Beginner to Amiga coding.

[TroyWilkins]

Originally Posted by Zetr0

How I would write this, as if I was explaining it to a student I would initially write is thus -

void main( void )
{
    int ppoints[ 9 ];
    
    ppoints[ 0 ] = 0;             // since these are assigns 
    ppoints[ 3 ] = 0;             // and do not change in the code
    ppoints[ 4 ] = 200;             // its worthless to repeat the assign
    ppoints[ 7 ] = 200;             // better to assign then first 
    ppoints[ 8 ] = 0;              // and take only those that change into the loop
                                          // = 100 less commands ;)

    for (int candy=0 ; candy < 201; candy +=10 )
    {
        ppoints[ 1 ] = candy;
        ppoints[ 2 ] = (200 - candy);
        ppoints[ 5 ] = (200 - candy);
        ppoints[ 6 ] = candy;
        ppoints[ 9 ] = candy;
        PolyDraw( &rastPort, 5, ppoints);
    }
}

Have a look at that, obviously there are better / quicker ways - but since this is your first foray in C/C++ we should keep it simple to begin =)

Ahh, yes, of course, thank you, good thinking.

Now, I'm almost at my wits end, I can't work out what I';m doing wrong with PolyDraw, I keep getting:

argument type incorrect; expecting "WORD *", found "int *"

I'm sure it's something silly, but I'm stuck as to what I'm doing wrong, or what to do about it.

[Zetr0]

*double post*

[Zetr0]

Originally Posted by TroyWilkins

Ahh, yes, of course, thank you, good thinking.

This is the point of C/C++ it gets you to have a process of thinking - modularization, implementation then continual refinement

Originally Posted by TroyWilkins

Now, I'm almost at my wits end, I can't work out what I';m doing wrong with PolyDraw, I keep getting:

argument type incorrect; expecting "WORD *", found "int *"

I'm sure it's something silly, but I'm stuck as to what I'm doing wrong, or what to do about it.

WORD is usually 16bits, an INT is usually 32bits on a 32bit system - this is where we start treading on compiler and machine specific differences but fear not its quite intuitive after a while of hair pulling -

try the following -

    WORD ppoints[ 9 ];
    
    ppoints[ 0 ] = 0;         // since these are assigns 
    ppoints[ 3 ] = 0;         // and do not change in the code
    ppoints[ 4 ] = 200;       // its worthless to repeat the assign
    ppoints[ 7 ] = 200;       // better to assign then first 
    ppoints[ 8 ] = 0;         // and take only those that change into the loop
                              // = 100 less commands ;)

    for (WORD candy=0 ; candy < 201; candy +=10 )
    {
        ppoints[ 1 ] = candy;
        ppoints[ 2 ] = (200 - candy);
        ppoints[ 5 ] = (200 - candy);
        ppoints[ 6 ] = candy;
        ppoints[ 9 ] = candy;
        PolyDraw( &rastPort, 5, ppoints);
    }
}

This is known as *Casting and it refers to the container type of your variables. Different containers have different methods of operations that can be done on them.

[TroyWilkins]

Ahh, thank you. Now it builds correctly, without errors, but hangs straight after opening the window. Hmm...

[Daedalus]

@The Goto argument

Look, this argument breaks out all over the internet (and in the pub too if your friends are so inclined) whenever Goto is mentioned, and rarely, if ever, reaches an agreement.

To be fair, I can't remember ever using it to break out of nested loops - I see how it would make the code simpler but I don't think I've needed to get out of such heavy nesting in such a critical hurry as to warrant it. I'm trying to remember the last time I actually used a Goto, but probably when trying to write super tight timing in an interrupt while developing firmware. ASM might've been a better choice there, but then you're back to using JMP, effectively what Goto compiles to.

Also, for me anyway, line numbers went out with the 8-bits. Using Goto with a label is of course totally different readability- and maintenance-wise than a line number, similar to constants versus magic numbers and not dissimilar to calling a function by name. Not sure where line numbers come into it at all.

