Re: Another minor thing

[Miles Bader]

Eli Zaretskii <address@hidden> writes:
> FWIW, I'm not sure this is worth fixing, amazingly enough.  There
> might be other circumstances where we will have such effects (meaning
> that the cursor isn't at the locus where characters are inserted).
> The bidi text editing is one such case.  (I can describe the details,
> if you are interested.)

The reason I thought about fixing this is because I thought about using
overlay strings as another implementation method for my holy-grail
project, messages appended to the minibuffer (the implementation I did
before works 99% of the time, but very occasionally exhibits unnatural
behavior, with code that modifies the minibuffer after displaying a
message [quail does this very occasionally]).

In particular, I thought maybe I could put a message on an overlay
string on an overlay at the end of the buffer, and have it appear as
being `after' the buffer (I wanted to do this instead of actually
inserting the text because some code actually modifies the minibuffer at
such a time).  

However because of the odd behavior of overlay strings, any such message
would force the cursor to be after it if the cursor was at the end of
the buffer, which is just too bogus for words...

> Fixing this might have an adverse side effect of breaking the rule
> that the cursor is displayed on the character *after* point.

I think `natural' insertion behavior is more important, at least from a
user's perspective.

> What I think _is_ important is that the user should be able to
> understand why does this situation happen.  I think your case
> satisfies this condition.

I'm not sure what you're saying here.  What does my case do?

-Miles