Using 'd' in Summary buffer produces surprising results (sometimes)

Using 'd' in Summary buffer produces surprising results (sometimes)

[Nick Dokos]

If I try to 'd' my way through a group, much of the time I get what I
expect: the message is marked 'r' and the next message is shown.

However, sometimes a 'd' will take me back to a previous message
(usually, perhaps always, in the thread I'm reading - I don't remember
any cases where it has taken me to a message in a previous thread, but
I'm not sure), one that I've already read. Successive 'd's will step
through all the intermediate messages until I get to a new message,
where another 'd' will take me back to the same previous message as
above and the whole process is repeated a few times. It usually stops
when I get to the first message of the next thread. Then later on it
might start again with a message in a different thread.

Has anybody seen this behavior? Anybody know what causes it? Is there
some way to avoid it? It does not always happen but it happens often
enough to be annoying. If anybody has any suggestions about debugging
this or about more information I can provide, I'd appreciate the help.
The main problem is that I don't know when this will happen: next time
it happens, I plan to edebug through gnus-summary-mark-as-read-forward,
but I'm wondering if there is a better way to go.

GNU Emacs 24.3.50.2 (x86_64-unknown-linux-gnu, GTK+ Version 2.24.10) of
2013-06-17 on pierrot

Gnus v5.13 (bundled with above - haven't tried upstream yet).

gnus-show-threads is t. 

Thanks!
-- 
Nick

Re: Using 'd' in Summary buffer produces surprising results (sometimes)

[Ted Zlatanov]

On Sat, 29 Jun 2013 22:44:04 -0400 Nick Dokos <ndokos@gmail.com> wrote: 

ND> If I try to 'd' my way through a group, much of the time I get what I
ND> expect: the message is marked 'r' and the next message is shown.

ND> However, sometimes a 'd' will take me back to a previous message
ND> (usually, perhaps always, in the thread I'm reading - I don't remember
ND> any cases where it has taken me to a message in a previous thread, but
ND> I'm not sure), one that I've already read. Successive 'd's will step
ND> through all the intermediate messages until I get to a new message,
ND> where another 'd' will take me back to the same previous message as
ND> above and the whole process is repeated a few times. It usually stops
ND> when I get to the first message of the next thread. Then later on it
ND> might start again with a message in a different thread.

ND> Has anybody seen this behavior? Anybody know what causes it? Is there
ND> some way to avoid it? It does not always happen but it happens often
ND> enough to be annoying. If anybody has any suggestions about debugging
ND> this or about more information I can provide, I'd appreciate the help.
ND> The main problem is that I don't know when this will happen: next time
ND> it happens, I plan to edebug through gnus-summary-mark-as-read-forward,
ND> but I'm wondering if there is a better way to go.

I've never seen this, myself.  If possible, try to replicate it with a
bare .emacs.

Ted

Re: Using 'd' in Summary buffer produces surprising results (sometimes)

[Nick Dokos]

Ted Zlatanov <tzz@lifelogs.com> writes:

> On Sat, 29 Jun 2013 22:44:04 -0400 Nick Dokos <ndokos@gmail.com> wrote: 
>
> ND> If I try to 'd' my way through a group, much of the time I get what I
> ND> expect: the message is marked 'r' and the next message is shown.
>
> ND> However, sometimes a 'd' will take me back to a previous message
> ND> (usually, perhaps always, in the thread I'm reading - I don't remember
> ND> any cases where it has taken me to a message in a previous thread, but
> ND> I'm not sure), one that I've already read. Successive 'd's will step
> ND> through all the intermediate messages until I get to a new message,
> ND> where another 'd' will take me back to the same previous message as
> ND> above and the whole process is repeated a few times. It usually stops
> ND> when I get to the first message of the next thread. Then later on it
> ND> might start again with a message in a different thread.
>
> ND> Has anybody seen this behavior? Anybody know what causes it? Is there
> ND> some way to avoid it? It does not always happen but it happens often
> ND> enough to be annoying. If anybody has any suggestions about debugging
> ND> this or about more information I can provide, I'd appreciate the help.
> ND> The main problem is that I don't know when this will happen: next time
> ND> it happens, I plan to edebug through gnus-summary-mark-as-read-forward,
> ND> but I'm wondering if there is a better way to go.
>
> I've never seen this, myself.  If possible, try to replicate it with a
> bare .emacs.
>
> Ted
>

Big screw-up on my part: what I was seeing was *not* with the gnus that
is bundled with emacs (I am on GNU Emacs 24.3.50.2
(x86_64-unknown-linux-gnu, GTK+ Version 2.24.10) of 2013-06-17 on
pierrot). It was with the devel version of gnus which I must have
enabled at some point and then promptly forgot about. Sorry about that.

I've gone back to the bundled version and I have not seen this behavior
even going through a fairly large group. I'll keep an eye out for it
though.  The next question then is: has anybody seen the above behavior
with the development version? `git describe' says

$ git describe
m0-7-150-g9cd1966

but I've just pulled and will try again with a more modern version
and (as Ted suggests) a simpler configuration (although my gnus config
is *very* simple).

Thanks!
-- 
Nick