At the end of the day, it's heavily discouraged but it works fine. It's up to the individual programmer whether they want to get into it or not. Use it provided you're happy with the potential consequences, or don't if you want to follow best practices.

[Zetr0]

Originally Posted by TroyWilkins

Ahh, thank you. Now it builds correctly, without errors, but hangs straight after opening the window. Hmm...

Okay thats interesting, lets try something static

void main( void )
{
      WORD point_array[] = {4, 4, 19, 4, 19,19, 4,19, 4, 4    };
      PolyDraw( &rastPort, 5, point_array);
}

See if that displays anything - if so it might be something I have not done correctly in the logic.

Originally Posted by Daedalus

@The Goto argument

Look, this argument breaks out all over the internet (and in the pub too if your friends are so inclined) whenever Goto is mentioned, and rarely, if ever, reaches an agreement.

I am sure wars have been fought for less =)

Originally Posted by Daedalus

To be fair, I can't remember ever using it to break out of nested loops - I see how it would make the code simpler but I don't think I've needed to get out of such heavy nesting in such a critical hurry as to warrant it. I'm trying to remember the last time I actually used a Goto, but probably when trying to write super tight timing in an interrupt while developing firmware. ASM might've been a better choice there, but then you're back to using JMP, effectively what Goto compiles to.

Indeed, if you are going to use "goto" you might as well use ASM or in-line ASM - which I have seen is about as readable as tea-leaves in a cyclone on even medium sized projects

Originally Posted by Daedalus

Also, for me anyway, line numbers went out with the 8-bits. Using Goto with a label is of course totally different readability- and maintenance-wise than a line number, similar to constants versus magic numbers and not dissimilar to calling a function by name. Not sure where line numbers come into it at all.

"Back in the Day" it was a line number based statement, once Borland Turbo C++ (v2) came out you could use essentially a jump into point - but as you mentioned there is no difference from this and a regular function call.

Originally Posted by Daedalus

At the end of the day, it's heavily discouraged but it works fine. It's up to the individual programmer whether they want to get into it or not. Use it provided you're happy with the potential consequences, or don't if you want to follow best practices.

In a nut shell, the use of "goto" is just poor methodology - there are better ways to do it.

[TroyWilkins]

I decided to do it with move and draw instead, and got it working.



I'll come back to PolyDraw and work out what is going on there another day.

[Zetr0]

@TroyWilkins

I am assuming you have set the raster port and pens?

[edit]
Yes you have - well done indeed!


here's some code that "should work" - its been a while since I have Amiga Coded

#define BLACK 1
int main ( void )
{
	WORD ppoints[ 9 ];
	struct BitMap bitMap = {0};
	struct RastPort rastPort = {0};
 
	/* Initialize the RastPort and link the BitMap to it. */
	InitRastPort(&rastPort);
	rastPort.BitMap = &bitMap;

	SetAPen(&rastPort, BLACK)
	
	ppoints[ 0 ] = 0;   // since these are assigns 
	ppoints[ 3 ] = 0;   // and do not change in the code
	ppoints[ 4 ] = 200; // its worthless to repeat the assign
	ppoints[ 7 ] = 200; // better to assign then first 
	ppoints[ 8 ] = 0;   // and take only those that change into the loop
			    // = 100 less commands ;)

	for (WORD candy=0 ; candy < 201; candy +=10 )
	{
		ppoints[ 1 ] = candy;
		ppoints[ 2 ] = (200 - candy);
		ppoints[ 5 ] = (200 - candy);
		ppoints[ 6 ] = candy;
		ppoints[ 9 ] = candy;
		PolyDraw( &rastPort, 5, ppoints);
	}

	// Dump data to CLI
	for( peter = 0; peter < 10; peter ++)
		print("point[%d] = %d\n", peter, ppoints[ peter ]);
	
	return 0;
}

[TroyWilkins]

This is what I've got:

#include <stdio.h>
#include <exec/types.h>
#include <graphics/gfx.h>
#include <intuition/intuition.h>
#include <dos/dos.h>

