Making SECRET_KEY the same in test and default

[craig-rueda]

CATEGORY

Choose one

• Bug Fix
• Enhancement (new features, refinement)
• Refactor
• Add tests
• Build / Development Environment
• Documentation

SUMMARY

The Flask secret key was different in the default config as compared to the test config. This caused issues when running the "load examples" and then subsequently running unit tests against the same DB in order to repro issues, etc.

REVIEWERS

@dpgaspar @john-bodley

[codecov-io]

Codecov Report

Merging #8838 into master will decrease coverage by 7.14%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #8838      +/-   ##
==========================================
- Coverage   65.88%   58.73%   -7.15%     
==========================================
  Files         483      359     -124     
  Lines       24176    11324   -12852     
  Branches     2779     2779              
==========================================
- Hits        15928     6651    -9277     
+ Misses       8070     4495    -3575     
  Partials      178      178

Impacted Files	Coverage Δ
superset/connectors/druid/__init__.py
superset/views/datasource.py
superset/models/sql_lab.py
superset/connectors/druid/models.py
superset/utils/import_datasource.py
superset/views/database/forms.py
superset/tasks/schedules.py
superset/db_engine_specs/bigquery.py
superset/examples/tabbed_dashboard.py
superset/db_engine_specs/athena.py
... and 114 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b6f31ab...ce295cb. Read the comment docs.

[mistercrunch]

Seems reasonable but this will break existing test environments.

[craig-rueda]

Seems reasonable but this will break existing test environments.

Howso? This file is used to provide overrides when running tox. People don't use it for running stuff in any "real" environments, do they?

[mistercrunch]

As discussed if people have an existing tox env (msyql / postgres / persisted sqlite) they'll need to nuke them first and reload examples.

I think it's worth doing but it effectively breaks all tox environments for every contributor :(

I wonder how cryptic the error message is? Maybe if what's happening is pretty clear it's not as bad?

[john-bodley]

I'm onboard with these changes. I think it's not too much effort to have developers recreate their tox environments.