4:09:02lokeRight now the actualy rendering function doewsn't return anything. That means that the calling code needs to first render the text, and then call the string-text-extents function to determine where to place the cursor afterwards. There issue is that drawing the string already computed this, and performing this measurement requires you to perform all the rendering (except not actually drawing it to the screen), meaning that you end up
4:10:15jack_rabbitThat shouldn't be a breaking change, though, so I would think it's ok.
4:11:28lokejack_rabbit: My proposed change is to have the render function return the same values as string-text-extents does, so the caller can reuse it.
4:12:00lokeI guess the functionc an be defined such that if it return NIL, the caller will call the string-extends function instead
4:13:37jack_rabbitThat makes sense. I mean a change from an API function that returns NIL to one that returns *something* is probably the easiest, least dangerous API change one can make.
4:14:10lokejack_rabbit: well, the current function doesn't return anything at all
4:14:21loke(at least the return value isn't used, yet.)
4:14:49jack_rabbitmmm Is it documented to return something, or is it just some coincidental object?
4:15:20lokeIt's part of clim-clx. So it's not documentet at all.
4:15:52lokeThe return value ends up being returned by the call to MEDIUM-DRAW-TEXT*
4:18:23lokeI think what happens is that it calls the medium-text-extents when creating the output record.