1//===--- PPLexerChange.cpp - Handle changing lexers in the preprocessor ---===//
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 implements pieces of the Preprocessor interface that manage the
10// current lexer stack.
11//
12//===----------------------------------------------------------------------===//
13
14#include "clang/Basic/FileManager.h"
15#include "clang/Basic/SourceManager.h"
16#include "clang/Lex/HeaderSearch.h"
17#include "clang/Lex/LexDiagnostic.h"
18#include "clang/Lex/MacroInfo.h"
19#include "clang/Lex/Preprocessor.h"
20#include "clang/Lex/PreprocessorOptions.h"
21#include "llvm/ADT/StringSwitch.h"
22#include "llvm/Support/MemoryBufferRef.h"
23#include "llvm/Support/Path.h"
24#include <optional>
25
26using namespace clang;
27
28//===----------------------------------------------------------------------===//
29// Miscellaneous Methods.
30//===----------------------------------------------------------------------===//
31
32/// isInPrimaryFile - Return true if we're in the top-level file, not in a
33/// \#include. This looks through macro expansions and active _Pragma lexers.
34bool Preprocessor::isInPrimaryFile() const {
35 if (IsFileLexer())
36 return IncludeMacroStack.empty();
37
38 // If there are any stacked lexers, we're in a #include.
39 assert(IsFileLexer(IncludeMacroStack[0]) &&
40 "Top level include stack isn't our primary lexer?");
41 return llvm::none_of(
42 Range: llvm::drop_begin(RangeOrContainer: IncludeMacroStack),
43 P: [&](const IncludeStackInfo &ISI) -> bool { return IsFileLexer(I: ISI); });
44}
45
46/// getCurrentLexer - Return the current file lexer being lexed from. Note
47/// that this ignores any potentially active macro expansions and _Pragma
48/// expansions going on at the time.
49PreprocessorLexer *Preprocessor::getCurrentFileLexer() const {
50 if (IsFileLexer())
51 return CurPPLexer;
52
53 // Look for a stacked lexer.
54 for (const IncludeStackInfo &ISI : llvm::reverse(C: IncludeMacroStack)) {
55 if (IsFileLexer(I: ISI))
56 return ISI.ThePPLexer;
57 }
58 return nullptr;
59}
60
61
62//===----------------------------------------------------------------------===//
63// Methods for Entering and Callbacks for leaving various contexts
64//===----------------------------------------------------------------------===//
65
66/// EnterSourceFile - Add a source file to the top of the include stack and
67/// start lexing tokens from it instead of the current buffer.
68bool Preprocessor::EnterSourceFile(FileID FID, ConstSearchDirIterator CurDir,
69 SourceLocation Loc,
70 bool IsFirstIncludeOfFile) {
71 assert(!CurTokenLexer && "Cannot #include a file inside a macro!");
72 ++NumEnteredSourceFiles;
73
74 if (MaxIncludeStackDepth < IncludeMacroStack.size())
75 MaxIncludeStackDepth = IncludeMacroStack.size();
76
77 // Get the MemoryBuffer for this FID, if it fails, we fail.
78 std::optional<llvm::MemoryBufferRef> InputFile =
79 getSourceManager().getBufferOrNone(FID, Loc);
80 if (!InputFile) {
81 SourceLocation FileStart = SourceMgr.getLocForStartOfFile(FID);
82 Diag(Loc, DiagID: diag::err_pp_error_opening_file)
83 << std::string(SourceMgr.getBufferName(Loc: FileStart)) << "";
84 return true;
85 }
86
87 if (isCodeCompletionEnabled() &&
88 SourceMgr.getFileEntryForID(FID) == CodeCompletionFile) {
89 CodeCompletionFileLoc = SourceMgr.getLocForStartOfFile(FID);
90 CodeCompletionLoc =
91 CodeCompletionFileLoc.getLocWithOffset(Offset: CodeCompletionOffset);
92 }
93
94 Lexer *TheLexer = new Lexer(FID, *InputFile, *this, IsFirstIncludeOfFile);
95 if (GetDependencyDirectives && FID != PredefinesFileID)
96 if (OptionalFileEntryRef File = SourceMgr.getFileEntryRefForID(FID))
97 if (auto MaybeDepDirectives = (*GetDependencyDirectives)(*File))
98 TheLexer->DepDirectives = *MaybeDepDirectives;
99
100 EnterSourceFileWithLexer(TheLexer, Dir: CurDir);
101 return false;
102}
103
104/// EnterSourceFileWithLexer - Add a source file to the top of the include stack
105/// and start lexing tokens from it instead of the current buffer.
106void Preprocessor::EnterSourceFileWithLexer(Lexer *TheLexer,
107 ConstSearchDirIterator CurDir) {
108 PreprocessorLexer *PrevPPLexer = CurPPLexer;
109
110 // Add the current lexer to the include stack.
111 if (CurPPLexer || CurTokenLexer)
112 PushIncludeMacroStack();
113
114 CurLexer.reset(p: TheLexer);
115 CurPPLexer = TheLexer;
116 CurDirLookup = CurDir;
117 CurLexerSubmodule = nullptr;
118 if (CurLexerCallback != CLK_LexAfterModuleImport)
119 CurLexerCallback = TheLexer->isDependencyDirectivesLexer()
120 ? CLK_DependencyDirectivesLexer
121 : CLK_Lexer;
122
123 // Notify the client, if desired, that we are in a new source file.
