Support For More Treo 650 Hard Buttons

[sharkus]

First off, this is an amazing app!

Would it be possible to support more hard buttons on the Treo 650?

Currently, only the d-pad, phone, calendar, mail, and off buttons can be assigned. Some requested options (in order of greatness):

- "home" and "menu pulldown" button support
- Side buttons (rocker/select)
- Full keyboard support

Also, is it possible to show a small overlay on the screen that shows the soft button locations? It seems to be hit or miss when trying to press these buttons (and a lot of the time the menu overlay appears, rather than the expected button press).

Thanks!
Sharkus

[sharkus]

Sorry, I just noticed this has been covered in other threads... Please feel free to flame at will...

[axiomjunglist]

it's cool...I'm waiting patiently myself

I'm sure it's on a to-do-list...just a matter of time

[sharkus]

I played around with the following application and my Treo 650:

www.geocities.com/nagamatu/KeyReplace/KeyEvent.prc

Some of the source from the app can be found here:

developers.palmone.com/pe/forums/displaypost.jsp?postID=10138135

Most of the keys behaved exactly the same with some differences in the lower 16 bits of "modifiers".

Does LJP use a different method for capturing key presses? Would it help if I posted results for the various keypresses as displayed by this application?

[Tinnus]

Could you please copy and paste the code here? It needs you to register and I'm not in the mood to go through a lenghtly process just to read some code.

[sharkus]

Following code borrowed from here:

developers.palmone.com/pe/forums/displaypost.jsp?postID=10138135:



Posted Apr 17, 2005 by Tatsuo Nagamatsu
This is a core part of KeyReplace application.


/******************************************************************************
*
* Copyright (c) 2004 Tatsuo Nagamatsu ([e-mail])
* All rights reserved.
*
* File: KeyReplace.c
*
* Release: Palm OS SDK 5.0 (111823)
*
*****************************************************************************/

static Boolean PrvSysHandleEvent(EventPtr eventP)
{
KeyReplaceGlobalsPtr gP;
UInt16 eType;
struct _KeyDownEventType keyDown;
Boolean handled = false;

eType = eventP->eType;
eType = SwapEndian16(eType);

switch (eType) {

case keyDownEvent:
keyDown = eventP->data.keyDown;
keyDown.chr = SwapEndian16(keyDown.chr);
keyDown.keyCode = SwapEndian16(keyDown.keyCode);
keyDown.modifiers = SwapEndian16(keyDown.modifiers);

if (MatchWithTableEntryBinarySearch(&keyDown) != -1) {
eventP->data.keyDown.chr = SwapEndian16(keyDown.chr);
eventP->data.keyDown.keyCode = SwapEndian16(keyDown.keyCode);
eventP->data.keyDown.modifiers = SwapEndian16(keyDown.modifiers);
}
break;

}

// Modify event type to nilEvent if handled is true
if (handled)
eventP->eType = SwapEndian16(nilEvent);

return (handled);
}


static void SetupNotification(Boolean bEnable)
{
UInt16 cardNo;
LocalID dbID;
static Err notifyCallback(SysNotifyParamType *notifyParamsP);
KeyReplaceGlobalsPtr gP = GetGlobalsPtr();

GetDatabaseInfo(&cardNo, &dbID);

if (bEnable) {
SysNotifyRegister(cardNo, dbID, sysNotifyEventDequeuedEvent, notifyCallback, sysNotifyKeyReplacePriority, gP);
SysNotifyRegister(cardNo, dbID, sysNotifyDeleteProtectedEvent, notifyCallback, sysNotifyNormalPriority, gP);
} else {
SysNotifyUnregister(cardNo, dbID, sysNotifyEventDequeuedEvent, sysNotifyKeyReplacePriority);
SysNotifyUnregister(cardNo, dbID, sysNotifyDeleteProtectedEvent, sysNotifyNormalPriority);
}
}

static Err notifyCallback(SysNotifyParamType *notifyParamsP)
{
KeyReplaceGlobalsPtr gP = (KeyReplaceGlobalsPtr)notifyParamsP->userDataP;

switch (notifyParamsP->notifyType) {
case sysNotifyEventDequeuedEvent:
PrvSysHandleEvent((EventPtr)notifyParamsP->notifyDetailsP);
break;
case sysNotifyDeleteProtectedEvent:
DoEnableDisable(false);
break;
}

return (errNone);
}



[Tinnus]

This app uses events to capture keypresses while LJP checks each key's status directly.

Checking directly is a lot faster because well.... it's direct

Events means you'll have to get through a lot of layers until you get the key press... a lot slower.

[sharkus]

Browsing the PalmOne SDK I came across the following API definition... It looks like this could be used to extend the KeyCurrentState() function call to support extra hardware buttons on the Treo (or other PalmOne devices).



UInt16 PmKeyKeysPressed   (   UInt16   refNum,
UInt16   count,
const UInt16   keyCodes[],
Boolean   pressed[]
)  
 

Get the current state of specified keys

Parameters:
    refNum: IN: library reference number (returned by SysLibFind() / SysLibLoad() )
    count: IN: number of elements in keyCodes[] and pressed[]
    keyCodes: IN: array of keyCodes to test
    pressed: OUT: Each element is set to true if the key specified by the corresponding element in keyCodes[] is pressed, and set to false if it is not pressed. It is acceptable to pass NULL if only the return value is of interest.

Return values:
    UInt16 count of keys that are pressed.



[metaview]

What I will add to MyUAE, when I get the time:


/*
if (g_treo600_sdk || g_treo650_sdk)
{
UInt16 keyCodes[50] = {keyBackspace, keyTab, keyReturn, keyEscape, keySpace, keySingleQuote,
keyComma, keyDash, keyPeriod, keySlash, keyZero, keyOne, keyTwo, keyThree,
keyFour, keyFive, keySix, keySeven, keyEight, keyNine, keySemiColon, keyLessThan,
keyEquals, keyBackslash, keyA, keyB, keyC, keyD, keyE, keyF, keyG, keyH, keyI,
keyJ, keyK, keyL, keyM, keyN, keyO, keyP, keyQ, keyR, keyS, keyT, keyU, keyV,
keyW, keyX, keyY, keyZ};
Boolean keyPressed[50];
stub_HsKeysPressed(50, keyCodes, keyPressed);
for (i=0; i<50; i++)
{
if (keyPressed)
{
// handle the keypress
break;
}
}
}
*/

Regards
Henk

[metaview]

sharkus said:

Browsing the PalmOne SDK I came across the following API definition... It looks like this could be used to extend the KeyCurrentState() function call to support extra hardware buttons on the Treo (or other PalmOne devices).



UInt16 PmKeyKeysPressed   (   UInt16   refNum,
UInt16   count,
const UInt16   keyCodes[],
Boolean   pressed[]
)  
 

Get the current state of specified keys

Parameters:
    refNum: IN: library reference number (returned by SysLibFind() / SysLibLoad() )
    count: IN: number of elements in keyCodes[] and pressed[]
    keyCodes: IN: array of keyCodes to test
    pressed: OUT: Each element is set to true if the key specified by the corresponding element in keyCodes[] is pressed, and set to false if it is not pressed. It is acceptable to pass NULL if only the return value is of interest.

Return values:
    UInt16 count of keys that are pressed.



Sounds interesting, but I guess for now, the Hs stuff is as good as the Pm stuff. The question is if the keyboard of the Tungsten C can be polled by thePm functions.

Regards
Henk