BugReporter.cpp revision 9b663716449b618ba0390b1dbebc54fa8e971124 
1// BugReporter.cpp - Generate PathDiagnostics for Bugs ------------*- C++ -*--//
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 defines BugReporter, a utility class for generating
11//  PathDiagnostics.
12//
13//===----------------------------------------------------------------------===//
14
15#include "clang/StaticAnalyzer/Core/BugReporter/BugReporter.h"
16#include "clang/StaticAnalyzer/Core/BugReporter/BugType.h"
17#include "clang/StaticAnalyzer/Core/PathSensitive/ExprEngine.h"
18#include "clang/AST/ASTContext.h"
19#include "clang/Analysis/CFG.h"
20#include "clang/AST/Expr.h"
21#include "clang/AST/ParentMap.h"
22#include "clang/AST/StmtObjC.h"
23#include "clang/Basic/SourceManager.h"
24#include "clang/Analysis/ProgramPoint.h"
25#include "clang/StaticAnalyzer/Core/BugReporter/PathDiagnostic.h"
26#include "llvm/Support/raw_ostream.h"
27#include "llvm/ADT/DenseMap.h"
28#include "llvm/ADT/STLExtras.h"
29#include "llvm/ADT/OwningPtr.h"
30#include <queue>
31
32using namespace clang;
33using namespace ento;
34
35BugReporterVisitor::~BugReporterVisitor() {}
36BugReporterContext::~BugReporterContext() {
37  for (visitor_iterator I = visitor_begin(), E = visitor_end(); I != E; ++I)
38    if ((*I)->isOwnedByReporterContext()) delete *I;
39}
40
41void BugReporterContext::addVisitor(BugReporterVisitor* visitor) {
42  if (!visitor)
43    return;
44
45  llvm::FoldingSetNodeID ID;
46  visitor->Profile(ID);
47  void *InsertPos;
48
49  if (CallbacksSet.FindNodeOrInsertPos(ID, InsertPos)) {
50    delete visitor;
51    return;
52  }
53
54  CallbacksSet.InsertNode(visitor, InsertPos);
55  Callbacks = F.add(visitor, Callbacks);
56}
57
58//===----------------------------------------------------------------------===//
59// Helper routines for walking the ExplodedGraph and fetching statements.
60//===----------------------------------------------------------------------===//
61
62static inline const Stmt* GetStmt(const ProgramPoint &P) {
63  if (const StmtPoint* SP = dyn_cast<StmtPoint>(&P))
64    return SP->getStmt();
65  else if (const BlockEdge* BE = dyn_cast<BlockEdge>(&P))
66    return BE->getSrc()->getTerminator();
67
68  return 0;
69}
70
71static inline const ExplodedNode*
72GetPredecessorNode(const ExplodedNode* N) {
73  return N->pred_empty() ? NULL : *(N->pred_begin());
74}
75
76static inline const ExplodedNode*
77GetSuccessorNode(const ExplodedNode* N) {
78  return N->succ_empty() ? NULL : *(N->succ_begin());
79}
80
81static const Stmt* GetPreviousStmt(const ExplodedNode* N) {
82  for (N = GetPredecessorNode(N); N; N = GetPredecessorNode(N))
83    if (const Stmt *S = GetStmt(N->getLocation()))
84      return S;
85
86  return 0;
87}
88
89static const Stmt* GetNextStmt(const ExplodedNode* N) {
90  for (N = GetSuccessorNode(N); N; N = GetSuccessorNode(N))
91    if (const Stmt *S = GetStmt(N->getLocation())) {
92      // Check if the statement is '?' or '&&'/'||'.  These are "merges",
93      // not actual statement points.
94      switch (S->getStmtClass()) {
95        case Stmt::ChooseExprClass:
96        case Stmt::ConditionalOperatorClass: continue;
97        case Stmt::BinaryOperatorClass: {
98          BinaryOperatorKind Op = cast<BinaryOperator>(S)->getOpcode();
99          if (Op == BO_LAnd || Op == BO_LOr)
100            continue;
101          break;
102        }
103        default:
104          break;
105      }
106
107      // Some expressions don't have locations.
108      if (S->getLocStart().isInvalid())
109        continue;
110
111      return S;
112    }
113
114  return 0;
115}
116
117static inline const Stmt*
118GetCurrentOrPreviousStmt(const ExplodedNode* N) {
119  if (const Stmt *S = GetStmt(N->getLocation()))
120    return S;
121
122  return GetPreviousStmt(N);
123}
124
125static inline const Stmt*
126GetCurrentOrNextStmt(const ExplodedNode* N) {
127  if (const Stmt *S = GetStmt(N->getLocation()))
128    return S;
129
130  return GetNextStmt(N);
131}
132
133//===----------------------------------------------------------------------===//
134// PathDiagnosticBuilder and its associated routines and helper objects.
135//===----------------------------------------------------------------------===//
136
137typedef llvm::DenseMap<const ExplodedNode*,
138const ExplodedNode*> NodeBackMap;
139
140namespace {
141class NodeMapClosure : public BugReport::NodeResolver {
142  NodeBackMap& M;
143public:
144  NodeMapClosure(NodeBackMap *m) : M(*m) {}
145  ~NodeMapClosure() {}
146
147  const ExplodedNode* getOriginalNode(const ExplodedNode* N) {
148    NodeBackMap::iterator I = M.find(N);
149    return I == M.end() ? 0 : I->second;
150  }
151};
152
153class PathDiagnosticBuilder : public BugReporterContext {
154  BugReport *R;
155  PathDiagnosticClient *PDC;
156  llvm::OwningPtr<ParentMap> PM;
157  NodeMapClosure NMC;
158public:
159  PathDiagnosticBuilder(GRBugReporter &br,
160                        BugReport *r, NodeBackMap *Backmap,
161                        PathDiagnosticClient *pdc)
162    : BugReporterContext(br),
163      R(r), PDC(pdc), NMC(Backmap) {
164    addVisitor(R);
165  }
166
167  PathDiagnosticLocation ExecutionContinues(const ExplodedNode* N);
168
169  PathDiagnosticLocation ExecutionContinues(llvm::raw_string_ostream& os,
170                                            const ExplodedNode* N);
171
172  Decl const &getCodeDecl() { return R->getErrorNode()->getCodeDecl(); }
173
174  ParentMap& getParentMap() { return R->getErrorNode()->getParentMap(); }
175
176  const Stmt *getParent(const Stmt *S) {
177    return getParentMap().getParent(S);
178  }
179
180  virtual NodeMapClosure& getNodeResolver() { return NMC; }
181
182  PathDiagnosticLocation getEnclosingStmtLocation(const Stmt *S);
183
184  PathDiagnosticClient::PathGenerationScheme getGenerationScheme() const {
185    return PDC ? PDC->getGenerationScheme() : PathDiagnosticClient::Extensive;
186  }
187
188  bool supportsLogicalOpControlFlow() const {
189    return PDC ? PDC->supportsLogicalOpControlFlow() : true;
190  }
191};
192} // end anonymous namespace
193
194PathDiagnosticLocation
195PathDiagnosticBuilder::ExecutionContinues(const ExplodedNode* N) {
196  if (const Stmt *S = GetNextStmt(N))
197    return PathDiagnosticLocation(S, getSourceManager());
198
199  return FullSourceLoc(N->getLocationContext()->getDecl()->getBodyRBrace(),
200                       getSourceManager());
201}
202
203PathDiagnosticLocation
204PathDiagnosticBuilder::ExecutionContinues(llvm::raw_string_ostream& os,
205                                          const ExplodedNode* N) {
206
207  // Slow, but probably doesn't matter.
208  if (os.str().empty())
209    os << ' ';
210
211  const PathDiagnosticLocation &Loc = ExecutionContinues(N);
212
213  if (Loc.asStmt())
214    os << "Execution continues on line "
215       << getSourceManager().getInstantiationLineNumber(Loc.asLocation())
216       << '.';
217  else {
218    os << "Execution jumps to the end of the ";
219    const Decl *D = N->getLocationContext()->getDecl();
220    if (isa<ObjCMethodDecl>(D))
221      os << "method";
222    else if (isa<FunctionDecl>(D))
223      os << "function";
224    else {
225      assert(isa<BlockDecl>(D));
226      os << "anonymous block";
227    }
228    os << '.';
229  }
230
231  return Loc;
232}
233
234static bool IsNested(const Stmt *S, ParentMap &PM) {
235  if (isa<Expr>(S) && PM.isConsumedExpr(cast<Expr>(S)))
236    return true;
237
238  const Stmt *Parent = PM.getParentIgnoreParens(S);
239
240  if (Parent)
241    switch (Parent->getStmtClass()) {
242      case Stmt::ForStmtClass:
243      case Stmt::DoStmtClass:
244      case Stmt::WhileStmtClass:
245        return true;
246      default:
247        break;
248    }
249
250  return false;
251}
252
253PathDiagnosticLocation
254PathDiagnosticBuilder::getEnclosingStmtLocation(const Stmt *S) {
255  assert(S && "Null Stmt* passed to getEnclosingStmtLocation");
256  ParentMap &P = getParentMap();
257  SourceManager &SMgr = getSourceManager();
258
259  while (IsNested(S, P)) {
260    const Stmt *Parent = P.getParentIgnoreParens(S);
261
262    if (!Parent)
263      break;
264
265    switch (Parent->getStmtClass()) {
266      case Stmt::BinaryOperatorClass: {
267        const BinaryOperator *B = cast<BinaryOperator>(Parent);
268        if (B->isLogicalOp())
269          return PathDiagnosticLocation(S, SMgr);
270        break;
271      }
272      case Stmt::CompoundStmtClass:
273      case Stmt::StmtExprClass:
274        return PathDiagnosticLocation(S, SMgr);
275      case Stmt::ChooseExprClass:
276        // Similar to '?' if we are referring to condition, just have the edge
277        // point to the entire choose expression.
