Re: How to make word at/before point always be a candidate for quail input method?, Re: How to make word at/before point always be a candidate for quail input method?

Daan via Users list for the GNU Emacs text editor
<[email protected]> writes:

> For example if I have the following rules:
> - "aa" -> "é"
> - "éb" -> "è"
>
> After typing "aa", "é" is completed and I can't just add "b" to have
> "è"
> any more.
>
> Extending the rules to include "aab" -> "é" isn't really practical
> since
> there will too many rules to list.
>
> ---
> Dan

Eli Zaretskii <[email protected]> writes:

>> Date: Tue, 23 Sep 2025 03:47:35 +0000
>> From: Daan via Users list for the GNU Emacs text editor
>> <[email protected]>
>> 
>> For example if I have the following rules:
>> - "aa" -> "é"
>> - "éb" -> "è"
>> 
>> After typing "aa", "é" is completed and I can't just add "b" to have
>> "è"
>> any more.
>> 
>> Extending the rules to include "aab" -> "é" isn't really practical
>> since
>> there will too many rules to list.
>
> The way Emacs input methods are implemented, they AFAIK work by
> collecting the keys you type, and don't look at the contents of the
> buffer.
>
> But I'm far from being an expert of writing input methods, so maybe
> there's some trick to achieve what you want.  Basically, you need to
> force the input method to push back the character before point onto
> unread-command-events.  Of course, there's then the problem of where
> to stop: for example, what if the buffer has "aaa" and you type "b"?

You can do this with a custom 'update translation' function which, in
addition, does not 'end' the translation after 'aa' (if all this is
worth it, of course). See Japanese, Malayalam etc. for an example.

--