IXUS 100 IS (Europe) / PowerShot SD780 IS - Porting Thread

HarpoMa

@jestik - I will look and see.
@Microfunguy - Give me a day or so and I'll put out a better copy of the source. I had forgotten I was mid-multi partition setup. I was unable to get a SD card to make 2 partitions so I couldn't finish it.

Harpo

sivalley

This is a great model to work on since it is compact and has pretty good quality HD video recording. Enabling optical zoom during HD recording via CHDK would really make this a unique camera. Thanks for all the work on this.

Microfunguy

Well, I have been able to port this to SDM with only minor issues.
The code for briefly overriding DISPLAY button when used as the ALT-exit button does not seem to work.
The functions 'kbd_get_autoclicked_key()' and 'kbd_get_held_key()' are identical.

I tried to compare with a compile of the CHDK-version source-code but get many errors about using hardware FP or using FPA instructions.

whim

@HarpoMa

I build on a Vista box and if I recall Vista is not supported in the build tree

CHDK-Shell supports Vista and Windows 7 since version 2.22, released in August (current is version 2.47)
if you prefer a command line environment, i've also added the GCC 4.4.0 kit as a separate download
on my drop.io (= the same as currently included with full versions of CHDK-Shell)

http://drop.io/gcc_for_chdkshell/asset/gcc440env-forchdk-sfx-exe

hth,

wim

Jestik

2 sivalley

Hm..why you say "would"? it has this function with CHDK

voxelman

Jestik
2 sivalley

Hm..why you say "would"? it has this function with CHDK

I think it even has zoom during video mode without chdk. Maybe he means auto focus during video recording. But I think the lack of auto focus during video recording is a limitation in the hardware.

HarpoMa

@whim

When I started to setup my build environment Vista wasn't supported. I haven't bothered to update since then.

Harpo

HarpoMa

@Microfunguy

Can't comment on the FP / FPA thing don't know what it is.

As for the two functions being identical....hmmm seems a bit of a mystery to me but I'm sure it made sence at the time...

Harpo

Microfunguy

HarpoMa
I'm sure it made sence at the time...

I am saying that the CHDK binary works but my version using your published source-code does not :-

	if (kbd_get_held_key() == KEY_PRINT) isHeldDisp=1;
	
	if (isHeldDisp != 1)
	{

		if (kbd_process() == 0)
		{
			physw_status[0] = kbd_new_state[0];
			physw_status[1] = kbd_new_state[1];
			physw_status[2] = (physw_status[2]  & (~KEYS_MASK2)) |
			(kbd_new_state[2] & KEYS_MASK2);
			physw_status[2] = physw_status[2] & ~SD_READONLY_FLAG;
		}
		else
		{
		// override keys
			physw_status[0] = (kbd_new_state[0] & (~KEYS_MASK0)) |
					  (kbd_mod_state[0] & KEYS_MASK0);
			physw_status[1] = (kbd_new_state[1] & (~KEYS_MASK1)) |
					  (kbd_mod_state[1] & KEYS_MASK1);
			physw_status[2] = (physw_status[2]  & (~KEYS_MASK2)) |
						(kbd_mod_state[2] & KEYS_MASK2);  //fix needed as below
			physw_status[2] = physw_status[2] & ~SD_READONLY_FLAG;
		}

      }
	else
	{
		isHeldDisp = 0;
		*mmio2 = *mmio2 & ~(0x00000040);
	}

etc

long kbd_get_held_key() 
{
    static long last_kbd_time = 0, press_count = 0;
    register long key, t;

    key=kbd_get_clicked_key();
    if (key) 
    {
        last_kbd_key = key;
        press_count = 0;
        last_kbd_time = get_tick_count();
        return 0;
    } 
    else 
    {
      if (last_kbd_key && kbd_is_key_pressed(last_kbd_key)) 
       {
        t = get_tick_count();
        if (t-last_kbd_time>((press_count)?175:500)) 
         {
          ++press_count;
          last_kbd_time = t;
          return last_kbd_key;
         } 
         else 
         {
          return 0;
         }
       } 
       else 
       {
        last_kbd_key = 0;
        return 0;
       }
    }
}

Can you publish the code that you currently use ?

David

HarpoMa

Not certain what doesn't work for you.

You have a current dump of my entire build environment (as of my work prior to yesterday). You mentioned you had to "fix" stuff maybe it broke during that process.

