A test using a deep copy of a model should reset the initial parameter values.

[dmbates]

Because the initial values for the theta parameters are derived from the values stored in the reterms they will be the final parameter estimates if those terms are deep copied to create a new model.

The NLopt optimizers don't object to this but the PRIMA optimizers do.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (6fde5ca) 96.33% compared to head (040d50d) 96.33%.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #744   +/-   ##
=======================================
  Coverage   96.33%   96.33%           
=======================================
  Files          34       34           
  Lines        3353     3353           
=======================================
  Hits         3230     3230           
  Misses        123      123           

Flag	Coverage Δ
current	96.33% <ø> (ø)
minimum	96.23% <ø> (ø)
nightly	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[dmbates]

The nightly tests are failing consistently on this PR. It appears that the download of the datasets fails when installing MixedModelsDatasets package.

[palday]

The nightly tests are failing consistently on this PR. It appears that the download of the datasets fails when installing MixedModelsDatasets package.

Yeah, I've run into this on other repositories as well and I don't know why it's happening -- I can't replicate with a local nightly build.

[dmbates]

@palday Okay to merge this PR without bumping the patch number? This only affects a test and doesn't change the result of the test - just the run time.

[palday]

@dmbates Go for merge without patch bump. 😄