Messages

Somebody should start a service to allow people to remote desktop into a sandboxed Mac, (with predefined development environments for different packages).
You could then compile your app, sign it with your certificate, and download/save the results.
Rent a Mac by the hour.
Only how would you get it onto your iPhone for testing!?  Darn, so close, ha.
(Would it be possible using 'Ad Hoc distribution'?)

It may be feasible if you run a computer cafe, and set up a Mac to allow development, you would get the locals at least.
Talk to the owner, he/she may see the potential for a new client type.

Or, if you can't buy a Mac, convince a friend to buy one. 

It would be very handy to automatically use iXors3D when compiling to the iPhone, and Xors3D when compiling for the PC (only two platforms are supported).
Then you could develop and test on your PC, and when you build for the iPhone, no changes would be needed and it would pick up the proper library.
But, I only looked into the Xors3D library briefly, and I'm not sure how compatible the Types or Functions are between each version.

iXors3D should link without modification, actually there's no need to 'fool' the glbasic compiler to copy it onto the Mac, ala current Win32. It should be even simpler to install and use.

Good work Truci, still very interested in your results here! 

According to:
http://iphonedevelopment.blogspot.com/2010/04/ipad-max-touches.html

they say the iPhone has 5 maximum touches, and the iPad has 11 maximum touches.

(11 touches eh? um . . . where does Steve Jobs buy his gloves?)

Ok, here it is converted to a Generic 'LookAt' function:

Code (glbasic) Select Expand

TYPE TVector
x
y
z
ENDTYPE

// matrix[]    - matrix array to hold the results, for use later at rendering
// v_look_from - Point to look from
// v_look_to   - Point to look at
FUNCTION Vector_LookAt: matrix[], v_look_from AS TVector, v_look_to AS TVector
LOCAL v_at AS TVector
LOCAL v_up AS TVector
LOCAL v_xaxis AS TVector

v_at = Vector_Subtract(v_look_to, v_look_from)
Vector_Normalize(v_at)

v_up = Vector_New(0.0, 1.0, 0.0)
v_xaxis = Vector_CrossProduct(v_at, v_up)
Vector_Normalize(v_xaxis)
v_up = Vector_CrossProduct(v_xaxis, v_at)
Vector_Normalize(v_up)

DIM matrix[16]
matrix[ 0] = v_xaxis.x
matrix[ 1] = v_xaxis.y
matrix[ 2] = v_xaxis.z
matrix[ 3] = 0

matrix[ 4] = v_up.x
matrix[ 5] = v_up.y
matrix[ 6] = v_up.z
matrix[ 7] = 0

matrix[ 8] = v_at.x
matrix[ 9] = v_at.y
matrix[10] = v_at.z
matrix[11] = 0

matrix[12] = v_look_from.x
matrix[13] = v_look_from.y
matrix[14] = v_look_from.z
matrix[15] = 1
ENDFUNCTION

FUNCTION Vector_New AS TVector: x, y, z
LOCAL v AS TVector
v.x = x
v.y = y
v.z = z
RETURN v
ENDFUNCTION

// Subtract v2 from v1
FUNCTION Vector_Subtract AS TVector: v1 AS TVector, v2 AS TVector
   LOCAL v AS TVector
   v.x = v1.x - v2.x
   v.y = v1.y - v2.y
   v.z = v1.z - v2.z
   RETURN v
ENDFUNCTION

// Normalize a vector so it's length = 1
FUNCTION Vector_Normalize: v AS TVector
   LOCAL m
   m = SQR(v.x*v.x + v.y*v.y + v.z*v.z)

   IF m <= kTollerance THEN m = 1
   v.x = v.x / m
   v.y = v.y / m
   v.z = v.z / m
   IF ABS(v.x) < kTollerance THEN v.x = 0
   IF ABS(v.y) < kTollerance THEN v.y = 0
   IF ABS(v.z) < kTollerance THEN v.z = 0
ENDFUNCTION

// Takes v1 AND v2 AND returns the cross product v1 X v2.
// The cross product is a vector perpendicular TO both v1 AND v2
// This is the normal of 2 vectors
FUNCTION Vector_CrossProduct AS TVector: v1 AS TVector, v2 AS TVector
   LOCAL rv AS TVector
   rv.x = (v1.y * v2.z) - (v1.z * v2.y)
   rv.y = (v1.z * v2.x) - (v1.x * v2.z)
   rv.z = (v1.x * v2.y) - (v1.y * v2.x)
   RETURN rv
ENDFUNCTION


