Re: [Licq-devel] Some GUI patches

Jörg Mensmann <[email protected] Thu, 17 Nov 2011 21:59:40 +0100
Newsgroups gmane.network.licq.devel
Message-ID <[email protected]>
Anders Olofsson <[email protected]> wrote:
> As Licq doesn't have any state handling other than normal
> configuration I'm not sure it worth pretending anything by having
> support for the parameter at all.

We at least need to provide a command line to the session manager with
which it restart the client with the same configuration (which is
implemented right now). Besides that, I agree that Licq does not need to
store more session state.

> Do you know if the window manager does anything useful for us when the
> session parameter is present?

AFAIK the session id is just passed to the client so that it can restore
any state it previously stored under that id. Licq doesn't use this
mechanism at all right now, and therefore the parameter could be removed
in LicqGui::commitData(). Actually, this would also fix the problem that
an additional "-session" parameter is added to the command line on each
session restore, i.e. Licq's command line grows longer and longer.

> I did some testing and found that the current event grabbing with X
> only worked if Caps Lock and Num Lock were off. I just committed a fix
> for this, please check if it works better for you now.

Right, that was the problem, it works now.

Fiddling with this a bit, I noticed a related issue, which does not
appear with the KDE-style shortcut handling: When setting the show/hide
shortcut to Alt+I or Meta+I, pressing this shortcut in the Licq main
window will switch to the "Ignore" group. Looks like the key event is
not consumed completely.

ciao
  J=F6rg