core: FileStorage reads integers above INT_MAX into float/double without truncation

FileNode::operator double() and operator float() read an INT node via readInt()
(32-bit), truncating values above INT_MAX -- e.g. an integer 6662329666 from an
externally-produced json/yaml/xml is read back as -1927604926. The node stores
the value as int64 (operator int64_t() already reads it correctly via readLong),
so use readLong() for the floating-point conversions too. Values that fit in
int32 are unchanged (sign-extended); larger ones are now correct.

Reader side of #29363 (the writer side was #29364).

modules/core/src/persistence.cpp

@@ -2376,7 +2376,7 @@ FileNode::operator float() const
 
     if( type == INT )
     {
-        return (float)readInt(p);
+        return (float)readLong(p);
     }
     else if( type == REAL )
     {
@@ -2397,7 +2397,7 @@ FileNode::operator double() const
 
     if( type == INT )
     {
-        return (double)readInt(p);
+        return (double)readLong(p);
     }
     else if( type == REAL )
     {

modules/core/test/test_io.cpp

@@ -2099,6 +2099,24 @@ TEST(Core_InputOutput, FileStorage_int64_26829)
     }
 }
 
+TEST(Core_InputOutput, FileStorage_read_bigint_as_real_29363)
+{
+    // An integer above INT_MAX (e.g. from externally-produced json/yaml/xml) must
+    // convert to float/double without truncating to int32. Regression for #29363.
+    String content =
+        "%YAML:1.0\n"
+        "a: 6662329666\n"   // ~6.6e9, exact in double
+        "b: -9876543210\n"
+        "c: 4294967296\n";  // 2^32, exact in double and float
+    FileStorage fs(content, FileStorage::READ | FileStorage::MEMORY);
+
+    EXPECT_EQ(6662329666.0,  (double)fs["a"]);
+    EXPECT_EQ(-9876543210.0, (double)fs["b"]);
+    EXPECT_EQ(4294967296.0,  (double)fs["c"]);
+    EXPECT_EQ(4294967296.0f, (float)fs["c"]);
+    EXPECT_EQ((int64_t)6662329666LL, (int64_t)fs["a"]);  // int64 path unchanged
+}
+
 template <typename T>
 T fsWriteRead(const T& expectedValue, const char* ext)
 {