1 //===- extra/modularize/Modularize.cpp - Check modularized headers --------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // This file implements a tool that checks whether a set of headers provides
11 // the consistent definitions required to use modules. For example, it detects
12 // whether the same entity (say, a NULL macro or size_t typedef) is defined in
13 // multiple headers or whether a header produces different definitions under
14 // different circumstances. These conditions cause modules built from the
15 // headers to behave poorly, and should be fixed before introducing a module
16 // map.
17 //
18 // Modularize takes as argument a file name for a file containing the
19 // newline-separated list of headers to check with respect to each other.
20 // Lines beginning with '#' and empty lines are ignored.
21 // Header file names followed by a colon and other space-separated
22 // file names will include those extra files as dependencies.
23 // The file names can be relative or full paths, but must be on the
24 // same line.
25 //
26 // Modularize also accepts regular front-end arguments.
27 //
28 // Usage:   modularize [-prefix (optional header path prefix)]
29 //   (include-files_list) [(front-end-options) ...]
30 //
31 // Note that unless a "-prefix (header path)" option is specified,
32 // non-absolute file paths in the header list file will be relative
33 // to the header list file directory.  Use -prefix to specify a different
34 // directory.
35 //
36 // Note that by default, the underlying Clang front end assumes .h files
37 // contain C source.  If your .h files in the file list contain C++ source,
38 // you should append the following to your command lines: -x c++
39 //
40 // Modularize will do normal parsing, reporting normal errors and warnings,
41 // but will also report special error messages like the following:
42 //
43 //   error: '(symbol)' defined at multiple locations:
44 //       (file):(row):(column)
45 //       (file):(row):(column)
46 //
47 //   error: header '(file)' has different contents depending on how it was
48 //     included
49 //
50 // The latter might be followed by messages like the following:
51 //
52 //   note: '(symbol)' in (file) at (row):(column) not always provided
53 //
54 // Checks will also be performed for macro expansions, defined(macro)
55 // expressions, and preprocessor conditional directives that evaluate
56 // inconsistently, and can produce error messages like the following:
57 //
58 //   (...)/SubHeader.h:11:5:
59 //   #if SYMBOL == 1
60 //       ^
61 //   error: Macro instance 'SYMBOL' has different values in this header,
62 //          depending on how it was included.
63 //     'SYMBOL' expanded to: '1' with respect to these inclusion paths:
64 //       (...)/Header1.h
65 //         (...)/SubHeader.h
66 //   (...)/SubHeader.h:3:9:
67 //   #define SYMBOL 1
68 //             ^
69 //   Macro defined here.
70 //     'SYMBOL' expanded to: '2' with respect to these inclusion paths:
71 //       (...)/Header2.h
72 //           (...)/SubHeader.h
73 //   (...)/SubHeader.h:7:9:
74 //   #define SYMBOL 2
75 //             ^
76 //   Macro defined here.
77 //
78 // Checks will also be performed for '#include' directives that are
79 // nested inside 'extern "C/C++" {}' or 'namespace (name) {}' blocks,
80 // and can produce error message like the following:
81 //
82 // IncludeInExtern.h:2:3
83 //   #include "Empty.h"
84 //   ^
85 // error: Include directive within extern "C" {}.
86 // IncludeInExtern.h:1:1
87 // extern "C" {
88 // ^
89 // The "extern "C" {}" block is here.
90 //
91 // See PreprocessorTracker.cpp for additional details.
92 //
93 // Modularize also has an option ("-module-map-path=module.modulemap") that will
94 // skip the checks, and instead act as a module.modulemap generation assistant,
95 // generating a module map file based on the header list.  An optional
96 // "-root-module=(rootName)" argument can specify a root module to be
97 // created in the generated module.modulemap file.  Note that you will likely
98 // need to edit this file to suit the needs of your headers.
99 //
100 // An example command line for generating a module.modulemap file:
101 //
102 //   modularize -module-map-path=module.modulemap -root-module=myroot \
103 //      headerlist.txt
104 //
105 // Note that if the headers in the header list have partial paths, sub-modules
106 // will be created for the subdirectires involved, assuming that the
107 // subdirectories contain headers to be grouped into a module, but still with
108 // individual modules for the headers in the subdirectory.
