Just saw a discussion about Linux package dependency issues. Software dependencies can indeed be a headache, especially for beginners. Personally, I rely on three main methods to resolve these dependency problems.
(The examples below all use Fedora.)
1. Try to Install Software Already Available in Your Repositories
This might sound like stating the obvious, but it happens often: we casually download a package from some random site, struggle through resolving all its dependencies, and then accidentally discover ――surprise!―― that this software was actually available in the system repositories all along. A simple yum install filename would have automatically analyzed and resolved dependencies and set everything up for us. This is a common pitfall but easily avoided: just check whether the software you want is already in your repositories before downloading it from elsewhere.
Of course, we shouldn’t rely solely on the default repositories, which don’t have a particularly rich selection. We need repositories like Livna and Freshrpms. Take the Livna repository as an example: once connected to the internet, enter the following in the terminal:
sudo rpm -ivh http://rpm.livna.org/livna-release-8.rpm
sudo rpm –import /etc/pki/rpm-gpg/RPM-GPG-KEY-livna
And that’s it.
This gives you easy access to the rich software selection on Livna. The installation method for Freshrpms and the Fedora_cn repository is similar; if you run into issues, a quick Google search should help.
2. Use yum localinstall
If the software we want is not found in any repository, we can download the package from elsewhere first, and then try using the yum localinstall command to install it. With this command, the system can automatically analyze dependencies and download the corresponding packages from repositories to resolve them.
The difference between this method and the previous one is that this method uses a locally downloaded package as a base, analyzes its dependencies, and then downloads the required packages from the repositories to satisfy them. In the previous method, both the dependency-resolving packages and the software package we actually want are downloaded from the repositories. That’s the only real distinction.
Once you’ve configured the repositories mentioned earlier, this method works perfectly in most situations. But if it still doesn’t solve the problem, we have to resort to the final method.
3. Leverage Web Search Effectively
Many websites offer RPM search services, like rpm find and others. My personal favorite, however, is rpm search at http://rpm.pbone.net. This site not only provides RPM package searches but also allows you to search for packages based on the required .so files.
For instance, if a downloaded package needs a particular .so file during installation, we can search for it by filename on Rpm Search. We can also filter search results by OS type, such as Fedora 7, Fedora 8, CentOS, and so on. The URL for this detailed search is: http://rpm.pbone.net/index.php3/stat/2/simple/2 . After finding the results, we can download and install them. Of course,