clang: lib/Tooling/Transformer/SourceCode.cpp Source File (original) (raw)
1
2
3
4
5
6
7
8
9
10
11
22#include "llvm/Support/Errc.h"
23#include "llvm/Support/Error.h"
24#include
25
26using namespace clang;
27
28using llvm::errc;
29using llvm::StringError;
30
34 Context.getLangOpts());
35}
36
41 Context.getLangOpts());
45 bool Err =
47 Context.getLangOpts(), true);
51}
52
55 bool AllowSystemHeaders) {
56 if (Range.isInvalid())
57 return llvm::make_error(errc::invalid_argument,
58 "Invalid range");
59
60 if (Range.getBegin().isMacroID() || Range.getEnd().isMacroID())
61 return llvm::make_error(
62 errc::invalid_argument, "Range starts or ends in a macro expansion");
63
64 if (!AllowSystemHeaders) {
65 if (SM.isInSystemHeader(Range.getBegin()) ||
66 SM.isInSystemHeader(Range.getEnd()))
67 return llvm::make_error(errc::invalid_argument,
68 "Range is in system header");
69 }
70
73 if (BeginInfo.first != EndInfo.first)
74 return llvm::make_error(
75 errc::invalid_argument, "Range begins and ends in different files");
76
77 if (BeginInfo.second > EndInfo.second)
78 return llvm::make_error(errc::invalid_argument,
79 "Range's begin is past its end");
80
81 return llvm::Error::success();
82}
83
88
89
90
91
94 assert(Loc.isMacroID() && "Location must be in a macro");
97 const auto &Expansion = SM.getSLocEntry(SM.getFileID(Loc)).getExpansion();
98 if (Expansion.isMacroArgExpansion()) {
99
100
101 Loc = Expansion.getSpellingLoc();
102 ArgLocs.push_back(Expansion.getExpansionLocStart());
103 } else {
104 return {};
105 }
106 }
107 return ArgLocs;
108}
109
112 if (Range.getBegin().isMacroID() && Range.getEnd().isMacroID()) {
113
114
117 return B.empty() || B != E;
118 }
119
120 return Range.getBegin().isMacroID() || Range.getEnd().isMacroID();
121}
122
123
124
125
126static std::optional
131 auto &SLoc = SM.getSLocEntry(SM.getFileID(Loc), &Invalid);
133 return std::nullopt;
134 if (auto &Expansion = SLoc.getExpansion();
135 !Expansion.isExpansionTokenRange()) {
136
137
138
139 return Expansion.getExpansionLocRange();
140 }
141 }
142 return std::nullopt;
143}
144
145
146
150 if (Range.isTokenRange()) {
153 if (EndToken) {
155 BeginToken ? BeginToken->getBegin() : Range.getBegin();
156
157
158
160 } else if (BeginToken) {
161
162
163
165 Range.getEnd());
166 }
167 }
168 return Range;
169}
170
174 bool IncludeMacroExpansion) {
176 if (IncludeMacroExpansion) {
178 } else {
181 return {};
182
183 auto B = SM.getSpellingLoc(AdjustedRange.getBegin());
184 auto E = SM.getSpellingLoc(AdjustedRange.getEnd());
185 if (AdjustedRange.isTokenRange())
188 }
189 return Range;
190}
191
194 const LangOptions &LangOpts, bool IncludeMacroExpansion) {
196 getRange(EditRange, SM, LangOpts, IncludeMacroExpansion);
198 if (IsInvalid)
199 return std::nullopt;
201}
202
205 const LangOptions &LangOpts, bool IncludeMacroExpansion) {
207 getRange(EditRange, SM, LangOpts, IncludeMacroExpansion);
208 bool IsInvalid =
210 if (IsInvalid)
211 return std::nullopt;
213}
214
218
219static bool contains(const std::settok::TokenKind &Terminators,
221 return Terminators.count(Tok.getKind()) > 0;
222}
223
224
225
226
227
228
229
230
231
234 const std::settok::TokenKind &Terminators,
236 assert(EntityLast.isValid() && "Invalid end location found.");
237
238
239
240
241
242
243 CharSourceRange ExpansionRange = SM.getExpansionRange(EntityLast);
244
245
246 std::unique_ptr Lexer = [&]() {
248 auto FileOffset = SM.getDecomposedLoc(ExpansionRange.getEnd());
249 llvm::StringRef File = SM.getBufferData(FileOffset.first, &Invalid);
250 assert( && "Cannot get file/offset");
251 return std::make_uniqueclang::Lexer(
252 SM.getLocForStartOfFile(FileOffset.first), LangOpts, File.begin(),
