The m4 error noted in the title and again below is another error I received quite awhile back with very little detail in the notes. I had mentioned that I resolved the error by uninstalling the x86_64 version of m4 and installing the i386 version. After further investigation I am sure the issue was not related to being 32-bit versus 64-bit but instead related to the m4 version instead. Below I describe note the error and how to resolve the error.
I received the below error when attempting to run a setup.py script to finish installing a certain application. The application was being installed on a 64-bit CentOS Linux system and during the initial part of the install I had attempted to use 32-bit python which ended up causing a bunch of errors. After I believed that all of the errors were corrected I still received an i386 error message. Since I am not very familiar with Python setup scripts it took me a couple minutes to figure out what needed to be done.
Installing the EPEL repository on CentOS Linux is easy to do so the below is simply a reference point if needed by anyone. EPEL stands for Extra Packages for Enterprise Linux and is a community based repository that was created for add-on Fedora Linux packages. There are many cases where installation instructions assume that the EPEL repo is already installed so if you run into an error because of a missing package or dependency package the EPEL repository is a great place to start. Below I have listed the i386(32-bit) and x86_64(64-bit) EPEL repository package list links which will allow you to easily verify that the packages you need are available before installing the EPEL repository. After the EPEL repository package list links there are separate commands listed to install the EPEL repository on 32-bit CentOS Linux 5.X and to install the EPEL repository on 64-bit CentOS Linux 5.X.
CUDA-Multiforcer: error while loading shared libraries: libcudart.so.2: cannot open shared object file: No such file or directory
After installing a new server recently that uses Nvidia video cards to take advantage of their GPU capabilities I ran into an error while running a program called CUDA-Multiforcer. The CUDA-Multiforcer application is a security audit tool used to check the strength of password hashes encrypted using MD4, MD5, NTLM, or SHA1. After compiling the CUDA-Multiforcer and installing various Nvidia packages I ran into an error with the libcudart.so.2 library. Below I describe the error in more detail as well as what I did to resolve the error.
Recently we installed a development server running 64-Bit (x86_64) CentOS Linux. One of the things we needed to install was Python version 2.6 without overwriting the default Python version 2.4 on the server. The reason for not wanting to overwrite the default Python version is because other packages such as the Yum Package Manager depend on Python version 2.4. Anyhow in the past I have used the CHL repository for the Python version 2.6 packages because they are compiled to run as python26 or python2.6 and using the python command still launches python 2.4. Anyhow after installing the CHL yum repository as noted here we had issues installing python version 2.6 which I describe below along with how to resolve the issue.