278        if (cast<ChooseExpr>(Parent)->getCond() == S)
279          return PathDiagnosticLocation(Parent, SMgr);
280        else
281          return PathDiagnosticLocation(S, SMgr);
282      case Stmt::ConditionalOperatorClass:
283        // For '?', if we are referring to condition, just have the edge point
284        // to the entire '?' expression.
285        if (cast<ConditionalOperator>(Parent)->getCond() == S)
286          return PathDiagnosticLocation(Parent, SMgr);
287        else
288          return PathDiagnosticLocation(S, SMgr);
289      case Stmt::DoStmtClass:
290          return PathDiagnosticLocation(S, SMgr);
291      case Stmt::ForStmtClass:
292        if (cast<ForStmt>(Parent)->getBody() == S)
293          return PathDiagnosticLocation(S, SMgr);
294        break;
295      case Stmt::IfStmtClass:
296        if (cast<IfStmt>(Parent)->getCond() != S)
297          return PathDiagnosticLocation(S, SMgr);
298        break;
299      case Stmt::ObjCForCollectionStmtClass:
300        if (cast<ObjCForCollectionStmt>(Parent)->getBody() == S)
301          return PathDiagnosticLocation(S, SMgr);
302        break;
303      case Stmt::WhileStmtClass:
304        if (cast<WhileStmt>(Parent)->getCond() != S)
305          return PathDiagnosticLocation(S, SMgr);
306        break;
307      default:
308        break;
309    }
310
311    S = Parent;
312  }
313
314  assert(S && "Cannot have null Stmt for PathDiagnosticLocation");
315
316  // Special case: DeclStmts can appear in for statement declarations, in which
317  //  case the ForStmt is the context.
318  if (isa<DeclStmt>(S)) {
319    if (const Stmt *Parent = P.getParent(S)) {
320      switch (Parent->getStmtClass()) {
321        case Stmt::ForStmtClass:
322        case Stmt::ObjCForCollectionStmtClass:
323          return PathDiagnosticLocation(Parent, SMgr);
324        default:
325          break;
326      }
327    }
328  }
329  else if (isa<BinaryOperator>(S)) {
330    // Special case: the binary operator represents the initialization
331    // code in a for statement (this can happen when the variable being
332    // initialized is an old variable.
333    if (const ForStmt *FS =
334          dyn_cast_or_null<ForStmt>(P.getParentIgnoreParens(S))) {
335      if (FS->getInit() == S)
336        return PathDiagnosticLocation(FS, SMgr);
337    }
338  }
339
340  return PathDiagnosticLocation(S, SMgr);
341}
342
343//===----------------------------------------------------------------------===//
344// ScanNotableSymbols: closure-like callback for scanning Store bindings.
345//===----------------------------------------------------------------------===//
346
347static const VarDecl*
348GetMostRecentVarDeclBinding(const ExplodedNode* N,
349                            GRStateManager& VMgr, SVal X) {
350
351  for ( ; N ; N = N->pred_empty() ? 0 : *N->pred_begin()) {
352
353    ProgramPoint P = N->getLocation();
354
355    if (!isa<PostStmt>(P))
356      continue;
357
358    const DeclRefExpr* DR = dyn_cast<DeclRefExpr>(cast<PostStmt>(P).getStmt());
359
360    if (!DR)
361      continue;
362
363    SVal Y = N->getState()->getSVal(DR);
364
365    if (X != Y)
366      continue;
367
368    const VarDecl* VD = dyn_cast<VarDecl>(DR->getDecl());
369
370    if (!VD)
371      continue;
372
373    return VD;
374  }
375
376  return 0;
377}
378
379namespace {
380class NotableSymbolHandler
381: public StoreManager::BindingsHandler {
382
383  SymbolRef Sym;
384  const GRState* PrevSt;
385  const Stmt* S;
386  GRStateManager& VMgr;
387  const ExplodedNode* Pred;
388  PathDiagnostic& PD;
389  BugReporter& BR;
390
391public:
392
393  NotableSymbolHandler(SymbolRef sym, const GRState* prevst, const Stmt* s,
394                       GRStateManager& vmgr, const ExplodedNode* pred,
395                       PathDiagnostic& pd, BugReporter& br)
396  : Sym(sym), PrevSt(prevst), S(s), VMgr(vmgr), Pred(pred), PD(pd), BR(br) {}
397
398  bool HandleBinding(StoreManager& SMgr, Store store, const MemRegion* R,
399                     SVal V) {
400
401    SymbolRef ScanSym = V.getAsSymbol();
402
403    if (ScanSym != Sym)
404      return true;
405
406    // Check if the previous state has this binding.
407    SVal X = PrevSt->getSVal(loc::MemRegionVal(R));
408
409    if (X == V) // Same binding?
410      return true;
411
412    // Different binding.  Only handle assignments for now.  We don't pull
413    // this check out of the loop because we will eventually handle other
414    // cases.
415
416    VarDecl *VD = 0;
417
418    if (const BinaryOperator* B = dyn_cast<BinaryOperator>(S)) {
419      if (!B->isAssignmentOp())
420        return true;
421
422      // What variable did we assign to?
423      DeclRefExpr* DR = dyn_cast<DeclRefExpr>(B->getLHS()->IgnoreParenCasts());
424
425      if (!DR)
426        return true;
427
428      VD = dyn_cast<VarDecl>(DR->getDecl());
429    }
430    else if (const DeclStmt* DS = dyn_cast<DeclStmt>(S)) {
431      // FIXME: Eventually CFGs won't have DeclStmts.  Right now we
432      //  assume that each DeclStmt has a single Decl.  This invariant
433      //  holds by contruction in the CFG.
434      VD = dyn_cast<VarDecl>(*DS->decl_begin());
435    }
436
437    if (!VD)
438      return true;
439
440    // What is the most recently referenced variable with this binding?
441    const VarDecl* MostRecent = GetMostRecentVarDeclBinding(Pred, VMgr, V);
442
443    if (!MostRecent)
444      return true;
445
446    // Create the diagnostic.
447    FullSourceLoc L(S->getLocStart(), BR.getSourceManager());
448
449    if (Loc::IsLocType(VD->getType())) {
450      std::string msg = "'" + std::string(VD->getNameAsString()) +
451      "' now aliases '" + MostRecent->getNameAsString() + "'";
452
453      PD.push_front(new PathDiagnosticEventPiece(L, msg));
454    }
455
456    return true;
457  }
458};
459}
460
461static void HandleNotableSymbol(const ExplodedNode* N,
462                                const Stmt* S,
463                                SymbolRef Sym, BugReporter& BR,
464                                PathDiagnostic& PD) {
465
466  const ExplodedNode* Pred = N->pred_empty() ? 0 : *N->pred_begin();
467  const GRState* PrevSt = Pred ? Pred->getState() : 0;
468
469  if (!PrevSt)
470    return;
471
472  // Look at the region bindings of the current state that map to the
473  // specified symbol.  Are any of them not in the previous state?