To use it:

Code (glbasic) Select Expand


LOCAL mat[]
LOCAL v_from AS TVector
LOCAL v_to AS TVector

v_from = Vector_New(1.0, 1.0, 1.0)
v_to = Vector_New(20.0, 5.0, 3.0)
Vector_LookAt(mat[], v_from, v_to)

X_MOVEMENT 0, 0, 0
X_ROTATION 0, 1, 0, 0
X_ROTATION 0, 0, 1, 0
X_ROTATION 0, 0, 0, 1
// Is there a shorter way to do the above? (I find if I don't, the current matrix affects it)

X_MULTMATRIX mat[]
X_DRAWOBJ  object_id, 0


I'm not sure why, but if I change the Cull Mode back to '0' (both sides) it works.

So I'm guessing that when the matrix is multiplied onto the stack, either:
- my normals get altered (but then why would I see them with Cull Mode '0'?)
- my normals get flipped inside out (but I would think you would see nothing, and not black)

I'm not sure of the performance hit setting Cull Mode to '0' (both sides), but it works, so I'm happy.

Oh, and the Entity Type and functions my code refers to is my own Entity system (very basic), and not the main one from the forums.
It should be very simple to convert the 'Entity_LookAt()' function to work for any vectors/points.

Thanks for all the help.

Getting close.  I found a good explanation of how to do this online at:
http://www.opengl.org/resources/faq/technical/lookat.cpp

I created a 'Look At' function, that takes the id of the Entity to rotate, and a vector (v) as a point to look at.

Code (glbasic) Select Expand

FUNCTION Entity_LookAt: id%, v AS TVector
LOCAL ex%
LOCAL v_at AS TVector
LOCAL v_up AS TVector
LOCAL v_xaxis AS TVector

ex = Entity_Find(id)
IF ex <= 0 THEN RETURN

v_up = Vector_New(0.0, 1.0, 0.0)

v_at = Vector_Subtract(v, _Entity[ex].position)
Vector_Normalize(v_at)

v_xaxis = Vector_CrossProduct(v_at, v_up)
Vector_Normalize(v_xaxis)

v_up = Vector_CrossProduct(v_xaxis, v_at)
Vector_Normalize(v_up)

DIM _Entity[ex].matrix[16]
_Entity[ex].matrix[ 0] = v_xaxis.x
_Entity[ex].matrix[ 1] = v_xaxis.y
_Entity[ex].matrix[ 2] = v_xaxis.z
_Entity[ex].matrix[ 3] = 0

_Entity[ex].matrix[ 4] = v_up.x
_Entity[ex].matrix[ 5] = v_up.y
_Entity[ex].matrix[ 6] = v_up.z
_Entity[ex].matrix[ 7] = 0

_Entity[ex].matrix[ 8] = v_at.x
_Entity[ex].matrix[ 9] = v_at.y
_Entity[ex].matrix[10] = v_at.z
_Entity[ex].matrix[11] = 0

_Entity[ex].matrix[12] = _Entity[ex].position.x
_Entity[ex].matrix[13] = _Entity[ex].position.y
_Entity[ex].matrix[14] = _Entity[ex].position.z
_Entity[ex].matrix[15] = 1
ENDFUNCTION


// Subtract v2 from v1
FUNCTION Vector_Subtract AS TVector: v1 AS TVector, v2 AS TVector
   LOCAL rv AS TVector
   rv.x = v1.x - v2.x
   rv.y = v1.y - v2.y
   rv.z = v1.z - v2.z
   RETURN rv
ENDFUNCTION

// Normalize a vector so it's length = 1
FUNCTION Vector_Normalize: v AS TVector
LOCAL m
m = SQR(v.x*v.x + v.y*v.y + v.z*v.z)

IF m <= kTollerance THEN m = 1
v.x = v.x / m
v.y = v.y / m
v.z = v.z / m
IF ABS(v.x) < kTollerance THEN v.x = 0
IF ABS(v.y) < kTollerance THEN v.y = 0
IF ABS(v.z) < kTollerance THEN v.z = 0
ENDFUNCTION

