kbhit optimized for telestrat

[jedeoric]

As we discuss, here is kbhit optimized

[oliverschmidt]

Now that your original code is merged this pull request can't be applied as-is. Please update your repo and add a commit to this pull request.

[oliverschmidt]

Does A have a defined value at this point from BRK_TELEMON XRD0? If yes, it might be worth mentioning in a comment.

[jedeoric]

It uses carry to set return value, but i have to look to source code to know what A returns. I try to have a look tonight

[oliverschmidt]

Sorry, my comment above THAT useful. What I actually wanted to express and what I probably should have written is: "I consider your recent proposal broken. The only way it could potentially work is if BRK_TELEMON XRD0 would be well known to return a certain value in A that you only have to override if necessary. But I rather presume that this isn't the case." You see, I'm not actually interested in the fact what BRK_TELEMON XRD0 returns in A. There's no need that you try to find out. Rather just go ahead and write code here which works independently from BRK_TELEMON XRD0' setting of A.

[groessler]

Sorry, Oliver, I cannot follow you.

[groessler]

This needs to be "txa"

[jedeoric]

Yes good point. I will fix it tonight. Thanks for the point :)

[oliverschmidt]

This needs to be "txa"

I don't agree.

[groessler]

why not?

[oliverschmidt]

Just to make sure we're talking about the same code. This ias what you want to see? tax replaced with txa, right?

_kbhit:
        BRK_TELEMON XRD0
        ldx     #$00 
        bcs     @no_char_action
        lda     #$01
        rts
@no_char_action:
        lda     #$00
        txa        
        rts

My take is that the tax shouldn't be replaced with txa. Rather it should be removed altogether.

[groessler]

I was thinking about something like:

_kbhit:
        BRK_TELEMON XRD0
        ldx     #$00 
        bcs     @no_char_action
        lda     #$01  ; return one
        rts
@no_char_action:
        txa        ; return zero
        rts

[oliverschmidt]

But that's not what you wrote above - or at least what I thought you wrote ;-) Anyhow, I guess the OP now has enough input on how to get it right...

[oliverschmidt]

Trying to bring this back on track...

A simple, straight forward option to write this - as suggested in #1392 (comment) - is:

.import return0, return1

_kbhit:
        BRK_TELEMON XRD0
        bcs     @no_char_action	
        jmp     return1
@no_char_action:
        jmp     return0

A "trying to be cool" option to write this is:

_kbhit:
        BRK_TELEMON XRD0
        ldx     #$00
        txa
        asl
        eor     #$01
        rts

[groessler]

That's the best version IMO. Smaller and faster than the others. But, asl should be rol :-)

[oliverschmidt]

But, asl should be rol :-)

How embarrassing :-(

I just directly pushed the fixed version: 8551431

[jedeoric]

Thanks. For information, XRD0 returns in A ascii code of the key pressed

[greg-king5]

For information, XRD0 returns in A ascii code of the key pressed.

In that case, both kbhit() and cgetc() must be redesigned. They must work together.

kbhit():
It sees that a key was tapped because it got a character! kbhit() must not lose that character. It must save that character somewhere. If kbhit() is called again before cgetc() is called, then it must return "true" immediately because of that saved character -- it mustn't ask Telemon for a new character.

cgetc():
It must behave in a similar way. If a saved character exists, then the function must return that character -- it mustn't ask Telemon for a new character (and, it can skip the cursor code). The function must "remove" the old character, in some way, so that the two functions will call Telemon the next time that they are called.