another _Pointsize issue in OpenGl!

[TempodiBasic]

Hi guys and gals
I was playing with a code posted by Ashish about plasma and while I's modding the code I lost the output!
I have spent different time (hours) thinking and searching about a bug in my mode...
when I was to abandoning the modded code  I've tryed to change _glPointSize from default 1 to 2 and you see what I have got:
  [ You are not allowed to view this attachment ]  
What do you think of these 2 different output changing the Point Size from default 1 to  2.
Thanks to read

[TempodiBasic]

Here the code to try and check

Code: QB64: [Select]

1. 1 CONST R = 0, G = 1, B = 2 'CODED BY ASHISH  modded by TempodiBasic
2. SCREEN _NEWIMAGE(300, 300, 32): DIM init AS INTEGER, screen2(299, 299, 2) AS INTEGER: init = 0
3. 3
4. IF init <> -1 THEN
5.     FOR y = 0 TO 299
6.        4 FOR x = 0 TO 299
7.            5 PSET (x, y), _RGB32(screen2(x, y, R), screen2(x, y, G), screen2(x, y, B))
8.        6 NEXT x
9.    7 NEXT y
10.     AdjournMatrix
11. END IF
12. a$ = INKEY$
13. IF a$ = CHR$(32) THEN SOUND 41, 18: init = 999: PRINT " Standard mode": _DELAY 1: CLS ' standard graphic mode
14. IF a$ = CHR$(13) THEN SOUND 141, 18: init = 999: PRINT " Opengl mode": _DELAY 1: CLS: init = -1 ' Opengl graphic mode
15. 10 IF a$ = CHR$(27) THEN END ELSE GOTO 3
16.  
17. SUB AdjournMatrix
18.     SHARED screen2() AS INTEGER
19.     STATIC i, j
20.     i = i + .1
21.     j = j - .1
22.     FOR Y = 0 TO 299
23.         FOR x = 0 TO 299
24.             screen2(x, Y, R) = 230 * (COS(j)) * ABS(SIN(i + x / 20) - COS(j + Y / 20))
25.             screen2(x, Y, G) = 200 * ABS(SIN(i) + COS(j))
26.             screen2(x, Y, B) = 50 * ABS(SIN(i)) * ABS(COS((x + Y) / 20) - SIN((x - Y) / 20))
27.         NEXT x
28.     NEXT Y
29. END SUB
30.  
31. SUB _GL
32.     SHARED init AS INTEGER, screen2() AS INTEGER
33.  
34.     IF init = 999 THEN
35.         EXIT SUB
36.     END IF
37.  
38.     IF init = 0 THEN ' init = 0 actives Opengl inizialization
39.         _glViewport 0, 0, 299, 299
40.         _glClearColor 0, 0, 0, 1
41.         _glPointSize 1
42.         ga = 1
43.         init = 999
44.     END IF
45.  
46.     _glClear _GL_COLOR_BUFFER_BIT
47.     ' draw matrix
48.     FOR y = 0 TO 299
49.         FOR x = 0 TO 299
50.             gr = Proportion(0, 255, 0, 1, screen2(x, y, R))
51.             gb = Proportion(0, 255, 0, 1, screen2(x, y, B))
52.             gg = Proportion(0, 255, 0, 1, screen2(x, y, G))
53.             _glColor4f gr, gg, gb, ga
54.             _glBegin _GL_POINTS
55.             _glVertex2f Proportion(0, 299, -1, 1, x), Proportion(0, 299, 1, -1, y)
56.             _glEnd
57.         NEXT x
58.     NEXT y
59.     AdjournMatrix
60.     _glFlush
61. END SUB
62.  
63. FUNCTION Proportion (min1, max1, min2, max2, value)
64.     'delta1: value = Delta2: x
65.     Proportion = ((value * (max2 - min2)) / (max1 - min1)) + min2
66. END FUNCTION
67.  

As you can see the original code is the ShortestPlasmaGenerator from Ashish. I have worked on the output and the structure data.
Pressing Space Bar key you active the standard graphic mode
pressing Enter key you active the OpenGl mode
pressing ESCape key you quit the demonstration code

If you set _GlPointSize to 1 you'll get the issue, while if _GlPointSize is bigger it is ok
Now I'll test the code into Linux Ubuntu Os on the same machine in a VM.

[TempodiBasic]

Hi
here from into Lubuntu 14  in VM  running QB64 x32 version 1.3 I got the same results using
_GlPointSize 1 or 2
here screenshot
  [ You are not allowed to view this attachment ]  
It lasts to think if it is OS dependent or QB64 version dependent.
Only further tests on other machines can solve the issue.
Thanks to read

[TempodiBasic]

Hi Guys and Gals

Attention please!

here the result of running the above code into QB64 1.3 32bit in Windows 10.... it works so something has been broken from 1.3 to 1.4!
see screenshot
  [ You are not allowed to view this attachment ]  
but the tale has not ended for now.
If you take the build version 1.5  64bit in windows10 it works again!  Something broken has beed fixed!
see screenshot
  [ You are not allowed to view this attachment ]  

Good Job QB64 developers team !