Re: address@hidden: Re: vertical-motion bug]

[Kim F. Storm]

Richard Stallman <address@hidden> writes:

> Would you please DTRT and ack?  I think this is a real bug,
> because the property values of the three chars are not eq,
> so they ought to be handled one by one.

I have installed some changes to fix this.


>
> From: Chong Yidong <address@hidden>
> Subject: Re: vertical-motion bug
> To: address@hidden
> Cc: address@hidden
> Date: Sun, 16 Jul 2006 08:45:06 -0400
>
> Richard Stallman <address@hidden> writes:
>
>>       M-<
>>       M-: (insert (propertize "a" 'display "a\nb\nc\n")) RET
>>       M-<
>>       C-n (or M-: (vertical-motion 1))
>>
>>     Result: point moves down three lines.  The expected behavior, based on
>>     the `vertical-motion' docstring, is to move just one line.
>>
>> There is no way to move down just one line, since no buffer
>> position corresponds to that screen position.  Emacs can either
>> move three lines or not move.  So I think its actual behavior
>> is the best possible thing it could do.
>
> Sorry, I gave a bad test case.  Try this:
>
>   M-:  (let ((pos (point-min)))
>          (dotimes (i 10) (insert "a"))
>          (while (< pos (point-max))
>            (put-text-property pos (1+ pos) 'display (propertize "a\n"))
>            (setq pos (1+ pos))))
>
> This puts a display property on each of the characters in the buffer,
> so there is a valid buffer position.  C-n and C-p skip past all of
> them.
> ----------
>

-- 
Kim F. Storm <address@hidden> http://www.cua.dk