adjust moving code

[GoogleCodeExporter]

Here is my first contribution to the code:
The method "findWalkablePosition" shown here:
    private boolean findWalkablePosition(int dx, int dy) {
        if (tryWalkablePosition(sgn(dx), sgn(dy))) return true;
        if (dx == 0 || dy == 0) return false;
        if (abs(dx) > abs(dy)) return tryWalkablePosition(sgn(dx), 0);
        return tryWalkablePosition(0, sgn(dy));
    }

Should be replaced by:
    private boolean findWalkablePosition(int dx, int dy) {
        if (tryWalkablePosition(sgn(dx), sgn(dy))) return true; // try getting into the direction he is pointing at

        if (dx == 0) { // he wants to go north or south but there is an obstacle
            if (tryWalkablePosition( 1, sgn(dy)) return true; // try getting north-east (or south-east)
            if (tryWalkablePosition(-1, sgn(dy)) return true; // try getting north-west (or south-west)
            return false;
        }

        if (dy == 0) { // he wants to go east or west but there is an obstacle
            if (tryWalkablePosition(sgn(dx), 1) return true; // try getting north-east (or north-west)
            if (tryWalkablePosition(sgn(dx),-1) return true; // try getting south-east (or south-west)
            return false;
        }

        if (abs(dx) >= abs(dy)) { // he wants to go more horizontally
            if (tryWalkablePosition(sgn(dx), 0) return true; // try to go horizontally
            if (tryWalkablePosition(0, sgn(dy)) return true; // try to go vertically
            return false
        } else { // he wants to go more vertically
            if (tryWalkablePosition(0, sgn(dy)) return true; // try to go vertically
            if (tryWalkablePosition(sgn(dx), 0) return true; // try to go horizontally
            return false
        }
    }

I would like to see that in action. Could you provide a new alpha with this 
moving code?

I want to make clear what is improved. Currently the following 2 situations are 
not possible:
 .
P#..C

P#####
.#####
.#####
.....C

P - player   . - the expected way   # - obstacle   C - the point where I click

Im just curious if it works.

Original issue reported on code.google.com by SamuelPl...@gmail.com on 11 Dec 2010 at 2:20

[GoogleCodeExporter]

You can download the source code and compile it on your computer to try it out. 
I would really encourage you to try it, so you can test other things as well if 
you want. I'll email you more info.

My very brief tests show that example 2 above actually does work with the 
current implementation. Example 1 would require a pathfinder to select the 
correct movement position. Compare to:
  1#
 P## C
  2..
should the player move to 1 or 2 when clicking C? To determine that, you would 
have to find a correct path from P to C.

Original comment by oskar.wi...@gmail.com on 12 Dec 2010 at 11:19

[GoogleCodeExporter]

Perhaps I will try to compile on my own.

My tests show that example 2 does not work. (Crossglen village tavern is the 
"#"-building)
P#####
.#####
.#E###
..c.C

It works if I click "c" but it doesn't work if I click "C".

I'm aware of the pathfinder situation. But I more often face a situation where 
a single obstacle is in my way (a), or where the other way is not possible. (b)

a)
 .
P#..C

b)
########
P##...C
 ..

IMO the pathfinder would have disadvantages on hold to move. Consider the 
following situation: (c)
c1)       c2)       c3)
       # |      ## |     ##
   P..c# |   P.Sc  | ..PS#c
   ##### |  #####  |.#####.
         |         | .....

If I hold to move and the display scrolls, then every time I accidentally cross 
a border my hero will move in a wrong direction. (P wants to go to S; c is 
where the user holds to move)

IMO a direction oriented system would be better here. (As it is now.)

Original comment by SamuelPl...@gmail.com on 12 Dec 2010 at 3:55

[GoogleCodeExporter]

i really like how u move around on the game i would like to keep it that way if 
u were to change is it possible to have it as setup the way u like it?

Original comment by breid3...@hotmail.com on 11 Jan 2011 at 5:54

[GoogleCodeExporter]

This would not significantly change the way you move.
It would just reduce the times you get stuck in a tree or fence.

But I agree to you, that it should be optional.

Original comment by SamuelPl...@gmail.com on 13 Jan 2011 at 11:30

[GoogleCodeExporter]

I would like to see more touch-based movement options.

