Summary

From the conversation on https://github.com/wiredtiger/wiredtiger/pull/7384:

Can we get an official ruling from luke.chen : is it our intent that the test suite run in both diagnostic and non-diagnostic modes?

If we only run in diagnostic mode, we can simplify the test suite and remove the checks for a diagnostic build. If we want to run in both modes, then evergreen.yml should clearly run on at least one common (fast) platform in non-diagnostic mode, and we should fix this test instead of reworking the build itself.

My personal inclination would be to run in both modes, although it's more work. Diagnostic builds are significantly different from non-diagnostic builds, and I think we're better off if we handle that in our approach to testing.

I don't know what capabilities we have in evergreen.yml, but randomly choosing a diagnostic or non-diagnostic build when we run the test suite would be a nice way to go.

Motivation

Does this affect any team outside of WT?

No

How likely is it that this use case or problem will occur?

N/A

If the problem does occur, what are the consequences and how severe are they?

N/A

Is this issue urgent?

Nothing strict, but more testing is more good.

Acceptance Criteria (Definition of Done)

Make a call on whether test suites should always run with diagnostics, without, both, or randomly choose.

Testing

N/A

Documentation update

N/A