Paint program abilities

Quote:

---

Originally posted by joey o.
Zaei, memory is cheep now and even if you got a little vb's ok with it. In vb you might as well save the pics to a memdc wether in a control or device independent 'cause your gonna loose on processing trying to calculate various regions anyway. I tryed that and found no gain. Also once a pic is drawn over if you depend on arrays to undraw what you have you will again be clobbered by memory issues if someone draws for a full minute without lifting the mouse. Using hidden pics and their dc's uses abrox 2% of my active memory (56k, win98 on a 333)to record a full screen pic. I have no reason to open 35 full screen bmp's but I can if I want to. This shouldn't be a concern. XP shows no difference at all for me but I've got 256 on that one.

---

And what If I wanted a 50 level undo? Going by CBs numbers, that would require 61.2 megabytes of memory, just for the picture. This doesnt include the overhead required for windows to store everything in its own DC, nor does it include VBs own overhead.

Using arrays, you add a single state to your undo stack/linked list when the mouse is released. As I stated above, The only real challenge is figuring out how to store the states in a generalized manner.

Z.

Please bear with me as I want to learn: I keep hearing about "states". What exactly what are these these states? Is it a memory location (memory dc), an array of past actions, or what? If it is an array of past actions then as far as I know you need to store the locations and pixels you may be covering which means you have to record them AS you draw. I want to know a better way but an example might help.

I know if you simply write the region of leftmost-topmost x,y - bottommost-rightmost of the area drawn in a single mousedown/move/up action you might save some memory but I still fail to see how a few lines coverred with big lines can magically reappear without writing and calling their pixels and locations from memory weather from a control's device context or directly from a created memory object.

Also I tryed two other methods.
1. I wrote every pixel drawn in a mousemove to a two dimentional array (one for each action the other for the pixels positions using point api and color as long) and thoroughly stopped dead after 2 minutes of drawing without lifting the mousebutton. I also tryed streeming them into a file after a certain size was reached-stil slow, I then tryed a timer to reduce the pixels recorded. That worked but I lost quality in my drawing. I'd rather know my number of pictures available for display than to chance crashing because someone wants to draw a big long never ending scribble.

2. The other method was to draw using a backbuffer. A lot of the on line code I've seen uses this method but drawing to to dc's at the same time was delayed quite a bit on my machine.

Please don't get me wrong. I want a better way. Please help to provide us with that way by way of example.