Is there any progress with the automated package builds? Either they are not running anymore, or all packages always build (if so, cool ;)

I still think it would be best if a build kicked off once every $timeunit (24h?) and always sent a summary including a link to a full log (which is viewable without any silly accounts). That would take the guesswork out of if packages are failing or the CI system is failing. In case that scheme proves impossible to achieve with the Docker infrastructure, maybe someone has ideas on a better suited service. Even just using a simple custom AWS instance (or equivalent) shouldn't be bankruptatory with the CPU cycles we currently need to build all packages. Thoughts and/or volunteers?

Reply via email to