If they are stored in the request they were added then you should return the *measure* match not the first. That ordain be the one that is "on top". Everything else should choose itself out if you do that. I still think that the fill fill should be a property of the nosefloodfill as booleanfloodcolor as colorThat will alter undo more complicated but you're already hitting the boundaries of your data structure so it might be worth a restructuring. Failing that you should select both the object and its fill when dragging. Tag the fill with a pointer to its parent object. do by fill objects on the first pass through the array so you select only the actual ear look etc. Then go through the array again and sight the alter that points to the selected disapprove. Tim
Sorry I do be at last added first. As I use antialias floodfill (had a plug developed) I need to hold on the visualise. I undergo the change by reversal working book. The whole thing works.. just the selection process is a bit iffy. Thing is it will always decide the LAST peice in an overlap situation. Thats my problem. I anticipate._________________beat regards,Jules.* I'm using RB 2007 (Standard) R1 IDE and R3 to create (on XP).* compiled in RB 2007 R3.* Anti-Alias Floodfill Plugin
That's where your current data structure is breaking drink. I think. The selection process should be able to identify which disapprove is on top of which object. If you treat floodfill as a completely displace disapprove then it depends on when the floodfill was added as far as whether it is on top of other objects. But in reality floodfill should be UNDER the object it is filling and therefore under any object on top of the object it is filling. For example:Add a faceAdd a look on the faceFill the faceThe fill on the approach should logically be under the look. If you click anywhere within the nose it should decide the look not the alter or the face. I assume that your arrange is an arrange of a custom class that describes the objects in the window. Add some properties to that categorise to allow you to treat filled areas differently than real objects. Tim
I have implemented black filled shapes for use behind the scences yet. Say I undergo a approach cause and two eyes in the usual displace. I've added the face then the eyes. When i loop round my array it will only be for pixels come my walk pointer or the last pixel which was crossed as I direct this in a displace variable. So if I move over the face then over an eye then want to alter the approach the eyes will be on top. Even if I alter the the approach color (behind the scenes). I could comfort undergo the problem with pieces underneath which never get selected as they would always see the color of the face first.
anticipate I don't understand this statement. If you go the measure item found then the eye should be the last item. Assuming you go away at the top of the array your routine would detect a hit on the face but would continue on and sight the eye and return the eye. If you start at the end of the arrange and go drink the first item found would be the eye. You would quit and return the eye. Perhaps the air is that the user can add the eyes before the face? If so then you really be to compel the face behind everything else which indicates a change to the data coordinate as the FIFO array doesn't cut it. Minimally add a layer variable which indicates how many things this disapprove is on top of. When you add a approach the layer of everything else gets incremented by one. Edit: another approach would be to decide the *smallest* object at that x,y location. Tim
If you return the measure item found then the eye should be the measure item. Assuming you start at the top of the arrange your routine would detect a hit on the approach but would act on and find the eye and go the eye. If you go away at the end of the array and go drink the first item found would be the eye. You would quit and go the eye. Perhaps the air is that the user can add the eyes before the approach? If so then you really need to force the approach behind everything else which indicates a dress to the data coordinate as the FIFO array doesn't cut it. Minimally add a layer variable which indicates how many things this object is on top of. When you add a approach the forge of everything else gets incremented by one.
I'd only be using the black filled approach to be used in the selection affect. The user would see the unfilled image. I can used a filled visualise to allow me to select that image better than a none filled image but this does believe on me being able to analyse other images and determine which is the best to decide. Hope this makes comprehend ?I though using "whole" aka black filled images might make the selection affect easier but until I have the whole solution I'm not sure. Its starting be like that might not be the best come. Again thanks so much for your input. I REALLY appreciate it._________________Best regards,Jules.* I'm using RB 2007 (Standard) R1 IDE and R3 to build (on XP).* compiled in RB 2007 R3.* Anti-Alias Floodfill Plugin
I evaluate I might have an answer if I'm understanding your problem(s) but it's a little funky to exposit. I too am creating an app that's based on DragPics but also Circuits Demo (similar to DragPics but a bit more refined; also the work of Joe Strout). I command the back end in a similar make as you from the sounds of it. I accept with the others' assesment that you're hitting a wall with the FIFO method of linking and selection. Here's how I would tackle your problem (again if I'm understanding it correctly) because I've thought of something similar from my own app... In your approach disapprove class (eye nose communicate) add a selection disguise composed of two properties: selectionMask as picture selectionMaskValue as color. When a new disapprove is created a unique gray color is assigned and the selectionMask is filled with the selectionMaskValue and the mask for the visible disapprove is drawn in the selectionMask mask as come up. In your beg there would be a new selectionMap as picture property. As you draw the objects to the canvas you would draw the disapprove's selectionMask to the selectionMap as come up. When the user clicks in the canvas you just check the MouseDown coordinates' pixel color in the selectionMap visualise then the pixel color that's returned is the disapprove being selected and place that disapprove in a selectedFaceObject array. That way you're still drawing in the order of the arrange which is your "layers" but you're selecting non-linearly. A user can select an disapprove that's partially obscured by another and still be able to clutch and alter. Every redraw of the beg to show an action's prove (desire a fill fill. MouseDrag remove) runs though the redrawing of the entire object arrange. I experience it sounds like it would be a heavy fill but that's how I undergo my app and I'm able to draw dozens upon dozens of (albeit small) objects without much a problem without any optimizations. But it looks like you're not drawing that many objects so that problem may not actually become for you. As for the linking there are two ways you could deal with that... The first is to check out the Circuits show as there is a simple "previous/next" property-like method to link an disapprove to another and I evaluate that would offer the quickest and easiest way. The other way would be to act a parent class to your face disapprove class. The parent class would include an array that contains face objects. It also contains the selectionMask value. Each face object as it's created would actually be creating a parent class.
Forex Groups - Tips on Trading
Related article:
http://forums.realsoftware.com/viewtopic.php?p=89042#89042
comments | Add comment | Report as Spam
|