109 //
110 // See the ModuleAssistant.cpp file comments for additional details about the
111 // implementation of the assistant mode.
112 //
113 // Future directions:
114 //
115 // Basically, we want to add new checks for whatever we can check with respect
116 // to checking headers for module'ability.
117 //
118 // Some ideas:
119 //
120 // 1. Omit duplicate "not always provided" messages
121 //
122 // 2. Add options to disable any of the checks, in case
123 // there is some problem with them, or the messages get too verbose.
124 //
125 // 3. Try to figure out the preprocessor conditional directives that
126 // contribute to problems and tie them to the inconsistent definitions.
127 //
128 // 4. There are some legitimate uses of preprocessor macros that
129 // modularize will flag as errors, such as repeatedly #include'ing
130 // a file and using interleaving defined/undefined macros
131 // to change declarations in the included file.  Is there a way
132 // to address this?  Maybe have modularize accept a list of macros
133 // to ignore.  Otherwise you can just exclude the file, after checking
134 // for legitimate errors.
135 //
136 // 5. What else?
137 //
138 // General clean-up and refactoring:
139 //
140 // 1. The Location class seems to be something that we might
141 // want to design to be applicable to a wider range of tools, and stick it
142 // somewhere into Tooling/ in mainline
143 //
144 //===----------------------------------------------------------------------===//
145 
146 #include "Modularize.h"
147 #include "PreprocessorTracker.h"
148 #include "clang/AST/ASTConsumer.h"
149 #include "clang/AST/ASTContext.h"
150 #include "clang/AST/RecursiveASTVisitor.h"
151 #include "clang/Basic/SourceManager.h"
152 #include "clang/Driver/Options.h"
153 #include "clang/Frontend/CompilerInstance.h"
154 #include "clang/Frontend/FrontendActions.h"
155 #include "clang/Lex/Preprocessor.h"
156 #include "clang/Tooling/CompilationDatabase.h"
157 #include "clang/Tooling/Tooling.h"
158 #include "llvm/Option/Arg.h"
159 #include "llvm/Option/ArgList.h"
160 #include "llvm/Option/OptTable.h"
161 #include "llvm/Option/Option.h"
162 #include "llvm/Support/CommandLine.h"
163 #include "llvm/Support/FileSystem.h"
164 #include "llvm/Support/MemoryBuffer.h"
165 #include "llvm/Support/Path.h"
166 #include <algorithm>
167 #include <fstream>
168 #include <iterator>
169 #include <string>
170 #include <vector>
171 
172 using namespace clang;
173 using namespace clang::driver;
174 using namespace clang::driver::options;
175 using namespace clang::tooling;
176 using namespace llvm;
177 using namespace llvm::opt;
178 using namespace Modularize;
179 
180 // Option to specify a file name for a list of header files to check.
181 cl::list<std::string>
182 ListFileNames(cl::Positional, cl::value_desc("list"),
183               cl::desc("<list of one or more header list files>"),
184               cl::CommaSeparated);
185 
186 // Collect all other arguments, which will be passed to the front end.
187 cl::list<std::string>
188 CC1Arguments(cl::ConsumeAfter,
189              cl::desc("<arguments to be passed to front end>..."));
190 
191 // Option to specify a prefix to be prepended to the header names.
192 cl::opt<std::string> HeaderPrefix(
193     "prefix", cl::init(""),
194     cl::desc(
195         "Prepend header file paths with this prefix."
196         " If not specified,"
197         " the files are considered to be relative to the header list file."));
198 
199 // Option for assistant mode, telling modularize to output a module map
200 // based on the headers list, and where to put it.
201 cl::opt<std::string> ModuleMapPath(
202     "module-map-path", cl::init(""),
203     cl::desc("Turn on module map output and specify output path or file name."
204              " If no path is specified and if prefix option is specified,"
205              " use prefix for file path."));
206 
207 // Option for assistant mode, telling modularize to output a module map
208 // based on the headers list, and where to put it.
209 cl::opt<std::string>
210 RootModule("root-module", cl::init(""),
211            cl::desc("Specify the name of the root module."));
212 
213 // Option for limiting the #include-inside-extern-or-namespace-block
214 // check to only those headers explicitly listed in the header list.
