About me: My name is Solène Rapenne, pronouns she/her. I like learning and sharing knowledge. Hobbies: '(BSD OpenBSD Qubes OS Lisp cmdline gaming security QubesOS internet-stuff). I love percent and lambda characters. OpenBSD developer solene@. No AI is involved in this blog.

Contact me: solene at dataswamp dot org or @solene@bsd.network (mastodon).

I'm a freelance OpenBSD, FreeBSD, Linux and Qubes OS consultant, this includes DevOps, DevSecOps, technical writing or documentation work. If you enjoy this blog, you can sponsor my open source work financially so I can write this blog and contribute to Free Software as my daily job.

Reuse of OpenBSD packages for trying runtime

Written by Solène, on 19 September 2021.
Tags: #openbsd #unix

Comments on Fediverse/Mastodon

1. Introduction §

So, I'm currently playing with OpenBSD trying each end user package (providing binaries) and see if they work when installed alone. I needed a simple way to keep packages downloaded and I didn't want to go the hard way by using rsync on a package mirror because it would waste too much bandwidth and would take too much time.

The most efficient way I found rely on a cache and ordering the source of packages.

2. pkg_add mastery §

pkg_add has a special variable named PKG_CACHE that when it's set, downloaded packages are copied in this directory. This is handy because every time I will install a package, all the packages downloaded by will kept in that directory.

The other variable that interests us for the job is PKG_PATH because we want pkg_add to first look up in $PKG_CACHE and if not found, in the usual mirror.

I've set this in my /root/.profile

export PKG_CACHE=/home/packages/
export PKG_PATH=${PKG_CACHE}:http://ftp.fr.openbsd.org/pub/OpenBSD/snapshots/packages/amd64/

Every time pkg_add will have to get a package, it will first look in the cache, if not there it will download it in the mirror and then store it in the cache.

3. Saving time removing packages §

Because I try packages one by one, installing and removing dependencies takes a lot of time (I'm using old hardware for the job). Instead of installing a package, deleting it and removing its dependencies, it's easier to work with manually installed packages and once done, remove dependencies, this way you will keep already installed dependencies that will be required for the next package.

#!/bin/sh

# prepare the packages passed as parameter as a regex for grep
KEEP=$(echo $* | awk '{ gsub(" ","|",$0); printf("(%s)", $0) }')

# iterate among the manually installed packages
# but skip the packages passed as parameter
for pkg in $(pkg_info -mz | grep -vE "$KEEP")
do
	# instead of deleting the package
	# mark it installed automatically
	pkg_add -aa $pkg
done

# install the packages given as parameter
pkg_add $*

# remove packages not required anymore
pkg_delete -a

This way, I can use this script (named add.sh) "./add.sh gnome" and then reuse it with "./add.sh xfce", the common dependencies between gnome and xfce packages won't be removed and reinstalled, they will be kept in place.

4. Conclusion §

There are always tricks to make bandwidth and storage more efficient, it's not complicated and it's always a good opportunity to understand simple mechanisms available in our daily tools.