Kindle VNC viewer (native app, GPLv2)

[twobob]

Originally Posted by PoP

You got it. My 5 way moves the character cursor too, as expected. I want to also move the mouse pointer.

this thread - obliquely - discusses the issue

http://osdir.com/ml/emulators.qemu/2.../msg00085.html

http://osdir.com/ml/emulators.qemu/2.../msg00093.html

etc..


other thoughts:

http://stackoverflow.com/questions/3...ordinates-in-x


get cursor position

#include <X11/Xlib.h>
#include <assert.h>
#include <unistd.h>
#include <stdio.h>
#include <malloc.h>

static int _XlibErrorHandler(Display *display, XErrorEvent *event) {
    fprintf(stderr, "An error occured detecting the mouse position\n");
    return True;
}

int main(void) {
    int number_of_screens;
    int i;
    Bool result;
    Window *root_windows;
    Window window_returned;
    int root_x, root_y;
    int win_x, win_y;
    unsigned int mask_return;

    Display *display = XOpenDisplay(NULL);
    assert(display);
    XSetErrorHandler(_XlibErrorHandler);
    number_of_screens = XScreenCount(display);
    fprintf(stderr, "There are %d screens available in this X session\n", number_of_screens);
    root_windows = malloc(sizeof(Window) * number_of_screens);
    for (i = 0; i < number_of_screens; i++) {
        root_windows[i] = XRootWindow(display, i);
    }
    for (i = 0; i < number_of_screens; i++) {
        result = XQueryPointer(display, root_windows[i], &window_returned,
                &window_returned, &root_x, &root_y, &win_x, &win_y,
                &mask_return);
        if (result == True) {
            break;
        }
    }
    if (result != True) {
        fprintf(stderr, "No mouse found.\n");
        return -1;
    }
    printf("Mouse is at (%d,%d)\n", root_x, root_y);

    free(root_windows);
    return 0;
}

So... XQueryPointer(display, root_windows[0], &window_returned,
&window_returned, &root_x, &root_y, &win_x, &win_y,
&mask_return);

perhaps that could be "massaged"

Setting x y

..., a flexible X-automation tool is xte, part of the xautomation package. Use xte -h to see a list of the X events it supports. This example goes to a point 320 pixels from the left edge of the screen and 50 pixels down and then fakes a click there with button 1:


xte 'mousemove 320 50' 'mousedown 1'
'mouseup 1'

The mousedown and mouseup events are separate, so you can drag as well as click.

also http://www.slack.com/sw/xwit-man.html

and http://tronche.com/gui/x/xlib/input/XWarpPointer.html

xdotool may also be useful (This I have)

HTH. no answers, but thoughts in a direction

[PoP]

Thanks for the research/references twobob! I will read with great attention. I also reached the conclusion that I would need Xorg extensions to LibVNC... and not discounting the speed problem to keep queries timely synchronized with the mouse position... I am a bit overwhelmed but not quitting yet

[twobob]

Originally Posted by PoP

Thanks for the research/references twobob! I will read with great attention. I also reached the conclusion that I would need Xorg extensions to LibVNC... and not discounting the speed problem to keep queries timely synchronized with the mouse position...

Think it is doable. Yes. my way would be some horrific hack or kludge as per usual but meh, if it gets the job done...

will have a play : ) see if I can get some of those tools for the 3

Originally Posted by PoP

I am a bit overwhelmed but not quitting yet

Never give up! teehehe

[hawhill]

PoP: don't get distracted with the X stuff - VNC can probably do what you want. I think we need to serve LibVNCClient's HandleCursorPosProc callback properly. I probably won't get to work on this for the next few days, so feel free to give it a try!

[twobob]

Originally Posted by hawhill

PoP: don't get distracted with the X stuff - VNC can probably do what you want. I think we need to serve LibVNCClient's HandleCursorPosProc callback properly. I probably won't get to work on this for the next few days, so feel free to give it a try!

: ) warm fuzzy feeling

for ref: http://libvncserver.sourceforge.net/...f5bfa7007dc0f6

00162 typedef rfbBool (*HandleCursorPosProc)(struct _rfbClient* client, int x, int y);

from http://libvncserver.sourceforge.net/...8h_source.html

[twobob]

Hi I was reading http://bisqwit.iki.fi/story/howto/dither/jy/ and then http://www.efg2.com/Lab/Library/Imag...sing/DHALF.TXT and

was wondering if this (almost complete) snippet would be of use to try a cheap 8bpp -dithered target?

1) This code (in the C programming language) dithers a 256-level
monochrome image onto a black-and-white display with the Bayer ordered
dither.

Spoiler:

PHP Code:

/* Bayer-method ordered dither.  The array line[] contains the intensity
** values for the line being processed.  As you can see, the ordered
** dither is much simpler than the error dispersion dither.  It is also
** many times faster, but it is not as accurate and produces cross-hatch
** patterns on the output.
*/

unsigned char line[WIDTH];

int pattern[8][8] = {
    { 0, 32,  8, 40,  2, 34, 10, 42},   /* 8x8 Bayer ordered dithering  */
    {48, 16, 56, 24, 50, 18, 58, 26},   /* pattern.  Each input pixel   */
    {12, 44,  4, 36, 14, 46,  6, 38},   /* is scaled to the 0..63 range */
    {60, 28, 52, 20, 62, 30, 54, 22},   /* before looking in this table */
    { 3, 35, 11, 43,  1, 33,  9, 41},   /* to determine the action.     */
    {51, 19, 59, 27, 49, 17, 57, 25},
    {15, 47,  7, 39, 13, 45,  5, 37},
    {63, 31, 55, 23, 61, 29, 53, 21} };

int getline();               /* Function to read line[] from image      */
                             /* file; must return EOF when done.        */

putdot(int x, int y);        /* Plot white dot at given x, y.           */

dither()
{
    int x, y;

    while (getline() != EOF) {
        for (x=0; x<WIDTH; ++x) {
            c = line[x] >> 2;           /* Scale value to 0..63 range   */

            if (c > pattern[x & 7][y & 7]) putdot(x, y);
        }
        ++y;
    }
} 


and

Spoiler:

void rfb16ToFramebuffer8(int x, int y, int w, int h) {
	int cx, cy;