124 if (Callbacks && !CurLexer->Is_PragmaLexer) {
125 SrcMgr::CharacteristicKind FileType =
126 SourceMgr.getFileCharacteristic(Loc: CurLexer->getFileLoc());
127
128 FileID PrevFID;
129 SourceLocation EnterLoc;
130 if (PrevPPLexer) {
131 PrevFID = PrevPPLexer->getFileID();
132 EnterLoc = PrevPPLexer->getSourceLocation();
133 }
134 Callbacks->FileChanged(Loc: CurLexer->getFileLoc(), Reason: PPCallbacks::EnterFile,
135 FileType, PrevFID);
136 Callbacks->LexedFileChanged(FID: CurLexer->getFileID(),
137 Reason: PPCallbacks::LexedFileChangeReason::EnterFile,
138 FileType, PrevFID, Loc: EnterLoc);
139 }
140}
141
142/// EnterMacro - Add a Macro to the top of the include stack and start lexing
143/// tokens from it instead of the current buffer.
144void Preprocessor::EnterMacro(Token &Tok, SourceLocation ILEnd,
145 MacroInfo *Macro, MacroArgs *Args) {
146 std::unique_ptr<TokenLexer> TokLexer;
147 if (NumCachedTokenLexers == 0) {
148 TokLexer = std::make_unique<TokenLexer>(args&: Tok, args&: ILEnd, args&: Macro, args&: Args, args&: *this);
149 } else {
150 TokLexer = std::move(TokenLexerCache[--NumCachedTokenLexers]);
151 TokLexer->Init(Tok, ELEnd: ILEnd, MI: Macro, Actuals: Args);
152 }
153
154 PushIncludeMacroStack();
155 CurDirLookup = nullptr;
156 CurTokenLexer = std::move(TokLexer);
157 if (CurLexerCallback != CLK_LexAfterModuleImport)
158 CurLexerCallback = CLK_TokenLexer;
159}
160
161/// EnterTokenStream - Add a "macro" context to the top of the include stack,
162/// which will cause the lexer to start returning the specified tokens.
163///
164/// If DisableMacroExpansion is true, tokens lexed from the token stream will
165/// not be subject to further macro expansion. Otherwise, these tokens will
166/// be re-macro-expanded when/if expansion is enabled.
167///
168/// If OwnsTokens is false, this method assumes that the specified stream of
169/// tokens has a permanent owner somewhere, so they do not need to be copied.
170/// If it is true, it assumes the array of tokens is allocated with new[] and
171/// must be freed.
172///
173void Preprocessor::EnterTokenStream(const Token *Toks, unsigned NumToks,
174 bool DisableMacroExpansion, bool OwnsTokens,
175 bool IsReinject) {
176 if (CurLexerCallback == CLK_CachingLexer) {
177 if (CachedLexPos < CachedTokens.size()) {
178 assert(IsReinject && "new tokens in the middle of cached stream");
179 // We're entering tokens into the middle of our cached token stream. We
180 // can't represent that, so just insert the tokens into the buffer.
181 CachedTokens.insert(I: CachedTokens.begin() + CachedLexPos,
182 From: Toks, To: Toks + NumToks);
183 if (OwnsTokens)
184 delete [] Toks;
185 return;
186 }
187
188 // New tokens are at the end of the cached token sequnece; insert the
189 // token stream underneath the caching lexer.
190 ExitCachingLexMode();
191 EnterTokenStream(Toks, NumToks, DisableMacroExpansion, OwnsTokens,
192 IsReinject);
193 EnterCachingLexMode();
194 return;
195 }
196
197 // Create a macro expander to expand from the specified token stream.
198 std::unique_ptr<TokenLexer> TokLexer;
199 if (NumCachedTokenLexers == 0) {
200 TokLexer = std::make_unique<TokenLexer>(
201 args&: Toks, args&: NumToks, args&: DisableMacroExpansion, args&: OwnsTokens, args&: IsReinject, args&: *this);
202 } else {
203 TokLexer = std::move(TokenLexerCache[--NumCachedTokenLexers]);
204 TokLexer->Init(TokArray: Toks, NumToks, DisableMacroExpansion, OwnsTokens,
205 IsReinject);
206 }
207
208 // Save our current state.
209 PushIncludeMacroStack();
210 CurDirLookup = nullptr;
211 CurTokenLexer = std::move(TokLexer);
212 if (CurLexerCallback != CLK_LexAfterModuleImport)
213 CurLexerCallback = CLK_TokenLexer;
214}
215
216/// Compute the relative path that names the given file relative to
217/// the given directory.