215 // This is a work-around for private includes that purposefully get
216 // included inside blocks.
217 static cl::opt<bool>
218 BlockCheckHeaderListOnly("block-check-header-list-only", cl::init(false),
219 cl::desc("Only warn if #include directives are inside extern or namespace"
220   " blocks if the included header is in the header list."));
221 
222 // Save the program name for error messages.
223 const char *Argv0;
224 // Save the command line for comments.
225 std::string CommandLine;
226 
227 // Read the header list file and collect the header file names and
228 // optional dependencies.
229 std::error_code
230 getHeaderFileNames(SmallVectorImpl<std::string> &HeaderFileNames,
231                    DependencyMap &Dependencies, StringRef ListFileName,
232                    StringRef HeaderPrefix) {
233   // By default, use the path component of the list file name.
234   SmallString<256> HeaderDirectory(ListFileName);
235   sys::path::remove_filename(HeaderDirectory);
236   SmallString<256> CurrentDirectory;
237   sys::fs::current_path(CurrentDirectory);
238 
239   // Get the prefix if we have one.
240   if (HeaderPrefix.size() != 0)
241     HeaderDirectory = HeaderPrefix;
242 
243   // Read the header list file into a buffer.
244   ErrorOr<std::unique_ptr<MemoryBuffer>> listBuffer =
245       MemoryBuffer::getFile(ListFileName);
246   if (std::error_code EC = listBuffer.getError())
247     return EC;
248 
249   // Parse the header list into strings.
250   SmallVector<StringRef, 32> Strings;
251   listBuffer.get()->getBuffer().split(Strings, "\n", -1, false);
252 
253   // Collect the header file names from the string list.
254   for (SmallVectorImpl<StringRef>::iterator I = Strings.begin(),
255                                             E = Strings.end();
256        I != E; ++I) {
257     StringRef Line = I->trim();
258     // Ignore comments and empty lines.
259     if (Line.empty() || (Line[0] == '#'))
260       continue;
261     std::pair<StringRef, StringRef> TargetAndDependents = Line.split(':');
262     SmallString<256> HeaderFileName;
263     // Prepend header file name prefix if it's not absolute.
264     if (sys::path::is_absolute(TargetAndDependents.first))
265       llvm::sys::path::native(TargetAndDependents.first, HeaderFileName);
266     else {
267       if (HeaderDirectory.size() != 0)
268         HeaderFileName = HeaderDirectory;
269       else
270         HeaderFileName = CurrentDirectory;
271       sys::path::append(HeaderFileName, TargetAndDependents.first);
272       sys::path::native(HeaderFileName);
273     }
274     // Handle optional dependencies.
275     DependentsVector Dependents;
276     SmallVector<StringRef, 4> DependentsList;
277     TargetAndDependents.second.split(DependentsList, " ", -1, false);
278     int Count = DependentsList.size();
279     for (int Index = 0; Index < Count; ++Index) {
280       SmallString<256> Dependent;
281       if (sys::path::is_absolute(DependentsList[Index]))
282         Dependent = DependentsList[Index];
283       else {
284         if (HeaderDirectory.size() != 0)
285           Dependent = HeaderDirectory;
286         else
287           Dependent = CurrentDirectory;
288         sys::path::append(Dependent, DependentsList[Index]);
289       }
290       sys::path::native(Dependent);
291       Dependents.push_back(Dependent.str());
292     }
293     // Save the resulting header file path and dependencies.
294     HeaderFileNames.push_back(HeaderFileName.str());
295     Dependencies[HeaderFileName.str()] = Dependents;
296   }
297 
298   return std::error_code();
299 }
300 
301 // Helper function for finding the input file in an arguments list.
