BugReporter.cpp revision 56ca35d396d8692c384c785f9aeebcf22563fe1e 
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::BinaryConditionalOperatorClass: continue;
97        case Stmt::ConditionalOperatorClass: continue;
98        case Stmt::BinaryOperatorClass: {
99          BinaryOperatorKind Op = cast<BinaryOperator>(S)->getOpcode();
100          if (Op == BO_LAnd || Op == BO_LOr)
101            continue;
102          break;
103        }
104        default:
105          break;
106      }
107
108      // Some expressions don't have locations.
109      if (S->getLocStart().isInvalid())
110        continue;
111
112      return S;
113    }
114
115  return 0;
116}
117
118static inline const Stmt*
119GetCurrentOrPreviousStmt(const ExplodedNode* N) {
120  if (const Stmt *S = GetStmt(N->getLocation()))
121    return S;
122
123  return GetPreviousStmt(N);
124}
125
126static inline const Stmt*
127GetCurrentOrNextStmt(const ExplodedNode* N) {
128  if (const Stmt *S = GetStmt(N->getLocation()))
129    return S;
130
131  return GetNextStmt(N);
132}
133
134//===----------------------------------------------------------------------===//
135// PathDiagnosticBuilder and its associated routines and helper objects.
136//===----------------------------------------------------------------------===//
137
138typedef llvm::DenseMap<const ExplodedNode*,
139const ExplodedNode*> NodeBackMap;
140
141namespace {
142class NodeMapClosure : public BugReport::NodeResolver {
143  NodeBackMap& M;
144public:
145  NodeMapClosure(NodeBackMap *m) : M(*m) {}
146  ~NodeMapClosure() {}
147
148  const ExplodedNode* getOriginalNode(const ExplodedNode* N) {
149    NodeBackMap::iterator I = M.find(N);
150    return I == M.end() ? 0 : I->second;
151  }
152};
153
154class PathDiagnosticBuilder : public BugReporterContext {
155  BugReport *R;
156  PathDiagnosticClient *PDC;
157  llvm::OwningPtr<ParentMap> PM;
158  NodeMapClosure NMC;
159public:
160  PathDiagnosticBuilder(GRBugReporter &br,
161                        BugReport *r, NodeBackMap *Backmap,
162                        PathDiagnosticClient *pdc)
163    : BugReporterContext(br),
164      R(r), PDC(pdc), NMC(Backmap) {
165    addVisitor(R);
166  }
167
168  PathDiagnosticLocation ExecutionContinues(const ExplodedNode* N);
169
170  PathDiagnosticLocation ExecutionContinues(llvm::raw_string_ostream& os,
171                                            const ExplodedNode* N);
172
173  Decl const &getCodeDecl() { return R->getErrorNode()->getCodeDecl(); }
174
175  ParentMap& getParentMap() { return R->getErrorNode()->getParentMap(); }
176
177  const Stmt *getParent(const Stmt *S) {
178    return getParentMap().getParent(S);
179  }
180
181  virtual NodeMapClosure& getNodeResolver() { return NMC; }
182
183  PathDiagnosticLocation getEnclosingStmtLocation(const Stmt *S);
184
185  PathDiagnosticClient::PathGenerationScheme getGenerationScheme() const {
186    return PDC ? PDC->getGenerationScheme() : PathDiagnosticClient::Extensive;
187  }
188
189  bool supportsLogicalOpControlFlow() const {
190    return PDC ? PDC->supportsLogicalOpControlFlow() : true;
191  }
192};
193} // end anonymous namespace
194
195PathDiagnosticLocation
196PathDiagnosticBuilder::ExecutionContinues(const ExplodedNode* N) {
197  if (const Stmt *S = GetNextStmt(N))
198    return PathDiagnosticLocation(S, getSourceManager());
199
200  return FullSourceLoc(N->getLocationContext()->getDecl()->getBodyRBrace(),
201                       getSourceManager());
202}
203
204PathDiagnosticLocation
205PathDiagnosticBuilder::ExecutionContinues(llvm::raw_string_ostream& os,
206                                          const ExplodedNode* N) {
207
208  // Slow, but probably doesn't matter.
209  if (os.str().empty())
210    os << ' ';
211
212  const PathDiagnosticLocation &Loc = ExecutionContinues(N);
213
214  if (Loc.asStmt())
215    os << "Execution continues on line "
216       << getSourceManager().getInstantiationLineNumber(Loc.asLocation())
217       << '.';
218  else {
219    os << "Execution jumps to the end of the ";
220    const Decl *D = N->getLocationContext()->getDecl();
221    if (isa<ObjCMethodDecl>(D))
222      os << "method";
223    else if (isa<FunctionDecl>(D))
224      os << "function";
225    else {
226      assert(isa<BlockDecl>(D));
227      os << "anonymous block";
228    }
229    os << '.';
230  }
231
232  return Loc;
233}
234
235static bool IsNested(const Stmt *S, ParentMap &PM) {
236  if (isa<Expr>(S) && PM.isConsumedExpr(cast<Expr>(S)))
237    return true;
238
239  const Stmt *Parent = PM.getParentIgnoreParens(S);
240
241  if (Parent)
242    switch (Parent->getStmtClass()) {
243      case Stmt::ForStmtClass:
244      case Stmt::DoStmtClass:
245      case Stmt::WhileStmtClass:
246        return true;
247      default:
248        break;
249    }
250
251  return false;
252}
253
254PathDiagnosticLocation
255PathDiagnosticBuilder::getEnclosingStmtLocation(const Stmt *S) {
256  assert(S && "Null Stmt* passed to getEnclosingStmtLocation");
257  ParentMap &P = getParentMap();
258  SourceManager &SMgr = getSourceManager();
259
260  while (IsNested(S, P)) {
261    const Stmt *Parent = P.getParentIgnoreParens(S);
262
263    if (!Parent)
264      break;
265
266    switch (Parent->getStmtClass()) {
267      case Stmt::BinaryOperatorClass: {
268        const BinaryOperator *B = cast<BinaryOperator>(Parent);
269        if (B->isLogicalOp())
270          return PathDiagnosticLocation(S, SMgr);
271        break;
272      }
273      case Stmt::CompoundStmtClass:
274      case Stmt::StmtExprClass:
275        return PathDiagnosticLocation(S, SMgr);
276      case Stmt::ChooseExprClass:
277        // Similar to '?' if we are referring to condition, just have the edge
278        // point to the entire choose expression.
279        if (cast<ChooseExpr>(Parent)->getCond() == S)
280          return PathDiagnosticLocation(Parent, SMgr);
281        else
282          return PathDiagnosticLocation(S, SMgr);
283      case Stmt::BinaryConditionalOperatorClass:
284      case Stmt::ConditionalOperatorClass:
285        // For '?', if we are referring to condition, just have the edge point
286        // to the entire '?' expression.
287        if (cast<AbstractConditionalOperator>(Parent)->getCond() == S)
288          return PathDiagnosticLocation(Parent, SMgr);
289        else
290          return PathDiagnosticLocation(S, SMgr);
291      case Stmt::DoStmtClass:
292          return PathDiagnosticLocation(S, SMgr);
293      case Stmt::ForStmtClass:
294        if (cast<ForStmt>(Parent)->getBody() == S)
295          return PathDiagnosticLocation(S, SMgr);
296        break;
297      case Stmt::IfStmtClass:
298        if (cast<IfStmt>(Parent)->getCond() != S)
299          return PathDiagnosticLocation(S, SMgr);
300        break;
301      case Stmt::ObjCForCollectionStmtClass:
302        if (cast<ObjCForCollectionStmt>(Parent)->getBody() == S)
303          return PathDiagnosticLocation(S, SMgr);
304        break;
305      case Stmt::WhileStmtClass:
306        if (cast<WhileStmt>(Parent)->getCond() != S)
307          return PathDiagnosticLocation(S, SMgr);
308        break;
309      default:
310        break;
311    }
312
313    S = Parent;
314  }
315
316  assert(S && "Cannot have null Stmt for PathDiagnosticLocation");
317
318  // Special case: DeclStmts can appear in for statement declarations, in which
319  //  case the ForStmt is the context.
