On Sat, Oct 20, 2001 at 08:58:07PM +0200, Ralf Habacker wrote: >> On Sat, Oct 20, 2001 at 10:16:46AM +0200, Ralf Habacker wrote: >> >4a. Some application (kdevelop,) couldn't be started >> > because of a runtime loader error. (Windows Error 0xc0000142) >> > I have tried to analyze this, but without success. >> > >> > This error is critical for this project, which implicated to me >> > to look after another way to port kde2. >> > Currently I'm trying out if "line" the linux emulator for windows >> > is a possible way for doing this. If have got running already some >> > applications. >> >> Hmm. A google search on this error doesn't unearth much interesting >> besides the fact that a DLL failed to initialize properly. Doesn't gdb >> tell you which one is having problems? I remember having similar >> problems, though, and they are really tricky to track down. >> >As I recognized it occurs after loading the last dll. In >http://sources.redhat.com/ml/cygwin-apps/2001-09/msg00135.html I have send >some informations. Do you have looked into ? >There are 4 threads and I'm not able to identify, which thread caused the >message box. In thread 1 an execve is called and hangs in kernel32. The other >threads are located in kernel32 too. Currently I'm out of ideas. :-(
I forgot that you had sent this previously. I now remember that you posted a lot of gdb info. I'm trying to get out of the business of having people send me their gdb stack traces and ask me to analyze them. There have been a flurry of these over the last few months and they are very tiriing. However, from looking at this message, I don't see anything out of the ordinary. To respond to your observations: Yes, every thread does have its own stack. Most of the threads were sitting in unknown system functions. Thread 1 was waiting for notification from a subprocess. I assume that you attached to a bash that had just invoked a process. If the process that bash invoked was displaying the 0xc0000142 popup then this would make sense. I am not sure why you attached to bash at this point. I don't think that bash is the problem. I assume that there is some problem in a DLL in a program that bash is invoking (./kjezz.exe in this case). If the program hadn't gone through initialization then it wouldn't show up in "ps" output unless you used "ps -W". In short, I don't see anything in your gdb session that would indicate that this is a cygwin problem. cgf