302 std::string findInputFile(const CommandLineArguments &CLArgs) {
303   std::unique_ptr<OptTable> Opts(createDriverOptTable());
304   const unsigned IncludedFlagsBitmask = options::CC1Option;
305   unsigned MissingArgIndex, MissingArgCount;
306   SmallVector<const char *, 256> Argv;
307   for (CommandLineArguments::const_iterator I = CLArgs.begin(),
308                                             E = CLArgs.end();
309        I != E; ++I)
310     Argv.push_back(I->c_str());
311   std::unique_ptr<InputArgList> Args(
312       Opts->ParseArgs(Argv.data(), Argv.data() + Argv.size(), MissingArgIndex,
313                       MissingArgCount, IncludedFlagsBitmask));
314   std::vector<std::string> Inputs = Args->getAllArgValues(OPT_INPUT);
315   return Inputs.back();
316 }
317 
318 // This arguments adjuster inserts "-include (file)" arguments for header
319 // dependencies.
320 ArgumentsAdjuster getAddDependenciesAdjuster(DependencyMap &Dependencies) {
321   return [&Dependencies](const CommandLineArguments &Args) {
322     std::string InputFile = findInputFile(Args);
323     DependentsVector &FileDependents = Dependencies[InputFile];
324     CommandLineArguments NewArgs(Args);
325     if (int Count = FileDependents.size()) {
326       for (int Index = 0; Index < Count; ++Index) {
327         NewArgs.push_back("-include");
328         std::string File(std::string("\"") + FileDependents[Index] +
329                          std::string("\""));
330         NewArgs.push_back(FileDependents[Index]);
331       }
332     }
333     return NewArgs;
334   };
335 }
336 
337 // FIXME: The Location class seems to be something that we might
338 // want to design to be applicable to a wider range of tools, and stick it
339 // somewhere into Tooling/ in mainline
340 struct Location {
341   const FileEntry *File;
342   unsigned Line, Column;
343 
344   Location() : File(), Line(), Column() {}
345 
346   Location(SourceManager &SM, SourceLocation Loc) : File(), Line(), Column() {
347     Loc = SM.getExpansionLoc(Loc);
348     if (Loc.isInvalid())
349       return;
350 
351     std::pair<FileID, unsigned> Decomposed = SM.getDecomposedLoc(Loc);
352     File = SM.getFileEntryForID(Decomposed.first);
353     if (!File)
354       return;
355 
356     Line = SM.getLineNumber(Decomposed.first, Decomposed.second);
357     Column = SM.getColumnNumber(Decomposed.first, Decomposed.second);
358   }
359 
360   operator bool() const { return File != nullptr; }
361 
362   friend bool operator==(const Location &X, const Location &Y) {
363     return X.File == Y.File && X.Line == Y.Line && X.Column == Y.Column;
364   }
365 
366   friend bool operator!=(const Location &X, const Location &Y) {
367     return !(X == Y);
368   }
369 
370   friend bool operator<(const Location &X, const Location &Y) {
371     if (X.File != Y.File)
372       return X.File < Y.File;
373     if (X.Line != Y.Line)
374       return X.Line < Y.Line;
375     return X.Column < Y.Column;
376   }
377   friend bool operator>(const Location &X, const Location &Y) { return Y < X; }
378   friend bool operator<=(const Location &X, const Location &Y) {
379     return !(Y < X);
380   }
381   friend bool operator>=(const Location &X, const Location &Y) {
382     return !(X < Y);
383   }
384 };
385 
386 struct Entry {
387   enum EntryKind {
388     EK_Tag,
389     EK_Value,
390     EK_Macro,
391 
392     EK_NumberOfKinds
393   } Kind;
394 
395   Location Loc;
396 
397   StringRef getKindName() { return getKindName(Kind); }
398   static StringRef getKindName(EntryKind kind);
399 };
400 
401 // Return a string representing the given kind.
