Re: [emacs-bidi] Input method with diacritics.

[Yair Friedman (Jerusalem)]

Kenichi Handa <address@hidden> writes:

> "Yair Friedman (Jerusalem)" <address@hidden> writes:
> > Kenichi Handa <address@hidden> writes:
> It's very difficult because there's no way to identify that
> '[' is a prefix key.   For instance, we have this kind of
> rule in "latin-1-postfix" input method.
>
>  ("A`" ?À)
>  ("A'" ?Á)
>  ("A^" ?Â)
>  ("A~" ?Ã)
>
> But, we don't want to treat `A' as a prefix key.
>

In a way it is, it's a prefix key with 6 keys on its map. ("'/^`~). In
the same way in latin-1-prefix ' is a prefix key with 12 keys on its map
(aeiouyAEIOUY).  The difference is in the semantics of the characters,
but as far as input methods go, it the same thing.

But both latin-1-prefix and latin1 postfix are methods with many prefix
keys each has a relatively small number of key on its map.  For these
kind of input methods the guidance parameter of quail-define-package is
sufficient. 

The hebrew "prefix" method is different in a way it has few prefix keys
each has many keys on its map. guidance is not too useful in these
cases.  cyrillic-yawerty is another input method like this.

> We may be able to add a new argument PREFIX-CHARS to
> quail-define-package, and modify quail-help to check it and
> generate separate keyboard tables.  It may require a gross
> change in quail-help.  Do you want to work on implementing
> it?

I will see, I'm not so sure my list is good enough for serious
programming.

> No.  My question is "do you want that pressing CAPS sends
> Emacs an independent event that can be treated as a prefix
> event as the same way as the current `['?".
>
> Or, do you want that pressing CAPS and some other key KEY at
> the same time sends Emacs an event and that event is handled
> by quail to insert the same character as what '[' followed
> by KEY inserts?

I don't really care as long a the map is there - there is no need to
signal the user when CAPS is pressed.

With this method I want for example:
a -> Shin
Shift-a -> A
CAPS on-a -> A
CAPS on-Shift-a -> Masora circle.

I need to tell emacs that the shifted characters when CAPS is pressed
should not produce the unshifted characters, but instead a different
set.

Do you think that this discussion should be held to the mule-list?
There is nothing bidi about it.