#55 jabber-activity-switch-to behaves incorrectly

[Anonymous]

This function documentation says: "switch back to the last non Jabber chat buffer used." Though it is expected to return to last (don't matter is it Jabber or not) buffer used.

Currently it behaves accordingly to documentation - if buffer, from which this function (through hotkey C-x C-j C-l) was used, was jabber buffer, you'll be not switched back to that buffer.

[Evgenii Terechkov]

Logged In: YES
user_id=1560441
Originator: NO

How do you see correct use case? More details, please.

[Nobody/Anonymous]

Logged In: NO

I've switched to some jabber conference (c1) manually. Message in other conference (c2) appeared, I've pressed C-x C-j C-l, which switched me to buffer c2. Next press of hotkey must switch me to buffer c1.

[Evgenii Terechkov]

Logged In: YES
user_id=1560441
Originator: NO

Currently, use scheme is n1->c1->c2->c3->n1 ("n" for normal buffers and "c" for chat, "->" for C-x C-j C-l press).

Do you want to use scheme n1->c1->n1->c2->n1->c3 ? IMHO, it is lead to useless keypressing.

Maybe, you can use function like this:
(defun switch-to-other-buffer () (interactive) (switch-to-buffer (other-buffer)))
instead of C-x C-j C-l?

[Nobody/Anonymous]

Logged In: NO

I have switched to c1 manually (by pressing C-x b) from another buffer (say, n1). Then c2 got some activity, I have pressed C-x C-j C-l, switched to c2 (this part is good). But second press will direct me to n1, while I want to switch to c1 (because that was latest buffer I've switched to manually).

[Evgenii Terechkov]

Logged In: YES
user_id=1560441
Originator: NO

You show only one "n" and two "c" buffers in your usage scheme (scheme you describe again is clean from your former comment), but how do you want to organize more complex scheme (with more that three "c" buffers)?

[Alexander Solovyov]

Logged In: YES
user_id=2024957
Originator: NO

Hm. At the end of chain of buffers with activity it just must switch me to last manually enabled buffer.

So if I was in c1 (manually enabled), it must drop me to that buffer irregardless of number of buffers in this chain.

[Evgenii Terechkov]

Logged In: YES
user_id=1560441
Originator: NO

How can we distinguish manualy switched jabber buffer from any of c(n) buffer? Where is the 'beggining' of chain? What buffer is 'origin'?

[Alexander Solovyov]

Logged In: YES
user_id=2024957
Originator: NO

That's a source of problem. Not sure how, but tracking.el (which is used in circe) solved this trouble.

I think I'll try to contact author and ask him about algorithm, though maybe tracking.el can be used itself (of course, needs testing).

[Evgenii Terechkov]

Logged In: YES
user_id=1560441
Originator: NO

My tests of tracking.el give poor results: some notifycations in modeline apear once and no more.