PRS-505 FSK Test-Environment for Autorun Applications under MS-Windows

[Mark Nord]

Originally Posted by m-land

1) Do I understand well that I can use Core calls as I do on "real" device? Any limits?

Yes you can.
Execpt of runScript, as I haven't ported "libfskLoad.so" to Windows-Dll and all "real-device-kbook" related stuff.
With release x.02 Core is loaded via JS new Function, therewith "sandboxed" and can be called from autorun-js direct.
With release x.03 Core is loaded via prsp.compile and executed in the scope of bytecode. So all nonscriptable fsk and fskin objects could be adressed, but Core-functions are invisible to "sandbox" JS like in autorun.js.
Here you have to used the wrapped functions _Core. or reference it via var Core = _Core. (See the dictionary code in the last sim package)

2) I need to have my "keyboard driver" loaded "on startup" (e.g. just after another core files. Is it enough to put my vk.js file to directory .../FSK Test\eBookSim\core? Or is there any other way?

This should do it.

3) What to do to use it on notebook (1024x768 screen res)?

You can try to modify the eBookSimMain800.xml to have it work with 768 lines, but you have to decide if you want to cut off top or bottom.

Any other ideas?

1. Eventualy you will place your keyboard in a dialog, which will automaticaly collect all keyboard actions.
2. You may have a look on defsked kbook.so for the PRS-600/650
   Firmware-packages are over at the-ebook.org
   [Edit/Add] Keyword is "KEYS" and the function which pop up additional characters to chose if a key is hold. E.g. hold "A" and get ÄÂÀÁÃÅ [/Edit]
3. I'am fantasizing of porting this keyboard (and maybe dictionary) to the 505/300 and have it controled by events send to it.

[Mark Nord]

Originally Posted by Mark Nord

For the latter, and assuming that my Calc will work on a 600, at least sudoku should be easily upgraded with a nummeric-keyboard, based on the doCommand (if I recall correctly) property of sprites.
Same applies to mahjong and should go for fiveballs, but for those two the gamecontrol is not so obvious.

My recollection was wrong. Sprites don't have a doCommand method, but baisc mouse-interaction is there.
Although I wasn't able to define new handler via <code>, like it is working for buttons, we can tweak the behavior!

The "nonTouch" menu for the 505 and 300 is moved to so named group and set to invisible, another group "Touch" with a standard fsk-menu is introduced.
(I havn't checked if kartu's PRS+ popup-menu is touch enabled and suitable for this.)

This menu seams to exist in the scope of DOCUMENT, even if definde in the group "Touch". Thus leading to a nasty behavior of the Sim. After exiting the app the Sim will lose focus. No more clicks are processed. But the app is not frozen, as you can close it via right click on the Windows-Programm-Bar.
[EDIT] Issue solved by calling menuBar.endLoop(ev) for both Sim and PRS-600[/EDIT]

And so there is a first preview of a click (maybe touch) enabled Sudoku.
Known issues:

• Select Game# not imlemented yet
• Zero-Sprite (for delete) has no glyph
• Sim lose focus, so exit via Systemmenu "X" or via "Menu" "left" "center"
  "center"
• Code is based on my x.03 version with "compiled/elevated" Core-functions and the wrapper to _Core for sandboxed calls. Havn't checked with the new 2.0.3 codes!!
• WinGame-Animation isn't stoped by fsk-menu-action
• Background for WinGame-Animation in wrong order in respect of sprites
• labels.png has no palette transparency set yet

[EDIT 21.12.2010]
Sudoku_menubar_endLoop_v3.zip attached, all Issues cleared, full functional in Sim and PRS-505
getSoValue is explicit declared, so there are no dependencies to PRS+, should even run as Autorun.App
Possible further enhancements:

• have 505/300 interact with the new fsk_menu
• try to create the sprites in code (cloneObject or xs.newInstanceOf) instead of defining them in XML

[/EDIT]
[ADD]
in target.init = function ()

if(_Core.system.getSoValue == undefined){

had to be replaced with

if(!_Core || !_Core.system || !_Core.system.getSoValue){

for correct operation outside the Sim
21.12.2010
To get Button-Input for models 505 and 300 each <group> needs at least on "focusable" <SPRITE> to process the event.
[/ADD]
[EDIT 14.02.2011]
..endloop_v4.zip
Thanks to hcripe's findings now usable at PRS-600
[/EDIT]
[ADD 21.2.2011]
improved code (menus are visible with the Win-Animation) can be found in post #165
previous versions are momentary kept for reference
[/ADD]

[kartu]

Mark, unfortunatelly Calc doesn't work, but I suspect I've got an outdated version. Could you please have a look?

http://code.google.com/p/prs-plus/so...pt/addons/Calc

PS
PRS+ popup menu is not touchscreen enabled yet.

[Mark Nord]

You are right the version in the Repository is the very basic one for the 505/300 even without topleft icon and without translation.

Lastest version (but I think w/o translation for the string "Calculator") is included in the x.03 download in this thread or in its own thread here

Please note that the button-actions are realized really quick 'n dirty, and should be redone if the concept proofed working.

There is also the problem of calling compiled PRS+Core function from within sandboxed JS-code. For the Calc I have explicit included the needed code for soGetValue.

Have you already decided how to deal with this - also in regards of the dictionary-code for the 300?

[kartu]

One of the easy ways to solve sandbox issue would be to bind handlers code from corresponding "starter" js files, as their code is not sandboxed. On the other hand, it will probably make it difficult to develop using your test environment, or will it not?