474  GRStateManager& VMgr = cast<GRBugReporter>(BR).getStateManager();
475  NotableSymbolHandler H(Sym, PrevSt, S, VMgr, Pred, PD, BR);
476  cast<GRBugReporter>(BR).getStateManager().iterBindings(N->getState(), H);
477}
478
479namespace {
480class ScanNotableSymbols
481: public StoreManager::BindingsHandler {
482
483  llvm::SmallSet<SymbolRef, 10> AlreadyProcessed;
484  const ExplodedNode* N;
485  const Stmt* S;
486  GRBugReporter& BR;
487  PathDiagnostic& PD;
488
489public:
490  ScanNotableSymbols(const ExplodedNode* n, const Stmt* s,
491                     GRBugReporter& br, PathDiagnostic& pd)
492  : N(n), S(s), BR(br), PD(pd) {}
493
494  bool HandleBinding(StoreManager& SMgr, Store store,
495                     const MemRegion* R, SVal V) {
496
497    SymbolRef ScanSym = V.getAsSymbol();
498
499    if (!ScanSym)
500      return true;
501
502    if (!BR.isNotable(ScanSym))
503      return true;
504
505    if (AlreadyProcessed.count(ScanSym))
506      return true;
507
508    AlreadyProcessed.insert(ScanSym);
509
510    HandleNotableSymbol(N, S, ScanSym, BR, PD);
511    return true;
512  }
513};
514} // end anonymous namespace
515
516//===----------------------------------------------------------------------===//
517// "Minimal" path diagnostic generation algorithm.
518//===----------------------------------------------------------------------===//
519
520static void CompactPathDiagnostic(PathDiagnostic &PD, const SourceManager& SM);
521
522static void GenerateMinimalPathDiagnostic(PathDiagnostic& PD,
523                                          PathDiagnosticBuilder &PDB,
524                                          const ExplodedNode *N) {
525
526  SourceManager& SMgr = PDB.getSourceManager();
527  const ExplodedNode* NextNode = N->pred_empty()
528                                        ? NULL : *(N->pred_begin());
529  while (NextNode) {
530    N = NextNode;
531    NextNode = GetPredecessorNode(N);
532
533    ProgramPoint P = N->getLocation();
534
535    if (const BlockEdge* BE = dyn_cast<BlockEdge>(&P)) {
536      const CFGBlock* Src = BE->getSrc();
537      const CFGBlock* Dst = BE->getDst();
538      const Stmt* T = Src->getTerminator();
539
540      if (!T)
541        continue;
542
543      FullSourceLoc Start(T->getLocStart(), SMgr);
544
545      switch (T->getStmtClass()) {
546        default:
547          break;
548
549        case Stmt::GotoStmtClass:
550        case Stmt::IndirectGotoStmtClass: {
551          const Stmt* S = GetNextStmt(N);
552
553          if (!S)
554            continue;
555
556          std::string sbuf;
557          llvm::raw_string_ostream os(sbuf);
558          const PathDiagnosticLocation &End = PDB.getEnclosingStmtLocation(S);
559
560          os << "Control jumps to line "
561          << End.asLocation().getInstantiationLineNumber();
562          PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
563                                                           os.str()));
564          break;
565        }
566
567        case Stmt::SwitchStmtClass: {
568          // Figure out what case arm we took.
569          std::string sbuf;
570          llvm::raw_string_ostream os(sbuf);
571
572          if (const Stmt* S = Dst->getLabel()) {
573            PathDiagnosticLocation End(S, SMgr);
574
575            switch (S->getStmtClass()) {
576              default:
577                os << "No cases match in the switch statement. "
578                "Control jumps to line "
579                << End.asLocation().getInstantiationLineNumber();
580                break;
581              case Stmt::DefaultStmtClass:
582                os << "Control jumps to the 'default' case at line "
583                << End.asLocation().getInstantiationLineNumber();
584                break;
585
586              case Stmt::CaseStmtClass: {
587                os << "Control jumps to 'case ";
588                const CaseStmt* Case = cast<CaseStmt>(S);
589                const Expr* LHS = Case->getLHS()->IgnoreParenCasts();
590
591                // Determine if it is an enum.
592                bool GetRawInt = true;
593
594                if (const DeclRefExpr* DR = dyn_cast<DeclRefExpr>(LHS)) {
595                  // FIXME: Maybe this should be an assertion.  Are there cases
596                  // were it is not an EnumConstantDecl?
597                  const EnumConstantDecl* D =
598                    dyn_cast<EnumConstantDecl>(DR->getDecl());
599
600                  if (D) {
601                    GetRawInt = false;
602                    os << D;
603                  }
604                }
605
606                if (GetRawInt)
607                  os << LHS->EvaluateAsInt(PDB.getASTContext());
608
609                os << ":'  at line "
610                << End.asLocation().getInstantiationLineNumber();
611                break;
612              }
613            }
614            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
615                                                             os.str()));
616          }
617          else {
618            os << "'Default' branch taken. ";
619            const PathDiagnosticLocation &End = PDB.ExecutionContinues(os, N);
620            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
621                                                             os.str()));
622          }
623
624          break;
625        }
626
627        case Stmt::BreakStmtClass:
628        case Stmt::ContinueStmtClass: {
629          std::string sbuf;
630          llvm::raw_string_ostream os(sbuf);
631          PathDiagnosticLocation End = PDB.ExecutionContinues(os, N);
632          PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
633                                                           os.str()));
634          break;
635        }
636
637          // Determine control-flow for ternary '?'.
638        case Stmt::ConditionalOperatorClass: {
639          std::string sbuf;
640          llvm::raw_string_ostream os(sbuf);
641          os << "'?' condition is ";
642
643          if (*(Src->succ_begin()+1) == Dst)
644            os << "false";
645          else
646            os << "true";
647
648          PathDiagnosticLocation End = PDB.ExecutionContinues(N);
649
650          if (const Stmt *S = End.asStmt())
651            End = PDB.getEnclosingStmtLocation(S);
652
653          PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
654                                                           os.str()));
655          break;
656        }
657
658          // Determine control-flow for short-circuited '&&' and '||'.
659        case Stmt::BinaryOperatorClass: {
660          if (!PDB.supportsLogicalOpControlFlow())
661            break;
662
663          const BinaryOperator *B = cast<BinaryOperator>(T);
664          std::string sbuf;
665          llvm::raw_string_ostream os(sbuf);
666          os << "Left side of '";
667
668          if (B->getOpcode() == BO_LAnd) {
669            os << "&&" << "' is ";
670
671            if (*(Src->succ_begin()+1) == Dst) {
672              os << "false";
673              PathDiagnosticLocation End(B->getLHS(), SMgr);
674              PathDiagnosticLocation Start(B->getOperatorLoc(), SMgr);
675              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
676                                                               os.str()));
677            }
678            else {
679              os << "true";
680              PathDiagnosticLocation Start(B->getLHS(), SMgr);
681              PathDiagnosticLocation End = PDB.ExecutionContinues(N);
682              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
683                                                               os.str()));
684            }
685          }
686          else {
687            assert(B->getOpcode() == BO_LOr);
688            os << "||" << "' is ";
689
690            if (*(Src->succ_begin()+1) == Dst) {
691              os << "false";
692              PathDiagnosticLocation Start(B->getLHS(), SMgr);
693              PathDiagnosticLocation End = PDB.ExecutionContinues(N);
694              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
695                                                               os.str()));
696            }
697            else {
698              os << "true";
699              PathDiagnosticLocation End(B->getLHS(), SMgr);
700              PathDiagnosticLocation Start(B->getOperatorLoc(), SMgr);
701              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
702                                                               os.str()));
703            }
704          }
705
706          break;
707        }
708
709        case Stmt::DoStmtClass:  {
710          if (*(Src->succ_begin()) == Dst) {
711            std::string sbuf;
712            llvm::raw_string_ostream os(sbuf);
713
714            os << "Loop condition is true. ";
715            PathDiagnosticLocation End = PDB.ExecutionContinues(os, N);
716
717            if (const Stmt *S = End.asStmt())
718              End = PDB.getEnclosingStmtLocation(S);
719
720            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
721                                                             os.str()));
722          }
723          else {
724            PathDiagnosticLocation End = PDB.ExecutionContinues(N);
725
726            if (const Stmt *S = End.asStmt())
727              End = PDB.getEnclosingStmtLocation(S);
728
729            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
730                              "Loop condition is false.  Exiting loop"));
731          }
732
733          break;
734        }
735
736        case Stmt::WhileStmtClass:
737        case Stmt::ForStmtClass: {
738          if (*(Src->succ_begin()+1) == Dst) {
739            std::string sbuf;
740            llvm::raw_string_ostream os(sbuf);
741
742            os << "Loop condition is false. ";
743            PathDiagnosticLocation End = PDB.ExecutionContinues(os, N);
744            if (const Stmt *S = End.asStmt())
745              End = PDB.getEnclosingStmtLocation(S);
746
747            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
748                                                             os.str()));
749          }
750          else {
751            PathDiagnosticLocation End = PDB.ExecutionContinues(N);
752            if (const Stmt *S = End.asStmt())
753              End = PDB.getEnclosingStmtLocation(S);
754
755            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
756                            "Loop condition is true.  Entering loop body"));
757          }
758
759          break;
760        }
761
762        case Stmt::IfStmtClass: {
763          PathDiagnosticLocation End = PDB.ExecutionContinues(N);
764
765          if (const Stmt *S = End.asStmt())
766            End = PDB.getEnclosingStmtLocation(S);
767
768          if (*(Src->succ_begin()+1) == Dst)
769            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
770                                                        "Taking false branch"));
771          else
772            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
773                                                         "Taking true branch"));
774
775          break;
776        }
777      }
778    }
779
780    if (NextNode) {
781      for (BugReporterContext::visitor_iterator I = PDB.visitor_begin(),
782           E = PDB.visitor_end(); I!=E; ++I) {
783        if (PathDiagnosticPiece* p = (*I)->VisitNode(N, NextNode, PDB))
784          PD.push_front(p);
785      }
786    }
787
788    if (const PostStmt* PS = dyn_cast<PostStmt>(&P)) {
789      // Scan the region bindings, and see if a "notable" symbol has a new
790      // lval binding.
