Start using apiv4 in test setup

[eileenmcnaughton]

Overview

This is a tidy up PR but it strikes me we should be migrating code that does set up for tests to apiv4 & I'm wondering how that should look - do we want helpers or should it just look like the change in line 52 here. I feel like the main thing that might be missing from the helpers is that we ensured debug output was on in the helpers & somewhat improved test fail messaging

Before

After

Technical Details

@colemanw @seamuslee001 thoughts - specifically on this line https://github.com/civicrm/civicrm-core/pull/17020/files#diff-5c38127f768355a7d41757435987de99R52

Comments

[civibot]

(Standard links)

• If this is your first pull-request for CiviCRM, please browse CONTRIBUTING.md for information about the development and testing processes.
• If you are reviewing this pull-request, you may wish to consult the test sites and the Review Standards (long template, short template).

[totten]

(From MatterMost) I feel like we should be using apiv4 where we can now for everything other than testing apiv3 - but our reliance / comfort with helper functions makes that surprisingly hard

For this kind of thing - setting up a couple records - I agree that it makes more sense to start using APIv4 instead of APIv3. +1

IMHO, it's nice that civicrm_api4() doesn't seem to wrap-up exceptions in the same way that civicrm_api3() does. So I don't think we need an analog for callAPISuccess()?

Regarding test-helpers generally... it's not monolithic IMHO. There probably are some that could be replaced with API calls. But consider a helper like individualCreate() or participantCreate() which bakes-in some defaults/examples. These are not meaningful for a general audience or runtime users, so it doesn't feel right to present them as a public "API". Never-the-less, they're useful in writing tests (because it otherwise gets quite redundant to define examples).

[eileenmcnaughton]

@totten I was imagining 'fixing' individualCreate to itself call apiv4 though

[totten]

Yup, that makes sense. 👍

[colemanw]

@eileenmcnaughton - style checker is strangely picky about some things.

[colemanw]

@eileenmcnaughton I fixed the style thing and also added ->setCheckPermissions(FALSE). If we did add a wrapper around civicrm_api4 for the tests, disabling permission checks would be a sensible default to set in it.