320  if (isa<DeclStmt>(S)) {
321    if (const Stmt *Parent = P.getParent(S)) {
322      switch (Parent->getStmtClass()) {
323        case Stmt::ForStmtClass:
324        case Stmt::ObjCForCollectionStmtClass:
325          return PathDiagnosticLocation(Parent, SMgr);
326        default:
327          break;
328      }
329    }
330  }
331  else if (isa<BinaryOperator>(S)) {
332    // Special case: the binary operator represents the initialization
333    // code in a for statement (this can happen when the variable being
334    // initialized is an old variable.
335    if (const ForStmt *FS =
336          dyn_cast_or_null<ForStmt>(P.getParentIgnoreParens(S))) {
337      if (FS->getInit() == S)
338        return PathDiagnosticLocation(FS, SMgr);
339    }
340  }
341
342  return PathDiagnosticLocation(S, SMgr);
343}
344
345//===----------------------------------------------------------------------===//
346// ScanNotableSymbols: closure-like callback for scanning Store bindings.
347//===----------------------------------------------------------------------===//
348
349static const VarDecl*
350GetMostRecentVarDeclBinding(const ExplodedNode* N,
351                            GRStateManager& VMgr, SVal X) {
352
353  for ( ; N ; N = N->pred_empty() ? 0 : *N->pred_begin()) {
354
355    ProgramPoint P = N->getLocation();
356
357    if (!isa<PostStmt>(P))
358      continue;
359
360    const DeclRefExpr* DR = dyn_cast<DeclRefExpr>(cast<PostStmt>(P).getStmt());
361
362    if (!DR)
363      continue;
364
365    SVal Y = N->getState()->getSVal(DR);
366
367    if (X != Y)
368      continue;
369
370    const VarDecl* VD = dyn_cast<VarDecl>(DR->getDecl());
371
372    if (!VD)
373      continue;
374
375    return VD;
376  }
377
378  return 0;
379}
380
381namespace {
382class NotableSymbolHandler
383: public StoreManager::BindingsHandler {
384
385  SymbolRef Sym;
386  const GRState* PrevSt;
387  const Stmt* S;
388  GRStateManager& VMgr;
389  const ExplodedNode* Pred;
390  PathDiagnostic& PD;
391  BugReporter& BR;
392
393public:
394
395  NotableSymbolHandler(SymbolRef sym, const GRState* prevst, const Stmt* s,
396                       GRStateManager& vmgr, const ExplodedNode* pred,
397                       PathDiagnostic& pd, BugReporter& br)
398  : Sym(sym), PrevSt(prevst), S(s), VMgr(vmgr), Pred(pred), PD(pd), BR(br) {}
399
400  bool HandleBinding(StoreManager& SMgr, Store store, const MemRegion* R,
401                     SVal V) {
402
403    SymbolRef ScanSym = V.getAsSymbol();
404
405    if (ScanSym != Sym)
406      return true;
407
408    // Check if the previous state has this binding.
409    SVal X = PrevSt->getSVal(loc::MemRegionVal(R));
410
411    if (X == V) // Same binding?
412      return true;
413
414    // Different binding.  Only handle assignments for now.  We don't pull
415    // this check out of the loop because we will eventually handle other
416    // cases.
417
418    VarDecl *VD = 0;
419
420    if (const BinaryOperator* B = dyn_cast<BinaryOperator>(S)) {
421      if (!B->isAssignmentOp())
422        return true;
423
424      // What variable did we assign to?
425      DeclRefExpr* DR = dyn_cast<DeclRefExpr>(B->getLHS()->IgnoreParenCasts());
426
427      if (!DR)
428        return true;
429
430      VD = dyn_cast<VarDecl>(DR->getDecl());
431    }
432    else if (const DeclStmt* DS = dyn_cast<DeclStmt>(S)) {
433      // FIXME: Eventually CFGs won't have DeclStmts.  Right now we
434      //  assume that each DeclStmt has a single Decl.  This invariant
435      //  holds by contruction in the CFG.
436      VD = dyn_cast<VarDecl>(*DS->decl_begin());
437    }
438
439    if (!VD)
440      return true;
441
442    // What is the most recently referenced variable with this binding?
443    const VarDecl* MostRecent = GetMostRecentVarDeclBinding(Pred, VMgr, V);
444
445    if (!MostRecent)
446      return true;
447
448    // Create the diagnostic.
449    FullSourceLoc L(S->getLocStart(), BR.getSourceManager());
450
451    if (Loc::isLocType(VD->getType())) {
452      std::string msg = "'" + std::string(VD->getNameAsString()) +
453      "' now aliases '" + MostRecent->getNameAsString() + "'";
454
455      PD.push_front(new PathDiagnosticEventPiece(L, msg));
456    }
457
458    return true;
459  }
460};
461}
462
463static void HandleNotableSymbol(const ExplodedNode* N,
464                                const Stmt* S,
465                                SymbolRef Sym, BugReporter& BR,
466                                PathDiagnostic& PD) {
467
468  const ExplodedNode* Pred = N->pred_empty() ? 0 : *N->pred_begin();
469  const GRState* PrevSt = Pred ? Pred->getState() : 0;
470
471  if (!PrevSt)
472    return;
473
474  // Look at the region bindings of the current state that map to the
475  // specified symbol.  Are any of them not in the previous state?
476  GRStateManager& VMgr = cast<GRBugReporter>(BR).getStateManager();
477  NotableSymbolHandler H(Sym, PrevSt, S, VMgr, Pred, PD, BR);
478  cast<GRBugReporter>(BR).getStateManager().iterBindings(N->getState(), H);
479}
480
481namespace {
482class ScanNotableSymbols
483: public StoreManager::BindingsHandler {
484
485  llvm::SmallSet<SymbolRef, 10> AlreadyProcessed;
486  const ExplodedNode* N;
487  const Stmt* S;
488  GRBugReporter& BR;
489  PathDiagnostic& PD;
490
491public:
492  ScanNotableSymbols(const ExplodedNode* n, const Stmt* s,
493                     GRBugReporter& br, PathDiagnostic& pd)
494  : N(n), S(s), BR(br), PD(pd) {}
495
496  bool HandleBinding(StoreManager& SMgr, Store store,
497                     const MemRegion* R, SVal V) {
498
499    SymbolRef ScanSym = V.getAsSymbol();
500
501    if (!ScanSym)
502      return true;
503
504    if (!BR.isNotable(ScanSym))
505      return true;
506
507    if (AlreadyProcessed.count(ScanSym))
508      return true;
509
510    AlreadyProcessed.insert(ScanSym);
511
512    HandleNotableSymbol(N, S, ScanSym, BR, PD);
513    return true;
514  }
515};
516} // end anonymous namespace
517
518//===----------------------------------------------------------------------===//
519// "Minimal" path diagnostic generation algorithm.
520//===----------------------------------------------------------------------===//
521
522static void CompactPathDiagnostic(PathDiagnostic &PD, const SourceManager& SM);
523
524static void GenerateMinimalPathDiagnostic(PathDiagnostic& PD,
525                                          PathDiagnosticBuilder &PDB,
526                                          const ExplodedNode *N) {
527
528  SourceManager& SMgr = PDB.getSourceManager();
529  const ExplodedNode* NextNode = N->pred_empty()
530                                        ? NULL : *(N->pred_begin());
531  while (NextNode) {
532    N = NextNode;
533    NextNode = GetPredecessorNode(N);
534
535    ProgramPoint P = N->getLocation();
536
537    if (const BlockEdge* BE = dyn_cast<BlockEdge>(&P)) {
538      const CFGBlock* Src = BE->getSrc();
539      const CFGBlock* Dst = BE->getDst();
540      const Stmt* T = Src->getTerminator();
541
542      if (!T)
543        continue;
544
545      FullSourceLoc Start(T->getLocStart(), SMgr);
546
547      switch (T->getStmtClass()) {
548        default:
549          break;
550
551        case Stmt::GotoStmtClass:
552        case Stmt::IndirectGotoStmtClass: {
553          const Stmt* S = GetNextStmt(N);
554
555          if (!S)
556            continue;
557
558          std::string sbuf;
559          llvm::raw_string_ostream os(sbuf);
560          const PathDiagnosticLocation &End = PDB.getEnclosingStmtLocation(S);
561
562          os << "Control jumps to line "
563          << End.asLocation().getInstantiationLineNumber();
564          PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
565                                                           os.str()));
566          break;
567        }
568
569        case Stmt::SwitchStmtClass: {
570          // Figure out what case arm we took.
