Restarting an app triggers reupload of systemd/upstart scripts

[roback]

If you are not on latest master when deploying you will upload old Procfiles which could cause the app not to function correctly.

Discovered after updating the Ruby version in nile (from 2.4 -> 2.6). When the app later was restarted, using an older commit, the app was started with Ruby 2.4. Luckily Ruby 2.4 was still installed and the project was still compatible with 2.4 so the app worked.

This could be solved by just removing the following line (I'm not sure whether it will have any other implication as haven't tested this):

capistrano-twingly/lib/capistrano/twingly/tasks/service.rake

Line 55 in c39af85

invoke "deploy:export_service"

[dentarg]

If you are not on latest master when deploying

I think we would like to stop this from happening (still possibly fix this issue too), so I created #53