253 File.data() + FileOffset.second, File.end());
254 }();
255
256
258
259
260
261
262
263
264
266 bool Terminated = false;
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288 bool TerminatedByMacro = false;
290 if (Terminators.empty() || contains(Terminators, Tok))
291 Terminated = true;
292 else if (EntityLast.isMacroID()) {
293 Terminated = true;
294 TerminatedByMacro = true;
295 }
296
297
299
300 while (!Terminated) {
301
303
304 switch (Tok.getKind()) {
305 case tok::eof:
306
307 case tok::l_brace:
308 case tok::r_brace:
309 case tok::comma:
310 return End;
311
312 case tok::unknown:
314
315 End = Tok.getEndLoc();
316 break;
317 default:
319 Terminated = true;
320 End = Tok.getEndLoc();
321 break;
322 }
323 }
324
325 do {
326
328
329 switch (Tok.getKind()) {
330 case tok::unknown:
332
333 return Tok.getEndLoc();
334 break;
335 case tok::comment:
336
337 End = Tok.getEndLoc();
338 break;
339 case tok:🚛
340 case tok::comma:
341 if (TerminatedByMacro && contains(Terminators, Tok)) {
342 End = Tok.getEndLoc();
343
344 TerminatedByMacro = false;
345 break;
346 }
347
348 return End;
349 default:
350
351 return End;
352 }
353 } while (true);
354}
355
356
357
358
359
360
361
363 if (llvm::isa(D) || llvm::isa(D))
364 return {tok::semi};
365
366 if (llvm::isa(D) || llvm::isa(D))
367 return {tok::r_brace, tok::semi};
368
369 if (llvm::isa(D) || llvm::isa(D))
370 return {tok::comma, tok::semi};
371
372 return {};
373}
374
375
376
377
381 const char *LocChars = SM.getCharacterData(Loc);
382 int i = 0;
384 ++i;
386 ++i;
388}
389
390
391
392
393
396
397
398
399
401 const char *LocChars =
404 "Loc must be a valid character and not the first of the source file.");
406 for (int i = 1; isWhitespace(LocChars[i]); ++i)
408 return true;
409 }
410
411
414 true);
415 if (Failed)
416
417 return true;
418
419 switch (Tok.getKind()) {
420 case tok::comment:
421 case tok::l_brace:
422 case tok::r_brace:
423 case tok::eof:
424 return true;
425 default:
426 return false;
427 }
428}
429
433 const LangOptions &LangOpts = Context.getLangOpts();
435
436
437 if (const auto *Record = llvm::dyn_cast(&Decl)) {
438 if (const auto *T = Record->getDescribedClassTemplate())
439 if (SM.isBeforeInTranslationUnit(T->getBeginLoc(), Range.getBegin()))
440 Range.setBegin(T->getBeginLoc());
441 } else if (const auto *F = llvm::dyn_cast(&Decl)) {
442 if (const auto *T = F->getDescribedFunctionTemplate())
443 if (SM.isBeforeInTranslationUnit(T->getBeginLoc(), Range.getBegin()))
444 Range.setBegin(T->getBeginLoc());
445 }
446
447
448
451 Range.setTokenRange(false);
452
453
454
455
458
459
460
461
462
463
464 if (SM.isBeforeInTranslationUnit(Comment->getBeginLoc(),
465 Range.getBegin()) &&
468 LangOpts) &&
470 const StringRef CommentText = Comment->getRawText(SM);
471 if (!CommentText.contains("LINT.IfChange") &&
472 !CommentText.contains("LINT.ThenChange"))
473 Range.setBegin(Comment->getBeginLoc());
474 }
475
479 continue;
481
482
483
485 StringRef Source =
488 continue;
489 llvm::StringRef BeforeAttr =
490 Source.substr(0, SM.getFileOffset(Range.getBegin()));
491 llvm::StringRef BeforeAttrStripped = BeforeAttr.rtrim();
492
493 for (llvm::StringRef Prefix : {"[[", "__attribute__(("}) {
494
495 if (BeforeAttrStripped.ends_with(Prefix)) {
496
497
498
499 Range.setBegin(Range.getBegin().getLocWithOffset(static_cast<int>(
500 -BeforeAttr.size() + BeforeAttrStripped.size() - Prefix.size())));
501 break;
502
503
504
505 }
506 }
507 }
508
509
510
512}
Defines the clang::ASTContext interface.
Defines the C++ Decl subclasses, other than those for templates (found in DeclTemplate....
Defines the C++ template declaration subclasses.