571          std::string sbuf;
572          llvm::raw_string_ostream os(sbuf);
573
574          if (const Stmt* S = Dst->getLabel()) {
575            PathDiagnosticLocation End(S, SMgr);
576
577            switch (S->getStmtClass()) {
578              default:
579                os << "No cases match in the switch statement. "
580                "Control jumps to line "
581                << End.asLocation().getInstantiationLineNumber();
582                break;
583              case Stmt::DefaultStmtClass:
584                os << "Control jumps to the 'default' case at line "
585                << End.asLocation().getInstantiationLineNumber();
586                break;
587
588              case Stmt::CaseStmtClass: {
589                os << "Control jumps to 'case ";
590                const CaseStmt* Case = cast<CaseStmt>(S);
591                const Expr* LHS = Case->getLHS()->IgnoreParenCasts();
592
593                // Determine if it is an enum.
594                bool GetRawInt = true;
595
596                if (const DeclRefExpr* DR = dyn_cast<DeclRefExpr>(LHS)) {
597                  // FIXME: Maybe this should be an assertion.  Are there cases
598                  // were it is not an EnumConstantDecl?
599                  const EnumConstantDecl* D =
600                    dyn_cast<EnumConstantDecl>(DR->getDecl());
601
602                  if (D) {
603                    GetRawInt = false;
604                    os << D;
605                  }
606                }
607
608                if (GetRawInt)
609                  os << LHS->EvaluateAsInt(PDB.getASTContext());
610
611                os << ":'  at line "
612                << End.asLocation().getInstantiationLineNumber();
613                break;
614              }
615            }
616            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
617                                                             os.str()));
618          }
619          else {
620            os << "'Default' branch taken. ";
621            const PathDiagnosticLocation &End = PDB.ExecutionContinues(os, N);
622            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
623                                                             os.str()));
624          }
625
626          break;
627        }
628
629        case Stmt::BreakStmtClass:
630        case Stmt::ContinueStmtClass: {
631          std::string sbuf;
632          llvm::raw_string_ostream os(sbuf);
633          PathDiagnosticLocation End = PDB.ExecutionContinues(os, N);
634          PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
635                                                           os.str()));
636          break;
637        }
638
639          // Determine control-flow for ternary '?'.
640        case Stmt::BinaryConditionalOperatorClass:
641        case Stmt::ConditionalOperatorClass: {
642          std::string sbuf;
643          llvm::raw_string_ostream os(sbuf);
644          os << "'?' condition is ";
645
646          if (*(Src->succ_begin()+1) == Dst)
647            os << "false";
648          else
649            os << "true";
650
651          PathDiagnosticLocation End = PDB.ExecutionContinues(N);
652
653          if (const Stmt *S = End.asStmt())
654            End = PDB.getEnclosingStmtLocation(S);
655
656          PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
657                                                           os.str()));
658          break;
659        }
660
661          // Determine control-flow for short-circuited '&&' and '||'.
662        case Stmt::BinaryOperatorClass: {
663          if (!PDB.supportsLogicalOpControlFlow())
664            break;
665
666          const BinaryOperator *B = cast<BinaryOperator>(T);
667          std::string sbuf;
668          llvm::raw_string_ostream os(sbuf);
669          os << "Left side of '";
670
671          if (B->getOpcode() == BO_LAnd) {
672            os << "&&" << "' is ";
673
674            if (*(Src->succ_begin()+1) == Dst) {
675              os << "false";
676              PathDiagnosticLocation End(B->getLHS(), SMgr);
677              PathDiagnosticLocation Start(B->getOperatorLoc(), SMgr);
678              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
679                                                               os.str()));
680            }
681            else {
682              os << "true";
683              PathDiagnosticLocation Start(B->getLHS(), SMgr);
684              PathDiagnosticLocation End = PDB.ExecutionContinues(N);
685              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
686                                                               os.str()));
687            }
688          }
689          else {
690            assert(B->getOpcode() == BO_LOr);
691            os << "||" << "' is ";
692
693            if (*(Src->succ_begin()+1) == Dst) {
694              os << "false";
695              PathDiagnosticLocation Start(B->getLHS(), SMgr);
696              PathDiagnosticLocation End = PDB.ExecutionContinues(N);
697              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
698                                                               os.str()));
699            }
700            else {
701              os << "true";
702              PathDiagnosticLocation End(B->getLHS(), SMgr);
703              PathDiagnosticLocation Start(B->getOperatorLoc(), SMgr);
704              PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
705                                                               os.str()));
706            }
707          }
708
709          break;
710        }
711
712        case Stmt::DoStmtClass:  {
713          if (*(Src->succ_begin()) == Dst) {
714            std::string sbuf;
715            llvm::raw_string_ostream os(sbuf);
716
717            os << "Loop condition is true. ";
718            PathDiagnosticLocation End = PDB.ExecutionContinues(os, N);
719
720            if (const Stmt *S = End.asStmt())
721              End = PDB.getEnclosingStmtLocation(S);
722
723            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
724                                                             os.str()));
725          }
726          else {
727            PathDiagnosticLocation End = PDB.ExecutionContinues(N);
728
729            if (const Stmt *S = End.asStmt())
730              End = PDB.getEnclosingStmtLocation(S);
731
732            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
733                              "Loop condition is false.  Exiting loop"));
734          }
735
736          break;
737        }
738
739        case Stmt::WhileStmtClass:
740        case Stmt::ForStmtClass: {
741          if (*(Src->succ_begin()+1) == Dst) {
742            std::string sbuf;
743            llvm::raw_string_ostream os(sbuf);
744
745            os << "Loop condition is false. ";
746            PathDiagnosticLocation End = PDB.ExecutionContinues(os, N);
747            if (const Stmt *S = End.asStmt())
748              End = PDB.getEnclosingStmtLocation(S);
749
750            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
751                                                             os.str()));
752          }
753          else {
754            PathDiagnosticLocation End = PDB.ExecutionContinues(N);
755            if (const Stmt *S = End.asStmt())
756              End = PDB.getEnclosingStmtLocation(S);
757
758            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
759                            "Loop condition is true.  Entering loop body"));
760          }
761
762          break;
763        }
764
765        case Stmt::IfStmtClass: {
766          PathDiagnosticLocation End = PDB.ExecutionContinues(N);
767
768          if (const Stmt *S = End.asStmt())
769            End = PDB.getEnclosingStmtLocation(S);
770
771          if (*(Src->succ_begin()+1) == Dst)
772            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
773                                                        "Taking false branch"));
774          else
775            PD.push_front(new PathDiagnosticControlFlowPiece(Start, End,
776                                                         "Taking true branch"));
777
778          break;
779        }
780      }
781    }
782
783    if (NextNode) {
784      for (BugReporterContext::visitor_iterator I = PDB.visitor_begin(),
785           E = PDB.visitor_end(); I!=E; ++I) {
786        if (PathDiagnosticPiece* p = (*I)->VisitNode(N, NextNode, PDB))
787          PD.push_front(p);
788      }
789    }
790
791    if (const PostStmt* PS = dyn_cast<PostStmt>(&P)) {
792      // Scan the region bindings, and see if a "notable" symbol has a new
793      // lval binding.
794      ScanNotableSymbols SNS(N, PS->getStmt(), PDB.getBugReporter(), PD);
795      PDB.getStateManager().iterBindings(N->getState(), SNS);
796    }
797  }
798
799  // After constructing the full PathDiagnostic, do a pass over it to compact
800  // PathDiagnosticPieces that occur within a macro.
801  CompactPathDiagnostic(PD, PDB.getSourceManager());
802}
803
804//===----------------------------------------------------------------------===//
805// "Extensive" PathDiagnostic generation.
