VariantChangeType bug or maybe not?
-
ok guys, this is driving me crazy:
#include "comdef.h" HRESULT fn(VARIANT v) { HRESULT hr = VariantChangeType(&v, &v, 0, VT_BSTR); return hr; } int main(int argc, char* argv[]) { variant_t v1(L"1"); variant_t v2(L"2"); fn(v1); fn(v2); // changes v1?? return 0; }now, when fn(v2); is called, the value of v1 is changed to "2". i believe that i'm doing something wrong, but i can't see what. -
ok guys, this is driving me crazy:
#include "comdef.h" HRESULT fn(VARIANT v) { HRESULT hr = VariantChangeType(&v, &v, 0, VT_BSTR); return hr; } int main(int argc, char* argv[]) { variant_t v1(L"1"); variant_t v2(L"2"); fn(v1); fn(v2); // changes v1?? return 0; }now, when fn(v2); is called, the value of v1 is changed to "2". i believe that i'm doing something wrong, but i can't see what.Perhaps pass v to fn() by reference? i.e. HRESULT fn(VARIANT& v); "You can stand all night at a redlight anywhere in town, hailing Marys left and right but none of them slow down. I've seen the best of men go past. I don't wanna be the last..."
-
Perhaps pass v to fn() by reference? i.e. HRESULT fn(VARIANT& v); "You can stand all night at a redlight anywhere in town, hailing Marys left and right but none of them slow down. I've seen the best of men go past. I don't wanna be the last..."
That would work correctly, and i would do that if it was a "real" code. but the code in the snippet is perfectly legal c++ (or at least i think so) so i'm not sure what's the reason for such behavior
-
That would work correctly, and i would do that if it was a "real" code. but the code in the snippet is perfectly legal c++ (or at least i think so) so i'm not sure what's the reason for such behavior
You are right. Do I feel like an idiot or what? :) However, the reason for this strange behaviour (I think) is that the BSTR is deallocated within ChangeType and then reallocated. (It probably doesn't check whether it is changing into the same type, i.e. BSTR->BSTR) This means that the new pointer is lost since it was assigned to a temp. object. The _variant_t contains the old pointer that now points to deallocated memory. I.e. the _variant_t BSTR pointer is dangling after the fn() call. Then the ghost value is a result of the memory allocation implementation. "You can stand all night at a redlight anywhere in town, hailing Marys left and right but none of them slow down. I've seen the best of men go past. I don't wanna be the last..."
-
You are right. Do I feel like an idiot or what? :) However, the reason for this strange behaviour (I think) is that the BSTR is deallocated within ChangeType and then reallocated. (It probably doesn't check whether it is changing into the same type, i.e. BSTR->BSTR) This means that the new pointer is lost since it was assigned to a temp. object. The _variant_t contains the old pointer that now points to deallocated memory. I.e. the _variant_t BSTR pointer is dangling after the fn() call. Then the ghost value is a result of the memory allocation implementation. "You can stand all night at a redlight anywhere in town, hailing Marys left and right but none of them slow down. I've seen the best of men go past. I don't wanna be the last..."
yes, i came to the same conclusion during the lunch :) thanks for help, anyway