218static void computeRelativePath(FileManager &FM, const DirectoryEntry *Dir,
219 FileEntryRef File, SmallString<128> &Result) {
220 Result.clear();
221
222 StringRef FilePath = File.getDir().getName();
223 StringRef Path = FilePath;
224 while (!Path.empty()) {
225 if (auto CurDir = FM.getOptionalDirectoryRef(DirName: Path)) {
226 if (*CurDir == Dir) {
227 Result = FilePath.substr(Start: Path.size());
228 llvm::sys::path::append(path&: Result,
229 a: llvm::sys::path::filename(path: File.getName()));
230 return;
231 }
232 }
233
234 Path = llvm::sys::path::parent_path(path: Path);
235 }
236
237 Result = File.getName();
238}
239
240void Preprocessor::PropagateLineStartLeadingSpaceInfo(Token &Result) {
241 if (CurTokenLexer) {
242 CurTokenLexer->PropagateLineStartLeadingSpaceInfo(Result);
243 return;
244 }
245 if (CurLexer) {
246 CurLexer->PropagateLineStartLeadingSpaceInfo(Result);
247 return;
248 }
249 // FIXME: Handle other kinds of lexers? It generally shouldn't matter,
250 // but it might if they're empty?
251}
252
253/// Determine the location to use as the end of the buffer for a lexer.
254///
255/// If the file ends with a newline, form the EOF token on the newline itself,
256/// rather than "on the line following it", which doesn't exist. This makes
257/// diagnostics relating to the end of file include the last file that the user
258/// actually typed, which is goodness.
259const char *Preprocessor::getCurLexerEndPos() {
260 const char *EndPos = CurLexer->BufferEnd;
261 if (EndPos != CurLexer->BufferStart &&
262 (EndPos[-1] == '\n' || EndPos[-1] == '\r')) {
263 --EndPos;
264
265 // Handle \n\r and \r\n:
266 if (EndPos != CurLexer->BufferStart &&
267 (EndPos[-1] == '\n' || EndPos[-1] == '\r') &&
268 EndPos[-1] != EndPos[0])
269 --EndPos;
270 }
271
272 return EndPos;
273}
274
275static void collectAllSubModulesWithUmbrellaHeader(
276 const Module &Mod, SmallVectorImpl<const Module *> &SubMods) {
277 if (Mod.getUmbrellaHeaderAsWritten())
278 SubMods.push_back(Elt: &Mod);
279 for (auto *M : Mod.submodules())
280 collectAllSubModulesWithUmbrellaHeader(Mod: *M, SubMods);
281}
282
283void Preprocessor::diagnoseMissingHeaderInUmbrellaDir(const Module &Mod) {
284 std::optional<Module::Header> UmbrellaHeader =
285 Mod.getUmbrellaHeaderAsWritten();
286 assert(UmbrellaHeader && "Module must use umbrella header");
287 const FileID &File = SourceMgr.translateFile(SourceFile: UmbrellaHeader->Entry);
288 SourceLocation ExpectedHeadersLoc = SourceMgr.getLocForEndOfFile(FID: File);
289 if (getDiagnostics().isIgnored(DiagID: diag::warn_uncovered_module_header,
290 Loc: ExpectedHeadersLoc))
291 return;
292
293 ModuleMap &ModMap = getHeaderSearchInfo().getModuleMap();
294 OptionalDirectoryEntryRef Dir = Mod.getEffectiveUmbrellaDir();
295 llvm::vfs::FileSystem &FS = FileMgr.getVirtualFileSystem();
296 std::error_code EC;
297 for (llvm::vfs::recursive_directory_iterator Entry(FS, Dir->getName(), EC),
298 End;
299 Entry != End && !EC; Entry.increment(EC)) {
300 using llvm::StringSwitch;
301
302 // Check whether this entry has an extension typically associated with
303 // headers.
304 if (!StringSwitch<bool>(llvm::sys::path::extension(path: Entry->path()))
305 .Cases(S0: ".h", S1: ".H", S2: ".hh", S3: ".hpp", Value: true)
306 .Default(Value: false))
307 continue;
308
309 if (auto Header = getFileManager().getOptionalFileRef(Filename: Entry->path()))
310 if (!getSourceManager().hasFileInfo(File: *Header)) {
311 if (!ModMap.isHeaderInUnavailableModule(Header: *Header)) {
312 // Find the relative path that would access this header.
313 SmallString<128> RelativePath;
314 computeRelativePath(FM&: FileMgr, Dir: *Dir, File: *Header, Result&: RelativePath);
315 Diag(Loc: ExpectedHeadersLoc, DiagID: diag::warn_uncovered_module_header)
316 << Mod.getFullModuleName() << RelativePath;
317 }
318 }
319 }
320}
321
322/// HandleEndOfFile - This callback is invoked when the lexer hits the end of
323/// the current file. This either returns the EOF token or pops a level off
324/// the include stack and keeps going.
325bool Preprocessor::HandleEndOfFile(Token &Result, bool isEndOfMacro) {
326 assert(!CurTokenLexer &&
327 "Ending a file when currently in a macro!");
328
329 SourceLocation UnclosedSafeBufferOptOutLoc;
330
331 if (IncludeMacroStack.empty() &&
332 isPPInSafeBufferOptOutRegion(StartLoc&: UnclosedSafeBufferOptOutLoc)) {
333 // To warn if a "-Wunsafe-buffer-usage" opt-out region is still open by the
334 // end of a file.
335 Diag(Loc: UnclosedSafeBufferOptOutLoc,
336 DiagID: diag::err_pp_unclosed_pragma_unsafe_buffer_usage);
337 }
338 // If we have an unclosed module region from a pragma at the end of a
339 // module, complain and close it now.