791      ScanNotableSymbols SNS(N, PS->getStmt(), PDB.getBugReporter(), PD);
792      PDB.getStateManager().iterBindings(N->getState(), SNS);
793    }
794  }
795
796  // After constructing the full PathDiagnostic, do a pass over it to compact
797  // PathDiagnosticPieces that occur within a macro.
798  CompactPathDiagnostic(PD, PDB.getSourceManager());
799}
800
801//===----------------------------------------------------------------------===//
802// "Extensive" PathDiagnostic generation.
803//===----------------------------------------------------------------------===//
804
805static bool IsControlFlowExpr(const Stmt *S) {
806  const Expr *E = dyn_cast<Expr>(S);
807
808  if (!E)
809    return false;
810
811  E = E->IgnoreParenCasts();
812
813  if (isa<ConditionalOperator>(E))
814    return true;
815
816  if (const BinaryOperator *B = dyn_cast<BinaryOperator>(E))
817    if (B->isLogicalOp())
818      return true;
819
820  return false;
821}
822
823namespace {
824class ContextLocation : public PathDiagnosticLocation {
825  bool IsDead;
826public:
827  ContextLocation(const PathDiagnosticLocation &L, bool isdead = false)
828    : PathDiagnosticLocation(L), IsDead(isdead) {}
829
830  void markDead() { IsDead = true; }
831  bool isDead() const { return IsDead; }
832};
833
834class EdgeBuilder {
835  std::vector<ContextLocation> CLocs;
836  typedef std::vector<ContextLocation>::iterator iterator;
837  PathDiagnostic &PD;
838  PathDiagnosticBuilder &PDB;
839  PathDiagnosticLocation PrevLoc;
840
841  bool IsConsumedExpr(const PathDiagnosticLocation &L);
842
843  bool containsLocation(const PathDiagnosticLocation &Container,
844                        const PathDiagnosticLocation &Containee);
845
846  PathDiagnosticLocation getContextLocation(const PathDiagnosticLocation &L);
847
848  PathDiagnosticLocation cleanUpLocation(PathDiagnosticLocation L,
849                                         bool firstCharOnly = false) {
850    if (const Stmt *S = L.asStmt()) {
851      const Stmt *Original = S;
852      while (1) {
853        // Adjust the location for some expressions that are best referenced
854        // by one of their subexpressions.
855        switch (S->getStmtClass()) {
856          default:
857            break;
858          case Stmt::ParenExprClass:
859            S = cast<ParenExpr>(S)->IgnoreParens();
860            firstCharOnly = true;
861            continue;
862          case Stmt::ConditionalOperatorClass:
863            S = cast<ConditionalOperator>(S)->getCond();
864            firstCharOnly = true;
865            continue;
866          case Stmt::ChooseExprClass:
867            S = cast<ChooseExpr>(S)->getCond();
868            firstCharOnly = true;
869            continue;
870          case Stmt::BinaryOperatorClass:
871            S = cast<BinaryOperator>(S)->getLHS();
872            firstCharOnly = true;
873            continue;
874        }
875
876        break;
877      }
878
879      if (S != Original)
880        L = PathDiagnosticLocation(S, L.getManager());
881    }
882
883    if (firstCharOnly)
884      L = PathDiagnosticLocation(L.asLocation());
885
886    return L;
887  }
888
889  void popLocation() {
890    if (!CLocs.back().isDead() && CLocs.back().asLocation().isFileID()) {
891      // For contexts, we only one the first character as the range.
892      rawAddEdge(cleanUpLocation(CLocs.back(), true));
893    }
894    CLocs.pop_back();
895  }
896
897public:
898  EdgeBuilder(PathDiagnostic &pd, PathDiagnosticBuilder &pdb)
899    : PD(pd), PDB(pdb) {
900
901      // If the PathDiagnostic already has pieces, add the enclosing statement
902      // of the first piece as a context as well.
903      if (!PD.empty()) {
904        PrevLoc = PD.begin()->getLocation();
905
906        if (const Stmt *S = PrevLoc.asStmt())
907          addExtendedContext(PDB.getEnclosingStmtLocation(S).asStmt());
908      }
909  }
910
911  ~EdgeBuilder() {
912    while (!CLocs.empty()) popLocation();
913
914    // Finally, add an initial edge from the start location of the first
915    // statement (if it doesn't already exist).
916    // FIXME: Should handle CXXTryStmt if analyser starts supporting C++.
917    if (const CompoundStmt *CS =
918          dyn_cast_or_null<CompoundStmt>(PDB.getCodeDecl().getBody()))
919      if (!CS->body_empty()) {
920        SourceLocation Loc = (*CS->body_begin())->getLocStart();
921        rawAddEdge(PathDiagnosticLocation(Loc, PDB.getSourceManager()));
922      }
923
924  }
925
926  void addEdge(PathDiagnosticLocation NewLoc, bool alwaysAdd = false);
927
928  void rawAddEdge(PathDiagnosticLocation NewLoc);
929
930  void addContext(const Stmt *S);
931  void addExtendedContext(const Stmt *S);
932};
933} // end anonymous namespace
934
935
936PathDiagnosticLocation
937EdgeBuilder::getContextLocation(const PathDiagnosticLocation &L) {
938  if (const Stmt *S = L.asStmt()) {
939    if (IsControlFlowExpr(S))
940      return L;
941
942    return PDB.getEnclosingStmtLocation(S);
943  }
944
945  return L;
946}
947
948bool EdgeBuilder::containsLocation(const PathDiagnosticLocation &Container,
949                                   const PathDiagnosticLocation &Containee) {
950
951  if (Container == Containee)
952    return true;
953
954  if (Container.asDecl())
955    return true;
956
957  if (const Stmt *S = Containee.asStmt())
958    if (const Stmt *ContainerS = Container.asStmt()) {
959      while (S) {
960        if (S == ContainerS)
961          return true;
962        S = PDB.getParent(S);
963      }
964      return false;
965    }
966
967  // Less accurate: compare using source ranges.
968  SourceRange ContainerR = Container.asRange();
969  SourceRange ContaineeR = Containee.asRange();
970
971  SourceManager &SM = PDB.getSourceManager();
972  SourceLocation ContainerRBeg = SM.getInstantiationLoc(ContainerR.getBegin());
973  SourceLocation ContainerREnd = SM.getInstantiationLoc(ContainerR.getEnd());
974  SourceLocation ContaineeRBeg = SM.getInstantiationLoc(ContaineeR.getBegin());
975  SourceLocation ContaineeREnd = SM.getInstantiationLoc(ContaineeR.getEnd());
976
977  unsigned ContainerBegLine = SM.getInstantiationLineNumber(ContainerRBeg);
978  unsigned ContainerEndLine = SM.getInstantiationLineNumber(ContainerREnd);
979  unsigned ContaineeBegLine = SM.getInstantiationLineNumber(ContaineeRBeg);
980  unsigned ContaineeEndLine = SM.getInstantiationLineNumber(ContaineeREnd);
981
982  assert(ContainerBegLine <= ContainerEndLine);
983  assert(ContaineeBegLine <= ContaineeEndLine);
984
985  return (ContainerBegLine <= ContaineeBegLine &&
986          ContainerEndLine >= ContaineeEndLine &&
987          (ContainerBegLine != ContaineeBegLine ||
988           SM.getInstantiationColumnNumber(ContainerRBeg) <=
989           SM.getInstantiationColumnNumber(ContaineeRBeg)) &&
990          (ContainerEndLine != ContaineeEndLine ||
991           SM.getInstantiationColumnNumber(ContainerREnd) >=
992           SM.getInstantiationColumnNumber(ContainerREnd)));
993}
994
995void EdgeBuilder::rawAddEdge(PathDiagnosticLocation NewLoc) {
996  if (!PrevLoc.isValid()) {
997    PrevLoc = NewLoc;
998    return;
999  }
1000
1001  const PathDiagnosticLocation &NewLocClean = cleanUpLocation(NewLoc);
1002  const PathDiagnosticLocation &PrevLocClean = cleanUpLocation(PrevLoc);
1003
1004  if (NewLocClean.asLocation() == PrevLocClean.asLocation())
1005    return;
1006
1007  // FIXME: Ignore intra-macro edges for now.
