--- In [email protected], "Sheri" <sheri...@...> wrote: > > --- In [email protected], "entropyreduction" > <alancampbelllists+yahoo@> wrote: > > > > --- In [email protected], "Sheri" <sherip99@> wrote:
> > accChild(i) will always return an exception if the child is not a full > > IAccessible object, i.e. if it's not associated with an IDispatch. > > > > http://msdn.microsoft.com/en-us/library/dd318475%28VS.85%29.aspx > > The page you referenced is for the function, get_accname, not the property > h.accname used in the referenced script. > None of the functions currently work. > I read somewhere that you cannot rely on com macros for determine the success > of accessible functions. > <http://msdn.microsoft.com/en-us/library/dd318012%28VS.85%29.aspx> Yup, that was the problem that was crashing IAccessible script when accessing wordpad. > Perhaps the plugin should have its own versio of each functions? That's where I'm going with the new options for get_description (state, role) and new numeric argument for get_accessible_child. > I think if the MSAA interface were being used correctly, the interface > functions would work and the complexity of the fact that some children are > simple elements and some objects are would be handled. MSAA interface were sensible, you could use it sensibily. But it isn't, it throws exceptions when you'd expect it to do otherwise, so yes, I'll need to wrap it in something vaguely useful. > > Quite common for an object not to have a default action. > > If a child doesn't have a property (such as default action) the function that > gets the property (but for, us no interface functions currently work) is > supposed to return a non-fatal "not implemented" result. I believe the invoke fails, and com plugin, if it sees invoke failing, sensibly calls it an error. I can't pick and choose with error codes to treat as error and which as soemthing else. Unless of course I wrap accessibility invokes around in their own special services, or parameterised versions of exisiting ones.