// Takes v1 AND v2 AND returns the cross product v1 X v2.
// The cross product is a vector perpendicular TO both v1 AND v2
// This is the normal of 2 vectors
FUNCTION Vector_CrossProduct AS TVector: v1 AS TVector, v2 AS TVector
   LOCAL rv AS TVector
   rv.x = (v1.y * v2.z) - (v1.z * v2.y)
   rv.y = (v1.z * v2.x) - (v1.x * v2.z)
   rv.z = (v1.x * v2.y) - (v1.y * v2.x)
   RETURN rv
ENDFUNCTION


And I use it like:

Code (glbasic) Select Expand


X_MOVEMENT 0, 0, 0
X_ROTATION 0, 1, 0, 0
X_ROTATION 0, 0, 1, 0
X_ROTATION 0, 0, 0, 1
X_MULTMATRIX _Entity[ex].matrix[]
X_DRAWOBJ  _Entity[ex].id_object, 0


(Is there a better way to reset the matrix?  (Identity Matrix?))
Now, it rotates and faces properly towards a point.

BUT now it displays the player object as black (no colours) while its using this code, but fine otherwise.
Is it something to do with the normals?
I've tried using 'X_AUTONORMALS' just before the X_MOVEMENT command, but with no luck.
Any ideas?

Thanks for the ideas.

I've been playing with the X_MULTMATRIX a bit this weekend, but so far no luck.
I've never used this command before, and I'm not sure exactly how to use it, or where to put what values in the mat[] array.
I'm not sure if I have to push/pop the matrix, if so, when?
I've tried different values in different matrix array positions. 
Is the last set of 4 values supposed to be zero?  Or can I put the position in there and skip the X_MOVEMENT command?

What I see (depending of what 'tweaks' I try) is either:
- the player disappears while this code runs [When I put the position into the matrix maybe, or maybe that causes the next problem]
- resets to the [0,0,0] position
- stretches and flickers beyond recognition [most common]

Here's what I've tried (using some code from the forums):
(my code looks very different, just simplified for the forum)

Code (glbasic) Select Expand


CONSTANT kTollerance = 0.0001

// Subtract v2 from v1
FUNCTION Vector_Subtract AS TVector: v1 AS TVector, v2 AS TVector
   LOCAL rv AS TVector
   rv.x = v1.x - v2.x
   rv.y = v1.y - v2.y
   rv.z = v1.z - v2.z
   RETURN rv
ENDFUNCTION

// Normalize a vector so it's length = 1 and apply tolerance
FUNCTION Vector_Normalize: v AS TVector
LOCAL m
m = SQR(v.x*v.x + v.y*v.y + v.z*v.z)

IF m <= kTollerance THEN m = 1
v.x = v.x / m
v.y = v.y / m
v.z = v.z / m
IF ABS(v.x) < kTollerance THEN v.x = 0
IF ABS(v.y) < kTollerance THEN v.y = 0
IF ABS(v.z) < kTollerance THEN v.z = 0
ENDFUNCTION

// p_start - starting point vector
// p_end   - end point vector
FUNCTION Player_Draw:
  LOCAL rotation as TVector
  rotation = Vector_Subtract(p_end, p_start)
  Vector_Normalize(rotation)

  DIM mat[16]
  mat[ 0] = rotation.x
  mat[ 1] = 0
  mat[ 2] = 0
  mat[ 3] = 0

  mat[ 4] = 0
  mat[ 5] = rotation.y
  mat[ 6] = 0
  mat[ 7] = 0

  mat[ 8] = 0
  mat[ 9] = 0
  mat[10] = rotation.z
  mat[11] = 1

  mat[12] = 0
  mat[13] = 0
  mat[14] = 0
  mat[15] = 0

  X_MOVEMENT position.x, position.y, position.z
// X_PUSHMATRIX here?
  X_MULTMATRIX mat[]
  X_DRAWOBJ player_id, 0
END FUNCTION


Just a thought.
I had the rotation on the Y axis working before the X_MULTMATRIX code was added.
Can I just use the X_PUSHMATRIX after I rotate on the Y, then rotate on the X to look up or down (determined using the ATAN command again) to rotate relative to the player and not the world?

Thanks for all your help!

Would using the Entity System make rotating easier?
ie: is there a built in 'Look At' option or something?
Or can I change an object's X axis relative to that object, and not the world?
So therefore I would only need to set its Y and X (and not Z)?