1008  if (NewLocClean.asLocation().getInstantiationLoc() ==
1009      PrevLocClean.asLocation().getInstantiationLoc())
1010    return;
1011
1012  PD.push_front(new PathDiagnosticControlFlowPiece(NewLocClean, PrevLocClean));
1013  PrevLoc = NewLoc;
1014}
1015
1016void EdgeBuilder::addEdge(PathDiagnosticLocation NewLoc, bool alwaysAdd) {
1017
1018  if (!alwaysAdd && NewLoc.asLocation().isMacroID())
1019    return;
1020
1021  const PathDiagnosticLocation &CLoc = getContextLocation(NewLoc);
1022
1023  while (!CLocs.empty()) {
1024    ContextLocation &TopContextLoc = CLocs.back();
1025
1026    // Is the top location context the same as the one for the new location?
1027    if (TopContextLoc == CLoc) {
1028      if (alwaysAdd) {
1029        if (IsConsumedExpr(TopContextLoc) &&
1030            !IsControlFlowExpr(TopContextLoc.asStmt()))
1031            TopContextLoc.markDead();
1032
1033        rawAddEdge(NewLoc);
1034      }
1035
1036      return;
1037    }
1038
1039    if (containsLocation(TopContextLoc, CLoc)) {
1040      if (alwaysAdd) {
1041        rawAddEdge(NewLoc);
1042
1043        if (IsConsumedExpr(CLoc) && !IsControlFlowExpr(CLoc.asStmt())) {
1044          CLocs.push_back(ContextLocation(CLoc, true));
1045          return;
1046        }
1047      }
1048
1049      CLocs.push_back(CLoc);
1050      return;
1051    }
1052
1053    // Context does not contain the location.  Flush it.
1054    popLocation();
1055  }
1056
1057  // If we reach here, there is no enclosing context.  Just add the edge.
1058  rawAddEdge(NewLoc);
1059}
1060
1061bool EdgeBuilder::IsConsumedExpr(const PathDiagnosticLocation &L) {
1062  if (const Expr *X = dyn_cast_or_null<Expr>(L.asStmt()))
1063    return PDB.getParentMap().isConsumedExpr(X) && !IsControlFlowExpr(X);
1064
1065  return false;
1066}
1067
1068void EdgeBuilder::addExtendedContext(const Stmt *S) {
1069  if (!S)
1070    return;
1071
1072  const Stmt *Parent = PDB.getParent(S);
1073  while (Parent) {
1074    if (isa<CompoundStmt>(Parent))
1075      Parent = PDB.getParent(Parent);
1076    else
1077      break;
1078  }
1079
1080  if (Parent) {
1081    switch (Parent->getStmtClass()) {
1082      case Stmt::DoStmtClass:
1083      case Stmt::ObjCAtSynchronizedStmtClass:
1084        addContext(Parent);
1085      default:
1086        break;
1087    }
1088  }
1089
1090  addContext(S);
1091}
1092
1093void EdgeBuilder::addContext(const Stmt *S) {
1094  if (!S)
1095    return;
1096
1097  PathDiagnosticLocation L(S, PDB.getSourceManager());
1098
1099  while (!CLocs.empty()) {
1100    const PathDiagnosticLocation &TopContextLoc = CLocs.back();
1101
1102    // Is the top location context the same as the one for the new location?
1103    if (TopContextLoc == L)
1104      return;
1105
1106    if (containsLocation(TopContextLoc, L)) {
1107      CLocs.push_back(L);
1108      return;
1109    }
1110
1111    // Context does not contain the location.  Flush it.
1112    popLocation();
1113  }
1114
1115  CLocs.push_back(L);
1116}
1117
1118static void GenerateExtensivePathDiagnostic(PathDiagnostic& PD,
1119                                            PathDiagnosticBuilder &PDB,
1120                                            const ExplodedNode *N) {
1121  EdgeBuilder EB(PD, PDB);
1122
1123  const ExplodedNode* NextNode = N->pred_empty() ? NULL : *(N->pred_begin());
1124  while (NextNode) {
1125    N = NextNode;
1126    NextNode = GetPredecessorNode(N);
1127    ProgramPoint P = N->getLocation();
1128
1129    do {
1130      // Block edges.
1131      if (const BlockEdge *BE = dyn_cast<BlockEdge>(&P)) {
1132        const CFGBlock &Blk = *BE->getSrc();
1133        const Stmt *Term = Blk.getTerminator();
1134
1135        // Are we jumping to the head of a loop?  Add a special diagnostic.
1136        if (const Stmt *Loop = BE->getDst()->getLoopTarget()) {
1137          PathDiagnosticLocation L(Loop, PDB.getSourceManager());
1138          const CompoundStmt *CS = NULL;
1139
1140          if (!Term) {
1141            if (const ForStmt *FS = dyn_cast<ForStmt>(Loop))
1142              CS = dyn_cast<CompoundStmt>(FS->getBody());
1143            else if (const WhileStmt *WS = dyn_cast<WhileStmt>(Loop))
1144              CS = dyn_cast<CompoundStmt>(WS->getBody());
1145          }
1146
1147          PathDiagnosticEventPiece *p =
1148            new PathDiagnosticEventPiece(L,
1149                                        "Looping back to the head of the loop");
1150
1151          EB.addEdge(p->getLocation(), true);
1152          PD.push_front(p);
1153
1154          if (CS) {
1155            PathDiagnosticLocation BL(CS->getRBracLoc(),
1156                                      PDB.getSourceManager());
1157            BL = PathDiagnosticLocation(BL.asLocation());
1158            EB.addEdge(BL);
1159          }
1160        }
1161
1162        if (Term)
1163          EB.addContext(Term);
1164
1165        break;
1166      }
1167
1168      if (const BlockEntrance *BE = dyn_cast<BlockEntrance>(&P)) {
1169        if (CFGStmt S = BE->getFirstElement().getAs<CFGStmt>()) {
1170          if (IsControlFlowExpr(S)) {
1171            // Add the proper context for '&&', '||', and '?'.
1172            EB.addContext(S);
1173          }
1174          else
1175            EB.addExtendedContext(PDB.getEnclosingStmtLocation(S).asStmt());
1176        }
1177
1178        break;
1179      }
1180    } while (0);
1181
1182    if (!NextNode)
1183      continue;
1184
1185    for (BugReporterContext::visitor_iterator I = PDB.visitor_begin(),
1186         E = PDB.visitor_end(); I!=E; ++I) {
1187      if (PathDiagnosticPiece* p = (*I)->VisitNode(N, NextNode, PDB)) {
1188        const PathDiagnosticLocation &Loc = p->getLocation();
1189        EB.addEdge(Loc, true);
1190        PD.push_front(p);
1191        if (const Stmt *S = Loc.asStmt())
1192          EB.addExtendedContext(PDB.getEnclosingStmtLocation(S).asStmt());
1193      }
1194    }
1195  }
1196}
1197
1198//===----------------------------------------------------------------------===//
1199// Methods for BugType and subclasses.
1200//===----------------------------------------------------------------------===//
1201BugType::~BugType() {
1202  // Free up the equivalence class objects.  Observe that we get a pointer to
1203  // the object first before incrementing the iterator, as destroying the
1204  // node before doing so means we will read from freed memory.
1205  for (iterator I = begin(), E = end(); I !=E; ) {
1206    BugReportEquivClass *EQ = &*I;
1207    ++I;
1208    delete EQ;
1209  }
1210}
1211void BugType::FlushReports(BugReporter &BR) {}
1212
1213//===----------------------------------------------------------------------===//
1214// Methods for BugReport and subclasses.
1215//===----------------------------------------------------------------------===//
1216BugReport::~BugReport() {}
1217RangedBugReport::~RangedBugReport() {}
1218
1219const Stmt* BugReport::getStmt() const {
1220  ProgramPoint ProgP = ErrorNode->getLocation();
1221  const Stmt *S = NULL;
1222
1223  if (BlockEntrance* BE = dyn_cast<BlockEntrance>(&ProgP)) {
1224    CFGBlock &Exit = ProgP.getLocationContext()->getCFG()->getExit();
1225    if (BE->getBlock() == &Exit)
1226      S = GetPreviousStmt(ErrorNode);
1227  }
1228  if (!S)
1229    S = GetStmt(ProgP);
1230
1231  return S;
1232}
1233
1234PathDiagnosticPiece*
1235BugReport::getEndPath(BugReporterContext& BRC,
1236                      const ExplodedNode* EndPathNode) {
1237
1238  const Stmt* S = getStmt();
1239
1240  if (!S)
1241    return NULL;
1242
1243  BugReport::ranges_iterator Beg, End;
1244  llvm::tie(Beg, End) = getRanges();
1245  PathDiagnosticLocation L(S, BRC.getSourceManager());
1246
1247  // Only add the statement itself as a range if we didn't specify any
1248  // special ranges for this report.