/* Prototypes for system functions. */
#include <proto/exec.h>
#include <proto/graphics.h>
#include <proto/intuition.h>
#include <proto/dos.h>

/* Prototypes for our functions. */
int program_init(void);
void program_loop(void);
void clean_up(void);

/* Global variables. */
struct GfxBase *GfxBase = NULL;
struct IntuitionBase *IntuitionBase = NULL;
struct Window *MyWindow = NULL;
int c;

int main(void)
   {
      int error;
      error = program_init(); /* Open needed libraries and the main window. */
      if ( !error )
      {
         program_loop(); /* If everything was initialized, execute the main program code. */
      }
      clean_up(); /* We must clean up before we exit. */
      return 0;
   }
int program_init(void)
   {
      int result = 0;
      /* First open the graphics.library. */
      GfxBase = (struct GfxBase *)OpenLibrary( "graphics.library", 0L );
      if ( GfxBase != NULL)
      {
         /* Second open the intuition.library. */
         IntuitionBase = (struct IntuitionBase *)OpenLibrary( "intuition.library", 0L );
         /* Was the library opened? */
         if ( IntuitionBase != NULL )
         {
            /* Since the library was opened, we can open a window. */
            MyWindow = (struct Window *)OpenWindowTags( NULL,
                            WA_Left, 20,
                            WA_Top, 20,
                            WA_Width, 210,
                            WA_Height, 214,
                            WA_Title, (ULONG)"Troys Draw Demo Win",
                            WA_DepthGadget, TRUE,
                            WA_CloseGadget, TRUE,
                            WA_SizeGadget, FALSE,
                            WA_DragBar, TRUE,
                            WA_GimmeZeroZero, TRUE,
                            WA_IDCMP, IDCMP_CLOSEWINDOW | IDCMP_MOUSEBUTTONS | IDCMP_MOUSEMOVE,
                            TAG_END );
            /* Was the window opened? */
            if ( MyWindow == NULL )
            {
               /* The window was not opened so display a message. */
               printf( "Unable to open the window!\n" );
               result = -1;
            }
         }
         else
         {
            /* The intuition.library was not opened so display a message. */
            printf( "Unable to open the intuition.library!\n" );
            result = -2;
         }
      }
      else
      {
         /* The graphics.library was not opened so display a message. */
         printf( "Unable to open the graphics.library!\n" );
         result = -3;
      }
      return result;
   }
void clean_up(void)
   {
      /* If the window is open, close it. */
      if ( MyWindow != NULL )
      { 
         CloseWindow( MyWindow );
      }
      /* If the intuition.library is open, close it. */
      if ( IntuitionBase != NULL )
      {
         CloseLibrary( (struct Library *)IntuitionBase );
      }
      /* If the graphics.library is open, close it. */
      if ( GfxBase != NULL )
      {
         CloseLibrary( (struct Library *)GfxBase );
      }
      return;
   }
void program_loop(void)
   {
      /* Main program loop. */
		for ( c = 0; c < 201; c +=5 ){
		Move(MyWindow->RPort, 0, c);
		Draw(MyWindow->RPort, 200-c,0);
		Draw(MyWindow->RPort, 200,200-c);
		Draw(MyWindow->RPort, c,200);
		Draw(MyWindow->RPort, 0,c);
		}
		Delay( 500L );
	}

Which gives this output (on a hires non-interlaced PAL screen):

[TroyWilkins]

I found this while trying to find examples of PolyDraw actually being used, which looks interesting:
http://www.memphisamigagroup.net/dis....c?noconvert=1

[Zetr0]

