Christian Faulhammer wrote: > That's a nice starting point to have a look if they aren't installed > they are unpopular or because they fail to build (which makes them a > candidate for removal).
I'm not following - how would we find out about the reason a package is never reported installed? I've been working on the least-installed reporting stuff today. For now I have split it in two halves: (1) a top N least-installed packages table (only looking at packages with one install or more) (2) an plaintext file with a list of _all_ packages with zero installations reported Latter is linked from the report's TOC on the main page, too. Benefits I see with this approach include: - Complete zero-installs data is presented, not excerpts - Still, the report does not explode - Suitable for further machine-processing Small adjustments to this behaviour should be no problem. Sebastian (1) http://smolt.hartwork.org:45678/static/stats/gentoo.html#installed_packages_least_installed_main_tree (2) http://smolt.hartwork.org:45678/static/stats/gentoo_zero_installs_packages.txt