Summary: | javaws hangs (strace reveals segfault) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Robin Smidsrød <robin> |
Component: | Current packages | Assignee: | Java team <java> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | ||
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | strace -ff javaws |
Description
Robin Smidsrød
2004-12-17 07:52:47 UTC
Created attachment 46218 [details]
strace -ff javaws
This trace covers it all. I pressed Ctrl-C to break the application.
Does this also happen on a gcc 3.3 system? I'm not sure, because I don't have a GCC 3.3 system to check against. And it would take ages to recompile this system against GCC 3.3 again. But my hunch is that if I recompiled my whole system with -march=pentium3 and GCC 3.3 it would probably go away. I've had similar problems with KDE packages, and they applied some bugfixes to them, and the problem is gone. Could anyone else with a Intel Pentium-M-based machine give some inside info? PS: Isn't sun-jdk/javaws a binary build? How could it be affected? Linking with glibc? Since I'm running glibc-2.3.4.20041102-r0, that is quite possibly the problem. As he noted in that bug, the problem is still present in sun-jdk 1.5 (5.0) as installed by emerge on gentoo (masked as of know). Though, in my trace file I see no reference to waitid. Do they mean waitpid? If so, I cannot verify what they say about calling waitpid without arguments. But I'm not C-guru... :) But you should find several references to waitpid in my trace file, but I don't know if they're wrong or not. Someone more experienced need to check the trace. Yes, this hit me too... |