402 StringRef Entry::getKindName(Entry::EntryKind kind) {
403   switch (kind) {
404   case EK_Tag:
405     return "tag";
406   case EK_Value:
407     return "value";
408   case EK_Macro:
409     return "macro";
410   case EK_NumberOfKinds:
411     break;
412   }
413   llvm_unreachable("invalid Entry kind");
414 }
415 
416 struct HeaderEntry {
417   std::string Name;
418   Location Loc;
419 
420   friend bool operator==(const HeaderEntry &X, const HeaderEntry &Y) {
421     return X.Loc == Y.Loc && X.Name == Y.Name;
422   }
423   friend bool operator!=(const HeaderEntry &X, const HeaderEntry &Y) {
424     return !(X == Y);
425   }
426   friend bool operator<(const HeaderEntry &X, const HeaderEntry &Y) {
427     return X.Loc < Y.Loc || (X.Loc == Y.Loc && X.Name < Y.Name);
428   }
429   friend bool operator>(const HeaderEntry &X, const HeaderEntry &Y) {
430     return Y < X;
431   }
432   friend bool operator<=(const HeaderEntry &X, const HeaderEntry &Y) {
433     return !(Y < X);
434   }
435   friend bool operator>=(const HeaderEntry &X, const HeaderEntry &Y) {
436     return !(X < Y);
437   }
438 };
439 
440 typedef std::vector<HeaderEntry> HeaderContents;
441 
442 class EntityMap : public StringMap<SmallVector<Entry, 2> > {
443 public:
444   DenseMap<const FileEntry *, HeaderContents> HeaderContentMismatches;
445 
446   void add(const std::string &Name, enum Entry::EntryKind Kind, Location Loc) {
447     // Record this entity in its header.
448     HeaderEntry HE = { Name, Loc };
449     CurHeaderContents[Loc.File].push_back(HE);
450 
451     // Check whether we've seen this entry before.
452     SmallVector<Entry, 2> &Entries = (*this)[Name];
453     for (unsigned I = 0, N = Entries.size(); I != N; ++I) {
454       if (Entries[I].Kind == Kind && Entries[I].Loc == Loc)
455         return;
456     }
457 
458     // We have not seen this entry before; record it.
459     Entry E = { Kind, Loc };
460     Entries.push_back(E);
461   }
462 
463   void mergeCurHeaderContents() {
464     for (DenseMap<const FileEntry *, HeaderContents>::iterator
465              H = CurHeaderContents.begin(),
466              HEnd = CurHeaderContents.end();
467          H != HEnd; ++H) {
468       // Sort contents.
469       std::sort(H->second.begin(), H->second.end());
470 
471       // Check whether we've seen this header before.
472       DenseMap<const FileEntry *, HeaderContents>::iterator KnownH =
473           AllHeaderContents.find(H->first);
474       if (KnownH == AllHeaderContents.end()) {
475         // We haven't seen this header before; record its contents.
476         AllHeaderContents.insert(*H);
477         continue;
478       }
479 
480       // If the header contents are the same, we're done.
481       if (H->second == KnownH->second)
482         continue;
483 
484       // Determine what changed.
485       std::set_symmetric_difference(
486           H->second.begin(), H->second.end(), KnownH->second.begin(),
487           KnownH->second.end(),
488           std::back_inserter(HeaderContentMismatches[H->first]));
489     }
490 
491     CurHeaderContents.clear();
492   }
493 
494 private:
495   DenseMap<const FileEntry *, HeaderContents> CurHeaderContents;
496   DenseMap<const FileEntry *, HeaderContents> AllHeaderContents;
497 };
498 
499 class CollectEntitiesVisitor
500     : public RecursiveASTVisitor<CollectEntitiesVisitor> {
501 public:
502   CollectEntitiesVisitor(SourceManager &SM, EntityMap &Entities,
503                          Preprocessor &PP, PreprocessorTracker &PPTracker,
504                          int &HadErrors)
505       : SM(SM), Entities(Entities), PP(PP), PPTracker(PPTracker),
506         HadErrors(HadErrors) {}
507 
508   bool TraverseStmt(Stmt *S) { return true; }
509   bool TraverseType(QualType T) { return true; }
510   bool TraverseTypeLoc(TypeLoc TL) { return true; }
511   bool TraverseNestedNameSpecifier(NestedNameSpecifier *NNS) { return true; }
512   bool TraverseNestedNameSpecifierLoc(NestedNameSpecifierLoc NNS) {
513     return true;
514   }
515   bool TraverseDeclarationNameInfo(DeclarationNameInfo NameInfo) {
516     return true;
517   }
518   bool TraverseTemplateName(TemplateName Template) { return true; }
519   bool TraverseTemplateArgument(const TemplateArgument &Arg) { return true; }
520   bool TraverseTemplateArgumentLoc(const TemplateArgumentLoc &ArgLoc) {
521     return true;
522   }
523   bool TraverseTemplateArguments(const TemplateArgument *Args,
524                                  unsigned NumArgs) {
525     return true;
526   }
527   bool TraverseConstructorInitializer(CXXCtorInitializer *Init) { return true; }
528   bool TraverseLambdaCapture(LambdaCapture C) { return true; }
529 
530   // Check 'extern "*" {}' block for #include directives.