806//===----------------------------------------------------------------------===//
807
808static bool IsControlFlowExpr(const Stmt *S) {
809  const Expr *E = dyn_cast<Expr>(S);
810
811  if (!E)
812    return false;
813
814  E = E->IgnoreParenCasts();
815
816  if (isa<AbstractConditionalOperator>(E))
817    return true;
818
819  if (const BinaryOperator *B = dyn_cast<BinaryOperator>(E))
820    if (B->isLogicalOp())
821      return true;
822
823  return false;
824}
825
826namespace {
827class ContextLocation : public PathDiagnosticLocation {
828  bool IsDead;
829public:
830  ContextLocation(const PathDiagnosticLocation &L, bool isdead = false)
831    : PathDiagnosticLocation(L), IsDead(isdead) {}
832
833  void markDead() { IsDead = true; }
834  bool isDead() const { return IsDead; }
835};
836
837class EdgeBuilder {
838  std::vector<ContextLocation> CLocs;
839  typedef std::vector<ContextLocation>::iterator iterator;
840  PathDiagnostic &PD;
841  PathDiagnosticBuilder &PDB;
842  PathDiagnosticLocation PrevLoc;
843
844  bool IsConsumedExpr(const PathDiagnosticLocation &L);
845
846  bool containsLocation(const PathDiagnosticLocation &Container,
847                        const PathDiagnosticLocation &Containee);
848
849  PathDiagnosticLocation getContextLocation(const PathDiagnosticLocation &L);
850
851  PathDiagnosticLocation cleanUpLocation(PathDiagnosticLocation L,
852                                         bool firstCharOnly = false) {
853    if (const Stmt *S = L.asStmt()) {
854      const Stmt *Original = S;
855      while (1) {
856        // Adjust the location for some expressions that are best referenced
857        // by one of their subexpressions.
858        switch (S->getStmtClass()) {
859          default:
860            break;
861          case Stmt::ParenExprClass:
862            S = cast<ParenExpr>(S)->IgnoreParens();
863            firstCharOnly = true;
864            continue;
865          case Stmt::BinaryConditionalOperatorClass:
866          case Stmt::ConditionalOperatorClass:
867            S = cast<AbstractConditionalOperator>(S)->getCond();
868            firstCharOnly = true;
869            continue;
870          case Stmt::ChooseExprClass:
871            S = cast<ChooseExpr>(S)->getCond();
872            firstCharOnly = true;
873            continue;
874          case Stmt::BinaryOperatorClass:
875            S = cast<BinaryOperator>(S)->getLHS();
876            firstCharOnly = true;
877            continue;
878        }
879
880        break;
881      }
882
883      if (S != Original)
884        L = PathDiagnosticLocation(S, L.getManager());
885    }
886
887    if (firstCharOnly)
888      L = PathDiagnosticLocation(L.asLocation());
889
890    return L;
891  }
892
893  void popLocation() {
894    if (!CLocs.back().isDead() && CLocs.back().asLocation().isFileID()) {
895      // For contexts, we only one the first character as the range.
896      rawAddEdge(cleanUpLocation(CLocs.back(), true));
897    }
898    CLocs.pop_back();
899  }
900
901public:
902  EdgeBuilder(PathDiagnostic &pd, PathDiagnosticBuilder &pdb)
903    : PD(pd), PDB(pdb) {
904
905      // If the PathDiagnostic already has pieces, add the enclosing statement
906      // of the first piece as a context as well.
907      if (!PD.empty()) {
908        PrevLoc = PD.begin()->getLocation();
909
910        if (const Stmt *S = PrevLoc.asStmt())
911          addExtendedContext(PDB.getEnclosingStmtLocation(S).asStmt());
912      }
913  }
914
915  ~EdgeBuilder() {
916    while (!CLocs.empty()) popLocation();
917
918    // Finally, add an initial edge from the start location of the first
919    // statement (if it doesn't already exist).
920    // FIXME: Should handle CXXTryStmt if analyser starts supporting C++.
921    if (const CompoundStmt *CS =
922          dyn_cast_or_null<CompoundStmt>(PDB.getCodeDecl().getBody()))
923      if (!CS->body_empty()) {
924        SourceLocation Loc = (*CS->body_begin())->getLocStart();
925        rawAddEdge(PathDiagnosticLocation(Loc, PDB.getSourceManager()));
926      }
927
928  }
929
930  void addEdge(PathDiagnosticLocation NewLoc, bool alwaysAdd = false);
931
932  void rawAddEdge(PathDiagnosticLocation NewLoc);
933
934  void addContext(const Stmt *S);
935  void addExtendedContext(const Stmt *S);
936};
937} // end anonymous namespace
938
939
940PathDiagnosticLocation
941EdgeBuilder::getContextLocation(const PathDiagnosticLocation &L) {
942  if (const Stmt *S = L.asStmt()) {
943    if (IsControlFlowExpr(S))
944      return L;
945
946    return PDB.getEnclosingStmtLocation(S);
947  }
948
949  return L;
950}
951
952bool EdgeBuilder::containsLocation(const PathDiagnosticLocation &Container,
953                                   const PathDiagnosticLocation &Containee) {
954
955  if (Container == Containee)
956    return true;
957
958  if (Container.asDecl())
959    return true;
960
961  if (const Stmt *S = Containee.asStmt())
962    if (const Stmt *ContainerS = Container.asStmt()) {
963      while (S) {
964        if (S == ContainerS)
965          return true;
966        S = PDB.getParent(S);
967      }
968      return false;
969    }
970
971  // Less accurate: compare using source ranges.
972  SourceRange ContainerR = Container.asRange();
973  SourceRange ContaineeR = Containee.asRange();
974
975  SourceManager &SM = PDB.getSourceManager();
976  SourceLocation ContainerRBeg = SM.getInstantiationLoc(ContainerR.getBegin());
977  SourceLocation ContainerREnd = SM.getInstantiationLoc(ContainerR.getEnd());
978  SourceLocation ContaineeRBeg = SM.getInstantiationLoc(ContaineeR.getBegin());
979  SourceLocation ContaineeREnd = SM.getInstantiationLoc(ContaineeR.getEnd());
980
981  unsigned ContainerBegLine = SM.getInstantiationLineNumber(ContainerRBeg);
982  unsigned ContainerEndLine = SM.getInstantiationLineNumber(ContainerREnd);
983  unsigned ContaineeBegLine = SM.getInstantiationLineNumber(ContaineeRBeg);
984  unsigned ContaineeEndLine = SM.getInstantiationLineNumber(ContaineeREnd);
985
986  assert(ContainerBegLine <= ContainerEndLine);
987  assert(ContaineeBegLine <= ContaineeEndLine);
988
989  return (ContainerBegLine <= ContaineeBegLine &&
990          ContainerEndLine >= ContaineeEndLine &&
991          (ContainerBegLine != ContaineeBegLine ||
992           SM.getInstantiationColumnNumber(ContainerRBeg) <=
993           SM.getInstantiationColumnNumber(ContaineeRBeg)) &&
994          (ContainerEndLine != ContaineeEndLine ||
995           SM.getInstantiationColumnNumber(ContainerREnd) >=
996           SM.getInstantiationColumnNumber(ContainerREnd)));
997}
998
999void EdgeBuilder::rawAddEdge(PathDiagnosticLocation NewLoc) {
1000  if (!PrevLoc.isValid()) {
1001    PrevLoc = NewLoc;
1002    return;
1003  }
1004
1005  const PathDiagnosticLocation &NewLocClean = cleanUpLocation(NewLoc);
1006  const PathDiagnosticLocation &PrevLocClean = cleanUpLocation(PrevLoc);
1007
1008  if (NewLocClean.asLocation() == PrevLocClean.asLocation())
1009    return;
1010
1011  // FIXME: Ignore intra-macro edges for now.
1012  if (NewLocClean.asLocation().getInstantiationLoc() ==
1013      PrevLocClean.asLocation().getInstantiationLoc())
1014    return;
1015
1016  PD.push_front(new PathDiagnosticControlFlowPiece(NewLocClean, PrevLocClean));
1017  PrevLoc = NewLoc;
1018}
1019
1020void EdgeBuilder::addEdge(PathDiagnosticLocation NewLoc, bool alwaysAdd) {
1021
1022  if (!alwaysAdd && NewLoc.asLocation().isMacroID())
1023    return;
1024
1025  const PathDiagnosticLocation &CLoc = getContextLocation(NewLoc);
1026
1027  while (!CLocs.empty()) {
1028    ContextLocation &TopContextLoc = CLocs.back();
1029
1030    // Is the top location context the same as the one for the new location?
1031    if (TopContextLoc == CLoc) {
1032      if (alwaysAdd) {
1033        if (IsConsumedExpr(TopContextLoc) &&
1034            !IsControlFlowExpr(TopContextLoc.asStmt()))
1035            TopContextLoc.markDead();
1036
1037        rawAddEdge(NewLoc);
1038      }
1039
1040      return;
1041    }
1042
1043    if (containsLocation(TopContextLoc, CLoc)) {
1044      if (alwaysAdd) {
1045        rawAddEdge(NewLoc);
1046
1047        if (IsConsumedExpr(CLoc) && !IsControlFlowExpr(CLoc.asStmt())) {
1048          CLocs.push_back(ContextLocation(CLoc, true));
1049          return;
1050        }
1051      }
1052
1053      CLocs.push_back(CLoc);
1054      return;
1055    }
1056
1057    // Context does not contain the location.  Flush it.