Or, is there a fancy way to use 'X_ROTATION' to do this?
I don't understand the parameters, besides using it the standard way such as:
X_ROTATION 90, 0, 1, 0 (would rotate on object 90 degrees on its Y axis).

But I've seem people use numbers other than 1 for the axis, such as "0, -1, 0", or "3, 4, 5", what would that do?

If I tried:
X_ROTATION 45, 1, 0, 1 - would that just rotate an object 45 degrees on both X and Z each?  Like the following:

X_ROTATION 45, 1, 0, 0
X_ROTATION 45, 0, 0, 1

Or does it split the value between the axis, or something else?
(ie: creates a new axis to rotate around, somehow combing X and Z?)

I hope I don't have to get into matrix math!!

Thanks again.
- Doug

[v_rotate.y = ATAN(v.z, v.x)]

I'm very new to 3D programming, and therefore 3D math.

I'm creating a path system where you create a bunch of path points and have an object (model, camera) follow that path.
Its working fairly well and follows the path, interpolating the position between the current two points based on those points time duration.  (It's only a linear interpolating for now, no smoothing).

I'm trying to implement an 'Orient To Path' option that will rotate the object to face the direction of the current path.
(Similar to the camera's 'LookAt' parameters).
So the numbers I have are the path start point and end point, and they are stored in a 3D vector TYPE (x,y,z).
I can take the difference of the two (end point - start point) to end up with one vector (named 'v') starting at (0,0,0).

Now, I want to rotate my object to match that vector's angle/direction.
I have the 'Y' rotation working with this:
   v_rotate.y = ATAN(v.z, v.x)
But even that acts weird and I had to add 180 or something to get it working as expected.
I'm not even sure of the range of numbers ATAN returns?  (-180 to 180?)  So then just add 180?
Or I've seen this?:  v_rotate.y = MOD(ATAN(v.z, v.x), 360)

But, I can't even come close to getting the 'X' rotation working.  (And I figured out that the 'Z' rotation is needed too?)
I had thought that if I rotate on the 'Y' axis first so I'm heading in the desired direction, then I could just figure out my X angle (relative to the direction I'm facing, ie. how much to tilt up or down).  It would be the same formula somewhat.  But I think I've come to learn that the 'X' rotation is always 'World' aligned?  And I can't make it relative to the new 'Y' rotation?  I hope that makes sense.

I think I have to somehow figure out this 'Angle' for the 'X' rotation, and split that value between 'X' and 'Z' somehow?

I've tried this many times and have failed.  (I find so many 2D examples, but that third dimension is eluding me!)
Any ideas?

Thanks,
-Doug

Are you asking how to use this feature, or to verify if it's working?

To use it, copy the above code into your project, and try (I've never verified this):

Code (glbasic) Select Expand


IF iPhoneIsCracked() = TRUE
   DisplayNagScreen()
ELSE
   SaveHighScore()
ENDIF


To test it, that would be harder.  One method is to actually release your game, and wait for somebody to crack it.
Jailbreak your iPhone, and install the cracked copy, and run it.
Otherwise, you can assume its working if it returns FALSE.
Maybe you can edit the plist file to test? (I have no idea really, and have never even seen this file).

Anybody know how to force this check into thinking it's a cracked iPhone?

Try this:

Code (glbasic) Select Expand


?IFDEF IPHONE
IMPORT int iPhoneIsCracked()
?ELSE
FUNCTION iPhoneIsCracked:
   RETURN FALSE
ENDFUNCTION
?ENDIF


Be sure to test if it works properly. I check for the SignerIdentity key.

How check this feature?

Sorry for my dumb question... if are.... 

I'm are on 70% of development status of my iphone game project... and I are very interested on this question. I have 2 iphones... one jailbreaked and one clean... two avaliable for test my project...

TIA, Hark0

[enums site:www.glbasic.com]

I resorted to using CONSTANTS myself.
Not that big of a deal, since we don't have any form of 'intellisense' functionality, which would make enums way more convenient.
But enums can work like a type for functions, which would be handy.

In c++, do enums get converted to the constant value during pre-processing? 
Or are they left intact internally for function parameters?
I was just wondering how easy it would be to add them to GLBasic.  On the surface it seems simple, but I have no idea.

