Re: [hatari-devel] Suggestion for enhanced keymap table format

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 25.10.2021 13.31, Vincent Barrilliot wrote:
> CKBD definitely supports it under TOS, and EmuTOS supports it too 
> (search kb_altnum in bios\ikbd.c),

Ah, it works only when numbers are entered from
the numeric keypad (and in case of PC / Hatari,
only with numlock enabled).

I had been trying with the normal number keys.


> maybe TOS 4.04 supports it too as I 
> remember my sister using it to type Spanish characters whenwas studying 
> Spanish (but can't remember if CKBD was running then, I'll check that 
> later on my Falcon)
>
> As you say, there are probably limitations/edge cases with keys being up 
> or down, the same as there were in other situations with shift/alt,

EmuTOS seems to support all alt-xxx keys going
down before any of them are released.

But alt-xxx support is actually easier to
implement by injecting the press & release events
for each of the digits together.  Without
releasing key immediately after it's pressed,
alt sequences having same number multiple times
(e.g. '111') would not work.

This means that at worst 7 key events (Alt press
+ 3 press & release events for digits) could
be inserted to IKBD buffer before emulation
is continued, but that seems to work OK.

(At least in EmuTOS console.)


I'll do a bit more testing & cleanup for the patch
series before sending it on the list.


	- Eero

> but 
> in general I think there's more value having that in than out, to offer 
> complete support of the PC's keyboard. The edge cases  where people do 
> ALT-µ are pretty slim... And we can have a switch to just disable 
> support for it.
>
> What doesn't work in what I did ?
>
>
> Vincent
>
> Le 25/10/2021 à 01:01, Eero Tamminen a écrit :
>
> > Hi,
> >
> > On 14.10.2021 21.19, Eero Tamminen wrote:
> > > On 14.10.2021 9.44, Vincent Barrilliot wrote:
> > > > This applies to PressedModifiers only. STScanCodes still needs to be 
> > > > an array so we can specify ALT-xxx sequences.
> > >
> > > Ah, that's a very good point!
> > >
> > > (Somehow I had not understood that your
> > > "Max is alt-a-b-c" comment in code meant those.)
> >
> > Which Atari SW supports that?  Neither EmuTOS
> > nor real TOS seems to support it, so I could
> > not have tested it.
> >
> > Also, I assume that each of the XXX digit keys
> > must be raised before simulating pressing the next
> > digit key down, and there probably needs to be
> > some emulated time to have elapsed between those,
> > for them to register properly inside emulation.
> >
> > Implementing something like that (emulation
> > running between inserting of the ALT-xxx digits)
> > would be *way* too complex.
> >
> >
> >     - Eero
> >
> > PS. except for this last item, the new patches are
> > ready.  I have not tested them yet, except that
> > they do build.