340 const bool LeavingSubmodule = CurLexer && CurLexerSubmodule;
341 if ((LeavingSubmodule || IncludeMacroStack.empty()) &&
342 !BuildingSubmoduleStack.empty() &&
343 BuildingSubmoduleStack.back().IsPragma) {
344 Diag(Loc: BuildingSubmoduleStack.back().ImportLoc,
345 DiagID: diag::err_pp_module_begin_without_module_end);
346 Module *M = LeaveSubmodule(/*ForPragma*/true);
347
348 Result.startToken();
349 const char *EndPos = getCurLexerEndPos();
350 CurLexer->BufferPtr = EndPos;
351 CurLexer->FormTokenWithChars(Result, TokEnd: EndPos, Kind: tok::annot_module_end);
352 Result.setAnnotationEndLoc(Result.getLocation());
353 Result.setAnnotationValue(M);
354 return true;
355 }
356
357 // See if this file had a controlling macro.
358 if (CurPPLexer) { // Not ending a macro, ignore it.
359 if (const IdentifierInfo *ControllingMacro =
360 CurPPLexer->MIOpt.GetControllingMacroAtEndOfFile()) {
361 // Okay, this has a controlling macro, remember in HeaderFileInfo.
362 if (OptionalFileEntryRef FE = CurPPLexer->getFileEntry()) {
363 HeaderInfo.SetFileControllingMacro(File: *FE, ControllingMacro);
364 if (MacroInfo *MI = getMacroInfo(II: ControllingMacro))
365 MI->setUsedForHeaderGuard(true);
366 if (const IdentifierInfo *DefinedMacro =
367 CurPPLexer->MIOpt.GetDefinedMacro()) {
368 if (!isMacroDefined(II: ControllingMacro) &&
369 DefinedMacro != ControllingMacro &&
370 CurLexer->isFirstTimeLexingFile()) {
371
372 // If the edit distance between the two macros is more than 50%,
373 // DefinedMacro may not be header guard, or can be header guard of
374 // another header file. Therefore, it maybe defining something
375 // completely different. This can be observed in the wild when
376 // handling feature macros or header guards in different files.
377
378 const StringRef ControllingMacroName = ControllingMacro->getName();
379 const StringRef DefinedMacroName = DefinedMacro->getName();
380 const size_t MaxHalfLength = std::max(a: ControllingMacroName.size(),
381 b: DefinedMacroName.size()) / 2;
382 const unsigned ED = ControllingMacroName.edit_distance(
383 Other: DefinedMacroName, AllowReplacements: true, MaxEditDistance: MaxHalfLength);
384 if (ED <= MaxHalfLength) {
385 // Emit a warning for a bad header guard.
386 Diag(Loc: CurPPLexer->MIOpt.GetMacroLocation(),
387 DiagID: diag::warn_header_guard)
388 << CurPPLexer->MIOpt.GetMacroLocation() << ControllingMacro;
389 Diag(Loc: CurPPLexer->MIOpt.GetDefinedLocation(),
390 DiagID: diag::note_header_guard)
391 << CurPPLexer->MIOpt.GetDefinedLocation() << DefinedMacro
392 << ControllingMacro
393 << FixItHint::CreateReplacement(
394 RemoveRange: CurPPLexer->MIOpt.GetDefinedLocation(),
395 Code: ControllingMacro->getName());
396 }
397 }
398 }
399 }
400 }
401 }
402
403 // Complain about reaching a true EOF within arc_cf_code_audited.
404 // We don't want to complain about reaching the end of a macro
405 // instantiation or a _Pragma.
406 if (PragmaARCCFCodeAuditedInfo.getLoc().isValid() && !isEndOfMacro &&
407 !(CurLexer && CurLexer->Is_PragmaLexer)) {
408 Diag(Loc: PragmaARCCFCodeAuditedInfo.getLoc(),
409 DiagID: diag::err_pp_eof_in_arc_cf_code_audited);
410
411 // Recover by leaving immediately.
412 PragmaARCCFCodeAuditedInfo = IdentifierLoc();
413 }
414
415 // Complain about reaching a true EOF within assume_nonnull.
416 // We don't want to complain about reaching the end of a macro
417 // instantiation or a _Pragma.
418 if (PragmaAssumeNonNullLoc.isValid() &&
419 !isEndOfMacro && !(CurLexer && CurLexer->Is_PragmaLexer)) {
420 // If we're at the end of generating a preamble, we should record the
421 // unterminated \#pragma clang assume_nonnull so we can restore it later
422 // when the preamble is loaded into the main file.
423 if (isRecordingPreamble() && isInPrimaryFile())
424 PreambleRecordedPragmaAssumeNonNullLoc = PragmaAssumeNonNullLoc;
425 else
426 Diag(Loc: PragmaAssumeNonNullLoc, DiagID: diag::err_pp_eof_in_assume_nonnull);
427 // Recover by leaving immediately.
428 PragmaAssumeNonNullLoc = SourceLocation();
429 }
430
431 bool LeavingPCHThroughHeader = false;
432
433 // If this is a #include'd file, pop it off the include stack and continue
434 // lexing the #includer file.
