nnimap closing connections on gnus-group-suspend

nnimap closing connections on gnus-group-suspend

[Florian Ragwitz]

[-- Attachment #1: Type: text/plain, Size: 584 bytes --]

On gnus-group-suspend, the old nnimap seems to have kept open the
connection to the IMAP server, whereas the new one closes them and
reconnects when Gnus is started the next time.

I've grown quite used to suspending Gnus a lot, so the extra time spent
on connecting again on the next time I want to read mail has been
slightly annoying me with the new nnimap code.

Is this change intentional and should I be writing some code to use
instead of gnus-group-suspend so I can just bury the group buffer and
close all the other gnus buffers without closing established IMAP
connections?

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: nnimap closing connections on gnus-group-suspend

[Steinar Bang]

>>>>> Florian Ragwitz <rafl@debian.org>:

> On gnus-group-suspend, the old nnimap seems to have kept open the
> connection to the IMAP server, whereas the new one closes them and
> reconnects when Gnus is started the next time.

FWIW I think the new behaviour is the correct one, and I've found the
old behaviour annoying, and I've had to start a new gnus in a new emacs,
after it has gotten confused after a suspend-resume.

Re: nnimap closing connections on gnus-group-suspend

[Florian Ragwitz]

[-- Attachment #1: Type: text/plain, Size: 1205 bytes --]

Steinar Bang <sb@dod.no> writes:
>>>>>> Florian Ragwitz <rafl@debian.org>:
>> On gnus-group-suspend, the old nnimap seems to have kept open the
>> connection to the IMAP server, whereas the new one closes them and
>> reconnects when Gnus is started the next time.
>
> FWIW I think the new behaviour is the correct one, and I've found the
> old behaviour annoying, and I've had to start a new gnus in a new
> emacs, after it has gotten confused after a suspend-resume.

How come? Automatic reconnection after the IMAP connection got closed
for whatever reason seems to work just fine with both the old and the
new nnimap. (I do remember issues with the old one and starttls that I
managed to work around by just using ssl instead of starttls.)

Is there another reason why Gnus would get confused when keeping
connections open while suspending?

While personally I still prefer the old behaviour, I don't have any
strong feelings about it either way. I'd be fine with writing a couple
more lines in my configuration to keep it the way I like if the new
behaviour is decided to be the preferred one. However, disconnecting on
suspend only to work around other unrelated bugs doesn't seem quite
right to me.

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: nnimap closing connections on gnus-group-suspend

[Steinar Bang]

>>>>> Florian Ragwitz <rafl@debian.org>:

> ... However, disconnecting on suspend only to work around other
> unrelated bugs doesn't seem quite right to me.

Well, er... nobody has worked around my suspend/resume issues, because
I've never reported them (I think).  I may have sent an occasional
message to this list, but I can't remember any responses.

I've mostly just shrugged and started a new gnus in a new emacs process.

Note that for me suspend/resume means waking up in a different LAN than
the one I left, or maybe no LAN at all, so disconnecting befor suspend
seems like the smartest thing to do (and the nicest thing to do to the
server, instead of letting a connection dangle).

But that's just my 80 øre (~ .10 Euros).

Re: nnimap closing connections on gnus-group-suspend

[Lars Magne Ingebrigtsen]

Florian Ragwitz <rafl@debian.org> writes:

> Is this change intentional and should I be writing some code to use
> instead of gnus-group-suspend so I can just bury the group buffer and
> close all the other gnus buffers without closing established IMAP
> connections?

It wasn't intentional in that I thought about it.  :-)  But I think it's
the correct behaviour.  If the user taps `z', then it seems likely that
they want Gnus to go to sleep, and that includes severing any network
connections, I think.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: nnimap closing connections on gnus-group-suspend

[Steinar Bang]

>>>>> Lars Magne Ingebrigtsen <larsi@gnus.org>:

> ...  If the user taps `z', then it seems likely that they want Gnus to
> go to sleep, and that includes severing any network connections, I
> think.

Ah... I misread.  I've never used that command.  Somehow I got the
impression that this was about gnus offline/online `J j'.

Please disregard my comments.

Re: nnimap closing connections on gnus-group-suspend

[Florian Ragwitz]

[-- Attachment #1: Type: text/plain, Size: 861 bytes --]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Florian Ragwitz <rafl@debian.org> writes:
>
>> Is this change intentional and should I be writing some code to use
>> instead of gnus-group-suspend so I can just bury the group buffer and
>> close all the other gnus buffers without closing established IMAP
>> connections?
>
> It wasn't intentional in that I thought about it.  :-) But I think
> it's the correct behaviour.  If the user taps `z', then it seems
> likely that they want Gnus to go to sleep, and that includes severing
> any network connections, I think.

OK, I can live with suspend behaving that way. Thanks for clarifying
this.

Would there happen to be another command similar to gnus-group-suspend,
but does imply not closing connections? If not, would patches for such a
thing be accepted? Are there name suggestions for such a thing even?

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: nnimap closing connections on gnus-group-suspend

[Lars Magne Ingebrigtsen]

Florian Ragwitz <rafl@debian.org> writes:

> Would there happen to be another command similar to gnus-group-suspend,
> but does imply not closing connections? If not, would patches for such a
> thing be accepted? Are there name suggestions for such a thing even?

If you don't want Gnus to be suspended, then why do anything?  Just
switch to a different buffer...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen