erratic behavior with mouse messages

im having problems with properly catching mouse click messages, eg WM_RBUTTONUP and WM_RBUTTONDOWN. for example, when you simply click a mouse button, the appropriate mouse button up code is not called. to make it call, you either have to double click, or press it down, move the mouse a small distance, and release. (all this is without any other mouse capture code, just WM_RBUTTONUP). I am mystified as to why this is occurring. has any body got any ideas about this?

i just found something interesting with spy++. right after the WM_RBUTTONDOWN message, a WM_CAPTURECHANGED message sends and returns. i looked it up, and i have been trying to use ReleaseCapture() to reverse whatever is happening, but it isnt working.

just thought that may be useful.

i found out the relevance of the WM_CAPTURECHANGED message. here is a crappy diagram of the message sequence:

Code:

---

WM_RBUTTONDOWN
    S WM_CAPTURECHANGED
    R WM_CAPTURECHANGED
    WM_MOUSEMOVE (single message)
WM_RBUTTONUP
S WM_CAPTURECHANGED
R WM_CAPTURECHANGED

---

the above diagram describes an instance where the code is not captured.

the next diagram shows the sequence of messages when the mouse is dragged a bit, thus, when the code captures effectively.

Code:

---

WM_RBUTTONDOWN
    S WM_CAPTURECHANGED
    R WM_CAPTURECHANGED
    WM_MOUSEMOVE (multiple messages)
    S WM_CAPTURECHANGED
    R WM_CAPTURECHANGED
    WM_MOUSEMOVE (multiple messages)
WM_RBUTTONUP

---

clearly, you have to wait for the second set of WM_CAPTURECHANGED messages before releasing the mouse button, but why? and how do you stop this happening?