435 if (!IncludeMacroStack.empty()) {
436
437 // If we lexed the code-completion file, act as if we reached EOF.
438 if (isCodeCompletionEnabled() && CurPPLexer &&
439 SourceMgr.getLocForStartOfFile(FID: CurPPLexer->getFileID()) ==
440 CodeCompletionFileLoc) {
441 assert(CurLexer && "Got EOF but no current lexer set!");
442 Result.startToken();
443 CurLexer->FormTokenWithChars(Result, TokEnd: CurLexer->BufferEnd, Kind: tok::eof);
444 CurLexer.reset();
445
446 CurPPLexer = nullptr;
447 recomputeCurLexerKind();
448 return true;
449 }
450
451 if (!isEndOfMacro && CurPPLexer &&
452 (SourceMgr.getIncludeLoc(FID: CurPPLexer->getFileID()).isValid() ||
453 // Predefines file doesn't have a valid include location.
454 (PredefinesFileID.isValid() &&
455 CurPPLexer->getFileID() == PredefinesFileID))) {
456 // Notify SourceManager to record the number of FileIDs that were created
457 // during lexing of the #include'd file.
458 unsigned NumFIDs =
459 SourceMgr.local_sloc_entry_size() -
460 CurPPLexer->getInitialNumSLocEntries() + 1/*#include'd file*/;
461 SourceMgr.setNumCreatedFIDsForFileID(FID: CurPPLexer->getFileID(), NumFIDs);
462 }
463
464 bool ExitedFromPredefinesFile = false;
465 FileID ExitedFID;
466 if (!isEndOfMacro && CurPPLexer) {
467 ExitedFID = CurPPLexer->getFileID();
468
469 assert(PredefinesFileID.isValid() &&
470 "HandleEndOfFile is called before PredefinesFileId is set");
471 ExitedFromPredefinesFile = (PredefinesFileID == ExitedFID);
472 }
473
474 if (LeavingSubmodule) {
475 // We're done with this submodule.
476 Module *M = LeaveSubmodule(/*ForPragma*/false);
477
478 // Notify the parser that we've left the module.
479 const char *EndPos = getCurLexerEndPos();
480 Result.startToken();
481 CurLexer->BufferPtr = EndPos;
482 CurLexer->FormTokenWithChars(Result, TokEnd: EndPos, Kind: tok::annot_module_end);
483 Result.setAnnotationEndLoc(Result.getLocation());
484 Result.setAnnotationValue(M);
485 }
486
487 bool FoundPCHThroughHeader = false;
488 if (CurPPLexer && creatingPCHWithThroughHeader() &&
489 isPCHThroughHeader(
490 FE: SourceMgr.getFileEntryForID(FID: CurPPLexer->getFileID())))
491 FoundPCHThroughHeader = true;
492
493 // We're done with the #included file.
494 RemoveTopOfLexerStack();
495
496 // Propagate info about start-of-line/leading white-space/etc.
497 PropagateLineStartLeadingSpaceInfo(Result);
498
499 // Notify the client, if desired, that we are in a new source file.
500 if (Callbacks && !isEndOfMacro && CurPPLexer) {
501 SourceLocation Loc = CurPPLexer->getSourceLocation();
502 SrcMgr::CharacteristicKind FileType =
503 SourceMgr.getFileCharacteristic(Loc);
504 Callbacks->FileChanged(Loc, Reason: PPCallbacks::ExitFile, FileType, PrevFID: ExitedFID);
505 Callbacks->LexedFileChanged(FID: CurPPLexer->getFileID(),
506 Reason: PPCallbacks::LexedFileChangeReason::ExitFile,
507 FileType, PrevFID: ExitedFID, Loc);
508 }
509
510 // Restore conditional stack as well as the recorded
511 // \#pragma clang assume_nonnull from the preamble right after exiting
512 // from the predefines file.
513 if (ExitedFromPredefinesFile) {
514 replayPreambleConditionalStack();
515 if (PreambleRecordedPragmaAssumeNonNullLoc.isValid())
516 PragmaAssumeNonNullLoc = PreambleRecordedPragmaAssumeNonNullLoc;
517 }
518
519 if (!isEndOfMacro && CurPPLexer && FoundPCHThroughHeader &&
520 (isInPrimaryFile() ||
521 CurPPLexer->getFileID() == getPredefinesFileID())) {
522 // Leaving the through header. Continue directly to end of main file
523 // processing.
524 LeavingPCHThroughHeader = true;
525 } else {
526 // Client should lex another token unless we generated an EOM.
527 return LeavingSubmodule;
528 }
529 }
530 // If this is the end of the main file, form an EOF token.
531 assert(CurLexer && "Got EOF but no current lexer set!");
532 const char *EndPos = getCurLexerEndPos();
533 Result.startToken();
534 CurLexer->BufferPtr = EndPos;
535
536 if (getLangOpts().IncrementalExtensions) {
537 CurLexer->FormTokenWithChars(Result, TokEnd: EndPos, Kind: tok::annot_repl_input_end);
538 Result.setAnnotationEndLoc(Result.getLocation());
539 Result.setAnnotationValue(nullptr);
540 } else {
541 CurLexer->FormTokenWithChars(Result, TokEnd: EndPos, Kind: tok::eof);
542 }
543
544 if (isCodeCompletionEnabled()) {
545 // Inserting the code-completion point increases the source buffer by 1,
546 // but the main FileID was created before inserting the point.
