IncludeInsertsAtEndOfRange);
65 StartOff = RB.getMappedOffset(OrigOffset: StartOff, AfterInserts: !opts.IncludeInsertsAtBeginOfRange);
66 }
67
68 // Adjust the end offset to the end of the last token, instead of being the
69 // start of the last token if this is a token range.
70 if (Range.isTokenRange())
71 EndOff += Lexer::MeasureTokenLength(Loc: Range.getEnd(), SM: *SourceMgr, LangOpts: *LangOpts);
72
73 return EndOff-StartOff;
74}
75
76int Rewriter::getRangeSize(SourceRange Range, RewriteOptions opts) const {
77 return getRangeSize(Range: CharSourceRange::getTokenRange(R: Range), opts);
78}
79
80/// getRewrittenText - Return the rewritten form of the text in the specified
81/// range. If the start or end of the range was unrewritable or if they are
82/// in different buffers, this returns an empty string.
83///
84/// Note that this method is not particularly efficient.
85std::string Rewriter::getRewrittenText(CharSourceRange Range) const {
86 if (!isRewritable(Loc: Range.getBegin()) ||
87 !isRewritable(Loc: Range.getEnd()))
88 return {};
89
90 FileID StartFileID, EndFileID;
91 unsigned StartOff, EndOff;
92 StartOff = getLocationOffsetAndFileID(Loc: Range.getBegin(), FID&: StartFileID);
93 EndOff = getLocationOffsetAndFileID(Loc: Range.getEnd(), FID&: EndFileID);
94
95 if (StartFileID != EndFileID)
96 return {}; // Start and end in different buffers.
97
98 // If edits have been made to this buffer, the delta between the range may
99 // have changed.
100 std::map<FileID, RewriteBuffer>::const_iterator I =
101 RewriteBuffers.find(x: StartFileID);
102 if (I == RewriteBuffers.end()) {
103 // If the buffer hasn't been rewritten, just return the text from the input.
104 const char *Ptr = SourceMgr->getCharacterData(SL: Range.getBegin());
105
106 // Adjust the end offset to the end of the last token, instead of being the
107 // start of the last token.
108 if (Range.isTokenRange())
109 EndOff +=
110 Lexer::MeasureTokenLength(Loc: Range.getEnd(), SM: *SourceMgr, LangOpts: *LangOpts);
111 return std::string(Ptr, Ptr+EndOff-StartOff);
112 }
113
114 const RewriteBuffer &RB = I->second;
115 EndOff = RB.getMappedOffset(OrigOffset: EndOff, AfterInserts: true);
116 StartOff = RB.getMappedOffset(OrigOffset: StartOff);
117
118 // Adjust the end offset to the end of the last token, instead of being the
119 // start of the last token.
120 if (Range.isTokenRange())
121 EndOff += Lexer::MeasureTokenLength(Loc: Range.getEnd(), SM: *SourceMgr, LangOpts: *LangOpts);
122
123 // Advance the iterators to the right spot, yay for linear time algorithms.
124 RewriteBuffer::iterator Start = RB.begin();
125 std::advance(i&: Start, n: StartOff);
126 RewriteBuffer::iterator End = Start;
127 assert(EndOff >= StartOff && "Invalid iteration distance");
128 std::advance(i&: End, n: EndOff-StartOff);
129
130 return std::string(Start, End);
131}
132
133unsigned Rewriter::getLocationOffsetAndFileID(SourceLocation Loc,
134 FileID &FID) const {
135 assert(Loc.isValid() && "Invalid location");
136 FileIDAndOffset V = SourceMgr->getDecomposedLoc(Loc);
137 FID = V.first;
138 return V.second;
139}
140
141/// getEditBuffer - Get or create a RewriteBuffer for the specified FileID.
142RewriteBuffer &Rewriter::getEditBuffer(FileID FID) {
143 std::map<FileID, RewriteBuffer>::iterator I =
144 RewriteBuffers.lower_bound(x: FID);
145 if (I != RewriteBuffers.end() && I->first == FID)
146 return I->second;
147 I = RewriteBuffers.insert(position: I, x: std::make_pair(x&: FID, y: RewriteBuffer()));
148
149 StringRef MB = SourceMgr->getBufferData(FID);
150 I->second.Initialize(BufStart: MB.begin(), BufEnd: MB.end());
151
152 return I->second;
153}
154
155/// InsertText - Insert the specified string at the specified location in the
156/// original buffer.