	/* we read single pixels */
	uint16_t *src = (uint16_t*)(client->frameBuffer) + client->width*y + x;
	/* and we write single pixels */
	uint8_t *dest = (uint8_t*)fbdata + finfo.line_length*y + x;

	for(cy = 0; cy < h; cy++) {
		for(cx = 0; cx < w; cx++) {
			uint32_t c;
			uint16_t v;
			uint8_t dval;

			v = *(src + cx);
#ifdef PERFECT_COLOR_CONVERSION
			c = ((v & 0x001F) * 77 // red
				+ ((v & 0x03E0) >> 5) * 151 // green
				+ ((v & 0x7C00) >> 10) * 28 // blue
			    ) >> (8 /* from multipl. above */ + 1 /* 5 -> 4 */ );
#else
			c = ((v & 0x001F) // red
				+ (((v & 0x03E0) >> 5) << 1) // green counts 2x
				+ ((v & 0x7C00) >> 10) // blue
			    ) >> (2 /* from shifts above */ + 1 /* 5 -> 4 */ );
#endif
			*(dest+cx) = (((uint8_t)c << 4) + ((uint8_t)c & 0xF)) ^ 255; /* repeat value in lower nibble */
		}
		dest += finfo.line_length;
		src += client->width;
	}
}

I guess?

Would dearly love to see a (cheap) dithered target for the touch as the one on the 3 does such a great job. Did think about getting it done myself but once again the maths has stumbled me.

Anyways, I know you are busy but I documented the thought.

Also:


Is the GIT the latest code?

I don't understand how I can get an error message that I cant find in the code???
The one about "4bpp only being supported for now sorry" Am I being stupid?


also this was worth a read http://www.virtualdub.org/blog/pivot/entry.php?id=151

[geekmaster]

Originally Posted by twobob

Hi I was reading http://bisqwit.iki.fi/story/howto/dither/jy/ and then http://www.efg2.com/Lab/Library/Imag...sing/DHALF.TXT and

was wondering if this (almost complete) snippet would be of use to try a cheap 8bpp -dithered target? ...

They should work, but they would be horribly slow compared to my code. Good luck trying to do smooth video or animation with that, while also trying to do audio...

[twobob]

Originally Posted by geekmaster

They should work, but they would be horribly slow compared to my code. Good luck trying to do smooth video or animation with that, while also trying to do audio...

I did have another go at getting my head around welding your code into either a) Hawhills kindlevncviewer code - This would be the simplest in terms of code overhead. The code is very clean.

or b) into Xephyr or SDL.

will revist a) again then with your code in mind.

[twobob]

Hmm. Thought I would share how I use my viewer.

[root@kindle root]# piano 2> /dev/null

very helpful thanks

[root@kindle root]# nano /mnt/us/usr/bin/piano
(pictured simple script)

x11vnc runs on my host.

[knc1]

Ah, Norma Jean
I read your PM before I saw the post above, but now I am up to speed.

[PoP]

Originally Posted by hawhill

PoP: ... I think we need to serve LibVNCClient's HandleCursorPosProc callback properly. I probably won't get to work on this for the next few days, so feel free to give it a try!

Ahh, good hawhill. Sorry, I would try but I am not experienced enough with the Lua and C integration (let alone with LibVNC).

I would start by adding a new function to config.lua

  GetPointerPos(x, y)
      get mouse coordinates from client

Which I imagine would automagically be "linked" to a new GetPointerPos function declaration in your kindlevncviewer.c . This function would return x and y mouse position after referring?/calling? typedef rfbBool (*HandleCursorPosProc)(struct _rfbClient* client, int x, int y); per rfbclient.h.

But I just don't know how to code these pieces together.

Would be fun to have but not a priority (don't sweat it). Thanks.

[Rockets00]

Video tutorial for Windows 7 is ready:
http://youtu.be/7ecuTxcHwgE

[twobob]

Helpful. I didn't catch 100% of everything. And I now rarely use windows, but I suspect I will probably use this guide with the DX at some point. so thanks for that.

Appreciate the discussion link and will even hazard a look in your bundle. appreciated.

[twobob]

From the docs: https://github.com/hwhw/kindlevncviewer

Building:
---------

It is suggested to compile LibVNCclient in a very minimal version. Personally, I use the following call to configure:

PKG_CONFIG_PATH=/home/hw/x-tools/arm-unknown-linux-gnueabi/arm-unknown-linux-gnueabi/sys-root/usr/lib/pkgconfig/ \
./configure --without-gcrypt --without-client-gcrypt \
--without-gnutls --without-client-tls \
--without-ipv6 --host=arm-unknown-linux-gnueabi \
--host=arm-unknown-linux-gnueabi \
--prefix=/home/hw/x-tools/arm-unknown-linux-gnueabi/arm-unknown-linux-gnueabi/sys-root/usr/

I think perhaps one of those --host calls should be something else. --build=i686-pc-linux-gnu maybe or something

Anyways. Thought I would share.

[hawhill]

Oh, right! Got to remember to check this, I won't probably have the time for the next 3 days. Thank you for the notice!