I've read through a number of bug reports that have been showing up throughout the internets. So far the community has come up with the following:
* Its not quite clear which yet, but its some combination of having smp and preempt compiled into the kernel. Ubuntu turns both of these on by default. So far, it looks like a kernel bug that is only triggered by skype.
* There has been more than one complaint from the community that skype client includes evil behaviours that make debugging the problem with debuggers difficult if not impossible.
* The bug shows up much more frequently if the sound device(s) used by skype are touched by something non-skype -- particularly esd. Users that either disable esd or disable Skype usually see an end to their problem.
This is a pretty disastrous thing for me. Not only do I not get incoming calls, but I usually don't even get a record that someone tried to call at all. For the few incoming calls that do get logged, those are indecipherable as the client's relativistic listing ("today", "yesterday" and such) is hopelessly confused because the calculation is based upon the current time rather than midnight.
Let us fix the problem so that you can make money from the service This problem is hurting a lot of people that are capable of diagnosing and providing fixes for this problem. Wouldn't it be nice to get some help? It doesn't hurt taking advantage of a free workforce...
I don't think opening up a little bit would hurt the company. The company apparently intends to make money not from phone service but software sales. The company also surely owns the copyright to its own codebase and probably has a zillion patents covering the unique ideas that are present. Any pain that the company did feel would certainly hurt less than encouraging potential long term users to migrate to tools such as asterisk and ekiga...
I know it won't happen. I'm sure that you guys have fought an uphill battle to convince the company to develop for this platform at all. Anyways, thanks and good luck tracing down the bugs. I'll try to hang in there for a bit longer.
QUOTE(berkus @ Fri Oct 13 2006, 09:16) [snapback]307578[/snapback]
Hi Günther,
we are puzzled by this no less than you. This seems to happen not only on latest versions, but more often on latest. I personally blame this on audio library we are using (GIPS), but it highly likely that this is a combination of factors on their side and on our side. We are trying to find a proper pattern now, but it seems to be far more easily reproducible for other people.
If you ping me on skype, we'll try to arrange some brainstorming session with you and some other people who have experienced this.