157bool Rewriter::InsertText(SourceLocation Loc, StringRef Str,
158 bool InsertAfter, bool indentNewLines) {
159 if (!isRewritable(Loc)) return true;
160 FileID FID;
161 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
162
163 SmallString<128> indentedStr;
164 if (indentNewLines && Str.contains(C: '\n')) {
165 StringRef MB = SourceMgr->getBufferData(FID);
166
167 unsigned lineNo = SourceMgr->getLineNumber(FID, FilePos: StartOffs) - 1;
168 const SrcMgr::ContentCache *Content =
169 &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
170 unsigned lineOffs = Content->SourceLineCache[lineNo];
171
172 // Find the whitespace at the start of the line.
173 StringRef indentSpace;
174 {
175 unsigned i = lineOffs;
176 while (isWhitespaceExceptNL(c: MB[i]))
177 ++i;
178 indentSpace = MB.substr(Start: lineOffs, N: i-lineOffs);
179 }
180
181 SmallVector<StringRef, 4> lines;
182 Str.split(A&: lines, Separator: "\n");
183
184 for (unsigned i = 0, e = lines.size(); i != e; ++i) {
185 indentedStr += lines[i];
186 if (i < e-1) {
187 indentedStr += '\n';
188 indentedStr += indentSpace;
189 }
190 }
191 Str = indentedStr.str();
192 }
193
194 getEditBuffer(FID).InsertText(OrigOffset: StartOffs, Str, InsertAfter);
195 return false;
196}
197
198bool Rewriter::InsertTextAfterToken(SourceLocation Loc, StringRef Str) {
199 if (!isRewritable(Loc)) return true;
200 FileID FID;
201 unsigned StartOffs = getLocationOffsetAndFileID(Loc, FID);
202 RewriteOptions rangeOpts;
203 rangeOpts.IncludeInsertsAtBeginOfRange = false;
204 StartOffs += getRangeSize(Range: SourceRange(Loc, Loc), opts: rangeOpts);
205 getEditBuffer(FID).InsertText(OrigOffset: StartOffs, Str, /*InsertAfter*/true);
206 return false;
207}
208
209/// RemoveText - Remove the specified text region.
210bool Rewriter::RemoveText(SourceLocation Start, unsigned Length,
211 RewriteOptions opts) {
212 if (!isRewritable(Loc: Start)) return true;
213 FileID FID;
214 unsigned StartOffs = getLocationOffsetAndFileID(Loc: Start, FID);
215 getEditBuffer(FID).RemoveText(OrigOffset: StartOffs, Size: Length, removeLineIfEmpty: opts.RemoveLineIfEmpty);
216 return false;
217}
218
219/// ReplaceText - This method replaces a range of characters in the input
220/// buffer with a new string. This is effectively a combined "remove/insert"
221/// operation.
222bool Rewriter::ReplaceText(SourceLocation Start, unsigned OrigLength,
223 StringRef NewStr) {
224 if (!isRewritable(Loc: Start)) return true;
225 FileID StartFileID;
226 unsigned StartOffs = getLocationOffsetAndFileID(Loc: Start, FID&: StartFileID);
227
228 getEditBuffer(FID: StartFileID).ReplaceText(OrigOffset: StartOffs, OrigLength, NewStr);
229 return false;
230}
231
232bool Rewriter::ReplaceText(SourceRange range, SourceRange replacementRange) {
233 if (!isRewritable(Loc: range.getBegin())) return true;
234 if (!isRewritable(Loc: range.getEnd())) return true;
235 if (replacementRange.isInvalid()) return true;
236 SourceLocation start = range.getBegin();
237 unsigned origLength = getRangeSize(Range: range);
238 unsigned newLength = getRangeSize(Range: replacementRange);
239 FileID FID;
240 unsigned newOffs = getLocationOffsetAndFileID(Loc: replacementRange.getBegin(),
241 FID);
242 StringRef MB = SourceMgr->getBufferData(FID);
243 return ReplaceText(Start: start, OrigLength: origLength, NewStr: MB.substr(Start: newOffs, N: newLength));
244}
245