1058    popLocation();
1059  }
1060
1061  // If we reach here, there is no enclosing context.  Just add the edge.
1062  rawAddEdge(NewLoc);
1063}
1064
1065bool EdgeBuilder::IsConsumedExpr(const PathDiagnosticLocation &L) {
1066  if (const Expr *X = dyn_cast_or_null<Expr>(L.asStmt()))
1067    return PDB.getParentMap().isConsumedExpr(X) && !IsControlFlowExpr(X);
1068
1069  return false;
1070}
1071
1072void EdgeBuilder::addExtendedContext(const Stmt *S) {
1073  if (!S)
1074    return;
1075
1076  const Stmt *Parent = PDB.getParent(S);
1077  while (Parent) {
1078    if (isa<CompoundStmt>(Parent))
1079      Parent = PDB.getParent(Parent);
1080    else
1081      break;
1082  }
1083
1084  if (Parent) {
1085    switch (Parent->getStmtClass()) {
1086      case Stmt::DoStmtClass:
1087      case Stmt::ObjCAtSynchronizedStmtClass:
1088        addContext(Parent);
1089      default:
1090        break;
1091    }
1092  }
1093
1094  addContext(S);
1095}
1096
1097void EdgeBuilder::addContext(const Stmt *S) {
1098  if (!S)
1099    return;
1100
1101  PathDiagnosticLocation L(S, PDB.getSourceManager());
1102
1103  while (!CLocs.empty()) {
1104    const PathDiagnosticLocation &TopContextLoc = CLocs.back();
1105
1106    // Is the top location context the same as the one for the new location?
1107    if (TopContextLoc == L)
1108      return;
1109
1110    if (containsLocation(TopContextLoc, L)) {
1111      CLocs.push_back(L);
1112      return;
1113    }
1114
1115    // Context does not contain the location.  Flush it.
1116    popLocation();
1117  }
1118
1119  CLocs.push_back(L);
1120}
1121
1122static void GenerateExtensivePathDiagnostic(PathDiagnostic& PD,
1123                                            PathDiagnosticBuilder &PDB,
1124                                            const ExplodedNode *N) {
1125  EdgeBuilder EB(PD, PDB);
1126
1127  const ExplodedNode* NextNode = N->pred_empty() ? NULL : *(N->pred_begin());
1128  while (NextNode) {
1129    N = NextNode;
1130    NextNode = GetPredecessorNode(N);
1131    ProgramPoint P = N->getLocation();
1132
1133    do {
1134      // Block edges.
1135      if (const BlockEdge *BE = dyn_cast<BlockEdge>(&P)) {
1136        const CFGBlock &Blk = *BE->getSrc();
1137        const Stmt *Term = Blk.getTerminator();
1138
1139        // Are we jumping to the head of a loop?  Add a special diagnostic.
1140        if (const Stmt *Loop = BE->getDst()->getLoopTarget()) {
1141          PathDiagnosticLocation L(Loop, PDB.getSourceManager());
1142          const CompoundStmt *CS = NULL;
1143
1144          if (!Term) {
1145            if (const ForStmt *FS = dyn_cast<ForStmt>(Loop))
1146              CS = dyn_cast<CompoundStmt>(FS->getBody());
1147            else if (const WhileStmt *WS = dyn_cast<WhileStmt>(Loop))
1148              CS = dyn_cast<CompoundStmt>(WS->getBody());
1149          }
1150
1151          PathDiagnosticEventPiece *p =
1152            new PathDiagnosticEventPiece(L,
1153                                        "Looping back to the head of the loop");
1154
1155          EB.addEdge(p->getLocation(), true);
1156          PD.push_front(p);
1157
1158          if (CS) {
1159            PathDiagnosticLocation BL(CS->getRBracLoc(),
1160                                      PDB.getSourceManager());
1161            BL = PathDiagnosticLocation(BL.asLocation());
1162            EB.addEdge(BL);
1163          }
1164        }
1165
1166        if (Term)
1167          EB.addContext(Term);
1168
1169        break;
1170      }
1171
1172      if (const BlockEntrance *BE = dyn_cast<BlockEntrance>(&P)) {
1173        if (CFGStmt S = BE->getFirstElement().getAs<CFGStmt>()) {
1174          if (IsControlFlowExpr(S)) {
1175            // Add the proper context for '&&', '||', and '?'.
1176            EB.addContext(S);
1177          }
1178          else
1179            EB.addExtendedContext(PDB.getEnclosingStmtLocation(S).asStmt());
1180        }
1181
1182        break;
1183      }
1184    } while (0);
1185
1186    if (!NextNode)
1187      continue;
1188
1189    for (BugReporterContext::visitor_iterator I = PDB.visitor_begin(),
1190         E = PDB.visitor_end(); I!=E; ++I) {
1191      if (PathDiagnosticPiece* p = (*I)->VisitNode(N, NextNode, PDB)) {
1192        const PathDiagnosticLocation &Loc = p->getLocation();
1193        EB.addEdge(Loc, true);
1194        PD.push_front(p);
1195        if (const Stmt *S = Loc.asStmt())
1196          EB.addExtendedContext(PDB.getEnclosingStmtLocation(S).asStmt());
1197      }
1198    }
1199  }
1200}
1201
1202//===----------------------------------------------------------------------===//
1203// Methods for BugType and subclasses.
1204//===----------------------------------------------------------------------===//
1205BugType::~BugType() {
1206  // Free up the equivalence class objects.  Observe that we get a pointer to
1207  // the object first before incrementing the iterator, as destroying the
1208  // node before doing so means we will read from freed memory.
1209  for (iterator I = begin(), E = end(); I !=E; ) {
1210    BugReportEquivClass *EQ = &*I;
1211    ++I;
1212    delete EQ;
1213  }
1214}
1215void BugType::FlushReports(BugReporter &BR) {}
1216
1217//===----------------------------------------------------------------------===//
1218// Methods for BugReport and subclasses.
1219//===----------------------------------------------------------------------===//
1220BugReport::~BugReport() {}
1221RangedBugReport::~RangedBugReport() {}
1222
1223const Stmt* BugReport::getStmt() const {
1224  ProgramPoint ProgP = ErrorNode->getLocation();
1225  const Stmt *S = NULL;
1226
1227  if (BlockEntrance* BE = dyn_cast<BlockEntrance>(&ProgP)) {
1228    CFGBlock &Exit = ProgP.getLocationContext()->getCFG()->getExit();
1229    if (BE->getBlock() == &Exit)
1230      S = GetPreviousStmt(ErrorNode);
1231  }
1232  if (!S)
1233    S = GetStmt(ProgP);
1234
1235  return S;
1236}
1237
1238PathDiagnosticPiece*
1239BugReport::getEndPath(BugReporterContext& BRC,
1240                      const ExplodedNode* EndPathNode) {
1241
1242  const Stmt* S = getStmt();
1243
1244  if (!S)
1245    return NULL;
1246
1247  BugReport::ranges_iterator Beg, End;
1248  llvm::tie(Beg, End) = getRanges();
1249  PathDiagnosticLocation L(S, BRC.getSourceManager());
1250
1251  // Only add the statement itself as a range if we didn't specify any
1252  // special ranges for this report.
1253  PathDiagnosticPiece* P = new PathDiagnosticEventPiece(L, getDescription(),
1254                                                        Beg == End);
1255
1256  for (; Beg != End; ++Beg)
1257    P->addRange(*Beg);
1258
1259  return P;
1260}
1261
1262std::pair<BugReport::ranges_iterator, BugReport::ranges_iterator>
1263BugReport::getRanges() const {
1264  if (const Expr* E = dyn_cast_or_null<Expr>(getStmt())) {
1265    R = E->getSourceRange();
1266    assert(R.isValid());
1267    return std::make_pair(&R, &R+1);
1268  }
1269  else
1270    return std::make_pair(ranges_iterator(), ranges_iterator());
1271}
1272
1273SourceLocation BugReport::getLocation() const {
1274  if (ErrorNode)
1275    if (const Stmt* S = GetCurrentOrPreviousStmt(ErrorNode)) {
1276      // For member expressions, return the location of the '.' or '->'.
