> At the time I thought it was failing everywhere and even now we know it fails > on non-KDE which makes it unusable for a large percentage of users. Knowing > that it works on KDE and the fact that it's a KDE tray thing, maybe the > severity should be important or serious.
I've marked it important -- I don't believe it's RC, because the problem is along the lines of "the application breaks when used in a way that was not intended". As mentioned earlier, I will do a new upload shortly that makes it clearer that this is a KDE-specific thing. After that I will reassign this as a wishlist bug to kdelibs for better cross-WM support for systray applets (since they all [should] derive from the same KSystemTray base class). > > (ii) install kcpuload and let me know if _that_ starts for you? > > It starts but doesn't bring anything up, I guess because of the lack of KDE > panel to go in? Mm, I expect so -- I just did a couple of tests in other WMs also and found the same thing. There was an old package (kdetrayproxy) which provided KDE systray support for other WMs -- this is now part of kdelibs proper, but i'm not sure what the magic invocation is to get it working. > After running kcpuload I'm not longer seeing the kbuildsycoca errors (in the > original report) when I invoke kteatime from a shell. FWIW, I believe those errors to be harmless in this context. > Can I expect to be able to run these things (kteatime and kcpuload) if > I don't have a KDE panel? Unless you have the kdetrayproxy magic working, I believe no. > And even then, if there is no KDE tray present when invoked it should > probably point out that the tray is required. Otherwise how are users > expected to know it's not just a bug? You're quite right; this is a worthwhile thing to add. This is common to all systray applets -- essentially it is a request for better error handling in the KSystemTray class from kdelibs. > One other thing I noticed, kcpuload puts itself in the background and returns > the prompt, kteatime doesn't. It might be nice to make them consistant (but > that's another wishlist bug). FWIW, I expect that this is because kcpuload uses KUniqueApplication (only run one copy at a time), whereas kteatime just uses KApplication (run as many as you like). b. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
