How Pacman works
Arch linux uses binary packages. This means that the code is already compiled when you download it, and accounts for the speed by which packages are installed. Packages live in repositories. The repositories are named in:
/etc/pacman.conf. For example the
core repository can be specified as follows:
[core] SigLevel = Never Include = /etc/pacman.d/mirrorlist
This means the actual location of the
core repository is specified by the URL's in
The actual packages are downloaded to the cache, which is normally located at:
/var/cache/pacman/pkg. You might find several versions of a given package here:
% cd /var/cache/pacman/pkg pkg % ls which* which-2.20-5-x86_64.pkg.tar.xz which-2.20-6-x86_64.pkg.tar.xz
The second important data structure is the pacman database. This is normally located at:
/var/lib/pacman/sync. For each repository you specify in
/etc/pacman.conf, there will be a corresponding database file located here. These files are tar gzipped. When you extract them, you will get one directory for each package.
% tree which-2.20-6 which-2.20-6 |-- depends `-- desc
depends lists the packages this package depends on. In the case above we are looking at the package called:
which. We can see the version that is installed is:
desc file, has a description of the file, such as the file size, MD5 hash, etc.
When you run:
pacman -Syu, pacman will try to sync the packages on your computer to the latest ones in the external repositories. It will compare the hashes on your local computer versus the ones on the external repos, if they are different then they download the newer one, install it, and update the corresponding database.
How is the master list of which packages are installed generated/stored? For example are the pacman databases built from somewhere else? Does the system get corrupted if they are deleted?