Re: [hatari-devel] Recent symbolic keyboard mapping change

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 13.10.2024 20.07, Miro Kropáček wrote:
> On Sun, 13 Oct 2024 at 19:04, Thorsten Otto <admin@xxxxxxxxxxx> wrote:
>
> > And how long do you want to wait for that text, that maybe never arrives?
> I'd assume that SDL does not play dice. :) Either it works (i.e. both
> events are sent) or not (i.e. there's a rule that empty text is never
> sent), to be tested.

Problem is that it depends on the user inputting the rest of the things 
before something may be generated as text input.

(In some cases input methods may e.g. compose whole words, or even 
paragraphs, before they forward the composed text to the application.)

For example, first user presses Shift down, then searches the key he 
wants to be shifted, until finally, potentially several seconds later, 
he presses that key, release it, and releases also Shift.

At which point you inject Shift to emulation?


There are two alternatives:

A) Either you can wait some time to see whether also key-to-be-shifted 
is pressed, before injecting either Shift, or whatever TOS keycode host 
Shift+key would map (which on TOS may be key that cannot be inserted 
with Shift pressed)

Or B), you always inject modifier key immediate, and extra key later on, 
if it arrives.


Case A), screws with timings for modifiers, and randomly breaks for 
search-and-peck typists.

Case B) break for everything when same keys are not shifted in TOS 
keymap, as on host keymap.


A) is just broken, but B) is not much of an improvement either, just a 
complication.


	- Eero