bug#19930: 25.0.50; REGRESSION: accessing minibuffer from 2nd frame for same buffer scrolls window back to point of 1st frame

bug#19930: 25.0.50; REGRESSION: accessing minibuffer from 2nd frame for same buffer scrolls window back to point of 1st frame

[Drew Adams]

This regression was introduced in Emacs 24.3.  There is no such problem
in Emacs 20 through 24.2.

To reproduce:

1. runemacs.exe -Q --debug-init  -l "onetest.el" -f "1on1-emacs"
   File onetest.el is attached.

   This opens Emacs with a standalone minibuffer frame.

2. Visit a file with some text in it, say foo.el, in the original
   (non-minibuffer) frame displayed.

3. Put point on some function name, e.g. near bob.

3. C-x 5 2, so you now have two frames showing the same buffer.

4. In the frame just created, scroll down a ways (or otherwise move down
   - it is unimportant how you do this), and put point on some other
   symbol (name).

5. M-x grep RET

   As soon as you hit RET, the current window (in the second frame)
   scrolls back to the point position of the first frame.  And if you
   then use M-n to grab a name at point for `grep', it grabs the symbol
   name under the point for the other frame.

In GNU Emacs 25.0.50.1 (i686-pc-mingw32)
 of 2014-10-20 on LEG570
Bzr revision: 118168 rgm@gnu.org-20141020195941-icp42t8ttcnud09g
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --enable-checking=yes,glyphs CPPFLAGS=-DGLYPH_DEBUG=1'

bug#19930: 25.0.50; REGRESSION: accessing minibuffer from 2nd frame for same buffer scrolls window back to point of 1st frame

[Noam Postavsky]

Drew Adams <drew.adams@oracle.com> writes:

> This regression was introduced in Emacs 24.3.  There is no such problem
> in Emacs 20 through 24.2.
>
> To reproduce:
>
> 1. runemacs.exe -Q --debug-init  -l "onetest.el" -f "1on1-emacs"
>    File onetest.el is attached.

Attachment seems to be missing.  I guess this could be the same as #31695?

bug#19930: 25.0.50; REGRESSION: accessing minibuffer from 2nd frame for same buffer scrolls window back to point of 1st frame

[Eli Zaretskii]

> From: Noam Postavsky <npostavs@gmail.com>
> Date: Tue, 05 Jun 2018 21:02:53 -0400
> Cc: 19930@debbugs.gnu.org
> 
> Drew Adams <drew.adams@oracle.com> writes:
> 
> > This regression was introduced in Emacs 24.3.  There is no such problem
> > in Emacs 20 through 24.2.
> >
> > To reproduce:
> >
> > 1. runemacs.exe -Q --debug-init  -l "onetest.el" -f "1on1-emacs"
> >    File onetest.el is attached.
> 
> Attachment seems to be missing.  I guess this could be the same as #31695?

Yes, all the signs are consistent with that one.

bug#19930: 25.0.50; REGRESSION: accessing minibuffer from 2nd frame for same buffer scrolls window back to point of 1st frame

[Drew Adams]

> > > This regression was introduced in Emacs 24.3.  There is no such
> > > problem in Emacs 20 through 24.2.
> > >
> > > To reproduce:
> > >
> > > 1. runemacs.exe -Q --debug-init  -l "onetest.el" -f "1on1-emacs"
> > >    File onetest.el is attached.
> >
> > Attachment seems to be missing.  I guess this could be the same as
> > #31695?
> 
> Yes, all the signs are consistent with that one.

Yes, sorry about that.  I didn't realize that I forgot the
attachment.  I looked for the file, but I no longer have it.
But it does sound like exactly the same bug.

The only odd thing in comparing the two reports is that the
thread of #31695 says that it is NOT a problem in Emacs 23.4.1,
but bug #19930 reports that it is a problem that started with
Emacs 24.3.  But I guess I could have been mistaken about that,
and #19930 is reported from 25.0.50.1 of 2014-10-20.

I'm glad there is a second report of the same problem, and that
the cause was found.

I still see this, and it can be quite annoying.  I end up having
to delete a frame that I really want to keep open to a particular
part of the buffer, just to be able to do `grep' etc.  Or else I
need to work around it some other way.

bug#19930: 25.0.50; REGRESSION: accessing minibuffer from 2nd frame for same buffer scrolls window back to point of 1st frame

[Noam Postavsky]

merge 19930 31695
quit

Drew Adams <drew.adams@oracle.com> writes:

>> > > This regression was introduced in Emacs 24.3.  There is no such
>> > > problem in Emacs 20 through 24.2.
>> > >
>> > > To reproduce:
>> > >
>> > > 1. runemacs.exe -Q --debug-init  -l "onetest.el" -f "1on1-emacs"
>> > >    File onetest.el is attached.
>> >
>> > Attachment seems to be missing.  I guess this could be the same as
>> > #31695?
>> 
>> Yes, all the signs are consistent with that one.
>
> Yes, sorry about that.  I didn't realize that I forgot the
> attachment.  I looked for the file, but I no longer have it.
> But it does sound like exactly the same bug.
>
> The only odd thing in comparing the two reports is that the
> thread of #31695 says that it is NOT a problem in Emacs 23.4.1,
> but bug #19930 reports that it is a problem that started with
> Emacs 24.3.  But I guess I could have been mistaken about that,
> and #19930 is reported from 25.0.50.1 of 2014-10-20.

No contradiction there, the range was gradually narrowed down, so both
bugs now say the problem started with 24.3.  I'll go ahead and merge.

| good |  broken |          |
|------+---------+----------|
|   ?? | 27.0.50 |  31695#5 |
| 23.4 | 24.5.50 | 31695#10 |
| 24.2 | 24.3    | 31695#13 |