bug#42663: 28.0.50; rectangle-mark-mode does not behave as documented when transient-mark-mode is off

[Eli Zaretskii]

> From: Sean Whitton <spwhitton@spwhitton.name>
> Date: Sat, 01 Aug 2020 21:19:22 -0700
> 
> rectangle-mark-mode's docstring says
> 
>     Toggle the region as rectangular.  Activates the region if needed.
>     Only lasts until the region is deactivated.
> 
> However, when Transient Mark mode is off,
> 
> 1) the region does not get activated by C-x SPC;
> 
> 2) the mode lasts until the mark is subsequently activated and then
>    deactivated again; and
> 
> 3) rectangle-mark-mode works inconsistently until the mark is activated:
>    killing the region kills the rectangle, but C-x C-x does not rotate
>    through the four corners of the rectangle.
> 
> I think the simplest fix might be for rectangle-mark-mode to activate
> the mark even when Transient Mark mode is off.  I think I'll probably
> add that to my own configuration, where Transient Mark mode is off.

I'm not sure the current behavior you described is a bug.  You seem to
be so confident it is that you haven't even provide an explanation why
you think so, and a naïve reading of the doc string doesn't make that
obvious (e.g., "if needed" could cover item 1).  In general, when
transient-mark-mode is off, Emacs behaves differently than when it's
on, and that's by design.

So could you please provide a more detailed explanation why you think
these are bugs?

Thanks.