What DID happened to the searching?
I've resorted to using Google searches like:
enums site:www.glbasic.com
to restrict searches to this site, it works OK guess.

I notice when you fold code, it replaces all [CR][LF] pairs with [BS][FF]  (ascii 08 & 12).
Looks like a bug in the 'folding' / 'unfolding' code and is two characters off somewhere.
(does the line end with [CR][LF] when NOT folded?  may be saved in a different format somehow?)

Does the compile error happen if you have a comment at the end?  (ie. push the problem to the right more)
Is it the same error if you remove the % from rand?  (it may insert the [BS] character in another position then).

And why not use 'INC tot%, rand%'.  (not as pretty, I guess, but I assume it converts to 'tot += rand' in c++.)
(just nit-picking, ignore this!! ha)

I'm using GLBasic 7.341, demo.
I can recreate this using these 4 lines:

In main.gbas:

Code (glbasic) Select Expand

LOCAL speed = kFast

And in library.gbas:

Code (glbasic) Select Expand

CONSTANT kSlow = 1
CONSTANT kFast = 2
CONSTANT kSpeedDefault = kFast


This will trigger the compile error.
But if I simply change the main.gbas to:

Code (glbasic) Select Expand

LOCAL speed = kSlow
everything compiles fine, because I'm not referring to the value assigned to kSpeedDefault.
If all of the above code is in the same module, I don't get any compile errors.

The work-around in this case is just to never use other constants as a default value for constants, but hard code them, as in:

Code (glbasic) Select Expand

CONSTANT kSpeedDefault = 2

Thanks for your comments and suggestions.

I have a library that I call from my main project module.
In this library, I set up some global constants.
Then, I tried to make another constant for a default value that assigns the value of one of the above constants, like this: (In Tween.gbas)

Code (glbasic) Select Expand

CONSTANT kTweenTransition_Linear = 1
CONSTANT kTweenTransition_Spring = 2
CONSTANT kTweenTransition_Bounce = 3

CONSTANT kTweenTransition_Default = kTweenTransition_Spring


But, in my main module, when I refer to the constant that was assigned as default (kTweenTransition_Spring in this case), I get this C++ error:
error: `kTweenTransition_Spring' was not declared in this scope.

But when I assign a hard coded value like:

Code (glbasic) Select Expand

CONSTANT kTweenTransition_Default = 2
everything works fine.

It seems assigning a constant to a constant causes the original constant to be declared earlier in the compiling cycle?

[edit]
Or never gets created in the first place? 
Or because it would have been created as a standard float in the first pass?

1. When I <ALT>+<TAB> back into the GLBasic editor, it removes my editing cursor so I have to keep re-clicking the code.
    - This works fine initially, but I have noticed quite frequently that once this behavior starts, it persists.
    - I haven't noticed what brings it back to normal, I assume by exiting and restarting.

2. <CTRL>+<TAB>, like above, doesn't always work for switching documents, and nothing happens.
    - It usually works fine for a while when GLBasic is freshly started.

3. <CTRL>+<TAB> again.  Not so much a bug, but a request.
    - When holding <CTRL> and repeatedly pressing <TAB>, it cycles through all the windows fine, as expected.
    - But if I hit <CTRL>+<TAB> once, it switches to the next window.  If I repeat the key sequence, it SHOULD return to the previously edited window, but it instead displays the NEXT window in the queue, like if I use the key sequence above.
    - ie. <CTRL>+(<TAB>, <TAB>) should be different than <CTRL>+<TAB>, <CTRL>+<TAB>
    - This feature would be handy when going back and forth between two documents, now I have to cycle through all the others first before I return to the one I was just editing.  (ha, I *could* close all the other windows first, but it'd be a pain)
    - I hope I explained that clearly.

4. When changing the display colors in the Options menu, it doesn't always get reflected back in the editor.  Reload fixes it.

5. Request: Add more color options:  Comment background, Margin / Editor separator line,  Function Hilighter (uses 'Normal Text' now), how about the Function Hilighter width?

6. Request: 'Jumps' tab: Space/line/(or different color) below 'Show All', 'Sort: 1.2.3' to show Functions better.

This happens when I'm using the demo version (didn't purchase full version yet) in both Windows XP Pro 32bit, and Windows 7 64bit.

Thanks, and great product, amazing price.