1277      if (const MemberExpr *ME = dyn_cast<MemberExpr>(S))
1278        return ME->getMemberLoc();
1279      // For binary operators, return the location of the operator.
1280      if (const BinaryOperator *B = dyn_cast<BinaryOperator>(S))
1281        return B->getOperatorLoc();
1282
1283      return S->getLocStart();
1284    }
1285
1286  return FullSourceLoc();
1287}
1288
1289PathDiagnosticPiece* BugReport::VisitNode(const ExplodedNode* N,
1290                                          const ExplodedNode* PrevN,
1291                                          BugReporterContext &BRC) {
1292  return NULL;
1293}
1294
1295//===----------------------------------------------------------------------===//
1296// Methods for BugReporter and subclasses.
1297//===----------------------------------------------------------------------===//
1298
1299BugReportEquivClass::~BugReportEquivClass() {
1300  for (iterator I=begin(), E=end(); I!=E; ++I) delete *I;
1301}
1302
1303GRBugReporter::~GRBugReporter() { }
1304BugReporterData::~BugReporterData() {}
1305
1306ExplodedGraph &GRBugReporter::getGraph() { return Eng.getGraph(); }
1307
1308GRStateManager&
1309GRBugReporter::getStateManager() { return Eng.getStateManager(); }
1310
1311BugReporter::~BugReporter() { FlushReports(); }
1312
1313void BugReporter::FlushReports() {
1314  if (BugTypes.isEmpty())
1315    return;
1316
1317  // First flush the warnings for each BugType.  This may end up creating new
1318  // warnings and new BugTypes.  Because ImmutableSet is a functional data
1319  // structure, we do not need to worry about the iterators being invalidated.
1320  for (BugTypesTy::iterator I=BugTypes.begin(), E=BugTypes.end(); I!=E; ++I)
1321    const_cast<BugType*>(*I)->FlushReports(*this);
1322
1323  // Iterate through BugTypes a second time.  BugTypes may have been updated
1324  // with new BugType objects and new warnings.
1325  for (BugTypesTy::iterator I=BugTypes.begin(), E=BugTypes.end(); I!=E; ++I) {
1326    BugType *BT = const_cast<BugType*>(*I);
1327
1328    typedef llvm::FoldingSet<BugReportEquivClass> SetTy;
1329    SetTy& EQClasses = BT->EQClasses;
1330
1331    for (SetTy::iterator EI=EQClasses.begin(), EE=EQClasses.end(); EI!=EE;++EI){
1332      BugReportEquivClass& EQ = *EI;
1333      FlushReport(EQ);
1334    }
1335
1336    // Delete the BugType object.
1337    delete BT;
1338  }
1339
1340  // Remove all references to the BugType objects.
1341  BugTypes = F.getEmptySet();
1342}
1343
1344//===----------------------------------------------------------------------===//
1345// PathDiagnostics generation.
1346//===----------------------------------------------------------------------===//
1347
1348static std::pair<std::pair<ExplodedGraph*, NodeBackMap*>,
1349                 std::pair<ExplodedNode*, unsigned> >
1350MakeReportGraph(const ExplodedGraph* G,
1351                llvm::SmallVectorImpl<const ExplodedNode*> &nodes) {
1352
1353  // Create the trimmed graph.  It will contain the shortest paths from the
1354  // error nodes to the root.  In the new graph we should only have one
1355  // error node unless there are two or more error nodes with the same minimum
1356  // path length.
1357  ExplodedGraph* GTrim;
1358  InterExplodedGraphMap* NMap;
1359
1360  llvm::DenseMap<const void*, const void*> InverseMap;
1361  llvm::tie(GTrim, NMap) = G->Trim(nodes.data(), nodes.data() + nodes.size(),
1362                                   &InverseMap);
1363
1364  // Create owning pointers for GTrim and NMap just to ensure that they are
1365  // released when this function exists.
1366  llvm::OwningPtr<ExplodedGraph> AutoReleaseGTrim(GTrim);
1367  llvm::OwningPtr<InterExplodedGraphMap> AutoReleaseNMap(NMap);
1368
1369  // Find the (first) error node in the trimmed graph.  We just need to consult
1370  // the node map (NMap) which maps from nodes in the original graph to nodes
1371  // in the new graph.
1372
1373  std::queue<const ExplodedNode*> WS;
1374  typedef llvm::DenseMap<const ExplodedNode*, unsigned> IndexMapTy;
1375  IndexMapTy IndexMap;
1376
1377  for (unsigned nodeIndex = 0 ; nodeIndex < nodes.size(); ++nodeIndex) {
1378    const ExplodedNode *originalNode = nodes[nodeIndex];
1379    if (const ExplodedNode *N = NMap->getMappedNode(originalNode)) {
1380      WS.push(N);
1381      IndexMap[originalNode] = nodeIndex;
1382    }
1383  }
1384
1385  assert(!WS.empty() && "No error node found in the trimmed graph.");
1386
1387  // Create a new (third!) graph with a single path.  This is the graph
1388  // that will be returned to the caller.
1389  ExplodedGraph *GNew = new ExplodedGraph();
1390
1391  // Sometimes the trimmed graph can contain a cycle.  Perform a reverse BFS
1392  // to the root node, and then construct a new graph that contains only
1393  // a single path.
1394  llvm::DenseMap<const void*,unsigned> Visited;
1395
1396  unsigned cnt = 0;
1397  const ExplodedNode* Root = 0;
1398
1399  while (!WS.empty()) {
1400    const ExplodedNode* Node = WS.front();
1401    WS.pop();
1402
1403    if (Visited.find(Node) != Visited.end())
1404      continue;
1405
1406    Visited[Node] = cnt++;
1407
1408    if (Node->pred_empty()) {
1409      Root = Node;
1410      break;
1411    }
1412
1413    for (ExplodedNode::const_pred_iterator I=Node->pred_begin(),
1414         E=Node->pred_end(); I!=E; ++I)
1415      WS.push(*I);
1416  }
1417
1418  assert(Root);
1419
1420  // Now walk from the root down the BFS path, always taking the successor
1421  // with the lowest number.
1422  ExplodedNode *Last = 0, *First = 0;
1423  NodeBackMap *BM = new NodeBackMap();
1424  unsigned NodeIndex = 0;
1425
1426  for ( const ExplodedNode *N = Root ;;) {
1427    // Lookup the number associated with the current node.
1428    llvm::DenseMap<const void*,unsigned>::iterator I = Visited.find(N);
1429    assert(I != Visited.end());
1430
1431    // Create the equivalent node in the new graph with the same state
1432    // and location.
1433    ExplodedNode* NewN = GNew->getNode(N->getLocation(), N->getState());
1434
1435    // Store the mapping to the original node.
1436    llvm::DenseMap<const void*, const void*>::iterator IMitr=InverseMap.find(N);
1437    assert(IMitr != InverseMap.end() && "No mapping to original node.");
1438    (*BM)[NewN] = (const ExplodedNode*) IMitr->second;
1439
1440    // Link up the new node with the previous node.
1441    if (Last)
1442      NewN->addPredecessor(Last, *GNew);
1443
1444    Last = NewN;
1445
1446    // Are we at the final node?
1447    IndexMapTy::iterator IMI =
1448      IndexMap.find((const ExplodedNode*)(IMitr->second));
1449    if (IMI != IndexMap.end()) {
1450      First = NewN;
1451      NodeIndex = IMI->second;
1452      break;
1453    }
1454
1455    // Find the next successor node.  We choose the node that is marked
1456    // with the lowest DFS number.
1457    ExplodedNode::const_succ_iterator SI = N->succ_begin();
1458    ExplodedNode::const_succ_iterator SE = N->succ_end();
1459    N = 0;
1460
1461    for (unsigned MinVal = 0; SI != SE; ++SI) {
1462
1463      I = Visited.find(*SI);
1464
1465      if (I == Visited.end())
1466        continue;
1467
1468      if (!N || I->second < MinVal) {
1469        N = *SI;
1470        MinVal = I->second;
1471      }
1472    }
1473
1474    assert(N);
1475  }
1476
1477  assert(First);
1478
1479  return std::make_pair(std::make_pair(GNew, BM),
1480                        std::make_pair(First, NodeIndex));
1481}
1482
1483/// CompactPathDiagnostic - This function postprocesses a PathDiagnostic object
1484///  and collapses PathDiagosticPieces that are expanded by macros.
