[regression] core#1688 - false positive on missing custom field check

[MegaphoneJon]

https://lab.civicrm.org/dev/core/-/issues/1688

Overview

Any custom field that can take multiple values (multi-select, checkboxes, etc) will return a false positive for missing custom fields on the API call System.check.

Before

False positives

After

No false positives

Technical Details

The current code assumes that any saved search that has a form value starting with custom_ will end with an integer (e.g. custom_18). However, multi-select fields' form values have an operator - e.g. custom_18_operator). The check grabs 18_operator and compares it to a list of existing custom field IDs, and of course comes up empty.

Comments

I confirmed that deleting this field will still correctly trigger a warning. I gave detailed replicated steps on the Gitlab ticket.

[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).

[kcristiano]

I've done a quick r-run. I can confirm that the false positives appear on a clean 5.24.0 install. Once the PR is applied no more false positives.

[seamuslee001]

@MegaphoneJon can you put this against the 5.25 branch please and then we can look to incorporate into 5.24 branch separately

[MegaphoneJon]

@seamuslee001 #16961