1249  PathDiagnosticPiece* P = new PathDiagnosticEventPiece(L, getDescription(),
1250                                                        Beg == End);
1251
1252  for (; Beg != End; ++Beg)
1253    P->addRange(*Beg);
1254
1255  return P;
1256}
1257
1258std::pair<BugReport::ranges_iterator, BugReport::ranges_iterator>
1259BugReport::getRanges() const {
1260  if (const Expr* E = dyn_cast_or_null<Expr>(getStmt())) {
1261    R = E->getSourceRange();
1262    assert(R.isValid());
1263    return std::make_pair(&R, &R+1);
1264  }
1265  else
1266    return std::make_pair(ranges_iterator(), ranges_iterator());
1267}
1268
1269SourceLocation BugReport::getLocation() const {
1270  if (ErrorNode)
1271    if (const Stmt* S = GetCurrentOrPreviousStmt(ErrorNode)) {
1272      // For member expressions, return the location of the '.' or '->'.
1273      if (const MemberExpr *ME = dyn_cast<MemberExpr>(S))
1274        return ME->getMemberLoc();
1275      // For binary operators, return the location of the operator.
1276      if (const BinaryOperator *B = dyn_cast<BinaryOperator>(S))
1277        return B->getOperatorLoc();
1278
1279      return S->getLocStart();
1280    }
1281
1282  return FullSourceLoc();
1283}
1284
1285PathDiagnosticPiece* BugReport::VisitNode(const ExplodedNode* N,
1286                                          const ExplodedNode* PrevN,
1287                                          BugReporterContext &BRC) {
1288  return NULL;
1289}
1290
1291//===----------------------------------------------------------------------===//
1292// Methods for BugReporter and subclasses.
1293//===----------------------------------------------------------------------===//
1294
1295BugReportEquivClass::~BugReportEquivClass() {
1296  for (iterator I=begin(), E=end(); I!=E; ++I) delete *I;
1297}
1298
1299GRBugReporter::~GRBugReporter() { }
1300BugReporterData::~BugReporterData() {}
1301
1302ExplodedGraph &GRBugReporter::getGraph() { return Eng.getGraph(); }
1303
1304GRStateManager&
1305GRBugReporter::getStateManager() { return Eng.getStateManager(); }
1306
1307BugReporter::~BugReporter() { FlushReports(); }
1308
1309void BugReporter::FlushReports() {
1310  if (BugTypes.isEmpty())
1311    return;
1312
1313  // First flush the warnings for each BugType.  This may end up creating new
1314  // warnings and new BugTypes.  Because ImmutableSet is a functional data
1315  // structure, we do not need to worry about the iterators being invalidated.
1316  for (BugTypesTy::iterator I=BugTypes.begin(), E=BugTypes.end(); I!=E; ++I)
1317    const_cast<BugType*>(*I)->FlushReports(*this);
1318
1319  // Iterate through BugTypes a second time.  BugTypes may have been updated
1320  // with new BugType objects and new warnings.
1321  for (BugTypesTy::iterator I=BugTypes.begin(), E=BugTypes.end(); I!=E; ++I) {
1322    BugType *BT = const_cast<BugType*>(*I);
1323
1324    typedef llvm::FoldingSet<BugReportEquivClass> SetTy;
1325    SetTy& EQClasses = BT->EQClasses;
1326
1327    for (SetTy::iterator EI=EQClasses.begin(), EE=EQClasses.end(); EI!=EE;++EI){
1328      BugReportEquivClass& EQ = *EI;
1329      FlushReport(EQ);
1330    }
1331
1332    // Delete the BugType object.
1333    delete BT;
1334  }
1335
1336  // Remove all references to the BugType objects.
1337  BugTypes = F.getEmptySet();
1338}
1339
1340//===----------------------------------------------------------------------===//
1341// PathDiagnostics generation.
1342//===----------------------------------------------------------------------===//
1343
1344static std::pair<std::pair<ExplodedGraph*, NodeBackMap*>,
1345                 std::pair<ExplodedNode*, unsigned> >
1346MakeReportGraph(const ExplodedGraph* G,
1347                llvm::SmallVectorImpl<const ExplodedNode*> &nodes) {
1348
1349  // Create the trimmed graph.  It will contain the shortest paths from the
1350  // error nodes to the root.  In the new graph we should only have one
1351  // error node unless there are two or more error nodes with the same minimum
1352  // path length.
1353  ExplodedGraph* GTrim;
1354  InterExplodedGraphMap* NMap;
1355
1356  llvm::DenseMap<const void*, const void*> InverseMap;
1357  llvm::tie(GTrim, NMap) = G->Trim(nodes.data(), nodes.data() + nodes.size(),
1358                                   &InverseMap);
1359
1360  // Create owning pointers for GTrim and NMap just to ensure that they are
1361  // released when this function exists.
1362  llvm::OwningPtr<ExplodedGraph> AutoReleaseGTrim(GTrim);
1363  llvm::OwningPtr<InterExplodedGraphMap> AutoReleaseNMap(NMap);
1364
1365  // Find the (first) error node in the trimmed graph.  We just need to consult
1366  // the node map (NMap) which maps from nodes in the original graph to nodes
1367  // in the new graph.
1368
1369  std::queue<const ExplodedNode*> WS;
1370  typedef llvm::DenseMap<const ExplodedNode*, unsigned> IndexMapTy;
1371  IndexMapTy IndexMap;
1372
1373  for (unsigned nodeIndex = 0 ; nodeIndex < nodes.size(); ++nodeIndex) {
1374    const ExplodedNode *originalNode = nodes[nodeIndex];
1375    if (const ExplodedNode *N = NMap->getMappedNode(originalNode)) {
1376      WS.push(N);
1377      IndexMap[originalNode] = nodeIndex;
1378    }
1379  }
1380
1381  assert(!WS.empty() && "No error node found in the trimmed graph.");
1382
1383  // Create a new (third!) graph with a single path.  This is the graph
1384  // that will be returned to the caller.
1385  ExplodedGraph *GNew = new ExplodedGraph();
1386
1387  // Sometimes the trimmed graph can contain a cycle.  Perform a reverse BFS
1388  // to the root node, and then construct a new graph that contains only
1389  // a single path.
1390  llvm::DenseMap<const void*,unsigned> Visited;
1391
1392  unsigned cnt = 0;
1393  const ExplodedNode* Root = 0;
1394
1395  while (!WS.empty()) {
1396    const ExplodedNode* Node = WS.front();
1397    WS.pop();
1398
1399    if (Visited.find(Node) != Visited.end())
1400      continue;
1401
1402    Visited[Node] = cnt++;
1403
1404    if (Node->pred_empty()) {
1405      Root = Node;
1406      break;
1407    }
1408
1409    for (ExplodedNode::const_pred_iterator I=Node->pred_begin(),
1410         E=Node->pred_end(); I!=E; ++I)
1411      WS.push(*I);
1412  }
1413
1414  assert(Root);
1415
1416  // Now walk from the root down the BFS path, always taking the successor
1417  // with the lowest number.
1418  ExplodedNode *Last = 0, *First = 0;
1419  NodeBackMap *BM = new NodeBackMap();
1420  unsigned NodeIndex = 0;
1421
1422  for ( const ExplodedNode *N = Root ;;) {
1423    // Lookup the number associated with the current node.
1424    llvm::DenseMap<const void*,unsigned>::iterator I = Visited.find(N);
1425    assert(I != Visited.end());
1426
1427    // Create the equivalent node in the new graph with the same state
1428    // and location.
1429    ExplodedNode* NewN = GNew->getNode(N->getLocation(), N->getState());
1430
1431    // Store the mapping to the original node.
1432    llvm::DenseMap<const void*, const void*>::iterator IMitr=InverseMap.find(N);
1433    assert(IMitr != InverseMap.end() && "No mapping to original node.");
1434    (*BM)[NewN] = (const ExplodedNode*) IMitr->second;
1435
1436    // Link up the new node with the previous node.
1437    if (Last)
1438      NewN->addPredecessor(Last, *GNew);
1439
1440    Last = NewN;
1441
1442    // Are we at the final node?
1443    IndexMapTy::iterator IMI =
1444      IndexMap.find((const ExplodedNode*)(IMitr->second));
1445    if (IMI != IndexMap.end()) {
1446      First = NewN;
1447      NodeIndex = IMI->second;
1448      break;
1449    }
1450
1451    // Find the next successor node.  We choose the node that is marked
1452    // with the lowest DFS number.
