https://bugs.documentfoundation.org/show_bug.cgi?id=168628
--- Comment #6 from jan d <[email protected]> --- The behavior the the selection rectangle makes sense to me, it behaves as I expected it: It is a rectangle enclosing all objects selected, and its shape is based on the current view as I see it (so it changes with different page layouts). As Heiko pointed out, not seeing the whole rectangle at any point implies that there are more shapes selected that are currently not fully in view. As for what should happen in the realignment: It seems that the command does what I expect it to do when I have the pages vertically layed out and then do a center horizontally: Both object are aligned with each other. Same for middle-vertically when pages are layed out horizontally, based on the current view mode and selection; the model seems to be that the whole document as one space and the new positions are calculated based on this. This makes sense, since the align-action is inherently visual. If I have the pages layout out vertically and ask to middle-vertically align objects on different pages, this model is not entirely upheld, though: The algorithm seems to try do a "try to act upon the command as good as possible but stop at the page boundaries", which seems to be a saveguard for objects jumping to other pages. I.e. The current behavior of the selection makes sense to me, the aligment behavior mostly makes sense to me, and the problems seem to be created by compromizes that make different concepts of how objects should behave (mostly) work well in parallel. -- You are receiving this mail because: You are on the CC list for the bug.
