Every example assumes you have an xcb_connection and possibly other variables at hand. For illustrating how xcb_get_property works, you need the window of which you want to get the property, for example. To make it clear that these variables are your responsibility, these examples consist of a single function which takes the necessary variables as parameters.
Flushing means calling xcb_flush to clear the XCB-internal write buffer and send all pending requests to the X11 server. You don't explicitly need to flush before using a reply function (like xcb_query_pointer_reply), but you do need to flush before entering the event loop of your program.
There are only two cases when XCB flushes by itself. The first case is when its write buffer becomes full, the second case is when you are asking for the reply of a request which wasn't flushed out yet (like xcb_query_pointer_reply). This last point also includes xcb_request_check(). Please note that waiting for an event does NOT flush.
Examples generally include the xcb_flush call where appropriate (for example after setting a property). Therefore, including these functions and calling them in your application should just work. However, you might get better results when flushing outside of the function, depending on the architecture of your program.
If an example does not compile (without warnings) when using -std=c99, that is considered a documentation bug. Similarly, not handling errors or leaking memory is also considered a documentation bug. Please inform us about it on xcb@lists.freedesktop.org.
Every example uses 4 spaces for indentation.
Comments are in asterisks, like /* this */.
No line is longer than 80 characters (including indentation).