Re: Issue with Filter Folder and wl-folder-empty-trash

[Kazuhiro Ito <kzhr@d1.dion.ne.jp>]

> I've added the follwing filter folders to my config:
> 
> /flag:unread/%INBOX	       "Unread"
> /subject:[Spam]/%INBOX	       "Spam"
> 
> While the first folder works, the 2nd one does not show the correct
> message numbers:
> 
> Inbox:0/275/1977
> Unread:0/275/275
> Spam:*/*/1977
> 
> When I open&close the Spam folder I get the correct counter:

I think it would be intended behaviours.  Unread count is one of
folder's internal information and updated while checking (e.g.
wl-folder-check-current-entity, `s' in Folder mode).  But Field search
result is unknown until actual search is done and it may need much
time.  So you are required to start search explicitly.

> But after pressing "s" the counter is wrong again (*/*/1977).
> With "S" (wl-folder-sync-current-entity) it also works.

I imagine that the target folder is recognized as `updated', therefore
search result is expired, but I have not confirmed at all.

> And why does '[Spam]' also match 'Spam'?
> Do I need some regex magic here?

Although I don't know that is intentional, value for field search
('[Spam]', in your case) is used as regexp.  wl-summary-pick also
treats field search value as regexp beacause the same routine is
used.

> There is a bug in wl-folder-empty-trash in Emacs 24.x.
> The folder stays open in Folder buffer.

Not reproducible for me in emacs-trunk.

> What's the best way to disable the "*trace of SMTP session" buffer in
> flim? At the moment I use:
> 
> diff --git a/smtp.el b/smtp.el
> index fe511e9..dfa07b0 100644
> --- a/smtp.el
> +++ b/smtp.el
> @@ -374,7 +374,8 @@ BUFFER may be a buffer or a buffer name which contains mail 
>           (smtp-open-connection (current-buffer) server smtp-service))
>         (make-local-variable 'smtp-read-point)
>         (setq smtp-read-point (point-min))
> -       (funcall smtp-submit-package-function package)))))
> +       (funcall smtp-submit-package-function package)
> +       (kill-buffer)))))

I think it would be better that you could customize keeping trace
buffer.  How about modification that trace buffer is killed only when
smtp-debug is nil?

-- 
Kazuhiro Ito