• North, south, east, west (no diagonals, and touch-to-move disabled)
• Swipe-to-move (distance of swipe determines distance of travel, 
configurable as "sensitivity")
• An on-screen directional controller pad

Hopefully, these are easy to implement.

Original comment by theGl...@bigfoot.com on 17 Jan 2011 at 2:39

[GoogleCodeExporter]

Nah, the movement controls are just perfect (galaxy s). Only a pathfinder would 
be brilliant.

Original comment by michisch...@web.de on 18 Jan 2011 at 1:42

[GoogleCodeExporter]

An optional translucent on screen D-Pad would be awesome IMO. If I get some 
free time this weekend I will try to build one and see how it goes, I may need 
some help getting the option into the options menu at some point.

While I'm doing this I'll look at refinement options for the movement code and 
maybe build out some pathfinder logic where the pathfinder is disabled by press 
and hold.

Original comment by divilsp...@gmail.com on 18 Jan 2011 at 5:43

[GoogleCodeExporter]

any word on an upcoming update for this game? been waiting forever 

Original comment by jesscahi...@gmail.com on 18 Jan 2011 at 6:23

[GoogleCodeExporter]

@divilsp...@gmail.com:
I already implemented an pathfinder. (See attachment)
But AFAIK Oskar did not yet decide if he should include it or not. (performance 
reasons)

IMO both the improved movement code and the pathfinder work very well so far.

@Jfaulkne...@yahoo.com:
Only Oskar knows. AFAIK he is working on content.
I expect an alpha or beta in this or in the next week.
But remember: I dont know anything.

Original comment by SamuelPl...@gmail.com on 18 Jan 2011 at 9:28

Attachments:

• MovementController.java

[GoogleCodeExporter]

At first glance this looks pretty awesome. Being as you have both pieces 
working how you like, have you tried breaking out of the pathfinder on press 
and hold? I think this would be a nice balance between performance and 
usability. It might also be worth considering a range limitation on the 
pathfinder if performance does suffer that badly; though it could always be 
masked by having move animation between the tiles and doing the compute during 
the animation.

For the sake of my first programming instructor.. "MORE COMMENTS!!!" :-) Even 
when the code was 80% comments he wanted more.

Original comment by divilsp...@gmail.com on 18 Jan 2011 at 9:38

[GoogleCodeExporter]

IMO the performance is good.
I tried an earlier version of it with my wildfire and did not have problems at 
all. After that I improves the code for even more performance. (I think I 
somewhat doubled the speed) Did not experience any difference.

@MORE COMMENTS:
I try to comment by writing readable code. IMO good code doesn't need many 
comments.
Just my 2 cents. (And... Oskar did comment rarely too ;-))

Original comment by SamuelPl...@gmail.com on 18 Jan 2011 at 9:52

[GoogleCodeExporter]

pathfinding is the one, A* ?

Original comment by rqpa...@gmail.com on 3 Feb 2011 at 6:11

[GoogleCodeExporter]

As the target can not necessarily be reached I cannot use A* in generally.
Perhaps A* could be adjusted somehow, but IMO you need to visit each 
(reachable) field anyway.

Original comment by SamuelPl...@gmail.com on 3 Feb 2011 at 7:14

[GoogleCodeExporter]

Waiting for a review

Original comment by sdeva...@gmail.com on 10 Feb 2011 at 1:00

• Added labels: Priority-Low, Type-Enhancement
• Removed labels: Priority-Medium, Type-Defect

[GoogleCodeExporter]

What do you mean with: "Waiting for a review"?

Original comment by SamuelPl...@gmail.com on 10 Feb 2011 at 1:08

[GoogleCodeExporter]

It was just a reminder to myself to review the submitted code.

Oskar explained it uses too much memory and will need to be improved before it 
is included but it definitely could be a great enhancement.

Original comment by sdeva...@gmail.com on 10 Feb 2011 at 5:33

• Changed state: Started

[GoogleCodeExporter]

Mh I think its not needed anymore, right?
I dont miss it and it looked strange sometimes.

Original comment by SamuelPl...@gmail.com on 21 Jul 2011 at 8:41