Thanks Jeremy for the patch. I worked a bit more on doc. On Tue, 24 Nov 2009 08:28:43 -0600, Jeremy Hankins wrote: > But we should obviously get rid of either window-relative-position > or window-absolute-position.
I noticed they're not exactly the same. window-absolute-position returns the adjusted position only if it's completely outside of the current VP, but window-relative-position always uses the window-viewport. They have to be united into one, perhaps using an optional arg, but I don't know exactly what this difference means. Is it possible for you to complete it? > But what should the name be? To my mind > window-absolute-position seems like a very strange and non-intuitive > name for something that gives the position of a window relative to the > vp it occupies, but what do others think? I don't feel like another "incompatible change" notice (since there remain tasks, many for me, for 1.6 release). But my best candidate is "window-position-in-viewport" >>>> I've renamed viewport-offset to viewport-offset-pixel. > [...] > I disagree. [...] Ok, then I'll change it to "viewport-offset-coord". Anyway the suffix is necessary, and "pixel" in doc will suffice. (In fact, when I browsed viewport.jl before, I couldn't tell if the warp-viewport's args are pixels or vp slot indices even though it is written "coordinates" in the doc-string. So "pixel" in doc is needed.) By the way, I've noticed one more confusing factor in VP. I think your explanation "think of vp as a hole" is correct. But the word "vp" is used also for the vp cell. It's sometimes useful, like "go to the next vp", rather than "move the vp to the next slot", but you have to be aware that it's an extended use. (So, "the viewport" is ok, but viewport"s" means you have two heads or 4 eyes. ::-) Regards, Teika (Teika kazura)