531   bool VisitLinkageSpecDecl(LinkageSpecDecl *D) {
532     // Bail if not a block.
533     if (!D->hasBraces())
534       return true;
535     SourceRange BlockRange = D->getSourceRange();
536     const char *LinkageLabel;
537     switch (D->getLanguage()) {
538     case LinkageSpecDecl::lang_c:
539       LinkageLabel = "extern \"C\" {}";
540       break;
541     case LinkageSpecDecl::lang_cxx:
542       LinkageLabel = "extern \"C++\" {}";
543       break;
544     }
545     if (!PPTracker.checkForIncludesInBlock(PP, BlockRange, LinkageLabel,
546                                            errs()))
547       HadErrors = 1;
548     return true;
549   }
550 
551   // Check 'namespace (name) {}' block for #include directives.
552   bool VisitNamespaceDecl(const NamespaceDecl *D) {
553     SourceRange BlockRange = D->getSourceRange();
554     std::string Label("namespace ");
555     Label += D->getName();
556     Label += " {}";
557     if (!PPTracker.checkForIncludesInBlock(PP, BlockRange, Label.c_str(),
558                                            errs()))
559       HadErrors = 1;
560     return true;
561   }
562 
563   // Collect definition entities.
564   bool VisitNamedDecl(NamedDecl *ND) {
565     // We only care about file-context variables.
566     if (!ND->getDeclContext()->isFileContext())
567       return true;
568 
569     // Skip declarations that tend to be properly multiply-declared.
570     if (isa<NamespaceDecl>(ND) || isa<UsingDirectiveDecl>(ND) ||
571         isa<NamespaceAliasDecl>(ND) ||
572         isa<ClassTemplateSpecializationDecl>(ND) || isa<UsingDecl>(ND) ||
573         isa<ClassTemplateDecl>(ND) || isa<TemplateTypeParmDecl>(ND) ||
574         isa<TypeAliasTemplateDecl>(ND) || isa<UsingShadowDecl>(ND) ||
575         isa<FunctionDecl>(ND) || isa<FunctionTemplateDecl>(ND) ||
576         (isa<TagDecl>(ND) &&
577          !cast<TagDecl>(ND)->isThisDeclarationADefinition()))
578       return true;
579 
580     // Skip anonymous declarations.
581     if (!ND->getDeclName())
582       return true;
583 
584     // Get the qualified name.
585     std::string Name;
586     llvm::raw_string_ostream OS(Name);
587     ND->printQualifiedName(OS);
588     OS.flush();
589     if (Name.empty())
590       return true;
591 
592     Location Loc(SM, ND->getLocation());
593     if (!Loc)
594       return true;
595 
596     Entities.add(Name, isa<TagDecl>(ND) ? Entry::EK_Tag : Entry::EK_Value, Loc);
597     return true;
598   }
599 
600 private:
601   SourceManager &SM;
602   EntityMap &Entities;
603   Preprocessor &PP;
604   PreprocessorTracker &PPTracker;
605   int &HadErrors;
606 };
607 
608 class CollectEntitiesConsumer : public ASTConsumer {
609 public:
610   CollectEntitiesConsumer(EntityMap &Entities,
611                           PreprocessorTracker &preprocessorTracker,
612                           Preprocessor &PP, StringRef InFile, int &HadErrors)
613       : Entities(Entities), PPTracker(preprocessorTracker), PP(PP),
614         HadErrors(HadErrors) {
615     PPTracker.handlePreprocessorEntry(PP, InFile);
616   }
617 
618   ~CollectEntitiesConsumer() { PPTracker.handlePreprocessorExit(); }
619 
620   virtual void HandleTranslationUnit(ASTContext &Ctx) {
621     SourceManager &SM = Ctx.getSourceManager();
622 
623     // Collect declared entities.
