Ignore CI variable in roulette experiment default specs

Jay McCurerequested to merge
jmc-roulette-spec-ignore-ci-var into master

What does this MR do and why?

The project-level CI variable ROULETTE_HIDE_REVIEWER_COLUMN_PERCENT is set to 100, which leaks into spec/tooling/danger/roulette_experiment_spec.rb. The two examples that assert the default 50% split fail on every pipeline that runs this spec.

This stubs the variable to nil for the .hide_reviewer_column? examples so the default applies. Contexts that set their own value still override it.

Evidence
  • Failing job: https://gitlab.com/gitlab-org/gitlab/-/jobs/17187979576 (roulette_experiment_spec.rb:14 and :21, failed on retry too; hidden_percent evaluated to 100).
  • Locally, with ROULETTE_HIDE_REVIEWER_COLUMN_PERCENT=100: 2 failures before, 0 after (11 examples). Also passes with the variable unset.

References

Screenshots or screen recordings

N/A, spec-only change.

How to set up and validate locally

N/A, spec-only change.

MR acceptance checklist

Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.

🤖 Generated with Claude Code

Merge request reports