* Added codegen.cpp to source validator whitelist, since we have the ability to call printf() from codegen for debug.

* Beefed up overflow checks in NumericRuntime.
* Fixed tests.

Author: pmenon

script/validators/source_validator.py

@@ -58,12 +58,12 @@
     "src/network/protocol.cpp",
     "src/include/common/macros.h",
     "src/common/stack_trace.cpp",
-    "src/include/parser/sql_scanner.h", # There is a free() in comments
     "src/include/index/bloom_filter.h",
     "src/include/index/compact_ints_key.h",
     "src/include/index/bwtree.h",
     "src/codegen/util/oa_hash_table.cpp",
-    "src/codegen/util/cc_hash_table.cpp"
+    "src/codegen/util/cc_hash_table.cpp",
+    "src/codegen/codegen.cpp",              # We allow calling printf() from codegen for debugging
 ]
 
 ## ==============================================

src/function/numeric_functions.cpp

@@ -216,13 +216,26 @@ T ParseInteger(const char *ptr, uint32_t len) {
   }
 
   // Convert
+  uint64_t cutoff =
+      static_cast<uint64_t>(negative ? -std::numeric_limits<int64_t>::min()
+                                     : std::numeric_limits<int64_t>::max());
+  uint64_t cutlimit = cutoff % 10;
+  cutoff /= 10;
+
   int64_t num = 0;
   while (start < end) {
     if (*start < '0' || *start > '9') {
       break;
     }
 
-    num = (num * 10) + (*start - '0');
+    uint32_t c = static_cast<uint32_t>(*start - '0');
+
+    if (static_cast<uint64_t>(num) > cutoff ||
+        (static_cast<uint64_t>(num) == cutoff && c > cutlimit)) {
+      goto overflow;
+    }
+
+    num = (num * 10) + c;
 
     start++;
   }
@@ -234,8 +247,7 @@ T ParseInteger(const char *ptr, uint32_t len) {
 
   // If we haven't consumed everything at this point, it was an invalid input
   if (start < end) {
-    codegen::RuntimeFunctions::ThrowInvalidInputStringException();
-    __builtin_unreachable();
+    goto invalid;
   }
 
   // Negate number if we need to
@@ -246,12 +258,19 @@ T ParseInteger(const char *ptr, uint32_t len) {
   // Range check
   if (num <= std::numeric_limits<T>::min() ||
       num >= std::numeric_limits<T>::max()) {
-    codegen::RuntimeFunctions::ThrowOverflowException();
-    __builtin_unreachable();
+    goto overflow;
   }
 
   // Done
   return static_cast<T>(num);
+
+overflow:
+  codegen::RuntimeFunctions::ThrowOverflowException();
+  __builtin_unreachable();
+
+invalid:
+  codegen::RuntimeFunctions::ThrowInvalidInputStringException();
+  __builtin_unreachable();
 }
 
 }  // namespace

test/codegen/value_integrity_test.cpp

@@ -198,8 +198,9 @@ void TestInputIntegral(
 
   // Default overflow tests
   std::vector<std::string> overflow_tests = {
-      std::to_string(static_cast<int64_t>(std::numeric_limits<T>::min()) - 1),
-      std::to_string(static_cast<int64_t>(std::numeric_limits<T>::max()) + 1)};
+      std::to_string(std::numeric_limits<T>::min()) + "1",
+      std::to_string(std::numeric_limits<T>::max()) + "1",
+      "123456789123456789123456789"};
   overflow_tests.insert(overflow_tests.end(), extra_overflow_tests.begin(),
                         extra_overflow_tests.end());
 

test/common/internal_types_test.cpp

@@ -325,8 +325,8 @@ TEST_F(InternalTypesTests, PlanNodeTypeTest) {
       PlanNodeType::ORDERBY, PlanNodeType::PROJECTION,
       PlanNodeType::MATERIALIZE, PlanNodeType::LIMIT, PlanNodeType::DISTINCT,
       PlanNodeType::SETOP, PlanNodeType::APPEND, PlanNodeType::AGGREGATE_V2,
-      PlanNodeType::HASH, PlanNodeType::RESULT, PlanNodeType::COPY,
-      PlanNodeType::MOCK};
+      PlanNodeType::HASH, PlanNodeType::RESULT,
+      PlanNodeType::EXPORT_EXTERNAL_FILE, PlanNodeType::MOCK};
 
   // Make sure that ToString and FromString work
   for (auto val : list) {