As I have said before it would be unwise to use my code in it's current state. If you insist on adding it to SDM right now you are likely going to waste a lot of time. Certainly that is your choice but you are on your own for any issues that arn't also in my build.

Harpo

Microfunguy

I am simply asking you if the keyboard code to handle DISPLAY button is correct in the published source.

I have other things working, including edge-overlay and zebra.

You can see your code that I am using above

Is that what you use in the working build ?

HarpoMa

You have the code from my latest build (minus support for partitions). The keyboard handling code hasn't change for quite some time. Explain... how are you using the DISP key? Explicit please and I'll try to assist.

harpo

Microfunguy

We are only referring to the key code, it is a very localised issue.

The tester is using the DISP key to enter ALT mode (as it is an alias for the Print key).

When he tries to leave ALT mode by pressing the DISP key he has to be VERY fast or the normal Canon action will occur.

i.e. the screen will blank.

Your code attempts to avoid this problem.

The tester claims it does not work as expected.

Normally, I do not try to port until a version has been 'officially' released.

With the ixus100 I allowed myself to be persuaded to do it immediately.

It will not happen again.

HarpoMa

Ahh I understand.... Please send me the kbd.c you are using (in full pls)

The heldDISP is actually for multitasking the DISP key as ALT & DISP. ALT is a single normal press and DISP is a long held press of DISP. ( I think you know this ) The key

{ 2, KEY_PRINT , 0x00000040 }, //doesn't exist
{ 2, KEY_DISPLAY , 0x00000040 }, //swapped for print atm

mapping works fine for the single press as listed in this order.

So the heldDISP is used to emulate a DISP key by allowing a long hold of DISP to mean DISP. So I think the mapping above has been swapped around as a first guess. But send the kbd.c in full and i'll look. It is not actually to avoid any problem as you describe but rather to add a "new" key as the sd100 has no print key. I'm noting that new cameras are having lesss and less keys.

Harpo

Microfunguy

Thanks, I will do that now.

I want to get this camera off my back and spend December on other things !

David

Microfunguy

I have sent file to your email address.

David

HarpoMa

The code as sent to me is broke. This will not correctly handle the DISP long held correctly. The code that is commented out is exactly the code required to handle long DISP and to NOT process the short DISP as DISP. It would seem to me that what your tester is seeing is exactly what they should expect to see given this code snippet...

// if (isHeldDisp != 1)
// {

if (kbd_process() == 0)
{
physw_status[0] = kbd_new_state[0];
physw_status[1] = kbd_new_state[1];
physw_status[2] = (physw_status[2] & (~KEYS_MASK2)) |
(kbd_new_state[2] & KEYS_MASK2);
physw_status[2] = physw_status[2] & ~SD_READONLY_FLAG;
}
else
{
// override keys
physw_status[0] = (kbd_new_state[0] & (~KEYS_MASK0)) |
(kbd_mod_state[0] & KEYS_MASK0);
physw_status[1] = (kbd_new_state[1] & (~KEYS_MASK1)) |
(kbd_mod_state[1] & KEYS_MASK1);
physw_status[2] = (physw_status[2] & (~KEYS_MASK2)) |
(kbd_mod_state[2] & KEYS_MASK2); //fix needed as below
physw_status[2] = physw_status[2] & ~SD_READONLY_FLAG;
if (switch_override == 1)
{
physw_status[1] = (physw_status[1] & (~SWITCH_MASK1)) | (switch_mod_state[1] & SWITCH_MASK1); // set relevant bits to '0' then add new pattern
physw_status[2] = (physw_status[2] & (~SWITCH_MASK2)) | (switch_mod_state[2] & SWITCH_MASK2); // set relevant bits to '0' then add new pattern
}
}

/*
}
else
{
isHeldDisp = 0;
*mmio2 = *mmio2 & ~(0x00000040);
}
*/

Microfunguy

No, that was not commented-out when I sent it to the tester.

I only did that later.

Microfunguy

You seem to have made a copy of kbd_get_autoclicked_key with the intention of creating kbd_get_held_key.

HarpoMa

Well I'm gonna go with the classic.... "It works for me it must be something you have changed." I am still thinking that the tester or builder has commented out some of the important code. As for the function that is the same I think I just didn't need to change anything. BUT I will rebuild my code tonight and verify 100% that I have no DISP/ALT issue.

Harpo

« Previous Page Next Page »