Fix environment mismatch error in docker env tests

[indiebrain]

Fixes ActiveRecord::EnvironmentMismatchError after a
bin/dev/bootstrap and bin/dev/test.

TLDR; The development and test environments were sharing the
mutualaid database instead of being isolated in their own databases.

References

• Remaining issues with docker development #718

[grenewode]

This looks pretty good to me - although I'm really only a casual dev-opser, and I can already tell your bash-foo is stronger than mine 😉 I do think it would be good to keep the ability to set the database with an environment variable when the docker image is built though.

[solebared]

Awesome! Thanks for looking into this 🙏🏾.
TBH, i'm still not sure the exact sequence of how all the pieces fit together, but i trust you and @grenewode on this 🧙🏾‍♂️.

[solebared]

To make sure i'm tracking: we're switching to presence because we want to ignore blank strings, whereas fetch would only ignore nils, correct?

[indiebrain]

💯That is correct.

The idea is that presence will return the contents of the ENV var when the ENV var contains what Ruby considers a truth-y value:

# WUMPUS=1 
ENV['WUMPUS'].presence 
# => "1"

# WUMPUS=found_me 
ENV['WUMPUS'].presence  
# => "found_me"

# WUMPUS=true 
ENV['WUMPUS'].presence  
# => "true"

# WUMPUS=false
ENV['WUMPUS'].presence  
# => "false"

but will return nil when the ENV var is unset, or contains a false-y value:

# WUMPUS=
ENV['WUMPUS'].presence  
# => nil

# WUMPUS=''
ENV['WUMPUS'].presence  
# => nil

The presence method is functionally equivalent to object.present? ? object : nil.

[indiebrain]

Thanks all, for the great conversations, and guidance.