1453    ExplodedNode::const_succ_iterator SI = N->succ_begin();
1454    ExplodedNode::const_succ_iterator SE = N->succ_end();
1455    N = 0;
1456
1457    for (unsigned MinVal = 0; SI != SE; ++SI) {
1458
1459      I = Visited.find(*SI);
1460
1461      if (I == Visited.end())
1462        continue;
1463
1464      if (!N || I->second < MinVal) {
1465        N = *SI;
1466        MinVal = I->second;
1467      }
1468    }
1469
1470    assert(N);
1471  }
1472
1473  assert(First);
1474
1475  return std::make_pair(std::make_pair(GNew, BM),
1476                        std::make_pair(First, NodeIndex));
1477}
1478
1479/// CompactPathDiagnostic - This function postprocesses a PathDiagnostic object
1480///  and collapses PathDiagosticPieces that are expanded by macros.
1481static void CompactPathDiagnostic(PathDiagnostic &PD, const SourceManager& SM) {
1482  typedef std::vector<std::pair<PathDiagnosticMacroPiece*, SourceLocation> >
1483          MacroStackTy;
1484
1485  typedef std::vector<PathDiagnosticPiece*>
1486          PiecesTy;
1487
1488  MacroStackTy MacroStack;
1489  PiecesTy Pieces;
1490
1491  for (PathDiagnostic::iterator I = PD.begin(), E = PD.end(); I!=E; ++I) {
1492    // Get the location of the PathDiagnosticPiece.
1493    const FullSourceLoc Loc = I->getLocation().asLocation();
1494
1495    // Determine the instantiation location, which is the location we group
1496    // related PathDiagnosticPieces.
1497    SourceLocation InstantiationLoc = Loc.isMacroID() ?
1498                                      SM.getInstantiationLoc(Loc) :
1499                                      SourceLocation();
1500
1501    if (Loc.isFileID()) {
1502      MacroStack.clear();
1503      Pieces.push_back(&*I);
1504      continue;
1505    }
1506
1507    assert(Loc.isMacroID());
1508
1509    // Is the PathDiagnosticPiece within the same macro group?
1510    if (!MacroStack.empty() && InstantiationLoc == MacroStack.back().second) {
1511      MacroStack.back().first->push_back(&*I);
1512      continue;
1513    }
1514
1515    // We aren't in the same group.  Are we descending into a new macro
1516    // or are part of an old one?
1517    PathDiagnosticMacroPiece *MacroGroup = 0;
1518
1519    SourceLocation ParentInstantiationLoc = InstantiationLoc.isMacroID() ?
1520                                          SM.getInstantiationLoc(Loc) :
1521                                          SourceLocation();
1522
1523    // Walk the entire macro stack.
1524    while (!MacroStack.empty()) {
1525      if (InstantiationLoc == MacroStack.back().second) {
1526        MacroGroup = MacroStack.back().first;
1527        break;
1528      }
1529
1530      if (ParentInstantiationLoc == MacroStack.back().second) {
1531        MacroGroup = MacroStack.back().first;
1532        break;
1533      }
1534
1535      MacroStack.pop_back();
1536    }
1537
1538    if (!MacroGroup || ParentInstantiationLoc == MacroStack.back().second) {
1539      // Create a new macro group and add it to the stack.
1540      PathDiagnosticMacroPiece *NewGroup = new PathDiagnosticMacroPiece(Loc);
1541
1542      if (MacroGroup)
1543        MacroGroup->push_back(NewGroup);
1544      else {
1545        assert(InstantiationLoc.isFileID());
1546        Pieces.push_back(NewGroup);
1547      }
1548
1549      MacroGroup = NewGroup;
1550      MacroStack.push_back(std::make_pair(MacroGroup, InstantiationLoc));
1551    }
1552
1553    // Finally, add the PathDiagnosticPiece to the group.
1554    MacroGroup->push_back(&*I);
1555  }
1556
1557  // Now take the pieces and construct a new PathDiagnostic.
1558  PD.resetPath(false);
1559
1560  for (PiecesTy::iterator I=Pieces.begin(), E=Pieces.end(); I!=E; ++I) {
1561    if (PathDiagnosticMacroPiece *MP=dyn_cast<PathDiagnosticMacroPiece>(*I))
1562      if (!MP->containsEvent()) {
1563        delete MP;
1564        continue;
1565      }
1566
1567    PD.push_back(*I);
1568  }
1569}
1570
1571void GRBugReporter::GeneratePathDiagnostic(PathDiagnostic& PD,
1572                        llvm::SmallVectorImpl<BugReport *> &bugReports) {
1573
1574  assert(!bugReports.empty());
1575  llvm::SmallVector<const ExplodedNode *, 10> errorNodes;
1576  for (llvm::SmallVectorImpl<BugReport*>::iterator I = bugReports.begin(),
1577    E = bugReports.end(); I != E; ++I) {
1578      errorNodes.push_back((*I)->getErrorNode());
1579  }
1580
1581  // Construct a new graph that contains only a single path from the error
1582  // node to a root.
1583  const std::pair<std::pair<ExplodedGraph*, NodeBackMap*>,
1584  std::pair<ExplodedNode*, unsigned> >&
1585    GPair = MakeReportGraph(&getGraph(), errorNodes);
1586
1587  // Find the BugReport with the original location.
1588  assert(GPair.second.second < bugReports.size());
1589  BugReport *R = bugReports[GPair.second.second];
1590  assert(R && "No original report found for sliced graph.");
1591
1592  llvm::OwningPtr<ExplodedGraph> ReportGraph(GPair.first.first);
1593  llvm::OwningPtr<NodeBackMap> BackMap(GPair.first.second);
1594  const ExplodedNode *N = GPair.second.first;
1595
1596  // Start building the path diagnostic...
1597  PathDiagnosticBuilder PDB(*this, R, BackMap.get(), getPathDiagnosticClient());
1598
1599  if (PathDiagnosticPiece* Piece = R->getEndPath(PDB, N))
1600    PD.push_back(Piece);
1601  else
1602    return;
1603
1604  // Register node visitors.
1605  R->registerInitialVisitors(PDB, N);
1606  bugreporter::registerNilReceiverVisitor(PDB);
1607
1608  switch (PDB.getGenerationScheme()) {
1609    case PathDiagnosticClient::Extensive:
1610      GenerateExtensivePathDiagnostic(PD, PDB, N);
1611      break;
1612    case PathDiagnosticClient::Minimal:
1613      GenerateMinimalPathDiagnostic(PD, PDB, N);
1614      break;
1615  }
1616}
1617
1618void BugReporter::Register(BugType *BT) {
1619  BugTypes = F.add(BugTypes, BT);
1620}
1621
1622void BugReporter::EmitReport(BugReport* R) {
1623  // Compute the bug report's hash to determine its equivalence class.
1624  llvm::FoldingSetNodeID ID;
1625  R->Profile(ID);
1626
1627  // Lookup the equivance class.  If there isn't one, create it.
1628  BugType& BT = R->getBugType();
1629  Register(&BT);
1630  void *InsertPos;
1631  BugReportEquivClass* EQ = BT.EQClasses.FindNodeOrInsertPos(ID, InsertPos);
1632
1633  if (!EQ) {
1634    EQ = new BugReportEquivClass(R);
1635    BT.EQClasses.InsertNode(EQ, InsertPos);
1636  }
1637  else
1638    EQ->AddReport(R);
1639}
1640
1641
1642//===----------------------------------------------------------------------===//
1643// Emitting reports in equivalence classes.
1644//===----------------------------------------------------------------------===//
1645
1646namespace {
1647struct FRIEC_WLItem {
1648  const ExplodedNode *N;
1649  ExplodedNode::const_succ_iterator I, E;
1650
1651  FRIEC_WLItem(const ExplodedNode *n)
1652  : N(n), I(N->succ_begin()), E(N->succ_end()) {}
1653};
1654}
1655
1656static BugReport *
1657FindReportInEquivalenceClass(BugReportEquivClass& EQ,
1658                             llvm::SmallVectorImpl<BugReport*> &bugReports) {
1659
1660  BugReportEquivClass::iterator I = EQ.begin(), E = EQ.end();
1661  assert(I != E);
1662  BugReport *R = *I;
1663  BugType& BT = R->getBugType();
1664
1665  // If we don't need to suppress any of the nodes because they are
1666  // post-dominated by a sink, simply add all the nodes in the equivalence class
1667  // to 'Nodes'.  Any of the reports will serve as a "representative" report.
