M-. on a symbol doesn't always move the cursor to the right position

[lokedhs]

If you press M-. on a symbol that is defined in a file that is already opened, the cursor doesn't move to the correct location. The problem only happens on recent versions of Emacs.

The cause of the problem is a change in behaviour of goto-char. It no longer can be used to move the cursor in an inactive buffer. The old behaviour can be restored by setting switch-to-buffer-preserve-window-point to nil.

The root cause is is in SLIME the cursor is set using goto-char, and later the buffer is displayed using pop-to-buffer. This no longer works, and according to a discussion on emacs-devel, the function set-window-point should be used instead.

Here is the discussion on emacs-devel: https://lists.gnu.org/archive/html/emacs-devel/2017-06/msg00227.html

[luismbo]

Thanks for looking into this. Do you know whether the new recommendation is backwards compatible with older Emacs versions?

…

On Thu, Jun 15, 2017, 11:08 Elias Mårtenson ***@***.***> wrote: If you press M-. on a symbol that is defined in a file that is already opened, the cursor doesn't move to the correct location. The problem only happens on recent versions of Emacs. The cause of the problem is a change in behaviour of goto-char. It no longer can be used to move the cursor in an inactive buffer. The old behaviour can be restored by setting switch-to-buffer-preserve-window-point to nil. The root cause is is in SLIME the cursor is set using goto-char, and later the buffer is displayed using pop-to-buffer. This no longer works, and according to a discussion on emacs-devel, the function set-window-point should be used instead. Here is the discussion on emacs-devel: https://lists.gnu.org/archive/html/emacs-devel/2017-06/msg00227.html — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#391>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACyv_gNzP3XchItvFqKcASy9XUHrRAZks5sEQKcgaJpZM4N6_9R> .

[lokedhs]

Based on the discussion on the mailing list, I think the only reliable solution to this is to change the behaviour of the code so that it switches to the buffer before moving the cursor. I think the propoer solution is to use pop-to-buffer first, and then place the cursor at the correct location using goto-char.

[luismbo]

Sounds like a plan. Have you got time for preparing a pull request?

…

On Thu, Jun 15, 2017, 19:48 Elias Mårtenson ***@***.***> wrote: Based on the discussion on the mailing list, I think the only reliable solution to this is to change the behaviour of the code so that it switches to the buffer before moving the cursor. I think the propoer solution is to use pop-to-buffer first, and then place the cursor at the correct location using goto-char. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#391 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACyv7RPkFRTTIaQi_rQn-Xexri4w3J_ks5sEXEjgaJpZM4N6_9R> .

[lokedhs]

I can give it a try. I'll look at it in the next few days or so.

[luismbo]

@lokedhs any news? :-)

[lokedhs]

Actually, I haven't looked at it since this was discussed last. If I recall correctly, the code was quite complictaed, and there was no obvious simple way to make the change.

I changed the Emacs configuration variable to revert to the old behaviour and I haven't bothered with this since. This is of course not ideal, and a proper solution should be implemented.