mjp-00.10-win32.tar.gz | Gzipped tar file of precompiled MJP components for Windows NT/2000/XP that requires some manual installation steps. |
mjp-00.10-win32.msi | Windows Installer file for Windows NT/2000/XP that provides a standard Windows setup program GUI to guide installation. |
mjp-00.10-i386-linux.tar.gz | Gzipped tar file of precompiled MJP components for i386 Linux that requires some manual installation steps. |
Source code is available via CVS.
The MJP 00.10 profiler agent has been tested on the following platforms:
OS | Compiled With | JRE |
---|---|---|
Windows 2000 SP4 | gcc 3.3.1 -mno-cygwin | Sun JDK 1.3.1-b24 |
Windows 2000 SP4 | gcc 3.3.1 -mno-cygwin | Sun JDK 1.4.0-b92 |
Windows 2000 SP4 | gcc 3.3.1 -mno-cygwin | Sun JDK 1.4.1_01-b01 |
Windows 2000 SP4 | gcc 3.3.1 -mno-cygwin | Sun JDK 1.4.2_02-b03 |
Windows 2000 SP4 | gcc 3.3.1 -mno-cygwin | Sun JDK 1.4.2_03-b02 |
Windows XP | gcc 3.3.1 -mno-cygwin | Sun JDK 1.4.1_01-b01 |
Red Hat Linux 9.0 | gcc 3.2.2 | Sun JDK 1.4.2_03-b02 |
The MJP 00.10 GUI has been tested on the following platforms:
OS | JRE |
---|---|
Windows 2000 SP4 | Sun JDK 1.4.0-b92 |
Windows 2000 SP4 | Sun JDK 1.4.1_01-b01 |
Windows 2000 SP4 | Sun JDK 1.4.2_02-b03 |
Windows 2000 SP4 | Sun JDK 1.4.2_03-b02 |
Windows XP | Sun JDK 1.4.1_01-b01 |
Red Hat Linux 9.0 | Sun JDK 1.4.2_03-b02 |
Note that I don't have a enough time to test MJP thoroughly in every configuration, so most of the testing is done on the configuration I happen to be using right before the release. In particular, testing on Linux is sparse because I rarely do any Java development on that platform. I welcome contributions by others in the form of confirmed working configurations.
The JavaHelp HTML viewer leaves a lot to be desired. Until a better viewer is found, the MJP online help will be rough around the edges. If you avoid following external links, you should avoid most of the problems.
No online help index yet exists, so there is nothing to match. Use the full text search in the mean time.
There is a known bug in Sun's JVM that causes the call stack to appear empty when the thread is sleeping. See http://developer.java.sun.com/developer/bugParade/bugs/4860655.html for more information about the bug. As of JDK 1.4.2_03, the bug is still not fixed. The bug causes MJP to not charge samples to threads that are sleeping. Obviously this will throw off the results considerably.