bug#36779: 25.1; mouse click not recognized for frames with large left position

[Eijiro Sumii]

On Tue, Aug 13, 2019 at 11:27 PM Eli Zaretskii <eliz@gnu.org> wrote:
> From: Eijiro Sumii <sumii@ecei.tohoku.ac.jp>
...
> > * X runtime problems
> >
> > ** General X problems
> >
> > *** Coordinates of mouse clicks not recognized correctly on multiple 
> > monitors
> >
> > This problem happens on a proprietary X server ( http://www.astec-x.com/ )
> > when the number of monitors is changed after the server has started.
> > It is a limitation of the server according to the manufacturer, but
> > known to exhibit only in combination with Emacs (not other X clients
> > or GNOME applications) for some unknown reason.  See
> > https://debbugs.gnu.org/cgi/bugreport.cgi?bug=36779 for details.
> > ----------
>
> Thanks.  Some comments:
>
>   . We generally avoid naming proprietary products, and instead say
>     something like "some proprietary X servers".

I gave the (link to) details of the X server (while avoiding to write
the product name, though the address includes it, which in fact may be
useful when searching the file for the former as long as the case is
ignored) since I was asked to do so.  The description "some
proprietary X servers" is not factual since (at least) I am not aware
of any other X server that exhibits the problem.  I had considered
"some proprietary X server" of course but that does not seem to give
useful information (if we keep the description vague, what is the
purpose of the file?).

>   . The part that this problem is only known to happen with Emacs
>     doesn't seem important (this is, after all, a file that deals
>     solely with Emacs-related problems), and I'd omit it.

That part is important since, while the manufacturer says the problem
is due to a limitation of the X server, it seems to happen only to
Emacs, which greatly confused me and is crucial information for
possible users in the same situation.

>   . I'd prefer not to send the reader to read the bug discussion, and
>     instead would tell here the important parts of what was said there
>     that clarify the issue (if there are such details there).

I linked to this discussion since the "details" are too complex to be
summarized in "a few lines" (as I was instructed).  Personally, I wish
if other entries in the file also included links to more detailed
discussions!

>   . We usually try to describe some workarounds -- do they exist in
>     this case?

The workaround is to avoid changing the monitor configuration after
the X server has started (or, equivalently, restart the X server every
time after such a change), which I believe is already clear from my
description.

Cheers,

        Eijiro