bug#35624: log-view-diff regression

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Date: Wed, 08 May 2019 22:52:41 +0300
> Cc: 35624@debbugs.gnu.org
> 
> 1. with the expanded last visible revision
> 
>   * h8..: 2019-05-08 Revision h8.
> b * g7..: 2019-05-07 Revision g7.
>   * f6..: 2019-05-06 Revision f6.
>   * e5..: 2019-05-05 Revision e5.
>   commit e5
>   Date:   2019-05-05
>   first line
>   second line
> e third line
> 
> should compare e5 and g7
> 
> 2. but when region's end (e) is after the last line
>    of the last expanded revision:
> 
>   * h8..: 2019-05-08 Revision h8.
> b * g7..: 2019-05-07 Revision g7.
>   * f6..: 2019-05-06 Revision f6.
>   * e5..: 2019-05-05 Revision e5.
>   commit e5
>   Date:   2019-05-05
>   first line
>   second line
>   third line
> e
> 
> should compare d4 (a previous revision of the last revision) and g7.

You seem to be assigning some meaning to the case where the region
includes only part of a revision's display.  Why is that meaning
useful?  From my POV, it just increases the probability of user errors
if they position the region end inaccurately.  Why not consider a
revision included if even some part of it is in the region, let alone
its header line?