1485static void CompactPathDiagnostic(PathDiagnostic &PD, const SourceManager& SM) {
1486  typedef std::vector<std::pair<PathDiagnosticMacroPiece*, SourceLocation> >
1487          MacroStackTy;
1488
1489  typedef std::vector<PathDiagnosticPiece*>
1490          PiecesTy;
1491
1492  MacroStackTy MacroStack;
1493  PiecesTy Pieces;
1494
1495  for (PathDiagnostic::iterator I = PD.begin(), E = PD.end(); I!=E; ++I) {
1496    // Get the location of the PathDiagnosticPiece.
1497    const FullSourceLoc Loc = I->getLocation().asLocation();
1498
1499    // Determine the instantiation location, which is the location we group
1500    // related PathDiagnosticPieces.
1501    SourceLocation InstantiationLoc = Loc.isMacroID() ?
1502                                      SM.getInstantiationLoc(Loc) :
1503                                      SourceLocation();
1504
1505    if (Loc.isFileID()) {
1506      MacroStack.clear();
1507      Pieces.push_back(&*I);
1508      continue;
1509    }
1510
1511    assert(Loc.isMacroID());
1512
1513    // Is the PathDiagnosticPiece within the same macro group?
1514    if (!MacroStack.empty() && InstantiationLoc == MacroStack.back().second) {
1515      MacroStack.back().first->push_back(&*I);
1516      continue;
1517    }
1518
1519    // We aren't in the same group.  Are we descending into a new macro
1520    // or are part of an old one?
1521    PathDiagnosticMacroPiece *MacroGroup = 0;
1522
1523    SourceLocation ParentInstantiationLoc = InstantiationLoc.isMacroID() ?
1524                                          SM.getInstantiationLoc(Loc) :
1525                                          SourceLocation();
1526
1527    // Walk the entire macro stack.
1528    while (!MacroStack.empty()) {
1529      if (InstantiationLoc == MacroStack.back().second) {
1530        MacroGroup = MacroStack.back().first;
1531        break;
1532      }
1533
1534      if (ParentInstantiationLoc == MacroStack.back().second) {
1535        MacroGroup = MacroStack.back().first;
1536        break;
1537      }
1538
1539      MacroStack.pop_back();
1540    }
1541
1542    if (!MacroGroup || ParentInstantiationLoc == MacroStack.back().second) {
1543      // Create a new macro group and add it to the stack.
1544      PathDiagnosticMacroPiece *NewGroup = new PathDiagnosticMacroPiece(Loc);
1545
1546      if (MacroGroup)
1547        MacroGroup->push_back(NewGroup);
1548      else {
1549        assert(InstantiationLoc.isFileID());
1550        Pieces.push_back(NewGroup);
1551      }
1552
1553      MacroGroup = NewGroup;
1554      MacroStack.push_back(std::make_pair(MacroGroup, InstantiationLoc));
1555    }
1556
1557    // Finally, add the PathDiagnosticPiece to the group.
1558    MacroGroup->push_back(&*I);
1559  }
1560
1561  // Now take the pieces and construct a new PathDiagnostic.
1562  PD.resetPath(false);
1563
1564  for (PiecesTy::iterator I=Pieces.begin(), E=Pieces.end(); I!=E; ++I) {
1565    if (PathDiagnosticMacroPiece *MP=dyn_cast<PathDiagnosticMacroPiece>(*I))
1566      if (!MP->containsEvent()) {
1567        delete MP;
1568        continue;
1569      }
1570
1571    PD.push_back(*I);
1572  }
1573}
1574
1575void GRBugReporter::GeneratePathDiagnostic(PathDiagnostic& PD,
1576                        llvm::SmallVectorImpl<BugReport *> &bugReports) {
1577
1578  assert(!bugReports.empty());
1579  llvm::SmallVector<const ExplodedNode *, 10> errorNodes;
1580  for (llvm::SmallVectorImpl<BugReport*>::iterator I = bugReports.begin(),
1581    E = bugReports.end(); I != E; ++I) {
1582      errorNodes.push_back((*I)->getErrorNode());
1583  }
1584
1585  // Construct a new graph that contains only a single path from the error
1586  // node to a root.
1587  const std::pair<std::pair<ExplodedGraph*, NodeBackMap*>,
1588  std::pair<ExplodedNode*, unsigned> >&
1589    GPair = MakeReportGraph(&getGraph(), errorNodes);
1590
1591  // Find the BugReport with the original location.
1592  assert(GPair.second.second < bugReports.size());
1593  BugReport *R = bugReports[GPair.second.second];
1594  assert(R && "No original report found for sliced graph.");
1595
1596  llvm::OwningPtr<ExplodedGraph> ReportGraph(GPair.first.first);
1597  llvm::OwningPtr<NodeBackMap> BackMap(GPair.first.second);
1598  const ExplodedNode *N = GPair.second.first;
1599
1600  // Start building the path diagnostic...
1601  PathDiagnosticBuilder PDB(*this, R, BackMap.get(), getPathDiagnosticClient());
1602
1603  if (PathDiagnosticPiece* Piece = R->getEndPath(PDB, N))
1604    PD.push_back(Piece);
1605  else
1606    return;
1607
1608  // Register node visitors.
1609  R->registerInitialVisitors(PDB, N);
1610  bugreporter::registerNilReceiverVisitor(PDB);
1611
1612  switch (PDB.getGenerationScheme()) {
1613    case PathDiagnosticClient::Extensive:
1614      GenerateExtensivePathDiagnostic(PD, PDB, N);
1615      break;
1616    case PathDiagnosticClient::Minimal:
1617      GenerateMinimalPathDiagnostic(PD, PDB, N);
1618      break;
1619  }
1620}
1621
1622void BugReporter::Register(BugType *BT) {
1623  BugTypes = F.add(BugTypes, BT);
1624}
1625
1626void BugReporter::EmitReport(BugReport* R) {
1627  // Compute the bug report's hash to determine its equivalence class.
1628  llvm::FoldingSetNodeID ID;
1629  R->Profile(ID);
1630
1631  // Lookup the equivance class.  If there isn't one, create it.
1632  BugType& BT = R->getBugType();
1633  Register(&BT);
1634  void *InsertPos;
1635  BugReportEquivClass* EQ = BT.EQClasses.FindNodeOrInsertPos(ID, InsertPos);
1636
1637  if (!EQ) {
1638    EQ = new BugReportEquivClass(R);
1639    BT.EQClasses.InsertNode(EQ, InsertPos);
1640  }
1641  else
1642    EQ->AddReport(R);
1643}
1644
1645
1646//===----------------------------------------------------------------------===//
1647// Emitting reports in equivalence classes.
1648//===----------------------------------------------------------------------===//
1649
1650namespace {
1651struct FRIEC_WLItem {
1652  const ExplodedNode *N;
1653  ExplodedNode::const_succ_iterator I, E;
1654
1655  FRIEC_WLItem(const ExplodedNode *n)
1656  : N(n), I(N->succ_begin()), E(N->succ_end()) {}
1657};
1658}
1659
1660static BugReport *
1661FindReportInEquivalenceClass(BugReportEquivClass& EQ,
1662                             llvm::SmallVectorImpl<BugReport*> &bugReports) {
1663
1664  BugReportEquivClass::iterator I = EQ.begin(), E = EQ.end();
1665  assert(I != E);
1666  BugReport *R = *I;
1667  BugType& BT = R->getBugType();
1668
1669  // If we don't need to suppress any of the nodes because they are
1670  // post-dominated by a sink, simply add all the nodes in the equivalence class
1671  // to 'Nodes'.  Any of the reports will serve as a "representative" report.
1672  if (!BT.isSuppressOnSink()) {
1673    for (BugReportEquivClass::iterator I=EQ.begin(), E=EQ.end(); I!=E; ++I) {
1674      const ExplodedNode* N = I->getErrorNode();
1675      if (N) {
1676        R = *I;
1677        bugReports.push_back(R);
1678      }
1679    }
1680    return R;
1681  }
1682
1683  // For bug reports that should be suppressed when all paths are post-dominated
1684  // by a sink node, iterate through the reports in the equivalence class
1685  // until we find one that isn't post-dominated (if one exists).  We use a
1686  // DFS traversal of the ExplodedGraph to find a non-sink node.  We could write
1687  // this as a recursive function, but we don't want to risk blowing out the
1688  // stack for very long paths.
