- Joined
- Aug 7, 2013
- Messages
- 1,338
Hi,
A follow up post from the previous one.
I have a struct which is recursive, called a Quad. A Quad is basically a wrapper around a JASS rect, but each Quad also an array of Quads which in term can have an array of Quads and so on. In addition to an array of Quads, Quads come up an array of JASS lightning of size 4 (you need 4 lightnings to make any rect visible).
I have an algorithm that divides up each Quad into 2 - 5 Quads (random everytime). The algorithm also takes advantage of the recursive structure of Quads, as depending on how deep it goes (a parameter you pass on the algorithm call), it will apply the algorithm to each child Quad, and to the children of those child Quads, and so on and so forth.
Now I notice that for whatever reason, my game either stops making rects, lightning effects, or both if I call the algorithm too many times. I have been of course debugging it by calling the algorithm with some parameter, and then when I reset the parameter, it cleans up all the previous work, and does a new one. But clearly it's not successful in deallocating all the memory, because the algorithm stops working.
Here is my procedure that frees up a Quad, and then calls the procedure on any of its non-zero Quad array members (pseudo-recursive).
Here is my interface with using the algorithm in game (a player chat string event).
The highlight() method does what its name says--it illuminates the boundaries of a given rect with four JASS lightnings. Here it is.
Now, I don't provide the patitionAlg() because it suffices to know how it works. For each Quad.partitionAlg(...), the algorithm creates anywhere from 2 - 5 Quads. Thus, at the end of each call for a given Quad, we can get up to 6 Quads total (since we count the Quad itself). Each Quad is a rect, and each Quad has an array of 4 JASS lightning.
So in the 'worst' case scenario, we will have six rects and 24 lightning effects for each Quad.
Given a level of iterations (when the algorithm will stop applying to the Quad array members), the maximum rects created is so:
iter = 0 : 6 rects
iter = 1: 6 + (6 x 6) rects
iter = 2: 6 + (6 x 6) + (6 x 6 x 6) rects
...
iter = n: 6 + (62) + (63) + ... + (6n + 1) rects
which we can condense into (changing our notation by +1 as well):
The amount of JASS lightning is simple 4 x the summation above, since every Quad gets 4 lightnings.
The algorithm breaks in this total. What happens is that it only partially cleans up the the last series of Quads, and then subsequent calls don't do anything
Sum(2) + Sum(3) + Sum(3) = 522 rects and 2088 lightning effects.
If I then do Sum(3), here is what happens.
Note that these are all called linearly temporally--I do Sum(2) then Sum(3) then Sum(3) then Sum(3). When I do a new call, I clean up the screen using the flush() method I showed.
Here is where it breaks. If you notice, this actually just the remnants of the above call. But it doesn't clean up all of the rects. And if I try any other calls, these lightning effects/rects are here for good. Everything stops at this point.
So from what I showed, at some point it fails to clean up all the rects/lightning effects. And then it won't make any more rects or lightning effects. This is peculiar to me. Any ideas what's going on?
A follow up post from the previous one.
I have a struct which is recursive, called a Quad. A Quad is basically a wrapper around a JASS rect, but each Quad also an array of Quads which in term can have an array of Quads and so on. In addition to an array of Quads, Quads come up an array of JASS lightning of size 4 (you need 4 lightnings to make any rect visible).
I have an algorithm that divides up each Quad into 2 - 5 Quads (random everytime). The algorithm also takes advantage of the recursive structure of Quads, as depending on how deep it goes (a parameter you pass on the algorithm call), it will apply the algorithm to each child Quad, and to the children of those child Quads, and so on and so forth.
Now I notice that for whatever reason, my game either stops making rects, lightning effects, or both if I call the algorithm too many times. I have been of course debugging it by calling the algorithm with some parameter, and then when I reset the parameter, it cleans up all the previous work, and does a new one. But clearly it's not successful in deallocating all the memory, because the algorithm stops working.
