[PATCH] D62883: [analyzer] Track conditions of terminator statements on which the reported node depends on

[Artem Dergachev via Phabricator via cfe-commits]

NoQ added inline comments.


================
Comment at: clang/lib/StaticAnalyzer/Core/BugReporterVisitors.cpp:1609-1613
+  if (B->rbegin()->getKind() != CFGElement::Kind::Statement)
+    return nullptr;
+
+  // This should be the condition of the terminator block.
+  const Stmt *S = B->rbegin()->castAs<CFGStmt>().getStmt();
----------------
A bit clearner:

```lang=c++
auto StmtElem = B->rbegin().getAs<CFGStmt>();
if (!StmtElem)
  return nullptr;

const Stmt *S = StmtElem->getStmt();
```

Also how about `CFGBlock::getTerminatorCondition()`?


================
Comment at: clang/lib/StaticAnalyzer/Core/BugReporterVisitors.cpp:1640-1642
+  CFGBlock *OriginB = GetRelevantBlock(Origin);
+  if (!OriginB || !NB)
+    return nullptr;
----------------
`// TODO: This can be cached.`


================
Comment at: clang/lib/StaticAnalyzer/Core/BugReporterVisitors.cpp:1646
+    if (const Expr *Condition = getTerminatorCondition(NB))
+      if (BR.addTrackedCondition(Condition))
+        bugreporter::trackExpressionValue(
----------------
All right, i still don't understand this caching based on condition expression.

You mean, like, if we're encountering the same condition multiple times (say, in a loop), we should only track it once? Why? Like, its values (which are the thing we'll really be tracking) may be different (say, on different loop iterations).


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D62883/new/

https://reviews.llvm.org/D62883