I'm trying to manage better my downloads and packages. From previous posts to this list, I've concluded that I should always prefer an Mandrake RPM, but if a suitable RPM is not available then I should build an RPM from source tarball and install it that way, and that checkinstall is an easy way to do that (with thanks to Derek for that suggestion).
If I decide to keep the source after building the RPM, it belongs in /usr/local/src -- correct? And is there any reason to save the source? Charlie suggested in an earlier post that "If you're building RPMs from tarballs you'll want to set up a build directory and add that as a local source as well. Then you can install with urpmi from there...", which sounds like a good idea. Yet checkinstall does not place the RPMs it creates in a single directory but rather in several subdirectories under /usr/src/RPM/RPMS/, classified by architecture. I am tempted to find how to change checkinstall's behavior to place RPMs in a single directory and follow Charlie's advice, but before I do I'd like to understand what the logic is in having RPMs for different architectures in different directories. This issue becomes more relevant considering that I also download pre-rolled RPMs and want to save them in the same place, and those RPMs come in different architectures (at the moment, I have RPMs for i386, i586, and noarch). It would be more convenient for me to have all my RPMs (both pre-rolled and those I rolled myself, regardless of architecture) in one place, and so be able to define one local source for urpmi rather than many. Perhaps this is all much ado about nothing, but I want to understand what is going on and why before I start messing things up or reinventing the wheel. All advice welcome. -- Warren Post, Registered Linux user 241394 Santa Rosa de Copán, Honduras http://srcopan.vze.com/
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com