Kanit Therdsteerasukdi wrote:
I'm a student at UCLA taking CS130: Software Engineering. For CS130, we are writing conformance tests for Common Controls (comctl32). I will be working on the Date and Time Picker. Attached is an initial test, testing the getters and setters.
Looks like good work.
+#include <stdio.h>
This is possibly not needed.
+static HINSTANCE g_hinst;
You should be able to use NULL instead of the hinst when creating the window, which would eliminate the need for g_hinst.
+ assert(hWndDateTime!=NULL); + return hWndDateTime;
Maybe it would be nicer to check hWndDateTime with an ok() macro in test_datetime_control() then use skip() if it's NULL. This way the test will fail without crashing in the case that Window creation fails, for example, when there's no X display available. When tests crash, make doesn't detect that they failed.
+ return; +}
The return is redundant, and probably better off removed.
+ r = SendMessage(hWndDateTime, DTM_SETRANGE, GDTR_MIN | GDTR_MAX, (LPARAM)st); + ok(r != 0, "Expected nonzero, got %ld\n", r);
MSDN does say "returns nonzero if successful", but it's probably better to check the exact value that it does return, in case there's a program out there that does the same. ie.
ok(r == 1, "Expected nonzero, got %ld\n", r); Mike
