Bug 110610
Summary: | Java supplied with Matlab 6.1 has a big problem with glibc in Fedora 1 | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | bednar <borkows> |
Component: | glibc | Assignee: | Jakub Jelinek <jakub> |
Status: | CLOSED NOTABUG | QA Contact: | |
Severity: | high | Docs Contact: | |
Priority: | high | ||
Version: | 1 | CC: | david_kewley |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | i686 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2003-11-30 23:36:27 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
bednar
2003-11-21 18:30:29 UTC
Try using some less buggy JVM. Latest Sun JDK should work just fine for example. Or, as a workaround for the buggy JVM, you can try: gcc -O2 -shared -o ~/libcwait.so -fpic -xc - <<\EOF #include <errno.h> #include <sys/syscall.h> #include <sys/types.h> #include <sys/wait.h> pid_t __libc_wait (int *status) { int res; asm volatile ("pushl %%ebx\n\t" "movl %2, %%ebx\n\t" "movl %1, %%eax\n\t" "int $0x80\n\t" "popl %%ebx" : "=a" (res) : "i" (__NR_wait4), "0" (WAIT_ANY), "c" (status), "d" (0), "S" (0)); return res; } EOF and run the program with LD_PRELOAD=~/libcwait.so LD_ASSUME_KERNEL=2.4.19 in environment. Hi! That solution works really good! I wonder if this piece of code will by used as patch on glibc int the next wersion of glibc rpm available in updated of Fedora Core 1. Will it? No, it will not. We have been applying it as workaround for 2 years and vendors did not bother to fix their buggy software. Great :) I though I'd chime in here to describe what I see, in case it helps others who come across this bug report. I am not asking for anyone (Jakub) to fix anything. I find the following in RHL 9: First of all, I see basically the same symptoms that bednar sees with Matlab R12.1. I didn't use the fix that Jakub suggested. A fix for me is to install the latest release of Sun JRE 1.3.1, and tell matlab to use it. Matlab R12.1 comes only with JRE 1.1.8 included, and it runs fine on RHL 8, but it doesn't run on RHL 9. Providing JRE 1.3.1 makes Matlab R12.1 run on RHL 9; perhaps this fix would work on FC 1 as well. Incidentally Matlab R13 and R13SP1 both come with JRE 1.3.1, and use it by default. They also come with 1.1.8, but do not use it by default (at least in my installs). I have been trying to get J2RE 1.4.2 to work with Matlab, and have had no success. Googling suggests that all other reporters have failed as well. I was intrigued by Jakub's fix, but this did not help me use Matlab R13 with J2RE 1.4.2 on RHL 9. It has the same problem that bednar described when using J2RE 1.4.2: you can open e.g. the Matlab desktop Java window just fine, but if you open e.g. an editor window (which also uses Java), the editor window doesn't fully render, and Matlab permanently hangs (I've let it run overnight with no return to normal state). I had the same problem running java launcher of JRE 1.3.1_08. We have big java project that currently runs only on JRE 1.3.1. I installed JRE 1.3.1_12 and it works! Best regards, Ciril |