246bool Rewriter::IncreaseIndentation(CharSourceRange range,
247 SourceLocation parentIndent) {
248 if (range.isInvalid()) return true;
249 if (!isRewritable(Loc: range.getBegin())) return true;
250 if (!isRewritable(Loc: range.getEnd())) return true;
251 if (!isRewritable(Loc: parentIndent)) return true;
252
253 FileID StartFileID, EndFileID, parentFileID;
254 unsigned StartOff, EndOff, parentOff;
255
256 StartOff = getLocationOffsetAndFileID(Loc: range.getBegin(), FID&: StartFileID);
257 EndOff = getLocationOffsetAndFileID(Loc: range.getEnd(), FID&: EndFileID);
258 parentOff = getLocationOffsetAndFileID(Loc: parentIndent, FID&: parentFileID);
259
260 if (StartFileID != EndFileID || StartFileID != parentFileID)
261 return true;
262 if (StartOff > EndOff)
263 return true;
264
265 FileID FID = StartFileID;
266 StringRef MB = SourceMgr->getBufferData(FID);
267
268 unsigned parentLineNo = SourceMgr->getLineNumber(FID, FilePos: parentOff) - 1;
269 unsigned startLineNo = SourceMgr->getLineNumber(FID, FilePos: StartOff) - 1;
270 unsigned endLineNo = SourceMgr->getLineNumber(FID, FilePos: EndOff) - 1;
271
272 const SrcMgr::ContentCache *Content =
273 &SourceMgr->getSLocEntry(FID).getFile().getContentCache();
274
275 // Find where the lines start.
276 unsigned parentLineOffs = Content->SourceLineCache[parentLineNo];
277 unsigned startLineOffs = Content->SourceLineCache[startLineNo];
278
279 // Find the whitespace at the start of each line.
280 StringRef parentSpace, startSpace;
281 {
282 unsigned i = parentLineOffs;
283 while (isWhitespaceExceptNL(c: MB[i]))
284 ++i;
285 parentSpace = MB.substr(Start: parentLineOffs, N: i-parentLineOffs);
286
287 i = startLineOffs;
288 while (isWhitespaceExceptNL(c: MB[i]))
289 ++i;
290 startSpace = MB.substr(Start: startLineOffs, N: i-startLineOffs);
291 }
292 if (parentSpace.size() >= startSpace.size())
293 return true;
294 if (!startSpace.starts_with(Prefix: parentSpace))
295 return true;
296
297 StringRef indent = startSpace.substr(Start: parentSpace.size());
298
299 // Indent the lines between start/end offsets.
300 RewriteBuffer &RB = getEditBuffer(FID);
301 for (unsigned lineNo = startLineNo; lineNo <= endLineNo; ++lineNo) {
302 unsigned offs = Content->SourceLineCache[lineNo];
303 unsigned i = offs;
304 while (isWhitespaceExceptNL(c: MB[i]))
305 ++i;
306 StringRef origIndent = MB.substr(Start: offs, N: i-offs);
307 if (origIndent.starts_with(Prefix: startSpace))
308 RB.InsertText(OrigOffset: offs, Str: indent, /*InsertAfter=*/false);
309 }
310
311 return false;
312}
313
314bool Rewriter::overwriteChangedFiles() {
315 bool AllWritten = true;
316 auto& Diag = getSourceMgr().getDiagnostics();
317 unsigned OverwriteFailure = Diag.getCustomDiagID(
318 L: DiagnosticsEngine::Error, FormatString: "unable to overwrite file %0: %1");
319 for (buffer_iterator I = buffer_begin(), E = buffer_end(); I != E; ++I) {
320 OptionalFileEntryRef Entry = getSourceMgr().getFileEntryRefForID(FID: I->first);
321 llvm::SmallString<128> Path(Entry->getName());
322 getSourceMgr().getFileManager().makeAbsolutePath(Path);
323 if (auto Error = llvm::writeToOutput(OutputFileName: Path, Write: [&](llvm::raw_ostream &OS) {
324 I->second.write(Stream&: OS);
325 return llvm::Error::success();
326 })) {
327 Diag.Report(DiagID: OverwriteFailure)
328 << Entry->getName() << llvm::toString(E: std::move(Error));
329 AllWritten = false;
330 }
331 }
332 return !AllWritten;
333}
334