Hi there, I am using Gnus from Emacs always. I have noticed that point could move randomly when going back to summary buffer by typing `=' in the article buffer. The issue started to appear probably since Emacs 25 but I didn't stumble upon a recipe to reproduce the issue until just now. See the attached screenshot of a thread. 1. Visit the first article in a thread and make sure to move point to the article buffer. 2. Press `n' to move to next article while point is still in the article buffer. 3. Press `n' again 4. Press `=' At this point the article buffer is hidden and point is moved back to the summary buffer. But instead of being placed on the article you last read it is moved to the article before entering the article buffer. I found this new behaviour very confusing. Sometimes I have pressed `n' many times in the article buffer and the summary buffer indicates where the point will be (by showing an empty square) but as soon as pressing `=' it changes. Leo
On Monday, 25 Feb 2019 at 07:40, Leo Liu wrote:
> I am using Gnus from Emacs always. I have noticed that point could move
> randomly when going back to summary buffer by typing `=' in the article
> buffer. The issue started to appear probably since Emacs 25 but I didn't
> stumble upon a recipe to reproduce the issue until just now.
Yes, I have had this behaviour for a long time now. Annoying. I've no
idea why nor how to fix it.
--
Eric S Fraga via Emacs 27.0.50 & org 9.2.1 on Debian 9.7
Leo Liu <sdl.web@gmail.com> writes:
> Hi there,
>
> I am using Gnus from Emacs always. I have noticed that point could move
> randomly when going back to summary buffer by typing `=' in the article
> buffer. The issue started to appear probably since Emacs 25 but I didn't
> stumble upon a recipe to reproduce the issue until just now.
>
> See the attached screenshot of a thread.
>
> 1. Visit the first article in a thread and make sure to move point to
> the article buffer.
>
> 2. Press `n' to move to next article while point is still in the article
> buffer.
>
> 3. Press `n' again
>
> 4. Press `='
Oddly, if you use "h" in the same scenario, it behaves correctly. I
looked briefly at "=", which gets into Gnus' windowing systems, and
couldn't immediately see what might be going wrong.
Maybe report this as a Gnus bug, so it gets added to the tracker?
On 2019-02-25 21:12 +0000, Eric S Fraga wrote: > Yes, I have had this behaviour for a long time now. Annoying. I've no > idea why nor how to fix it. On 2019-02-25 13:24 -0800, Eric Abrahamsen wrote: > Oddly, if you use "h" in the same scenario, it behaves correctly. I > looked briefly at "=", which gets into Gnus' windowing systems, and > couldn't immediately see what might be going wrong. > > Maybe report this as a Gnus bug, so it gets added to the tracker? > I forgot to attach the screenshot but you two seem to get it anyway ;) I think the issue might be `gnus-summary-position-point' in `gnus-summary-next-article'. When the window of the summary buffer is not selected goto-char won't work properly, instead one must use `set-window-point' to move point. Leo
Leo Liu <sdl.web@gmail.com> writes: > 1. Visit the first article in a thread and make sure to move point to > the article buffer. > > 2. Press `n' to move to next article while point is still in the article > buffer. > > 3. Press `n' again > > 4. Press `=' > > At this point the article buffer is hidden and point is moved back to > the summary buffer. But instead of being placed on the article you last > read it is moved to the article before entering the article buffer. I'm unable to reproduce this bug in Emacs 27 -- when hitting `=', point moves to the last read article in the summary buffer... -- (domestic pets only, the antidote for overdose, milk.) bloggy blog: http://lars.ingebrigtsen.no
On 2019-06-22 12:38 +0200, Lars Ingebrigtsen wrote:
> I'm unable to reproduce this bug in Emacs 27 -- when hitting `=', point
> moves to the last read article in the summary buffer...
To reproduce the bug both buffers, article and summary must be visible.
It seems what I wrote on <m1va17mhsh.fsf@gmail.com> is still true with
27 which I quote:
,----
| I think the issue might be `gnus-summary-position-point' in
| `gnus-summary-next-article'. When the window of the summary buffer is
| not selected goto-char won't work properly, instead one must use
| `set-window-point' to move point.
`----
Leo Liu <sdl.web@gmail.com> writes: > On 2019-06-22 12:38 +0200, Lars Ingebrigtsen wrote: >> I'm unable to reproduce this bug in Emacs 27 -- when hitting `=', point >> moves to the last read article in the summary buffer... > > To reproduce the bug both buffers, article and summary must be visible. Yes, both are visible for me (but `=' makes only the summary buffer visible). -- (domestic pets only, the antidote for overdose, milk.) bloggy blog: http://lars.ingebrigtsen.no
Leo Liu <sdl.web@gmail.com> writes: > On 2019-06-22 12:38 +0200, Lars Ingebrigtsen wrote: >> I'm unable to reproduce this bug in Emacs 27 -- when hitting `=', point >> moves to the last read article in the summary buffer... > > To reproduce the bug both buffers, article and summary must be visible. Yes, both are visible for me (but `=' makes only the summary buffer visible). -- (domestic pets only, the antidote for overdose, milk.) bloggy blog: http://lars.ingebrigtsen.no
On 2019-06-24 12:53 +0200, Lars Ingebrigtsen wrote:
> Yes, both are visible for me (but `=' makes only the summary buffer
> visible).
Indeed the bug is fixed in 27.1 which I just started using today ;)