1668  if (!BT.isSuppressOnSink()) {
1669    for (BugReportEquivClass::iterator I=EQ.begin(), E=EQ.end(); I!=E; ++I) {
1670      const ExplodedNode* N = I->getErrorNode();
1671      if (N) {
1672        R = *I;
1673        bugReports.push_back(R);
1674      }
1675    }
1676    return R;
1677  }
1678
1679  // For bug reports that should be suppressed when all paths are post-dominated
1680  // by a sink node, iterate through the reports in the equivalence class
1681  // until we find one that isn't post-dominated (if one exists).  We use a
1682  // DFS traversal of the ExplodedGraph to find a non-sink node.  We could write
1683  // this as a recursive function, but we don't want to risk blowing out the
1684  // stack for very long paths.
1685  BugReport *exampleReport = 0;
1686
1687  for (; I != E; ++I) {
1688    R = *I;
1689    const ExplodedNode *errorNode = R->getErrorNode();
1690
1691    if (!errorNode)
1692      continue;
1693    if (errorNode->isSink()) {
1694      assert(false &&
1695           "BugType::isSuppressSink() should not be 'true' for sink end nodes");
1696      return 0;
1697    }
1698    // No successors?  By definition this nodes isn't post-dominated by a sink.
1699    if (errorNode->succ_empty()) {
1700      bugReports.push_back(R);
1701      if (!exampleReport)
1702        exampleReport = R;
1703      continue;
1704    }
1705
1706    // At this point we know that 'N' is not a sink and it has at least one
1707    // successor.  Use a DFS worklist to find a non-sink end-of-path node.
1708    typedef FRIEC_WLItem WLItem;
1709    typedef llvm::SmallVector<WLItem, 10> DFSWorkList;
1710    llvm::DenseMap<const ExplodedNode *, unsigned> Visited;
1711
1712    DFSWorkList WL;
1713    WL.push_back(errorNode);
1714    Visited[errorNode] = 1;
1715
1716    while (!WL.empty()) {
1717      WLItem &WI = WL.back();
1718      assert(!WI.N->succ_empty());
1719
1720      for (; WI.I != WI.E; ++WI.I) {
1721        const ExplodedNode *Succ = *WI.I;
1722        // End-of-path node?
1723        if (Succ->succ_empty()) {
1724          // If we found an end-of-path node that is not a sink.
1725          if (!Succ->isSink()) {
1726            bugReports.push_back(R);
1727            if (!exampleReport)
1728              exampleReport = R;
1729            WL.clear();
1730            break;
1731          }
1732          // Found a sink?  Continue on to the next successor.
1733          continue;
1734        }
1735        // Mark the successor as visited.  If it hasn't been explored,
1736        // enqueue it to the DFS worklist.
1737        unsigned &mark = Visited[Succ];
1738        if (!mark) {
1739          mark = 1;
1740          WL.push_back(Succ);
1741          break;
1742        }
1743      }
1744
1745      // The worklist may have been cleared at this point.  First
1746      // check if it is empty before checking the last item.
1747      if (!WL.empty() && &WL.back() == &WI)
1748        WL.pop_back();
1749    }
1750  }
1751
1752  // ExampleReport will be NULL if all the nodes in the equivalence class
1753  // were post-dominated by sinks.
1754  return exampleReport;
1755}
1756
1757//===----------------------------------------------------------------------===//
1758// DiagnosticCache.  This is a hack to cache analyzer diagnostics.  It
1759// uses global state, which eventually should go elsewhere.
1760//===----------------------------------------------------------------------===//
1761namespace {
1762class DiagCacheItem : public llvm::FoldingSetNode {
1763  llvm::FoldingSetNodeID ID;
1764public:
1765  DiagCacheItem(BugReport *R, PathDiagnostic *PD) {
1766    ID.AddString(R->getBugType().getName());
1767    ID.AddString(R->getBugType().getCategory());
1768    ID.AddString(R->getDescription());
1769    ID.AddInteger(R->getLocation().getRawEncoding());
1770    PD->Profile(ID);
1771  }
1772
1773  void Profile(llvm::FoldingSetNodeID &id) {
1774    id = ID;
1775  }
1776
1777  llvm::FoldingSetNodeID &getID() { return ID; }
1778};
1779}
1780
1781static bool IsCachedDiagnostic(BugReport *R, PathDiagnostic *PD) {
1782  // FIXME: Eventually this diagnostic cache should reside in something
1783  // like AnalysisManager instead of being a static variable.  This is
1784  // really unsafe in the long term.
1785  typedef llvm::FoldingSet<DiagCacheItem> DiagnosticCache;
1786  static DiagnosticCache DC;
1787
1788  void *InsertPos;
1789  DiagCacheItem *Item = new DiagCacheItem(R, PD);
1790
1791  if (DC.FindNodeOrInsertPos(Item->getID(), InsertPos)) {
1792    delete Item;
1793    return true;
1794  }
1795
1796  DC.InsertNode(Item, InsertPos);
1797  return false;
1798}
1799
1800void BugReporter::FlushReport(BugReportEquivClass& EQ) {
1801  llvm::SmallVector<BugReport*, 10> bugReports;
1802  BugReport *exampleReport = FindReportInEquivalenceClass(EQ, bugReports);
1803  if (!exampleReport)
1804    return;
1805
1806  PathDiagnosticClient* PD = getPathDiagnosticClient();
1807
1808  // FIXME: Make sure we use the 'R' for the path that was actually used.
1809  // Probably doesn't make a difference in practice.
1810  BugType& BT = exampleReport->getBugType();
1811
1812  llvm::OwningPtr<PathDiagnostic>
1813    D(new PathDiagnostic(exampleReport->getBugType().getName(),
1814                         !PD || PD->useVerboseDescription()
1815                         ? exampleReport->getDescription()
1816                         : exampleReport->getShortDescription(),
1817                         BT.getCategory()));
1818
1819  if (!bugReports.empty())
1820    GeneratePathDiagnostic(*D.get(), bugReports);
1821
1822  if (IsCachedDiagnostic(exampleReport, D.get()))
1823    return;
1824
1825  // Get the meta data.
1826  std::pair<const char**, const char**> Meta =
1827    exampleReport->getExtraDescriptiveText();
1828  for (const char** s = Meta.first; s != Meta.second; ++s)
1829    D->addMeta(*s);
1830
1831  // Emit a summary diagnostic to the regular Diagnostics engine.
1832  BugReport::ranges_iterator Beg, End;
1833  llvm::tie(Beg, End) = exampleReport->getRanges();
1834  Diagnostic &Diag = getDiagnostic();
1835  FullSourceLoc L(exampleReport->getLocation(), getSourceManager());
1836
1837  // Search the description for '%', as that will be interpretted as a
1838  // format character by FormatDiagnostics.
1839  llvm::StringRef desc = exampleReport->getShortDescription();
1840  unsigned ErrorDiag;
1841  {
1842    llvm::SmallString<512> TmpStr;
1843    llvm::raw_svector_ostream Out(TmpStr);
1844    for (llvm::StringRef::iterator I=desc.begin(), E=desc.end(); I!=E; ++I)
1845      if (*I == '%')
1846        Out << "%%";
1847      else
1848        Out << *I;
1849
1850    Out.flush();
1851    ErrorDiag = Diag.getCustomDiagID(Diagnostic::Warning, TmpStr);
1852  }
1853
1854  {
1855    DiagnosticBuilder diagBuilder = Diag.Report(L, ErrorDiag);
1856    for (BugReport::ranges_iterator I = Beg; I != End; ++I)
1857      diagBuilder << *I;
1858  }
1859
1860  // Emit a full diagnostic for the path if we have a PathDiagnosticClient.
1861  if (!PD)
1862    return;
1863
1864  if (D->empty()) {
1865    PathDiagnosticPiece* piece =
1866      new PathDiagnosticEventPiece(L, exampleReport->getDescription());
1867
1868    for ( ; Beg != End; ++Beg) piece->addRange(*Beg);
1869    D->push_back(piece);
1870  }
1871
1872  PD->HandlePathDiagnostic(D.take());
1873}
1874
1875void BugReporter::EmitBasicReport(llvm::StringRef name, llvm::StringRef str,
1876                                  SourceLocation Loc,
1877                                  SourceRange* RBeg, unsigned NumRanges) {
1878  EmitBasicReport(name, "", str, Loc, RBeg, NumRanges);
1879}
1880
1881void BugReporter::EmitBasicReport(llvm::StringRef name,
1882                                  llvm::StringRef category,
1883                                  llvm::StringRef str, SourceLocation Loc,
1884                                  SourceRange* RBeg, unsigned NumRanges) {
1885
1886  // 'BT' will be owned by BugReporter as soon as we call 'EmitReport'.
1887  BugType *BT = new BugType(name, category);
1888  FullSourceLoc L = getContext().getFullLoc(Loc);
1889  RangedBugReport *R = new DiagBugReport(*BT, str, L);
1890  for ( ; NumRanges > 0 ; --NumRanges, ++RBeg) R->addRange(*RBeg);
1891  EmitReport(R);
1892}
1893