2013/9/30 Nikolay Sivov <bunglehead@gmail.com>

On 9/30/2013 00:51, Daniel Jeliñski wrote:

+struct progress_list {
+    const DWORD progress_retval_init;  /* value to return from progress routine */
+    const BOOL cancel_init;            /* value to set Cancel flag to */
+    const DWORD progress_retval_end;   /* value to return from progress routine */
+    const BOOL cancel_end;             /* value to set Cancel flag to */
+    const DWORD progress_count;        /* number of times progress is invoked */
+    const BOOL copy_retval;            /* expected CopyFileEx result */
+    const DWORD lastError;             /* expected CopyFileEx error code */
+} ;

 I don't see a point making them 'const'.

+static DWORD WINAPI progress(LARGE_INTEGER TotalFileSize,
+        LARGE_INTEGER TotalBytesTransferred,
+        LARGE_INTEGER StreamSize,
+        LARGE_INTEGER StreamBytesTransferred,
+        DWORD dwStreamNumber,
+        DWORD dwCallbackReason,
+        HANDLE hSourceFile,
+        HANDLE hDestinationFile,
+        LPVOID lpData)
+{
+    progressInvoked++;

Please pass all globals as context data with lpData, and please use 'void*' instead of LPVOID.

Good point about lpData. Still, does that make the patch invalid? Why didn't you mention that on the first review?