Hi Marcus,
On 09/29/12 11:44, Marcus Meissner wrote:
> Hi,
>
> Various coverity issues complain about user-after-free scenarios,
> all involving this code path.
>
> I strongly think if call_ret signals error, we also need to return
> an error condition to avoid the callers from proceeding as if nothing
> happened.
Agreed, thanks for catching it.
> Ciao, Marcus
> ---
> dlls/rpcrt4/rpc_transport.c | 2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/dlls/rpcrt4/rpc_transport.c b/dlls/rpcrt4/rpc_transport.c
> index 686ce68..e572b78 100644
> --- a/dlls/rpcrt4/rpc_transport.c
> +++ b/dlls/rpcrt4/rpc_transport.c
> @@ -1882,7 +1882,7 @@ static RPC_STATUS wait_async_request(RpcHttpAsyncData
> *async_data, BOOL call_ret
>
> if(call_ret) {
> RpcHttpAsyncData_Release(async_data);
> - return RPC_S_OK;
> + return call_ret; /* The Http* status codes map into the RPC_S
> namespace */
Not really, Http* APIs return BOOL and GetLastError() call is needed to
get error code. AFAIR, when I tested this code, most HTTP-related errors
were translated to RPC_S_SERVER_UNAVAILABLE by rpcrt4. That may be a
better choice here as well.
Jacek