547 // Compensate by reducing the EOF location by 1, otherwise the location
548 // will point to the next FileID.
549 // FIXME: This is hacky, the code-completion point should probably be
550 // inserted before the main FileID is created.
551 if (CurLexer->getFileLoc() == CodeCompletionFileLoc)
552 Result.setLocation(Result.getLocation().getLocWithOffset(Offset: -1));
553 }
554
555 if (creatingPCHWithThroughHeader() && !LeavingPCHThroughHeader) {
556 // Reached the end of the compilation without finding the through header.
557 Diag(Loc: CurLexer->getFileLoc(), DiagID: diag::err_pp_through_header_not_seen)
558 << PPOpts.PCHThroughHeader << 0;
559 }
560
561 if (!isIncrementalProcessingEnabled())
562 // We're done with lexing.
563 CurLexer.reset();
564
565 if (!isIncrementalProcessingEnabled())
566 CurPPLexer = nullptr;
567
568 if (TUKind == TU_Complete) {
569 // This is the end of the top-level file. 'WarnUnusedMacroLocs' has
570 // collected all macro locations that we need to warn because they are not
571 // used.
572 for (WarnUnusedMacroLocsTy::iterator
573 I=WarnUnusedMacroLocs.begin(), E=WarnUnusedMacroLocs.end();
574 I!=E; ++I)
575 Diag(Loc: *I, DiagID: diag::pp_macro_not_used);
576 }
577
578 // If we are building a module that has an umbrella header, make sure that
579 // each of the headers within the directory, including all submodules, is
580 // covered by the umbrella header was actually included by the umbrella
581 // header.
582 if (Module *Mod = getCurrentModule()) {
583 llvm::SmallVector<const Module *, 4> AllMods;
584 collectAllSubModulesWithUmbrellaHeader(Mod: *Mod, SubMods&: AllMods);
585 for (auto *M : AllMods)
586 diagnoseMissingHeaderInUmbrellaDir(Mod: *M);
587 }
588
589 return true;
590}
591
592/// HandleEndOfTokenLexer - This callback is invoked when the current TokenLexer
593/// hits the end of its token stream.
594bool Preprocessor::HandleEndOfTokenLexer(Token &Result) {
595 assert(CurTokenLexer && !CurPPLexer &&
596 "Ending a macro when currently in a #include file!");
597
598 if (!MacroExpandingLexersStack.empty() &&
599 MacroExpandingLexersStack.back().first == CurTokenLexer.get())
600 removeCachedMacroExpandedTokensOfLastLexer();
601
602 // Delete or cache the now-dead macro expander.
603 if (NumCachedTokenLexers == TokenLexerCacheSize)
604 CurTokenLexer.reset();
605 else
606 TokenLexerCache[NumCachedTokenLexers++] = std::move(CurTokenLexer);
607
608 // Handle this like a #include file being popped off the stack.
609 return HandleEndOfFile(Result, isEndOfMacro: true);
610}
611
612/// RemoveTopOfLexerStack - Pop the current lexer/macro exp off the top of the
613/// lexer stack. This should only be used in situations where the current
614/// state of the top-of-stack lexer is unknown.
615void Preprocessor::RemoveTopOfLexerStack() {
616 assert(!IncludeMacroStack.empty() && "Ran out of stack entries to load");
617
618 if (CurTokenLexer) {
619 // Delete or cache the now-dead macro expander.
620 if (NumCachedTokenLexers == TokenLexerCacheSize)
621 CurTokenLexer.reset();
622 else
623 TokenLexerCache[NumCachedTokenLexers++] = std::move(CurTokenLexer);
624 }
625
626 PopIncludeMacroStack();
627}
628
629/// HandleMicrosoftCommentPaste - When the macro expander pastes together a
630/// comment (/##/) in microsoft mode, this method handles updating the current
631/// state, returning the token on the next source line.
632void Preprocessor::HandleMicrosoftCommentPaste(Token &Tok) {
633 assert(CurTokenLexer && !CurPPLexer &&
634 "Pasted comment can only be formed from macro");
635 // We handle this by scanning for the closest real lexer, switching it to
636 // raw mode and preprocessor mode. This will cause it to return \n as an
637 // explicit EOD token.
638 PreprocessorLexer *FoundLexer = nullptr;
639 bool LexerWasInPPMode = false;
640 for (const IncludeStackInfo &ISI : llvm::reverse(C&: IncludeMacroStack)) {
641 if (ISI.ThePPLexer == nullptr) continue; // Scan for a real lexer.
642
643 // Once we find a real lexer, mark it as raw mode (disabling macro
644 // expansions) and preprocessor mode (return EOD). We know that the lexer
645 // was *not* in raw mode before, because the macro that the comment came
646 // from was expanded. However, it could have already been in preprocessor
647 // mode (#if COMMENT) in which case we have to return it to that mode and
648 // return EOD.