624     CollectEntitiesVisitor(SM, Entities, PP, PPTracker, HadErrors)
625         .TraverseDecl(Ctx.getTranslationUnitDecl());
626 
627     // Collect macro definitions.
628     for (Preprocessor::macro_iterator M = PP.macro_begin(),
629                                       MEnd = PP.macro_end();
630          M != MEnd; ++M) {
631       Location Loc(SM, M->second->getLocation());
632       if (!Loc)
633         continue;
634 
635       Entities.add(M->first->getName().str(), Entry::EK_Macro, Loc);
636     }
637 
638     // Merge header contents.
639     Entities.mergeCurHeaderContents();
640   }
641 
642 private:
643   EntityMap &Entities;
644   PreprocessorTracker &PPTracker;
645   Preprocessor &PP;
646   int &HadErrors;
647 };
648 
649 class CollectEntitiesAction : public SyntaxOnlyAction {
650 public:
651   CollectEntitiesAction(EntityMap &Entities,
652                         PreprocessorTracker &preprocessorTracker,
653                         int &HadErrors)
654       : Entities(Entities), PPTracker(preprocessorTracker),
655         HadErrors(HadErrors) {}
656 
657 protected:
658   std::unique_ptr<clang::ASTConsumer>
659   CreateASTConsumer(CompilerInstance &CI, StringRef InFile) override {
660     return llvm::make_unique<CollectEntitiesConsumer>(
661         Entities, PPTracker, CI.getPreprocessor(), InFile, HadErrors);
662   }
663 
664 private:
665   EntityMap &Entities;
666   PreprocessorTracker &PPTracker;
667   int &HadErrors;
668 };
669 
670 class ModularizeFrontendActionFactory : public FrontendActionFactory {
671 public:
672   ModularizeFrontendActionFactory(EntityMap &Entities,
673                                   PreprocessorTracker &preprocessorTracker,
674                                   int &HadErrors)
675       : Entities(Entities), PPTracker(preprocessorTracker),
676         HadErrors(HadErrors) {}
677 
678   virtual CollectEntitiesAction *create() {
679     return new CollectEntitiesAction(Entities, PPTracker, HadErrors);
680   }
681 
682 private:
683   EntityMap &Entities;
684   PreprocessorTracker &PPTracker;
685   int &HadErrors;
686 };
687 
688 int main(int Argc, const char **Argv) {
689 
690   // Save program name for error messages.
691   Argv0 = Argv[0];
692 
693   // Save program arguments for use in module.modulemap comment.
694   CommandLine = sys::path::stem(sys::path::filename(Argv0));
695   for (int ArgIndex = 1; ArgIndex < Argc; ArgIndex++) {
696     CommandLine.append(" ");
697     CommandLine.append(Argv[ArgIndex]);
698   }
699 
700   // This causes options to be parsed.
701   cl::ParseCommandLineOptions(Argc, Argv, "modularize.\n");
702 
703   // No go if we have no header list file.
704   if (ListFileNames.size() == 0) {
705     cl::PrintHelpMessage();
706     return 1;
707   }
708 
709   // Get header file names and dependencies.
710   SmallVector<std::string, 32> Headers;
711   DependencyMap Dependencies;
712   for (auto I : ListFileNames) {
713     if (std::error_code EC =
714             getHeaderFileNames(Headers, Dependencies, I, HeaderPrefix)) {
715       errs() << Argv[0] << ": error: Unable to get header list '" << I
716              << "': " << EC.message() << '\n';
717       return 1;
718     }
719   }
720 
721   // If we are in assistant mode, output the module map and quit.
722   if (ModuleMapPath.length() != 0) {
723     if (!createModuleMap(ModuleMapPath, Headers, Dependencies, HeaderPrefix,
724                          RootModule))
725       return 1; // Failed.
726     return 0;   // Success - Skip checks in assistant mode.
727   }
728 
729   // Create the compilation database.
730   SmallString<256> PathBuf;
731   sys::fs::current_path(PathBuf);
732   std::unique_ptr<CompilationDatabase> Compilations;
733   Compilations.reset(
734       new FixedCompilationDatabase(Twine(PathBuf), CC1Arguments));
735 
736   // Create preprocessor tracker, to watch for macro and conditional problems.
