Hi Thomas, On September 22, 2010 01:55:55 pm T. Modes wrote: > Please don't at the current status.
OK, will think of something so that it won't interfere. > First, it does not work correctly if you undo/redo a command which > changes the number or order of the images (add/remove image, swap > images and so on). In this case you switch the active state of not > affected images. Yes. I find this to be a "cosmetic" issue. I might as well make them all visible. I don't need the visibility history all together. Or... > Second, I don't think it is a good idea to modify the panorama object > directly (even in the CommandHistory::Undo/Redo). All modifications of > the panorama object should happen inside a PanoCommands, which are > push to the command stack. All other can produce side effects which we > don't consider at the first glance. Doesn't it become a chicken and egg problem? I mean, I want to skip over those PanoCommands that toggle visibility. But then, if I simplify the patch, give up completely the visibility thing and just simulate the multiple repeated undo/redo? then it would not modify the panorama object directly? I can always issue a "show all" PanoCommands at the end of the sequence. Should I? Yuv
signature.asc
Description: This is a digitally signed message part.