649 FoundLexer = ISI.ThePPLexer;
650 FoundLexer->LexingRawMode = true;
651 LexerWasInPPMode = FoundLexer->ParsingPreprocessorDirective;
652 FoundLexer->ParsingPreprocessorDirective = true;
653 break;
654 }
655
656 // Okay, we either found and switched over the lexer, or we didn't find a
657 // lexer. In either case, finish off the macro the comment came from, getting
658 // the next token.
659 if (!HandleEndOfTokenLexer(Result&: Tok)) Lex(Result&: Tok);
660
661 // Discarding comments as long as we don't have EOF or EOD. This 'comments
662 // out' the rest of the line, including any tokens that came from other macros
663 // that were active, as in:
664 // #define submacro a COMMENT b
665 // submacro c
666 // which should lex to 'a' only: 'b' and 'c' should be removed.
667 while (Tok.isNot(K: tok::eod) && Tok.isNot(K: tok::eof))
668 Lex(Result&: Tok);
669
670 // If we got an eod token, then we successfully found the end of the line.
671 if (Tok.is(K: tok::eod)) {
672 assert(FoundLexer && "Can't get end of line without an active lexer");
673 // Restore the lexer back to normal mode instead of raw mode.
674 FoundLexer->LexingRawMode = false;
675
676 // If the lexer was already in preprocessor mode, just return the EOD token
677 // to finish the preprocessor line.
678 if (LexerWasInPPMode) return;
679
680 // Otherwise, switch out of PP mode and return the next lexed token.
681 FoundLexer->ParsingPreprocessorDirective = false;
682 return Lex(Result&: Tok);
683 }
684
685 // If we got an EOF token, then we reached the end of the token stream but
686 // didn't find an explicit \n. This can only happen if there was no lexer
687 // active (an active lexer would return EOD at EOF if there was no \n in
688 // preprocessor directive mode), so just return EOF as our token.
689 assert(!FoundLexer && "Lexer should return EOD before EOF in PP mode");
690}
691
692void Preprocessor::EnterSubmodule(Module *M, SourceLocation ImportLoc,
693 bool ForPragma) {
694 if (!getLangOpts().ModulesLocalVisibility) {
695 // Just track that we entered this submodule.
696 BuildingSubmoduleStack.push_back(
697 Elt: BuildingSubmoduleInfo(M, ImportLoc, ForPragma, CurSubmoduleState,
698 PendingModuleMacroNames.size()));
699 if (Callbacks)
700 Callbacks->EnteredSubmodule(M, ImportLoc, ForPragma);
701 return;
702 }
703
704 // Resolve as much of the module definition as we can now, before we enter
705 // one of its headers.
706 // FIXME: Can we enable Complain here?
707 // FIXME: Can we do this when local visibility is disabled?
708 ModuleMap &ModMap = getHeaderSearchInfo().getModuleMap();
709 ModMap.resolveExports(Mod: M, /*Complain=*/false);
710 ModMap.resolveUses(Mod: M, /*Complain=*/false);
711 ModMap.resolveConflicts(Mod: M, /*Complain=*/false);
712
713 // If this is the first time we've entered this module, set up its state.
714 auto R = Submodules.try_emplace(k: M);
715 auto &State = R.first->second;
716 bool FirstTime = R.second;
717 if (FirstTime) {
718 // Determine the set of starting macros for this submodule; take these
719 // from the "null" module (the predefines buffer).
720 //
721 // FIXME: If we have local visibility but not modules enabled, the
722 // NullSubmoduleState is polluted by #defines in the top-level source
723 // file.
724 auto &StartingMacros = NullSubmoduleState.Macros;
725
726 // Restore to the starting state.
727 // FIXME: Do this lazily, when each macro name is first referenced.
728 for (auto &Macro : StartingMacros) {
729 // Skip uninteresting macros.
730 if (!Macro.second.getLatest() &&
731 Macro.second.getOverriddenMacros().empty())
732 continue;
733
734 MacroState MS(Macro.second.getLatest());
735 MS.setOverriddenMacros(PP&: *this, Overrides: Macro.second.getOverriddenMacros());
736 State.Macros.insert(KV: std::make_pair(x&: Macro.first, y: std::move(MS)));
737 }
738 }
739
740 // Track that we entered this module.
741 BuildingSubmoduleStack.push_back(
742 Elt: BuildingSubmoduleInfo(M, ImportLoc, ForPragma, CurSubmoduleState,
743 PendingModuleMacroNames.size()));
744
745 if (Callbacks)
746 Callbacks->EnteredSubmodule(M, ImportLoc, ForPragma);
747
748 // Switch to this submodule as the current submodule.
749 CurSubmoduleState = &State;
750
751 // This module is visible to itself, but exports should not be made visible
752 // until they are imported.
753 if (FirstTime)
754 makeModuleVisible(M, Loc: ImportLoc, /*IncludeExports=*/false);
755}
756
757bool Preprocessor::needModuleMacros() const {
758 // If we're not within a submodule, we never need to create ModuleMacros.
759 if (BuildingSubmoduleStack.empty())
760 return false;
761 // If we are tracking module macro visibility even for textually-included
762 // headers, we need ModuleMacros.
763 if (getLangOpts().ModulesLocalVisibility)
764 return true;
765 // Otherwise, we only need module macros if we're actually compiling a module
766 // interface.
767 return getLangOpts().isCompilingModule();
768}
769
770Module *Preprocessor::LeaveSubmodule(bool ForPragma) {
771 if (BuildingSubmoduleStack.empty() ||
772 BuildingSubmoduleStack.back().IsPragma != ForPragma) {
773 assert(ForPragma && "non-pragma module enter/leave mismatch");
774 return nullptr;
775 }
776
777 auto &Info = BuildingSubmoduleStack.back();
778
779 Module *LeavingMod = Info.M;
780 SourceLocation ImportLoc = Info.ImportLoc;
781
782 if (!needModuleMacros() ||
783 (!getLangOpts().ModulesLocalVisibility &&
784 LeavingMod->getTopLevelModuleName() != getLangOpts().CurrentModule)) {
785 // If we don't need module macros, or this is not a module for which we
786 // are tracking macro visibility, don't build any, and preserve the list
787 // of pending names for the surrounding submodule.
788 BuildingSubmoduleStack.pop_back();
789
790 if (Callbacks)
791 Callbacks->LeftSubmodule(M: LeavingMod, ImportLoc, ForPragma);
792
793 makeModuleVisible(M: LeavingMod, Loc: ImportLoc);
794 return LeavingMod;
795 }
796
797 // Create ModuleMacros for any macros defined in this submodule.
798 llvm::SmallPtrSet<const IdentifierInfo*, 8> VisitedMacros;
799 for (unsigned I = Info.OuterPendingModuleMacroNames;
800 I != PendingModuleMacroNames.size(); ++I) {
801 auto *II = PendingModuleMacroNames[I];
802 if (!VisitedMacros.insert(Ptr: II).second)
803 continue;
804
805 auto MacroIt = CurSubmoduleState->Macros.find(Val: II);
806 if (MacroIt == CurSubmoduleState->Macros.end())
807 continue;
808 auto &Macro = MacroIt->second;
809
810 // Find the starting point for the MacroDirective chain in this submodule.
811 MacroDirective *OldMD = nullptr;
812 auto *OldState = Info.OuterSubmoduleState;
813 if (getLangOpts().ModulesLocalVisibility)
814 OldState = &NullSubmoduleState;
815 if (OldState && OldState != CurSubmoduleState) {
816 // FIXME: It'd be better to start at the state from when we most recently
817 // entered this submodule, but it doesn't really matter.
818 auto &OldMacros = OldState->Macros;
819 auto OldMacroIt = OldMacros.find(Val: II);
820 if (OldMacroIt == OldMacros.end())
821 OldMD = nullptr;
822 else
823 OldMD = OldMacroIt->second.getLatest();
824 }
825
826 // This module may have exported a new macro. If so, create a ModuleMacro
827 // representing that fact.
828 bool ExplicitlyPublic = false;
829 for (auto *MD = Macro.getLatest(); MD != OldMD; MD = MD->getPrevious()) {
830 assert(MD && "broken macro directive chain");
831
832 if (auto *VisMD = dyn_cast<VisibilityMacroDirective>(Val: MD)) {
833 // The latest visibility directive for a name in a submodule affects
834 // all the directives that come before it.
835 if (VisMD->isPublic())
836 ExplicitlyPublic = true;
837 else if (!ExplicitlyPublic)
838 // Private with no following public directive: not exported.
839 break;
840 } else {
841 MacroInfo *Def = nullptr;
842 if (DefMacroDirective *DefMD = dyn_cast<DefMacroDirective>(Val: MD))
843 Def = DefMD->getInfo();
844
845 // FIXME: Issue a warning if multiple headers for the same submodule
846 // define a macro, rather than silently ignoring all but the first.
847 bool IsNew;
848 // Don't bother creating a module macro if it would represent a #undef
849 // that doesn't override anything.
850 if (Def || !Macro.getOverriddenMacros().empty())
851 addModuleMacro(Mod: LeavingMod, II, Macro: Def, Overrides: Macro.getOverriddenMacros(),
852 IsNew);
853
854 if (!getLangOpts().ModulesLocalVisibility) {
855 // This macro is exposed to the rest of this compilation as a
856 // ModuleMacro; we don't need to track its MacroDirective any more.
857 Macro.setLatest(nullptr);
858 Macro.setOverriddenMacros(PP&: *this, Overrides: {});
859 }
860 break;
861 }
862 }
863 }
864 PendingModuleMacroNames.resize(N: Info.OuterPendingModuleMacroNames);
865
866 // FIXME: Before we leave this submodule, we should parse all the other
867 // headers within it. Otherwise, we're left with an inconsistent state
868 // where we've made the module visible but don't yet have its complete
869 // contents.
870
871 // Put back the outer module's state, if we're tracking it.
872 if (getLangOpts().ModulesLocalVisibility)
873 CurSubmoduleState = Info.OuterSubmoduleState;
874
875 BuildingSubmoduleStack.pop_back();
876
877 if (Callbacks)
878 Callbacks->LeftSubmodule(M: LeavingMod, ImportLoc, ForPragma);
879
880 // A nested #include makes the included submodule visible.
881 makeModuleVisible(M: LeavingMod, Loc: ImportLoc);
882 return LeavingMod;
883}
884