737   std::unique_ptr<PreprocessorTracker> PPTracker(
738     PreprocessorTracker::create(Headers, BlockCheckHeaderListOnly));
739 
740   // Parse all of the headers, detecting duplicates.
741   EntityMap Entities;
742   ClangTool Tool(*Compilations, Headers);
743   Tool.appendArgumentsAdjuster(getAddDependenciesAdjuster(Dependencies));
744   int HadErrors = 0;
745   ModularizeFrontendActionFactory Factory(Entities, *PPTracker, HadErrors);
746   HadErrors |= Tool.run(&Factory);
747 
748   // Create a place to save duplicate entity locations, separate bins per kind.
749   typedef SmallVector<Location, 8> LocationArray;
750   typedef SmallVector<LocationArray, Entry::EK_NumberOfKinds> EntryBinArray;
751   EntryBinArray EntryBins;
752   int KindIndex;
753   for (KindIndex = 0; KindIndex < Entry::EK_NumberOfKinds; ++KindIndex) {
754     LocationArray Array;
755     EntryBins.push_back(Array);
756   }
757 
758   // Check for the same entity being defined in multiple places.
759   for (EntityMap::iterator E = Entities.begin(), EEnd = Entities.end();
760        E != EEnd; ++E) {
761     // If only one occurrence, exit early.
762     if (E->second.size() == 1)
763       continue;
764     // Clear entity locations.
765     for (EntryBinArray::iterator CI = EntryBins.begin(), CE = EntryBins.end();
766          CI != CE; ++CI) {
767       CI->clear();
768     }
769     // Walk the entities of a single name, collecting the locations,
770     // separated into separate bins.
771     for (unsigned I = 0, N = E->second.size(); I != N; ++I) {
772       EntryBins[E->second[I].Kind].push_back(E->second[I].Loc);
773     }
774     // Report any duplicate entity definition errors.
775     int KindIndex = 0;
776     for (EntryBinArray::iterator DI = EntryBins.begin(), DE = EntryBins.end();
777          DI != DE; ++DI, ++KindIndex) {
778       int ECount = DI->size();
779       // If only 1 occurrence of this entity, skip it, as we only report duplicates.
780       if (ECount <= 1)
781         continue;
782       LocationArray::iterator FI = DI->begin();
783       StringRef kindName = Entry::getKindName((Entry::EntryKind)KindIndex);
784       errs() << "error: " << kindName << " '" << E->first()
785              << "' defined at multiple locations:\n";
786       for (LocationArray::iterator FE = DI->end(); FI != FE; ++FI) {
787         errs() << "    " << FI->File->getName() << ":" << FI->Line << ":"
788                << FI->Column << "\n";
789       }
790       HadErrors = 1;
791     }
792   }
793 
794   // Complain about macro instance in header files that differ based on how
795   // they are included.
796   if (PPTracker->reportInconsistentMacros(errs()))
797     HadErrors = 1;
798 
799   // Complain about preprocessor conditional directives in header files that
800   // differ based on how they are included.
801   if (PPTracker->reportInconsistentConditionals(errs()))
802     HadErrors = 1;
803 
804   // Complain about any headers that have contents that differ based on how
805   // they are included.
806   // FIXME: Could we provide information about which preprocessor conditionals
807   // are involved?
808   for (DenseMap<const FileEntry *, HeaderContents>::iterator
809            H = Entities.HeaderContentMismatches.begin(),
810            HEnd = Entities.HeaderContentMismatches.end();
811        H != HEnd; ++H) {
812     if (H->second.empty()) {
813       errs() << "internal error: phantom header content mismatch\n";
814       continue;
815     }
816 
817     HadErrors = 1;
818     errs() << "error: header '" << H->first->getName()
819            << "' has different contents depending on how it was included.\n";
820     for (unsigned I = 0, N = H->second.size(); I != N; ++I) {
821       errs() << "note: '" << H->second[I].Name << "' in "
822              << H->second[I].Loc.File->getName() << " at "
823              << H->second[I].Loc.Line << ":" << H->second[I].Loc.Column
824              << " not always provided\n";
825     }
826   }
827 
828   return HadErrors;
829 }
830