From kde-devel  Sat Jun 28 10:38:24 2014
From: Ian Wadham <iandw.au () gmail ! com>
Date: Sat, 28 Jun 2014 10:38:24 +0000
To: kde-devel
Subject: Should KDED4 run all the time?
Message-Id: <2B05B730-14E2-45D2-922E-6AE82CD04ACF () gmail ! com>
X-MARC-Message: https://marc.info/?l=kde-devel&m=140395194825994

Hi guys,

When fixing some bugs in the KCrash-DrKonqi sequence on Apple OS X,
I have come to a point where Dr Konqi attempts to call kded4, using DBus,
and issues a message "Failed to communicate with kded. Make sure it is running."

This is actually a great advance: it has rarely been possible to run Dr Konqi
and get even a backtrace on Apple OS X until now.  However progress towards
reporting the crash. on bugs.kde.org, is blocked at the point when you log in to
BKO.  Before fixing this, I have some questions.

1. KDED4 runs all the time on a KDE desktop I believe, judging from what
    is said on http://kdepepo.wordpress.com/2011/05/11/troubleshooting-kded4-bugs/,
    but does it need to on other desktops, such as Gnome, Windows and OS X?

    ATM it never runs in MacPorts with KDE apps on Apple OS X, and AFAIK
    that has had little effect on running KDE apps there.

2. The reason Dr Konqi calls KDED4 is to load KCookieJar and check 
    whether cookies are enabled for bugs.kde.org on this user's system.
    If they are not, Dr Konqi will not process the bug report any further.
    Is this approach really necessary?

3. AFAIK cookies are needed on BKO only so that the user can stay logged
    in to BKO if he or she wishes.  So would it be enough just to issue a warning
    message about that and move on, if KDED4 or KCookieJar or the required
    cookies are unavailable?

4. If cookies are enabled, is there any way to attempt to connect immediately
    to bugs.kde.org from Dr Konqi and skip entering the ID and password, in case
    the user is already logged in there?

I am just wondering what is the best way to fix these issues and make proper
crash-reporting possible in KDE applications on Apple OS X.

Cheers,  Ian W.



>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<