Here is my procedure that frees up a Quad, and then calls the procedure on any of its non-zero Quad array members (pseudo-recursive).
JASS:
method flush takes nothing returns nothing
local integer i = 0
loop
exitwhen i == BOLTS_PER_RECT
call DestroyLightning(bolts[i])
set bolts[i] = null
set i = i + 1
endloop
set i = 0
loop
exitwhen i == MAX_QUADS
if subQuads[i] != 0 then
call subQuads[i].flush()
endif
set i = i + 1
endloop
call RemoveRect(r)
set r = null
call this.destroy()
endmethodHere is my interface with using the algorithm in game (a player chat string event).
JASS:
scope RectPlay initializer init
globals
Quad q
rect r
endglobals
private function main takes nothing returns boolean
local integer iter = S2I(SubString(GetEventPlayerChatString(), 1, 2))
call q.flush()
set q = Quad.create(Rect(-13440, 9184, -11744, 10688))
call q.highlight()
call q.partitionAlg(iter)
return false
endfunction
private function init takes nothing returns nothing
local trigger t = CreateTrigger()
local Quad p
call TriggerRegisterPlayerChatEvent(t, Player(0), "-", false)
call TriggerAddCondition(t, Condition(function main))
set q = Quad.create(Rect(-13440, 9184, -11744, 10688))
call q.highlight()
call q.partitionAlg(1)
endfunction
endscopeThe highlight() method does what its name says--it illuminates the boundaries of a given rect with four JASS lightnings. Here it is.
JASS:
method highlight takes nothing returns nothing
local integer i = GetRandomInt(0, TOTAL_LIGHTNINGS)
set bolts[0] = AddLightningEx(LIGHTNING[i], false, minX, minY, 0.0, minX, maxY, 0.0)
set bolts[1] = AddLightningEx(LIGHTNING[i], false, maxX, minY, 0.0, maxX, maxY, 0.0)
set bolts[2] = AddLightningEx(LIGHTNING[i], false, minX, minY, 0.0, maxX, minY, 0.0)
set bolts[3] = AddLightningEx(LIGHTNING[i], false, minX, maxY, 0.0, maxX, maxY, 0.0)
endmethodNow, I don't provide the patitionAlg() because it suffices to know how it works. For each Quad.partitionAlg(...), the algorithm creates anywhere from 2 - 5 Quads. Thus, at the end of each call for a given Quad, we can get up to 6 Quads total (since we count the Quad itself). Each Quad is a rect, and each Quad has an array of 4 JASS lightning.
So in the 'worst' case scenario, we will have six rects and 24 lightning effects for each Quad.
Given a level of iterations (when the algorithm will stop applying to the Quad array members), the maximum rects created is so:
iter = 0 : 6 rects
iter = 1: 6 + (6 x 6) rects
iter = 2: 6 + (6 x 6) + (6 x 6 x 6) rects
...
iter = n: 6 + (62) + (63) + ... + (6n + 1) rects
which we can condense into (changing our notation by +1 as well):
The amount of JASS lightning is simple 4 x the summation above, since every Quad gets 4 lightnings.
The algorithm breaks in this total. What happens is that it only partially cleans up the the last series of Quads, and then subsequent calls don't do anything
Sum(2) + Sum(3) + Sum(3) = 522 rects and 2088 lightning effects.
If I then do Sum(3), here is what happens.
Note that these are all called linearly temporally--I do Sum(2) then Sum(3) then Sum(3) then Sum(3). When I do a new call, I clean up the screen using the flush() method I showed.
Here is where it breaks. If you notice, this actually just the remnants of the above call. But it doesn't clean up all of the rects. And if I try any other calls, these lightning effects/rects are here for good. Everything stops at this point.
So from what I showed, at some point it fails to clean up all the rects/lightning effects. And then it won't make any more rects or lightning effects. This is peculiar to me. Any ideas what's going on?
Last edited:
