Created attachment 311784 [details] [review]
Doc: fix some gtk-doc warnings

With that patch and the one from bug #744061, remaining issues are:

1) g_set_object() inline confuses gtk-doc and it complains for unused old_object.

2) gio has undocumented win32 APIs:
GWin32RegistryKey
GWin32RegistryKeyWatcherFlags (<items>)
GWin32RegistrySubkeyIter
GWin32RegistryValueIter
GWin32RegistryValueType (<items>)

Created attachment 311791 [details] [review]
Doc: fix some gtk-doc warnings

Fixed the doc of g_file_monitor_emit_event() that was removed in commit 737ab3201163631be152801a859b3874a667f10. It is now the implementation that is responsible to call that method from the right thread if I understand correctly.

Review of attachment 311791 [details] [review]:

Looks good.  Thanks for playing this game :)

::: gio/gfilemonitor.c
@@ +273,3 @@
+ * Implementations are responsible to call this method from the
+ * [thread-default main context][g-main-context-push-thread-default] of the
+ * thread that the monitor was created in.

This is not _strictly_ true, but it's good enough.  In fact, I think it's true for all external users of the API...

Thanks, pushed. Also fixed missing g_sequence_is_empty() and g_socket_connectable_to_string() while at it.