Just realized that Calc included into 2.0.3preview got there from 505/300's (I didn't plan to add it in the preview version for 600 before testing it myself, but it showed itself in menu customizer / key bindings and someone has used it ).

What makes me wonder is why doesn't it show any UI on 600s. Since it works on 300, XML file doesn't contain any errors, so I would expect it to show some UI.

[Mark Nord]

What makes me wonder is why doesn't it show any UI on 600s. Since it works on 300, XML file doesn't contain any errors, so I would expect it to show some UI.

By checking the code in resources.img I found this line in addons.js:

tmp=function(){var a={name:"Calculator",title:"Calculator",description:"scientific calculator",icon:"CROSSED_BOX",activate:function(){kbook.autoRunRoot.path=Core.config.addonsPath+"Calc/calculator.xml";kbook.autoRunRoot.enterIf(kbook.model);},

But as far as I can see the needed files in the addonsPath are missing. So the app couldn't show any UI.
For the content of recources.img see attachment - But I'm sure you know what you packed into

BTW I would prefere to use surquizu's Calc-Icon instead of "CROSSED_BOX"

One of the easy ways to solve sandbox issue would be to bind handlers code from corresponding "starter" js files, as their code is not sandboxed. On the other hand, it will probably make it difficult to develop using your test environment, or will it not?

As it is done by _PRSPlus.js in the Sim?
If so, there will be no problems with the Sim.

[kartu]

Originally Posted by Mark Nord

By checking the code in resources.img I found this line in addons.js:

tmp=function(){var a={name:"Calculator",title:"Calculator",description:"scientific calculator",icon:"CROSSED_BOX",activate:function(){kbook.autoRunRoot.path=Core.config.addonsPath+"Calc/calculator.xml";kbook.autoRunRoot.enterIf(kbook.model);},

But as far as I can see the needed files in the addonsPath are missing. So the app couldn't show any UI.
For the content of recources.img see attachment - But I'm sure you know what you packed into

There is actually a bug in the build script.

BTW I would prefere to use surquizu's Calc-Icon instead of "CROSSED_BOX"

Me too. Remind me where it was again please, and I'll create an "issue" on projects site, not to forget about it again.

As it is done by _PRSPlus.js in the Sim?
If so, there will be no problems with the Sim.

I meant even skipping the:

<function id="doHoldCenter">
this.doNextF();
</function>

in the XML, directly going:

var AUTORUN = kbook.model.container.sandbox.AUTORUN_GROUP.sandbo x.AUTORUN;
// not sure if it should be .doHoldCenter or .sandbox.doHoldCenter, most likely the latter
AUTORUN.doHoldCenter = function () {
...
}

but your way would certainly work.

[Mark Nord]

Originally Posted by kartu

Me too. Remind me where it was again please, and I'll create an "issue" on projects site, not to forget about it again.

The link is here

var AUTORUN = kbook.model.container.sandbox.AUTORUN_GROUP.sandbo x.AUTORUN;
// not sure if it should be .doHoldCenter or .sandbox.doHoldCenter, most likely the latter
AUTORUN.doHoldCenter = function () {
...
}

Thanks, will try this.
But in the Sim the app dosn't reside in kbook.model.container, so I will have to find the correct sandbox

[kartu]

Originally Posted by Mark Nord

The link is here

Thanks, that's Issue #27

But in the Sim the app dosn't reside in kbook.model.container, so I will have to find the correct sandbox

Fskin.model.container might work.

[Mark Nord]

Just want to announce that I uploaded a "final" version of the click-enabled Sudoku-code in posting #32

[kartu]

Mark, I'm debugging Calc on 600 now, cheching _Core get so value should look like this (or if _Core is not defined, call fails):

_Core || !_Core.system || !_Core.system.getSoValue

Conversion from coordinates to button index doesn't seem to work. Trying to figure out why. Most likely it's because buttons are shifted a bit.

Is there a reason not to use sender's id for operation lookup?

[Mark Nord]

Your are right, the check for _Core.system.getSoValue is fare to deep.

Originally Posted by kartu

Is there a reason not to use sender's id for operation lookup?

No, as I wrote, the code is quick 'n dirty
You will have to set gridCursor to invisible, too!!

Maybe we need a trunc or round

	var x = _Core.system.getSoValue(sender,"x");
	x = (x - 121 ) / 60 ;

here.

PS: _Core check in Sudoko is also wrong.

[kartu]

There seems to be 81/37 pixel difference between test environment and the reader.

What is the purpose of popen/pclose functions?

[Mark Nord]

Originally Posted by kartu

There seems to be 81/37 pixel difference between test environment and the reader.

OK, this is the offset of EINK (the output-panel) in Sim to the application window.
<panel id="EINK" skin="panel" left="81" top="37" width="600" height="800"> //E-Ink Display 600x800

What is the purpose of popen/pclose functions?

Parenthesisopen and Parenthesisclose, just inherit form http://www.motionnet.com/calculator/

[Edit]
Good eye! This should be openp() and closep()!

We could define
var simDX=0;
var simDY=0;
Append init() like this

	if(_Core == undefined){
		simDX = 81;
		SimDY = 37;

and modify

target.doButtonClick = function (sender) {
	this.showTime();
	var x = _Core.system.getSoValue(sender,"x");
	x = (x - 121 + simDX) / 60 ;
	var y = _Core.system.getSoValue(sender,"y");
	y = (y - 287 + simDY) / 50;

[/Edit]

[Mark Nord]

Parenthesisopen and Parenthesisclose, just inherit form http://www.motionnet.com/calculator/

Good eye! This should be openp() and closep()!

Without the correction (1+2)*3=7 !!