This is a quick way to set up an Arch Linux Pacman repository, hosted on Amazon S3 or an equivalent service. It can also be easily modified to use SSH/SCP to upload files to a more traditional web server, or a local machine.
It allows you to run your own repository, so that you can use
install your own packages, and have them easily upgraded when you publish
- Create one subdirectory for each repo you wish to establish.
- Inside this folder, create another one for each architecture you wish to
Makefile.configand update as required.
- Build your packages as normal, and put the output
.xzfiles into the relevant repo's architecture folder from step 2.
maketo construct the repo's database and upload it.
- As packages are updated, copy them in as per steps 4 and 5. You can delete
.xzfiles if you wish, or leave them around for a while in case anyone is looking for older versions if they need to downgrade.
If uploading to
s3://example.com/hello/ and the repo defined in
Makefile.config is called
myrepo, then anyone wishing to use the repo will
need to add this to
[myrepo] SigLevel = Optional TrustAll Server = http://example.com/hello/$repo/$arch/
pacman -Sy to read in the new repo, after which you can install any
packages from it with the usual
pacman -S packagename.