FormatToken * Next
The next token in the unwrapped line.
llvm::MachO::Record Record
static llvm::SmallVector< SourceLocation, 2 > getMacroArgumentExpansionLocs(SourceLocation Loc, const SourceManager &SM)
Definition SourceCode.cpp:93
static bool contains(const std::set< tok::TokenKind > &Terminators, const Token &Tok)
Definition SourceCode.cpp:219
static bool atOrBeforeSeparation(const SourceManager &SM, SourceLocation Loc, const LangOptions &LangOpts)
Definition SourceCode.cpp:394
static bool startsWithNewline(const SourceManager &SM, const Token &Tok)
Definition SourceCode.cpp:215
static SourceLocation getEntityEndLoc(const SourceManager &SM, SourceLocation EntityLast, const std::set< tok::TokenKind > &Terminators, const LangOptions &LangOpts)
Definition SourceCode.cpp:233
static CharSourceRange getRangeForSplitTokens(CharSourceRange Range, const SourceManager &SM, const LangOptions &LangOpts)
Definition SourceCode.cpp:147
static SourceLocation skipWhitespaceAndNewline(const SourceManager &SM, SourceLocation Loc, const LangOptions &LangOpts)
Definition SourceCode.cpp:378
static std::optional< CharSourceRange > getExpansionForSplitToken(SourceLocation Loc, const SourceManager &SM, const LangOptions &LangOpts)
Definition SourceCode.cpp:127
static CharSourceRange getRange(const CharSourceRange &EditRange, const SourceManager &SM, const LangOptions &LangOpts, bool IncludeMacroExpansion)
Definition SourceCode.cpp:171
static bool spelledInMacroDefinition(CharSourceRange Range, const SourceManager &SM)
Definition SourceCode.cpp:110
static std::set< tok::TokenKind > getTerminators(const Decl &D)
Definition SourceCode.cpp:362
Defines the SourceManager interface.
Holds long-lived AST nodes (such as types and decls) that can be referred to throughout the semantic ...
RawComment * getRawCommentForDeclNoCache(const Decl *D) const
Return the documentation comment attached to a given declaration, without looking into cache.
Attr - This represents one attribute.
SourceLocation getLocation() const
Represents a character-granular source range.
static CharSourceRange getCharRange(SourceRange R)
static CharSourceRange getTokenRange(SourceRange R)
SourceLocation getEnd() const
Decl - This represents one declaration (or definition), e.g.
SourceLocation getEndLoc() const LLVM_READONLY
ASTContext & getASTContext() const LLVM_READONLY
virtual SourceRange getSourceRange() const LLVM_READONLY
Source range that this declaration covers.
Keeps track of the various options that can be enabled, which controls the dialect of C or C++ that i...
Lexer - This provides a simple interface that turns a text buffer into a stream of tokens.
static StringRef getSourceText(CharSourceRange Range, const SourceManager &SM, const LangOptions &LangOpts, bool *Invalid=nullptr)
Returns a string for the source that the range encompasses.
void SetKeepWhitespaceMode(bool Val)
SetKeepWhitespaceMode - This method lets clients enable or disable whitespace retention mode.
bool LexFromRawLexer(Token &Result)
LexFromRawLexer - Lex a token from a designated raw lexer (one with no associated preprocessor object...
static CharSourceRange getAsCharRange(SourceRange Range, const SourceManager &SM, const LangOptions &LangOpts)
Given a token range, produce a corresponding CharSourceRange that is not a token range.
static CharSourceRange makeFileCharRange(CharSourceRange Range, const SourceManager &SM, const LangOptions &LangOpts)
Accepts a range and returns a character range with file locations.
static bool getRawToken(SourceLocation Loc, Token &Result, const SourceManager &SM, const LangOptions &LangOpts, bool IgnoreWhiteSpace=false)
Relex the token at the specified location.
static SourceLocation getLocForEndOfToken(SourceLocation Loc, unsigned Offset, const SourceManager &SM, const LangOptions &LangOpts)
Computes the source location just past the end of the token at this source location.
Encodes a location in the source.
bool isValid() const
Return true if this is a valid SourceLocation object.
SourceLocation getLocWithOffset(IntTy Offset) const
Return a source location with the specified offset from this SourceLocation.
This class handles loading and caching of source files into memory.
Token - This structure provides full information about a lexed token.
TokenKind
Provides a simple uniform namespace for tokens from all C languages.
The JSON file list parser is used to communicate input to InstallAPI.
LLVM_READONLY bool isVerticalWhitespace(unsigned char c)
Returns true if this character is vertical ASCII whitespace: '\n', '\r'.
std::pair< FileID, unsigned > FileIDAndOffset
LLVM_READONLY bool isHorizontalWhitespace(unsigned char c)
Returns true if this character is horizontal ASCII whitespace: ' ', '\t', '\f', '\v'.
const FunctionProtoType * T
LLVM_READONLY bool isWhitespace(unsigned char c)
Return true if this character is horizontal or vertical ASCII whitespace: ' ', '\t',...