[clang] [clang][dataflow] Add support for serialization and deserialization. (PR #152487)

[Utkarsh Saxena via cfe-commits]

================
@@ -208,6 +208,27 @@ bool DataflowAnalysisContext::equivalentFormulas(const Formula &Val1,
   return isUnsatisfiable(std::move(Constraints));
 }
 
+llvm::DenseSet<Atom> DataflowAnalysisContext::getTransitiveClosure(
+    const llvm::DenseSet<Atom> &Tokens) const {
+  llvm::DenseSet<Atom> VisitedTokens;
+  // Use a worklist algorithm, with `Remaining` holding the worklist.
+  std::vector<Atom> Remaining(Tokens.begin(), Tokens.end());
+
+  // For each token in `Remaining`, add its dependencies to the worklist.
+  while (!Remaining.empty()) {
+    auto Token = Remaining.back();
+    Remaining.pop_back();
+    if (!VisitedTokens.insert(Token).second)
+      continue;
+    if (auto DepsIt = FlowConditionDeps.find(Token);
+        DepsIt != FlowConditionDeps.end())
+      for (Atom A : DepsIt->second)
+        Remaining.push_back(A);
----------------
usx95 wrote:

If we talk about the latest suggestion, I don't see how we will push `B` again. The idea is that when we push a token, we mark it as `TokensSeen` before even visiting it. And if a token was already "seen" it is never pushed again.

| Current           | TokensSeen | Remaining |
| :---------------- | :--------: | :-------: |
| initial condition |     A      |     A     |
| A                 |    ABC     |    BC     |
| B                 |    ABC     |     C     |
| C                 |    ABC     |   null    |


https://github.com/llvm/llvm-project/pull/152487