1689  BugReport *exampleReport = 0;
1690
1691  for (; I != E; ++I) {
1692    R = *I;
1693    const ExplodedNode *errorNode = R->getErrorNode();
1694
1695    if (!errorNode)
1696      continue;
1697    if (errorNode->isSink()) {
1698      assert(false &&
1699           "BugType::isSuppressSink() should not be 'true' for sink end nodes");
1700      return 0;
1701    }
1702    // No successors?  By definition this nodes isn't post-dominated by a sink.
1703    if (errorNode->succ_empty()) {
1704      bugReports.push_back(R);
1705      if (!exampleReport)
1706        exampleReport = R;
1707      continue;
1708    }
1709
1710    // At this point we know that 'N' is not a sink and it has at least one
1711    // successor.  Use a DFS worklist to find a non-sink end-of-path node.
1712    typedef FRIEC_WLItem WLItem;
1713    typedef llvm::SmallVector<WLItem, 10> DFSWorkList;
1714    llvm::DenseMap<const ExplodedNode *, unsigned> Visited;
1715
1716    DFSWorkList WL;
1717    WL.push_back(errorNode);
1718    Visited[errorNode] = 1;
1719
1720    while (!WL.empty()) {
1721      WLItem &WI = WL.back();
1722      assert(!WI.N->succ_empty());
1723
1724      for (; WI.I != WI.E; ++WI.I) {
1725        const ExplodedNode *Succ = *WI.I;
1726        // End-of-path node?
1727        if (Succ->succ_empty()) {
1728          // If we found an end-of-path node that is not a sink.
1729          if (!Succ->isSink()) {
1730            bugReports.push_back(R);
1731            if (!exampleReport)
1732              exampleReport = R;
1733            WL.clear();
1734            break;
1735          }
1736          // Found a sink?  Continue on to the next successor.
1737          continue;
1738        }
1739        // Mark the successor as visited.  If it hasn't been explored,
1740        // enqueue it to the DFS worklist.
1741        unsigned &mark = Visited[Succ];
1742        if (!mark) {
1743          mark = 1;
1744          WL.push_back(Succ);
1745          break;
1746        }
1747      }
1748
1749      // The worklist may have been cleared at this point.  First
1750      // check if it is empty before checking the last item.
1751      if (!WL.empty() && &WL.back() == &WI)
1752        WL.pop_back();
1753    }
1754  }
1755
1756  // ExampleReport will be NULL if all the nodes in the equivalence class
1757  // were post-dominated by sinks.
1758  return exampleReport;
1759}
1760
1761//===----------------------------------------------------------------------===//
1762// DiagnosticCache.  This is a hack to cache analyzer diagnostics.  It
1763// uses global state, which eventually should go elsewhere.
1764//===----------------------------------------------------------------------===//
1765namespace {
1766class DiagCacheItem : public llvm::FoldingSetNode {
1767  llvm::FoldingSetNodeID ID;
1768public:
1769  DiagCacheItem(BugReport *R, PathDiagnostic *PD) {
1770    ID.AddString(R->getBugType().getName());
1771    ID.AddString(R->getBugType().getCategory());
1772    ID.AddString(R->getDescription());
1773    ID.AddInteger(R->getLocation().getRawEncoding());
1774    PD->Profile(ID);
1775  }
1776
1777  void Profile(llvm::FoldingSetNodeID &id) {
1778    id = ID;
1779  }
1780
1781  llvm::FoldingSetNodeID &getID() { return ID; }
1782};
1783}
1784
1785static bool IsCachedDiagnostic(BugReport *R, PathDiagnostic *PD) {
1786  // FIXME: Eventually this diagnostic cache should reside in something
1787  // like AnalysisManager instead of being a static variable.  This is
1788  // really unsafe in the long term.
1789  typedef llvm::FoldingSet<DiagCacheItem> DiagnosticCache;
1790  static DiagnosticCache DC;
1791
1792  void *InsertPos;
1793  DiagCacheItem *Item = new DiagCacheItem(R, PD);
1794
1795  if (DC.FindNodeOrInsertPos(Item->getID(), InsertPos)) {
1796    delete Item;
1797    return true;
1798  }
1799
1800  DC.InsertNode(Item, InsertPos);
1801  return false;
1802}
1803
1804void BugReporter::FlushReport(BugReportEquivClass& EQ) {
1805  llvm::SmallVector<BugReport*, 10> bugReports;
1806  BugReport *exampleReport = FindReportInEquivalenceClass(EQ, bugReports);
1807  if (!exampleReport)
1808    return;
1809
1810  PathDiagnosticClient* PD = getPathDiagnosticClient();
1811
1812  // FIXME: Make sure we use the 'R' for the path that was actually used.
1813  // Probably doesn't make a difference in practice.
1814  BugType& BT = exampleReport->getBugType();
1815
1816  llvm::OwningPtr<PathDiagnostic>
1817    D(new PathDiagnostic(exampleReport->getBugType().getName(),
1818                         !PD || PD->useVerboseDescription()
1819                         ? exampleReport->getDescription()
1820                         : exampleReport->getShortDescription(),
1821                         BT.getCategory()));
1822
1823  if (!bugReports.empty())
1824    GeneratePathDiagnostic(*D.get(), bugReports);
1825
1826  if (IsCachedDiagnostic(exampleReport, D.get()))
1827    return;
1828
1829  // Get the meta data.
1830  std::pair<const char**, const char**> Meta =
1831    exampleReport->getExtraDescriptiveText();
1832  for (const char** s = Meta.first; s != Meta.second; ++s)
1833    D->addMeta(*s);
1834
1835  // Emit a summary diagnostic to the regular Diagnostics engine.
1836  BugReport::ranges_iterator Beg, End;
1837  llvm::tie(Beg, End) = exampleReport->getRanges();
1838  Diagnostic &Diag = getDiagnostic();
1839  FullSourceLoc L(exampleReport->getLocation(), getSourceManager());
1840
1841  // Search the description for '%', as that will be interpretted as a
1842  // format character by FormatDiagnostics.
1843  llvm::StringRef desc = exampleReport->getShortDescription();
1844  unsigned ErrorDiag;
1845  {
1846    llvm::SmallString<512> TmpStr;
1847    llvm::raw_svector_ostream Out(TmpStr);
1848    for (llvm::StringRef::iterator I=desc.begin(), E=desc.end(); I!=E; ++I)
1849      if (*I == '%')
1850        Out << "%%";
1851      else
1852        Out << *I;
1853
1854    Out.flush();
1855    ErrorDiag = Diag.getCustomDiagID(Diagnostic::Warning, TmpStr);
1856  }
1857
1858  {
1859    DiagnosticBuilder diagBuilder = Diag.Report(L, ErrorDiag);
1860    for (BugReport::ranges_iterator I = Beg; I != End; ++I)
1861      diagBuilder << *I;
1862  }
1863
1864  // Emit a full diagnostic for the path if we have a PathDiagnosticClient.
1865  if (!PD)
1866    return;
1867
1868  if (D->empty()) {
1869    PathDiagnosticPiece* piece =
1870      new PathDiagnosticEventPiece(L, exampleReport->getDescription());
1871
1872    for ( ; Beg != End; ++Beg) piece->addRange(*Beg);
1873    D->push_back(piece);
1874  }
1875
1876  PD->HandlePathDiagnostic(D.take());
1877}
1878
1879void BugReporter::EmitBasicReport(llvm::StringRef name, llvm::StringRef str,
1880                                  SourceLocation Loc,
1881                                  SourceRange* RBeg, unsigned NumRanges) {
1882  EmitBasicReport(name, "", str, Loc, RBeg, NumRanges);
1883}
1884
1885void BugReporter::EmitBasicReport(llvm::StringRef name,
1886                                  llvm::StringRef category,
1887                                  llvm::StringRef str, SourceLocation Loc,
1888                                  SourceRange* RBeg, unsigned NumRanges) {
1889
1890  // 'BT' will be owned by BugReporter as soon as we call 'EmitReport'.
1891  BugType *BT = new BugType(name, category);
1892  FullSourceLoc L = getContext().getFullLoc(Loc);
1893  RangedBugReport *R = new DiagBugReport(*BT, str, L);
1894  for ( ; NumRanges > 0 ; --NumRanges, ++RBeg) R->addRange(*RBeg);
1895  EmitReport(R);
1896}
1897