@TroyWilkins

I see you are coding under the power of darkness - this is how nature intended such dark arts sir

void program_loop( void )
{
      /* Main program loop. */

	WORD ppoints[ 9 ];
	
	ppoints[ 0 ] = 0;   // since these are assigns 
	ppoints[ 3 ] = 0;   // and do not change in the code
	ppoints[ 4 ] = 200; // its worthless to repeat the assign
	ppoints[ 7 ] = 200; // better to assign then first 
	ppoints[ 8 ] = 0;   // and take only those that change into the loop
			    // = 100 less commands ;)

	for (WORD candy=0 ; candy < 201; candy +=10 )
	{
		ppoints[ 1 ] = candy;
		ppoints[ 2 ] = (200 - candy);
		ppoints[ 5 ] = (200 - candy);
		ppoints[ 6 ] = candy;
		ppoints[ 9 ] = candy;
		// lets try this dereference -> Mmmm Pointers ;)
		PolyDraw(MyWindow->RPort, 5, ppoints);
	}

	/* Dump data to CLI
	for( fred = 0; fred < 10; fred ++)
		print("%d", ppoints[ fred ]);*/	
}

[TroyWilkins]

Originally Posted by Zetr0

@TroyWilkins

I see you are coding under the power of darkness - this is how nature intended such dark arts sir

Power of darkness? Sorry my friend, you've lost me there?!?

I tried the changes you suggested, and here's what I got:

[Zetr0]

@TroyWilkins

In Tasmania / Australia it must be about 10pm ?

Yes my mistake, I dereferenced the RastPort pointer incorrectly.

[TroyWilkins]

Originally Posted by Zetr0

@TroyWilkins

In Tasmania / Australia it must be about 10pm ?

Yes my mistake, I dereferenced the RastPort pointer incorrectly.

Ahh, yes, 10:15pm right now Hence why I'm about to call it quits for now.

I've been threatening to quit this since about 7pm, I think I should do so soon, hahaha

[Zetr0]

@TroyWilkins

Fear not I do most of my best work under the power of darkness

Lets try this -

void program_loop( void )
{
      /* Main program loop. */

	WORD ppoints[ 9 ];
	
	ppoints[ 0 ] = 0;   // since these are assigns 
	ppoints[ 3 ] = 0;   // and do not change in the code
	ppoints[ 4 ] = 200; // its worthless to repeat the assign
	ppoints[ 7 ] = 200; // better to assign then first 
	ppoints[ 8 ] = 0;   // and take only those that change into the loop
			    // = 100 less commands ;)

	Move (&RPort, 0, 0 ); // make sure we are in the right place

	for (WORD candy=0 ; candy < 201; candy +=10 )
	{
		ppoints[ 1 ] = candy;
		ppoints[ 2 ] = (200 - candy);
		ppoints[ 5 ] = (200 - candy);
		ppoints[ 6 ] = candy;
		ppoints[ 9 ] = candy;
		PolyDraw(&RPort, 5, ppoints);
	}
	
	// I think the data was initially outside either the X/Y boundary
	// of the Raster Port and or BitMap
	
}

been a while coding on the miggy - looks like I know what I am going to be doing this afternoon =D

[alkis]

I am pretty sure SAS C doesn't accept the C99 feature of declaring a 'for' index inside the 'for'.

Just keep your declarations on top of the function.

On your working example, bring the declaration of 'c' variable inside the void program_loop(void). It's good practise to keep it local if you can.

[Thorham]

Originally Posted by TroyWilkins

Oooookay, I think enough has been said about the merits (or lack thereof) of goto now...

Indeed. Sorry for bringing it up, I should've known better than to open that can of worms.

[Zetr0]

Originally Posted by alkis

I am pretty sure SAS C doesn't accept the C99 feature of declaring a 'for' index inside the 'for'.

Just keep your declarations on top of the function.

On your working example, bring the declaration of 'c' variable inside the void program_loop(void). It's good practise to keep it local if you can.

hmmm I will test this later as I might be wrong but I do remember writing out the following syntax quite often.

for( int index =0; blah ; blah )

and if memory serves I would write this loop out a lot, when chunking data from FAST to CHIP

for( register int index =0; blah ; blah )

I used to use Dice / Lattice C as well as SAS - hmmm

[davideo]

@Zetr0

I found some of your coding at one point. I believe it was on Aminet but can't seem to locate it now!!

[Samurai_Crow]

In C89 you have to start a new code block to declare a new local variable. Of course you can just use a set of curly braces to do that...