From:	Ludvig Larsson <ludvig AT club-internet DOT fr>
Newsgroups:	comp.os.msdos.djgpp
Subject:	The First __dpmi_int in prog doesn't work!
Date:	Wed, 17 Dec 1997 14:09:18 +0100
Organization:	Hey!!
Lines:	69
Message-ID:	<3497CEFE.1112@club-internet.fr>
Reply-To:	ludvig AT club-internet DOT fr
NNTP-Posting-Host:	ppp-135-38.romans.club-internet.fr
Mime-Version:	1.0
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Hi everybody!

Back after moving to a new town, with new questions!

I am building a graphic interface that I will use in my
game one day and I'v run into a strange problem.

First I sometimes got really strange problems with the
graphic initialisation(I'm using 640x480x65536colors),
sometimes it's a lot of graphics,trash, on top of the graphic
screen(like 80-90% of the screen is covered), I can write and read
on vram and everything works perfect exept that images and stuff
only can be seen in the holes between all this strange stuff
that is covering screen. No graphic operation is actually
interfered with, if I make a get_bitmap from a blank screen(blank
or blanc, all this wierd stuff is there anyway) i get a bitmap
that is actually blank.
So, I changed my code, if I push the Q-button, the graphic is
reinstalled and that works perfectly. I start tha game, only shit
on the screen, I push "q" and everything goes normal. *huh*.
Aha! I tried to initialize the graphic twice in the beginning of the
program, three four times, doesn't work. Have to push the "q"
button to get the graphics right.
But!
Then i changed place of the init_mouse(that was just after the
init_graph) and init_graph; No more problems!!
That was what I thought. Now, every time I start the game the first
time when I have switched the computer on, I got no mouse. I quit(ctrl
alt del) restart the program and everything works perfect.
It's like the computer don't like the first interruptcall.

What is wrong?

I'm using __dpmi_int with
__dpmi_regs reg;

like:
int setmode(short mode)
{
 int ret;
 __dpmi_regs reg;
 __dpmi_meminfo mi;
 int vesa_err;
  
 reg.h.ah=0x4f;
 reg.h.al=0x02;
 reg.h.bl=mode;
 if(mode>255)
  reg.h.bh=1;
 else
  reg.h.bl=0;
 __dpmi_int(0x10, &reg);
 ret=reg.h.ah;
 return(ret);
}
And it's actually working, if it isn't the first interrupt the program
launches.

I'm working under windows, actually, the graphic problems tended to
be worse when I had a 'short-cut' to the .exe.

Any guesses?

BTW when I started building this package I almost never got these wierd
screens, but I haven't changed the code to make the initialisations.

So, If anyone have a guess, please send me a mail!!

<ludvig AT club-internet DOT fr>

- Raw text -