[llvm] [IR2Vec] Refactor vocabulary to use section-based storage (PR #158376)

[Kirill Stoimenov via llvm-commits]

================
@@ -43,8 +43,11 @@ class FunctionPropertiesAnalysisTest : public testing::Test {
 public:
   FunctionPropertiesAnalysisTest() {
     auto VocabVector = ir2vec::Vocabulary::createDummyVocabForTest(1);
-    MAM.registerPass([&] { return IR2VecVocabAnalysis(VocabVector); });
-    IR2VecVocab = ir2vec::Vocabulary(std::move(VocabVector));
+    MAM.registerPass([VocabVector = std::move(VocabVector)]() mutable {
+      return IR2VecVocabAnalysis(std::move(VocabVector));
+    });
+    IR2VecVocab =
----------------
kstoimenov wrote:

@svkeerthy this introduced a memory leak: https://lab.llvm.org/buildbot/#/builders/52/builds/11612

Could you please fix it? Thanks! 

https://github.com/llvm/llvm-project/pull/158376